1 /*-
2  * Copyright (c) 2003-2007 Tim Kientzle
3  * All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
8  * 1. Redistributions of source code must retain the above copyright
9  *    notice, this list of conditions and the following disclaimer.
10  * 2. Redistributions in binary form must reproduce the above copyright
11  *    notice, this list of conditions and the following disclaimer in the
12  *    documentation and/or other materials provided with the distribution.
13  *
14  * THIS SOFTWARE IS PROVIDED BY THE AUTHOR(S) ``AS IS'' AND ANY EXPRESS OR
15  * IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE IMPLIED WARRANTIES
16  * OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE ARE DISCLAIMED.
17  * IN NO EVENT SHALL THE AUTHOR(S) BE LIABLE FOR ANY DIRECT, INDIRECT,
18  * INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL DAMAGES (INCLUDING, BUT
19  * NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS OR SERVICES; LOSS OF USE,
20  * DATA, OR PROFITS; OR BUSINESS INTERRUPTION) HOWEVER CAUSED AND ON ANY
21  * THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT LIABILITY, OR TORT
22  * (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY OUT OF THE USE OF
23  * THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF SUCH DAMAGE.
24  */
25 
26 #include "archive_platform.h"
27 
28 __FBSDID("$FreeBSD: head/lib/libarchive/archive_write_set_compression_gzip.c 201081 2009-12-28 02:04:42Z kientzle $");
29 
30 #ifdef HAVE_ERRNO_H
31 #include <errno.h>
32 #endif
33 #ifdef HAVE_STDLIB_H
34 #include <stdlib.h>
35 #endif
36 #ifdef HAVE_STRING_H
37 #include <string.h>
38 #endif
39 #include <time.h>
40 #ifdef HAVE_ZLIB_H
41 #include <zlib.h>
42 #endif
43 
44 #include "archive.h"
45 #include "archive_private.h"
46 #include "archive_string.h"
47 #include "archive_write_private.h"
48 
49 #if ARCHIVE_VERSION_NUMBER < 4000000
50 int
51 archive_write_set_compression_gzip(struct archive *a)
52 {
53 	__archive_write_filters_free(a);
54 	return (archive_write_add_filter_gzip(a));
55 }
56 #endif
57 
58 /* Don't compile this if we don't have zlib. */
59 
60 struct private_data {
61 	int		 compression_level;
62 	int		 timestamp;
63 #ifdef HAVE_ZLIB_H
64 	z_stream	 stream;
65 	int64_t		 total_in;
66 	unsigned char	*compressed;
67 	size_t		 compressed_buffer_size;
68 	unsigned long	 crc;
69 #else
70 	struct archive_write_program_data *pdata;
71 #endif
72 };
73 
74 /*
75  * Yuck.  zlib.h is not const-correct, so I need this one bit
76  * of ugly hackery to convert a const * pointer to a non-const pointer.
77  */
78 #define	SET_NEXT_IN(st,src)					\
79 	(st)->stream.next_in = (Bytef *)(uintptr_t)(const void *)(src)
80 
81 static int archive_compressor_gzip_options(struct archive_write_filter *,
82 		    const char *, const char *);
83 static int archive_compressor_gzip_open(struct archive_write_filter *);
84 static int archive_compressor_gzip_write(struct archive_write_filter *,
85 		    const void *, size_t);
86 static int archive_compressor_gzip_close(struct archive_write_filter *);
87 static int archive_compressor_gzip_free(struct archive_write_filter *);
88 #ifdef HAVE_ZLIB_H
89 static int drive_compressor(struct archive_write_filter *,
90 		    struct private_data *, int finishing);
91 #endif
92 
93 
94 /*
95  * Add a gzip compression filter to this write handle.
