xref: /freebsd/contrib/libarchive/libarchive/archive_read_support_filter_lzop.c (revision bd66c1b43e33540205dbc1187c2f2a15c58b57ba)
1 /*-
2  * Copyright (c) 2003-2007 Tim Kientzle
3  * Copyright (c) 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_UNISTD_H
30 #include <unistd.h>
31 #endif
32 #ifdef HAVE_ERRNO_H
33 #include <errno.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 #ifdef HAVE_UNISTD_H
42 #include <unistd.h>
43 #endif
44 #ifdef HAVE_LZO_LZOCONF_H
45 #include <lzo/lzoconf.h>
46 #endif
47 #ifdef HAVE_LZO_LZO1X_H
48 #include <lzo/lzo1x.h>
49 #endif
50 #ifdef HAVE_ZLIB_H
51 #include <zlib.h> /* for crc32 and adler32 */
52 #endif
53 
54 #include "archive.h"
55 #if !defined(HAVE_ZLIB_H) &&\
56      defined(HAVE_LZO_LZOCONF_H) && defined(HAVE_LZO_LZO1X_H)
57 #include "archive_crc32.h"
58 #endif
59 #include "archive_endian.h"
60 #include "archive_private.h"
61 #include "archive_read_private.h"
62 
63 #ifndef HAVE_ZLIB_H
64 #define adler32	lzo_adler32
65 #endif
66 
67 #define LZOP_HEADER_MAGIC "\x89\x4c\x5a\x4f\x00\x0d\x0a\x1a\x0a"
68 #define LZOP_HEADER_MAGIC_LEN 9
69 
70 #if defined(HAVE_LZO_LZOCONF_H) && defined(HAVE_LZO_LZO1X_H)
71 struct read_lzop {
72 	unsigned char	*out_block;
73 	size_t		 out_block_size;
74 	int64_t		 total_out;
75 	int		 flags;
76 	uint32_t	 compressed_cksum;
77 	uint32_t	 uncompressed_cksum;
78 	size_t		 compressed_size;
79 	size_t		 uncompressed_size;
80 	size_t		 unconsumed_bytes;
81 	char		 in_stream;
82 	char		 eof; /* True = found end of compressed data. */
83 };
84 
85 #define FILTER			0x0800
86 #define CRC32_HEADER		0x1000
87 #define EXTRA_FIELD		0x0040
88 #define ADLER32_UNCOMPRESSED	0x0001
89 #define ADLER32_COMPRESSED	0x0002
90 #define CRC32_UNCOMPRESSED	0x0100
91 #define CRC32_COMPRESSED	0x0200
92 #define MAX_BLOCK_SIZE		(64 * 1024 * 1024)
93 
94 static ssize_t  lzop_filter_read(struct archive_read_filter *, const void **);
95 static int	lzop_filter_close(struct archive_read_filter *);
96 #endif
97 
98 static int lzop_bidder_bid(struct archive_read_filter_bidder *,
99     struct archive_read_filter *);
100 static int lzop_bidder_init(struct archive_read_filter *);
101 
102 static const struct archive_read_filter_bidder_vtable
103 lzop_bidder_vtable = {
104 	.bid = lzop_bidder_bid,
105 	.init = lzop_bidder_init,
106 };
107 
108 int
archive_read_support_filter_lzop(struct archive * _a)109 archive_read_support_filter_lzop(struct archive *_a)
110 {
111 	struct archive_read *a = (struct archive_read *)_a;
112 
113 	if (__archive_read_register_bidder(a, NULL, NULL,
114 				&lzop_bidder_vtable) != ARCHIVE_OK)
115 		return (ARCHIVE_FATAL);
116 
117 	/* Signal the extent of lzop support with the return value here. */
118 #if defined(HAVE_LZO_LZOCONF_H) && defined(HAVE_LZO_LZO1X_H)
119 	return (ARCHIVE_OK);
120 #else
121 	/* Return ARCHIVE_WARN since this always uses an external program. */
122 	archive_set_error(_a, ARCHIVE_ERRNO_MISC,
123 	    "Using external lzop program for lzop decompression");
124 	return (ARCHIVE_WARN);
125 #endif
126 }
127 
128 /*
129  * Bidder just verifies the header and returns the number of verified bits.
