1/*
2 * LZMA2 decoder
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#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 */
44struct 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 */
95struct 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. */
115struct 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
132struct 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 dervied from the number of literal
151	 * position bits, and a mask dervied 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
214struct 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
253struct xz_dec_lzma2 {
254	/*
255	 * The order below is important on x86 to reduce code size and
256	 * it shouldn't hurt on other platforms. Everything up to and
257	 * including lzma.pos_mask are in the first 128 bytes on x86-32,
258	 * which allows using smaller instructions to access those
259	 * variables. On x86-64, fewer variables fit into the first 128
260	 * bytes, but this is still the best order without sacrificing
261	 * the readability by splitting the structures.
262	 */
263	struct rc_dec rc;
264	struct dictionary dict;
265	struct lzma2_dec lzma2;
266	struct lzma_dec lzma;
267
268	/*
269	 * Temporary buffer which holds small number of input bytes between
270	 * decoder calls. See lzma2_lzma() for details.
271	 */
272	struct {
273		uint32_t size;
274		uint8_t buf[3 * LZMA_IN_REQUIRED];
275	} temp;
276};
277
278/**************
279 * Dictionary *
280 **************/
281
282/*
283 * Reset the dictionary state. When in single-call mode, set up the beginning
284 * of the dictionary to point to the actual output buffer.
285 */
286static void dict_reset(struct dictionary *dict, struct xz_buf *b)
287{
288	if (DEC_IS_SINGLE(dict->mode)) {
289		dict->buf = b->out + b->out_pos;
290		dict->end = b->out_size - b->out_pos;
291	}
292
293	dict->start = 0;
294	dict->pos = 0;
295	dict->limit = 0;
296	dict->full = 0;
297}
298
299/* Set dictionary write limit */
300static void dict_limit(struct dictionary *dict, size_t out_max)
301{
302	if (dict->end - dict->pos <= out_max)
303		dict->limit = dict->end;
304	else
305		dict->limit = dict->pos + out_max;
306}
307
308/* Return true if at least one byte can be written into the dictionary. */
309static inline bool dict_has_space(const struct dictionary *dict)
310{
311	return dict->pos < dict->limit;
312}
313
314/*
315 * Get a byte from the dictionary at the given distance. The distance is
316 * assumed to valid, or as a special case, zero when the dictionary is
317 * still empty. This special case is needed for single-call decoding to
318 * avoid writing a '\0' to the end of the destination buffer.
319 */
320static inline uint32_t dict_get(const struct dictionary *dict, uint32_t dist)
321{
322	size_t offset = dict->pos - dist - 1;
323
324	if (dist >= dict->pos)
325		offset += dict->end;
326
327	return dict->full > 0 ? dict->buf[offset] : 0;
328}
329
330/*
331 * Put one byte into the dictionary. It is assumed that there is space for it.
332 */
333static inline void dict_put(struct dictionary *dict, uint8_t byte)
334{
335	dict->buf[dict->pos++] = byte;
336
337	if (dict->full < dict->pos)
338		dict->full = dict->pos;
339}
340
341/*
342 * Repeat given number of bytes from the given distance. If the distance is
343 * invalid, false is returned. On success, true is returned and *len is
344 * updated to indicate how many bytes were left to be repeated.
