1 /* Lziprecover - Data recovery tool for the lzip format
2 Copyright (C) 2009-2021 Antonio Diaz Diaz.
3
4 This program is free software: you can redistribute it and/or modify
5 it under the terms of the GNU General Public License as published by
6 the Free Software Foundation, either version 2 of the License, or
7 (at your option) any later version.
8
9 This program is distributed in the hope that it will be useful,
10 but WITHOUT ANY WARRANTY; without even the implied warranty of
11 MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the
12 GNU General Public License for more details.
13
14 You should have received a copy of the GNU General Public License
15 along with this program. If not, see <http://www.gnu.org/licenses/>.
16 */
17
18 class State
19 {
20 int st;
21
22 public:
23 enum { states = 12 };
State()24 State() : st( 0 ) {}
operator()25 int operator()() const { return st; }
is_char()26 bool is_char() const { return st < 7; }
27
set_char()28 void set_char()
29 {
30 static const int next[states] = { 0, 0, 0, 0, 1, 2, 3, 4, 5, 6, 4, 5 };
31 st = next[st];
32 }
is_char_set_char()33 bool is_char_set_char()
34 {
35 if( st < 7 ) { st -= ( st < 4 ) ? st : 3; return true; }
36 else { st -= ( st < 10 ) ? 3 : 6; return false; }
37 }
set_match()38 void set_match() { st = ( st < 7 ) ? 7 : 10; }
set_rep()39 void set_rep() { st = ( st < 7 ) ? 8 : 11; }
set_short_rep()40 void set_short_rep() { st = ( st < 7 ) ? 9 : 11; }
41 };
42
43
44 enum {
45 min_dictionary_bits = 12,
46 min_dictionary_size = 1 << min_dictionary_bits, // >= modeled_distances
47 max_dictionary_bits = 29,
48 max_dictionary_size = 1 << max_dictionary_bits,
49 min_member_size = 36,
50 literal_context_bits = 3,
51 literal_pos_state_bits = 0, // not used
52 pos_state_bits = 2,
53 pos_states = 1 << pos_state_bits,
54 pos_state_mask = pos_states - 1,
55
56 len_states = 4,
57 dis_slot_bits = 6,
58 start_dis_model = 4,
59 end_dis_model = 14,
60 modeled_distances = 1 << (end_dis_model / 2), // 128
61 dis_align_bits = 4,
62 dis_align_size = 1 << dis_align_bits,
63
64 len_low_bits = 3,
65 len_mid_bits = 3,
66 len_high_bits = 8,
67 len_low_symbols = 1 << len_low_bits,
68 len_mid_symbols = 1 << len_mid_bits,
69 len_high_symbols = 1 << len_high_bits,
70 max_len_symbols = len_low_symbols + len_mid_symbols + len_high_symbols,
71
72 min_match_len = 2, // must be 2
73 max_match_len = min_match_len + max_len_symbols - 1, // 273
74 min_match_len_limit = 5 };
75
get_len_state(const int len)76 inline int get_len_state( const int len )
77 { return std::min( len - min_match_len, len_states - 1 ); }
78
get_lit_state(const uint8_t prev_byte)79 inline int get_lit_state( const uint8_t prev_byte )
80 { return prev_byte >> ( 8 - literal_context_bits ); }
81
82
83 enum { bit_model_move_bits = 5,
84 bit_model_total_bits = 11,
85 bit_model_total = 1 << bit_model_total_bits };
86
87 struct Bit_model
88 {
89 int probability;
Bit_modelBit_model90 Bit_model() : probability( bit_model_total / 2 ) {}
91 };
92
93 struct Len_model
94 {
95 Bit_model choice1;
96 Bit_model choice2;
97 Bit_model bm_low[pos_states][len_low_symbols];
98 Bit_model bm_mid[pos_states][len_mid_symbols];
99 Bit_model bm_high[len_high_symbols];
100 };
101
102
103 // defined in main.cc
104 extern int verbosity;
105
106 class Pretty_print // requires global var 'int verbosity'
107 {
108 std::string name_;
109 std::string padded_name;
110 const char * const stdin_name;
111 unsigned longest_name;
112 mutable bool first_post;
113
114 public:
Pretty_print(const std::vector<std::string> & filenames)115 Pretty_print( const std::vector< std::string > & filenames )
116 : stdin_name( "(stdin)" ), longest_name( 0 ), first_post( false )
117 {
118 if( verbosity <= 0 ) return;
119 const unsigned stdin_name_len = std::strlen( stdin_name );
120 for( unsigned i = 0; i < filenames.size(); ++i )
121 {
122 const std::string & s = filenames[i];