130  */
131 static int
lzop_bidder_bid(struct archive_read_filter_bidder * self,struct archive_read_filter * filter)132 lzop_bidder_bid(struct archive_read_filter_bidder *self,
133     struct archive_read_filter *filter)
134 {
135 	const unsigned char *p;
136 	ssize_t avail;
137 
138 	(void)self; /* UNUSED */
139 
140 	p = __archive_read_filter_ahead(filter, LZOP_HEADER_MAGIC_LEN, &avail);
141 	if (p == NULL || avail == 0)
142 		return (0);
143 
144 	if (memcmp(p, LZOP_HEADER_MAGIC, LZOP_HEADER_MAGIC_LEN))
145 		return (0);
146 
147 	return (LZOP_HEADER_MAGIC_LEN * 8);
148 }
149 
150 #if !defined(HAVE_LZO_LZOCONF_H) || !defined(HAVE_LZO_LZO1X_H)
151 /*
152  * If we don't have the library on this system, we can't do the
153  * decompression directly.  We can, however, try to run "lzop -d"
154  * in case that's available.
155  */
156 static int
lzop_bidder_init(struct archive_read_filter * self)157 lzop_bidder_init(struct archive_read_filter *self)
158 {
159 	int r;
160 
161 	r = __archive_read_program(self, "lzop -d");
162 	/* Note: We set the format here even if __archive_read_program()
163 	 * above fails.  We do, after all, know what the format is
164 	 * even if we weren't able to read it. */
165 	self->code = ARCHIVE_FILTER_LZOP;
166 	self->name = "lzop";
167 	return (r);
168 }
169 #else
170 
171 static const struct archive_read_filter_vtable
172 lzop_reader_vtable = {
173 	.read = lzop_filter_read,
174 	.close = lzop_filter_close
175 };
176 
177 /*
178  * Initialize the filter object.
179  */
180 static int
lzop_bidder_init(struct archive_read_filter * self)181 lzop_bidder_init(struct archive_read_filter *self)
182 {
183 	struct read_lzop *state;
184 
185 	self->code = ARCHIVE_FILTER_LZOP;
186 	self->name = "lzop";
187 
188 	state = calloc(1, sizeof(*state));
189 	if (state == NULL) {
190 		archive_set_error(&self->archive->archive, ENOMEM,
191 		    "Can't allocate data for lzop decompression");
192 		return (ARCHIVE_FATAL);
193 	}
194 
195 	self->data = state;
196 	self->vtable = &lzop_reader_vtable;
197 
198 	return (ARCHIVE_OK);
199 }
200 
201 static int
consume_header(struct archive_read_filter * self)202 consume_header(struct archive_read_filter *self)
203 {
204 	struct read_lzop *state = (struct read_lzop *)self->data;
205 	const unsigned char *p, *_p;
206 	unsigned checksum, flags, len, method, version;
207 
208 	/*
209 	 * Check LZOP magic code.
210 	 */
211 	p = __archive_read_filter_ahead(self->upstream,
212 		LZOP_HEADER_MAGIC_LEN, NULL);
213 	if (p == NULL)
214 		return (ARCHIVE_EOF);
215 
216 	if (memcmp(p, LZOP_HEADER_MAGIC, LZOP_HEADER_MAGIC_LEN))
217 		return (ARCHIVE_EOF);
218 	__archive_read_filter_consume(self->upstream,
219 	    LZOP_HEADER_MAGIC_LEN);
220 
221 	p = __archive_read_filter_ahead(self->upstream, 29, NULL);
222 	if (p == NULL)
223 		goto truncated;
224 	_p = p;
225 	version = archive_be16dec(p);
226 	p += 4;/* version(2 bytes) + library version(2 bytes) */
227 
228 	if (version >= 0x940) {
229 		unsigned reqversion = archive_be16dec(p); p += 2;
230 		if (reqversion < 0x900) {
231 			archive_set_error(&self->archive->archive,
232 			    ARCHIVE_ERRNO_MISC, "Invalid required version");
233 			return (ARCHIVE_FAILED);
234 		}
235 	}
236 
237 	method = *p++;
238 	if (method < 1 || method > 3) {
239 		archive_set_error(&self->archive->archive, ARCHIVE_ERRNO_MISC,
240 		    "Unsupported method");
241 		return (ARCHIVE_FAILED);
242 	}
243 
244 	if (version >= 0x940) {
245 		unsigned level = *p++;
246 #if 0
247 		unsigned default_level[] = {0, 3, 1, 9};
248 #endif
249 		if (level == 0)
250 			/* Method is 1..3 here due to check above. */
251 #if 0	/* Avoid an error Clang Static Analyzer claims
252 	  "Value stored to 'level' is never read". */
253 			level = default_level[method];
254 #else
255 			;/* NOP */
256 #endif
257 		else if (level > 9) {
258 			archive_set_error(&self->archive->archive,
259 			    ARCHIVE_ERRNO_MISC, "Invalid level");
