aboutsummaryrefslogtreecommitdiffstats
path: root/lib/decompress_unlzma.c
diff options
context:
space:
mode:
Diffstat (limited to 'lib/decompress_unlzma.c')
-rw-r--r--lib/decompress_unlzma.c647
1 files changed, 647 insertions, 0 deletions
diff --git a/lib/decompress_unlzma.c b/lib/decompress_unlzma.c
new file mode 100644
index 000000000000..546f2f4c157e
--- /dev/null
+++ b/lib/decompress_unlzma.c
@@ -0,0 +1,647 @@
1/* Lzma decompressor for Linux kernel. Shamelessly snarfed
2 *from busybox 1.1.1
3 *
4 *Linux kernel adaptation
5 *Copyright (C) 2006 Alain < alain@knaff.lu >
6 *
7 *Based on small lzma deflate implementation/Small range coder
8 *implementation for lzma.
9 *Copyright (C) 2006 Aurelien Jacobs < aurel@gnuage.org >
10 *
11 *Based on LzmaDecode.c from the LZMA SDK 4.22 (http://www.7-zip.org/)
12 *Copyright (C) 1999-2005 Igor Pavlov
13 *
14 *Copyrights of the parts, see headers below.
15 *
16 *
17 *This program is free software; you can redistribute it and/or
18 *modify it under the terms of the GNU Lesser General Public
19 *License as published by the Free Software Foundation; either
20 *version 2.1 of the License, or (at your option) any later version.
21 *
22 *This program is distributed in the hope that it will be useful,
23 *but WITHOUT ANY WARRANTY; without even the implied warranty of
24 *MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
25 *Lesser General Public License for more details.
26 *
27 *You should have received a copy of the GNU Lesser General Public
28 *License along with this library; if not, write to the Free Software
29 *Foundation, Inc., 51 Franklin St, Fifth Floor, Boston, MA 02110-1301 USA
30 */
31
32#ifndef STATIC
33#include <linux/decompress/unlzma.h>
34#endif /* STATIC */
35
36#include <linux/decompress/mm.h>
37
38#define MIN(a, b) (((a) < (b)) ? (a) : (b))
39
40static long long INIT read_int(unsigned char *ptr, int size)
41{
42 int i;
43 long long ret = 0;
44
45 for (i = 0; i < size; i++)
46 ret = (ret << 8) | ptr[size-i-1];
47 return ret;
48}
49
50#define ENDIAN_CONVERT(x) \
51 x = (typeof(x))read_int((unsigned char *)&x, sizeof(x))
52
53
54/* Small range coder implementation for lzma.
55 *Copyright (C) 2006 Aurelien Jacobs < aurel@gnuage.org >
56 *
57 *Based on LzmaDecode.c from the LZMA SDK 4.22 (http://www.7-zip.org/)
58 *Copyright (c) 1999-2005 Igor Pavlov
59 */
60
61#include <linux/compiler.h>
62
63#define LZMA_IOBUF_SIZE 0x10000
64
65struct rc {
66 int (*fill)(void*, unsigned int);
67 uint8_t *ptr;
68 uint8_t *buffer;
69 uint8_t *buffer_end;
70 int buffer_size;
71 uint32_t code;
72 uint32_t range;
73 uint32_t bound;
74};
75
76
77#define RC_TOP_BITS 24
78#define RC_MOVE_BITS 5
79#define RC_MODEL_TOTAL_BITS 11
80
81
82/* Called twice: once at startup and once in rc_normalize() */
83static void INIT rc_read(struct rc *rc)
84{
85 rc->buffer_size = rc->fill((char *)rc->buffer, LZMA_IOBUF_SIZE);
86 if (rc->buffer_size <= 0)
87 error("unexpected EOF");
88 rc->ptr = rc->buffer;
89 rc->buffer_end = rc->buffer + rc->buffer_size;
90}
91
92/* Called once */
93static inline void INIT rc_init(struct rc *rc,
94 int (*fill)(void*, unsigned int),
95 char *buffer, int buffer_size)
96{
97 rc->fill = fill;
98 rc->buffer = (uint8_t *)buffer;
99 rc->buffer_size = buffer_size;
100 rc->buffer_end = rc->buffer + rc->buffer_size;
101 rc->ptr = rc->buffer;
102
103 rc->code = 0;
104 rc->range = 0xFFFFFFFF;
105}
106
107static inline void INIT rc_init_code(struct rc *rc)
108{
109 int i;
110
111 for (i = 0; i < 5; i++) {
112 if (rc->ptr >= rc->buffer_end)
113 rc_read(rc);
114 rc->code = (rc->code << 8) | *rc->ptr++;
115 }
116}
117
118
119/* Called once. TODO: bb_maybe_free() */
120static inline void INIT rc_free(struct rc *rc)
121{
122 free(rc->buffer);
123}
124
125/* Called twice, but one callsite is in inline'd rc_is_bit_0_helper() */
126static void INIT rc_do_normalize(struct rc *rc)
127{
128 if (rc->ptr >= rc->buffer_end)
129 rc_read(rc);
130 rc->range <<= 8;
131 rc->code = (rc->code << 8) | *rc->ptr++;
132}
133static inline void INIT rc_normalize(struct rc *rc)
134{
135 if (rc->range < (1 << RC_TOP_BITS))
136 rc_do_normalize(rc);
137}
138
139/* Called 9 times */
140/* Why rc_is_bit_0_helper exists?
