xref: /linux/Documentation/filesystems/erofs.rst (revision 805185b7c7a1069e407b6f7b3bc98e44d415f484)
1.. SPDX-License-Identifier: GPL-2.0
2
3======================================
4EROFS - Enhanced Read-Only File System
5======================================
6
7Overview
8========
9
10EROFS (Enhanced Read-Only File System) is a modern, efficient, and secure
11read-only kernel filesystem designed for various use cases including immutable
12system images, container images, application sandbox images, and dataset
13distribution.
14
15An immutable image filesystem can be regarded as an enhanced archive format
16which allows golden images to be built once and mounted everywhere -- images are
17bit-for-bit identical across all deployments and can be verified, audited, or
18shared without concerns about runtime modifications (in this model, all user
19writes should be redirected into another trusted filesystem, for example, via
20overlayfs for copy-on-write-style redirection, by design).
21
22EROFS is a dedicated implementation of the image filesystem idea above, with a
23flexible, hierarchical on-disk design so that needed features can be enabled on
24demand. Filesystem data in the core format is strictly block-aligned in order
25to perform optimally on all kinds of storage media, including block devices and
26memory-backed devices. The on-disk format is easy to parse and purposely avoids
27the unnecessary metadata redundancy found in generic writable filesystems, which
28can suffer from extra inconsistency issues -- making it ideal for security
29auditing and untrusted remote access. In addition, designs such as inline data,
30inline/shared extended attributes, and optimized (de)compression provide better
31space efficiency while maintaining high performance.
32
33In short, EROFS aims to be a better fit for the following scenarios:
34
35 - As part of a secure immutable storage solution, where it needs to be
36   immutable and bit-for-bit identical to the official golden image for
37   each individual copy, in order to meet security, data sharing, and/or
38   other requirements;
39
40 - Minimizing storage overhead with guaranteed end-to-end performance
41   by using compact (meta)data layout, optimized transparent data compression,
42   deduplication and direct access, especially for those embedded devices with
43   limited memory and high-density hosts with numerous containers.
44
45Here is the list of highlights:
46
47 - Little endian on-disk design with 48-bit block addressing, supporting up
48   to 1 EiB filesystem capacity with 4 KiB block size;
49
50 - Two compact inode metadata layouts for space and performance efficiency:
51
52   ========================  ========  ======================================
53                             compact   extended
54   ========================  ========  ======================================
55   Inode core metadata size  32 bytes  64 bytes
56   Max file size             4 GiB     16 EiB (also limited by max. vol size)
57   Max uids/gids             65536     4294967296
58   Nanosecond timestamps     no        yes
59   Max hardlinks             65536     4294967296
60   ========================  ========  ======================================
61
62 - Support tailpacking inline data for better space efficiency and reduce
63   unneeded I/O amplification;
64
65 - Block-based and file-backed distribution are both supported;
66
67 - Multiple devices to reference external data blobs: inode data can be
68   optionally placed into external blobs, which enables image layering and data
69   sharing among different filesystems;
70
71 - Inline and shared extended attributes with an optional bloom filter that
72   speeds up negative extended attribute lookups;
73
74 - POSIX.1e ACLs by using extended attributes;
75
76 - Transparent data compression as an option: Supported algorithms (LZ4,
77   MicroLZMA, DEFLATE and Zstandard) can be selected on a per-inode basis.
78   Both the on-disk metadata and decompression runtime have been heavily
79   optimized to minimize the overhead for better performance.
80
81 - Merging tail-end data into a special inode as fragments;
82
83 - Chunk-based deduplication and rolling-hash compressed data deduplication;
84
85 - Direct I/O and FSDAX support on uncompressed inodes for use cases such as
86   secure containers, loop devices, and ramdisks that do not need page caching;
87
88 - Page cache sharing among inodes with identical content fingerprints on
89   the same machine.