260 			return (ARCHIVE_FAILED);
261 		}
262 	}
263 
264 	flags = archive_be32dec(p); p += 4;
265 
266 	if (flags & FILTER)
267 		p += 4; /* Skip filter */
268 	p += 4; /* Skip mode */
269 	if (version >= 0x940)
270 		p += 8; /* Skip mtime */
271 	else
272 		p += 4; /* Skip mtime */
273 	len = *p++; /* Read filename length */
274 	len += p - _p;
275 	/* Make sure we have all bytes we need to calculate checksum. */
276 	p = __archive_read_filter_ahead(self->upstream, len + 4, NULL);
277 	if (p == NULL)
278 		goto truncated;
279 	if (flags & CRC32_HEADER)
280 		checksum = crc32(crc32(0, NULL, 0), p, len);
281 	else
282 		checksum = adler32(adler32(0, NULL, 0), p, len);
283 #ifndef DONT_FAIL_ON_CRC_ERROR
284 	if (archive_be32dec(p + len) != checksum)
285 		goto corrupted;
286 #endif
287 	__archive_read_filter_consume(self->upstream, len + 4);
288 	if (flags & EXTRA_FIELD) {
289 		/* Skip extra field */
290 		p = __archive_read_filter_ahead(self->upstream, 4, NULL);
291 		if (p == NULL)
292 			goto truncated;
293 		len = archive_be32dec(p);
294 		__archive_read_filter_consume(self->upstream,
295 		    (int64_t)len + 4 + 4);
296 	}
297 	state->flags = flags;
298 	state->in_stream = 1;
299 	return (ARCHIVE_OK);
300 truncated:
301 	archive_set_error(&self->archive->archive,
302 	    ARCHIVE_ERRNO_FILE_FORMAT, "Truncated lzop data");
303 	return (ARCHIVE_FAILED);
304 corrupted:
305 	archive_set_error(&self->archive->archive,
306 	    ARCHIVE_ERRNO_FILE_FORMAT, "Corrupted lzop header");
307 	return (ARCHIVE_FAILED);
308 }
309 
310 static int
consume_block_info(struct archive_read_filter * self)311 consume_block_info(struct archive_read_filter *self)
312 {
313 	struct read_lzop *state = (struct read_lzop *)self->data;
314 	const unsigned char *p;
315 	unsigned flags = state->flags;
316 
317 	p = __archive_read_filter_ahead(self->upstream, 4, NULL);
318 	if (p == NULL)
319 		goto truncated;
320 	state->uncompressed_size = archive_be32dec(p);
321 	__archive_read_filter_consume(self->upstream, 4);
322 	if (state->uncompressed_size == 0)
323 		return (ARCHIVE_EOF);
324 	if (state->uncompressed_size > MAX_BLOCK_SIZE)
325 		goto corrupted;
326 
327 	p = __archive_read_filter_ahead(self->upstream, 4, NULL);
328 	if (p == NULL)
329 		goto truncated;
330 	state->compressed_size = archive_be32dec(p);
331 	__archive_read_filter_consume(self->upstream, 4);
332 	if (state->compressed_size > state->uncompressed_size)
333 		goto corrupted;
334 
335 	if (flags & (CRC32_UNCOMPRESSED | ADLER32_UNCOMPRESSED)) {
336 		p = __archive_read_filter_ahead(self->upstream, 4, NULL);
337 		if (p == NULL)
338 			goto truncated;
339 		state->compressed_cksum = state->uncompressed_cksum =
340 		    archive_be32dec(p);
341 		__archive_read_filter_consume(self->upstream, 4);
342 	}
343 	if ((flags & (CRC32_COMPRESSED | ADLER32_COMPRESSED)) &&
344 	    state->compressed_size < state->uncompressed_size) {
345 		p = __archive_read_filter_ahead(self->upstream, 4, NULL);
346 		if (p == NULL)
347 			goto truncated;
348 		state->compressed_cksum = archive_be32dec(p);
349 		__archive_read_filter_consume(self->upstream, 4);
350 	}
351 	return (ARCHIVE_OK);
352 truncated:
353 	archive_set_error(&self->archive->archive,
354 	    ARCHIVE_ERRNO_FILE_FORMAT, "Truncated lzop data");
355 	return (ARCHIVE_FAILED);
356 corrupted:
357 	archive_set_error(&self->archive->archive,
358 	    ARCHIVE_ERRNO_FILE_FORMAT, "Corrupted lzop header");
359 	return (ARCHIVE_FAILED);
360 }
361 
362 static ssize_t
lzop_filter_read(struct archive_read_filter * self,const void ** p)363 lzop_filter_read(struct archive_read_filter *self, const void **p)
364 {
365 	struct read_lzop *state = (struct read_lzop *)self->data;
366 	const void *b;
367 	lzo_uint out_size;
368 	uint32_t cksum;
369 	int ret, r;
370 
371 	if (state->unconsumed_bytes) {
372 		__archive_read_filter_consume(self->upstream,
373 		    state->unconsumed_bytes);