141 *Because we want to always expose (rc->code < rc->bound) to optimizer
142 */
143static inline uint32_t INIT rc_is_bit_0_helper(struct rc *rc, uint16_t *p)
144{
145 rc_normalize(rc);
146 rc->bound = *p * (rc->range >> RC_MODEL_TOTAL_BITS);
147 return rc->bound;
148}
149static inline int INIT rc_is_bit_0(struct rc *rc, uint16_t *p)
150{
151 uint32_t t = rc_is_bit_0_helper(rc, p);
152 return rc->code < t;
153}
154
155/* Called ~10 times, but very small, thus inlined */
156static inline void INIT rc_update_bit_0(struct rc *rc, uint16_t *p)
157{
158 rc->range = rc->bound;
159 *p += ((1 << RC_MODEL_TOTAL_BITS) - *p) >> RC_MOVE_BITS;
160}
161static inline void rc_update_bit_1(struct rc *rc, uint16_t *p)
162{
163 rc->range -= rc->bound;
164 rc->code -= rc->bound;
165 *p -= *p >> RC_MOVE_BITS;
166}
167
168/* Called 4 times in unlzma loop */
169static int INIT rc_get_bit(struct rc *rc, uint16_t *p, int *symbol)
170{
171 if (rc_is_bit_0(rc, p)) {
172 rc_update_bit_0(rc, p);
173 *symbol *= 2;
174 return 0;
175 } else {
176 rc_update_bit_1(rc, p);
177 *symbol = *symbol * 2 + 1;
178 return 1;
179 }
180}
181
182/* Called once */
183static inline int INIT rc_direct_bit(struct rc *rc)
184{
185 rc_normalize(rc);
186 rc->range >>= 1;
187 if (rc->code >= rc->range) {
188 rc->code -= rc->range;
189 return 1;
190 }
191 return 0;
192}
193
194/* Called twice */
195static inline void INIT
196rc_bit_tree_decode(struct rc *rc, uint16_t *p, int num_levels, int *symbol)
197{
198 int i = num_levels;
199
200 *symbol = 1;
201 while (i--)
202 rc_get_bit(rc, p + *symbol, symbol);
203 *symbol -= 1 << num_levels;
204}
205
206
207/*
208 * Small lzma deflate implementation.
