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 #include <sys/uio.h>
21 #include <sys/socket.h>
22 #include <sys/syslimits.h>
23 #include <sys/wait.h>
24 
25 #include <errno.h>
26 #include <stdio.h>
27 #include <stdlib.h>
28 #include <string.h>
29 #include <stdint.h>
30 #include <sha1.h>
31 #include <zlib.h>
32 #include <ctype.h>
33 #include <limits.h>
34 #include <imsg.h>
35 #include <time.h>
36 #include <unistd.h>
37 
38 #include "got_error.h"
39 #include "got_object.h"
40 #include "got_repository.h"
41 #include "got_opentemp.h"
42 
43 #include "got_lib_sha1.h"
44 #include "got_lib_delta.h"
45 #include "got_lib_privsep.h"
46 #include "got_lib_pack.h"
47 #include "got_lib_inflate.h"
48 #include "got_lib_object.h"
49 #include "got_lib_object_cache.h"
50 #include "got_lib_repository.h"
51 
52 #ifndef nitems
53 #define nitems(_a) (sizeof(_a) / sizeof((_a)[0]))
54 #endif
55 
56 #define GOT_OBJ_TAG_COMMIT	"commit"
57 #define GOT_OBJ_TAG_TREE	"tree"
58 #define GOT_OBJ_TAG_BLOB	"blob"
59 
60 #define GOT_COMMIT_TAG_TREE		"tree "
61 #define GOT_COMMIT_TAG_PARENT		"parent "
62 #define GOT_COMMIT_TAG_AUTHOR		"author "
63 #define GOT_COMMIT_TAG_COMMITTER	"committer "
64 
65 int
66 got_object_id_cmp(const struct got_object_id *id1,
67     const struct got_object_id *id2)
68 {
69 	return memcmp(id1->sha1, id2->sha1, SHA1_DIGEST_LENGTH);
70 }
71 
72 const struct got_error *
73 got_object_qid_alloc_partial(struct got_object_qid **qid)
74 {
75 	const struct got_error *err = NULL;
76 
77 	*qid = malloc(sizeof(**qid));
78 	if (*qid == NULL)
79 		return got_error_from_errno();
80 
81 	(*qid)->id = malloc(sizeof(*((*qid)->id)));
82 	if ((*qid)->id == NULL) {
83 		err = got_error_from_errno();
84 		got_object_qid_free(*qid);
85 		*qid = NULL;
86 		return err;
87 	}
88 
89 	return NULL;
90 }
91 
92 const struct got_error *
93 got_object_id_str(char **outbuf, struct got_object_id *id)
94 {
95 	static const size_t len = SHA1_DIGEST_STRING_LENGTH;
96 
97 	*outbuf = malloc(len);
98 	if (*outbuf == NULL)
99 		return got_error_from_errno();
100 
101 	if (got_sha1_digest_to_str(id->sha1, *outbuf, len) == NULL) {
102 		free(*outbuf);
103 		*outbuf = NULL;
104 		return got_error(GOT_ERR_BAD_OBJ_ID_STR);
105 	}
106 
107 	return NULL;
108 }
109 
110 void
111 got_object_close(struct got_object *obj)
112 {
113 	if (obj->refcnt > 0) {
114 		obj->refcnt--;
115 		if (obj->refcnt > 0)
116 			return;
117 	}
118 
119 	if (obj->flags & GOT_OBJ_FLAG_DELTIFIED) {
120 		struct got_delta *delta;
121 		while (!SIMPLEQ_EMPTY(&obj->deltas.entries)) {
122 			delta = SIMPLEQ_FIRST(&obj->deltas.entries);
123 			SIMPLEQ_REMOVE_HEAD(&obj->deltas.entries, entry);
124 			got_delta_close(delta);
125 		}
126 	}
127 	if (obj->flags & GOT_OBJ_FLAG_PACKED)
128 		free(obj->path_packfile);
129 	free(obj);
130 }
131 
132 void
133 got_object_qid_free(struct got_object_qid *qid)
134 {
135 	free(qid->id);
136 	free(qid);
137 }
138 
139 struct got_commit_object *
140 got_object_commit_alloc_partial(void)
141 {
142 	struct got_commit_object *commit;
143 
144 	commit = calloc(1, sizeof(*commit));
145 	if (commit == NULL)
146 		return NULL;
147 	commit->tree_id = malloc(sizeof(*commit->tree_id));
148 	if (commit->tree_id == NULL) {
149 		free(commit);
150 		return NULL;
151 	}
152 
153 	SIMPLEQ_INIT(&commit->parent_ids);
154 
155 	return commit;
156 }
157 
158 const struct got_error *
159 got_object_commit_add_parent(struct got_commit_object *commit,
