4 * A decompressor for the LZX compression format, as used in WIM files.
8 * Copyright (C) 2012-2016 Eric Biggers
10 * This file is free software; you can redistribute it and/or modify it under
11 * the terms of the GNU Lesser General Public License as published by the Free
12 * Software Foundation; either version 3 of the License, or (at your option) any
15 * This file is distributed in the hope that it will be useful, but WITHOUT
16 * ANY WARRANTY; without even the implied warranty of MERCHANTABILITY or FITNESS
17 * FOR A PARTICULAR PURPOSE. See the GNU Lesser General Public License for more
20 * You should have received a copy of the GNU Lesser General Public License
21 * along with this file; if not, see http://www.gnu.org/licenses/.
25 * LZX is an LZ77 and Huffman-code based compression format that has many
26 * similarities to DEFLATE (the format used by zlib/gzip). The compression
27 * ratio is as good or better than DEFLATE. See lzx_compress.c for a format
28 * overview, and see https://en.wikipedia.org/wiki/LZX_(algorithm) for a
29 * historical overview. Here I make some pragmatic notes.
31 * The old specification for LZX is the document "Microsoft LZX Data Compression
32 * Format" (1997). It defines the LZX format as used in cabinet files. Allowed
33 * window sizes are 2^n where 15 <= n <= 21. However, this document contains
34 * several errors, so don't read too much into it...
36 * The new specification for LZX is the document "[MS-PATCH]: LZX DELTA
37 * Compression and Decompression" (2014). It defines the LZX format as used by
38 * Microsoft's binary patcher. It corrects several errors in the 1997 document
39 * and extends the format in several ways --- namely, optional reference data,
40 * up to 2^25 byte windows, and longer match lengths.
42 * WIM files use a more restricted form of LZX. No LZX DELTA extensions are
43 * present, the window is not "sliding", E8 preprocessing is done
44 * unconditionally with a fixed file size, and the maximum window size is always
45 * 2^15 bytes (equal to the size of each "chunk" in a compressed WIM resource).
46 * This code is primarily intended to implement this form of LZX. But although
47 * not compatible with WIMGAPI, this code also supports maximum window sizes up
50 * TODO: Add support for window sizes up to 2^25 bytes.
59 #include "wimlib/decompressor_ops.h"
60 #include "wimlib/decompress_common.h"
61 #include "wimlib/error.h"
62 #include "wimlib/lzx_common.h"
63 #include "wimlib/util.h"
65 /* These values are chosen for fast decompression. */
66 #define LZX_MAINCODE_TABLEBITS 11
67 #define LZX_LENCODE_TABLEBITS 10
68 #define LZX_PRECODE_TABLEBITS 6
69 #define LZX_ALIGNEDCODE_TABLEBITS 7
71 #define LZX_READ_LENS_MAX_OVERRUN 50
73 struct lzx_decompressor {
75 u16 maincode_decode_table[(1 << LZX_MAINCODE_TABLEBITS) +
76 (LZX_MAINCODE_MAX_NUM_SYMBOLS * 2)]
77 _aligned_attribute(DECODE_TABLE_ALIGNMENT);
78 u8 maincode_lens[LZX_MAINCODE_MAX_NUM_SYMBOLS + LZX_READ_LENS_MAX_OVERRUN];
81 u16 lencode_decode_table[(1 << LZX_LENCODE_TABLEBITS) +
82 (LZX_LENCODE_NUM_SYMBOLS * 2)]
83 _aligned_attribute(DECODE_TABLE_ALIGNMENT);
84 u8 lencode_lens[LZX_LENCODE_NUM_SYMBOLS + LZX_READ_LENS_MAX_OVERRUN];
87 u16 alignedcode_decode_table[(1 << LZX_ALIGNEDCODE_TABLEBITS) +
88 (LZX_ALIGNEDCODE_NUM_SYMBOLS * 2)]
89 _aligned_attribute(DECODE_TABLE_ALIGNMENT);
90 u8 alignedcode_lens[LZX_ALIGNEDCODE_NUM_SYMBOLS];
94 u16 precode_decode_table[(1 << LZX_PRECODE_TABLEBITS) +
95 (LZX_PRECODE_NUM_SYMBOLS * 2)]