123 const unsigned len = ( s == "-" ) ? stdin_name_len : s.size();
124 if( longest_name < len ) longest_name = len;
125 }
126 if( longest_name == 0 ) longest_name = stdin_name_len;
127 }
128
Pretty_print(const std::string & filename)129 Pretty_print( const std::string & filename )
130 : stdin_name( "(stdin)" ), first_post( false )
131 {
132 const unsigned stdin_name_len = std::strlen( stdin_name );
133 longest_name = ( filename == "-" ) ? stdin_name_len : filename.size();
134 if( longest_name == 0 ) longest_name = stdin_name_len;
135 set_name( filename );
136 }
137
set_name(const std::string & filename)138 void set_name( const std::string & filename )
139 {
140 if( filename.size() && filename != "-" ) name_ = filename;
141 else name_ = stdin_name;
142 padded_name = " "; padded_name += name_; padded_name += ": ";
143 if( longest_name > name_.size() )
144 padded_name.append( longest_name - name_.size(), ' ' );
145 first_post = true;
146 }
147
reset()148 void reset() const { if( name_.size() ) first_post = true; }
name()149 const char * name() const { return name_.c_str(); }
150 void operator()( const char * const msg = 0, FILE * const f = stderr ) const;
151 };
152
153
154 class CRC32
155 {
156 uint32_t data[256]; // Table of CRCs of all 8-bit messages.
157
158 public:
CRC32()159 CRC32()
160 {
161 for( unsigned n = 0; n < 256; ++n )
162 {
163 unsigned c = n;
164 for( int k = 0; k < 8; ++k )
165 { if( c & 1 ) c = 0xEDB88320U ^ ( c >> 1 ); else c >>= 1; }
166 data[n] = c;
167 }
168 }
169
170 uint32_t operator[]( const uint8_t byte ) const { return data[byte]; }
171
update_byte(uint32_t & crc,const uint8_t byte)172 void update_byte( uint32_t & crc, const uint8_t byte ) const
173 { crc = data[(crc^byte)&0xFF] ^ ( crc >> 8 ); }
174
update_buf(uint32_t & crc,const uint8_t * const buffer,const int size)175 void update_buf( uint32_t & crc, const uint8_t * const buffer,
176 const int size ) const
177 {
178 uint32_t c = crc;
179 for( int i = 0; i < size; ++i )
180 c = data[(c^buffer[i])&0xFF] ^ ( c >> 8 );
181 crc = c;
182 }
183
compute_crc(const uint8_t * const buffer,const long long size)184 uint32_t compute_crc( const uint8_t * const buffer,
185 const long long size ) const
186 {
187 uint32_t crc = 0xFFFFFFFFU;
188 for( long long i = 0; i < size; ++i )
189 crc = data[(crc^buffer[i])&0xFF] ^ ( crc >> 8 );
190 return crc ^ 0xFFFFFFFFU;
191 }
192 };
193
194 extern const CRC32 crc32;
195
196
isvalid_ds(const unsigned dictionary_size)197 inline bool isvalid_ds( const unsigned dictionary_size )
198 { return ( dictionary_size >= min_dictionary_size &&
199 dictionary_size <= max_dictionary_size ); }
200
201
real_bits(unsigned value)202 inline int real_bits( unsigned value )
203 {
204 int bits = 0;
205 while( value > 0 ) { value >>= 1; ++bits; }
206 return bits;
207 }
208
209
210 const uint8_t lzip_magic[4] = { 0x4C, 0x5A, 0x49, 0x50 }; // "LZIP"
211
212 struct Lzip_header
213 {
214 uint8_t data[6]; // 0-3 magic bytes
215 // 4 version
216 // 5 coded dictionary size
217 enum { size = 6 };
218
set_magicLzip_header219 void set_magic() { std::memcpy( data, lzip_magic, 4 ); data[4] = 1; }
verify_magicLzip_header220 bool verify_magic() const
221 { return ( std::memcmp( data, lzip_magic, 4 ) == 0 ); }
222
verify_prefixLzip_header223 bool verify_prefix( const int sz ) const // detect (truncated) header
224 {
225 for( int i = 0; i < sz && i < 4; ++i )
226 if( data[i] != lzip_magic[i] ) return false;
227 return ( sz > 0 );
228 }
verify_corruptLzip_header229 bool verify_corrupt() const // detect corrupt header
230 {
231 int matches = 0;
232 for( int i = 0; i < 4; ++i )
233 if( data[i] == lzip_magic[i] ) ++matches;
234 return ( matches > 1 && matches < 4 );
235 }
236
versionLzip_header237 uint8_t version() const { return data[4]; }
verify_versionLzip_header238 bool verify_version() const { return ( data[4] == 1 ); }
239
dictionary_sizeLzip_header240 unsigned dictionary_size() const