160     const char *id_str)
161 {
162 	const struct got_error *err = NULL;
163 	struct got_object_qid *qid;
164 
165 	err = got_object_qid_alloc_partial(&qid);
166 	if (err)
167 		return err;
168 
169 	if (!got_parse_sha1_digest(qid->id->sha1, id_str)) {
170 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
171 		free(qid->id);
172 		free(qid);
173 		return err;
174 	}
175 
176 	SIMPLEQ_INSERT_TAIL(&commit->parent_ids, qid, entry);
177 	commit->nparents++;
178 
179 	return NULL;
180 }
181 
182 static const struct got_error *
183 parse_gmtoff(time_t *gmtoff, const char *tzstr)
184 {
185 	int sign = 1;
186 	const char *p = tzstr;
187 	time_t h, m;
188 
189 	*gmtoff = 0;
190 
191 	if (*p == '-')
192 		sign = -1;
193 	else if (*p != '+')
194 		return got_error(GOT_ERR_BAD_OBJ_DATA);
195 	p++;
196 	if (!isdigit(*p) && !isdigit(*(p + 1)))
197 		return got_error(GOT_ERR_BAD_OBJ_DATA);
198 	h = (((*p - '0') * 10) + (*(p + 1) - '0'));
199 
200 	p += 2;
201 	if (!isdigit(*p) && !isdigit(*(p + 1)))
202 		return got_error(GOT_ERR_BAD_OBJ_DATA);
203 	m = ((*p - '0') * 10) + (*(p + 1) - '0');
204 
205 	*gmtoff = (h * 60 * 60 + m * 60) * sign;
206 	return NULL;
207 }
208 
209 static const struct got_error *
210 parse_commit_time(time_t *time, time_t *gmtoff, char *committer)
211 {
212 	const struct got_error *err = NULL;
213 	const char *errstr;
214 	char *space, *tzstr;
215 
216 	/* Parse and strip off trailing timezone indicator string. */
217 	space = strrchr(committer, ' ');
218 	if (space == NULL)
219 		return got_error(GOT_ERR_BAD_OBJ_DATA);
220 	tzstr = strdup(space + 1);
221 	if (tzstr == NULL)
222 		return got_error_from_errno();
223 	err = parse_gmtoff(gmtoff, tzstr);
224 	free(tzstr);
225 	if (err)
226 		return err;
227 	*space = '\0';
228 
229 	/* Timestamp is separated from committer name + email by space. */
230 	space = strrchr(committer, ' ');
231 	if (space == NULL)
232 		return got_error(GOT_ERR_BAD_OBJ_DATA);
233 
234 	/* Timestamp parsed here is expressed in comitter's local time. */
235 	*time = strtonum(space + 1, 0, INT64_MAX, &errstr);
236 	if (errstr)
237 		return got_error(GOT_ERR_BAD_OBJ_DATA);
238 
239 	/* Express the time stamp in UTC. */
240 	*time -= *gmtoff;
241 
242 	/* Strip off parsed time information, leaving just author and email. */
243 	*space = '\0';
244 
245 	return NULL;
246 }
247 
248 void
249 got_object_commit_close(struct got_commit_object *commit)
250 {
251 	struct got_object_qid *qid;
252 
253 	if (commit->refcnt > 0) {
254 		commit->refcnt--;
255 		if (commit->refcnt > 0)
256 			return;
257 	}
258 
259 	while (!SIMPLEQ_EMPTY(&commit->parent_ids)) {
260 		qid = SIMPLEQ_FIRST(&commit->parent_ids);
261 		SIMPLEQ_REMOVE_HEAD(&commit->parent_ids, entry);
262 		got_object_qid_free(qid);
263 	}
264 
265 	free(commit->tree_id);
266 	free(commit->author);
267 	free(commit->committer);
268 	free(commit->logmsg);
269 	free(commit);
270 }
271 
272 const struct got_error *
273 got_object_parse_commit(struct got_commit_object **commit, char *buf, size_t len)
274 {
275 	const struct got_error *err = NULL;
276 	char *s = buf;
277 	size_t tlen;
278 	ssize_t remain = (ssize_t)len;
279  
280 	*commit = got_object_commit_alloc_partial();
281 	if (*commit == NULL)
282 		return got_error_from_errno();
283 
284 	tlen = strlen(GOT_COMMIT_TAG_TREE);
285 	if (strncmp(s, GOT_COMMIT_TAG_TREE, tlen) == 0) {
286 		remain -= tlen;
