1/*
2 * XZ decompressor
3 *
4 * Authors: Lasse Collin <lasse.collin@tukaani.org>
5 *          Igor Pavlov <http://7-zip.org/>
6 *
7 * This file has been put into the public domain.
8 * You can do whatever you want with this file.
9 */
10
11#ifndef XZ_H
12#define XZ_H
13
14#ifdef __KERNEL__
15#	include <linux/stddef.h>
16#	include <linux/types.h>
17#else
18#	include <stddef.h>
19#	include <stdint.h>
20#endif
21
22#ifdef __cplusplus
23extern "C" {
24#endif
25
26/* In Linux, this is used to make extern functions static when needed. */
27#ifndef XZ_EXTERN
28#	define XZ_EXTERN extern
29#endif
30
31/**
32 * enum xz_mode - Operation mode
33 *
34 * @XZ_SINGLE:              Single-call mode. This uses less RAM than
35 *                          than multi-call modes, because the LZMA2
36 *                          dictionary doesn't need to be allocated as
37 *                          part of the decoder state. All required data
38 *                          structures are allocated at initialization,
39 *                          so xz_dec_run() cannot return XZ_MEM_ERROR.
40 * @XZ_PREALLOC:            Multi-call mode with preallocated LZMA2
41 *                          dictionary buffer. All data structures are
42 *                          allocated at initialization, so xz_dec_run()
43 *                          cannot return XZ_MEM_ERROR.
44 * @XZ_DYNALLOC:            Multi-call mode. The LZMA2 dictionary is
45 *                          allocated once the required size has been
46 *                          parsed from the stream headers. If the
47 *                          allocation fails, xz_dec_run() will return
48 *                          XZ_MEM_ERROR.
49 *
50 * It is possible to enable support only for a subset of the above
51 * modes at compile time by defining XZ_DEC_SINGLE, XZ_DEC_PREALLOC,
52 * or XZ_DEC_DYNALLOC. The xz_dec kernel module is always compiled
53 * with support for all operation modes, but the preboot code may
54 * be built with fewer features to minimize code size.
55 */
56enum xz_mode {
57	XZ_SINGLE,
58	XZ_PREALLOC,
59	XZ_DYNALLOC
60};
61
62/**
63 * enum xz_ret - Return codes
64 * @XZ_OK:                  Everything is OK so far. More input or more
65 *                          output space is required to continue. This
66 *                          return code is possible only in multi-call mode
67 *                          (XZ_PREALLOC or XZ_DYNALLOC).
68 * @XZ_STREAM_END:          Operation finished successfully.
69 * @XZ_UNSUPPORTED_CHECK:   Integrity check type is not supported. Decoding
70 *                          is still possible in multi-call mode by simply
71 *                          calling xz_dec_run() again.
72 *                          Note that this return value is used only if
73 *                          XZ_DEC_ANY_CHECK was defined at build time,
74 *                          which is not used in the kernel. Unsupported
75 *                          check types return XZ_OPTIONS_ERROR if
76 *                          XZ_DEC_ANY_CHECK was not defined at build time.
77 * @XZ_MEM_ERROR:           Allocating memory failed. This return code is
78 *                          possible only if the decoder was initialized
79 *                          with XZ_DYNALLOC. The amount of memory that was
80 *                          tried to be allocated was no more than the
81 *                          dict_max argument given to xz_dec_init().
82 * @XZ_MEMLIMIT_ERROR:      A bigger LZMA2 dictionary would be needed than
83 *                          allowed by the dict_max argument given to
84 *                          xz_dec_init(). This return value is possible
85 *                          only in multi-call mode (XZ_PREALLOC or
86 *                          XZ_DYNALLOC); the single-call mode (XZ_SINGLE)
87 *                          ignores the dict_max argument.
88 * @XZ_FORMAT_ERROR:        File format was not recognized (wrong magic
89 *                          bytes).
90 * @XZ_OPTIONS_ERROR:       This implementation doesn't support the requested
91 *                          compression options. In the decoder this means
92 *                          that the header CRC32 matches, but the header
93 *                          itself specifies something that we don't support.
94 * @XZ_DATA_ERROR:          Compressed data is corrupt.
95 * @XZ_BUF_ERROR:           Cannot make any progress. Details are slightly
96 *                          different between multi-call and single-call
97 *                          mode; more information below.
98 *
99 * In multi-call mode, XZ_BUF_ERROR is returned when two consecutive calls
100 * to XZ code cannot consume any input and cannot produce any new output.
