1 ///////////////////////////////////////////////////////////////////////////////
2 //
3 /// \file microlzma_decoder.c
4 /// \brief Decode MicroLZMA format
5 //
6 // Author: Lasse Collin
7 //
8 // This file has been put into the public domain.
9 // You can do whatever you want with this file.
10 //
11 ///////////////////////////////////////////////////////////////////////////////
12
13 #include "lzma_decoder.h"
14 #include "lz_decoder.h"
15
16
17 typedef struct {
18 /// LZMA1 decoder
19 lzma_next_coder lzma;
20
21 /// Compressed size of the stream as given by the application.
22 /// This must be exactly correct.
23 ///
24 /// This will be decremented when input is read.
25 uint64_t comp_size;
26
27 /// Uncompressed size of the stream as given by the application.
28 /// This may be less than the actual uncompressed size if
29 /// uncomp_size_is_exact is false.
30 ///
31 /// This will be decremented when output is produced.
32 lzma_vli uncomp_size;
33
34 /// LZMA dictionary size as given by the application
35 uint32_t dict_size;
36
37 /// If true, the exact uncompressed size is known. If false,
38 /// uncomp_size may be smaller than the real uncompressed size;
39 /// uncomp_size may never be bigger than the real uncompressed size.
40 bool uncomp_size_is_exact;
41
42 /// True once the first byte of the MicroLZMA stream
43 /// has been processed.
44 bool props_decoded;
45 } lzma_microlzma_coder;
46
47
48 static lzma_ret
49 microlzma_decode(void *coder_ptr, const lzma_allocator *allocator,
50 const uint8_t *restrict in, size_t *restrict in_pos,
51 size_t in_size, uint8_t *restrict out,
52 size_t *restrict out_pos, size_t out_size, lzma_action action)
53 {
54 lzma_microlzma_coder *coder = coder_ptr;
55
56 // Remember the in start position so that we can update comp_size.
57 const size_t in_start = *in_pos;
58
59 // Remember the out start position so that we can update uncomp_size.
60 const size_t out_start = *out_pos;
61
62 // Limit the amount of input so that the decoder won't read more than
63 // comp_size. This is required when uncomp_size isn't exact because
64 // in that case the LZMA decoder will try to decode more input even
65 // when it has no output space (it can be looking for EOPM).
66 if (in_size - *in_pos > coder->comp_size)
67 in_size = *in_pos + (size_t)(coder->comp_size);
68
69 // When the exact uncompressed size isn't known, we must limit
70 // the available output space to prevent the LZMA decoder from
71 // trying to decode too much.
72 if (!coder->uncomp_size_is_exact
73 && out_size - *out_pos > coder->uncomp_size)
74 out_size = *out_pos + (size_t)(coder->uncomp_size);
75
76 if (!coder->props_decoded) {
77 // There must be at least one byte of input to decode
78 // the properties byte.
79 if (*in_pos >= in_size)
80 return LZMA_OK;
81
82 lzma_options_lzma options = {
83 .dict_size = coder->dict_size,
84 .preset_dict = NULL,
85 .preset_dict_size = 0,
86 .ext_flags = 0, // EOPM not allowed when size is known
87 .ext_size_low = UINT32_MAX, // Unknown size by default
88 .ext_size_high = UINT32_MAX,
89 };
90
91 if (coder->uncomp_size_is_exact)
92 lzma_set_ext_size(options, coder->uncomp_size);
93
94 // The properties are stored as bitwise-negation
95 // of the typical encoding.
96 if (lzma_lzma_lclppb_decode(&options, ~in[*in_pos]))
97 return LZMA_OPTIONS_ERROR;
98
99 ++*in_pos;
100
101 // Initialize the decoder.
102 lzma_filter_info filters[2] = {
103 {
104 .id = LZMA_FILTER_LZMA1EXT,
105 .init = &lzma_lzma_decoder_init,
106 .options = &options,
107 }, {
108 .init = NULL,
109 }
110 };
111
112 return_if_error(lzma_next_filter_init(&coder->lzma,
113 allocator, filters));
114
115 // Pass one dummy 0x00 byte to the LZMA decoder since that
116 // is what it expects the first byte to be.
