1 /*-
2 * Copyright (c) 2003-2010 Tim Kientzle
3 * Copyright (c) 2009-2012 Michihiro NAKAJIMA
4 * All rights reserved.
5 *
6 * Redistribution and use in source and binary forms, with or without
7 * modification, are permitted provided that the following conditions
8 * are met:
9 * 1. Redistributions of source code must retain the above copyright
10 * notice, this list of conditions and the following disclaimer.
11 * 2. Redistributions in binary form must reproduce the above copyright
12 * notice, this list of conditions and the following disclaimer in the
13 * documentation and/or other materials provided with the distribution.
14 *
15 * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
16 * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
17 * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
18 * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
19 * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
20 * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
21 * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
22 * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
23 * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
24 * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
25 */
26
27 #include "archive_platform.h"
28
29 #ifdef HAVE_ERRNO_H
30 #include <errno.h>
31 #endif
32 #ifdef HAVE_LIMITS_H
33 #include <limits.h>
34 #endif
35 #ifdef HAVE_STDLIB_H
36 #include <stdlib.h>
37 #endif
38 #ifdef HAVE_STRING_H
39 #include <string.h>
40 #endif
41 #include <time.h>
42 #ifdef HAVE_LZMA_H
43 #include <lzma.h>
44 #endif
45
46 #include "archive.h"
47 #include "archive_endian.h"
48 #include "archive_private.h"
49 #include "archive_write_private.h"
50
51 #if ARCHIVE_VERSION_NUMBER < 4000000
52 int
archive_write_set_compression_lzip(struct archive * a)53 archive_write_set_compression_lzip(struct archive *a)
54 {
55 __archive_write_filters_free(a);
56 return (archive_write_add_filter_lzip(a));
57 }
58
59 int
archive_write_set_compression_lzma(struct archive * a)60 archive_write_set_compression_lzma(struct archive *a)
61 {
62 __archive_write_filters_free(a);
63 return (archive_write_add_filter_lzma(a));
64 }
65
66 int
archive_write_set_compression_xz(struct archive * a)67 archive_write_set_compression_xz(struct archive *a)
68 {
69 __archive_write_filters_free(a);
70 return (archive_write_add_filter_xz(a));
71 }
72
73 #endif
74
75 #ifndef HAVE_LZMA_H
76 int
archive_write_add_filter_xz(struct archive * a)77 archive_write_add_filter_xz(struct archive *a)
78 {
79 archive_set_error(a, ARCHIVE_ERRNO_MISC,
80 "xz compression not supported on this platform");
81 return (ARCHIVE_FATAL);
82 }
83
84 int
archive_write_add_filter_lzma(struct archive * a)85 archive_write_add_filter_lzma(struct archive *a)
86 {
87 archive_set_error(a, ARCHIVE_ERRNO_MISC,
88 "lzma compression not supported on this platform");
89 return (ARCHIVE_FATAL);
90 }
91
92 int
archive_write_add_filter_lzip(struct archive * a)93 archive_write_add_filter_lzip(struct archive *a)
94 {
95 archive_set_error(a, ARCHIVE_ERRNO_MISC,
96 "lzma compression not supported on this platform");
97 return (ARCHIVE_FATAL);
98 }
99 #else
100 /* Don't compile this if we don't have liblzma. */
101