345 */
346static bool dict_repeat(struct dictionary *dict, uint32_t *len, uint32_t dist)
347{
348	size_t back;
349	uint32_t left;
350
351	if (dist >= dict->full || dist >= dict->size)
352		return false;
353
354	left = min_t(size_t, dict->limit - dict->pos, *len);
355	*len -= left;
356
357	back = dict->pos - dist - 1;
358	if (dist >= dict->pos)
359		back += dict->end;
360
361	do {
362		dict->buf[dict->pos++] = dict->buf[back++];
363		if (back == dict->end)
364			back = 0;
365	} while (--left > 0);
366
367	if (dict->full < dict->pos)
368		dict->full = dict->pos;
369
370	return true;
371}
372
373/* Copy uncompressed data as is from input to dictionary and output buffers. */
374static void dict_uncompressed(struct dictionary *dict, struct xz_buf *b,
375			      uint32_t *left)
376{
377	size_t copy_size;
378
379	while (*left > 0 && b->in_pos < b->in_size
380			&& b->out_pos < b->out_size) {
381		copy_size = min(b->in_size - b->in_pos,
382				b->out_size - b->out_pos);
383		if (copy_size > dict->end - dict->pos)
384			copy_size = dict->end - dict->pos;
385		if (copy_size > *left)
386			copy_size = *left;
387
388		*left -= copy_size;
389
390		memcpy(dict->buf + dict->pos, b->in + b->in_pos, copy_size);
391		dict->pos += copy_size;
392
393		if (dict->full < dict->pos)
394			dict->full = dict->pos;
395
396		if (DEC_IS_MULTI(dict->mode)) {
397			if (dict->pos == dict->end)
398				dict->pos = 0;
399
400			memcpy(b->out + b->out_pos, b->in + b->in_pos,
401					copy_size);
402		}
403
404		dict->start = dict->pos;
405
406		b->out_pos += copy_size;
407		b->in_pos += copy_size;
408	}
409}
410
411/*
412 * Flush pending data from dictionary to b->out. It is assumed that there is
413 * enough space in b->out. This is guaranteed because caller uses dict_limit()
414 * before decoding data into the dictionary.
415 */
416static uint32_t dict_flush(struct dictionary *dict, struct xz_buf *b)
417{
418	size_t copy_size = dict->pos - dict->start;
419
420	if (DEC_IS_MULTI(dict->mode)) {
421		if (dict->pos == dict->end)
422			dict->pos = 0;
423
424		memcpy(b->out + b->out_pos, dict->buf + dict->start,
425				copy_size);
426	}
427
428	dict->start = dict->pos;
429	b->out_pos += copy_size;
430	return copy_size;
431}
432
433/*****************
434 * Range decoder *
435 *****************/
436
437/* Reset the range decoder. */
438static void rc_reset(struct rc_dec *rc)
439{
440	rc->range = (uint32_t)-1;
441	rc->code = 0;
442	rc->init_bytes_left = RC_INIT_BYTES;
443}
444
445/*
446 * Read the first five initial bytes into rc->code if they haven't been
447 * read already. (Yes, the first byte gets completely ignored.)
448 */
449static bool rc_read_init(struct rc_dec *rc, struct xz_buf *b)
450{
451	while (rc->init_bytes_left > 0) {
452		if (b->in_pos == b->in_size)
453			return false;
454
455		rc->code = (rc->code << 8) + b->in[b->in_pos++];
456		--rc->init_bytes_left;
457	}
458
459	return true;
460}
461
462/* Return true if there may not be enough input for the next decoding loop. */
463static inline bool rc_limit_exceeded(const struct rc_dec *rc)
464{
465	return rc->in_pos > rc->in_limit;
466}
467
468/*
469 * Return true if it is possible (from point of view of range decoder) that
470 * we have reached the end of the LZMA chunk.
471 */
472static inline bool rc_is_finished(const struct rc_dec *rc)
473{
474	return rc->code == 0;
475}
476
477/* Read the next input byte if needed. */
478static __always_inline void rc_normalize(struct rc_dec *rc)
479{
480	if (rc->range < RC_TOP_VALUE) {
481		rc->range <<= RC_SHIFT_BITS;
482		rc->code = (rc->code << RC_SHIFT_BITS) + rc->in[rc->in_pos++];
483	}
484}
485
486/*
487 * Decode one bit. In some versions, this function has been splitted in three
488 * functions so that the compiler is supposed to be able to more easily avoid
489 * an extra branch. In this particular version of the LZMA decoder, this
490 * doesn't seem to be a good idea (tested with GCC 3.3.6, 3.4.6, and 4.3.3
491 * on x86). Using a non-splitted version results in nicer looking code too.
492 *
493 * NOTE: This must return an int. Do not make it return a bool or the speed
494 * of the code generated by GCC 3.x decreases 10-15 %. (GCC 4.3 doesn't care,
495 * and it generates 10-20 % faster code than GCC 3.x from this file anyway.)
