1 /*
2  * LZMA2 decoder
3  *
4  * Authors: Lasse Collin <lasse.collin@tukaani.org>
5  *          Igor Pavlov <https://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 #include "xz_private.h"
12 #include "xz_lzma2.h"
13 
14 /*
15  * Range decoder initialization eats the first five bytes of each LZMA chunk.
16  */
17 #define RC_INIT_BYTES 5
18 
19 /*
20  * Minimum number of usable input buffer to safely decode one LZMA symbol.
21  * The worst case is that we decode 22 bits using probabilities and 26
22  * direct bits. This may decode at maximum of 20 bytes of input. However,
23  * lzma_main() does an extra normalization before returning, thus we
24  * need to put 21 here.
25  */
26 #define LZMA_IN_REQUIRED 21
27 
28 /*
29  * Dictionary (history buffer)
30  *
31  * These are always true:
32  *    start <= pos <= full <= end
33  *    pos <= limit <= end
34  *
35  * In multi-call mode, also these are true:
36  *    end == size
37  *    size <= size_max
38  *    allocated <= size
39  *
40  * Most of these variables are size_t to support single-call mode,
41  * in which the dictionary variables address the actual output
42  * buffer directly.
43  */
44 struct dictionary {
45 	/* Beginning of the history buffer */
46 	uint8_t *buf;
47 
48 	/* Old position in buf (before decoding more data) */
49 	size_t start;
50 
51 	/* Position in buf */
52 	size_t pos;
53 
54 	/*
55 	 * How full dictionary is. This is used to detect corrupt input that
56 	 * would read beyond the beginning of the uncompressed stream.
57 	 */
58 	size_t full;
59 
60 	/* Write limit; we don't write to buf[limit] or later bytes. */
61 	size_t limit;
62 
63 	/*
64 	 * End of the dictionary buffer. In multi-call mode, this is
65 	 * the same as the dictionary size. In single-call mode, this
66 	 * indicates the size of the output buffer.
67 	 */
68 	size_t end;
69 
70 	/*
71 	 * Size of the dictionary as specified in Block Header. This is used
72 	 * together with "full" to detect corrupt input that would make us
73 	 * read beyond the beginning of the uncompressed stream.
74 	 */
75 	uint32_t size;
76 
77 	/*
78 	 * Maximum allowed dictionary size in multi-call mode.
79 	 * This is ignored in single-call mode.
80 	 */
81 	uint32_t size_max;
82 
83 	/*
84 	 * Amount of memory currently allocated for the dictionary.
85 	 * This is used only with XZ_DYNALLOC. (With XZ_PREALLOC,
86 	 * size_max is always the same as the allocated size.)
87 	 */
88 	uint32_t allocated;
89 
90 	/* Operation mode */
91 	enum xz_mode mode;
92 };
93 
94 /* Range decoder */
95 struct rc_dec {
96 	uint32_t range;
97 	uint32_t code;
98 
99 	/*
100 	 * Number of initializing bytes remaining to be read
101 	 * by rc_read_init().
102 	 */
103 	uint32_t init_bytes_left;
104 
105 	/*
106 	 * Buffer from which we read our input. It can be either
107 	 * temp.buf or the caller-provided input buffer.
108 	 */
109 	const uint8_t *in;
110 	size_t in_pos;
111 	size_t in_limit;
112 };
113 
114 /* Probabilities for a length decoder. */
115 struct lzma_len_dec {
116 	/* Probability of match length being at least 10 */
117 	uint16_t choice;
118 
119 	/* Probability of match length being at least 18 */
120 	uint16_t choice2;
121 
122 	/* Probabilities for match lengths 2-9 */
123 	uint16_t low[POS_STATES_MAX][LEN_LOW_SYMBOLS];
124 
125 	/* Probabilities for match lengths 10-17 */
126 	uint16_t mid[POS_STATES_MAX][LEN_MID_SYMBOLS];
127 
128 	/* Probabilities for match lengths 18-273 */
129 	uint16_t high[LEN_HIGH_SYMBOLS];
130 };
131 
132 struct lzma_dec {
133 	/* Distances of latest four matches */
134 	uint32_t rep0;
135 	uint32_t rep1;
136 	uint32_t rep2;
137 	uint32_t rep3;
138 
139 	/* Types of the most recently seen LZMA symbols */
140 	enum lzma_state state;
141 
142 	/*
143 	 * Length of a match. This is updated so that dict_repeat can
144 	 * be called again to finish repeating the whole match.
145 	 */
146 	uint32_t len;
147 
148 	/*
149 	 * LZMA properties or related bit masks (number of literal
150 	 * context bits, a mask derived from the number of literal
151 	 * position bits, and a mask derived from the number
152 	 * position bits)
153 	 */
154 	uint32_t lc;
155 	uint32_t literal_pos_mask; /* (1 << lp) - 1 */
156 	uint32_t pos_mask;         /* (1 << pb) - 1 */
157 
158 	/* If 1, it's a match. Otherwise it's a single 8-bit literal. */
159 	uint16_t is_match[STATES][POS_STATES_MAX];
160 
161 	/* If 1, it's a repeated match. The distance is one of rep0 .. rep3. */
162 	uint16_t is_rep[STATES];
163 
164 	/*
165 	 * If 0, distance of a repeated match is rep0.
166 	 * Otherwise check is_rep1.
167 	 */
168 	uint16_t is_rep0[STATES];
169 
170 	/*
171 	 * If 0, distance of a repeated match is rep1.
172 	 * Otherwise check is_rep2.
173 	 */
174 	uint16_t is_rep1[STATES];
175 
176 	/* If 0, distance of a repeated match is rep2. Otherwise it is rep3. */
177 	uint16_t is_rep2[STATES];
178 
179 	/*
180 	 * If 1, the repeated match has length of one byte. Otherwise
181 	 * the length is decoded from rep_len_decoder.
182 	 */
183 	uint16_t is_rep0_long[STATES][POS_STATES_MAX];
184 
185 	/*
186 	 * Probability tree for the highest two bits of the match
187 	 * distance. There is a separate probability tree for match
188 	 * lengths of 2 (i.e. MATCH_LEN_MIN), 3, 4, and [5, 273].
189 	 */
190 	uint16_t dist_slot[DIST_STATES][DIST_SLOTS];
191 
192 	/*
193 	 * Probility trees for additional bits for match distance
194 	 * when the distance is in the range [4, 127].
195 	 */
196 	uint16_t dist_special[FULL_DISTANCES - DIST_MODEL_END];
197 
198 	/*
199 	 * Probability tree for the lowest four bits of a match
200 	 * distance that is equal to or greater than 128.
