xref: /linux/Documentation/filesystems/erofs.rst (revision c532de5a67a70f8533d495f8f2aaa9a0491c3ad0)
1.. SPDX-License-Identifier: GPL-2.0
2
3======================================
4EROFS - Enhanced Read-Only File System
5======================================
6
7Overview
8========
9
10EROFS filesystem stands for Enhanced Read-Only File System.  It aims to form a
11generic read-only filesystem solution for various read-only use cases instead
12of just focusing on storage space saving without considering any side effects
13of runtime performance.
14
15It is designed to meet the needs of flexibility, feature extendability and user
16payload friendly, etc.  Apart from those, it is still kept as a simple
17random-access friendly high-performance filesystem to get rid of unneeded I/O
18amplification and memory-resident overhead compared to similar approaches.
19
20It is implemented to be a better choice for the following scenarios:
21
22 - read-only storage media or
23
24 - part of a fully trusted read-only solution, which means it needs to be
25   immutable and bit-for-bit identical to the official golden image for
26   their releases due to security or other considerations and
27
28 - hope to minimize extra storage space with guaranteed end-to-end performance
29   by using compact layout, transparent file compression and direct access,
30   especially for those embedded devices with limited memory and high-density
31   hosts with numerous containers.
32
33Here are the main features of EROFS:
34
35 - Little endian on-disk design;
36
37 - Block-based distribution and file-based distribution over fscache are
38   supported;
39
40 - Support multiple devices to refer to external blobs, which can be used
41   for container images;
42
43 - 32-bit block addresses for each device, therefore 16TiB address space at
44   most with 4KiB block size for now;
45
46 - Two inode layouts for different requirements:
47
48   =====================  ============  ======================================
49                          compact (v1)  extended (v2)
50   =====================  ============  ======================================
51   Inode metadata size    32 bytes      64 bytes
52   Max file size          4 GiB         16 EiB (also limited by max. vol size)
53   Max uids/gids          65536         4294967296
54   Per-inode timestamp    no            yes (64 + 32-bit timestamp)
55   Max hardlinks          65536         4294967296
56   Metadata reserved      8 bytes       18 bytes
57   =====================  ============  ======================================
58
59 - Support extended attributes as an option;
60
61 - Support a bloom filter that speeds up negative extended attribute lookups;
62
63 - Support POSIX.1e ACLs by using extended attributes;
64
65 - Support transparent data compression as an option:
66   LZ4, MicroLZMA and DEFLATE algorithms can be used on a per-file basis; In
67   addition, inplace decompression is also supported to avoid bounce compressed
68   buffers and unnecessary page cache thrashing.
69
70 - Support chunk-based data deduplication and rolling-hash compressed data
71   deduplication;
72
73 - Support tailpacking inline compared to byte-addressed unaligned metadata
74   or smaller block size alternatives;
75
76 - Support merging tail-end data into a special inode as fragments.
77
78 - Support large folios to make use of THPs (Transparent Hugepages);
79
80 - Support direct I/O on uncompressed files to avoid double caching for loop
81   devices;
82
83 - Support FSDAX on uncompressed images for secure containers and ramdisks in
84   order to get rid of unnecessary page cache.
85
86 - Support file-based on-demand loading with the Fscache infrastructure.
87
88The following git tree provides the file system user-space tools under
89development, such as a formatting tool (mkfs.erofs), an on-disk consistency &
90compatibility checking tool (fsck.erofs), and a debugging tool (dump.erofs):
91
92- git://git.kernel.org/pub/scm/linux/kernel/git/xiang/erofs-utils.git
93
94For more information, please also refer to the documentation site:
95
96- https://erofs.docs.kernel.org
97
98Bugs and patches are welcome, please kindly help us and send to the following
99linux-erofs mailing list:
100
101- linux-erofs mailing list   <linux-erofs@lists.ozlabs.org>
102
103Mount options
104=============
105
106===================    =========================================================
107(no)user_xattr         Setup Extended User Attributes. Note: xattr is enabled
108                       by default if CONFIG_EROFS_FS_XATTR is selected.