96 _aligned_attribute(DECODE_TABLE_ALIGNMENT);
97 u8 precode_lens[LZX_PRECODE_NUM_SYMBOLS];
98 u8 extra_offset_bits[LZX_MAX_OFFSET_SLOTS];
101 unsigned window_order;
102 unsigned num_main_syms;
104 /* Like lzx_extra_offset_bits[], but does not include the entropy-coded
105 * bits of aligned offset blocks */
106 u8 extra_offset_bits_minus_aligned[LZX_MAX_OFFSET_SLOTS];
108 } _aligned_attribute(DECODE_TABLE_ALIGNMENT);
110 /* Read a Huffman-encoded symbol using the precode. */
111 static inline unsigned
112 read_presym(const struct lzx_decompressor *d, struct input_bitstream *is)
114 return read_huffsym(is, d->precode_decode_table,
115 LZX_PRECODE_TABLEBITS, LZX_MAX_PRE_CODEWORD_LEN);
118 /* Read a Huffman-encoded symbol using the main code. */
119 static inline unsigned
120 read_mainsym(const struct lzx_decompressor *d, struct input_bitstream *is)
122 return read_huffsym(is, d->maincode_decode_table,
123 LZX_MAINCODE_TABLEBITS, LZX_MAX_MAIN_CODEWORD_LEN);
126 /* Read a Huffman-encoded symbol using the length code. */
127 static inline unsigned
128 read_lensym(const struct lzx_decompressor *d, struct input_bitstream *is)
130 return read_huffsym(is, d->lencode_decode_table,
131 LZX_LENCODE_TABLEBITS, LZX_MAX_LEN_CODEWORD_LEN);
134 /* Read a Huffman-encoded symbol using the aligned offset code. */
135 static inline unsigned
136 read_alignedsym(const struct lzx_decompressor *d, struct input_bitstream *is)
138 return read_huffsym(is, d->alignedcode_decode_table,
139 LZX_ALIGNEDCODE_TABLEBITS, LZX_MAX_ALIGNED_CODEWORD_LEN);
143 * Read a precode from the compressed input bitstream, then use it to decode
144 * @num_lens codeword length values and write them to @lens.
147 lzx_read_codeword_lens(struct lzx_decompressor *d, struct input_bitstream *is,
148 u8 *lens, unsigned num_lens)
151 u8 *lens_end = lens + num_lens;
153 /* Read the lengths of the precode codewords. These are stored
155 for (int i = 0; i < LZX_PRECODE_NUM_SYMBOLS; i++) {
157 bitstream_read_bits(is, LZX_PRECODE_ELEMENT_SIZE);
160 /* Build the decoding table for the precode. */
161 if (make_huffman_decode_table(d->precode_decode_table,
162 LZX_PRECODE_NUM_SYMBOLS,
163 LZX_PRECODE_TABLEBITS,
165 LZX_MAX_PRE_CODEWORD_LEN))
168 /* Decode the codeword lengths. */
173 /* Read the next precode symbol. */
174 presym = read_presym(d, is);
176 /* Difference from old length */
177 len = *len_ptr - presym;
182 /* Special RLE values */
188 run_len = 4 + bitstream_read_bits(is, 4);
190 } else if (presym == 18) {
191 /* Longer run of 0's */
192 run_len = 20 + bitstream_read_bits(is, 5);
195 /* Run of identical lengths */
196 run_len = 4 + bitstream_read_bits(is, 1);
197 presym = read_presym(d, is);
198 if (unlikely(presym > 17))
200 len = *len_ptr - presym;
209 * The worst case overrun is when presym == 18,
210 * run_len == 20 + 31, and only 1 length was remaining.
211 * So LZX_READ_LENS_MAX_OVERRUN == 50.
213 * Overrun while reading the first half of maincode_lens
214 * can corrupt the previous values in the second half.
215 * This doesn't really matter because the resulting
216 * lengths will still be in range, and data that
217 * generates overruns is invalid anyway.
220 } while (len_ptr < lens_end);
226 * Read the header of an LZX block. For all block types, the block type and
227 * size is saved in *block_type_ret and *block_size_ret, respectively. For
228 * compressed blocks, the codeword lengths are also saved. For uncompressed
229 * blocks, the recent offsets queue is also updated.