201 	 */
202 	uint16_t dist_align[ALIGN_SIZE];
203 
204 	/* Length of a normal match */
205 	struct lzma_len_dec match_len_dec;
206 
207 	/* Length of a repeated match */
208 	struct lzma_len_dec rep_len_dec;
209 
210 	/* Probabilities of literals */
211 	uint16_t literal[LITERAL_CODERS_MAX][LITERAL_CODER_SIZE];
212 };
213 
214 struct lzma2_dec {
215 	/* Position in xz_dec_lzma2_run(). */
216 	enum lzma2_seq {
217 		SEQ_CONTROL,
218 		SEQ_UNCOMPRESSED_1,
219 		SEQ_UNCOMPRESSED_2,
220 		SEQ_COMPRESSED_0,
221 		SEQ_COMPRESSED_1,
222 		SEQ_PROPERTIES,
223 		SEQ_LZMA_PREPARE,
224 		SEQ_LZMA_RUN,
225 		SEQ_COPY
226 	} sequence;
227 
228 	/* Next position after decoding the compressed size of the chunk. */
229 	enum lzma2_seq next_sequence;
230 
231 	/* Uncompressed size of LZMA chunk (2 MiB at maximum) */
232 	uint32_t uncompressed;
233 
234 	/*
235 	 * Compressed size of LZMA chunk or compressed/uncompressed
236 	 * size of uncompressed chunk (64 KiB at maximum)
237 	 */
238 	uint32_t compressed;
239 
240 	/*
241 	 * True if dictionary reset is needed. This is false before
242 	 * the first chunk (LZMA or uncompressed).
243 	 */
244 	bool need_dict_reset;
245 
246 	/*
247 	 * True if new LZMA properties are needed. This is false
248 	 * before the first LZMA chunk.
249 	 */
250 	bool need_props;
251 
252 #ifdef XZ_DEC_MICROLZMA
253 	bool pedantic_microlzma;
254 #endif
255 };
256 
257 struct xz_dec_lzma2 {
258 	/*
259 	 * The order below is important on x86 to reduce code size and
260 	 * it shouldn't hurt on other platforms. Everything up to and
261 	 * including lzma.pos_mask are in the first 128 bytes on x86-32,
262 	 * which allows using smaller instructions to access those
263 	 * variables. On x86-64, fewer variables fit into the first 128
264 	 * bytes, but this is still the best order without sacrificing
265 	 * the readability by splitting the structures.
266 	 */
267 	struct rc_dec rc;
268 	struct dictionary dict;
269 	struct lzma2_dec lzma2;
270 	struct lzma_dec lzma;
271 
272 	/*
273 	 * Temporary buffer which holds small number of input bytes between
274 	 * decoder calls. See lzma2_lzma() for details.
275 	 */
276 	struct {
277 		uint32_t size;
278 		uint8_t buf[3 * LZMA_IN_REQUIRED];
279 	} temp;
280 };
281 
282 /**************
283  * Dictionary *
284  **************/
285 
286 /*
287  * Reset the dictionary state. When in single-call mode, set up the beginning
288  * of the dictionary to point to the actual output buffer.
289  */
dict_reset(struct dictionary * dict,struct xz_buf * b)290 static void dict_reset(struct dictionary *dict, struct xz_buf *b)
291 {
292 	if (DEC_IS_SINGLE(dict->mode)) {
293 		dict->buf = b->out + b->out_pos;
294 		dict->end = b->out_size - b->out_pos;
295 	}
296 
297 	dict->start = 0;
298 	dict->pos = 0;
299 	dict->limit = 0;
300 	dict->full = 0;
301 }
302 
303 /* Set dictionary write limit */
dict_limit(struct dictionary * dict,size_t out_max)304 static void dict_limit(struct dictionary *dict, size_t out_max)
305 {
306 	if (dict->end - dict->pos <= out_max)
307 		dict->limit = dict->end;
308 	else
309 		dict->limit = dict->pos + out_max;
310 }
311 
312 /* Return true if at least one byte can be written into the dictionary. */
dict_has_space(const struct dictionary * dict)313 static inline bool dict_has_space(const struct dictionary *dict)
314 {
315 	return dict->pos < dict->limit;
316 }
317 
318 /*
319  * Get a byte from the dictionary at the given distance. The distance is
320  * assumed to valid, or as a special case, zero when the dictionary is
321  * still empty. This special case is needed for single-call decoding to
322  * avoid writing a '\0' to the end of the destination buffer.
323  */
dict_get(const struct dictionary * dict,uint32_t dist)324 static inline uint32_t dict_get(const struct dictionary *dict, uint32_t dist)
325 {
326 	size_t offset = dict->pos - dist - 1;
327 
328 	if (dist >= dict->pos)
329 		offset += dict->end;
330 
331 	return dict->full > 0 ? dict->buf[offset] : 0;
332 }
333 
334 /*
335  * Put one byte into the dictionary. It is assumed that there is space for it.
336  */
dict_put(struct dictionary * dict,uint8_t byte)337 static inline void dict_put(struct dictionary *dict, uint8_t byte)
338 {
339 	dict->buf[dict->pos++] = byte;
340 
341 	if (dict->full < dict->pos)
342 		dict->full = dict->pos;
343 }
344 
345 /*
346  * Repeat given number of bytes from the given distance. If the distance is
347  * invalid, false is returned. On success, true is returned and *len is
348  * updated to indicate how many bytes were left to be repeated.
349  */
dict_repeat(struct dictionary * dict,uint32_t * len,uint32_t dist)350 static bool dict_repeat(struct dictionary *dict, uint32_t *len, uint32_t dist)
351 {
352 	size_t back;
353 	uint32_t left;
354 
355 	if (dist >= dict->full || dist >= dict->size)
356 		return false;
357 
358 	left = min_t(size_t, dict->limit - dict->pos, *len);
359 	*len -= left;
360 
361 	back = dict->pos - dist - 1;
362 	if (dist >= dict->pos)
363 		back += dict->end;
364 
365 	do {
366 		dict->buf[dict->pos++] = dict->buf[back++];
367 		if (back == dict->end)
368 			back = 0;
369 	} while (--left > 0);
370 
371 	if (dict->full < dict->pos)
372 		dict->full = dict->pos;
373 
374 	return true;
375 }
376 
377 /* Copy uncompressed data as is from input to dictionary and output buffers. */
dict_uncompressed(struct dictionary * dict,struct xz_buf * b,uint32_t * left)378 static void dict_uncompressed(struct dictionary *dict, struct xz_buf *b,
379 			      uint32_t *left)
380 {
381 	size_t copy_size;
382 
383 	while (*left > 0 && b->in_pos < b->in_size
384 			&& b->out_pos < b->out_size) {
385 		copy_size = min(b->in_size - b->in_pos,
386 				b->out_size - b->out_pos);
387 		if (copy_size > dict->end - dict->pos)
388 			copy_size = dict->end - dict->pos;
389 		if (copy_size > *left)
390 			copy_size = *left;
391 
392 		*left -= copy_size;
393 
394 		/*
395 		 * If doing in-place decompression in single-call mode and the
396 		 * uncompressed size of the file is larger than the caller
397 		 * thought (i.e. it is invalid input!), the buffers below may
398 		 * overlap and cause undefined behavior with memcpy().
