1 /*
2  * Copyright (c) 2023 Omar Polo <op@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 #include <sys/socket.h>
19 #include <sys/stat.h>
20 #include <sys/time.h>
21 #include <sys/tree.h>
22 #include <sys/types.h>
23 #include <sys/uio.h>
24 #include <sys/wait.h>
25 
26 #include <endian.h>
27 #include <limits.h>
28 #include <sha1.h>
29 #include <sha2.h>
30 #include <stdint.h>
31 #include <stdio.h>
32 #include <stdlib.h>
33 #include <string.h>
34 #include <unistd.h>
35 #include <imsg.h>
36 
37 #include "got_error.h"
38 #include "got_cancel.h"
39 #include "got_object.h"
40 #include "got_opentemp.h"
41 #include "got_path.h"
42 #include "got_reference.h"
43 #include "got_repository.h"
44 #include "got_repository_load.h"
45 
46 #include "got_lib_delta.h"
47 #include "got_lib_hash.h"
48 #include "got_lib_object.h"
49 #include "got_lib_object_cache.h"
50 #include "got_lib_pack.h"
51 #include "got_lib_ratelimit.h"
52 #include "got_lib_repository.h"
53 #include "got_lib_privsep.h"
54 
55 #define GIT_BUNDLE_SIGNATURE_V2 "# v2 git bundle\n"
56 
57 #ifndef nitems
58 #define nitems(_a)	(sizeof((_a)) / sizeof((_a)[0]))
59 #endif
60 
61 #ifndef ssizeof
62 #define ssizeof(_x) ((ssize_t)(sizeof(_x)))
63 #endif
64 
65 static const struct got_error *
66 temp_file(int *fd, char **path, const char *ext, struct got_repository *repo)
67 {
68 	const struct got_error *err;
69 	char p[PATH_MAX];
70 	int r;
71 
72 	*path = NULL;
73 
74 	r = snprintf(p, sizeof(p), "%s/%s/loading",
75 	    got_repo_get_path_git_dir(repo), GOT_OBJECTS_PACK_DIR);
76 	if (r < 0 || (size_t)r >= sizeof(p))
77 		return got_error_from_errno("snprintf");
78 
79 	err = got_opentemp_named_fd(path, fd, p, ext);
80 	if (err)
81 		return err;
82 
83 	if (fchmod(*fd, GOT_DEFAULT_FILE_MODE) == -1)
84 		return got_error_from_errno("fchmod");
85 
86 	return NULL;
87 }
88 
89 static const struct got_error *
90 load_report_progress(got_load_progress_cb progress_cb, void *progress_arg,
91     struct got_ratelimit *rl, off_t packsiz, int nobj_total,
92     int nobj_indexed, int nobj_loose, int nobj_resolved)
93 {
94 	const struct got_error *err;
95 	int elapsed;
96 
97 	if (progress_cb == NULL)
98 		return NULL;
99 
100 	err = got_ratelimit_check(&elapsed, rl);
101 	if (err || !elapsed)
102 		return err;
103 
104 	return progress_cb(progress_arg, packsiz, nobj_total, nobj_indexed,
105 	    nobj_loose, nobj_resolved);
106 }
107 
108 static const struct got_error *
109 copypack(FILE *in, int outfd, off_t *tot,
110     struct got_object_id *id, struct got_ratelimit *rl,
111     got_load_progress_cb progress_cb, void *progress_arg,
112     got_cancel_cb cancel_cb, void *cancel_arg)
113 {
114 	const struct got_error *err;
115 	struct got_hash hash;
116 	struct got_object_id expected_id;
117 	char buf[BUFSIZ], sha1buf[SHA1_DIGEST_LENGTH];
118 	size_t r, sha1buflen = 0;
119 
120 	*tot = 0;
121 	got_hash_init(&hash, GOT_HASH_SHA1);
122 
123 	for (;;) {
124 		err = cancel_cb(cancel_arg);
125 		if (err)
126 			return err;
127 
128 		r = fread(buf, 1, sizeof(buf), in);
129 		if (r == 0)
130 			break;
131 
132 		/*
133 		 * An expected SHA1 checksum sits at the end of the
134 		 * pack file.  Since we don't know the file size ahead
135 		 * of time we have to keep SHA1_DIGEST_LENGTH bytes
136 		 * buffered and avoid mixing those bytes int our hash
137 		 * computation until we know for sure that additional
138 		 * pack file data bytes follow.
139 		 *
140 		 * We can assume that BUFSIZE is greater than
141 		 * SHA1_DIGEST_LENGTH and that a short read means that
142 		 * we've reached EOF.