232 lzx_read_block_header(struct lzx_decompressor *d, struct input_bitstream *is,
233 u32 recent_offsets[], int *block_type_ret,
239 bitstream_ensure_bits(is, 4);
241 /* Read the block type. */
242 block_type = bitstream_pop_bits(is, 3);
244 /* Read the block size. */
245 if (bitstream_pop_bits(is, 1)) {
246 block_size = LZX_DEFAULT_BLOCK_SIZE;
248 block_size = bitstream_read_bits(is, 16);
249 if (d->window_order >= 16) {
251 block_size |= bitstream_read_bits(is, 8);
255 switch (block_type) {
257 case LZX_BLOCKTYPE_ALIGNED:
259 /* Read the aligned offset codeword lengths. */
261 for (int i = 0; i < LZX_ALIGNEDCODE_NUM_SYMBOLS; i++) {
262 d->alignedcode_lens[i] =
263 bitstream_read_bits(is,
264 LZX_ALIGNEDCODE_ELEMENT_SIZE);
267 /* Fall though, since the rest of the header for aligned offset
268 * blocks is the same as that for verbatim blocks. */
270 case LZX_BLOCKTYPE_VERBATIM:
272 /* Read the main codeword lengths, which are divided into two
273 * parts: literal symbols and match headers. */
275 if (lzx_read_codeword_lens(d, is, d->maincode_lens,
279 if (lzx_read_codeword_lens(d, is, d->maincode_lens + LZX_NUM_CHARS,
280 d->num_main_syms - LZX_NUM_CHARS))
284 /* Read the length codeword lengths. */
286 if (lzx_read_codeword_lens(d, is, d->lencode_lens,
287 LZX_LENCODE_NUM_SYMBOLS))
292 case LZX_BLOCKTYPE_UNCOMPRESSED:
294 * The header of an uncompressed block contains new values for
295 * the recent offsets queue, starting on the next 16-bit
296 * boundary in the bitstream. Careful: if the stream is
297 * *already* aligned, the correct thing to do is to throw away
298 * the next 16 bits (this is probably a mistake in the format).
300 bitstream_ensure_bits(is, 1);
302 recent_offsets[0] = bitstream_read_u32(is);
303 recent_offsets[1] = bitstream_read_u32(is);
304 recent_offsets[2] = bitstream_read_u32(is);
306 /* Offsets of 0 are invalid. */
307 if (recent_offsets[0] == 0 || recent_offsets[1] == 0 ||
308 recent_offsets[2] == 0)
313 /* Unrecognized block type. */
317 *block_type_ret = block_type;
318 *block_size_ret = block_size;
322 /* Decompress a block of LZX-compressed data. */
324 lzx_decompress_block(struct lzx_decompressor *d, struct input_bitstream *is,
325 int block_type, u32 block_size,
326 u8 * const out_begin, u8 *out_next, u32 recent_offsets[])
328 u8 * const block_end = out_next + block_size;
329 unsigned min_aligned_offset_slot;
332 * Build the Huffman decode tables. We always need to build the main
333 * and length decode tables. For aligned blocks we additionally need to
334 * build the aligned offset decode table.
337 if (make_huffman_decode_table(d->maincode_decode_table,
339 LZX_MAINCODE_TABLEBITS,
341 LZX_MAX_MAIN_CODEWORD_LEN))
344 if (make_huffman_decode_table(d->lencode_decode_table,
345 LZX_LENCODE_NUM_SYMBOLS,
346 LZX_LENCODE_TABLEBITS,
348 LZX_MAX_LEN_CODEWORD_LEN))
351 if (block_type == LZX_BLOCKTYPE_ALIGNED) {
352 if (make_huffman_decode_table(d->alignedcode_decode_table,
353 LZX_ALIGNEDCODE_NUM_SYMBOLS,
354 LZX_ALIGNEDCODE_TABLEBITS,
356 LZX_MAX_ALIGNED_CODEWORD_LEN))
358 min_aligned_offset_slot = LZX_MIN_ALIGNED_OFFSET_SLOT;
359 memcpy(d->extra_offset_bits, d->extra_offset_bits_minus_aligned,
360 sizeof(lzx_extra_offset_bits));
362 min_aligned_offset_slot = LZX_MAX_OFFSET_SLOTS;
363 memcpy(d->extra_offset_bits, lzx_extra_offset_bits,
364 sizeof(lzx_extra_offset_bits));
367 /* Decode the literals and matches. */
373 unsigned offset_slot;
375 mainsym = read_mainsym(d, is);
376 if (mainsym < LZX_NUM_CHARS) {
378 *out_next++ = mainsym;
384 /* Decode the length header and offset slot. */
385 STATIC_ASSERT(LZX_NUM_CHARS % LZX_NUM_LEN_HEADERS == 0);
386 length = mainsym % LZX_NUM_LEN_HEADERS;
387 offset_slot = (mainsym - LZX_NUM_CHARS) / LZX_NUM_LEN_HEADERS;
389 /* If needed, read a length symbol to decode the full length. */
390 if (length == LZX_NUM_PRIMARY_LENS)
391 length += read_lensym(d, is);
392 length += LZX_MIN_MATCH_LEN;