399 		 * With valid inputs memcpy() would be fine here.
400 		 */
401 		memmove(dict->buf + dict->pos, b->in + b->in_pos, copy_size);
402 		dict->pos += copy_size;
403 
404 		if (dict->full < dict->pos)
405 			dict->full = dict->pos;
406 
407 		if (DEC_IS_MULTI(dict->mode)) {
408 			if (dict->pos == dict->end)
409 				dict->pos = 0;
410 
411 			/*
412 			 * Like above but for multi-call mode: use memmove()
413 			 * to avoid undefined behavior with invalid input.
414 			 */
415 			memmove(b->out + b->out_pos, b->in + b->in_pos,
416 					copy_size);
417 		}
418 
419 		dict->start = dict->pos;
420 
421 		b->out_pos += copy_size;
422 		b->in_pos += copy_size;
423 	}
424 }
425 
426 #ifdef XZ_DEC_MICROLZMA
427 #	define DICT_FLUSH_SUPPORTS_SKIPPING true
428 #else
429 #	define DICT_FLUSH_SUPPORTS_SKIPPING false
430 #endif
431 
432 /*
433  * Flush pending data from dictionary to b->out. It is assumed that there is
434  * enough space in b->out. This is guaranteed because caller uses dict_limit()
435  * before decoding data into the dictionary.
436  */
dict_flush(struct dictionary * dict,struct xz_buf * b)437 static uint32_t dict_flush(struct dictionary *dict, struct xz_buf *b)
438 {
439 	size_t copy_size = dict->pos - dict->start;
440 
441 	if (DEC_IS_MULTI(dict->mode)) {
442 		if (dict->pos == dict->end)
443 			dict->pos = 0;
444 
445 		/*
446 		 * These buffers cannot overlap even if doing in-place
447 		 * decompression because in multi-call mode dict->buf
448 		 * has been allocated by us in this file; it's not
449 		 * provided by the caller like in single-call mode.
450 		 *
451 		 * With MicroLZMA, b->out can be NULL to skip bytes that
452 		 * the caller doesn't need. This cannot be done with XZ
453 		 * because it would break BCJ filters.
454 		 */
455 		if (!DICT_FLUSH_SUPPORTS_SKIPPING || b->out != NULL)
456 			memcpy(b->out + b->out_pos, dict->buf + dict->start,
457 					copy_size);
458 	}
459 
460 	dict->start = dict->pos;
461 	b->out_pos += copy_size;
462 	return copy_size;
463 }
464 
465 /*****************
466  * Range decoder *
467  *****************/
468 
469 /* Reset the range decoder. */
rc_reset(struct rc_dec * rc)470 static void rc_reset(struct rc_dec *rc)
471 {
472 	rc->range = (uint32_t)-1;
473 	rc->code = 0;
474 	rc->init_bytes_left = RC_INIT_BYTES;
475 }
476 
477 /*
478  * Read the first five initial bytes into rc->code if they haven't been
479  * read already. (Yes, the first byte gets completely ignored.)
480  */
rc_read_init(struct rc_dec * rc,struct xz_buf * b)481 static bool rc_read_init(struct rc_dec *rc, struct xz_buf *b)
482 {
483 	while (rc->init_bytes_left > 0) {
484 		if (b->in_pos == b->in_size)
485 			return false;
486 
487 		rc->code = (rc->code << 8) + b->in[b->in_pos++];
488 		--rc->init_bytes_left;
489 	}
490 
491 	return true;
492 }
493 
494 /* Return true if there may not be enough input for the next decoding loop. */
rc_limit_exceeded(const struct rc_dec * rc)495 static inline bool rc_limit_exceeded(const struct rc_dec *rc)
496 {
497 	return rc->in_pos > rc->in_limit;
498 }
499 
500 /*
501  * Return true if it is possible (from point of view of range decoder) that
502  * we have reached the end of the LZMA chunk.
503  */
rc_is_finished(const struct rc_dec * rc)504 static inline bool rc_is_finished(const struct rc_dec *rc)
505 {
506 	return rc->code == 0;
507 }
508 
509 /* Read the next input byte if needed. */
rc_normalize(struct rc_dec * rc)510 static __always_inline void rc_normalize(struct rc_dec *rc)
511 {
512 	if (rc->range < RC_TOP_VALUE) {
513 		rc->range <<= RC_SHIFT_BITS;
514 		rc->code = (rc->code << RC_SHIFT_BITS) + rc->in[rc->in_pos++];
515 	}
516 }
517 
518 /*
519  * Decode one bit. In some versions, this function has been split in three
520  * functions so that the compiler is supposed to be able to more easily avoid
521  * an extra branch. In this particular version of the LZMA decoder, this
522  * doesn't seem to be a good idea (tested with GCC 3.3.6, 3.4.6, and 4.3.3
523  * on x86). Using a non-split version results in nicer looking code too.
524  *
525  * NOTE: This must return an int. Do not make it return a bool or the speed
526  * of the code generated by GCC 3.x decreases 10-15 %. (GCC 4.3 doesn't care,
527  * and it generates 10-20 % faster code than GCC 3.x from this file anyway.)