102 struct private_data {
103 int compression_level;
104 uint32_t threads;
105 lzma_stream stream;
106 lzma_filter lzmafilters[2];
107 lzma_options_lzma lzma_opt;
108 int64_t total_in;
109 unsigned char *compressed;
110 size_t compressed_buffer_size;
111 int64_t total_out;
112 /* the CRC32 value of uncompressed data for lzip */
113 uint32_t crc32;
114 };
115
116 static int archive_compressor_xz_options(struct archive_write_filter *,
117 const char *, const char *);
118 static int archive_compressor_xz_open(struct archive_write_filter *);
119 static int archive_compressor_xz_write(struct archive_write_filter *,
120 const void *, size_t);
121 static int archive_compressor_xz_close(struct archive_write_filter *);
122 static int archive_compressor_xz_free(struct archive_write_filter *);
123 static int drive_compressor(struct archive_write_filter *,
124 struct private_data *, int finishing);
125
126 struct option_value {
127 uint32_t dict_size;
128 uint32_t nice_len;
129 lzma_match_finder mf;
130 };
131 static const struct option_value option_values[] = {
132 { 1 << 16, 32, LZMA_MF_HC3},
133 { 1 << 20, 32, LZMA_MF_HC3},
134 { 3 << 19, 32, LZMA_MF_HC4},
135 { 1 << 21, 32, LZMA_MF_BT4},
136 { 3 << 20, 32, LZMA_MF_BT4},
137 { 1 << 22, 32, LZMA_MF_BT4},
138 { 1 << 23, 64, LZMA_MF_BT4},
139 { 1 << 24, 64, LZMA_MF_BT4},
140 { 3 << 23, 64, LZMA_MF_BT4},
141 { 1 << 25, 64, LZMA_MF_BT4}
142 };
143
144 static int
common_setup(struct archive_write_filter * f)145 common_setup(struct archive_write_filter *f)
146 {
147 struct private_data *data;
148 struct archive_write *a = (struct archive_write *)f->archive;
149 data = calloc(1, sizeof(*data));
150 if (data == NULL) {
151 archive_set_error(&a->archive, ENOMEM, "Out of memory");
152 return (ARCHIVE_FATAL);
153 }
154 f->data = data;
155 data->compression_level = LZMA_PRESET_DEFAULT;
156 data->threads = 1;
157 f->open = &archive_compressor_xz_open;
158 f->close = archive_compressor_xz_close;
159 f->free = archive_compressor_xz_free;
160 f->options = &archive_compressor_xz_options;
161 return (ARCHIVE_OK);
162 }
163
164 /*
165 * Add an xz compression filter to this write handle.
166 */
167 int
archive_write_add_filter_xz(struct archive * _a)168 archive_write_add_filter_xz(struct archive *_a)
169 {
170 struct archive_write_filter *f;
171 int r;
172
173 archive_check_magic(_a, ARCHIVE_WRITE_MAGIC,
174 ARCHIVE_STATE_NEW, "archive_write_add_filter_xz");
175 f = __archive_write_allocate_filter(_a);
176 r = common_setup(f);
177 if (r == ARCHIVE_OK) {
178 f->code = ARCHIVE_FILTER_XZ;
179 f->name = "xz";
180 }
181 return (r);
182 }
183
184 /* LZMA is handled identically, we just need a different compression
185 * code set. (The liblzma setup looks at the code to determine
186 * the one place that XZ and LZMA require different handling.) */
187 int
archive_write_add_filter_lzma(struct archive * _a)188 archive_write_add_filter_lzma(struct archive *_a)
189 {
190 struct archive_write_filter *f;
191 int r;
192
193 archive_check_magic(_a, ARCHIVE_WRITE_MAGIC,
194 ARCHIVE_STATE_NEW, "archive_write_add_filter_lzma");
195 f = __archive_write_allocate_filter(_a);
196 r = common_setup(f);
197 if (r == ARCHIVE_OK) {
198 f->code = ARCHIVE_FILTER_LZMA;