287 		if (remain < SHA1_DIGEST_STRING_LENGTH) {
288 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
289 			goto done;
290 		}
291 		s += tlen;
292 		if (!got_parse_sha1_digest((*commit)->tree_id->sha1, s)) {
293 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
294 			goto done;
295 		}
296 		remain -= SHA1_DIGEST_STRING_LENGTH;
297 		s += SHA1_DIGEST_STRING_LENGTH;
298 	} else {
299 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
300 		goto done;
301 	}
302 
303 	tlen = strlen(GOT_COMMIT_TAG_PARENT);
304 	while (strncmp(s, GOT_COMMIT_TAG_PARENT, tlen) == 0) {
305 		remain -= tlen;
306 		if (remain < SHA1_DIGEST_STRING_LENGTH) {
307 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
308 			goto done;
309 		}
310 		s += tlen;
311 		err = got_object_commit_add_parent(*commit, s);
312 		if (err)
313 			goto done;
314 
315 		remain -= SHA1_DIGEST_STRING_LENGTH;
316 		s += SHA1_DIGEST_STRING_LENGTH;
317 	}
318 
319 	tlen = strlen(GOT_COMMIT_TAG_AUTHOR);
320 	if (strncmp(s, GOT_COMMIT_TAG_AUTHOR, tlen) == 0) {
321 		char *p;
322 		size_t slen;
323 
324 		remain -= tlen;
325 		if (remain <= 0) {
326 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
327 			goto done;
328 		}
329 		s += tlen;
330 		p = strchr(s, '\n');
331 		if (p == NULL) {
332 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
333 			goto done;
334 		}
335 		*p = '\0';
336 		slen = strlen(s);
337 		err = parse_commit_time(&(*commit)->author_time,
338 		    &(*commit)->author_gmtoff, s);
339 		if (err)
340 			goto done;
341 		(*commit)->author = strdup(s);
342 		if ((*commit)->author == NULL) {
343 			err = got_error_from_errno();
344 			goto done;
345 		}
346 		s += slen + 1;
347 		remain -= slen + 1;
348 	}
349 
350 	tlen = strlen(GOT_COMMIT_TAG_COMMITTER);
351 	if (strncmp(s, GOT_COMMIT_TAG_COMMITTER, tlen) == 0) {
352 		char *p;
353 		size_t slen;
354 
355 		remain -= tlen;
356 		if (remain <= 0) {
357 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
358 			goto done;
359 		}
360 		s += tlen;
361 		p = strchr(s, '\n');
362 		if (p == NULL) {
363 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
364 			goto done;
365 		}
366 		*p = '\0';
367 		slen = strlen(s);
368 		err = parse_commit_time(&(*commit)->committer_time,
369 		    &(*commit)->committer_gmtoff, s);
370 		if (err)
371 			goto done;
372 		(*commit)->committer = strdup(s);
373 		if ((*commit)->committer == NULL) {
374 			err = got_error_from_errno();
375 			goto done;
376 		}
377 		s += slen + 1;
378 		remain -= slen + 1;
379 	}
380 
381 	(*commit)->logmsg = strndup(s, remain);
382 	if ((*commit)->logmsg == NULL) {
383 		err = got_error_from_errno();
384 		goto done;
385 	}
386 done:
387 	if (err) {
388 		got_object_commit_close(*commit);
389 		*commit = NULL;
390 	}
391 	return err;
392 }
393 
394 void
395 got_object_tree_entry_close(struct got_tree_entry *te)
396 {
397 	free(te->id);
398 	free(te->name);
399 	free(te);
400 }
401 
402 void
403 got_object_tree_close(struct got_tree_object *tree)
404 {
405 	struct got_tree_entry *te;
406 
407 	if (tree->refcnt > 0) {
408 		tree->refcnt--;
409 		if (tree->refcnt > 0)
410 			return;
411 	}
412 
413 	while (!SIMPLEQ_EMPTY(&tree->entries.head)) {
414 		te = SIMPLEQ_FIRST(&tree->entries.head);
415 		SIMPLEQ_REMOVE_HEAD(&tree->entries.head, entry);
416 		got_object_tree_entry_close(te);
417 	}
418 
419 	free(tree);
420 }
421 
422 struct got_tree_entry *