528  */
rc_bit(struct rc_dec * rc,uint16_t * prob)529 static __always_inline int rc_bit(struct rc_dec *rc, uint16_t *prob)
530 {
531 	uint32_t bound;
532 	int bit;
533 
534 	rc_normalize(rc);
535 	bound = (rc->range >> RC_BIT_MODEL_TOTAL_BITS) * *prob;
536 	if (rc->code < bound) {
537 		rc->range = bound;
538 		*prob += (RC_BIT_MODEL_TOTAL - *prob) >> RC_MOVE_BITS;
539 		bit = 0;
540 	} else {
541 		rc->range -= bound;
542 		rc->code -= bound;
543 		*prob -= *prob >> RC_MOVE_BITS;
544 		bit = 1;
545 	}
546 
547 	return bit;
548 }
549 
550 /* Decode a bittree starting from the most significant bit. */
rc_bittree(struct rc_dec * rc,uint16_t * probs,uint32_t limit)551 static __always_inline uint32_t rc_bittree(struct rc_dec *rc,
552 					   uint16_t *probs, uint32_t limit)
553 {
554 	uint32_t symbol = 1;
555 
556 	do {
557 		if (rc_bit(rc, &probs[symbol]))
558 			symbol = (symbol << 1) + 1;
559 		else
560 			symbol <<= 1;
561 	} while (symbol < limit);
562 
563 	return symbol;
564 }
565 
566 /* Decode a bittree starting from the least significant bit. */
rc_bittree_reverse(struct rc_dec * rc,uint16_t * probs,uint32_t * dest,uint32_t limit)567 static __always_inline void rc_bittree_reverse(struct rc_dec *rc,
568 					       uint16_t *probs,
569 					       uint32_t *dest, uint32_t limit)
570 {
571 	uint32_t symbol = 1;
572 	uint32_t i = 0;
573 
574 	do {
575 		if (rc_bit(rc, &probs[symbol])) {
576 			symbol = (symbol << 1) + 1;
577 			*dest += 1 << i;
578 		} else {
579 			symbol <<= 1;
580 		}
581 	} while (++i < limit);
582 }
583 
584 /* Decode direct bits (fixed fifty-fifty probability) */
rc_direct(struct rc_dec * rc,uint32_t * dest,uint32_t limit)585 static inline void rc_direct(struct rc_dec *rc, uint32_t *dest, uint32_t limit)
586 {
587 	uint32_t mask;
588 
589 	do {
590 		rc_normalize(rc);
591 		rc->range >>= 1;
592 		rc->code -= rc->range;
593 		mask = (uint32_t)0 - (rc->code >> 31);
594 		rc->code += rc->range & mask;
595 		*dest = (*dest << 1) + (mask + 1);
596 	} while (--limit > 0);
597 }
598 
599 /********
600  * LZMA *
601  ********/
602 
603 /* Get pointer to literal coder probability array. */
lzma_literal_probs(struct xz_dec_lzma2 * s)604 static uint16_t *lzma_literal_probs(struct xz_dec_lzma2 *s)
605 {
606 	uint32_t prev_byte = dict_get(&s->dict, 0);
607 	uint32_t low = prev_byte >> (8 - s->lzma.lc);
608 	uint32_t high = (s->dict.pos & s->lzma.literal_pos_mask) << s->lzma.lc;
609 	return s->lzma.literal[low + high];
610 }
611 
612 /* Decode a literal (one 8-bit byte) */
lzma_literal(struct xz_dec_lzma2 * s)613 static void lzma_literal(struct xz_dec_lzma2 *s)
614 {
615 	uint16_t *probs;
616 	uint32_t symbol;
617 	uint32_t match_byte;
618 	uint32_t match_bit;
619 	uint32_t offset;
620 	uint32_t i;
621 
622 	probs = lzma_literal_probs(s);
623 
624 	if (lzma_state_is_literal(s->lzma.state)) {
625 		symbol = rc_bittree(&s->rc, probs, 0x100);
626 	} else {
627 		symbol = 1;
628 		match_byte = dict_get(&s->dict, s->lzma.rep0) << 1;
629 		offset = 0x100;
630 
631 		do {
632 			match_bit = match_byte & offset;
633 			match_byte <<= 1;
634 			i = offset + match_bit + symbol;
635 
636 			if (rc_bit(&s->rc, &probs[i])) {
637 				symbol = (symbol << 1) + 1;
638 				offset &= match_bit;
639 			} else {
640 				symbol <<= 1;
641 				offset &= ~match_bit;
642 			}
643 		} while (symbol < 0x100);
644 	}
645 
646 	dict_put(&s->dict, (uint8_t)symbol);
647 	lzma_state_literal(&s->lzma.state);
648 }
649 
650 /* Decode the length of the match into s->lzma.len. */
lzma_len(struct xz_dec_lzma2 * s,struct lzma_len_dec * l,uint32_t pos_state)651 static void lzma_len(struct xz_dec_lzma2 *s, struct lzma_len_dec *l,
652 		     uint32_t pos_state)
653 {
654 	uint16_t *probs;
655 	uint32_t limit;
656 
657 	if (!rc_bit(&s->rc, &l->choice)) {
658 		probs = l->low[pos_state];
659 		limit = LEN_LOW_SYMBOLS;
660 		s->lzma.len = MATCH_LEN_MIN;
661 	} else {
662 		if (!rc_bit(&s->rc, &l->choice2)) {
663 			probs = l->mid[pos_state];
664 			limit = LEN_MID_SYMBOLS;
665 			s->lzma.len = MATCH_LEN_MIN + LEN_LOW_SYMBOLS;
666 		} else {
667 			probs = l->high;
668 			limit = LEN_HIGH_SYMBOLS;
669 			s->lzma.len = MATCH_LEN_MIN + LEN_LOW_SYMBOLS
670 					+ LEN_MID_SYMBOLS;
671 		}
672 	}
673 
674 	s->lzma.len += rc_bittree(&s->rc, probs, limit) - limit;
675 }
676 
677 /* Decode a match. The distance will be stored in s->lzma.rep0. */
lzma_match(struct xz_dec_lzma2 * s,uint32_t pos_state)678 static void lzma_match(struct xz_dec_lzma2 *s, uint32_t pos_state)
679 {
680 	uint16_t *probs;
681 	uint32_t dist_slot;
682 	uint32_t limit;
683 
684 	lzma_state_match(&s->lzma.state);
685 
686 	s->lzma.rep3 = s->lzma.rep2;
687 	s->lzma.rep2 = s->lzma.rep1;
688 	s->lzma.rep1 = s->lzma.rep0;
689 
690 	lzma_len(s, &s->lzma.match_len_dec, pos_state);
691 
692 	probs = s->lzma.dist_slot[lzma_get_dist_state(s->lzma.len)];
693 	dist_slot = rc_bittree(&s->rc, probs, DIST_SLOTS) - DIST_SLOTS;
694 
695 	if (dist_slot < DIST_MODEL_START) {
696 		s->lzma.rep0 = dist_slot;
697 	} else {
698 		limit = (dist_slot >> 1) - 1;
699 		s->lzma.rep0 = 2 + (dist_slot & 1);
700 
701 		if (dist_slot < DIST_MODEL_END) {
702 			s->lzma.rep0 <<= limit;
703 			probs = s->lzma.dist_special + s->lzma.rep0
704 					- dist_slot - 1;
705 			rc_bittree_reverse(&s->rc, probs,
706 					&s->lzma.rep0, limit);
707 		} else {
708 			rc_direct(&s->rc, &s->lzma.rep0, limit - ALIGN_BITS);
709 			s->lzma.rep0 <<= ALIGN_BITS;
710 			rc_bittree_reverse(&s->rc, s->lzma.dist_align,
711 					&s->lzma.rep0, ALIGN_BITS);
712 		}
713 	}
714 }
715 
716 /*
717  * Decode a repeated match. The distance is one of the four most recently
718  * seen matches. The distance will be stored in s->lzma.rep0.
