VirtualBox

source: vbox/trunk/src/libs/liblzma-5.8.1/common/stream_decoder.c@ 108911

Last change on this file since 108911 was 108911, checked in by vboxsync, 4 weeks ago

libs/liblzma: Applied and adjusted our liblzma changes to 5.8.1 and export to OSE. jiraref:VBP-1635

  • Property svn:eol-style set to native
  • Property svn:keywords set to Author Date Id Revision
  • Property sync-process set to export
File size: 12.7 KB
Line 
1// SPDX-License-Identifier: 0BSD
2
3///////////////////////////////////////////////////////////////////////////////
4//
5/// \file stream_decoder.c
6/// \brief Decodes .xz Streams
7//
8// Author: Lasse Collin
9//
10///////////////////////////////////////////////////////////////////////////////
11
12#include "stream_decoder.h"
13#include "block_decoder.h"
14#include "index.h"
15
16
17typedef struct {
18 enum {
19 SEQ_STREAM_HEADER,
20 SEQ_BLOCK_HEADER,
21 SEQ_BLOCK_INIT,
22 SEQ_BLOCK_RUN,
23 SEQ_INDEX,
24 SEQ_STREAM_FOOTER,
25 SEQ_STREAM_PADDING,
26 } sequence;
27
28 /// Block decoder
29 lzma_next_coder block_decoder;
30
31 /// Block options decoded by the Block Header decoder and used by
32 /// the Block decoder.
33 lzma_block block_options;
34
35 /// Stream Flags from Stream Header
36 lzma_stream_flags stream_flags;
37
38 /// Index is hashed so that it can be compared to the sizes of Blocks
39 /// with O(1) memory usage.
40 lzma_index_hash *index_hash;
41
42 /// Memory usage limit
43 uint64_t memlimit;
44
45 /// Amount of memory actually needed (only an estimate)
46 uint64_t memusage;
47
48 /// If true, LZMA_NO_CHECK is returned if the Stream has
49 /// no integrity check.
50 bool tell_no_check;
51
52 /// If true, LZMA_UNSUPPORTED_CHECK is returned if the Stream has
53 /// an integrity check that isn't supported by this liblzma build.
54 bool tell_unsupported_check;
55
56 /// If true, LZMA_GET_CHECK is returned after decoding Stream Header.
57 bool tell_any_check;
58
59 /// If true, we will tell the Block decoder to skip calculating
60 /// and verifying the integrity check.
61 bool ignore_check;
62
63 /// If true, we will decode concatenated Streams that possibly have
64 /// Stream Padding between or after them. LZMA_STREAM_END is returned
65 /// once the application isn't giving us any new input (LZMA_FINISH),
66 /// and we aren't in the middle of a Stream, and possible
67 /// Stream Padding is a multiple of four bytes.
68 bool concatenated;
69
70 /// When decoding concatenated Streams, this is true as long as we
71 /// are decoding the first Stream. This is needed to avoid misleading
72 /// LZMA_FORMAT_ERROR in case the later Streams don't have valid magic
73 /// bytes.
74 bool first_stream;
75
76 /// Write position in buffer[] and position in Stream Padding
77 size_t pos;
78
79 /// Buffer to hold Stream Header, Block Header, and Stream Footer.
80 /// Block Header has biggest maximum size.
81 uint8_t buffer[LZMA_BLOCK_HEADER_SIZE_MAX];
82} lzma_stream_coder;
83
84
85static lzma_ret
86stream_decoder_reset(lzma_stream_coder *coder, const lzma_allocator *allocator)
87{
88 // Initialize the Index hash used to verify the Index.
89 coder->index_hash = lzma_index_hash_init(coder->index_hash, allocator);
90 if (coder->index_hash == NULL)
91 return LZMA_MEM_ERROR;
92
93 // Reset the rest of the variables.
94 coder->sequence = SEQ_STREAM_HEADER;
95 coder->pos = 0;
96
97 return LZMA_OK;
98}
99
100
101static lzma_ret
102stream_decode(void *coder_ptr, const lzma_allocator *allocator,
103 const uint8_t *restrict in, size_t *restrict in_pos,
104 size_t in_size, uint8_t *restrict out,
105 size_t *restrict out_pos, size_t out_size, lzma_action action)
106{
107 lzma_stream_coder *coder = coder_ptr;
108
109 // When decoding the actual Block, it may be able to produce more
110 // output even if we don't give it any new input.