394 if (offset_slot < LZX_NUM_RECENT_OFFSETS) {
397 /* Note: This isn't a real LRU queue, since using the R2
398 * offset doesn't bump the R1 offset down to R2. */
399 offset = recent_offsets[offset_slot];
400 recent_offsets[offset_slot] = recent_offsets[0];
402 /* Explicit offset */
403 offset = bitstream_read_bits(is, d->extra_offset_bits[offset_slot]);
404 if (offset_slot >= min_aligned_offset_slot) {
405 offset = (offset << LZX_NUM_ALIGNED_OFFSET_BITS) |
406 read_alignedsym(d, is);
408 offset += lzx_offset_slot_base[offset_slot];
410 /* Update the match offset LRU queue. */
411 STATIC_ASSERT(LZX_NUM_RECENT_OFFSETS == 3);
412 recent_offsets[2] = recent_offsets[1];
413 recent_offsets[1] = recent_offsets[0];
415 recent_offsets[0] = offset;
417 /* Validate the match, then copy it to the current position. */
419 if (unlikely(length > block_end - out_next))
422 if (unlikely(offset > out_next - out_begin))
425 lz_copy(out_next, length, offset, block_end, LZX_MIN_MATCH_LEN);
429 } while (out_next != block_end);
435 lzx_decompress(const void *restrict compressed_data, size_t compressed_size,
436 void *restrict uncompressed_data, size_t uncompressed_size,
439 struct lzx_decompressor *d = _d;
440 u8 * const out_begin = uncompressed_data;
441 u8 *out_next = out_begin;
442 u8 * const out_end = out_begin + uncompressed_size;
443 struct input_bitstream is;
444 STATIC_ASSERT(LZX_NUM_RECENT_OFFSETS == 3);
445 u32 recent_offsets[LZX_NUM_RECENT_OFFSETS] = {1, 1, 1};
446 unsigned may_have_e8_byte = 0;
448 init_input_bitstream(&is, compressed_data, compressed_size);
450 /* Codeword lengths begin as all 0's for delta encoding purposes. */
451 memset(d->maincode_lens, 0, d->num_main_syms);
452 memset(d->lencode_lens, 0, LZX_LENCODE_NUM_SYMBOLS);
454 /* Decompress blocks until we have all the uncompressed data. */
456 while (out_next != out_end) {
460 if (lzx_read_block_header(d, &is, recent_offsets,
461 &block_type, &block_size))
464 if (block_size < 1 || block_size > out_end - out_next)
467 if (likely(block_type != LZX_BLOCKTYPE_UNCOMPRESSED)) {
469 /* Compressed block */
470 if (lzx_decompress_block(d, &is, block_type, block_size,
475 /* If the first E8 byte was in this block, then it must
476 * have been encoded as a literal using mainsym E8. */
477 may_have_e8_byte |= d->maincode_lens[0xE8];
480 /* Uncompressed block */
481 if (bitstream_read_bytes(&is, out_next, block_size))
484 /* Re-align the bitstream if needed. */
486 bitstream_read_byte(&is);
488 /* There may have been an E8 byte in the block. */
489 may_have_e8_byte = 1;
491 out_next += block_size;
494 /* Postprocess the data unless it cannot possibly contain E8 bytes. */
495 if (may_have_e8_byte)
496 lzx_postprocess(uncompressed_data, uncompressed_size);
502 lzx_create_decompressor(size_t max_block_size, void **d_ret)
504 unsigned window_order;
505 struct lzx_decompressor *d;
507 window_order = lzx_get_window_order(max_block_size);
508 if (window_order == 0)
509 return WIMLIB_ERR_INVALID_PARAM;
511 d = ALIGNED_MALLOC(sizeof(*d), DECODE_TABLE_ALIGNMENT);
513 return WIMLIB_ERR_NOMEM;
515 d->window_order = window_order;
516 d->num_main_syms = lzx_get_num_main_syms(window_order);
518 /* Initialize 'd->extra_offset_bits_minus_aligned'. */
519 STATIC_ASSERT(sizeof(d->extra_offset_bits_minus_aligned) ==
520 sizeof(lzx_extra_offset_bits));
521 STATIC_ASSERT(sizeof(d->extra_offset_bits) ==
522 sizeof(lzx_extra_offset_bits));
523 memcpy(d->extra_offset_bits_minus_aligned, lzx_extra_offset_bits,
524 sizeof(lzx_extra_offset_bits));
525 for (unsigned offset_slot = LZX_MIN_ALIGNED_OFFSET_SLOT;
526 offset_slot < LZX_MAX_OFFSET_SLOTS; offset_slot++)
528 d->extra_offset_bits_minus_aligned[offset_slot] -=
529 LZX_NUM_ALIGNED_OFFSET_BITS;
537 lzx_free_decompressor(void *_d)
542 const struct decompressor_ops lzx_decompressor_ops = {
543 .create_decompressor = lzx_create_decompressor,
544 .decompress = lzx_decompress,
545 .free_decompressor = lzx_free_decompressor,