496 */
497static __always_inline int rc_bit(struct rc_dec *rc, uint16_t *prob)
498{
499	uint32_t bound;
500	int bit;
501
502	rc_normalize(rc);
503	bound = (rc->range >> RC_BIT_MODEL_TOTAL_BITS) * *prob;
504	if (rc->code < bound) {
505		rc->range = bound;
506		*prob += (RC_BIT_MODEL_TOTAL - *prob) >> RC_MOVE_BITS;
507		bit = 0;
508	} else {
509		rc->range -= bound;
510		rc->code -= bound;
511		*prob -= *prob >> RC_MOVE_BITS;
512		bit = 1;
513	}
514
515	return bit;
516}
517
518/* Decode a bittree starting from the most significant bit. */
519static __always_inline uint32_t rc_bittree(struct rc_dec *rc,
520					   uint16_t *probs, uint32_t limit)
521{
522	uint32_t symbol = 1;
523
524	do {
525		if (rc_bit(rc, &probs[symbol]))
526			symbol = (symbol << 1) + 1;
527		else
528			symbol <<= 1;
529	} while (symbol < limit);
530
531	return symbol;
532}
533
534/* Decode a bittree starting from the least significant bit. */
535static __always_inline void rc_bittree_reverse(struct rc_dec *rc,
536					       uint16_t *probs,
537					       uint32_t *dest, uint32_t limit)
538{
539	uint32_t symbol = 1;
540	uint32_t i = 0;
541
542	do {
543		if (rc_bit(rc, &probs[symbol])) {
544			symbol = (symbol << 1) + 1;
545			*dest += 1 << i;
546		} else {
547			symbol <<= 1;
548		}
549	} while (++i < limit);
550}
551
552/* Decode direct bits (fixed fifty-fifty probability) */
553static inline void rc_direct(struct rc_dec *rc, uint32_t *dest, uint32_t limit)
554{
555	uint32_t mask;
556
557	do {
558		rc_normalize(rc);
559		rc->range >>= 1;
560		rc->code -= rc->range;
561		mask = (uint32_t)0 - (rc->code >> 31);
562		rc->code += rc->range & mask;
563		*dest = (*dest << 1) + (mask + 1);
564	} while (--limit > 0);
565}
566
567/********
568 * LZMA *
569 ********/
570
571/* Get pointer to literal coder probability array. */
572static uint16_t *lzma_literal_probs(struct xz_dec_lzma2 *s)
573{
574	uint32_t prev_byte = dict_get(&s->dict, 0);
575	uint32_t low = prev_byte >> (8 - s->lzma.lc);
576	uint32_t high = (s->dict.pos & s->lzma.literal_pos_mask) << s->lzma.lc;
577	return s->lzma.literal[low + high];
578}
579
580/* Decode a literal (one 8-bit byte) */
581static void lzma_literal(struct xz_dec_lzma2 *s)
582{
583	uint16_t *probs;
584	uint32_t symbol;
585	uint32_t match_byte;
586	uint32_t match_bit;
587	uint32_t offset;
588	uint32_t i;
589
590	probs = lzma_literal_probs(s);
591
592	if (lzma_state_is_literal(s->lzma.state)) {
593		symbol = rc_bittree(&s->rc, probs, 0x100);
594	} else {
595		symbol = 1;
596		match_byte = dict_get(&s->dict, s->lzma.rep0) << 1;
597		offset = 0x100;
598
599		do {
600			match_bit = match_byte & offset;
601			match_byte <<= 1;
602			i = offset + match_bit + symbol;
603
604			if (rc_bit(&s->rc, &probs[i])) {
605				symbol = (symbol << 1) + 1;
606				offset &= match_bit;
607			} else {
608				symbol <<= 1;
609				offset &= ~match_bit;
610			}
611		} while (symbol < 0x100);
612	}
613
614	dict_put(&s->dict, (uint8_t)symbol);
615	lzma_state_literal(&s->lzma.state);
616}
617
618/* Decode the length of the match into s->lzma.len. */