111 while (true)
112 switch (coder->sequence) {
113 case SEQ_STREAM_HEADER: {
114 // Copy the Stream Header to the internal buffer.
115 lzma_bufcpy(in, in_pos, in_size, coder->buffer, &coder->pos,
116 LZMA_STREAM_HEADER_SIZE);
117
118 // Return if we didn't get the whole Stream Header yet.
119 if (coder->pos < LZMA_STREAM_HEADER_SIZE)
120 return LZMA_OK;
121
122 coder->pos = 0;
123
124 // Decode the Stream Header.
125 const lzma_ret ret = lzma_stream_header_decode(
126 &coder->stream_flags, coder->buffer);
127 if (ret != LZMA_OK)
128 return ret == LZMA_FORMAT_ERROR && !coder->first_stream
129 ? LZMA_DATA_ERROR : ret;
130
131 // If we are decoding concatenated Streams, and the later
132 // Streams have invalid Header Magic Bytes, we give
133 // LZMA_DATA_ERROR instead of LZMA_FORMAT_ERROR.
134 coder->first_stream = false;
135
136 // Copy the type of the Check so that Block Header and Block
137 // decoders see it.
138 coder->block_options.check = coder->stream_flags.check;
139
140 // Even if we return LZMA_*_CHECK below, we want
141 // to continue from Block Header decoding.
142 coder->sequence = SEQ_BLOCK_HEADER;
143
144 // Detect if there's no integrity check or if it is
145 // unsupported if those were requested by the application.
146 if (coder->tell_no_check && coder->stream_flags.check
147 == LZMA_CHECK_NONE)
148 return LZMA_NO_CHECK;
149
150 if (coder->tell_unsupported_check
151 && !lzma_check_is_supported(
152 coder->stream_flags.check))
153 return LZMA_UNSUPPORTED_CHECK;
154
155 if (coder->tell_any_check)
156 return LZMA_GET_CHECK;
157
158 FALLTHROUGH;
159 }
160
161 case SEQ_BLOCK_HEADER: {
162 if (*in_pos >= in_size)
163 return LZMA_OK;
164
165 if (coder->pos == 0) {
166 // Detect if it's Index.
167 if (in[*in_pos] == INDEX_INDICATOR) {
168 coder->sequence = SEQ_INDEX;
169 break;
170 }
171
172 // Calculate the size of the Block Header. Note that
173 // Block Header decoder wants to see this byte too
174 // so don't advance *in_pos.
175 coder->block_options.header_size
176 = lzma_block_header_size_decode(
177 in[*in_pos]);
178 }
179
180 // Copy the Block Header to the internal buffer.
181 lzma_bufcpy(in, in_pos, in_size, coder->buffer, &coder->pos,
182 coder->block_options.header_size);
183
184 // Return if we didn't get the whole Block Header yet.
185 if (coder->pos < coder->block_options.header_size)
186 return LZMA_OK;
187
188 coder->pos = 0;
189 coder->sequence = SEQ_BLOCK_INIT;
190 FALLTHROUGH;
191 }
192
193 case SEQ_BLOCK_INIT: {
194 // Checking memusage and doing the initialization needs
195 // its own sequence point because we need to be able to
196 // retry if we return LZMA_MEMLIMIT_ERROR.
197
198 // Version 1 is needed to support the .ignore_check option.
199 coder->block_options.version = 1;
200
201 // Set up a buffer to hold the filter chain. Block Header
202 // decoder will initialize all members of this array so
203 // we don't need to do it here.
204 lzma_filter filters[LZMA_FILTERS_MAX + 1];
205 coder->block_options.filters = filters;
206
207 // Decode the Block Header.
208 return_if_error(lzma_block_header_decode(&coder->block_options,
209 allocator, coder->buffer));
210
211 // If LZMA_IGNORE_CHECK was used, this flag needs to be set.
212 // It has to be set after lzma_block_header_decode() because
213 // it always resets this to false.
214 coder->block_options.ignore_check = coder->ignore_check;
215
216 // Check the memory usage limit.
217 const uint64_t memusage = lzma_raw_decoder_memusage(filters);
218 lzma_ret ret;
219
220 if (memusage == UINT64_MAX) {
221 // One or more unknown Filter IDs.