101 * This happens when there is no new input available, or the output buffer
102 * is full while at least one output byte is still pending. Assuming your
103 * code is not buggy, you can get this error only when decoding a compressed
104 * stream that is truncated or otherwise corrupt.
105 *
106 * In single-call mode, XZ_BUF_ERROR is returned only when the output buffer
107 * is too small or the compressed input is corrupt in a way that makes the
108 * decoder produce more output than the caller expected. When it is
109 * (relatively) clear that the compressed input is truncated, XZ_DATA_ERROR
110 * is used instead of XZ_BUF_ERROR.
111 */
112enum xz_ret {
113	XZ_OK,
114	XZ_STREAM_END,
115	XZ_UNSUPPORTED_CHECK,
116	XZ_MEM_ERROR,
117	XZ_MEMLIMIT_ERROR,
118	XZ_FORMAT_ERROR,
119	XZ_OPTIONS_ERROR,
120	XZ_DATA_ERROR,
121	XZ_BUF_ERROR
122};
123
124/**
125 * struct xz_buf - Passing input and output buffers to XZ code
126 * @in:         Beginning of the input buffer. This may be NULL if and only
127 *              if in_pos is equal to in_size.
128 * @in_pos:     Current position in the input buffer. This must not exceed
129 *              in_size.
130 * @in_size:    Size of the input buffer
131 * @out:        Beginning of the output buffer. This may be NULL if and only
132 *              if out_pos is equal to out_size.
133 * @out_pos:    Current position in the output buffer. This must not exceed
134 *              out_size.
135 * @out_size:   Size of the output buffer
136 *
137 * Only the contents of the output buffer from out[out_pos] onward, and
138 * the variables in_pos and out_pos are modified by the XZ code.
139 */
140struct xz_buf {
141	const uint8_t *in;
142	size_t in_pos;
143	size_t in_size;
144
145	uint8_t *out;
146	size_t out_pos;
147	size_t out_size;
148};
149
150/**
151 * struct xz_dec - Opaque type to hold the XZ decoder state
152 */
153struct xz_dec;
154
155/**
156 * xz_dec_init() - Allocate and initialize a XZ decoder state
157 * @mode:       Operation mode
158 * @dict_max:   Maximum size of the LZMA2 dictionary (history buffer) for
159 *              multi-call decoding. This is ignored in single-call mode
160 *              (mode == XZ_SINGLE). LZMA2 dictionary is always 2^n bytes
161 *              or 2^n + 2^(n-1) bytes (the latter sizes are less common
162 *              in practice), so other values for dict_max don't make sense.
163 *              In the kernel, dictionary sizes of 64 KiB, 128 KiB, 256 KiB,
164 *              512 KiB, and 1 MiB are probably the only reasonable values,
165 *              except for kernel and initramfs images where a bigger
166 *              dictionary can be fine and useful.
167 *
168 * Single-call mode (XZ_SINGLE): xz_dec_run() decodes the whole stream at
169 * once. The caller must provide enough output space or the decoding will
170 * fail. The output space is used as the dictionary buffer, which is why
171 * there is no need to allocate the dictionary as part of the decoder's
172 * internal state.
173 *
174 * Because the output buffer is used as the workspace, streams encoded using
175 * a big dictionary are not a problem in single-call mode. It is enough that
176 * the output buffer is big enough to hold the actual uncompressed data; it
177 * can be smaller than the dictionary size stored in the stream headers.
178 *
179 * Multi-call mode with preallocated dictionary (XZ_PREALLOC): dict_max bytes
180 * of memory is preallocated for the LZMA2 dictionary. This way there is no
181 * risk that xz_dec_run() could run out of memory, since xz_dec_run() will
182 * never allocate any memory. Instead, if the preallocated dictionary is too
183 * small for decoding the given input stream, xz_dec_run() will return
184 * XZ_MEMLIMIT_ERROR. Thus, it is important to know what kind of data will be
185 * decoded to avoid allocating excessive amount of memory for the dictionary.
186 *
187 * Multi-call mode with dynamically allocated dictionary (XZ_DYNALLOC):
188 * dict_max specifies the maximum allowed dictionary size that xz_dec_run()
189 * may allocate once it has parsed the dictionary size from the stream
190 * headers. This way excessive allocations can be avoided while still
191 * limiting the maximum memory usage to a sane value to prevent running the
192 * system out of memory when decompressing streams from untrusted sources.
