Gotweb

Blob

Date:: Fri Jan 12 22:00:06 2018 UTC
Message:: introduce got_ferror() and use it
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/types.h>
18 #include <sys/stat.h>
19 #include <sys/queue.h>
20 
21 #include <dirent.h>
22 #include <errno.h>
23 #include <stdio.h>
24 #include <stdint.h>
25 #include <stdlib.h>
26 #include <string.h>
27 #include <limits.h>
28 #include <sha1.h>
29 #include <endian.h>
30 #include <zlib.h>
31 
32 #include "got_error.h"
33 #include "got_object.h"
34 #include "got_repository.h"
35 #include "got_sha1.h"
36 #include "pack.h"
37 #include "path.h"
38 
39 #define GOT_PACK_PREFIX		"pack-"
40 #define GOT_PACKFILE_SUFFIX	".pack"
41 #define GOT_PACKIDX_SUFFIX		".idx"
42 #define GOT_PACKFILE_NAMELEN	(strlen(GOT_PACK_PREFIX) + \
43 				SHA1_DIGEST_STRING_LENGTH - 1 + \
44 				strlen(GOT_PACKFILE_SUFFIX))
45 #define GOT_PACKIDX_NAMELEN	(strlen(GOT_PACK_PREFIX) + \
46 				SHA1_DIGEST_STRING_LENGTH - 1 + \
47 				strlen(GOT_PACKIDX_SUFFIX))
48 
49 #ifndef MIN
50 #define	MIN(_a,_b) ((_a) < (_b) ? (_a) : (_b))
51 #endif
52 
53 static const struct got_error *
54 verify_fanout_table(uint32_t *fanout_table)
55 {
56 	int i;
57 
58 	for (i = 0; i < 0xff - 1; i++) {
59 		if (be32toh(fanout_table[i]) > be32toh(fanout_table[i + 1]))
60 			return got_error(GOT_ERR_BAD_PACKIDX);
61 	}
62 
63 	return NULL;
64 }
65 
66 static const struct got_error *
67 get_packfile_size(size_t *size, const char *path_idx)
68 {
69 	struct stat sb;
70 	char *path_pack;
71 	char base_path[PATH_MAX];
72 	char *dot;
73 
74 	if (strlcpy(base_path, path_idx, PATH_MAX) > PATH_MAX)
75 		return got_error(GOT_ERR_NO_SPACE);
76 
77 	dot = strrchr(base_path, '.');
78 	if (dot == NULL)
79 		return got_error(GOT_ERR_BAD_PATH);
80 	*dot = '\0';
81 	if (asprintf(&path_pack, "%s.pack", base_path) == -1)
82 		return got_error(GOT_ERR_NO_MEM);
83 
84 	if (stat(path_pack, &sb) != 0) {
85 		free(path_pack);
86 		return got_error_from_errno();
87 	}
88 
89 	free(path_pack);
90 	*size = sb.st_size;
91 	return 0;
92 }
93 
94 const struct got_error *
95 got_packidx_open(struct got_packidx_v2_hdr **packidx, const char *path)
96 {
97 	struct got_packidx_v2_hdr *p;
98 	FILE *f;
99 	const struct got_error *err = NULL;
100 	size_t n, nobj, packfile_size;
101 	SHA1_CTX ctx;
102 	uint8_t sha1[SHA1_DIGEST_LENGTH];
103 
104 	SHA1Init(&ctx);
105 
106 	f = fopen(path, "rb");
107 	if (f == NULL)
108 		return got_error(GOT_ERR_BAD_PATH);
109 
110 	err = get_packfile_size(&packfile_size, path);
111 	if (err)
112 		return err;
113 
114 	p = calloc(1, sizeof(*p));
115 	if (p == NULL) {
116 		err = got_error(GOT_ERR_NO_MEM);
117 		goto done;
118 	}
119 
120 	n = fread(&p->magic, sizeof(p->magic), 1, f);
121 	if (n != 1) {
122 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
123 		goto done;
124 	}
125 
126 	if (betoh32(p->magic) != GOT_PACKIDX_V2_MAGIC) {
127 		err = got_error(GOT_ERR_BAD_PACKIDX);
128 		goto done;
129 	}
130 
131 	SHA1Update(&ctx, (uint8_t *)&p->magic, sizeof(p->magic));
132 
133 	n = fread(&p->version, sizeof(p->version), 1, f);
134 	if (n != 1) {
135 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
136 		goto done;
137 	}
138 
139 	if (betoh32(p->version) != GOT_PACKIDX_VERSION) {
140 		err = got_error(GOT_ERR_BAD_PACKIDX);
141 		goto done;
142 	}
143 
144 	SHA1Update(&ctx, (uint8_t *)&p->version, sizeof(p->version));
145 
146 	n = fread(&p->fanout_table, sizeof(p->fanout_table), 1, f);