109(no)acl                Setup POSIX Access Control List. Note: acl is enabled
110                       by default if CONFIG_EROFS_FS_POSIX_ACL is selected.
111cache_strategy=%s      Select a strategy for cached decompression from now on:
112
113		       ==========  =============================================
114                         disabled  In-place I/O decompression only;
115                        readahead  Cache the last incomplete compressed physical
116                                   cluster for further reading. It still does
117                                   in-place I/O decompression for the rest
118                                   compressed physical clusters;
119                       readaround  Cache the both ends of incomplete compressed
120                                   physical clusters for further reading.
121                                   It still does in-place I/O decompression
122                                   for the rest compressed physical clusters.
123		       ==========  =============================================
124dax={always,never}     Use direct access (no page cache).  See
125                       Documentation/filesystems/dax.rst.
126dax                    A legacy option which is an alias for ``dax=always``.
127device=%s              Specify a path to an extra device to be used together.
128fsid=%s                Specify a filesystem image ID for Fscache back-end.
129domain_id=%s           Specify a domain ID in fscache mode so that different images
130                       with the same blobs under a given domain ID can share storage.
131===================    =========================================================
132
133Sysfs Entries
134=============
135
136Information about mounted erofs file systems can be found in /sys/fs/erofs.
137Each mounted filesystem will have a directory in /sys/fs/erofs based on its
138device name (i.e., /sys/fs/erofs/sda).
139(see also Documentation/ABI/testing/sysfs-fs-erofs)
140
141On-disk details
142===============
143
144Summary
145-------
146Different from other read-only file systems, an EROFS volume is designed
147to be as simple as possible::
148
149                                |-> aligned with the block size
150   ____________________________________________________________
151  | |SB| | ... | Metadata | ... | Data | Metadata | ... | Data |
152  |_|__|_|_____|__________|_____|______|__________|_____|______|
153  0 +1K
154
155All data areas should be aligned with the block size, but metadata areas
156may not. All metadatas can be now observed in two different spaces (views):
157
158 1. Inode metadata space
159
160    Each valid inode should be aligned with an inode slot, which is a fixed
161    value (32 bytes) and designed to be kept in line with compact inode size.
162
163    Each inode can be directly found with the following formula:
164         inode offset = meta_blkaddr * block_size + 32 * nid
165
166    ::
167
168                                 |-> aligned with 8B
169                                            |-> followed closely
170     + meta_blkaddr blocks                                      |-> another slot
171       _____________________________________________________________________
172     |  ...   | inode |  xattrs  | extents  | data inline | ... | inode ...
173     |________|_______|(optional)|(optional)|__(optional)_|_____|__________
174              |-> aligned with the inode slot size
175                   .                   .
176                 .                         .
177               .                              .
178             .                                    .
179           .                                         .
180         .                                              .
181       .____________________________________________________|-> aligned with 4B
182       | xattr_ibody_header | shared xattrs | inline xattrs |
183       |____________________|_______________|_______________|
184       |->    12 bytes    <-|->x * 4 bytes<-|               .
185                           .                .                 .
186                     .                      .                   .
187                .                           .                     .
188            ._______________________________.______________________.
189            | id | id | id | id |  ... | id | ent | ... | ent| ... |
190            |____|____|____|____|______|____|_____|_____|____|_____|
191                                            |-> aligned with 4B
192                                                        |-> aligned with 4B
193
194    Inode could be 32 or 64 bytes, which can be distinguished from a common
195    field which all inode versions have -- i_format::
196
197        __________________               __________________
198       |     i_format     |             |     i_format     |
199       |__________________|             |__________________|
200       |        ...       |             |        ...       |
201       |                  |             |                  |
202       |__________________| 32 bytes    |                  |
203                                        |                  |
204                                        |__________________| 64 bytes
205
206    Xattrs, extents, data inline are placed after the corresponding inode with
207    proper alignment, and they could be optional for different data mappings.