719  */
lzma_rep_match(struct xz_dec_lzma2 * s,uint32_t pos_state)720 static void lzma_rep_match(struct xz_dec_lzma2 *s, uint32_t pos_state)
721 {
722 	uint32_t tmp;
723 
724 	if (!rc_bit(&s->rc, &s->lzma.is_rep0[s->lzma.state])) {
725 		if (!rc_bit(&s->rc, &s->lzma.is_rep0_long[
726 				s->lzma.state][pos_state])) {
727 			lzma_state_short_rep(&s->lzma.state);
728 			s->lzma.len = 1;
729 			return;
730 		}
731 	} else {
732 		if (!rc_bit(&s->rc, &s->lzma.is_rep1[s->lzma.state])) {
733 			tmp = s->lzma.rep1;
734 		} else {
735 			if (!rc_bit(&s->rc, &s->lzma.is_rep2[s->lzma.state])) {
736 				tmp = s->lzma.rep2;
737 			} else {
738 				tmp = s->lzma.rep3;
739 				s->lzma.rep3 = s->lzma.rep2;
740 			}
741 
742 			s->lzma.rep2 = s->lzma.rep1;
743 		}
744 
745 		s->lzma.rep1 = s->lzma.rep0;
746 		s->lzma.rep0 = tmp;
747 	}
748 
749 	lzma_state_long_rep(&s->lzma.state);
750 	lzma_len(s, &s->lzma.rep_len_dec, pos_state);
751 }
752 
753 /* LZMA decoder core */
lzma_main(struct xz_dec_lzma2 * s)754 static bool lzma_main(struct xz_dec_lzma2 *s)
755 {
756 	uint32_t pos_state;
757 
758 	/*
759 	 * If the dictionary was reached during the previous call, try to
760 	 * finish the possibly pending repeat in the dictionary.
761 	 */
762 	if (dict_has_space(&s->dict) && s->lzma.len > 0)
763 		dict_repeat(&s->dict, &s->lzma.len, s->lzma.rep0);
764 
765 	/*
766 	 * Decode more LZMA symbols. One iteration may consume up to
767 	 * LZMA_IN_REQUIRED - 1 bytes.
768 	 */
769 	while (dict_has_space(&s->dict) && !rc_limit_exceeded(&s->rc)) {
770 		pos_state = s->dict.pos & s->lzma.pos_mask;
771 
772 		if (!rc_bit(&s->rc, &s->lzma.is_match[
773 				s->lzma.state][pos_state])) {
774 			lzma_literal(s);
775 		} else {
776 			if (rc_bit(&s->rc, &s->lzma.is_rep[s->lzma.state]))
777 				lzma_rep_match(s, pos_state);
778 			else
779 				lzma_match(s, pos_state);
780 
781 			if (!dict_repeat(&s->dict, &s->lzma.len, s->lzma.rep0))
782 				return false;
783 		}
784 	}
785 
786 	/*
787 	 * Having the range decoder always normalized when we are outside
788 	 * this function makes it easier to correctly handle end of the chunk.
789 	 */
790 	rc_normalize(&s->rc);
791 
792 	return true;
793 }
794 
795 /*
796  * Reset the LZMA decoder and range decoder state. Dictionary is not reset
797  * here, because LZMA state may be reset without resetting the dictionary.
798  */
lzma_reset(struct xz_dec_lzma2 * s)799 static void lzma_reset(struct xz_dec_lzma2 *s)
800 {
801 	uint16_t *probs;
802 	size_t i;
803 
804 	s->lzma.state = STATE_LIT_LIT;
805 	s->lzma.rep0 = 0;
806 	s->lzma.rep1 = 0;
807 	s->lzma.rep2 = 0;
808 	s->lzma.rep3 = 0;
809 	s->lzma.len = 0;
810 
811 	/*
812 	 * All probabilities are initialized to the same value. This hack
813 	 * makes the code smaller by avoiding a separate loop for each
814 	 * probability array.
815 	 *
816 	 * This could be optimized so that only that part of literal
817 	 * probabilities that are actually required. In the common case
818 	 * we would write 12 KiB less.
819 	 */
820 	probs = s->lzma.is_match[0];
821 	for (i = 0; i < PROBS_TOTAL; ++i)
822 		probs[i] = RC_BIT_MODEL_TOTAL / 2;
823 
824 	rc_reset(&s->rc);
825 }
826 
827 /*
828  * Decode and validate LZMA properties (lc/lp/pb) and calculate the bit masks
829  * from the decoded lp and pb values. On success, the LZMA decoder state is
830  * reset and true is returned.
831  */
lzma_props(struct xz_dec_lzma2 * s,uint8_t props)832 static bool lzma_props(struct xz_dec_lzma2 *s, uint8_t props)
833 {
834 	if (props > (4 * 5 + 4) * 9 + 8)
835 		return false;
836 
837 	s->lzma.pos_mask = 0;
838 	while (props >= 9 * 5) {
839 		props -= 9 * 5;
840 		++s->lzma.pos_mask;
841 	}
842 
843 	s->lzma.pos_mask = (1 << s->lzma.pos_mask) - 1;
844 
845 	s->lzma.literal_pos_mask = 0;
846 	while (props >= 9) {
847 		props -= 9;
848 		++s->lzma.literal_pos_mask;
849 	}
850 
851 	s->lzma.lc = props;
852 
853 	if (s->lzma.lc + s->lzma.literal_pos_mask > 4)
854 		return false;
855 
856 	s->lzma.literal_pos_mask = (1 << s->lzma.literal_pos_mask) - 1;
857 
858 	lzma_reset(s);
859 
860 	return true;
861 }
862 
863 /*********
864  * LZMA2 *
865  *********/
866 
867 /*
868  * The LZMA decoder assumes that if the input limit (s->rc.in_limit) hasn't
869  * been exceeded, it is safe to read up to LZMA_IN_REQUIRED bytes. This
870  * wrapper function takes care of making the LZMA decoder's assumption safe.
871  *
872  * As long as there is plenty of input left to be decoded in the current LZMA
873  * chunk, we decode directly from the caller-supplied input buffer until
874  * there's LZMA_IN_REQUIRED bytes left. Those remaining bytes are copied into
875  * s->temp.buf, which (hopefully) gets filled on the next call to this
876  * function. We decode a few bytes from the temporary buffer so that we can
877  * continue decoding from the caller-supplied input buffer again.