90
91For more detailed information, please refer to our documentation site:
92
93- https://erofs.docs.kernel.org
94
95The following git tree provides the file system user-space tools under
96development, such as a formatting tool (mkfs.erofs), an on-disk consistency &
97compatibility checking tool (fsck.erofs), and a debugging tool (dump.erofs):
98
99- git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs-utils.git
100
101Bugs and patches are welcome, please kindly help us and send to the following
102linux-erofs mailing list:
103
104- linux-erofs mailing list   <linux-erofs@lists.ozlabs.org>
105
106Mount options
107=============
108
109===================    =========================================================
110(no)user_xattr         Setup Extended User Attributes. Note: xattr is enabled
111                       by default if CONFIG_EROFS_FS_XATTR is selected.
112(no)acl                Setup POSIX Access Control List. Note: acl is enabled
113                       by default if CONFIG_EROFS_FS_POSIX_ACL is selected.
114cache_strategy=%s      Select a strategy for cached decompression from now on:
115
116		       ==========  =============================================
117                         disabled  In-place I/O decompression only;
118                        readahead  Cache the last incomplete compressed physical
119                                   cluster for further reading. It still does
120                                   in-place I/O decompression for the rest
121                                   compressed physical clusters;
122                       readaround  Cache both ends of incomplete compressed
123                                   physical clusters for further reading.
124                                   It still does in-place I/O decompression
125                                   for the rest compressed physical clusters.
126		       ==========  =============================================
127dax={always,never}     Use direct access (no page cache).  See
128                       Documentation/filesystems/dax.rst.
129dax                    A legacy option which is an alias for ``dax=always``.
130device=%s              Specify a path to an extra device to be used together.
131directio               (For file-backed mounts) Use direct I/O to access backing
132                       files, and asynchronous I/O will be enabled if supported.
133domain_id=%s           Specify a trusted domain ID. Filesystems sharing the same
134                       domain ID can share page cache across mounts when inode
135                       page sharing is enabled. (not shown in mountinfo output)
136fsoffset=%llu          Specify block-aligned filesystem offset for the primary device.
137inode_share            Enable inode page sharing for this filesystem.  Inodes with
138                       identical content within the same domain ID can share the
139                       page cache.
140===================    =========================================================
141
142Sysfs Entries
143=============
144
145Information about mounted erofs file systems can be found in /sys/fs/erofs.
146Each mounted filesystem will have a directory in /sys/fs/erofs based on its
147device name (i.e., /sys/fs/erofs/sda).
148(see also Documentation/ABI/testing/sysfs-fs-erofs)
149
150On-disk details
151===============
152
153Summary
154-------
155Different from other read-only file systems, an EROFS volume is designed
156to be as simple as possible::
157
158                                |-> aligned with the block size
159   ____________________________________________________________
160  | |SB| | ... | Metadata | ... | Data | Metadata | ... | Data |
161  |_|__|_|_____|__________|_____|______|__________|_____|______|
162  0 +1K
163
164All data areas should be aligned with the block size, but metadata areas
165may not. All metadata can be now observed in two different spaces (views):
166
167 1. Inode metadata space
168
169    Each valid inode should be aligned with an inode slot, which is a fixed
170    value (32 bytes) and designed to be kept in line with compact inode size.
171
172    Each inode can be directly found with the following formula:
173         inode offset = meta_blkaddr * block_size + 32 * nid
174
175    ::
176
177                                 |-> aligned with 8B
178                                            |-> followed closely
179     + meta_blkaddr blocks                                      |-> another slot
180       _____________________________________________________________________
181     |  ...   | inode |  xattrs  | extents  | data inline | ... | inode ...
182     |________|_______|(optional)|(optional)|__(optional)_|_____|__________
183              |-> aligned with the inode slot size
184                   .                   .
185                 .                         .
186               .                              .
187             .                                    .
188           .                                         .
189         .                                              .
190       .____________________________________________________|-> aligned with 4B
191       | xattr_ibody_header | shared xattrs | inline xattrs |
192       |____________________|_______________|_______________|
193       |->    12 bytes    <-|->x * 4 bytes<-|               .
194                           .                .                 .
195                     .                      .                   .
196                .                           .                     .
197            ._______________________________.______________________.