241 {
242 unsigned sz = ( 1 << ( data[5] & 0x1F ) );
243 if( sz > min_dictionary_size )
244 sz -= ( sz / 16 ) * ( ( data[5] >> 5 ) & 7 );
245 return sz;
246 }
247
dictionary_sizeLzip_header248 bool dictionary_size( const unsigned sz )
249 {
250 if( !isvalid_ds( sz ) ) return false;
251 data[5] = real_bits( sz - 1 );
252 if( sz > min_dictionary_size )
253 {
254 const unsigned base_size = 1 << data[5];
255 const unsigned fraction = base_size / 16;
256 for( unsigned i = 7; i >= 1; --i )
257 if( base_size - ( i * fraction ) >= sz )
258 { data[5] |= ( i << 5 ); break; }
259 }
260 return true;
261 }
262
verifyLzip_header263 bool verify( const bool ignore_bad_ds ) const
264 { return verify_magic() && verify_version() &&
265 ( ignore_bad_ds || isvalid_ds( dictionary_size() ) ); }
266 };
267
268
269 struct Lzip_trailer
270 {
271 uint8_t data[20]; // 0-3 CRC32 of the uncompressed data
272 // 4-11 size of the uncompressed data
273 // 12-19 member size including header and trailer
274 enum { size = 20 };
275
data_crcLzip_trailer276 unsigned data_crc() const
277 {
278 unsigned tmp = 0;
279 for( int i = 3; i >= 0; --i ) { tmp <<= 8; tmp += data[i]; }
280 return tmp;
281 }
282
data_crcLzip_trailer283 void data_crc( unsigned crc )
284 { for( int i = 0; i <= 3; ++i ) { data[i] = (uint8_t)crc; crc >>= 8; } }
285
data_sizeLzip_trailer286 unsigned long long data_size() const
287 {
288 unsigned long long tmp = 0;
289 for( int i = 11; i >= 4; --i ) { tmp <<= 8; tmp += data[i]; }
290 return tmp;
291 }
292
data_sizeLzip_trailer293 void data_size( unsigned long long sz )
294 { for( int i = 4; i <= 11; ++i ) { data[i] = (uint8_t)sz; sz >>= 8; } }
295
member_sizeLzip_trailer296 unsigned long long member_size() const
297 {
298 unsigned long long tmp = 0;
299 for( int i = 19; i >= 12; --i ) { tmp <<= 8; tmp += data[i]; }
300 return tmp;
301 }
302
member_sizeLzip_trailer303 void member_size( unsigned long long sz )
304 { for( int i = 12; i <= 19; ++i ) { data[i] = (uint8_t)sz; sz >>= 8; } }
305
verify_consistencyLzip_trailer306 bool verify_consistency() const // check internal consistency
307 {
308 const unsigned crc = data_crc();
309 const unsigned long long dsize = data_size();
310 if( ( crc == 0 ) != ( dsize == 0 ) ) return false;
311 const unsigned long long msize = member_size();
312 if( msize < min_member_size ) return false;
313 const unsigned long long mlimit = ( 9 * dsize + 7 ) / 8 + min_member_size;
314 if( mlimit > dsize && msize > mlimit ) return false;
315 const unsigned long long dlimit = 7090 * ( msize - 26 ) - 1;
316 if( dlimit > msize && dsize > dlimit ) return false;
317 return true;
318 }
319 };
320
321
322 struct Bad_byte
323 {
324 enum Mode { literal, delta, flip };
325 long long pos;
326 Mode mode;
327 uint8_t value;
328
Bad_byteBad_byte329 Bad_byte() : pos( -1 ), mode( literal ), value( 0 ) {}
operatorBad_byte330 uint8_t operator()( const uint8_t old_value ) const
331 {
332 if( mode == delta ) return old_value + value;
333 if( mode == flip ) return old_value ^ value;
334 return value;
335 }
336 };
337
338
339 #ifndef INT64_MAX
340 #define INT64_MAX 0x7FFFFFFFFFFFFFFFLL
341 #endif
342
343 class Block
344 {
345 long long pos_, size_; // pos + size <= INT64_MAX
346
347 public:
Block(const long long p,const long long s)348 Block( const long long p, const long long s ) : pos_( p ), size_( s ) {}
349
pos()350 long long pos() const { return pos_; }
size()351 long long size() const { return size_; }
end()352 long long end() const { return pos_ + size_; }
353
pos(const long long p)354 void pos( const long long p ) { pos_ = p; }
size(const long long s)355 void size( const long long s ) { size_ = s; }
356
357 bool operator==( const Block & b ) const
358 { return pos_ == b.pos_ && size_ == b.size_; }
359 bool operator!=( const Block & b ) const
360 { return pos_ != b.pos_ || size_ != b.size_; }
361