209 * Copyright (C) 2006 Aurelien Jacobs < aurel@gnuage.org >
210 *
211 * Based on LzmaDecode.c from the LZMA SDK 4.22 (http://www.7-zip.org/)
212 * Copyright (C) 1999-2005 Igor Pavlov
213 */
214
215
216struct lzma_header {
217 uint8_t pos;
218 uint32_t dict_size;
219 uint64_t dst_size;
220} __attribute__ ((packed)) ;
221
222
223#define LZMA_BASE_SIZE 1846
224#define LZMA_LIT_SIZE 768
225
226#define LZMA_NUM_POS_BITS_MAX 4
227
228#define LZMA_LEN_NUM_LOW_BITS 3
229#define LZMA_LEN_NUM_MID_BITS 3
230#define LZMA_LEN_NUM_HIGH_BITS 8
231
232#define LZMA_LEN_CHOICE 0
233#define LZMA_LEN_CHOICE_2 (LZMA_LEN_CHOICE + 1)
234#define LZMA_LEN_LOW (LZMA_LEN_CHOICE_2 + 1)
235#define LZMA_LEN_MID (LZMA_LEN_LOW \
236 + (1 << (LZMA_NUM_POS_BITS_MAX + LZMA_LEN_NUM_LOW_BITS)))
237#define LZMA_LEN_HIGH (LZMA_LEN_MID \
238 +(1 << (LZMA_NUM_POS_BITS_MAX + LZMA_LEN_NUM_MID_BITS)))
239#define LZMA_NUM_LEN_PROBS (LZMA_LEN_HIGH + (1 << LZMA_LEN_NUM_HIGH_BITS))
240
241#define LZMA_NUM_STATES 12
242#define LZMA_NUM_LIT_STATES 7
243
244#define LZMA_START_POS_MODEL_INDEX 4
245#define LZMA_END_POS_MODEL_INDEX 14
246#define LZMA_NUM_FULL_DISTANCES (1 << (LZMA_END_POS_MODEL_INDEX >> 1))
247
248#define LZMA_NUM_POS_SLOT_BITS 6
249#define LZMA_NUM_LEN_TO_POS_STATES 4
250
251#define LZMA_NUM_ALIGN_BITS 4
252
253#define LZMA_MATCH_MIN_LEN 2
254
255#define LZMA_IS_MATCH 0
256#define LZMA_IS_REP (LZMA_IS_MATCH + (LZMA_NUM_STATES << LZMA_NUM_POS_BITS_MAX))
257#define LZMA_IS_REP_G0 (LZMA_IS_REP + LZMA_NUM_STATES)
258#define LZMA_IS_REP_G1 (LZMA_IS_REP_G0 + LZMA_NUM_STATES)
259#define LZMA_IS_REP_G2 (LZMA_IS_REP_G1 + LZMA_NUM_STATES)
260#define LZMA_IS_REP_0_LONG (LZMA_IS_REP_G2 + LZMA_NUM_STATES)
261#define LZMA_POS_SLOT (LZMA_IS_REP_0_LONG \
262 + (LZMA_NUM_STATES << LZMA_NUM_POS_BITS_MAX))
263#define LZMA_SPEC_POS (LZMA_POS_SLOT \
264 +(LZMA_NUM_LEN_TO_POS_STATES << LZMA_NUM_POS_SLOT_BITS))
265#define LZMA_ALIGN (LZMA_SPEC_POS \
266 + LZMA_NUM_FULL_DISTANCES - LZMA_END_POS_MODEL_INDEX)
267#define LZMA_LEN_CODER (LZMA_ALIGN + (1 << LZMA_NUM_ALIGN_BITS))
268#define LZMA_REP_LEN_CODER (LZMA_LEN_CODER + LZMA_NUM_LEN_PROBS)
269#define LZMA_LITERAL (LZMA_REP_LEN_CODER + LZMA_NUM_LEN_PROBS)
270
271
272struct writer {
273 uint8_t *buffer;
274 uint8_t previous_byte;
275 size_t buffer_pos;
276 int bufsize;
277 size_t global_pos;
278 int(*flush)(void*, unsigned int);
279 struct lzma_header *header;
280};
281
282struct cstate {
283 int state;
284 uint32_t rep0, rep1, rep2, rep3;
285};
286
287static inline size_t INIT get_pos(struct writer *wr)
288{
289 return
290 wr->global_pos + wr->buffer_pos;
291}
292
293static inline uint8_t INIT peek_old_byte(struct writer *wr,
294 uint32_t offs)
295{
296 if (!wr->flush) {
297 int32_t pos;
298 while (offs > wr->header->dict_size)
299 offs -= wr->header->dict_size;
300 pos = wr->buffer_pos - offs;