147 	if (n != 1) {
148 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
149 		goto done;
150 	}
151 
152 	err = verify_fanout_table(p->fanout_table);
153 	if (err)
154 		goto done;
155 
156 	SHA1Update(&ctx, (uint8_t *)p->fanout_table, sizeof(p->fanout_table));
157 
158 	nobj = betoh32(p->fanout_table[0xff]);
159 
160 	p->sorted_ids = calloc(nobj, sizeof(*p->sorted_ids));
161 	if (p->sorted_ids == NULL) {
162 		err = got_error(GOT_ERR_NO_MEM);
163 		goto done;
164 	}
165 
166 	n = fread(p->sorted_ids, sizeof(*p->sorted_ids), nobj, f);
167 	if (n != nobj) {
168 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
169 		goto done;
170 	}
171 
172 	SHA1Update(&ctx, (uint8_t *)p->sorted_ids,
173 	    nobj * sizeof(*p->sorted_ids));
174 
175 	p->crc32 = calloc(nobj, sizeof(*p->crc32));
176 	if (p->crc32 == NULL) {
177 		err = got_error(GOT_ERR_NO_MEM);
178 		goto done;
179 	}
180 
181 	n = fread(p->crc32, sizeof(*p->crc32), nobj, f);
182 	if (n != nobj) {
183 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
184 		goto done;
185 	}
186 
187 	SHA1Update(&ctx, (uint8_t *)p->crc32, nobj * sizeof(*p->crc32));
188 
189 	p->offsets = calloc(nobj, sizeof(*p->offsets));
190 	if (p->offsets == NULL) {
191 		err = got_error(GOT_ERR_NO_MEM);
192 		goto done;
193 	}
194 
195 	n = fread(p->offsets, sizeof(*p->offsets), nobj, f);
196 	if (n != nobj) {
197 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
198 		goto done;
199 	}
200 
201 	SHA1Update(&ctx, (uint8_t *)p->offsets, nobj * sizeof(*p->offsets));
202 
203 	/* Large file offsets are contained only in files > 2GB. */
204 	if (packfile_size <= 0x80000000)
205 		goto checksum;
206 
207 	p->large_offsets = calloc(nobj, sizeof(*p->large_offsets));
208 	if (p->large_offsets == NULL) {
209 		err = got_error(GOT_ERR_NO_MEM);
210 		goto done;
211 	}
212 
213 	n = fread(p->large_offsets, sizeof(*p->large_offsets), nobj, f);
214 	if (n != nobj) {
215 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
216 		goto done;
217 	}
218 
219 	SHA1Update(&ctx, (uint8_t*)p->large_offsets,
220 	    nobj * sizeof(*p->large_offsets));
221 
222 checksum:
223 	n = fread(&p->trailer, sizeof(p->trailer), 1, f);
224 	if (n != 1) {
225 		err = got_ferror(f, GOT_ERR_BAD_PACKIDX);
226 		goto done;
227 	}
228 
229 	SHA1Update(&ctx, p->trailer.packfile_sha1, SHA1_DIGEST_LENGTH);
230 	SHA1Final(sha1, &ctx);
231 	if (memcmp(p->trailer.packidx_sha1, sha1, SHA1_DIGEST_LENGTH) != 0)
232 		err = got_error(GOT_ERR_PACKIDX_CSUM);
233 done:
234 	fclose(f);
235 	if (err)
236 		got_packidx_close(p);
237 	else
238 		*packidx = p;
239 	return err;
240 }
241 
242 void
243 got_packidx_close(struct got_packidx_v2_hdr *packidx)
244 {
245 	free(packidx->sorted_ids);
246 	free(packidx->offsets);
247 	free(packidx->crc32);
248 	free(packidx->large_offsets);
249 	free(packidx);
250 }
251 
252 static int
253 is_packidx_filename(const char *name, size_t len)
254 {
255 	if (len != GOT_PACKIDX_NAMELEN)
256 		return 0;
257 
258 	if (strncmp(name, GOT_PACK_PREFIX, strlen(GOT_PACK_PREFIX)) != 0)
259 		return 0;
260 
261 	if (strcmp(name + strlen(GOT_PACK_PREFIX) +
262 	    SHA1_DIGEST_STRING_LENGTH - 1, GOT_PACKIDX_SUFFIX) != 0)
263 		return 0;
264 
265 	return 1;
266 }
267 
268 static off_t
269 get_object_offset(struct got_packidx_v2_hdr *packidx, int idx)
270 {
271 	uint32_t totobj = betoh32(packidx->fanout_table[0xff]);
