1 /*-
2  * Copyright (c) 2017 Sean Purcell
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("$FreeBSD$");
29 
30 
31 #ifdef HAVE_ERRNO_H
32 #include <errno.h>
33 #endif
34 #ifdef HAVE_STDLIB_H
35 #include <stdlib.h>
36 #endif
37 #ifdef HAVE_STRING_H
38 #include <string.h>
39 #endif
40 #ifdef HAVE_ZSTD_H
41 #include <zstd.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 /* Don't compile this if we don't have zstd.h */
50 
51 struct private_data {
52 	int		 compression_level;
53 	int      threads;
54 #if HAVE_ZSTD_H && HAVE_LIBZSTD_COMPRESSOR
55 	ZSTD_CStream	*cstream;
56 	int64_t		 total_in;
57 	ZSTD_outBuffer	 out;
58 #else
59 	struct archive_write_program_data *pdata;
60 #endif
61 };
62 
63 /* If we don't have the library use default range values (zstdcli.c v1.4.0) */
64 #define CLEVEL_MIN -99
65 #define CLEVEL_STD_MIN 0 /* prior to 1.3.4 and more recent without using --fast */
66 #define CLEVEL_DEFAULT 3
67 #define CLEVEL_STD_MAX 19 /* without using --ultra */
68 #define CLEVEL_MAX 22
69 
70 #define MINVER_NEGCLEVEL 10304
71 #define MINVER_MINCLEVEL 10306
72 
73 static int archive_compressor_zstd_options(struct archive_write_filter *,
74 		    const char *, const char *);
75 static int archive_compressor_zstd_open(struct archive_write_filter *);
76 static int archive_compressor_zstd_write(struct archive_write_filter *,
77 		    const void *, size_t);
78 static int archive_compressor_zstd_close(struct archive_write_filter *);
79 static int archive_compressor_zstd_free(struct archive_write_filter *);
80 #if HAVE_ZSTD_H && HAVE_LIBZSTD_COMPRESSOR
81 static int drive_compressor(struct archive_write_filter *,
82 		    struct private_data *, int, const void *, size_t);
83 #endif
84 
85 
86 /*
87  * Add a zstd compression filter to this write handle.
88  */
89 int
90 archive_write_add_filter_zstd(struct archive *_a)
91 {
92 	struct archive_write *a = (struct archive_write *)_a;
93 	struct archive_write_filter *f = __archive_write_allocate_filter(_a);
94 	struct private_data *data;
95 	archive_check_magic(&a->archive, ARCHIVE_WRITE_MAGIC,
96 	    ARCHIVE_STATE_NEW, "archive_write_add_filter_zstd");
97 
98 	data = calloc(1, sizeof(*data));
99 	if (data == NULL) {
100 		archive_set_error(&a->archive, ENOMEM, "Out of memory");
101 		return (ARCHIVE_FATAL);
102 	}
103 	f->data = data;
104 	f->open = &archive_compressor_zstd_open;
105 	f->options = &archive_compressor_zstd_options;
106 	f->close = &archive_compressor_zstd_close;
107 	f->free = &archive_compressor_zstd_free;
108 	f->code = ARCHIVE_FILTER_ZSTD;
109 	f->name = "zstd";
110 	data->compression_level = CLEVEL_DEFAULT;
111 	data->threads = 0;
112 #if HAVE_ZSTD_H && HAVE_LIBZSTD_COMPRESSOR
113 	data->cstream = ZSTD_createCStream();
114 	if (data->cstream == NULL) {
115 		free(data);
116 		archive_set_error(&a->archive, ENOMEM,
117 		    "Failed to allocate zstd compressor object");
118 		return (ARCHIVE_FATAL);
119 	}
120 
121 	return (ARCHIVE_OK);
122 #else
123 	data->pdata = __archive_write_program_allocate("zstd");
124 	if (data->pdata == NULL) {
125 		free(data);
126 		archive_set_error(&a->archive, ENOMEM, "Out of memory");
127 		return (ARCHIVE_FATAL);
128 	}
129 	archive_set_error(&a->archive, ARCHIVE_ERRNO_MISC,
130 	    "Using external zstd program");
131 	return (ARCHIVE_WARN);
132 #endif
133 }
134 
135 static int
136 archive_compressor_zstd_free(struct archive_write_filter *f)
137 {
138 	struct private_data *data = (struct private_data *)f->data;
139 #if HAVE_ZSTD_H && HAVE_LIBZSTD_COMPRESSOR
140 	ZSTD_freeCStream(data->cstream);
141 	free(data->out.dst);
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 static int string_is_numeric (const char* value)
151 {
152        size_t len = strlen(value);
153        size_t i;
154 
155        if (len == 0) {
156                return (ARCHIVE_WARN);
157        }
158        else if (len == 1 && !(value[0] >= '0' && value[0] <= '9')) {
159                return (ARCHIVE_WARN);
160        }
161        else if (!(value[0] >= '0' && value[0] <= '9') &&
162                 value[0] != '-' && value[0] != '+') {
163                return (ARCHIVE_WARN);
164        }
165 
166        for (i = 1; i < len; i++) {
167                if (!(value[i] >= '0' && value[i] <= '9')) {
168                        return (ARCHIVE_WARN);
169                }
170        }
171 
172        return (ARCHIVE_OK);
173 }
174 
175 /*
176  * Set write options.