222 ret = LZMA_OPTIONS_ERROR;
223 } else {
224 // Now we can set coder->memusage since we know that
225 // the filter chain is valid. We don't want
226 // lzma_memusage() to return UINT64_MAX in case of
227 // invalid filter chain.
228 coder->memusage = memusage;
229
230 if (memusage > coder->memlimit) {
231 // The chain would need too much memory.
232 ret = LZMA_MEMLIMIT_ERROR;
233 } else {
234 // Memory usage is OK.
235 // Initialize the Block decoder.
236 ret = lzma_block_decoder_init(
237 &coder->block_decoder,
238 allocator,
239 &coder->block_options);
240 }
241 }
242
243 // Free the allocated filter options since they are needed
244 // only to initialize the Block decoder.
245 lzma_filters_free(filters, allocator);
246 coder->block_options.filters = NULL;
247
248 // Check if memory usage calculation and Block decoder
249 // initialization succeeded.
250 if (ret != LZMA_OK)
251 return ret;
252
253 coder->sequence = SEQ_BLOCK_RUN;
254 FALLTHROUGH;
255 }
256
257 case SEQ_BLOCK_RUN: {
258 const lzma_ret ret = coder->block_decoder.code(
259 coder->block_decoder.coder, allocator,
260 in, in_pos, in_size, out, out_pos, out_size,
261 action);
262
263 if (ret != LZMA_STREAM_END)
264 return ret;
265
266 // Block decoded successfully. Add the new size pair to
267 // the Index hash.
268 return_if_error(lzma_index_hash_append(coder->index_hash,
269 lzma_block_unpadded_size(
270 &coder->block_options),
271 coder->block_options.uncompressed_size));
272
273 coder->sequence = SEQ_BLOCK_HEADER;
274 break;
275 }
276
277 case SEQ_INDEX: {
278 // If we don't have any input, don't call
279 // lzma_index_hash_decode() since it would return
280 // LZMA_BUF_ERROR, which we must not do here.
281 if (*in_pos >= in_size)
282 return LZMA_OK;
283
284 // Decode the Index and compare it to the hash calculated
285 // from the sizes of the Blocks (if any).
286 const lzma_ret ret = lzma_index_hash_decode(coder->index_hash,
287 in, in_pos, in_size);
288 if (ret != LZMA_STREAM_END)
289 return ret;
290
291 coder->sequence = SEQ_STREAM_FOOTER;
292 FALLTHROUGH;
293 }
294
295 case SEQ_STREAM_FOOTER: {
296 // Copy the Stream Footer to the internal buffer.
297 lzma_bufcpy(in, in_pos, in_size, coder->buffer, &coder->pos,
298 LZMA_STREAM_HEADER_SIZE);
299
300 // Return if we didn't get the whole Stream Footer yet.
301 if (coder->pos < LZMA_STREAM_HEADER_SIZE)
302 return LZMA_OK;
303
304 coder->pos = 0;
305
306 // Decode the Stream Footer. The decoder gives
307 // LZMA_FORMAT_ERROR if the magic bytes don't match,
308 // so convert that return code to LZMA_DATA_ERROR.
309 lzma_stream_flags footer_flags;
310 const lzma_ret ret = lzma_stream_footer_decode(
311 &footer_flags, coder->buffer);
312 if (ret != LZMA_OK)
313 return ret == LZMA_FORMAT_ERROR
314 ? LZMA_DATA_ERROR : ret;
315
316 // Check that Index Size stored in the Stream Footer matches
317 // the real size of the Index field.
318 if (lzma_index_hash_size(coder->index_hash)
319 != footer_flags.backward_size)
320 return LZMA_DATA_ERROR;
321
322 // Compare that the Stream Flags fields are identical in
323 // both Stream Header and Stream Footer.
324 return_if_error(lzma_stream_flags_compare(
325 &coder->stream_flags, &footer_flags));
326
327 if (!coder->concatenated)
328 return LZMA_STREAM_END;
329
330 coder->sequence = SEQ_STREAM_PADDING;
331 FALLTHROUGH;
332 }
333
334 case SEQ_STREAM_PADDING:
335 assert(coder->concatenated);
336
337 // Skip over possible Stream Padding.
338 while (true) {
339 if (*in_pos >= in_size) {
340 // Unless LZMA_FINISH was used, we cannot
341 // know if there's more input coming later.
342 if (action != LZMA_FINISH)
343 return LZMA_OK;
344
345 // Stream Padding must be a multiple of
346 // four bytes.