96  */
97 int
98 archive_write_add_filter_gzip(struct archive *_a)
99 {
100 	struct archive_write *a = (struct archive_write *)_a;
101 	struct archive_write_filter *f = __archive_write_allocate_filter(_a);
102 	struct private_data *data;
103 	archive_check_magic(&a->archive, ARCHIVE_WRITE_MAGIC,
104 	    ARCHIVE_STATE_NEW, "archive_write_add_filter_gzip");
105 
106 	data = calloc(1, sizeof(*data));
107 	if (data == NULL) {
108 		archive_set_error(&a->archive, ENOMEM, "Out of memory");
109 		return (ARCHIVE_FATAL);
110 	}
111 	f->data = data;
112 	f->open = &archive_compressor_gzip_open;
113 	f->options = &archive_compressor_gzip_options;
114 	f->close = &archive_compressor_gzip_close;
115 	f->free = &archive_compressor_gzip_free;
116 	f->code = ARCHIVE_FILTER_GZIP;
117 	f->name = "gzip";
118 #ifdef HAVE_ZLIB_H
119 	data->compression_level = Z_DEFAULT_COMPRESSION;
120 	return (ARCHIVE_OK);
121 #else
122 	data->pdata = __archive_write_program_allocate();
123 	if (data->pdata == NULL) {
124 		free(data);
125 		archive_set_error(&a->archive, ENOMEM, "Out of memory");
126 		return (ARCHIVE_FATAL);
127 	}
128 	data->compression_level = 0;
129 	archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
130 	    "Using external gzip program");
131 	return (ARCHIVE_WARN);
132 #endif
133 }
134 
135 static int
136 archive_compressor_gzip_free(struct archive_write_filter *f)
137 {
138 	struct private_data *data = (struct private_data *)f->data;
139 
140 #ifdef HAVE_ZLIB_H
141 	free(data->compressed);
142 #else
143 	__archive_write_program_free(data->pdata);
144 #endif
145 	free(data);
146 	f->data = NULL;
147 	return (ARCHIVE_OK);
148 }
149 
150 /*
151  * Set write options.
152  */
153 static int
154 archive_compressor_gzip_options(struct archive_write_filter *f, const char *key,
155     const char *value)
156 {
157 	struct private_data *data = (struct private_data *)f->data;
158 
159 	if (strcmp(key, "compression-level") == 0) {
160 		if (value == NULL || !(value[0] >= '0' && value[0] <= '9') ||
161 		    value[1] != '\0')
162 			return (ARCHIVE_WARN);
163 		data->compression_level = value[0] - '0';
164 		return (ARCHIVE_OK);
165 	}
166 	if (strcmp(key, "timestamp") == 0) {
167 		data->timestamp = (value == NULL)?-1:1;
168 		return (ARCHIVE_OK);
169 	}
170 
171 	/* Note: The "warn" return is just to inform the options
172 	 * supervisor that we didn't handle it.  It will generate
173 	 * a suitable error if no one used this option. */
174 	return (ARCHIVE_WARN);
175 }
176 
177 #ifdef HAVE_ZLIB_H
178 /*
179  * Setup callback.
180  */
181 static int
182 archive_compressor_gzip_open(struct archive_write_filter *f)
183 {
184 	struct private_data *data = (struct private_data *)f->data;
185 	int ret;
186 
187 	ret = __archive_write_open_filter(f->next_filter);
188 	if (ret != ARCHIVE_OK)
189 		return (ret);
190 
191 	if (data->compressed == NULL) {
192 		size_t bs = 65536, bpb;
193 		if (f->archive->magic == ARCHIVE_WRITE_MAGIC) {
194 			/* Buffer size should be a multiple number of
195 			 * the of bytes per block for performance. */
196 			bpb = archive_write_get_bytes_per_block(f->archive);
197 			if (bpb > bs)
198 				bs = bpb;
199 			else if (bpb != 0)
200 				bs -= bs % bpb;
201 		}
202 		data->compressed_buffer_size = bs;
203 		data->compressed
204 		    = (unsigned char *)malloc(data->compressed_buffer_size);
205 		if (data->compressed == NULL) {
206 			archive_set_error(f->archive, ENOMEM,
207 			    "Can't allocate data for compression buffer");
208 			return (ARCHIVE_FATAL);
209 		}
210 	}
211 
212 	data->crc = crc32(0L, NULL, 0);
213 	data->stream.next_out = data->compressed;
214 	data->stream.avail_out = (uInt)data->compressed_buffer_size;
215 
216 	/* Prime output buffer with a gzip header. */
217 	data->compressed[0] = 0x1f; /* GZip signature bytes */
218 	data->compressed[1] = 0x8b;
219 	data->compressed[2] = 0x08; /* "Deflate" compression */
220 	data->compressed[3] = 0; /* No options */
