op public repos

Blob

Date:: Tue Mar 24 15:15:51 2020 UTC
Message:: make got-index-pack compute and verify the pack file's SHA1 checksum
Actions:: History | Blame | Raw File
1 /*
2  * Copyright (c) 2018 Stefan Sperling <stsp@openbsd.org>
3  *
4  * Permission to use, copy, modify, and distribute this software for any
5  * purpose with or without fee is hereby granted, provided that the above
6  * copyright notice and this permission notice appear in all copies.
7  *
8  * THE SOFTWARE IS PROVIDED "AS IS" AND THE AUTHOR DISCLAIMS ALL WARRANTIES
9  * WITH REGARD TO THIS SOFTWARE INCLUDING ALL IMPLIED WARRANTIES OF
10  * MERCHANTABILITY AND FITNESS. IN NO EVENT SHALL THE AUTHOR BE LIABLE FOR
11  * ANY SPECIAL, DIRECT, INDIRECT, OR CONSEQUENTIAL DAMAGES OR ANY DAMAGES
12  * WHATSOEVER RESULTING FROM LOSS OF USE, DATA OR PROFITS, WHETHER IN AN
13  * ACTION OF CONTRACT, NEGLIGENCE OR OTHER TORTIOUS ACTION, ARISING OUT OF
14  * OR IN CONNECTION WITH THE USE OR PERFORMANCE OF THIS SOFTWARE.
15  */
16 
17 #include <sys/queue.h>
18 
19 #include <errno.h>
20 #include <stdio.h>
21 #include <stdlib.h>
22 #include <string.h>
23 #include <sha1.h>
24 #include <zlib.h>
25 #include <time.h>
26 
27 #include "got_error.h"
28 #include "got_object.h"
29 #include "got_path.h"
30 
31 #include "got_lib_inflate.h"
32 
33 #ifndef MIN
34 #define	MIN(_a,_b) ((_a) < (_b) ? (_a) : (_b))
35 #endif
36 
37 const struct got_error *
38 got_inflate_init(struct got_inflate_buf *zb, uint8_t *outbuf, size_t bufsize,
39     struct got_inflate_checksum *csum)
40 {
41 	const struct got_error *err = NULL;
42 	int zerr;
43 
44 	memset(&zb->z, 0, sizeof(zb->z));
45 
46 	zb->z.zalloc = Z_NULL;
47 	zb->z.zfree = Z_NULL;
48 	zerr = inflateInit(&zb->z);
49 	if (zerr != Z_OK) {
50 		if  (zerr == Z_ERRNO)
51 			return got_error_from_errno("inflateInit");
52 		if  (zerr == Z_MEM_ERROR) {
53 			errno = ENOMEM;
54 			return got_error_from_errno("inflateInit");
55 		}
56 		return got_error(GOT_ERR_DECOMPRESSION);
57 	}
58 
59 	zb->inlen = zb->outlen = bufsize;
60 
61 	zb->inbuf = calloc(1, zb->inlen);
62 	if (zb->inbuf == NULL) {
63 		err = got_error_from_errno("calloc");
64 		goto done;
65 	}
66 
67 	zb->flags = 0;
68 	if (outbuf == NULL) {
69 		zb->outbuf = calloc(1, zb->outlen);
70 		if (zb->outbuf == NULL) {
71 			err = got_error_from_errno("calloc");
72 			goto done;
73 		}
74 		zb->flags |= GOT_INFLATE_F_OWN_OUTBUF;
75 	} else
76 		zb->outbuf = outbuf;
77 
78 	zb->csum = csum;
79 done:
80 	if (err)
81 		got_inflate_end(zb);
82 	return err;
83 }
84 
85 static void
86 csum_input(struct got_inflate_checksum *csum, const char *buf, size_t len)
87 {
88 	if (csum->input_crc)
89 		*csum->input_crc = crc32(*csum->input_crc, buf, len);
90 
91 	if (csum->input_sha1)
92 		SHA1Update(csum->input_sha1, buf, len);
93 }
94 
95 const struct got_error *
96 got_inflate_read(struct got_inflate_buf *zb, FILE *f, size_t *outlenp,
97    size_t *consumed)
98 {
99 	size_t last_total_out = zb->z.total_out;
100 	size_t last_total_in = zb->z.total_in;
101 	z_stream *z = &zb->z;
102 	int ret = Z_ERRNO;
103 
104 	z->next_out = zb->outbuf;
105 	z->avail_out = zb->outlen;
106 
107 	*outlenp = 0;