208    _currently_ total 5 data layouts are supported:
209
210    ==  ====================================================================
211     0  flat file data without data inline (no extent);
212     1  fixed-sized output data compression (with non-compacted indexes);
213     2  flat file data with tail packing data inline (no extent);
214     3  fixed-sized output data compression (with compacted indexes, v5.3+);
215     4  chunk-based file (v5.15+).
216    ==  ====================================================================
217
218    The size of the optional xattrs is indicated by i_xattr_count in inode
219    header. Large xattrs or xattrs shared by many different files can be
220    stored in shared xattrs metadata rather than inlined right after inode.
221
222 2. Shared xattrs metadata space
223
224    Shared xattrs space is similar to the above inode space, started with
225    a specific block indicated by xattr_blkaddr, organized one by one with
226    proper align.
227
228    Each share xattr can also be directly found by the following formula:
229         xattr offset = xattr_blkaddr * block_size + 4 * xattr_id
230
231::
232
233                           |-> aligned by  4 bytes
234    + xattr_blkaddr blocks                     |-> aligned with 4 bytes
235     _________________________________________________________________________
236    |  ...   | xattr_entry |  xattr data | ... |  xattr_entry | xattr data  ...
237    |________|_____________|_____________|_____|______________|_______________
238
239Directories
240-----------
241All directories are now organized in a compact on-disk format. Note that
242each directory block is divided into index and name areas in order to support
243random file lookup, and all directory entries are _strictly_ recorded in
244alphabetical order in order to support improved prefix binary search
245algorithm (could refer to the related source code).
246
247::
248
249                  ___________________________
250                 /                           |
251                /              ______________|________________
252               /              /              | nameoff1       | nameoffN-1
253  ____________.______________._______________v________________v__________
254 | dirent | dirent | ... | dirent | filename | filename | ... | filename |
255 |___.0___|____1___|_____|___N-1__|____0_____|____1_____|_____|___N-1____|
256      \                           ^
257       \                          |                           * could have
258        \                         |                             trailing '\0'
259         \________________________| nameoff0
260                             Directory block
261
262Note that apart from the offset of the first filename, nameoff0 also indicates
263the total number of directory entries in this block since it is no need to
264introduce another on-disk field at all.
265
266Chunk-based files
267-----------------
268In order to support chunk-based data deduplication, a new inode data layout has
269been supported since Linux v5.15: Files are split in equal-sized data chunks
270with ``extents`` area of the inode metadata indicating how to get the chunk
271data: these can be simply as a 4-byte block address array or in the 8-byte
272chunk index form (see struct erofs_inode_chunk_index in erofs_fs.h for more
273details.)
274
275By the way, chunk-based files are all uncompressed for now.
276
277Long extended attribute name prefixes
278-------------------------------------
279There are use cases where extended attributes with different values can have
280only a few common prefixes (such as overlayfs xattrs).  The predefined prefixes
281work inefficiently in both image size and runtime performance in such cases.
282
283The long xattr name prefixes feature is introduced to address this issue.  The
284overall idea is that, apart from the existing predefined prefixes, the xattr
285entry could also refer to user-specified long xattr name prefixes, e.g.
286"trusted.overlay.".
287
288When referring to a long xattr name prefix, the highest bit (bit 7) of
289erofs_xattr_entry.e_name_index is set, while the lower bits (bit 0-6) as a whole
290represent the index of the referred long name prefix among all long name
291prefixes.  Therefore, only the trailing part of the name apart from the long
292xattr name prefix is stored in erofs_xattr_entry.e_name, which could be empty if
293the full xattr name matches exactly as its long xattr name prefix.
294
295All long xattr prefixes are stored one by one in the packed inode as long as
296the packed inode is valid, or in the meta inode otherwise.  The
297xattr_prefix_count (of the on-disk superblock) indicates the total number of
298long xattr name prefixes, while (xattr_prefix_start * 4) indicates the start
299offset of long name prefixes in the packed/meta inode.  Note that, long extended
300attribute name prefixes are disabled if xattr_prefix_count is 0.