362 bool operator<( const Block & b ) const { return pos_ < b.pos_; }
363
includes(const long long pos)364 bool includes( const long long pos ) const
365 { return ( pos_ <= pos && end() > pos ); }
overlaps(const Block & b)366 bool overlaps( const Block & b ) const
367 { return ( pos_ < b.end() && b.pos_ < end() ); }
overlaps(const long long pos,const long long size)368 bool overlaps( const long long pos, const long long size ) const
369 { return ( pos_ < pos + size && pos < end() ); }
370
shift(Block & b)371 void shift( Block & b ) { ++size_; ++b.pos_; --b.size_; }
372 Block split( const long long pos );
373 };
374
375
376 struct Member_list // members/gaps/tdata to be dumped/removed/stripped
377 {
378 bool damaged;
379 bool tdata;
380 bool in, rin;
381 std::vector< Block > range_vector, rrange_vector;
382
Member_listMember_list383 Member_list() : damaged( false ), tdata( false ), in( true ), rin( true ) {}
384 void parse( const char * p );
385
rangeMember_list386 bool range() const { return range_vector.size() || rrange_vector.size(); }
387
388 // blocks is the sum of members + gaps, excluding trailing data
includesMember_list389 bool includes( const long i, const long blocks ) const
390 {
391 for( unsigned j = 0; j < range_vector.size(); ++j )
392 {
393 if( range_vector[j].pos() > i ) break;
394 if( range_vector[j].end() > i ) return in;
395 }
396 if( i >= 0 && i < blocks )
397 for( unsigned j = 0; j < rrange_vector.size(); ++j )
398 {
399 if( rrange_vector[j].pos() > blocks - i - 1 ) break;
400 if( rrange_vector[j].end() > blocks - i - 1 ) return rin;
401 }
402 return !in || !rin;
403 }
404 };
405
406
407 struct Error
408 {
409 const char * const msg;
ErrorError410 explicit Error( const char * const s ) : msg( s ) {}
411 };
412
positive_diff(const unsigned long long x,const unsigned long long y)413 inline unsigned long long positive_diff( const unsigned long long x,
414 const unsigned long long y )
415 { return ( ( x > y ) ? x - y : 0 ); }
416
set_retval(int & retval,const int new_val)417 inline void set_retval( int & retval, const int new_val )
418 { if( retval < new_val ) retval = new_val; }
419
420 const char * const bad_magic_msg = "Bad magic number (file not in lzip format).";
421 const char * const bad_dict_msg = "Invalid dictionary size in member header.";
422 const char * const corrupt_mm_msg = "Corrupt header in multimember file.";
423 const char * const trailing_msg = "Trailing data not allowed.";
424
425 // defined in alone_to_lz.cc
426 int alone_to_lz( const int infd, const Pretty_print & pp );
427
428 // defined in decoder.cc
429 long long readblock( const int fd, uint8_t * const buf, const long long size );
430 long long writeblock( const int fd, const uint8_t * const buf,
431 const long long size );
432
433 // defined in dump_remove.cc
434 int dump_members( const std::vector< std::string > & filenames,
435 const std::string & default_output_filename,
436 const Member_list & member_list, const bool force,
437 bool ignore_errors, bool ignore_trailing,
438 const bool loose_trailing, const bool strip,
439 const bool to_stdout );
440 int remove_members( const std::vector< std::string > & filenames,
441 const Member_list & member_list, bool ignore_errors,
442 bool ignore_trailing, const bool loose_trailing );
443
444 // defined in list.cc
445 int list_files( const std::vector< std::string > & filenames,
446 const bool ignore_errors,
447 const bool ignore_trailing, const bool loose_trailing );
448
449 // defined in lzip_index.cc
450 int seek_read( const int fd, uint8_t * const buf, const int size,
451 const long long pos );
452
453 // defined in lunzcrash.cc
454 int lunzcrash( const std::string & input_filename );
455 int md5sum_files( const std::vector< std::string > & filenames );
456
457 // defined in main.cc
458 extern const char * const program_name;
459 extern std::string output_filename; // global vars for output file
460 extern int outfd;