108 	if (consumed)
109 		*consumed = 0;
110 	do {
111 		char *csum_in = NULL;
112 		size_t csum_avail = 0;
113 
114 		if (z->avail_in == 0) {
115 			size_t n = fread(zb->inbuf, 1, zb->inlen, f);
116 			if (n == 0) {
117 				if (ferror(f))
118 					return got_ferror(f, GOT_ERR_IO);
119 				/* EOF */
120 				ret = Z_STREAM_END;
121 				break;
122 			}
123 			z->next_in = zb->inbuf;
124 			z->avail_in = n;
125 		}
126 		if (zb->csum) {
127 			csum_in = z->next_in;
128 			csum_avail = z->avail_in;
129 		}
130 		ret = inflate(z, Z_SYNC_FLUSH);
131 		if (zb->csum)
132 			csum_input(zb->csum, csum_in, csum_avail - z->avail_in);
133 	} while (ret == Z_OK && z->avail_out > 0);
134 
135 	if (ret == Z_OK || ret == Z_BUF_ERROR) {
136 		zb->flags |= GOT_INFLATE_F_HAVE_MORE;
137 	} else {
138 		if (ret != Z_STREAM_END)
139 			return got_error(GOT_ERR_DECOMPRESSION);
140 		zb->flags &= ~GOT_INFLATE_F_HAVE_MORE;
141 	}
142 
143 	*outlenp = z->total_out - last_total_out;
144 	if (consumed)
145 		*consumed += z->total_in - last_total_in;
146 	return NULL;
147 }
148 
149 const struct got_error *
150 got_inflate_read_fd(struct got_inflate_buf *zb, int fd, size_t *outlenp,
151     size_t *consumed)
152 {
153 	size_t last_total_out = zb->z.total_out;
154 	size_t last_total_in = zb->z.total_in;
155 	z_stream *z = &zb->z;
156 	int ret = Z_ERRNO;
157 
158 	z->next_out = zb->outbuf;
159 	z->avail_out = zb->outlen;
160 
161 	*outlenp = 0;
162 	if (consumed)
163 		*consumed = 0;
164 	do {
165 		char *csum_in = NULL;
166 		size_t csum_avail = 0;
167 
168 		if (z->avail_in == 0) {
169 			ssize_t n = read(fd, zb->inbuf, zb->inlen);
170 			if (n < 0)
171 				return got_error_from_errno("read");
172 			else if (n == 0) {
173 				/* EOF */
174 				ret = Z_STREAM_END;
175 				break;
176 			}
177 			z->next_in = zb->inbuf;
178 			z->avail_in = n;
179 		}
180 		if (zb->csum) {
181 			csum_in = z->next_in;
182 			csum_avail = z->avail_in;
183 		}
184 		ret = inflate(z, Z_SYNC_FLUSH);
185 		if (zb->csum)
186 			csum_input(zb->csum, csum_in, csum_avail - z->avail_in);
187 	} while (ret == Z_OK && z->avail_out > 0);
188 
189 	if (ret == Z_OK || ret == Z_BUF_ERROR) {
190 		zb->flags |= GOT_INFLATE_F_HAVE_MORE;
191 	} else {
192 		if (ret != Z_STREAM_END)
193 			return got_error(GOT_ERR_DECOMPRESSION);
194 		zb->flags &= ~GOT_INFLATE_F_HAVE_MORE;
195 	}
196 
197 	*outlenp = z->total_out - last_total_out;
198 	if (consumed)
199 		*consumed += z->total_in - last_total_in;
200 	return NULL;
201 }
202 
203 const struct got_error *
204 got_inflate_read_mmap(struct got_inflate_buf *zb, uint8_t *map, size_t offset,
205     size_t len, size_t *outlenp, size_t *consumed)
206 {
207 	size_t last_total_out = zb->z.total_out;
208 	z_stream *z = &zb->z;
209 	int ret = Z_ERRNO;
210 
211 	z->next_out = zb->outbuf;
212 	z->avail_out = zb->outlen;
213 
214 	*outlenp = 0;
215 	*consumed = 0;
216 
217 	do {
218 		char *csum_in = NULL;
219 		size_t csum_avail = 0;
220 		size_t last_total_in = zb->z.total_in;
221 
222 		if (z->avail_in == 0) {
223 			if (len == 0) {
224 				/* EOF */
225 				ret = Z_STREAM_END;
226 				break;
227 			}
228 			z->next_in = map + offset + *consumed;