143 		 */
144 
145 		if (r >= sizeof(sha1buf)) {
146 			*tot += sha1buflen;
147 			got_hash_update(&hash, sha1buf, sha1buflen);
148 			if (write(outfd, sha1buf, sha1buflen) == -1)
149 				return got_error_from_errno("write");
150 
151 			r -= sizeof(sha1buf);
152 			memcpy(sha1buf, &buf[r], sizeof(sha1buf));
153 			sha1buflen = sizeof(sha1buf);
154 
155 			*tot += r;
156 			got_hash_update(&hash, buf, r);
157 			if (write(outfd, buf, r) == -1)
158 				return got_error_from_errno("write");
159 
160 			err = load_report_progress(progress_cb, progress_arg,
161 			    rl, *tot, 0, 0, 0, 0);
162 			if (err)
163 				return err;
164 
165 			continue;
166 		}
167 
168 		if (sha1buflen == 0)
169 			return got_error(GOT_ERR_BAD_PACKFILE);
170 
171 		/* short read, we've reached EOF */
172 		*tot += r;
173 		got_hash_update(&hash, sha1buf, r);
174 		if (write(outfd, sha1buf, r) == -1)
175 			return got_error_from_errno("write");
176 
177 		memmove(&sha1buf[0], &sha1buf[r], sizeof(sha1buf) - r);
178 		memcpy(&sha1buf[sizeof(sha1buf) - r], buf, r);
179 		break;
180 	}
181 
182 	if (sha1buflen == 0)
183 		return got_error(GOT_ERR_BAD_PACKFILE);
184 
185 	got_hash_final_object_id(&hash, id);
186 
187 	/* XXX SHA256 */
188 	memset(&expected_id, 0, sizeof(expected_id));
189 	memcpy(&expected_id.sha1, sha1buf, sizeof(expected_id.sha1));
190 
191 	if (got_object_id_cmp(id, &expected_id) != 0)
192 		return got_error(GOT_ERR_PACKIDX_CSUM);
193 
194 	/* re-add the expected hash at the end of the pack */
195 	if (write(outfd, sha1buf, sizeof(sha1buf)) == -1)
196 		return got_error_from_errno("write");
197 
198 	*tot += sizeof(sha1buf);
199 	err = progress_cb(progress_arg, *tot, 0, 0, 0, 0);
200 	if (err)
201 		return err;
202 
203 	return NULL;
204 }
205 
206 const struct got_error *
207 got_repo_load(FILE *in, struct got_pathlist_head *refs_found,
208     struct got_repository *repo, int list_refs_only, int noop,
209     got_load_progress_cb progress_cb, void *progress_arg,
210     got_cancel_cb cancel_cb, void *cancel_arg)
211 {
212 	const struct got_error *err = NULL;
213 	struct got_object_id id;
214 	struct got_object *obj;
215 	struct got_packfile_hdr pack_hdr;
216 	struct got_ratelimit rl;
217 	struct imsgbuf idxibuf;
218 	const char *repo_path;
219 	char *packpath = NULL, *idxpath = NULL;
220 	char *tmppackpath = NULL, *tmpidxpath = NULL;
221 	int packfd = -1, idxfd = -1;
222 	char *spc, *refname, *id_str = NULL;
223 	char *line = NULL;
224 	size_t linesize = 0;
225 	ssize_t linelen;
226 	size_t i;
227 	ssize_t n;
228 	off_t packsiz;
229 	int tmpfds[3] = {-1, -1, -1};
230 	int imsg_idxfds[2] = {-1, -1};
231 	int ch, done, nobj, idxstatus;
232 	pid_t idxpid;
233 
234 	got_ratelimit_init(&rl, 0, 500);
235 
236 	repo_path = got_repo_get_path_git_dir(repo);
237 
238 	linelen = getline(&line, &linesize, in);
239 	if (linelen == -1) {
240 		err = got_ferror(in, GOT_ERR_IO);
241 		goto done;
242 	}
243 
244 	if (strcmp(line, GIT_BUNDLE_SIGNATURE_V2) != 0) {
245 		err = got_error(GOT_ERR_BUNDLE_FORMAT);
246 		goto done;
247 	}
248 
249 	/* Parse the prerequisite */
250 	for (;;) {
251 		ch = fgetc(in);
252 		if (ch != '-') {
253 			if (ch != EOF)
254 				ungetc(ch, in);
255 			break;
256 		}
257 
258 		linelen = getline(&line, &linesize, in);
259 		if (linelen == -1) {
260 			err = got_ferror(in, GOT_ERR_IO);
261 			goto done;
262 		}
263 
264 		if (line[linelen - 1] == '\n')
265 			line[linelen - 1] = '\0';
266 
267 		if (!got_parse_object_id(&id, line, GOT_HASH_SHA1)) {
268 			err = got_error_path(line, GOT_ERR_BAD_OBJ_ID_STR);
269 			goto done;