878  */
lzma2_lzma(struct xz_dec_lzma2 * s,struct xz_buf * b)879 static bool lzma2_lzma(struct xz_dec_lzma2 *s, struct xz_buf *b)
880 {
881 	size_t in_avail;
882 	uint32_t tmp;
883 
884 	in_avail = b->in_size - b->in_pos;
885 	if (s->temp.size > 0 || s->lzma2.compressed == 0) {
886 		tmp = 2 * LZMA_IN_REQUIRED - s->temp.size;
887 		if (tmp > s->lzma2.compressed - s->temp.size)
888 			tmp = s->lzma2.compressed - s->temp.size;
889 		if (tmp > in_avail)
890 			tmp = in_avail;
891 
892 		memcpy(s->temp.buf + s->temp.size, b->in + b->in_pos, tmp);
893 
894 		if (s->temp.size + tmp == s->lzma2.compressed) {
895 			memzero(s->temp.buf + s->temp.size + tmp,
896 					sizeof(s->temp.buf)
897 						- s->temp.size - tmp);
898 			s->rc.in_limit = s->temp.size + tmp;
899 		} else if (s->temp.size + tmp < LZMA_IN_REQUIRED) {
900 			s->temp.size += tmp;
901 			b->in_pos += tmp;
902 			return true;
903 		} else {
904 			s->rc.in_limit = s->temp.size + tmp - LZMA_IN_REQUIRED;
905 		}
906 
907 		s->rc.in = s->temp.buf;
908 		s->rc.in_pos = 0;
909 
910 		if (!lzma_main(s) || s->rc.in_pos > s->temp.size + tmp)
911 			return false;
912 
913 		s->lzma2.compressed -= s->rc.in_pos;
914 
915 		if (s->rc.in_pos < s->temp.size) {
916 			s->temp.size -= s->rc.in_pos;
917 			memmove(s->temp.buf, s->temp.buf + s->rc.in_pos,
918 					s->temp.size);
919 			return true;
920 		}
921 
922 		b->in_pos += s->rc.in_pos - s->temp.size;
923 		s->temp.size = 0;
924 	}
925 
926 	in_avail = b->in_size - b->in_pos;
927 	if (in_avail >= LZMA_IN_REQUIRED) {
928 		s->rc.in = b->in;
929 		s->rc.in_pos = b->in_pos;
930 
931 		if (in_avail >= s->lzma2.compressed + LZMA_IN_REQUIRED)
932 			s->rc.in_limit = b->in_pos + s->lzma2.compressed;
933 		else
934 			s->rc.in_limit = b->in_size - LZMA_IN_REQUIRED;
935 
936 		if (!lzma_main(s))
937 			return false;
938 
939 		in_avail = s->rc.in_pos - b->in_pos;
940 		if (in_avail > s->lzma2.compressed)
941 			return false;
942 
943 		s->lzma2.compressed -= in_avail;
944 		b->in_pos = s->rc.in_pos;
945 	}
946 
947 	in_avail = b->in_size - b->in_pos;
948 	if (in_avail < LZMA_IN_REQUIRED) {
949 		if (in_avail > s->lzma2.compressed)
950 			in_avail = s->lzma2.compressed;
951 
952 		memcpy(s->temp.buf, b->in + b->in_pos, in_avail);
953 		s->temp.size = in_avail;
954 		b->in_pos += in_avail;
955 	}
956 
957 	return true;
958 }
959 
960 /*
961  * Take care of the LZMA2 control layer, and forward the job of actual LZMA
962  * decoding or copying of uncompressed chunks to other functions.
963  */
xz_dec_lzma2_run(struct xz_dec_lzma2 * s,struct xz_buf * b)964 XZ_EXTERN enum xz_ret xz_dec_lzma2_run(struct xz_dec_lzma2 *s,
965 				       struct xz_buf *b)
966 {
967 	uint32_t tmp;
968 
969 	while (b->in_pos < b->in_size || s->lzma2.sequence == SEQ_LZMA_RUN) {
970 		switch (s->lzma2.sequence) {
971 		case SEQ_CONTROL:
972 			/*
973 			 * LZMA2 control byte
974 			 *
975 			 * Exact values:
976 			 *   0x00   End marker
977 			 *   0x01   Dictionary reset followed by
978 			 *          an uncompressed chunk
979 			 *   0x02   Uncompressed chunk (no dictionary reset)
980 			 *
981 			 * Highest three bits (s->control & 0xE0):
982 			 *   0xE0   Dictionary reset, new properties and state
983 			 *          reset, followed by LZMA compressed chunk
984 			 *   0xC0   New properties and state reset, followed
985 			 *          by LZMA compressed chunk (no dictionary
986 			 *          reset)
987 			 *   0xA0   State reset using old properties,
988 			 *          followed by LZMA compressed chunk (no
989 			 *          dictionary reset)
990 			 *   0x80   LZMA chunk (no dictionary or state reset)
991 			 *
992 			 * For LZMA compressed chunks, the lowest five bits
993 			 * (s->control & 1F) are the highest bits of the
994 			 * uncompressed size (bits 16-20).
995 			 *
996 			 * A new LZMA2 stream must begin with a dictionary
997 			 * reset. The first LZMA chunk must set new
998 			 * properties and reset the LZMA state.
999 			 *
1000 			 * Values that don't match anything described above
1001 			 * are invalid and we return XZ_DATA_ERROR.
1002 			 */
1003 			tmp = b->in[b->in_pos++];
1004 
1005 			if (tmp == 0x00)
1006 				return XZ_STREAM_END;
1007 
1008 			if (tmp >= 0xE0 || tmp == 0x01) {
1009 				s->lzma2.need_props = true;
1010 				s->lzma2.need_dict_reset = false;
1011 				dict_reset(&s->dict, b);
1012 			} else if (s->lzma2.need_dict_reset) {
1013 				return XZ_DATA_ERROR;
1014 			}
1015 
1016 			if (tmp >= 0x80) {
1017 				s->lzma2.uncompressed = (tmp & 0x1F) << 16;
1018 				s->lzma2.sequence = SEQ_UNCOMPRESSED_1;
1019 
1020 				if (tmp >= 0xC0) {
1021 					/*
1022 					 * When there are new properties,
1023 					 * state reset is done at
1024 					 * SEQ_PROPERTIES.