619static void lzma_len(struct xz_dec_lzma2 *s, struct lzma_len_dec *l,
620		     uint32_t pos_state)
621{
622	uint16_t *probs;
623	uint32_t limit;
624
625	if (!rc_bit(&s->rc, &l->choice)) {
626		probs = l->low[pos_state];
627		limit = LEN_LOW_SYMBOLS;
628		s->lzma.len = MATCH_LEN_MIN;
629	} else {
630		if (!rc_bit(&s->rc, &l->choice2)) {
631			probs = l->mid[pos_state];
632			limit = LEN_MID_SYMBOLS;
633			s->lzma.len = MATCH_LEN_MIN + LEN_LOW_SYMBOLS;
634		} else {
635			probs = l->high;
636			limit = LEN_HIGH_SYMBOLS;
637			s->lzma.len = MATCH_LEN_MIN + LEN_LOW_SYMBOLS
638					+ LEN_MID_SYMBOLS;
639		}
640	}
641
642	s->lzma.len += rc_bittree(&s->rc, probs, limit) - limit;
643}
644
645/* Decode a match. The distance will be stored in s->lzma.rep0. */
646static void lzma_match(struct xz_dec_lzma2 *s, uint32_t pos_state)
647{
648	uint16_t *probs;
649	uint32_t dist_slot;
650	uint32_t limit;
651
652	lzma_state_match(&s->lzma.state);
653
654	s->lzma.rep3 = s->lzma.rep2;
655	s->lzma.rep2 = s->lzma.rep1;
656	s->lzma.rep1 = s->lzma.rep0;
657
658	lzma_len(s, &s->lzma.match_len_dec, pos_state);
659
660	probs = s->lzma.dist_slot[lzma_get_dist_state(s->lzma.len)];
661	dist_slot = rc_bittree(&s->rc, probs, DIST_SLOTS) - DIST_SLOTS;
662
663	if (dist_slot < DIST_MODEL_START) {
664		s->lzma.rep0 = dist_slot;
665	} else {
666		limit = (dist_slot >> 1) - 1;
667		s->lzma.rep0 = 2 + (dist_slot & 1);
668
669		if (dist_slot < DIST_MODEL_END) {
670			s->lzma.rep0 <<= limit;
671			probs = s->lzma.dist_special + s->lzma.rep0
672					- dist_slot - 1;
673			rc_bittree_reverse(&s->rc, probs,
674					&s->lzma.rep0, limit);
675		} else {
676			rc_direct(&s->rc, &s->lzma.rep0, limit - ALIGN_BITS);
677			s->lzma.rep0 <<= ALIGN_BITS;
678			rc_bittree_reverse(&s->rc, s->lzma.dist_align,
679					&s->lzma.rep0, ALIGN_BITS);
680		}
681	}
682}
683
684/*
685 * Decode a repeated match. The distance is one of the four most recently
686 * seen matches. The distance will be stored in s->lzma.rep0.
687 */
688static void lzma_rep_match(struct xz_dec_lzma2 *s, uint32_t pos_state)
689{
690	uint32_t tmp;
691
692	if (!rc_bit(&s->rc, &s->lzma.is_rep0[s->lzma.state])) {
693		if (!rc_bit(&s->rc, &s->lzma.is_rep0_long[
694				s->lzma.state][pos_state])) {
695			lzma_state_short_rep(&s->lzma.state);
696			s->lzma.len = 1;
697			return;
698		}
699	} else {
700		if (!rc_bit(&s->rc, &s->lzma.is_rep1[s->lzma.state])) {
701			tmp = s->lzma.rep1;
702		} else {
703			if (!rc_bit(&s->rc, &s->lzma.is_rep2[s->lzma.state])) {
704				tmp = s->lzma.rep2;
705			} else {
706				tmp = s->lzma.rep3;
707				s->lzma.rep3 = s->lzma.rep2;
708			}
709
710			s->lzma.rep2 = s->lzma.rep1;
711		}
712
713		s->lzma.rep1 = s->lzma.rep0;
714		s->lzma.rep0 = tmp;
715	}
716
717	lzma_state_long_rep(&s->lzma.state);
718	lzma_len(s, &s->lzma.rep_len_dec, pos_state);
719}
720
721/* LZMA decoder core */
722static bool lzma_main(struct xz_dec_lzma2 *s)
723{
724	uint32_t pos_state;
725
726	/*
727	 * If the dictionary was reached during the previous call, try to
728	 * finish the possibly pending repeat in the dictionary.