270 		}
271 
272 		err = got_object_open(&obj, repo, &id);
273 		if (err)
274 			goto done;
275 		got_object_close(obj);
276 	}
277 
278 	/* Read references */
279 	for (;;) {
280 		struct got_object_id *id;
281 		char *dup;
282 
283 		linelen = getline(&line, &linesize, in);
284 		if (linelen == -1) {
285 			err = got_ferror(in, GOT_ERR_IO);
286 			goto done;
287 		}
288 		if (line[linelen - 1] == '\n')
289 			line[linelen - 1] = '\0';
290 		if (*line == '\0')
291 			break;
292 
293 		spc = strchr(line, ' ');
294 		if (spc == NULL) {
295 			err = got_error(GOT_ERR_IO);
296 			goto done;
297 		}
298 		*spc = '\0';
299 
300 		refname = spc + 1;
301 		if (!got_ref_name_is_valid(refname)) {
302 			err = got_error(GOT_ERR_BAD_REF_DATA);
303 			goto done;
304 		}
305 
306 		id = malloc(sizeof(*id));
307 		if (id == NULL) {
308 			err = got_error_from_errno("malloc");
309 			goto done;
310 		}
311 
312 		if (!got_parse_object_id(id, line, GOT_HASH_SHA1)) {
313 			free(id);
314 			err = got_error(GOT_ERR_BAD_OBJ_ID_STR);
315 			goto done;
316 		}
317 
318 		dup = strdup(refname);
319 		if (dup == NULL) {
320 			free(id);
321 			err = got_error_from_errno("strdup");
322 			goto done;
323 		}
324 
325 		err = got_pathlist_append(refs_found, dup, id);
326 		if (err) {
327 			free(id);
328 			free(dup);
329 			goto done;
330 		}
331 	}
332 
333 	if (list_refs_only)
334 		goto done;
335 
336 	err = temp_file(&packfd, &tmppackpath, ".pack", repo);
337 	if (err)
338 		goto done;
339 
340 	err = temp_file(&idxfd, &tmpidxpath, ".idx", repo);
341 	if (err)
342 		goto done;
343 
344 	err = copypack(in, packfd, &packsiz, &id, &rl,
345 	    progress_cb, progress_arg, cancel_cb, cancel_arg);
346 	if (err)
347 		goto done;
348 
349 	if (lseek(packfd, 0, SEEK_SET) == -1) {
350 		err = got_error_from_errno("lseek");
351 		goto done;
352 	}
353 
354 	/* Safety checks on the pack' content. */
355 	if (packsiz <= ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH) {
356 		err = got_error_msg(GOT_ERR_BAD_PACKFILE, "short pack file");
357 		goto done;
358 	}
359 
360 	n = read(packfd, &pack_hdr, ssizeof(pack_hdr));
361 	if (n == -1) {
362 		err = got_error_from_errno("read");
363 		goto done;
364 	}
365 	if (n != ssizeof(pack_hdr)) {
366 		err = got_error(GOT_ERR_IO);
367 		goto done;
368 	}
369 	if (pack_hdr.signature != htobe32(GOT_PACKFILE_SIGNATURE)) {
370 		err = got_error_msg(GOT_ERR_BAD_PACKFILE,
371 		    "bad pack file signature");
372 		goto done;
373 	}
374 	if (pack_hdr.version != htobe32(GOT_PACKFILE_VERSION)) {
375 		err = got_error_msg(GOT_ERR_BAD_PACKFILE,
376 		    "bad pack file version");
377 		goto done;
378 	}
379 	nobj = be32toh(pack_hdr.nobjects);
380 	if (nobj == 0 &&
381 	    packsiz > ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH) {
382 		err = got_error_msg(GOT_ERR_BAD_PACKFILE,
383 		    "bad pack file with zero objects");
384 		goto done;
385 	}
386 	if (nobj != 0 &&
387 	    packsiz <= ssizeof(pack_hdr) + SHA1_DIGEST_LENGTH) {
388 		err = got_error_msg(GOT_ERR_BAD_PACKFILE,
389 		    "empty pack file with non-zero object count");
390 		goto done;
391 	}
392 
393 	/* nothing to do if there are no objects. */
394 	if (nobj == 0)
395 		goto done;
396 
397 	for (i = 0; i < nitems(tmpfds); i++) {
398 		tmpfds[i] = got_opentempfd();
399 		if (tmpfds[i] == -1) {
400 			err = got_error_from_errno("got_opentempfd");
401 			goto done;
402 		}
403 	}
404 
405 	if (lseek(packfd, 0, SEEK_SET) == -1) {
406 		err = got_error_from_errno("lseek");
407 		goto done;
408 	}
409 