229 			z->avail_in = len - *consumed;
230 		}
231 		if (zb->csum) {
232 			csum_in = z->next_in;
233 			csum_avail = z->avail_in;
234 		}
235 		ret = inflate(z, Z_SYNC_FLUSH);
236 		if (zb->csum)
237 			csum_input(zb->csum, csum_in, csum_avail - z->avail_in);
238 		*consumed += z->total_in - last_total_in;
239 	} while (ret == Z_OK && z->avail_out > 0);
240 
241 	if (ret == Z_OK || ret == Z_BUF_ERROR) {
242 		zb->flags |= GOT_INFLATE_F_HAVE_MORE;
243 	} else {
244 		if (ret != Z_STREAM_END)
245 			return got_error(GOT_ERR_DECOMPRESSION);
246 		zb->flags &= ~GOT_INFLATE_F_HAVE_MORE;
247 	}
248 
249 	*outlenp = z->total_out - last_total_out;
250 	return NULL;
251 }
252 
253 void
254 got_inflate_end(struct got_inflate_buf *zb)
255 {
256 	free(zb->inbuf);
257 	if (zb->flags & GOT_INFLATE_F_OWN_OUTBUF)
258 		free(zb->outbuf);
259 	inflateEnd(&zb->z);
260 }
261 
262 const struct got_error *
263 got_inflate_to_mem(uint8_t **outbuf, size_t *outlen,
264     size_t *consumed_total, FILE *f)
265 {
266 	const struct got_error *err;
267 	size_t avail, consumed;
268 	struct got_inflate_buf zb;
269 	void *newbuf;
270 	int nbuf = 1;
271 
272 	if (outbuf) {
273 		*outbuf = malloc(GOT_INFLATE_BUFSIZE);
274 		if (*outbuf == NULL)
275 			return got_error_from_errno("malloc");
276 		err = got_inflate_init(&zb, *outbuf, GOT_INFLATE_BUFSIZE, NULL);
277 	} else
278 		err = got_inflate_init(&zb, NULL, GOT_INFLATE_BUFSIZE, NULL);
279 	if (err)
280 		return err;
281 
282 	*outlen = 0;
283 	if (consumed_total)
284 		*consumed_total = 0;
285 
286 	do {
287 		err = got_inflate_read(&zb, f, &avail, &consumed);
288 		if (err)
289 			goto done;
290 		*outlen += avail;
291 		if (consumed_total)
292 			*consumed_total += consumed;
293 		if (zb.flags & GOT_INFLATE_F_HAVE_MORE) {
294 			if (outbuf == NULL)
295 				continue;
296 			zb.outlen = (nbuf * GOT_INFLATE_BUFSIZE) - *outlen;
297 			newbuf = reallocarray(*outbuf, ++nbuf,
298 			   GOT_INFLATE_BUFSIZE);
299 			if (newbuf == NULL) {
300 				err = got_error_from_errno("reallocarray");
301 				free(*outbuf);
302 				*outbuf = NULL;
303 				*outlen = 0;
304 				goto done;
305 			}
306 			*outbuf = newbuf;
307 			zb.outbuf = newbuf + *outlen;
308 		}
309 	} while (zb.flags & GOT_INFLATE_F_HAVE_MORE);
310 
311 done:
312 	got_inflate_end(&zb);
313 	return err;
314 }
315 
316 const struct got_error *
317 got_inflate_to_mem_fd(uint8_t **outbuf, size_t *outlen,
318     size_t *consumed_total, struct got_inflate_checksum *csum, 
319     size_t expected_size, int infd)
320 {
321 	const struct got_error *err;
322 	size_t avail, consumed;
323 	struct got_inflate_buf zb;
324 	void *newbuf;
325 	int nbuf = 1;
326 	size_t bufsize = GOT_INFLATE_BUFSIZE;
327 
328 	/* Optimize buffer size in case short reads should suffice. */
329 	if (expected_size > 0 && expected_size < bufsize)
330 		bufsize = expected_size;
331 		
332 	if (outbuf) {
333 		*outbuf = malloc(bufsize);
334 		if (*outbuf == NULL)
335 			return got_error_from_errno("malloc");
336 		err = got_inflate_init(&zb, *outbuf, GOT_INFLATE_BUFSIZE, csum);
337 	} else
338 		err = got_inflate_init(&zb, NULL, bufsize, csum);
339 	if (err)
340 		goto done;
341 