729	 */
730	if (dict_has_space(&s->dict) && s->lzma.len > 0)
731		dict_repeat(&s->dict, &s->lzma.len, s->lzma.rep0);
732
733	/*
734	 * Decode more LZMA symbols. One iteration may consume up to
735	 * LZMA_IN_REQUIRED - 1 bytes.
736	 */
737	while (dict_has_space(&s->dict) && !rc_limit_exceeded(&s->rc)) {
738		pos_state = s->dict.pos & s->lzma.pos_mask;
739
740		if (!rc_bit(&s->rc, &s->lzma.is_match[
741				s->lzma.state][pos_state])) {
742			lzma_literal(s);
743		} else {
744			if (rc_bit(&s->rc, &s->lzma.is_rep[s->lzma.state]))
745				lzma_rep_match(s, pos_state);
746			else
747				lzma_match(s, pos_state);
748
749			if (!dict_repeat(&s->dict, &s->lzma.len, s->lzma.rep0))
750				return false;
751		}
752	}
753
754	/*
755	 * Having the range decoder always normalized when we are outside
756	 * this function makes it easier to correctly handle end of the chunk.
757	 */
758	rc_normalize(&s->rc);
759
760	return true;
761}
762
763/*
764 * Reset the LZMA decoder and range decoder state. Dictionary is nore reset
765 * here, because LZMA state may be reset without resetting the dictionary.
766 */
767static void lzma_reset(struct xz_dec_lzma2 *s)
768{
769	uint16_t *probs;
770	size_t i;
771
772	s->lzma.state = STATE_LIT_LIT;
773	s->lzma.rep0 = 0;
774	s->lzma.rep1 = 0;
775	s->lzma.rep2 = 0;
776	s->lzma.rep3 = 0;
777
778	/*
779	 * All probabilities are initialized to the same value. This hack
780	 * makes the code smaller by avoiding a separate loop for each
781	 * probability array.
782	 *
783	 * This could be optimized so that only that part of literal
784	 * probabilities that are actually required. In the common case
785	 * we would write 12 KiB less.
786	 */
787	probs = s->lzma.is_match[0];
788	for (i = 0; i < PROBS_TOTAL; ++i)
789		probs[i] = RC_BIT_MODEL_TOTAL / 2;
790
791	rc_reset(&s->rc);
792}
793
794/*
795 * Decode and validate LZMA properties (lc/lp/pb) and calculate the bit masks
796 * from the decoded lp and pb values. On success, the LZMA decoder state is
797 * reset and true is returned.
798 */
799static bool lzma_props(struct xz_dec_lzma2 *s, uint8_t props)
800{
801	if (props > (4 * 5 + 4) * 9 + 8)
802		return false;
803
804	s->lzma.pos_mask = 0;
805	while (props >= 9 * 5) {
806		props -= 9 * 5;
807		++s->lzma.pos_mask;
808	}
809
810	s->lzma.pos_mask = (1 << s->lzma.pos_mask) - 1;
811
812	s->lzma.literal_pos_mask = 0;
813	while (props >= 9) {
814		props -= 9;
815		++s->lzma.literal_pos_mask;
816	}
817
818	s->lzma.lc = props;
819
820	if (s->lzma.lc + s->lzma.literal_pos_mask > 4)
821		return false;
822
823	s->lzma.literal_pos_mask = (1 << s->lzma.literal_pos_mask) - 1;
824
825	lzma_reset(s);
826
827	return true;
828}
829
830/*********
831 * LZMA2 *
832 *********/
833
834/*
835 * The LZMA decoder assumes that if the input limit (s->rc.in_limit) hasn't
836 * been exceeded, it is safe to read up to LZMA_IN_REQUIRED bytes. This
837 * wrapper function takes care of making the LZMA decoder's assumption safe.
838 *
839 * As long as there is plenty of input left to be decoded in the current LZMA
840 * chunk, we decode directly from the caller-supplied input buffer until
841 * there's LZMA_IN_REQUIRED bytes left. Those remaining bytes are copied into
842 * s->temp.buf, which (hopefully) gets filled on the next call to this
843 * function. We decode a few bytes from the temporary buffer so that we can
844 * continue decoding from the caller-supplied input buffer again.