221 	if (data->timestamp >= 0) {
222 		time_t t = time(NULL);
223 		data->compressed[4] = (uint8_t)(t)&0xff;  /* Timestamp */
224 		data->compressed[5] = (uint8_t)(t>>8)&0xff;
225 		data->compressed[6] = (uint8_t)(t>>16)&0xff;
226 		data->compressed[7] = (uint8_t)(t>>24)&0xff;
227 	} else
228 		memset(&data->compressed[4], 0, 4);
229 	data->compressed[8] = 0; /* No deflate options */
230 	data->compressed[9] = 3; /* OS=Unix */
231 	data->stream.next_out += 10;
232 	data->stream.avail_out -= 10;
233 
234 	f->write = archive_compressor_gzip_write;
235 
236 	/* Initialize compression library. */
237 	ret = deflateInit2(&(data->stream),
238 	    data->compression_level,
239 	    Z_DEFLATED,
240 	    -15 /* < 0 to suppress zlib header */,
241 	    8,
242 	    Z_DEFAULT_STRATEGY);
243 
244 	if (ret == Z_OK) {
245 		f->data = data;
246 		return (ARCHIVE_OK);
247 	}
248 
249 	/* Library setup failed: clean up. */
250 	archive_set_error(f->archive, ARCHIVE_ERRNO_MISC, "Internal error "
251 	    "initializing compression library");
252 
253 	/* Override the error message if we know what really went wrong. */
254 	switch (ret) {
255 	case Z_STREAM_ERROR:
256 		archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
257 		    "Internal error initializing "
258 		    "compression library: invalid setup parameter");
259 		break;
260 	case Z_MEM_ERROR:
261 		archive_set_error(f->archive, ENOMEM,
262 		    "Internal error initializing compression library");
263 		break;
264 	case Z_VERSION_ERROR:
265 		archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
266 		    "Internal error initializing "
267 		    "compression library: invalid library version");
268 		break;
269 	}
270 
271 	return (ARCHIVE_FATAL);
272 }
273 
274 /*
275  * Write data to the compressed stream.
276  */
277 static int
278 archive_compressor_gzip_write(struct archive_write_filter *f, const void *buff,
279     size_t length)
280 {
281 	struct private_data *data = (struct private_data *)f->data;
282 	int ret;
283 
284 	/* Update statistics */
285 	data->crc = crc32(data->crc, (const Bytef *)buff, (uInt)length);
286 	data->total_in += length;
287 
288 	/* Compress input data to output buffer */
289 	SET_NEXT_IN(data, buff);
290 	data->stream.avail_in = (uInt)length;
291 	if ((ret = drive_compressor(f, data, 0)) != ARCHIVE_OK)
292 		return (ret);
293 
294 	return (ARCHIVE_OK);
295 }
296 
297 /*
298  * Finish the compression...
299  */
300 static int
301 archive_compressor_gzip_close(struct archive_write_filter *f)
302 {
303 	unsigned char trailer[8];
304 	struct private_data *data = (struct private_data *)f->data;
305 	int ret, r1;
306 
307 	/* Finish compression cycle */
308 	ret = drive_compressor(f, data, 1);
309 	if (ret == ARCHIVE_OK) {
310 		/* Write the last compressed data. */
311 		ret = __archive_write_filter(f->next_filter,
312 		    data->compressed,
313 		    data->compressed_buffer_size - data->stream.avail_out);
314 	}
315 	if (ret == ARCHIVE_OK) {
316 		/* Build and write out 8-byte trailer. */
317 		trailer[0] = (uint8_t)(data->crc)&0xff;
318 		trailer[1] = (uint8_t)(data->crc >> 8)&0xff;
319 		trailer[2] = (uint8_t)(data->crc >> 16)&0xff;
320 		trailer[3] = (uint8_t)(data->crc >> 24)&0xff;
321 		trailer[4] = (uint8_t)(data->total_in)&0xff;
322 		trailer[5] = (uint8_t)(data->total_in >> 8)&0xff;
323 		trailer[6] = (uint8_t)(data->total_in >> 16)&0xff;
324 		trailer[7] = (uint8_t)(data->total_in >> 24)&0xff;
325 		ret = __archive_write_filter(f->next_filter, trailer, 8);
326 	}
327 
328 	switch (deflateEnd(&(data->stream))) {
329 	case Z_OK:
330 		break;
331 	default:
332 		archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
333 		    "Failed to clean up compressor");
334 		ret = ARCHIVE_FATAL;
335 	}
336 	r1 = __archive_write_close_filter(f->next_filter);
337 	return (r1 < ret ? r1 : ret);
338 }
339 
340 /*
341  * Utility function to push input data through compressor,
342  * writing full output blocks as necessary.