342 	*outlen = 0;
343 	if (consumed_total)
344 		*consumed_total = 0;
345 
346 	do {
347 		err = got_inflate_read_fd(&zb, infd, &avail, &consumed);
348 		if (err)
349 			goto done;
350 		*outlen += avail;
351 		if (consumed_total)
352 			*consumed_total += consumed;
353 		if (zb.flags & GOT_INFLATE_F_HAVE_MORE) {
354 			if (outbuf == NULL)
355 				continue;
356 			zb.outlen = (nbuf * GOT_INFLATE_BUFSIZE) - *outlen;
357 			newbuf = reallocarray(*outbuf, ++nbuf,
358 			    GOT_INFLATE_BUFSIZE);
359 			if (newbuf == NULL) {
360 				err = got_error_from_errno("reallocarray");
361 				free(*outbuf);
362 				*outbuf = NULL;
363 				*outlen = 0;
364 				goto done;
365 			}
366 			*outbuf = newbuf;
367 			zb.outbuf = newbuf + *outlen;
368 		}
369 	} while (zb.flags & GOT_INFLATE_F_HAVE_MORE);
370 
371 done:
372 	got_inflate_end(&zb);
373 	return err;
374 }
375 
376 const struct got_error *
377 got_inflate_to_mem_mmap(uint8_t **outbuf, size_t *outlen,
378     size_t *consumed_total, struct got_inflate_checksum *csum, uint8_t *map,
379     size_t offset, size_t len)
380 {
381 	const struct got_error *err;
382 	size_t avail, consumed;
383 	struct got_inflate_buf zb;
384 	void *newbuf;
385 	int nbuf = 1;
386 
387 	if (outbuf) {
388 		*outbuf = malloc(GOT_INFLATE_BUFSIZE);
389 		if (*outbuf == NULL)
390 			return got_error_from_errno("malloc");
391 		err = got_inflate_init(&zb, *outbuf, GOT_INFLATE_BUFSIZE, csum);
392 		if (err) {
393 			free(*outbuf);
394 			*outbuf = NULL;
395 			return err;
396 		}
397 	} else {
398 		err = got_inflate_init(&zb, NULL, GOT_INFLATE_BUFSIZE, csum);
399 	}
400 
401 	*outlen = 0;
402 	if (consumed_total)
403 		*consumed_total = 0;
404 	do {
405 		err = got_inflate_read_mmap(&zb, map, offset, len, &avail,
406 		    &consumed);
407 		if (err)
408 			goto done;
409 		offset += consumed;
410 		if (consumed_total)
411 			*consumed_total += consumed;
412 		len -= consumed;
413 		*outlen += avail;
414 		if (len == 0)
415 			break;
416 		if (zb.flags & GOT_INFLATE_F_HAVE_MORE) {
417 			if (outbuf == NULL)
418 				continue;
419 			newbuf = reallocarray(*outbuf, ++nbuf,
420 			    GOT_INFLATE_BUFSIZE);
421 			if (newbuf == NULL) {
422 				err = got_error_from_errno("reallocarray");
423 				free(*outbuf);
424 				*outbuf = NULL;
425 				*outlen = 0;
426 				goto done;
427 			}
428 			*outbuf = newbuf;
429 			zb.outbuf = newbuf + *outlen;
430 			zb.outlen = (nbuf * GOT_INFLATE_BUFSIZE) - *outlen;
431 		}
432 	} while (zb.flags & GOT_INFLATE_F_HAVE_MORE);
433 done:
434 	got_inflate_end(&zb);
435 	return err;
436 }
437 
438 const struct got_error *
439 got_inflate_to_fd(size_t *outlen, FILE *infile, int outfd)
440 {
441 	const struct got_error *err = NULL;
442 	size_t avail;
443 	struct got_inflate_buf zb;
444 
445 	err = got_inflate_init(&zb, NULL, GOT_INFLATE_BUFSIZE, NULL);
446 	if (err)
447 		goto done;
448 
449 	*outlen = 0;
450 
451 	do {
452 		err = got_inflate_read(&zb, infile, &avail, NULL);
453 		if (err)
454 			goto done;
455 		if (avail > 0) {
456 			ssize_t n;
457 			n = write(outfd, zb.outbuf, avail);
458 			if (n != avail) {
459 				err = got_error_from_errno("write");
460 				goto done;
461 			}