193 *
194 * On success, xz_dec_init() returns a pointer to struct xz_dec, which is
195 * ready to be used with xz_dec_run(). If memory allocation fails,
196 * xz_dec_init() returns NULL.
197 */
198XZ_EXTERN struct xz_dec *xz_dec_init(enum xz_mode mode, uint32_t dict_max);
199
200/**
201 * xz_dec_run() - Run the XZ decoder
202 * @s:          Decoder state allocated using xz_dec_init()
203 * @b:          Input and output buffers
204 *
205 * The possible return values depend on build options and operation mode.
206 * See enum xz_ret for details.
207 *
208 * Note that if an error occurs in single-call mode (return value is not
209 * XZ_STREAM_END), b->in_pos and b->out_pos are not modified and the
210 * contents of the output buffer from b->out[b->out_pos] onward are
211 * undefined. This is true even after XZ_BUF_ERROR, because with some filter
212 * chains, there may be a second pass over the output buffer, and this pass
213 * cannot be properly done if the output buffer is truncated. Thus, you
214 * cannot give the single-call decoder a too small buffer and then expect to
215 * get that amount valid data from the beginning of the stream. You must use
216 * the multi-call decoder if you don't want to uncompress the whole stream.
217 */
218XZ_EXTERN enum xz_ret xz_dec_run(struct xz_dec *s, struct xz_buf *b);
219
220/**
221 * xz_dec_reset() - Reset an already allocated decoder state
222 * @s:          Decoder state allocated using xz_dec_init()
223 *
224 * This function can be used to reset the multi-call decoder state without
225 * freeing and reallocating memory with xz_dec_end() and xz_dec_init().
226 *
227 * In single-call mode, xz_dec_reset() is always called in the beginning of
228 * xz_dec_run(). Thus, explicit call to xz_dec_reset() is useful only in
229 * multi-call mode.
230 */
231XZ_EXTERN void xz_dec_reset(struct xz_dec *s);
232
233/**
234 * xz_dec_end() - Free the memory allocated for the decoder state
235 * @s:          Decoder state allocated using xz_dec_init(). If s is NULL,
236 *              this function does nothing.
237 */
238XZ_EXTERN void xz_dec_end(struct xz_dec *s);
239
240/*
241 * Standalone build (userspace build or in-kernel build for boot time use)
242 * needs a CRC32 implementation. For normal in-kernel use, kernel's own
243 * CRC32 module is used instead, and users of this module don't need to
244 * care about the functions below.
245 */
246#ifndef XZ_INTERNAL_CRC32
247#	ifdef __KERNEL__
248#		define XZ_INTERNAL_CRC32 0
249#	else
250#		define XZ_INTERNAL_CRC32 1
251#	endif
252#endif
253
254/*
255 * If CRC64 support has been enabled with XZ_USE_CRC64, a CRC64
256 * implementation is needed too.
257 */
258#ifndef XZ_USE_CRC64
259#	undef XZ_INTERNAL_CRC64
260#	define XZ_INTERNAL_CRC64 0
261#endif
262#ifndef XZ_INTERNAL_CRC64
263#	ifdef __KERNEL__
264#		error Using CRC64 in the kernel has not been implemented.
265#	else
266#		define XZ_INTERNAL_CRC64 1
267#	endif
268#endif
269
270#if XZ_INTERNAL_CRC32
271/*
272 * This must be called before any other xz_* function to initialize
273 * the CRC32 lookup table.
274 */
275XZ_EXTERN void xz_crc32_init(void);
276
277/*
278 * Update CRC32 value using the polynomial from IEEE-802.3. To start a new
279 * calculation, the third argument must be zero. To continue the calculation,
280 * the previously returned value is passed as the third argument.
281 */
282XZ_EXTERN uint32_t xz_crc32(const uint8_t *buf, size_t size, uint32_t crc);
283#endif
284
285#if XZ_INTERNAL_CRC64
286/*
287 * This must be called before any other xz_* function (except xz_crc32_init())
288 * to initialize the CRC64 lookup table.
289 */
290XZ_EXTERN void xz_crc64_init(void);
291
292/*
293 * Update CRC64 value using the polynomial from ECMA-182. To start a new
294 * calculation, the third argument must be zero. To continue the calculation,
295 * the previously returned value is passed as the third argument.
296 */
297XZ_EXTERN uint64_t xz_crc64(const uint8_t *buf, size_t size, uint64_t crc);
298#endif
299
300#ifdef __cplusplus
301}
302#endif
303
304#endif
305