272 	uint32_t offset = betoh32(packidx->offsets[idx]);
273 	if (offset & GOT_PACKIDX_OFFSET_VAL_IS_LARGE_IDX) {
274 		uint64_t loffset;
275 		idx = offset & GOT_PACKIDX_OFFSET_VAL_MASK;
276 		if (idx < 0 || idx > totobj || packidx->large_offsets == NULL)
277 			return -1;
278 		loffset = betoh64(packidx->large_offsets[idx]);
279 		return (loffset > INT64_MAX ? -1 : (off_t)loffset);
280 	}
281 	return (off_t)(offset & GOT_PACKIDX_OFFSET_VAL_MASK);
282 }
283 
284 static int
285 get_object_idx(struct got_packidx_v2_hdr *packidx, struct got_object_id *id)
286 {
287 	u_int8_t id0 = id->sha1[0];
288 	uint32_t totobj = betoh32(packidx->fanout_table[0xff]);
289 	int i = 0;
290 
291 	if (id0 > 0)
292 		i = betoh32(packidx->fanout_table[id0 - 1]);
293 
294 	while (i < totobj) {
295 		struct got_object_id *oid = &packidx->sorted_ids[i++];
296 		uint32_t offset;
297 
298 		if (got_object_id_cmp(id, oid) < 0)
299 			continue;
300 		if (got_object_id_cmp(id, oid) > 0)
301 			break;
302 
303 		return i;
304 	}
305 
306 	return -1;
307 }
308 
309 const struct got_error *
310 read_packfile_hdr(FILE *f, struct got_packidx_v2_hdr *packidx)
311 {
312 	const struct got_error *err = NULL;
313 	uint32_t totobj = betoh32(packidx->fanout_table[0xff]);
314 	struct got_packfile_hdr hdr;
315 	size_t n;
316 
317 	n = fread(&hdr, sizeof(hdr), 1, f);
318 	if (n != 1)
319 		return got_ferror(f, GOT_ERR_BAD_PACKIDX);
320 
321 	if (betoh32(hdr.signature) != GOT_PACKFILE_SIGNATURE ||
322 	    betoh32(hdr.version) != GOT_PACKFILE_VERSION ||
323 	    betoh32(hdr.nobjects) != totobj)
324 		err = got_error(GOT_ERR_BAD_PACKFILE);
325 
326 	return err;
327 }
328 
329 static const struct got_error *
330 dump_packed_object(FILE **f, FILE *packfile, off_t offset)
331 {
332 	const struct got_error *err = NULL;
333 	const char *template = "/tmp/got.XXXXXXXXXX";
334 	uint64_t size = 0;
335 	uint8_t type = 0;
336 	uint8_t sizeN;
337 	int i;
338 	size_t n;
339 	const char *type_tag;
340 
341 	*f = got_opentemp();
342 	if (*f == NULL) {
343 		err = got_error(GOT_ERR_FILE_OPEN);
344 		goto done;
345 	}
346 
347 	if (fseeko(packfile, offset, SEEK_SET) != 0) {
348 		err = got_error_from_errno();
349 		goto done;
350 	}
351 
352 	i = 0;
353 	do {
354 		/* We do not support size values which don't fit in 64 bit. */
355 		if (i > 9) {
356 			err = got_error(GOT_ERR_NO_SPACE);
357 			goto done;
358 		}
359 
360 		n = fread(&sizeN, sizeof(sizeN), 1, packfile);
361 		if (n != 1) {
362 			err = got_ferror(packfile, GOT_ERR_BAD_PACKIDX);
363 			goto done;
364 		}
365 
366 		if (i == 0) {
367 			type = (sizeN & GOT_PACK_OBJ_SIZE0_TYPE_MASK) >>
368 			    GOT_PACK_OBJ_SIZE0_TYPE_MASK_SHIFT;
369 			size = (sizeN & GOT_PACK_OBJ_SIZE0_VAL_MASK);
370 		} else {
371 			size_t shift = 4 + 7 * (i - 1);
372 			size |= ((sizeN & GOT_PACK_OBJ_SIZE_VAL_MASK) << shift);
373 		}
374 		i++;
375 	} while (sizeN & GOT_PACK_OBJ_SIZE_MORE);
376 
377 	if (type == GOT_OBJ_TYPE_OFFSET_DELTA)
378 		printf("object type OFFSET_DELTA not yet implemented\n");
379 	else if (type == GOT_OBJ_TYPE_REF_DELTA)
380 		printf("object type REF_DELTA not yet implemented\n");
381 	else if (type == GOT_OBJ_TYPE_TAG)
382 		printf("object type TAG not yet implemented\n");
383 
384 	type_tag = got_object_get_type_tag(type);
385 	if (type_tag == NULL) {
386 		err = got_error(GOT_ERR_BAD_OBJ_HDR);
387 		goto done;
388 	}
389 
