README
1
2XZ Embedded
3===========
4
5 XZ Embedded is a relatively small, limited implementation of the .xz
6 file format. Currently only decoding is implemented.
7
8 XZ Embedded was written for use in the Linux kernel, but the code can
9 be easily used in other environments too, including regular userspace
10 applications. See userspace/xzminidec.c for an example program.
11
12 NOTE: The version of XZ Embedded in the Linux kernel lacks a few
13 build-time-selectable optional features that are present in the
14 upstream XZ Embedded project: support for concatated .xz files,
15 CRC64, and ignoring unsupported check. These aren't in Linux
16 because they don't seem useful there but they would add to the
17 code size.
18
19 This README contains information that is useful only when the copy
20 of XZ Embedded isn't part of the Linux kernel tree. You should also
21 read linux/Documentation/staging/xz.rst even if you aren't using
22 XZ Embedded as part of Linux; information in that file is not
23 repeated in this README.
24
25Conformance to the .xz file format specification
26
27 As of the .xz file format specification version 1.2.0, this
28 decompressor implementation has the following limitations:
29
30 - SHA-256 isn't supported. It can be ignored as an unsupported
31 checked type if that feature is enabled at build time.
32
33 - Delta filter is not included.
34
35 - BCJ filters don't support non-default start offset.
36
37 - LZMA2 supports at most 3 GiB dictionary.
38
39 There are a couple of corner cases where things have been simplified
40 at expense of detecting errors as early as possible. These should not
41 matter in practice at all since they don't cause security issues. But
42 it is good to know this if testing the code with the test files from
43 XZ Utils.
44
45Compiler requirements
46
47 XZ Embedded should compile with any C99 or C11 compiler. The code
48 should also be GNU-C89 compatible still. GNU-C89 was used in the
49 Linux kernel until 2022. GNU-C89 support likely will be dropped
50 at some point.
51
52Embedding into userspace applications
53
54 To embed the XZ decoder, copy the following files into a single
55 directory in your source code tree:
56
57 linux/include/linux/xz.h
58 linux/lib/xz/xz_crc32.c
59 linux/lib/xz/xz_dec_lzma2.c
60 linux/lib/xz/xz_dec_stream.c
61 linux/lib/xz/xz_lzma2.h
62 linux/lib/xz/xz_private.h
63 linux/lib/xz/xz_stream.h
64 userspace/xz_config.h
65
66 Alternatively, xz.h may be placed into a different directory but then
67 that directory must be in the compiler include path when compiling
68 the .c files.
69
70 Your code should use only the functions declared in xz.h. The rest of
71 the .h files are meant only for internal use in XZ Embedded.
72
73 You may want to modify xz_config.h to be more suitable for your build
74 environment. Probably you should at least skim through it even if the
75 default file works as is.
76
77Supporting concatenated .xz files
78
79 Regular .xz files can be concatenated as is and the xz command line
80 tool will decompress all streams from a concatenated file (a few
81 other popular formats and tools support this too). This kind of .xz
82 files are more common than one might think because pxz, an early
83 threaded XZ compressor, created this kind of .xz files.
84
85 The xz_dec_run() function will stop after decompressing one stream.
86 This is good when XZ data is stored inside some other file format.
87 However, if one is decompressing regular standalone .xz files, one
88 will want to decompress all streams in the file. This is easy with
89 xz_dec_catrun(). To include support for xz_dec_catrun(), you need
90 to #define XZ_DEC_CONCATENATED in xz_config.h or in compiler flags.
91
92Integrity check support
93
94 XZ Embedded always supports the integrity check types None and
95 CRC32. Support for CRC64 is optional. SHA-256 is currently not
96 supported in XZ Embedded although the .xz format does support it.
97 The xz tool from XZ Utils uses CRC64 by default, but CRC32 is usually
98 enough in embedded systems to keep the code size smaller.
99
100 If you want support for CRC64, you need to copy linux/lib/xz/xz_crc64.c
101 into your application, and #define XZ_USE_CRC64 in xz_config.h or in
102 compiler flags.
103
104 When using the internal CRC32 or CRC64, their lookup tables need to be
105 initialized with xz_crc32_init() and xz_crc64_init(), respectively.
106 See xz.h for details.
107
108 To use external CRC32 or CRC64 code instead of the code from
109 xz_crc32.c or xz_crc64.c, the following #defines may be used
110 in xz_config.h or in compiler flags:
111
112 #define XZ_INTERNAL_CRC32 0
113 #define XZ_INTERNAL_CRC64 0
114
115 Then it is up to you to provide compatible xz_crc32() or xz_crc64()
116 functions.
117
118 If the .xz file being decompressed uses an integrity check type that
119 isn't supported by XZ Embedded, it is treated as an error and the
120 file cannot be decompressed. For multi-call mode, this can be modified
121 by #defining XZ_DEC_ANY_CHECK. Then xz_dec_run() will return
122 XZ_UNSUPPORTED_CHECK when unsupported check type is detected. After
123 that decompression can be continued normally except that the
124 integrity check won't be verified. In single-call mode there's
125 no way to continue decoding, so XZ_DEC_ANY_CHECK is almost useless
126 in single-call mode.
127
128BCJ filter support
129
130 If you want support for one or more BCJ filters, you need to copy
131 linux/lib/xz/xz_dec_bcj.c into your application, and use appropriate
132 #defines in xz_config.h or in compiler flags. You don't need these
133 #defines in the code that just uses XZ Embedded via xz.h, but having
134 them always #defined doesn't hurt either.
135
136 #define Instruction set BCJ filter endianness
137 XZ_DEC_X86 x86-32 or x86-64 Little endian only
138 XZ_DEC_POWERPC PowerPC Big endian only
139 XZ_DEC_IA64 Itanium (IA-64) Big or little endian
140 XZ_DEC_ARM ARM Little endian instructions
141 XZ_DEC_ARMTHUMB ARM-Thumb Big or little endian
142 XZ_DEC_ARM64 ARM64 Big or little endian
143 XZ_DEC_SPARC SPARC Big or little endian
144 XZ_DEC_RISCV RISC-V Big or little endian
145
146 While some architectures are (partially) bi-endian, the endianness
147 setting doesn't change the endianness of the instructions on all
148 architectures. That's why many filters work for both big and little
149 endian executables (Itanium and ARM based architectures have little
150 endian instructions and SPARC has big endian instructions).
151
152Notes about shared libraries
153
154 If you are including XZ Embedded into a shared library, you should
155 rename the xz_* functions to prevent symbol conflicts in case your
156 library is linked against some other library or application that
157 also has XZ Embedded in it (which may even be a different version
158 of XZ Embedded).
159
160 Please don't create a shared library of XZ Embedded itself unless
161 it is fine to rebuild everything depending on that shared library
162 every time you upgrade to a newer version of XZ Embedded. There are
163 no API or ABI stability guarantees between different versions of
164 XZ Embedded.
165
166Contact information
167
168 Email: Lasse Collin <[email protected]>
169 IRC: Larhzu on #tukaani on Libera Chat
170 GitHub: https://github.com/tukaani-project/xz-embedded
171
172