177  */
178 static int
179 archive_compressor_zstd_options(struct archive_write_filter *f, const char *key,
180     const char *value)
181 {
182 	struct private_data *data = (struct private_data *)f->data;
183 
184 	if (strcmp(key, "compression-level") == 0) {
185 		int level = atoi(value);
186 		/* If we don't have the library, hard-code the max level */
187 		int minimum = CLEVEL_MIN;
188 		int maximum = CLEVEL_MAX;
189 		if (string_is_numeric(value) != ARCHIVE_OK) {
190 			return (ARCHIVE_WARN);
191 		}
192 #if HAVE_ZSTD_H && HAVE_LIBZSTD_COMPRESSOR
193 		maximum = ZSTD_maxCLevel();
194 #if ZSTD_VERSION_NUMBER >= MINVER_MINCLEVEL
195 		if (ZSTD_versionNumber() >= MINVER_MINCLEVEL) {
196 			minimum = ZSTD_minCLevel();
197 		}
198 		else
199 #endif
200 		if (ZSTD_versionNumber() < MINVER_NEGCLEVEL) {
201 			minimum = CLEVEL_STD_MIN;
202 		}
203 #endif
204 		if (level < minimum || level > maximum) {
205 			return (ARCHIVE_WARN);
206 		}
207 		data->compression_level = level;
208 		return (ARCHIVE_OK);
209 	} else if (strcmp(key, "threads") == 0) {
210 		int threads = atoi(value);
211 		if (string_is_numeric(value) != ARCHIVE_OK) {
212 			return (ARCHIVE_WARN);
213 		}
214 
215 		int minimum = 0;
216 
217 		if (threads < minimum) {
218 			return (ARCHIVE_WARN);
219 		}
220 
221 		data->threads = threads;
222 		return (ARCHIVE_OK);
223 	}
224 
225 	/* Note: The "warn" return is just to inform the options
226 	 * supervisor that we didn't handle it.  It will generate
227 	 * a suitable error if no one used this option. */
228 	return (ARCHIVE_WARN);
229 }
230 
231 #if HAVE_ZSTD_H && HAVE_LIBZSTD_COMPRESSOR
232 /*
233  * Setup callback.
234  */
235 static int
236 archive_compressor_zstd_open(struct archive_write_filter *f)
237 {
238 	struct private_data *data = (struct private_data *)f->data;
239 
240 	if (data->out.dst == NULL) {
241 		size_t bs = ZSTD_CStreamOutSize(), bpb;
242 		if (f->archive->magic == ARCHIVE_WRITE_MAGIC) {
243 			/* Buffer size should be a multiple number of
244 			 * the of bytes per block for performance. */
245 			bpb = archive_write_get_bytes_per_block(f->archive);
246 			if (bpb > bs)
247 				bs = bpb;
248 			else if (bpb != 0)
249 				bs -= bs % bpb;
250 		}
251 		data->out.size = bs;
252 		data->out.pos = 0;
253 		data->out.dst
254 		    = (unsigned char *)malloc(data->out.size);
255 		if (data->out.dst == NULL) {
256 			archive_set_error(f->archive, ENOMEM,
257 			    "Can't allocate data for compression buffer");
258 			return (ARCHIVE_FATAL);
259 		}
260 	}
261 
262 	f->write = archive_compressor_zstd_write;
263 
264 	if (ZSTD_isError(ZSTD_initCStream(data->cstream,
265 	    data->compression_level))) {
266 		archive_set_error(f->archive, ARCHIVE_ERRNO_MISC,
267 		    "Internal error initializing zstd compressor object");
268 		return (ARCHIVE_FATAL);
269 	}
270 
271 	ZSTD_CCtx_setParameter(data->cstream, ZSTD_c_nbWorkers, data->threads);
272 
273 	return (ARCHIVE_OK);
274 }
275 
276 /*
277  * Write data to the compressed stream.