343  *
344  * Note that this handles both the regular write case (finishing ==
345  * false) and the end-of-archive case (finishing == true).
346  */
347 static int
348 drive_compressor(struct archive_write_filter *f,
349     struct private_data *data, int finishing)
350 {
351 	int ret;
352 
353 	for (;;) {
354 		if (data->stream.avail_out == 0) {
355 			ret = __archive_write_filter(f->next_filter,
356 			    data->compressed,
357 			    data->compressed_buffer_size);
358 			if (ret != ARCHIVE_OK)
359 				return (ARCHIVE_FATAL);
360 			data->stream.next_out = data->compressed;
361 			data->stream.avail_out =
362 			    (uInt)data->compressed_buffer_size;
363 		}
364 
365 		/* If there's nothing to do, we're done. */
366 		if (!finishing && data->stream.avail_in == 0)
367 			return (ARCHIVE_OK);
368 
369 		ret = deflate(&(data->stream),
370 		    finishing ? Z_FINISH : Z_NO_FLUSH );
371 
372 		switch (ret) {
373 		case Z_OK:
374 			/* In non-finishing case, check if compressor
375 			 * consumed everything */
376 			if (!finishing && data->stream.avail_in == 0)
377 				return (ARCHIVE_OK);
378 			/* In finishing case, this return always means
379 			 * there's more work */
380 			break;
381 		case Z_STREAM_END:
382 			/* This return can only occur in finishing case. */
383 			return (ARCHIVE_OK);
384 		default:
385 			/* Any other return value indicates an error. */
386 			archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
387 			    "GZip compression failed:"
388 			    " deflate() call returned status %d",
389 			    ret);
390 			return (ARCHIVE_FATAL);
391 		}
392 	}
393 }
394 
395 #else /* HAVE_ZLIB_H */
396 
397 static int
398 archive_compressor_gzip_open(struct archive_write_filter *f)
399 {
400 	struct private_data *data = (struct private_data *)f->data;
401 	struct archive_string as;
402 	int r;
403 
404 	archive_string_init(&as);
405 	archive_strcpy(&as, "gzip");
406 
407 	/* Specify compression level. */
408 	if (data->compression_level > 0) {
409 		archive_strcat(&as, " -");
410 		archive_strappend_char(&as, '0' + data->compression_level);
411 	}
412 	if (data->timestamp < 0)
413 		/* Do not save timestamp. */
414 		archive_strcat(&as, " -n");
415 	else if (data->timestamp > 0)
416 		/* Save timestamp. */
417 		archive_strcat(&as, " -N");
418 
419 	f->write = archive_compressor_gzip_write;
420 	r = __archive_write_program_open(f, data->pdata, as.s);
421 	archive_string_free(&as);
422 	return (r);
423 }
424 
425 static int
426 archive_compressor_gzip_write(struct archive_write_filter *f, const void *buff,
427     size_t length)
428 {
429 	struct private_data *data = (struct private_data *)f->data;
430 
431 	return __archive_write_program_write(f, data->pdata, buff, length);
432 }
433 
434 static int
435 archive_compressor_gzip_close(struct archive_write_filter *f)
436 {
437 	struct private_data *data = (struct private_data *)f->data;
438 
439 	return __archive_write_program_close(f, data->pdata);
440 }
441 
442 #endif /* HAVE_ZLIB_H */
443