423 got_alloc_tree_entry_partial(void)
424 {
425 	struct got_tree_entry *te;
426 
427 	te = malloc(sizeof(*te));
428 	if (te == NULL)
429 		return NULL;
430 
431 	te->id = malloc(sizeof(*te->id));
432 	if (te->id == NULL) {
433 		free(te);
434 		te = NULL;
435 	}
436 	return te;
437 }
438 
439 static const struct got_error *
440 parse_tree_entry(struct got_tree_entry **te, size_t *elen, char *buf,
441     size_t maxlen)
442 {
443 	char *p = buf, *space;
444 	const struct got_error *err = NULL;
445 
446 	*te = got_alloc_tree_entry_partial();
447 	if (*te == NULL)
448 		return got_error_from_errno();
449 
450 	*elen = strlen(buf) + 1;
451 	if (*elen > maxlen) {
452 		free(*te);
453 		*te = NULL;
454 		return got_error(GOT_ERR_BAD_OBJ_DATA);
455 	}
456 
457 	space = strchr(buf, ' ');
458 	if (space == NULL) {
459 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
460 		free(*te);
461 		*te = NULL;
462 		return err;
463 	}
464 	(*te)->mode = 0;
465 	while (*p != ' ') {
466 		if (*p < '0' && *p > '7') {
467 			err = got_error(GOT_ERR_BAD_OBJ_DATA);
468 			goto done;
469 		}
470 		(*te)->mode <<= 3;
471 		(*te)->mode |= *p - '0';
472 		p++;
473 	}
474 
475 	(*te)->name = strdup(space + 1);
476 	if (*elen > maxlen || maxlen - *elen < SHA1_DIGEST_LENGTH) {
477 		err = got_error(GOT_ERR_BAD_OBJ_DATA);
478 		goto done;
479 	}
480 	buf += strlen(buf) + 1;
481 	memcpy((*te)->id->sha1, buf, SHA1_DIGEST_LENGTH);
482 	*elen += SHA1_DIGEST_LENGTH;
483 done:
484 	if (err) {
485 		got_object_tree_entry_close(*te);
486 		*te = NULL;
487 	}
488 	return err;
489 }
490 
491 const struct got_error *
492 got_object_parse_tree(struct got_tree_object **tree, uint8_t *buf, size_t len)
493 {
494 	const struct got_error *err;
495 	size_t remain = len;
496 
497 	*tree = calloc(1, sizeof(**tree));
498 	if (*tree == NULL)
499 		return got_error_from_errno();
500 
501 	SIMPLEQ_INIT(&(*tree)->entries.head);
502 
503 	while (remain > 0) {
504 		struct got_tree_entry *te;
505 		size_t elen;
506 
507 		err = parse_tree_entry(&te, &elen, buf, remain);
508 		if (err)
509 			return err;
510 		(*tree)->entries.nentries++;
511 		SIMPLEQ_INSERT_TAIL(&(*tree)->entries.head, te, entry);
512 		buf += elen;
513 		remain -= elen;
514 	}
515 
516 	if (remain != 0) {
517 		got_object_tree_close(*tree);
518 		return got_error(GOT_ERR_BAD_OBJ_DATA);
519 	}
520 
521 	return NULL;
522 }
523 
524 const struct got_error *
525 got_read_file_to_mem(uint8_t **outbuf, size_t *outlen, FILE *f)
526 {
527 	const struct got_error *err = NULL;
528 	static const size_t blocksize = 512;
529 	size_t n, total, remain;
530 	uint8_t *buf;
531 
532 	*outbuf = NULL;
533 	*outlen = 0;
534 
535 	buf = malloc(blocksize);
536 	if (buf == NULL)
537 		return got_error_from_errno();
538 
539 	remain = blocksize;
540 	total = 0;
541 	while (1) {
542 		if (remain == 0) {
543 			uint8_t *newbuf;
544 			newbuf = reallocarray(buf, 1, total + blocksize);
545 			if (newbuf == NULL) {
546 				err = got_error_from_errno();
547 				goto done;
548 			}
549 			buf = newbuf;
550 			remain += blocksize;
551 		}
552 		n = fread(buf + total, 1, remain, f);
553 		if (n == 0) {
554 			if (ferror(f)) {
555 				err = got_ferror(f, GOT_ERR_IO);
556 				goto done;
557 			}
558 			break; /* EOF */
559 		}
560 		remain -= n;
561 		total += n;
562 	};
563 
564 done:
565 	if (err == NULL) {
566 		*outbuf = buf;
567 		*outlen = total;
568 	} else
569 		free(buf);
570 	return err;
571 }