1025 					 */
1026 					s->lzma2.need_props = false;
1027 					s->lzma2.next_sequence
1028 							= SEQ_PROPERTIES;
1029 
1030 				} else if (s->lzma2.need_props) {
1031 					return XZ_DATA_ERROR;
1032 
1033 				} else {
1034 					s->lzma2.next_sequence
1035 							= SEQ_LZMA_PREPARE;
1036 					if (tmp >= 0xA0)
1037 						lzma_reset(s);
1038 				}
1039 			} else {
1040 				if (tmp > 0x02)
1041 					return XZ_DATA_ERROR;
1042 
1043 				s->lzma2.sequence = SEQ_COMPRESSED_0;
1044 				s->lzma2.next_sequence = SEQ_COPY;
1045 			}
1046 
1047 			break;
1048 
1049 		case SEQ_UNCOMPRESSED_1:
1050 			s->lzma2.uncompressed
1051 					+= (uint32_t)b->in[b->in_pos++] << 8;
1052 			s->lzma2.sequence = SEQ_UNCOMPRESSED_2;
1053 			break;
1054 
1055 		case SEQ_UNCOMPRESSED_2:
1056 			s->lzma2.uncompressed
1057 					+= (uint32_t)b->in[b->in_pos++] + 1;
1058 			s->lzma2.sequence = SEQ_COMPRESSED_0;
1059 			break;
1060 
1061 		case SEQ_COMPRESSED_0:
1062 			s->lzma2.compressed
1063 					= (uint32_t)b->in[b->in_pos++] << 8;
1064 			s->lzma2.sequence = SEQ_COMPRESSED_1;
1065 			break;
1066 
1067 		case SEQ_COMPRESSED_1:
1068 			s->lzma2.compressed
1069 					+= (uint32_t)b->in[b->in_pos++] + 1;
1070 			s->lzma2.sequence = s->lzma2.next_sequence;
1071 			break;
1072 
1073 		case SEQ_PROPERTIES:
1074 			if (!lzma_props(s, b->in[b->in_pos++]))
1075 				return XZ_DATA_ERROR;
1076 
1077 			s->lzma2.sequence = SEQ_LZMA_PREPARE;
1078 
1079 			fallthrough;
1080 
1081 		case SEQ_LZMA_PREPARE:
1082 			if (s->lzma2.compressed < RC_INIT_BYTES)
1083 				return XZ_DATA_ERROR;
1084 
1085 			if (!rc_read_init(&s->rc, b))
1086 				return XZ_OK;
1087 
1088 			s->lzma2.compressed -= RC_INIT_BYTES;
1089 			s->lzma2.sequence = SEQ_LZMA_RUN;
1090 
1091 			fallthrough;
1092 
1093 		case SEQ_LZMA_RUN:
1094 			/*
1095 			 * Set dictionary limit to indicate how much we want
1096 			 * to be encoded at maximum. Decode new data into the
1097 			 * dictionary. Flush the new data from dictionary to
1098 			 * b->out. Check if we finished decoding this chunk.
1099 			 * In case the dictionary got full but we didn't fill
1100 			 * the output buffer yet, we may run this loop
1101 			 * multiple times without changing s->lzma2.sequence.
1102 			 */
1103 			dict_limit(&s->dict, min_t(size_t,
1104 					b->out_size - b->out_pos,
1105 					s->lzma2.uncompressed));
1106 			if (!lzma2_lzma(s, b))
1107 				return XZ_DATA_ERROR;
1108 
1109 			s->lzma2.uncompressed -= dict_flush(&s->dict, b);
1110 
1111 			if (s->lzma2.uncompressed == 0) {
1112 				if (s->lzma2.compressed > 0 || s->lzma.len > 0
1113 						|| !rc_is_finished(&s->rc))
1114 					return XZ_DATA_ERROR;
1115 
1116 				rc_reset(&s->rc);
1117 				s->lzma2.sequence = SEQ_CONTROL;
1118 
1119 			} else if (b->out_pos == b->out_size
1120 					|| (b->in_pos == b->in_size
1121 						&& s->temp.size
1122 						< s->lzma2.compressed)) {
1123 				return XZ_OK;
1124 			}
1125 
1126 			break;
1127 
1128 		case SEQ_COPY:
1129 			dict_uncompressed(&s->dict, b, &s->lzma2.compressed);
1130 			if (s->lzma2.compressed > 0)
1131 				return XZ_OK;
1132 
1133 			s->lzma2.sequence = SEQ_CONTROL;
1134 			break;
1135 		}
1136 	}
1137 
1138 	return XZ_OK;
1139 }
1140 
xz_dec_lzma2_create(enum xz_mode mode,uint32_t dict_max)1141 XZ_EXTERN struct xz_dec_lzma2 *xz_dec_lzma2_create(enum xz_mode mode,
1142 						   uint32_t dict_max)
1143 {
1144 	struct xz_dec_lzma2 *s = kmalloc(sizeof(*s), GFP_KERNEL);
1145 	if (s == NULL)
1146 		return NULL;
1147 
1148 	s->dict.mode = mode;
1149 	s->dict.size_max = dict_max;
1150 
1151 	if (DEC_IS_PREALLOC(mode)) {
1152 		s->dict.buf = vmalloc(dict_max);
1153 		if (s->dict.buf == NULL) {
1154 			kfree(s);
1155 			return NULL;
1156 		}
1157 	} else if (DEC_IS_DYNALLOC(mode)) {
1158 		s->dict.buf = NULL;
1159 		s->dict.allocated = 0;
1160 	}
1161 
1162 	return s;
1163 }
1164 
xz_dec_lzma2_reset(struct xz_dec_lzma2 * s,uint8_t props)1165 XZ_EXTERN enum xz_ret xz_dec_lzma2_reset(struct xz_dec_lzma2 *s, uint8_t props)
1166 {
1167 	/* This limits dictionary size to 3 GiB to keep parsing simpler. */
1168 	if (props > 39)
1169 		return XZ_OPTIONS_ERROR;
1170 
1171 	s->dict.size = 2 + (props & 1);
1172 	s->dict.size <<= (props >> 1) + 11;
1173 
1174 	if (DEC_IS_MULTI(s->dict.mode)) {
1175 		if (s->dict.size > s->dict.size_max)
1176 			return XZ_MEMLIMIT_ERROR;
1177 
1178 		s->dict.end = s->dict.size;
1179 
1180 		if (DEC_IS_DYNALLOC(s->dict.mode)) {
1181 			if (s->dict.allocated < s->dict.size) {
1182 				s->dict.allocated = s->dict.size;
1183 				vfree(s->dict.buf);
1184 				s->dict.buf = vmalloc(s->dict.size);
1185 				if (s->dict.buf == NULL) {
1186 					s->dict.allocated = 0;
1187 					return XZ_MEM_ERROR;
1188 				}
1189 			}
1190 		}
1191 	}
1192 
1193 	s->lzma2.sequence = SEQ_CONTROL;
1194 	s->lzma2.need_dict_reset = true;
1195 
1196 	s->temp.size = 0;
1197 
1198 	return XZ_OK;
1199 }
1200 
xz_dec_lzma2_end(struct xz_dec_lzma2 * s)1201 XZ_EXTERN void xz_dec_lzma2_end(struct xz_dec_lzma2 *s)
1202 {
1203 	if (DEC_IS_MULTI(s->dict.mode))
1204 		vfree(s->dict.buf);
1205 
1206 	kfree(s);
1207 }
1208 
1209 #ifdef XZ_DEC_MICROLZMA
1210 /* This is a wrapper struct to have a nice struct name in the public API. */
1211 struct xz_dec_microlzma {
1212 	struct xz_dec_lzma2 s;
1213 };
1214 
xz_dec_microlzma_run(struct xz_dec_microlzma * s_ptr,struct xz_buf * b)1215 enum xz_ret xz_dec_microlzma_run(struct xz_dec_microlzma *s_ptr,
1216 				 struct xz_buf *b)
1217 {
1218 	struct xz_dec_lzma2 *s = &s_ptr->s;
1219 
1220 	/*
1221 	 * sequence is SEQ_PROPERTIES before the first input byte,
1222 	 * SEQ_LZMA_PREPARE until a total of five bytes have been read,
1223 	 * and SEQ_LZMA_RUN for the rest of the input stream.