347 return coder->pos == 0
348 ? LZMA_STREAM_END
349 : LZMA_DATA_ERROR;
350 }
351
352 // If the byte is not zero, it probably indicates
353 // beginning of a new Stream (or the file is corrupt).
354 if (in[*in_pos] != 0x00)
355 break;
356
357 ++*in_pos;
358 coder->pos = (coder->pos + 1) & 3;
359 }
360
361 // Stream Padding must be a multiple of four bytes (empty
362 // Stream Padding is OK).
363 if (coder->pos != 0) {
364 ++*in_pos;
365 return LZMA_DATA_ERROR;
366 }
367
368 // Prepare to decode the next Stream.
369 return_if_error(stream_decoder_reset(coder, allocator));
370 break;
371
372 default:
373 assert(0);
374 return LZMA_PROG_ERROR;
375 }
376
377 // Never reached
378}
379
380
381static void
382stream_decoder_end(void *coder_ptr, const lzma_allocator *allocator)
383{
384 lzma_stream_coder *coder = coder_ptr;
385 lzma_next_end(&coder->block_decoder, allocator);
386 lzma_index_hash_end(coder->index_hash, allocator);
387 lzma_free(coder, allocator);
388 return;
389}
390
391
392static lzma_check
393stream_decoder_get_check(const void *coder_ptr)
394{
395 const lzma_stream_coder *coder = coder_ptr;
396 return coder->stream_flags.check;
397}
398
399
400static lzma_ret
401stream_decoder_memconfig(void *coder_ptr, uint64_t *memusage,
402 uint64_t *old_memlimit, uint64_t new_memlimit)
403{
404 lzma_stream_coder *coder = coder_ptr;
405
406 *memusage = coder->memusage;
407 *old_memlimit = coder->memlimit;
408
409 if (new_memlimit != 0) {
410 if (new_memlimit < coder->memusage)
411 return LZMA_MEMLIMIT_ERROR;
412
413 coder->memlimit = new_memlimit;
414 }
415
416 return LZMA_OK;
417}
418
419
420extern lzma_ret
421lzma_stream_decoder_init(
422 lzma_next_coder *next, const lzma_allocator *allocator,
423 uint64_t memlimit, uint32_t flags)
424{
425 lzma_next_coder_init(&lzma_stream_decoder_init, next, allocator);
426
427 if (flags & ~LZMA_SUPPORTED_FLAGS)
428 return LZMA_OPTIONS_ERROR;
429
430 lzma_stream_coder *coder = next->coder;
431 if (coder == NULL) {
432 coder = lzma_alloc(sizeof(lzma_stream_coder), allocator);
433 if (coder == NULL)
434 return LZMA_MEM_ERROR;
435
436 next->coder = coder;
437 next->code = &stream_decode;
438 next->end = &stream_decoder_end;
439 next->get_check = &stream_decoder_get_check;
440 next->memconfig = &stream_decoder_memconfig;
441
442 coder->block_decoder = LZMA_NEXT_CODER_INIT;
443 coder->index_hash = NULL;
444 }
445
446 coder->memlimit = my_max(1, memlimit);
447 coder->memusage = LZMA_MEMUSAGE_BASE;
448 coder->tell_no_check = (flags & LZMA_TELL_NO_CHECK) != 0;
449 coder->tell_unsupported_check
450 = (flags & LZMA_TELL_UNSUPPORTED_CHECK) != 0;
451 coder->tell_any_check = (flags & LZMA_TELL_ANY_CHECK) != 0;
452 coder->ignore_check = (flags & LZMA_IGNORE_CHECK) != 0;
453 coder->concatenated = (flags & LZMA_CONCATENATED) != 0;
454 coder->first_stream = true;
455
456 return stream_decoder_reset(coder, allocator);
457}
458
459
460extern LZMA_API(lzma_ret)
461lzma_stream_decoder(lzma_stream *strm, uint64_t memlimit, uint32_t flags)
462{
463 lzma_next_strm_init(lzma_stream_decoder_init, strm, memlimit, flags);
464
465 strm->internal->supported_actions[LZMA_RUN] = true;
466 strm->internal->supported_actions[LZMA_FINISH] = true;
467
468 return LZMA_OK;
469}
Note: See TracBrowser for help on using the repository browser.

© 2025 Oracle Support Privacy / Do Not Sell My Info Terms of Use Trademark Policy Automated Access Etiquette