198            | id | id | id | id |  ... | id | ent | ... | ent| ... |
199            |____|____|____|____|______|____|_____|_____|____|_____|
200                                            |-> aligned with 4B
201                                                        |-> aligned with 4B
202
203    Inode could be 32 or 64 bytes, which can be distinguished from a common
204    field which all inode versions have -- i_format::
205
206        __________________               __________________
207       |     i_format     |             |     i_format     |
208       |__________________|             |__________________|
209       |        ...       |             |        ...       |
210       |                  |             |                  |
211       |__________________| 32 bytes    |                  |
212                                        |                  |
213                                        |__________________| 64 bytes
214
215    Xattrs, extents, data inline are placed after the corresponding inode with
216    proper alignment, and they could be optional for different data mappings.
217    _currently_ total 5 data layouts are supported:
218
219    ==  ====================================================================
220     0  flat file data without data inline (no extent);
221     1  fixed-sized output data compression (with non-compacted indexes);
222     2  flat file data with tail packing data inline (no extent);
223     3  fixed-sized output data compression (with compacted indexes, v5.3+);
224     4  chunk-based file (v5.15+).
225    ==  ====================================================================
226
227    The size of the optional xattrs is indicated by i_xattr_count in inode
228    header. Large xattrs or xattrs shared by many different files can be
229    stored in shared xattrs metadata rather than inlined right after inode.
230
231 2. Shared xattrs metadata space
232
233    Shared xattrs space is similar to the above inode space, started with
234    a specific block indicated by xattr_blkaddr, organized one by one with
235    proper align.
236
237    Each share xattr can also be directly found by the following formula:
238         xattr offset = xattr_blkaddr * block_size + 4 * xattr_id
239
240::
241
242                           |-> aligned by  4 bytes
243    + xattr_blkaddr blocks                     |-> aligned with 4 bytes
244     _________________________________________________________________________
245    |  ...   | xattr_entry |  xattr data | ... |  xattr_entry | xattr data  ...
246    |________|_____________|_____________|_____|______________|_______________
247
248Directories
249-----------
250All directories are now organized in a compact on-disk format. Note that
251each directory block is divided into index and name areas in order to support
252random file lookup, and all directory entries are _strictly_ recorded in
253alphabetical order in order to support improved prefix binary search
254algorithm (could refer to the related source code).
255
256::
257
258                  ___________________________
259                 /                           |
260                /              ______________|________________
261               /              /              | nameoff1       | nameoffN-1
262  ____________.______________._______________v________________v__________
263 | dirent | dirent | ... | dirent | filename | filename | ... | filename |
264 |___.0___|____1___|_____|___N-1__|____0_____|____1_____|_____|___N-1____|
265      \                           ^
266       \                          |                           * could have
267        \                         |                             trailing '\0'
268         \________________________| nameoff0
269                             Directory block
270
271Note that apart from the offset of the first filename, nameoff0 also indicates
272the total number of directory entries in this block since it is no need to
273introduce another on-disk field at all.
274
275Chunk-based files
276-----------------
277In order to support chunk-based data deduplication, a new inode data layout has
278been supported since Linux v5.15: Files are split in equal-sized data chunks
279with ``extents`` area of the inode metadata indicating how to get the chunk
280data: these can be simply as a 4-byte block address array or in the 8-byte
281chunk index form (see struct erofs_inode_chunk_index in erofs_fs.h for more
282details.)
283
284By the way, chunk-based files are all uncompressed for now.
285
286Long extended attribute name prefixes
287-------------------------------------
288There are use cases where extended attributes with different values can have
289only a few common prefixes (such as overlayfs xattrs).  The predefined prefixes
290work inefficiently in both image size and runtime performance in such cases.
291
292The long xattr name prefixes feature is introduced to address this issue.  The
293overall idea is that, apart from the existing predefined prefixes, the xattr
294entry could also refer to user-specified long xattr name prefixes, e.g.
295"trusted.overlay.".
296
297When referring to a long xattr name prefix, the highest bit (bit 7) of
298erofs_xattr_entry.e_name_index is set, while the lower bits (bit 0-6) as a whole
299represent the index of the referred long name prefix among all long name
300prefixes.  Therefore, only the trailing part of the name apart from the long
301xattr name prefix is stored in erofs_xattr_entry.e_name, which could be empty if
302the full xattr name matches exactly as its long xattr name prefix.