117 const uint8_t dummy_in = 0;
118 size_t dummy_in_pos = 0;
119 if (coder->lzma.code(coder->lzma.coder, allocator,
120 &dummy_in, &dummy_in_pos, 1,
121 out, out_pos, out_size, LZMA_RUN) != LZMA_OK)
122 return LZMA_PROG_ERROR;
123
124 assert(dummy_in_pos == 1);
125 coder->props_decoded = true;
126 }
127
128 // The rest is normal LZMA decoding.
129 lzma_ret ret = coder->lzma.code(coder->lzma.coder, allocator,
130 in, in_pos, in_size,
131 out, out_pos, out_size, action);
132
133 // Update the remaining compressed size.
134 assert(coder->comp_size >= *in_pos - in_start);
135 coder->comp_size -= *in_pos - in_start;
136
137 if (coder->uncomp_size_is_exact) {
138 // After successful decompression of the complete stream
139 // the compressed size must match.
140 if (ret == LZMA_STREAM_END && coder->comp_size != 0)
141 ret = LZMA_DATA_ERROR;
142 } else {
143 // Update the amount of output remaining.
144 assert(coder->uncomp_size >= *out_pos - out_start);
145 coder->uncomp_size -= *out_pos - out_start;
146
147 // - We must not get LZMA_STREAM_END because the stream
148 // shouldn't have EOPM.
149 // - We must use uncomp_size to determine when to
150 // return LZMA_STREAM_END.
151 if (ret == LZMA_STREAM_END)
152 ret = LZMA_DATA_ERROR;
153 else if (coder->uncomp_size == 0)
154 ret = LZMA_STREAM_END;
155 }
156
157 return ret;
158 }
159
160
161 static void
162 microlzma_decoder_end(void *coder_ptr, const lzma_allocator *allocator)
163 {
164 lzma_microlzma_coder *coder = coder_ptr;
165 lzma_next_end(&coder->lzma, allocator);
166 lzma_free(coder, allocator);
167 return;
168 }
169
170
171 static lzma_ret
172 microlzma_decoder_init(lzma_next_coder *next, const lzma_allocator *allocator,
173 uint64_t comp_size,
174 uint64_t uncomp_size, bool uncomp_size_is_exact,
175 uint32_t dict_size)
176 {
177 lzma_next_coder_init(µlzma_decoder_init, next, allocator);
178
179 lzma_microlzma_coder *coder = next->coder;
180
181 if (coder == NULL) {
182 coder = lzma_alloc(sizeof(lzma_microlzma_coder), allocator);
183 if (coder == NULL)
184 return LZMA_MEM_ERROR;
185
186 next->coder = coder;
187 next->code = µlzma_decode;
188 next->end = µlzma_decoder_end;
189
190 coder->lzma = LZMA_NEXT_CODER_INIT;
191 }
192
193 // The public API is uint64_t but the internal LZ decoder API uses
194 // lzma_vli.
195 if (uncomp_size > LZMA_VLI_MAX)
196 return LZMA_OPTIONS_ERROR;
197
198 coder->comp_size = comp_size;
199 coder->uncomp_size = uncomp_size;
200 coder->uncomp_size_is_exact = uncomp_size_is_exact;
201 coder->dict_size = dict_size;
202
203 coder->props_decoded = false;
204
205 return LZMA_OK;
206 }
207
208
209 extern LZMA_API(lzma_ret)
210 lzma_microlzma_decoder(lzma_stream *strm, uint64_t comp_size,
211 uint64_t uncomp_size, lzma_bool uncomp_size_is_exact,
212 uint32_t dict_size)
213 {
214 lzma_next_strm_init(microlzma_decoder_init, strm, comp_size,
215 uncomp_size, uncomp_size_is_exact, dict_size);
216
217 strm->internal->supported_actions[LZMA_RUN] = true;
218 strm->internal->supported_actions[LZMA_FINISH] = true;
219
220 return LZMA_OK;
221 }