845 */
846static bool lzma2_lzma(struct xz_dec_lzma2 *s, struct xz_buf *b)
847{
848	size_t in_avail;
849	uint32_t tmp;
850
851	in_avail = b->in_size - b->in_pos;
852	if (s->temp.size > 0 || s->lzma2.compressed == 0) {
853		tmp = 2 * LZMA_IN_REQUIRED - s->temp.size;
854		if (tmp > s->lzma2.compressed - s->temp.size)
855			tmp = s->lzma2.compressed - s->temp.size;
856		if (tmp > in_avail)
857			tmp = in_avail;
858
859		memcpy(s->temp.buf + s->temp.size, b->in + b->in_pos, tmp);
860
861		if (s->temp.size + tmp == s->lzma2.compressed) {
862			memzero(s->temp.buf + s->temp.size + tmp,
863					sizeof(s->temp.buf)
864						- s->temp.size - tmp);
865			s->rc.in_limit = s->temp.size + tmp;
866		} else if (s->temp.size + tmp < LZMA_IN_REQUIRED) {
867			s->temp.size += tmp;
868			b->in_pos += tmp;
869			return true;
870		} else {
871			s->rc.in_limit = s->temp.size + tmp - LZMA_IN_REQUIRED;
872		}
873
874		s->rc.in = s->temp.buf;
875		s->rc.in_pos = 0;
876
877		if (!lzma_main(s) || s->rc.in_pos > s->temp.size + tmp)
878			return false;
879
880		s->lzma2.compressed -= s->rc.in_pos;
881
882		if (s->rc.in_pos < s->temp.size) {
883			s->temp.size -= s->rc.in_pos;
884			memmove(s->temp.buf, s->temp.buf + s->rc.in_pos,
885					s->temp.size);
886			return true;
887		}
888
889		b->in_pos += s->rc.in_pos - s->temp.size;
890		s->temp.size = 0;
891	}
892
893	in_avail = b->in_size - b->in_pos;
894	if (in_avail >= LZMA_IN_REQUIRED) {
895		s->rc.in = b->in;
896		s->rc.in_pos = b->in_pos;
897
898		if (in_avail >= s->lzma2.compressed + LZMA_IN_REQUIRED)
899			s->rc.in_limit = b->in_pos + s->lzma2.compressed;
900		else
901			s->rc.in_limit = b->in_size - LZMA_IN_REQUIRED;
902
903		if (!lzma_main(s))
904			return false;
905
906		in_avail = s->rc.in_pos - b->in_pos;
907		if (in_avail > s->lzma2.compressed)
908			return false;
909
910		s->lzma2.compressed -= in_avail;
911		b->in_pos = s->rc.in_pos;
912	}
913
914	in_avail = b->in_size - b->in_pos;
915	if (in_avail < LZMA_IN_REQUIRED) {
916		if (in_avail > s->lzma2.compressed)
917			in_avail = s->lzma2.compressed;
918
919		memcpy(s->temp.buf, b->in + b->in_pos, in_avail);
920		s->temp.size = in_avail;
921		b->in_pos += in_avail;
922	}
923
924	return true;
925}
926
927/*
928 * Take care of the LZMA2 control layer, and forward the job of actual LZMA
929 * decoding or copying of uncompressed chunks to other functions.
930 */
931XZ_EXTERN enum xz_ret xz_dec_lzma2_run(struct xz_dec_lzma2 *s,
932				       struct xz_buf *b)
933{
934	uint32_t tmp;
935
936	while (b->in_pos < b->in_size || s->lzma2.sequence == SEQ_LZMA_RUN) {
937		switch (s->lzma2.sequence) {
938		case SEQ_CONTROL:
939			/*
940			 * LZMA2 control byte
941			 *
942			 * Exact values:
943			 *   0x00   End marker
944			 *   0x01   Dictionary reset followed by
945			 *          an uncompressed chunk
946			 *   0x02   Uncompressed chunk (no dictionary reset)
947			 *
948			 * Highest three bits (s->control & 0xE0):
949			 *   0xE0   Dictionary reset, new properties and state
950			 *          reset, followed by LZMA compressed chunk
951			 *   0xC0   New properties and state reset, followed
952			 *          by LZMA compressed chunk (no dictionary
953			 *          reset)
954			 *   0xA0   State reset using old properties,
955			 *          followed by LZMA compressed chunk (no
956			 *          dictionary reset)
957			 *   0x80   LZMA chunk (no dictionary or state reset)
958			 *
959			 * For LZMA compressed chunks, the lowest five bits
960			 * (s->control & 1F) are the highest bits of the
961			 * uncompressed size (bits 16-20).