303
304All long xattr prefixes are stored one by one in the packed inode as long as
305the packed inode is valid, or in the meta inode otherwise.  The
306xattr_prefix_count (of the on-disk superblock) indicates the total number of
307long xattr name prefixes, while (xattr_prefix_start * 4) indicates the start
308offset of long name prefixes in the packed/meta inode.  Note that, long extended
309attribute name prefixes are disabled if xattr_prefix_count is 0.
310
311Each long name prefix is stored in the format: ALIGN({__le16 len, data}, 4),
312where len represents the total size of the data part.  The data part is actually
313represented by 'struct erofs_xattr_long_prefix', where base_index represents the
314index of the predefined xattr name prefix, e.g. EROFS_XATTR_INDEX_TRUSTED for
315"trusted.overlay." long name prefix, while the infix string keeps the string
316after stripping the short prefix, e.g. "overlay." for the example above.
317
318Data compression
319----------------
320EROFS implements fixed-sized output compression which generates fixed-sized
321compressed data blocks from variable-sized input in contrast to other existing
322fixed-sized input solutions. Relatively higher compression ratios can be gotten
323by using fixed-sized output compression since nowadays popular data compression
324algorithms are mostly LZ77-based and such fixed-sized output approach can be
325benefited from the historical dictionary (aka. sliding window).
326
327In details, original (uncompressed) data is turned into several variable-sized
328extents and in the meanwhile, compressed into physical clusters (pclusters).
329In order to record each variable-sized extent, logical clusters (lclusters) are
330introduced as the basic unit of compress indexes to indicate whether a new
331extent is generated within the range (HEAD) or not (NONHEAD). Lclusters are now
332fixed in block size, as illustrated below::
333
334          |<-    variable-sized extent    ->|<-       VLE         ->|
335        clusterofs                        clusterofs              clusterofs
336          |                                 |                       |
337 _________v_________________________________v_______________________v________
338 ... |    .         |              |        .     |              |  .   ...
339 ____|____._________|______________|________.___ _|______________|__.________
340     |-> lcluster <-|-> lcluster <-|-> lcluster <-|-> lcluster <-|
341          (HEAD)        (NONHEAD)       (HEAD)        (NONHEAD)    .
342           .             CBLKCNT            .                    .
343            .                               .                  .
344             .                              .                .
345       _______._____________________________.______________._________________
346          ... |              |              |              | ...
347       _______|______________|______________|______________|_________________
348              |->      big pcluster       <-|-> pcluster <-|
349
350A physical cluster can be seen as a container of physical compressed blocks
351which contains compressed data. Previously, only lcluster-sized (4KB) pclusters
352were supported. After big pcluster feature is introduced (available since
353Linux v5.13), pcluster can be a multiple of lcluster size.
354
355For each HEAD lcluster, clusterofs is recorded to indicate where a new extent
356starts and blkaddr is used to seek the compressed data. For each NONHEAD
357lcluster, delta0 and delta1 are available instead of blkaddr to indicate the
358distance to its HEAD lcluster and the next HEAD lcluster. A PLAIN lcluster is
359also a HEAD lcluster except that its data is uncompressed. See the comments
360around "struct z_erofs_vle_decompressed_index" in erofs_fs.h for more details.
361
362If big pcluster is enabled, pcluster size in lclusters needs to be recorded as
363well. Let the delta0 of the first NONHEAD lcluster store the compressed block
364count with a special flag as a new called CBLKCNT NONHEAD lcluster. It's easy
365to understand its delta0 is constantly 1, as illustrated below::
366
367   __________________________________________________________
368  | HEAD |  NONHEAD  | NONHEAD | ... | NONHEAD | HEAD | HEAD |
369  |__:___|_(CBLKCNT)_|_________|_____|_________|__:___|____:_|
370     |<----- a big pcluster (with CBLKCNT) ------>|<--  -->|
371           a lcluster-sized pcluster (without CBLKCNT) ^
372
373If another HEAD follows a HEAD lcluster, there is no room to record CBLKCNT,
374but it's easy to know the size of such pcluster is 1 lcluster as well.
375
376Since Linux v6.1, each pcluster can be used for multiple variable-sized extents,
377therefore it can be used for compressed data deduplication.
378