390 	fprintf(*f, "%s %llu", type_tag, size);
391 	fputc('\0', *f);
392 
393 	while (size > 0) {
394 		uint8_t data[2048];
395 		size_t len = MIN(size, sizeof(data));
396 
397 		n = fread(data, len, 1, packfile);
398 		if (n != 1) {
399 			err = got_ferror(packfile, GOT_ERR_BAD_PACKIDX);
400 			goto done;
401 		}
402 
403 		n = fwrite(data, len, 1, *f);
404 		if (n != 1) {
405 			err = got_ferror(*f, GOT_ERR_BAD_PACKIDX);
406 			goto done;
407 		}
408 
409 		size -= len;
410 	}
411 
412 	printf("object type is %d\n", type);
413 	rewind(*f);
414 done:
415 	if (err && *f)
416 		fclose(*f);
417 	return err;
418 }
419 static const struct got_error *
420 extract_object(FILE **f, const char *path_packdir,
421     struct got_packidx_v2_hdr *packidx, struct got_object_id *id)
422 {
423 	const struct got_error *err = NULL;
424 	int idx = get_object_idx(packidx, id);
425 	off_t offset;
426 	char *path_packfile;
427 	FILE *packfile;
428 	char hex[SHA1_DIGEST_STRING_LENGTH];
429 	char *sha1str;
430 
431 	*f = NULL;
432 	if (idx == -1) /* object not found in pack index */
433 		return NULL;
434 
435 	offset = get_object_offset(packidx, idx);
436 	if (offset == (uint64_t)-1)
437 		return got_error(GOT_ERR_BAD_PACKIDX);
438 
439 	sha1str = got_sha1_digest_to_str(packidx->trailer.packfile_sha1,
440 	    hex, sizeof(hex));
441 	if (sha1str == NULL)
442 		return got_error(GOT_ERR_PACKIDX_CSUM);
443 
444 	if (asprintf(&path_packfile, "%s/%s%s%s", path_packdir,
445 	    GOT_PACK_PREFIX, sha1str, GOT_PACKFILE_SUFFIX) == -1)
446 		return got_error(GOT_ERR_NO_MEM);
447 
448 	packfile = fopen(path_packfile, "rb");
449 	if (packfile == NULL) {
450 		err = got_error_from_errno();
451 		goto done;
452 	}
453 
454 	err = read_packfile_hdr(packfile, packidx);
455 	if (err)
456 		goto done;
457 
458 	printf("Dumping object at offset %llu\n", offset);
459 	err = dump_packed_object(f, packfile, offset);
460 	if (err)
461 		goto done;
462 
463 done:
464 	free(path_packfile);
465 	if (packfile && fclose(packfile) == -1 && err == 0)
466 		err = got_error_from_errno();
467 	return err;
468 }
469 
470 const struct got_error *
471 got_packfile_extract_object(FILE **f, struct got_object_id *id,
472     struct got_repository *repo)
473 {
474 	const struct got_error *err = NULL;
475 	DIR *packdir = NULL;
476 	struct dirent *dent;
477 	char *path_packdir = got_repo_get_path_objects_pack(repo);
478 
479 	if (path_packdir == NULL) {
480 		err = got_error(GOT_ERR_NO_MEM);
481 		goto done;
482 	}
483 
484 	packdir = opendir(path_packdir);
485 	if (packdir == NULL) {
486 		err = got_error_from_errno();
487 		goto done;
488 	}
489 
490 	while ((dent = readdir(packdir)) != NULL) {
491 		struct got_packidx_v2_hdr *packidx;
492 		char *path_packidx, *path_object;
493 
494 		if (!is_packidx_filename(dent->d_name, dent->d_namlen))
495 			continue;
496 
497 		if (asprintf(&path_packidx, "%s/%s", path_packdir,
498 		    dent->d_name) == -1) {
499 			err = got_error(GOT_ERR_NO_MEM);
500 			goto done;
501 		}
502 
503 		err = got_packidx_open(&packidx, path_packidx);
504 		free(path_packidx);
505 		if (err)
506 			goto done;
507 
508 		err = extract_object(f, path_packdir, packidx, id);
509 		if (err)
510 			goto done;
511 		if (*f != NULL)
512 			break;
513 	}
514 
515 done:
516 	free(path_packdir);
517 	if (packdir && closedir(packdir) != 0 && err == 0)
518 		err = got_error_from_errno();
519 	return err;
520 }