461 struct stat;
462 const char * bad_version( const unsigned version );
463 const char * format_ds( const unsigned dictionary_size );
464 void show_header( const unsigned dictionary_size );
465 int open_instream( const char * const name, struct stat * const in_statsp,
466 const bool one_to_one, const bool reg_only = false );
467 int open_truncable_stream( const char * const name,
468 struct stat * const in_statsp );
469 bool open_outstream( const bool force, const bool protect,
470 const bool rw = false, const bool skipping = true );
471 bool file_exists( const std::string & filename );
472 void cleanup_and_fail( const int retval );
473 void set_signal_handler();
474 int close_outstream( const struct stat * const in_statsp );
475 std::string insert_fixed( std::string name );
476 void show_error( const char * const msg, const int errcode = 0,
477 const bool help = false );
478 void show_file_error( const char * const filename, const char * const msg,
479 const int errcode = 0 );
480 void internal_error( const char * const msg );
481 void show_2file_error( const char * const msg1, const char * const name1,
482 const char * const name2, const char * const msg2 );
483 class Range_decoder;
484 void show_dprogress( const unsigned long long cfile_size = 0,
485 const unsigned long long partial_size = 0,
486 const Range_decoder * const d = 0,
487 const Pretty_print * const p = 0 );
488
489 // defined in merge.cc
490 bool copy_file( const int infd, const int outfd,
491 const long long max_size = -1 );
492 int test_member_from_file( const int infd, const unsigned long long msize,
493 long long * const failure_posp = 0 );
494 int merge_files( const std::vector< std::string > & filenames,
495 const std::string & default_output_filename,
496 const char terminator, const bool force );
497
498 // defined in nrep_stats.cc
499 int print_nrep_stats( const std::vector< std::string > & filenames,
500 const int repeated_byte, const bool ignore_errors,
501 const bool ignore_trailing, const bool loose_trailing );
502
503 // defined in range_dec.cc
504 const char * format_num( unsigned long long num,
505 unsigned long long limit = -1ULL,
506 const int set_prefix = 0 );
507 bool safe_seek( const int fd, const long long pos );
508 int range_decompress( const std::string & input_filename,
509 const std::string & default_output_filename,
510 Block range, const bool force, const bool ignore_errors,
511 const bool ignore_trailing, const bool loose_trailing,
512 const bool to_stdout );
513
514 // defined in repair.cc
515 long long seek_write( const int fd, const uint8_t * const buf,
516 const long long size, const long long pos );
517 uint8_t * read_member( const int infd, const long long mpos,
518 const long long msize );
519 int repair_file( const std::string & input_filename,
520 const std::string & default_output_filename,
521 const char terminator, const bool force );
522 int debug_delay( const std::string & input_filename, Block range,
523 const char terminator );
524 int debug_repair( const std::string & input_filename,
525 const Bad_byte & bad_byte, const char terminator );
526 int debug_decompress( const std::string & input_filename,
527 const Bad_byte & bad_byte, const bool show_packets );
528
529 // defined in reproduce.cc
530 int reproduce_file( const std::string & input_filename,
531 const std::string & default_output_filename,
532 const char * const lzip_name,
533 const char * const reference_filename,
534 const int lzip_level, const char terminator,
535 const bool force );
536 int debug_reproduce_file( const std::string & input_filename,
537 const char * const lzip_name,
538 const char * const reference_filename,
539 const Block & range, const int sector_size,
540 const int lzip_level );
541
542 // defined in split.cc
543 int split_file( const std::string & input_filename,
544 const std::string & default_output_filename, const bool force );
545