301 return wr->buffer[pos];
302 } else {
303 uint32_t pos = wr->buffer_pos - offs;
304 while (pos >= wr->header->dict_size)
305 pos += wr->header->dict_size;
306 return wr->buffer[pos];
307 }
308
309}
310
311static inline void INIT write_byte(struct writer *wr, uint8_t byte)
312{
313 wr->buffer[wr->buffer_pos++] = wr->previous_byte = byte;
314 if (wr->flush && wr->buffer_pos == wr->header->dict_size) {
315 wr->buffer_pos = 0;
316 wr->global_pos += wr->header->dict_size;
317 wr->flush((char *)wr->buffer, wr->header->dict_size);
318 }
319}
320
321
322static inline void INIT copy_byte(struct writer *wr, uint32_t offs)
323{
324 write_byte(wr, peek_old_byte(wr, offs));
325}
326
327static inline void INIT copy_bytes(struct writer *wr,
328 uint32_t rep0, int len)
329{
330 do {
331 copy_byte(wr, rep0);
332 len--;
333 } while (len != 0 && wr->buffer_pos < wr->header->dst_size);
334}
335
336static inline void INIT process_bit0(struct writer *wr, struct rc *rc,
337 struct cstate *cst, uint16_t *p,
338 int pos_state, uint16_t *prob,
339 int lc, uint32_t literal_pos_mask) {
340 int mi = 1;
341 rc_update_bit_0(rc, prob);
342 prob = (p + LZMA_LITERAL +
343 (LZMA_LIT_SIZE
344 * (((get_pos(wr) & literal_pos_mask) << lc)
345 + (wr->previous_byte >> (8 - lc))))
346 );
347
348 if (cst->state >= LZMA_NUM_LIT_STATES) {
349 int match_byte = peek_old_byte(wr, cst->rep0);
350 do {
351 int bit;
352 uint16_t *prob_lit;
353
354 match_byte <<= 1;
355 bit = match_byte & 0x100;
356 prob_lit = prob + 0x100 + bit + mi;
357 if (rc_get_bit(rc, prob_lit, &mi)) {
358 if (!bit)
359 break;
360 } else {
361 if (bit)
362 break;
363 }
364 } while (mi < 0x100);
365 }
366 while (mi < 0x100) {
367 uint16_t *prob_lit = prob + mi;
368 rc_get_bit(rc, prob_lit, &mi);
369 }
370 write_byte(wr, mi);
371 if (cst->state < 4)
372 cst->state = 0;
373 else if (cst->state < 10)
374 cst->state -= 3;
375 else
376 cst->state -= 6;
377}
378
379static inline void INIT process_bit1(struct writer *wr, struct rc *rc,
380 struct cstate *cst, uint16_t *p,
381 int pos_state, uint16_t *prob) {
382 int offset;
383 uint16_t *prob_len;
384 int num_bits;
385 int len;
386
387 rc_update_bit_1(rc, prob);
388 prob = p + LZMA_IS_REP + cst->state;
389 if (rc_is_bit_0(rc, prob)) {
390 rc_update_bit_0(rc, prob);
391 cst->rep3 = cst->rep2;
392 cst->rep2 = cst->rep1;
393 cst->rep1 = cst->rep0;
394 cst->state = cst->state < LZMA_NUM_LIT_STATES ? 0 : 3;
395 prob = p + LZMA_LEN_CODER;
396 } else {
397 rc_update_bit_1(rc, prob);
398 prob = p + LZMA_IS_REP_G0 + cst->state;
399 if (rc_is_bit_0(rc, prob)) {
400 rc_update_bit_0(rc, prob);
401 prob = (p + LZMA_IS_REP_0_LONG
402 + (cst->state <<
403 LZMA_NUM_POS_BITS_MAX) +
404 pos_state);
405 if (rc_is_bit_0(rc, prob)) {
406 rc_update_bit_0(rc, prob);
407
408 cst->state = cst->state < LZMA_NUM_LIT_STATES ?