199 f->name = "lzma";
200 }
201 return (r);
202 }
203
204 int
archive_write_add_filter_lzip(struct archive * _a)205 archive_write_add_filter_lzip(struct archive *_a)
206 {
207 struct archive_write_filter *f;
208 int r;
209
210 archive_check_magic(_a, ARCHIVE_WRITE_MAGIC,
211 ARCHIVE_STATE_NEW, "archive_write_add_filter_lzip");
212 f = __archive_write_allocate_filter(_a);
213 r = common_setup(f);
214 if (r == ARCHIVE_OK) {
215 f->code = ARCHIVE_FILTER_LZIP;
216 f->name = "lzip";
217 }
218 return (r);
219 }
220
221 static int
archive_compressor_xz_init_stream(struct archive_write_filter * f,struct private_data * data)222 archive_compressor_xz_init_stream(struct archive_write_filter *f,
223 struct private_data *data)
224 {
225 static const lzma_stream lzma_stream_init_data = LZMA_STREAM_INIT;
226 int ret;
227 #ifdef HAVE_LZMA_STREAM_ENCODER_MT
228 lzma_mt mt_options;
229 #endif
230
231 data->stream = lzma_stream_init_data;
232 data->stream.next_out = data->compressed;
233 data->stream.avail_out = data->compressed_buffer_size;
234 if (f->code == ARCHIVE_FILTER_XZ) {
235 #ifdef HAVE_LZMA_STREAM_ENCODER_MT
236 if (data->threads != 1) {
237 memset(&mt_options, 0, sizeof(mt_options));
238 mt_options.threads = data->threads;
239 mt_options.timeout = 300;
240 mt_options.filters = data->lzmafilters;
241 mt_options.check = LZMA_CHECK_CRC64;
242 ret = lzma_stream_encoder_mt(&(data->stream),
243 &mt_options);
244 } else
245 #endif
246 ret = lzma_stream_encoder(&(data->stream),
247 data->lzmafilters, LZMA_CHECK_CRC64);
248 } else if (f->code == ARCHIVE_FILTER_LZMA) {
249 ret = lzma_alone_encoder(&(data->stream), &data->lzma_opt);
250 } else { /* ARCHIVE_FILTER_LZIP */
251 int dict_size = data->lzma_opt.dict_size;
252 int ds, log2dic, wedges;
253
254 /* Calculate a coded dictionary size */
255 if (dict_size < (1 << 12) || dict_size > (1 << 29)) {
256 archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
257 "Unacceptable dictionary size for lzip: %d",
258 dict_size);
259 return (ARCHIVE_FATAL);
260 }
261 for (log2dic = 29; log2dic >= 12; log2dic--) {
262 if (dict_size & (1 << log2dic))
263 break;
264 }
265 if (dict_size > (1 << log2dic)) {
266 log2dic++;
267 wedges =
268 ((1 << log2dic) - dict_size) / (1 << (log2dic - 4));
269 } else
270 wedges = 0;
271 ds = ((wedges << 5) & 0xe0) | (log2dic & 0x1f);
272
273 data->crc32 = 0;
274 /* Make a header */
275 data->compressed[0] = 0x4C;
276 data->compressed[1] = 0x5A;
277 data->compressed[2] = 0x49;
278 data->compressed[3] = 0x50;
279 data->compressed[4] = 1;/* Version */
280 data->compressed[5] = (unsigned char)ds;
281 data->stream.next_out += 6;
282 data->stream.avail_out -= 6;
283
284 ret = lzma_raw_encoder(&(data->stream), data->lzmafilters);
285 }
286 if (ret == LZMA_OK)
287 return (ARCHIVE_OK);
288
289 switch (ret) {
290 case LZMA_MEM_ERROR:
291 archive_set_error(f->archive, ENOMEM,
292 "Internal error initializing compression library: "
293 "Cannot allocate memory");
294 break;
295 default:
296 archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
297 "Internal error initializing compression library: "
298 "It's a bug in liblzma");
299 break;
300 }
301 return (ARCHIVE_FATAL);
302 }
303
304 /*
305 * Setup callback.