374 		state->unconsumed_bytes = 0;
375 	}
376 	if (state->eof)
377 		return (0);
378 
379 	for (;;) {
380 		if (!state->in_stream) {
381 			ret = consume_header(self);
382 			if (ret < ARCHIVE_OK)
383 				return (ret);
384 			if (ret == ARCHIVE_EOF) {
385 				state->eof = 1;
386 				return (0);
387 			}
388 		}
389 		ret = consume_block_info(self);
390 		if (ret < ARCHIVE_OK)
391 			return (ret);
392 		if (ret == ARCHIVE_EOF)
393 			state->in_stream = 0;
394 		else
395 			break;
396 	}
397 
398 	if (state->out_block == NULL ||
399 	    state->out_block_size < state->uncompressed_size) {
400 		void *new_block;
401 
402 		new_block = realloc(state->out_block, state->uncompressed_size);
403 		if (new_block == NULL) {
404 			archive_set_error(&self->archive->archive, ENOMEM,
405 			    "Can't allocate data for lzop decompression");
406 			return (ARCHIVE_FATAL);
407 		}
408 		state->out_block = new_block;
409 		state->out_block_size = state->uncompressed_size;
410 	}
411 
412 	b = __archive_read_filter_ahead(self->upstream,
413 		state->compressed_size, NULL);
414 	if (b == NULL) {
415 		archive_set_error(&self->archive->archive,
416 		    ARCHIVE_ERRNO_FILE_FORMAT, "Truncated lzop data");
417 		return (ARCHIVE_FATAL);
418 	}
419 	if (state->flags & CRC32_COMPRESSED)
420 		cksum = crc32(crc32(0, NULL, 0), b, state->compressed_size);
421 	else if (state->flags & ADLER32_COMPRESSED)
422 		cksum = adler32(adler32(0, NULL, 0), b, state->compressed_size);
423 	else
424 		cksum = state->compressed_cksum;
425 	if (cksum != state->compressed_cksum) {
426 		archive_set_error(&self->archive->archive,
427 		    ARCHIVE_ERRNO_MISC, "Corrupted data");
428 		return (ARCHIVE_FATAL);
429 	}
430 
431 	/*
432 	 * If the both uncompressed size and compressed size are the same,
433 	 * we do not decompress this block.
434 	 */
435 	if (state->uncompressed_size == state->compressed_size) {
436 		*p = b;
437 		state->total_out += state->compressed_size;
438 		state->unconsumed_bytes = state->compressed_size;
439 		return ((ssize_t)state->uncompressed_size);
440 	}
441 
442 	/*
443 	 * Drive lzo uncompression.
444 	 */
445 	out_size = (lzo_uint)state->uncompressed_size;
446 	r = lzo1x_decompress_safe(b, (lzo_uint)state->compressed_size,
447 		state->out_block, &out_size, NULL);
448 	switch (r) {
449 	case LZO_E_OK:
450 		if (out_size == state->uncompressed_size)
451 			break;
452 		archive_set_error(&self->archive->archive,
453 		    ARCHIVE_ERRNO_MISC, "Corrupted data");
454 		return (ARCHIVE_FATAL);
455 	case LZO_E_OUT_OF_MEMORY:
456 		archive_set_error(&self->archive->archive, ENOMEM,
457 		    "lzop decompression failed: out of memory");
458 		return (ARCHIVE_FATAL);
459 	default:
460 		archive_set_error(&self->archive->archive, ARCHIVE_ERRNO_MISC,
461 		    "lzop decompression failed: %d", r);
462 		return (ARCHIVE_FATAL);
463 	}
464 
465 	if (state->flags & CRC32_UNCOMPRESSED)
466 		cksum = crc32(crc32(0, NULL, 0), state->out_block,
467 		    state->uncompressed_size);
468 	else if (state->flags & ADLER32_UNCOMPRESSED)
469 		cksum = adler32(adler32(0, NULL, 0), state->out_block,
470 		    state->uncompressed_size);
471 	else
472 		cksum = state->uncompressed_cksum;
473 	if (cksum != state->uncompressed_cksum) {
474 		archive_set_error(&self->archive->archive,
475 		    ARCHIVE_ERRNO_MISC, "Corrupted data");
476 		return (ARCHIVE_FATAL);
477 	}
478 
479 	__archive_read_filter_consume(self->upstream, state->compressed_size);
480 	*p = state->out_block;
481 	state->total_out += out_size;
482 	return ((ssize_t)out_size);
483 }
484 
485 /*
486  * Clean up the decompressor.
487  */
488 static int
lzop_filter_close(struct archive_read_filter * self)489 lzop_filter_close(struct archive_read_filter *self)
490 {
491 	struct read_lzop *state = (struct read_lzop *)self->data;
492 
493 	free(state->out_block);
494 	free(state);
495 	return (ARCHIVE_OK);
496 }
497 
498 #endif
499