409 9 : 11;
410 copy_byte(wr, cst->rep0);
411 return;
412 } else {
413 rc_update_bit_1(rc, prob);
414 }
415 } else {
416 uint32_t distance;
417
418 rc_update_bit_1(rc, prob);
419 prob = p + LZMA_IS_REP_G1 + cst->state;
420 if (rc_is_bit_0(rc, prob)) {
421 rc_update_bit_0(rc, prob);
422 distance = cst->rep1;
423 } else {
424 rc_update_bit_1(rc, prob);
425 prob = p + LZMA_IS_REP_G2 + cst->state;
426 if (rc_is_bit_0(rc, prob)) {
427 rc_update_bit_0(rc, prob);
428 distance = cst->rep2;
429 } else {
430 rc_update_bit_1(rc, prob);
431 distance = cst->rep3;
432 cst->rep3 = cst->rep2;
433 }
434 cst->rep2 = cst->rep1;
435 }
436 cst->rep1 = cst->rep0;
437 cst->rep0 = distance;
438 }
439 cst->state = cst->state < LZMA_NUM_LIT_STATES ? 8 : 11;
440 prob = p + LZMA_REP_LEN_CODER;
441 }
442
443 prob_len = prob + LZMA_LEN_CHOICE;
444 if (rc_is_bit_0(rc, prob_len)) {
445 rc_update_bit_0(rc, prob_len);
446 prob_len = (prob + LZMA_LEN_LOW
447 + (pos_state <<
448 LZMA_LEN_NUM_LOW_BITS));
449 offset = 0;
450 num_bits = LZMA_LEN_NUM_LOW_BITS;
451 } else {
452 rc_update_bit_1(rc, prob_len);
453 prob_len = prob + LZMA_LEN_CHOICE_2;
454 if (rc_is_bit_0(rc, prob_len)) {
455 rc_update_bit_0(rc, prob_len);
456 prob_len = (prob + LZMA_LEN_MID
457 + (pos_state <<
458 LZMA_LEN_NUM_MID_BITS));
459 offset = 1 << LZMA_LEN_NUM_LOW_BITS;
460 num_bits = LZMA_LEN_NUM_MID_BITS;
461 } else {
462 rc_update_bit_1(rc, prob_len);
463 prob_len = prob + LZMA_LEN_HIGH;
464 offset = ((1 << LZMA_LEN_NUM_LOW_BITS)
465 + (1 << LZMA_LEN_NUM_MID_BITS));
466 num_bits = LZMA_LEN_NUM_HIGH_BITS;
467 }
468 }
469
470 rc_bit_tree_decode(rc, prob_len, num_bits, &len);
471 len += offset;
472
473 if (cst->state < 4) {
474 int pos_slot;
475
476 cst->state += LZMA_NUM_LIT_STATES;
477 prob =
478 p + LZMA_POS_SLOT +
479 ((len <
480 LZMA_NUM_LEN_TO_POS_STATES ? len :
481 LZMA_NUM_LEN_TO_POS_STATES - 1)
482 << LZMA_NUM_POS_SLOT_BITS);
483 rc_bit_tree_decode(rc, prob,
484 LZMA_NUM_POS_SLOT_BITS,
485 &pos_slot);
486 if (pos_slot >= LZMA_START_POS_MODEL_INDEX) {
487 int i, mi;
488 num_bits = (pos_slot >> 1) - 1;
489 cst->rep0 = 2 | (pos_slot & 1);
490 if (pos_slot < LZMA_END_POS_MODEL_INDEX) {
491 cst->rep0 <<= num_bits;
492 prob = p + LZMA_SPEC_POS +
493 cst->rep0 - pos_slot - 1;
494 } else {
495 num_bits -= LZMA_NUM_ALIGN_BITS;
496 while (num_bits--)
497 cst->rep0 = (cst->rep0 << 1) |
498 rc_direct_bit(rc);
499 prob = p + LZMA_ALIGN;
500 cst->rep0 <<= LZMA_NUM_ALIGN_BITS;
501 num_bits = LZMA_NUM_ALIGN_BITS;
502 }
503 i = 1;
504 mi = 1;
505 while (num_bits--) {
506 if (rc_get_bit(rc, prob + mi, &mi))
507 cst->rep0 |= i;
508 i <<= 1;
509 }
510 } else
511 cst->rep0 = pos_slot;
512 if (++(cst->rep0) == 0)
513 return;
514 }
515
516 len += LZMA_MATCH_MIN_LEN;
517
518 copy_bytes(wr, cst->rep0, len);
519}
520
521
522
523STATIC inline int INIT unlzma(unsigned char *buf, int in_len,