962			 *
963			 * A new LZMA2 stream must begin with a dictionary
964			 * reset. The first LZMA chunk must set new
965			 * properties and reset the LZMA state.
966			 *
967			 * Values that don't match anything described above
968			 * are invalid and we return XZ_DATA_ERROR.
969			 */
970			tmp = b->in[b->in_pos++];
971
972			if (tmp == 0x00)
973				return XZ_STREAM_END;
974
975			if (tmp >= 0xE0 || tmp == 0x01) {
976				s->lzma2.need_props = true;
977				s->lzma2.need_dict_reset = false;
978				dict_reset(&s->dict, b);
979			} else if (s->lzma2.need_dict_reset) {
980				return XZ_DATA_ERROR;
981			}
982
983			if (tmp >= 0x80) {
984				s->lzma2.uncompressed = (tmp & 0x1F) << 16;
985				s->lzma2.sequence = SEQ_UNCOMPRESSED_1;
986
987				if (tmp >= 0xC0) {
988					/*
989					 * When there are new properties,
990					 * state reset is done at
991					 * SEQ_PROPERTIES.
992					 */
993					s->lzma2.need_props = false;
994					s->lzma2.next_sequence
995							= SEQ_PROPERTIES;
996
997				} else if (s->lzma2.need_props) {
998					return XZ_DATA_ERROR;
999
1000				} else {
1001					s->lzma2.next_sequence
1002							= SEQ_LZMA_PREPARE;
1003					if (tmp >= 0xA0)
1004						lzma_reset(s);
1005				}
1006			} else {
1007				if (tmp > 0x02)
1008					return XZ_DATA_ERROR;
1009
1010				s->lzma2.sequence = SEQ_COMPRESSED_0;
1011				s->lzma2.next_sequence = SEQ_COPY;
1012			}
1013
1014			break;
1015
1016		case SEQ_UNCOMPRESSED_1:
1017			s->lzma2.uncompressed
1018					+= (uint32_t)b->in[b->in_pos++] << 8;
1019			s->lzma2.sequence = SEQ_UNCOMPRESSED_2;
1020			break;
1021
1022		case SEQ_UNCOMPRESSED_2:
1023			s->lzma2.uncompressed
1024					+= (uint32_t)b->in[b->in_pos++] + 1;
1025			s->lzma2.sequence = SEQ_COMPRESSED_0;
1026			break;
1027
1028		case SEQ_COMPRESSED_0:
1029			s->lzma2.compressed
1030					= (uint32_t)b->in[b->in_pos++] << 8;
1031			s->lzma2.sequence = SEQ_COMPRESSED_1;
1032			break;
1033
1034		case SEQ_COMPRESSED_1:
1035			s->lzma2.compressed
1036					+= (uint32_t)b->in[b->in_pos++] + 1;
1037			s->lzma2.sequence = s->lzma2.next_sequence;
1038			break;
1039
1040		case SEQ_PROPERTIES:
1041			if (!lzma_props(s, b->in[b->in_pos++]))
1042				return XZ_DATA_ERROR;
1043
1044			s->lzma2.sequence = SEQ_LZMA_PREPARE;
1045
1046		/* Fall through */
1047
1048		case SEQ_LZMA_PREPARE:
1049			if (s->lzma2.compressed < RC_INIT_BYTES)
1050				return XZ_DATA_ERROR;
1051
1052			if (!rc_read_init(&s->rc, b))
1053				return XZ_OK;
1054
1055			s->lzma2.compressed -= RC_INIT_BYTES;
1056			s->lzma2.sequence = SEQ_LZMA_RUN;
1057
1058		/* Fall through */
1059
1060		case SEQ_LZMA_RUN:
1061			/*
1062			 * Set dictionary limit to indicate how much we want
1063			 * to be encoded at maximum. Decode new data into the
1064			 * dictionary. Flush the new data from dictionary to
1065			 * b->out. Check if we finished decoding this chunk.