1224 	 */
1225 	if (s->lzma2.sequence != SEQ_LZMA_RUN) {
1226 		if (s->lzma2.sequence == SEQ_PROPERTIES) {
1227 			/* One byte is needed for the props. */
1228 			if (b->in_pos >= b->in_size)
1229 				return XZ_OK;
1230 
1231 			/*
1232 			 * Don't increment b->in_pos here. The same byte is
1233 			 * also passed to rc_read_init() which will ignore it.
1234 			 */
1235 			if (!lzma_props(s, ~b->in[b->in_pos]))
1236 				return XZ_DATA_ERROR;
1237 
1238 			s->lzma2.sequence = SEQ_LZMA_PREPARE;
1239 		}
1240 
1241 		/*
1242 		 * xz_dec_microlzma_reset() doesn't validate the compressed
1243 		 * size so we do it here. We have to limit the maximum size
1244 		 * to avoid integer overflows in lzma2_lzma(). 3 GiB is a nice
1245 		 * round number and much more than users of this code should
1246 		 * ever need.
1247 		 */
1248 		if (s->lzma2.compressed < RC_INIT_BYTES
1249 				|| s->lzma2.compressed > (3U << 30))
1250 			return XZ_DATA_ERROR;
1251 
1252 		if (!rc_read_init(&s->rc, b))
1253 			return XZ_OK;
1254 
1255 		s->lzma2.compressed -= RC_INIT_BYTES;
1256 		s->lzma2.sequence = SEQ_LZMA_RUN;
1257 
1258 		dict_reset(&s->dict, b);
1259 	}
1260 
1261 	/* This is to allow increasing b->out_size between calls. */
1262 	if (DEC_IS_SINGLE(s->dict.mode))
1263 		s->dict.end = b->out_size - b->out_pos;
1264 
1265 	while (true) {
1266 		dict_limit(&s->dict, min_t(size_t, b->out_size - b->out_pos,
1267 					   s->lzma2.uncompressed));
1268 
1269 		if (!lzma2_lzma(s, b))
1270 			return XZ_DATA_ERROR;
1271 
1272 		s->lzma2.uncompressed -= dict_flush(&s->dict, b);
1273 
1274 		if (s->lzma2.uncompressed == 0) {
1275 			if (s->lzma2.pedantic_microlzma) {
1276 				if (s->lzma2.compressed > 0 || s->lzma.len > 0
1277 						|| !rc_is_finished(&s->rc))
1278 					return XZ_DATA_ERROR;
1279 			}
1280 
1281 			return XZ_STREAM_END;
1282 		}
1283 
1284 		if (b->out_pos == b->out_size)
1285 			return XZ_OK;
1286 
1287 		if (b->in_pos == b->in_size
1288 				&& s->temp.size < s->lzma2.compressed)
1289 			return XZ_OK;
1290 	}
1291 }
1292 
xz_dec_microlzma_alloc(enum xz_mode mode,uint32_t dict_size)1293 struct xz_dec_microlzma *xz_dec_microlzma_alloc(enum xz_mode mode,
1294 						uint32_t dict_size)
1295 {
1296 	struct xz_dec_microlzma *s;
1297 
1298 	/* Restrict dict_size to the same range as in the LZMA2 code. */
1299 	if (dict_size < 4096 || dict_size > (3U << 30))
1300 		return NULL;
1301 
1302 	s = kmalloc(sizeof(*s), GFP_KERNEL);
1303 	if (s == NULL)
1304 		return NULL;
1305 
1306 	s->s.dict.mode = mode;
1307 	s->s.dict.size = dict_size;
1308 
1309 	if (DEC_IS_MULTI(mode)) {
1310 		s->s.dict.end = dict_size;
1311 
1312 		s->s.dict.buf = vmalloc(dict_size);
1313 		if (s->s.dict.buf == NULL) {
1314 			kfree(s);
1315 			return NULL;
1316 		}
1317 	}
1318 
1319 	return s;
1320 }
1321 
xz_dec_microlzma_reset(struct xz_dec_microlzma * s,uint32_t comp_size,uint32_t uncomp_size,int uncomp_size_is_exact)1322 void xz_dec_microlzma_reset(struct xz_dec_microlzma *s, uint32_t comp_size,
1323 			    uint32_t uncomp_size, int uncomp_size_is_exact)
1324 {
1325 	/*
1326 	 * comp_size is validated in xz_dec_microlzma_run().
1327 	 * uncomp_size can safely be anything.
1328 	 */
1329 	s->s.lzma2.compressed = comp_size;
1330 	s->s.lzma2.uncompressed = uncomp_size;
1331 	s->s.lzma2.pedantic_microlzma = uncomp_size_is_exact;
1332 
1333 	s->s.lzma2.sequence = SEQ_PROPERTIES;
1334 	s->s.temp.size = 0;
1335 }
1336 
xz_dec_microlzma_end(struct xz_dec_microlzma * s)1337 void xz_dec_microlzma_end(struct xz_dec_microlzma *s)
1338 {
1339 	if (DEC_IS_MULTI(s->s.dict.mode))
1340 		vfree(s->s.dict.buf);
1341 
1342 	kfree(s);
1343 }
1344 #endif
1345