410 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_idxfds) == -1) {
411 		err = got_error_from_errno("socketpair");
412 		goto done;
413 	}
414 	idxpid = fork();
415 	if (idxpid == -1) {
416 		err= got_error_from_errno("fork");
417 		goto done;
418 	} else if (idxpid == 0)
419 		got_privsep_exec_child(imsg_idxfds,
420 		    GOT_PATH_PROG_INDEX_PACK, tmppackpath);
421 	if (close(imsg_idxfds[1]) == -1) {
422 		err = got_error_from_errno("close");
423 		goto done;
424 	}
425 	imsg_idxfds[1] = -1;
426 	imsg_init(&idxibuf, imsg_idxfds[0]);
427 
428 	err = got_privsep_send_index_pack_req(&idxibuf, id.sha1, packfd);
429 	if (err)
430 		goto done;
431 	packfd = -1;
432 
433 	err = got_privsep_send_index_pack_outfd(&idxibuf, idxfd);
434 	if (err)
435 		goto done;
436 	idxfd = -1;
437 
438 	for (i = 0; i < nitems(tmpfds); i++) {
439 		err = got_privsep_send_tmpfd(&idxibuf, tmpfds[i]);
440 		if (err != NULL)
441 			goto done;
442 		tmpfds[i] = -1;
443 	}
444 
445 	done = 0;
446 	while (!done) {
447 		int nobj_total, nobj_indexed, nobj_loose, nobj_resolved;
448 
449 		err = got_privsep_recv_index_progress(&done, &nobj_total,
450 		    &nobj_indexed, &nobj_loose, &nobj_resolved, &idxibuf);
451 		if (err)
452 			goto done;
453 		if (nobj_indexed != 0) {
454 			err = load_report_progress(progress_cb, progress_arg,
455 			    &rl, packsiz, nobj_total, nobj_indexed,
456 			    nobj_loose, nobj_resolved);
457 			if (err)
458 				goto done;
459 		}
460 	}
461 	if (close(imsg_idxfds[0]) == -1) {
462 		err = got_error_from_errno("close");
463 		goto done;
464 	}
465 	imsg_idxfds[0] = -1;
466 	if (waitpid(idxpid, &idxstatus, 0) == -1) {
467 		err = got_error_from_errno("waitpid");
468 		goto done;
469 	}
470 
471 	if (noop)
472 		goto done;
473 
474 	err = got_object_id_str(&id_str, &id);
475 	if (err)
476 		goto done;
477 
478 	if (asprintf(&packpath, "%s/%s/pack-%s.pack", repo_path,
479 	    GOT_OBJECTS_PACK_DIR, id_str) == -1) {
480 		err = got_error_from_errno("asprintf");
481 		goto done;
482 	}
483 
484 	if (asprintf(&idxpath, "%s/%s/pack-%s.idx", repo_path,
485 	    GOT_OBJECTS_PACK_DIR, id_str) == -1) {
486 		err = got_error_from_errno("asprintf");
487 		goto done;
488 	}
489 
490 	if (rename(tmppackpath, packpath) == -1) {
491 		err = got_error_from_errno3("rename", tmppackpath, packpath);
492 		goto done;
493 	}
494 	free(tmppackpath);
495 	tmppackpath = NULL;
496 
497 	if (rename(tmpidxpath, idxpath) == -1) {
498 		err = got_error_from_errno3("rename", tmpidxpath, idxpath);
499 		goto done;
500 	}
501 	free(tmpidxpath);
502 	tmpidxpath = NULL;
503 
504  done:
505 	free(line);
506 	free(packpath);
507 	free(idxpath);
508 	free(id_str);
509 
510 	if (tmppackpath && unlink(tmppackpath) == -1 && err == NULL)
511 		err = got_error_from_errno2("unlink", tmppackpath);
512 	if (packfd != -1 && close(packfd) == -1 && err == NULL)
513 		err = got_error_from_errno("close");
514 	free(tmppackpath);
515 
516 	if (tmpidxpath && unlink(tmpidxpath) == -1 && err == NULL)
517 		err = got_error_from_errno2("unlink", tmpidxpath);
518 	if (idxfd != -1 && close(idxfd) == -1 && err == NULL)
519 		err = got_error_from_errno("close");
520 	free(tmpidxpath);
521 
522 	if (imsg_idxfds[0] != -1 && close(imsg_idxfds[0]) == -1 && err == NULL)
523 		err = got_error_from_errno("close");
524 	if (imsg_idxfds[1] != -1 && close(imsg_idxfds[1]) == -1 && err == NULL)
525 		err = got_error_from_errno("close");
526 
527 	for (i = 0; i < nitems(tmpfds); ++i)
528 		if (tmpfds[i] != -1 && close(tmpfds[i]) == -1 && err == NULL)
529 			err = got_error_from_errno("close");
530 
531 	return err;
532 }