306 */
307 static int
archive_compressor_xz_open(struct archive_write_filter * f)308 archive_compressor_xz_open(struct archive_write_filter *f)
309 {
310 struct private_data *data = f->data;
311 int ret;
312
313 if (data->compressed == NULL) {
314 size_t bs = 65536, bpb;
315 if (f->archive->magic == ARCHIVE_WRITE_MAGIC) {
316 /* Buffer size should be a multiple number of the bytes
317 * per block for performance. */
318 bpb = archive_write_get_bytes_per_block(f->archive);
319 if (bpb > bs)
320 bs = bpb;
321 else if (bpb != 0)
322 bs -= bs % bpb;
323 }
324 data->compressed_buffer_size = bs;
325 data->compressed = malloc(data->compressed_buffer_size);
326 if (data->compressed == NULL) {
327 archive_set_error(f->archive, ENOMEM,
328 "Can't allocate data for compression buffer");
329 return (ARCHIVE_FATAL);
330 }
331 }
332
333 f->write = archive_compressor_xz_write;
334
335 /* Initialize compression library. */
336 if (f->code == ARCHIVE_FILTER_LZIP) {
337 const struct option_value *val =
338 &option_values[data->compression_level];
339
340 data->lzma_opt.dict_size = val->dict_size;
341 data->lzma_opt.preset_dict = NULL;
342 data->lzma_opt.preset_dict_size = 0;
343 data->lzma_opt.lc = LZMA_LC_DEFAULT;
344 data->lzma_opt.lp = LZMA_LP_DEFAULT;
345 data->lzma_opt.pb = LZMA_PB_DEFAULT;
346 data->lzma_opt.mode =
347 data->compression_level<= 2? LZMA_MODE_FAST:LZMA_MODE_NORMAL;
348 data->lzma_opt.nice_len = val->nice_len;
349 data->lzma_opt.mf = val->mf;
350 data->lzma_opt.depth = 0;
351 data->lzmafilters[0].id = LZMA_FILTER_LZMA1;
352 data->lzmafilters[0].options = &data->lzma_opt;
353 data->lzmafilters[1].id = LZMA_VLI_UNKNOWN;/* Terminate */
354 } else {
355 if (lzma_lzma_preset(&data->lzma_opt, data->compression_level)) {
356 archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
357 "Internal error initializing compression library");
358 }
359 data->lzmafilters[0].id = LZMA_FILTER_LZMA2;
360 data->lzmafilters[0].options = &data->lzma_opt;
361 data->lzmafilters[1].id = LZMA_VLI_UNKNOWN;/* Terminate */
362 }
363 ret = archive_compressor_xz_init_stream(f, data);
364 if (ret == LZMA_OK) {
365 f->data = data;
366 return (0);
367 }
368 return (ARCHIVE_FATAL);
369 }
370
371 /*
372 * Set write options.
373 */
374 static int
archive_compressor_xz_options(struct archive_write_filter * f,const char * key,const char * value)375 archive_compressor_xz_options(struct archive_write_filter *f,
376 const char *key, const char *value)
377 {
378 struct private_data *data = (struct private_data *)f->data;
379
380 if (strcmp(key, "compression-level") == 0) {
381 if (value == NULL || !(value[0] >= '0' && value[0] <= '9') ||
382 value[1] != '\0') {
383 archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
384 "compression-level invalid");
385 return (ARCHIVE_FAILED);
386 }
387 data->compression_level = value[0] - '0';
388 if (data->compression_level > 9)
389 data->compression_level = 9;
390 return (ARCHIVE_OK);
391 } else if (strcmp(key, "threads") == 0) {
392 char *endptr;
393 unsigned long val;
394
395 if (value == NULL) {
396 archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
397 "threads option requires an argument");
398 return (ARCHIVE_FAILED);
399 }
400 errno = 0;
401 val = strtoul(value, &endptr, 10);
402 if (errno != 0 || *endptr != '\0' || val > (unsigned)INT_MAX) {
403 data->threads = 1;
404 archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
405 "threads invalid");
406 return (ARCHIVE_FAILED);
407 }
408 data->threads = (int)val;
409 if (data->threads == 0) {
410 #ifdef HAVE_LZMA_STREAM_ENCODER_MT
411 data->threads = lzma_cputhreads();
412 #else
413 data->threads = 1;
414 #endif
415 }
416 return (ARCHIVE_OK);
417 }
418
419 /* Note: The "warn" return is just to inform the options
420 * supervisor that we didn't handle it. It will generate
421 * a suitable error if no one used this option. */
422 return (ARCHIVE_WARN);
423 }
424
425 /*
426 * Write data to the compressed stream.
427 */
428 static int
archive_compressor_xz_write(struct archive_write_filter * f,const void * buff,size_t length)429 archive_compressor_xz_write(struct archive_write_filter *f,
430 const void *buff, size_t length)
431 {
432 struct private_data *data = (struct private_data *)f->data;
433 int ret;
434
435 /* Update statistics */
436 data->total_in += length;
437 if (f->code == ARCHIVE_FILTER_LZIP)
438 data->crc32 = lzma_crc32(buff, length, data->crc32);
439
440 /* Compress input data to output buffer */
441 data->stream.next_in = buff;
442 data->stream.avail_in = length;
443 if ((ret = drive_compressor(f, data, 0)) != ARCHIVE_OK)
444 return (ret);
445
446 return (ARCHIVE_OK);
447 }
448
449
450 /*
451 * Finish the compression...