301
302Each long name prefix is stored in the format: ALIGN({__le16 len, data}, 4),
303where len represents the total size of the data part.  The data part is actually
304represented by 'struct erofs_xattr_long_prefix', where base_index represents the
305index of the predefined xattr name prefix, e.g. EROFS_XATTR_INDEX_TRUSTED for
306"trusted.overlay." long name prefix, while the infix string keeps the string
307after stripping the short prefix, e.g. "overlay." for the example above.
308
309Data compression
310----------------
311EROFS implements fixed-sized output compression which generates fixed-sized
312compressed data blocks from variable-sized input in contrast to other existing
313fixed-sized input solutions. Relatively higher compression ratios can be gotten
314by using fixed-sized output compression since nowadays popular data compression
315algorithms are mostly LZ77-based and such fixed-sized output approach can be
316benefited from the historical dictionary (aka. sliding window).
317
318In details, original (uncompressed) data is turned into several variable-sized
319extents and in the meanwhile, compressed into physical clusters (pclusters).
320In order to record each variable-sized extent, logical clusters (lclusters) are
321introduced as the basic unit of compress indexes to indicate whether a new
322extent is generated within the range (HEAD) or not (NONHEAD). Lclusters are now
323fixed in block size, as illustrated below::
324
325          |<-    variable-sized extent    ->|<-       VLE         ->|
326        clusterofs                        clusterofs              clusterofs
327          |                                 |                       |
328 _________v_________________________________v_______________________v________
329 ... |    .         |              |        .     |              |  .   ...
330 ____|____._________|______________|________.___ _|______________|__.________
331     |-> lcluster <-|-> lcluster <-|-> lcluster <-|-> lcluster <-|
332          (HEAD)        (NONHEAD)       (HEAD)        (NONHEAD)    .
333           .             CBLKCNT            .                    .
334            .                               .                  .
335             .                              .                .
336       _______._____________________________.______________._________________
337          ... |              |              |              | ...
338       _______|______________|______________|______________|_________________
339              |->      big pcluster       <-|-> pcluster <-|
340
341A physical cluster can be seen as a container of physical compressed blocks
342which contains compressed data. Previously, only lcluster-sized (4KB) pclusters
343were supported. After big pcluster feature is introduced (available since
344Linux v5.13), pcluster can be a multiple of lcluster size.
345
346For each HEAD lcluster, clusterofs is recorded to indicate where a new extent
347starts and blkaddr is used to seek the compressed data. For each NONHEAD
348lcluster, delta0 and delta1 are available instead of blkaddr to indicate the
349distance to its HEAD lcluster and the next HEAD lcluster. A PLAIN lcluster is
350also a HEAD lcluster except that its data is uncompressed. See the comments
351around "struct z_erofs_vle_decompressed_index" in erofs_fs.h for more details.
352
353If big pcluster is enabled, pcluster size in lclusters needs to be recorded as
354well. Let the delta0 of the first NONHEAD lcluster store the compressed block
355count with a special flag as a new called CBLKCNT NONHEAD lcluster. It's easy
356to understand its delta0 is constantly 1, as illustrated below::
357
358   __________________________________________________________
359  | HEAD |  NONHEAD  | NONHEAD | ... | NONHEAD | HEAD | HEAD |
360  |__:___|_(CBLKCNT)_|_________|_____|_________|__:___|____:_|
361     |<----- a big pcluster (with CBLKCNT) ------>|<--  -->|
362           a lcluster-sized pcluster (without CBLKCNT) ^
363
364If another HEAD follows a HEAD lcluster, there is no room to record CBLKCNT,
365but it's easy to know the size of such pcluster is 1 lcluster as well.
366
367Since Linux v6.1, each pcluster can be used for multiple variable-sized extents,
368therefore it can be used for compressed data deduplication.
369