524 int(*fill)(void*, unsigned int),
525 int(*flush)(void*, unsigned int),
526 unsigned char *output,
527 int *posp,
528 void(*error_fn)(char *x)
529 )
530{
531 struct lzma_header header;
532 int lc, pb, lp;
533 uint32_t pos_state_mask;
534 uint32_t literal_pos_mask;
535 uint16_t *p;
536 int num_probs;
537 struct rc rc;
538 int i, mi;
539 struct writer wr;
540 struct cstate cst;
541 unsigned char *inbuf;
542 int ret = -1;
543
544 set_error_fn(error_fn);
545 if (!flush)
546 in_len -= 4; /* Uncompressed size hack active in pre-boot
547 environment */
548 if (buf)
549 inbuf = buf;
550 else
551 inbuf = malloc(LZMA_IOBUF_SIZE);
552 if (!inbuf) {
553 error("Could not allocate input bufer");
554 goto exit_0;
555 }
556
557 cst.state = 0;
558 cst.rep0 = cst.rep1 = cst.rep2 = cst.rep3 = 1;
559
560 wr.header = &header;
561 wr.flush = flush;
562 wr.global_pos = 0;
563 wr.previous_byte = 0;
564 wr.buffer_pos = 0;
565
566 rc_init(&rc, fill, inbuf, in_len);
567
568 for (i = 0; i < sizeof(header); i++) {
569 if (rc.ptr >= rc.buffer_end)
570 rc_read(&rc);
571 ((unsigned char *)&header)[i] = *rc.ptr++;
572 }
573
574 if (header.pos >= (9 * 5 * 5))
575 error("bad header");
576
577 mi = 0;
578 lc = header.pos;
579 while (lc >= 9) {
580 mi++;
581 lc -= 9;
582 }
583 pb = 0;
584 lp = mi;
585 while (lp >= 5) {
586 pb++;
587 lp -= 5;
588 }
589 pos_state_mask = (1 << pb) - 1;
590 literal_pos_mask = (1 << lp) - 1;
591
592 ENDIAN_CONVERT(header.dict_size);
593 ENDIAN_CONVERT(header.dst_size);
594
595 if (header.dict_size == 0)
596 header.dict_size = 1;
597
598 if (output)
599 wr.buffer = output;
600 else {
601 wr.bufsize = MIN(header.dst_size, header.dict_size);
602 wr.buffer = large_malloc(wr.bufsize);
603 }
604 if (wr.buffer == NULL)
605 goto exit_1;
606
607 num_probs = LZMA_BASE_SIZE + (LZMA_LIT_SIZE << (lc + lp));
608 p = (uint16_t *) large_malloc(num_probs * sizeof(*p));
609 if (p == 0)
610 goto exit_2;
611 num_probs = LZMA_LITERAL + (LZMA_LIT_SIZE << (lc + lp));
612 for (i = 0; i < num_probs; i++)
613 p[i] = (1 << RC_MODEL_TOTAL_BITS) >> 1;
614
615 rc_init_code(&rc);
616
617 while (get_pos(&wr) < header.dst_size) {
618 int pos_state = get_pos(&wr) & pos_state_mask;
619 uint16_t *prob = p + LZMA_IS_MATCH +
620 (cst.state << LZMA_NUM_POS_BITS_MAX) + pos_state;
621 if (rc_is_bit_0(&rc, prob))
622 process_bit0(&wr, &rc, &cst, p, pos_state, prob,
623 lc, literal_pos_mask);
624 else {
625 process_bit1(&wr, &rc, &cst, p, pos_state, prob);
626 if (cst.rep0 == 0)
627 break;
628 }
629 }
630
631 if (posp)
632 *posp = rc.ptr-rc.buffer;
633 if (wr.flush)
634 wr.flush(wr.buffer, wr.buffer_pos);
635 ret = 0;
636 large_free(p);
637exit_2:
638 if (!output)
639 large_free(wr.buffer);
640exit_1:
641 if (!buf)
642 free(inbuf);
643exit_0:
644 return ret;
645}
646
647#define decompress unlzma
generated by cgit v1.2.2 (git 2.25.0) at 2025-11-04 05:40:28 -0500