278  */
279 static int
280 archive_compressor_zstd_write(struct archive_write_filter *f, const void *buff,
281     size_t length)
282 {
283 	struct private_data *data = (struct private_data *)f->data;
284 	int ret;
285 
286 	/* Update statistics */
287 	data->total_in += length;
288 
289 	if ((ret = drive_compressor(f, data, 0, buff, length)) != ARCHIVE_OK)
290 		return (ret);
291 
292 	return (ARCHIVE_OK);
293 }
294 
295 /*
296  * Finish the compression...
297  */
298 static int
299 archive_compressor_zstd_close(struct archive_write_filter *f)
300 {
301 	struct private_data *data = (struct private_data *)f->data;
302 
303 	/* Finish zstd frame */
304 	return drive_compressor(f, data, 1, NULL, 0);
305 }
306 
307 /*
308  * Utility function to push input data through compressor,
309  * writing full output blocks as necessary.
310  *
311  * Note that this handles both the regular write case (finishing ==
312  * false) and the end-of-archive case (finishing == true).
313  */
314 static int
315 drive_compressor(struct archive_write_filter *f,
316     struct private_data *data, int finishing, const void *src, size_t length)
317 {
318 	ZSTD_inBuffer in = (ZSTD_inBuffer) { src, length, 0 };
319 
320 	for (;;) {
321 		if (data->out.pos == data->out.size) {
322 			const int ret = __archive_write_filter(f->next_filter,
323 			    data->out.dst, data->out.size);
324 			if (ret != ARCHIVE_OK)
325 				return (ARCHIVE_FATAL);
326 			data->out.pos = 0;
327 		}
328 
329 		/* If there's nothing to do, we're done. */
330 		if (!finishing && in.pos == in.size)
331 			return (ARCHIVE_OK);
332 
333 		{
334 			const size_t zstdret = !finishing ?
335 			    ZSTD_compressStream(data->cstream, &data->out, &in)
336 			    : ZSTD_endStream(data->cstream, &data->out);
337 
338 			if (ZSTD_isError(zstdret)) {
339 				archive_set_error(f->archive,
340 				    ARCHIVE_ERRNO_MISC,
341 				    "Zstd compression failed: %s",
342 				    ZSTD_getErrorName(zstdret));
343 				return (ARCHIVE_FATAL);
344 			}
345 
346 			/* If we're finishing, 0 means nothing left to flush */
347 			if (finishing && zstdret == 0) {
348 				const int ret = __archive_write_filter(f->next_filter,
349 				    data->out.dst, data->out.pos);
350 				return (ret);
351 			}
352 		}
353 	}
354 }
355 
356 #else /* HAVE_ZSTD_H && HAVE_LIBZSTD_COMPRESSOR */
357 
358 static int
359 archive_compressor_zstd_open(struct archive_write_filter *f)
360 {
361 	struct private_data *data = (struct private_data *)f->data;
362 	struct archive_string as;
363 	int r;
364 
365 	archive_string_init(&as);
366 	/* --no-check matches library default */
367 	archive_strcpy(&as, "zstd --no-check");
368 
369 	if (data->compression_level < CLEVEL_STD_MIN) {
370 		struct archive_string as2;
371 		archive_string_init(&as2);
372 		archive_string_sprintf(&as2, " --fast=%d", -data->compression_level);
373 		archive_string_concat(&as, &as2);
374 		archive_string_free(&as2);
375 	} else {
376 		struct archive_string as2;
377 		archive_string_init(&as2);
378 		archive_string_sprintf(&as2, " -%d", data->compression_level);
379 		archive_string_concat(&as, &as2);
380 		archive_string_free(&as2);
381 	}
382 
383 	if (data->compression_level > CLEVEL_STD_MAX) {
384 		archive_strcat(&as, " --ultra");
385 	}
386 
387 	if (data->threads != 0) {
388 		struct archive_string as2;
389 		archive_string_init(&as2);
390 		archive_string_sprintf(&as2, " --threads=%d", data->threads);
391 		archive_string_concat(&as, &as2);
392 		archive_string_free(&as2);
393 	}
394 
395 	f->write = archive_compressor_zstd_write;
396 	r = __archive_write_program_open(f, data->pdata, as.s);
397 	archive_string_free(&as);
398 	return (r);
399 }
400 
401 static int
402 archive_compressor_zstd_write(struct archive_write_filter *f, const void *buff,
403     size_t length)
404 {
405 	struct private_data *data = (struct private_data *)f->data;
406 
407 	return __archive_write_program_write(f, data->pdata, buff, length);
408 }
409 
410 static int
411 archive_compressor_zstd_close(struct archive_write_filter *f)
412 {
413 	struct private_data *data = (struct private_data *)f->data;
414 
415 	return __archive_write_program_close(f, data->pdata);
416 }
417 
418 #endif /* HAVE_ZSTD_H && HAVE_LIBZSTD_COMPRESSOR */
419