452 */
453 static int
archive_compressor_xz_close(struct archive_write_filter * f)454 archive_compressor_xz_close(struct archive_write_filter *f)
455 {
456 struct private_data *data = (struct private_data *)f->data;
457 int ret;
458
459 ret = drive_compressor(f, data, 1);
460 if (ret == ARCHIVE_OK) {
461 data->total_out +=
462 data->compressed_buffer_size - data->stream.avail_out;
463 ret = __archive_write_filter(f->next_filter,
464 data->compressed,
465 data->compressed_buffer_size - data->stream.avail_out);
466 if (f->code == ARCHIVE_FILTER_LZIP && ret == ARCHIVE_OK) {
467 archive_le32enc(data->compressed, data->crc32);
468 archive_le64enc(data->compressed+4, data->total_in);
469 archive_le64enc(data->compressed+12, data->total_out + 20);
470 ret = __archive_write_filter(f->next_filter,
471 data->compressed, 20);
472 }
473 }
474 lzma_end(&(data->stream));
475 return ret;
476 }
477
478 static int
archive_compressor_xz_free(struct archive_write_filter * f)479 archive_compressor_xz_free(struct archive_write_filter *f)
480 {
481 struct private_data *data = (struct private_data *)f->data;
482 free(data->compressed);
483 free(data);
484 f->data = NULL;
485 return (ARCHIVE_OK);
486 }
487
488 /*
489 * Utility function to push input data through compressor,
490 * writing full output blocks as necessary.
491 *
492 * Note that this handles both the regular write case (finishing ==
493 * false) and the end-of-archive case (finishing == true).
494 */
495 static int
drive_compressor(struct archive_write_filter * f,struct private_data * data,int finishing)496 drive_compressor(struct archive_write_filter *f,
497 struct private_data *data, int finishing)
498 {
499 int ret;
500
501 for (;;) {
502 if (data->stream.avail_out == 0) {
503 data->total_out += data->compressed_buffer_size;
504 ret = __archive_write_filter(f->next_filter,
505 data->compressed,
506 data->compressed_buffer_size);
507 if (ret != ARCHIVE_OK)
508 return (ARCHIVE_FATAL);
509 data->stream.next_out = data->compressed;
510 data->stream.avail_out = data->compressed_buffer_size;
511 }
512
513 /* If there's nothing to do, we're done. */
514 if (!finishing && data->stream.avail_in == 0)
515 return (ARCHIVE_OK);
516
517 ret = lzma_code(&(data->stream),
518 finishing ? LZMA_FINISH : LZMA_RUN );
519
520 switch (ret) {
521 case LZMA_OK:
522 /* In non-finishing case, check if compressor
523 * consumed everything */
524 if (!finishing && data->stream.avail_in == 0)
525 return (ARCHIVE_OK);
526 /* In finishing case, this return always means
527 * there's more work */
528 break;
529 case LZMA_STREAM_END:
530 /* This return can only occur in finishing case. */
531 if (finishing)
532 return (ARCHIVE_OK);
533 archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
534 "lzma compression data error");
535 return (ARCHIVE_FATAL);
536 case LZMA_MEMLIMIT_ERROR:
537 archive_set_error(f->archive, ENOMEM,
538 "lzma compression error: "
539 "%ju MiB would have been needed",
540 (uintmax_t)((lzma_memusage(&(data->stream))
541 + 1024 * 1024 -1)
542 / (1024 * 1024)));
543 return (ARCHIVE_FATAL);
544 default:
545 /* Any other return value indicates an error. */
546 archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
547 "lzma compression failed:"
548 " lzma_code() call returned status %d",
549 ret);
550 return (ARCHIVE_FATAL);
551 }
552 }
553 }
554
555 #endif /* HAVE_LZMA_H */
556