462 			*outlen += avail;
463 		}
464 	} while (zb.flags & GOT_INFLATE_F_HAVE_MORE);
465 
466 done:
467 	if (err == NULL) {
468 		if (lseek(outfd, SEEK_SET, 0) == -1)
469 			err = got_error_from_errno("lseek");
470 	}
471 	got_inflate_end(&zb);
472 	return err;
473 }
474 
475 const struct got_error *
476 got_inflate_to_file(size_t *outlen, FILE *infile, FILE *outfile)
477 {
478 	const struct got_error *err;
479 	size_t avail;
480 	struct got_inflate_buf zb;
481 
482 	err = got_inflate_init(&zb, NULL, GOT_INFLATE_BUFSIZE, NULL);
483 	if (err)
484 		goto done;
485 
486 	*outlen = 0;
487 
488 	do {
489 		err = got_inflate_read(&zb, infile, &avail, NULL);
490 		if (err)
491 			goto done;
492 		if (avail > 0) {
493 			size_t n;
494 			n = fwrite(zb.outbuf, avail, 1, outfile);
495 			if (n != 1) {
496 				err = got_ferror(outfile, GOT_ERR_IO);
497 				goto done;
498 			}
499 			*outlen += avail;
500 		}
501 	} while (zb.flags & GOT_INFLATE_F_HAVE_MORE);
502 
503 done:
504 	if (err == NULL)
505 		rewind(outfile);
506 	got_inflate_end(&zb);
507 	return err;
508 }
509 
510 const struct got_error *
511 got_inflate_to_file_fd(size_t *outlen, size_t *consumed_total,
512     struct got_inflate_checksum *csum, int infd, FILE *outfile)
513 {
514 	const struct got_error *err;
515 	size_t avail, consumed;
516 	struct got_inflate_buf zb;
517 
518 	err = got_inflate_init(&zb, NULL, GOT_INFLATE_BUFSIZE, csum);
519 	if (err)
520 		goto done;
521 
522 	*outlen = 0;
523 	if (consumed_total)
524 		*consumed_total = 0;
525 	do {
526 		err = got_inflate_read_fd(&zb, infd, &avail, &consumed);
527 		if (err)
528 			goto done;
529 		if (avail > 0) {
530 			size_t n;
531 			n = fwrite(zb.outbuf, avail, 1, outfile);
532 			if (n != 1) {
533 				err = got_ferror(outfile, GOT_ERR_IO);
534 				goto done;
535 			}
536 			*outlen += avail;
537 			if (consumed_total)
538 				*consumed_total += consumed;
539 		}
540 	} while (zb.flags & GOT_INFLATE_F_HAVE_MORE);
541 
542 done:
543 	if (err == NULL)
544 		rewind(outfile);
545 	got_inflate_end(&zb);
546 	return err;
547 }
548 
549 const struct got_error *
550 got_inflate_to_file_mmap(size_t *outlen, size_t *consumed_total,
551     struct got_inflate_checksum *csum, uint8_t *map, size_t offset,
552     size_t len, FILE *outfile)
553 {
554 	const struct got_error *err;
555 	size_t avail, consumed;
556 	struct got_inflate_buf zb;
557 
558 	err = got_inflate_init(&zb, NULL, GOT_INFLATE_BUFSIZE, csum);
559 	if (err)
560 		goto done;
561 
562 	*outlen = 0;
563 	if (consumed_total)
564 		*consumed_total = 0;
565 	do {
566 		err = got_inflate_read_mmap(&zb, map, offset, len, &avail,
567 		    &consumed);
568 		if (err)
569 			goto done;
570 		offset += consumed;
571 		if (consumed_total)
572 			*consumed_total += consumed;
573 		len -= consumed;
574 		if (avail > 0) {
575 			size_t n;
576 			n = fwrite(zb.outbuf, avail, 1, outfile);
577 			if (n != 1) {
578 				err = got_ferror(outfile, GOT_ERR_IO);
579 				goto done;
580 			}
581 			*outlen += avail;
582 		}
583 	} while (zb.flags & GOT_INFLATE_F_HAVE_MORE);
584 
585 done:
586 	if (err == NULL)
587 		rewind(outfile);
588 	got_inflate_end(&zb);
589 	return err;
590 }