1066			 * In case the dictionary got full but we didn't fill
1067			 * the output buffer yet, we may run this loop
1068			 * multiple times without changing s->lzma2.sequence.
1069			 */
1070			dict_limit(&s->dict, min_t(size_t,
1071					b->out_size - b->out_pos,
1072					s->lzma2.uncompressed));
1073			if (!lzma2_lzma(s, b))
1074				return XZ_DATA_ERROR;
1075
1076			s->lzma2.uncompressed -= dict_flush(&s->dict, b);
1077
1078			if (s->lzma2.uncompressed == 0) {
1079				if (s->lzma2.compressed > 0 || s->lzma.len > 0
1080						|| !rc_is_finished(&s->rc))
1081					return XZ_DATA_ERROR;
1082
1083				rc_reset(&s->rc);
1084				s->lzma2.sequence = SEQ_CONTROL;
1085
1086			} else if (b->out_pos == b->out_size
1087					|| (b->in_pos == b->in_size
1088						&& s->temp.size
1089						< s->lzma2.compressed)) {
1090				return XZ_OK;
1091			}
1092
1093			break;
1094
1095		case SEQ_COPY:
1096			dict_uncompressed(&s->dict, b, &s->lzma2.compressed);
1097			if (s->lzma2.compressed > 0)
1098				return XZ_OK;
1099
1100			s->lzma2.sequence = SEQ_CONTROL;
1101			break;
1102		}
1103	}
1104
1105	return XZ_OK;
1106}
1107
1108XZ_EXTERN struct xz_dec_lzma2 *xz_dec_lzma2_create(enum xz_mode mode,
1109						   uint32_t dict_max)
1110{
1111	struct xz_dec_lzma2 *s = kmalloc(sizeof(*s), GFP_KERNEL);
1112	if (s == NULL)
1113		return NULL;
1114
1115	s->dict.mode = mode;
1116	s->dict.size_max = dict_max;
1117
1118	if (DEC_IS_PREALLOC(mode)) {
1119		s->dict.buf = vmalloc(dict_max);
1120		if (s->dict.buf == NULL) {
1121			kfree(s);
1122			return NULL;
1123		}
1124	} else if (DEC_IS_DYNALLOC(mode)) {
1125		s->dict.buf = NULL;
1126		s->dict.allocated = 0;
1127	}
1128
1129	return s;
1130}
1131
1132XZ_EXTERN enum xz_ret xz_dec_lzma2_reset(struct xz_dec_lzma2 *s, uint8_t props)
1133{
1134	/* This limits dictionary size to 3 GiB to keep parsing simpler. */
1135	if (props > 39)
1136		return XZ_OPTIONS_ERROR;
1137
1138	s->dict.size = 2 + (props & 1);
1139	s->dict.size <<= (props >> 1) + 11;
1140
1141	if (DEC_IS_MULTI(s->dict.mode)) {
1142		if (s->dict.size > s->dict.size_max)
1143			return XZ_MEMLIMIT_ERROR;
1144
1145		s->dict.end = s->dict.size;
1146
1147		if (DEC_IS_DYNALLOC(s->dict.mode)) {
1148			if (s->dict.allocated < s->dict.size) {
1149				vfree(s->dict.buf);
1150				s->dict.buf = vmalloc(s->dict.size);
1151				if (s->dict.buf == NULL) {
1152					s->dict.allocated = 0;
1153					return XZ_MEM_ERROR;
1154				}
1155			}
1156		}
1157	}
1158
1159	s->lzma.len = 0;
1160
1161	s->lzma2.sequence = SEQ_CONTROL;
1162	s->lzma2.need_dict_reset = true;
1163
1164	s->temp.size = 0;
1165
1166	return XZ_OK;
1167}
1168
1169XZ_EXTERN void xz_dec_lzma2_end(struct xz_dec_lzma2 *s)
1170{
1171	if (DEC_IS_MULTI(s->dict.mode))
1172		vfree(s->dict.buf);
1173
1174	kfree(s);
1175}
1176