1 /*
2  * Copyright (c) 2018, 2019 Ori Bernstein <ori@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/wait.h>
23 #include <sys/syslimits.h>
24 #include <sys/resource.h>
25 #include <sys/socket.h>
26 
27 #include <errno.h>
28 #include <err.h>
29 #include <fcntl.h>
30 #include <stdio.h>
31 #include <stdlib.h>
32 #include <string.h>
33 #include <stdint.h>
34 #include <sha1.h>
35 #include <zlib.h>
36 #include <ctype.h>
37 #include <limits.h>
38 #include <imsg.h>
39 #include <time.h>
40 #include <uuid.h>
41 #include <netdb.h>
42 #include <netinet/in.h>
43 
44 #include "got_error.h"
45 #include "got_reference.h"
46 #include "got_repository.h"
47 #include "got_path.h"
48 #include "got_cancel.h"
49 #include "got_worktree.h"
50 #include "got_object.h"
51 #include "got_opentemp.h"
52 #include "got_fetch.h"
53 
54 #include "got_lib_delta.h"
55 #include "got_lib_inflate.h"
56 #include "got_lib_object.h"
57 #include "got_lib_object_parse.h"
58 #include "got_lib_object_create.h"
59 #include "got_lib_pack.h"
60 #include "got_lib_sha1.h"
61 #include "got_lib_privsep.h"
62 #include "got_lib_object_cache.h"
63 #include "got_lib_repository.h"
64 
65 #define GOT_PROTOMAX	64
66 #define GOT_HOSTMAX	256
67 #define GOT_PATHMAX	512
68 #define GOT_REPOMAX	256
69 #define GOT_PORTMAX	16
70 #define GOT_URIMAX	1024
71 
72 static int
73 hassuffix(char *base, char *suf)
74 {
75 	int nb, ns;
76 
77 	nb = strlen(base);
78 	ns = strlen(suf);
79 	if (ns <= nb && strcmp(base + (nb - ns), suf) == 0)
80 		return 1;
81 	return 0;
82 }
83 
84 static const struct got_error *
85 dial_ssh(int *fetchfd, const char *host, const char *port, const char *path,
86     const char *direction)
87 {
88 	const struct got_error *error = NULL;
89 	int pid, pfd[2];
90 	char cmd[64];
91 
92 	*fetchfd = -1;
93 
94 	if (pipe(pfd) == -1)
95 		return got_error_from_errno("pipe");
96 
97 	pid = fork();
98 	if (pid == -1) {
99 		error = got_error_from_errno("fork");
100 		close(pfd[0]);
101 		close(pfd[1]);
102 		return error;
103 	} else if (pid == 0) {
104 		int n;
105 		close(pfd[1]);
106 		dup2(pfd[0], 0);
107 		dup2(pfd[0], 1);
108 		n = snprintf(cmd, sizeof(cmd), "git-%s-pack", direction);
109 		if (n < 0 || n >= sizeof(cmd))
110 			err(1, "snprintf");
111 		if (execlp("ssh", "ssh", host, cmd, path, NULL) == -1)
112 			err(1, "execlp");
113 		abort(); /* not reached */
114 	} else {
115 		close(pfd[0]);
116 		*fetchfd = pfd[1];
117 		return NULL;
118 	}
119 }
120 
121 static const struct got_error *
122 dial_git(int *fetchfd, const char *host, const char *port, const char *path,
123     const char *direction)
124 {
125 	const struct got_error *err = NULL;
126 	struct addrinfo hints, *servinfo, *p;
127 	char *cmd = NULL, *pkt = NULL;
128 	int fd = -1, totlen, r, eaicode;
129 
130 	*fetchfd = -1;
131 
132 	memset(&hints, 0, sizeof hints);
133 	hints.ai_family = AF_UNSPEC;
134 	hints.ai_socktype = SOCK_STREAM;
135 	eaicode = getaddrinfo(host, port, &hints, &servinfo);
136 	if (eaicode)
137 		return got_error_msg(GOT_ERR_ADDRINFO, gai_strerror(eaicode));
138 
139 	for (p = servinfo; p != NULL; p = p->ai_next) {
140 		if ((fd = socket(p->ai_family, p->ai_socktype,
141 		    p->ai_protocol)) == -1)
142 			continue;
143 		if (connect(fd, p->ai_addr, p->ai_addrlen) == 0)
144 			break;
145 		err = got_error_from_errno("connect");
146 		close(fd);
147 	}
148 	if (p == NULL)
149 		goto done;
150 
151 	if (asprintf(&cmd, "git-%s-pack %s", direction, path) == -1) {
152 		err = got_error_from_errno("asprintf");
153 		goto done;
154 	}
155 	totlen = 4 + strlen(cmd) + 1 + strlen("host=") + strlen(host) + 1;
156 	if (asprintf(&pkt, "%04x%s", totlen, cmd) == -1) {
157 		err = got_error_from_errno("asprintf");
158 		goto done;
159 	}
160 	r = write(fd, pkt, strlen(pkt) + 1);
161 	if (r == -1) {
162 		err = got_error_from_errno("write");
163 		goto done;
164 	}
165 	if (asprintf(&pkt, "host=%s", host) == -1) {
166 		err = got_error_from_errno("asprintf");
167 		goto done;
168 	}
169 	r = write(fd, pkt, strlen(pkt) + 1);
170 	if (r == -1) {
171 		err = got_error_from_errno("write");
172 		goto done;
173 	}
174 done:
175 	free(cmd);
176 	free(pkt);
177 	if (err) {
178 		if (fd != -1)
179 			close(fd);
180 	} else
181 		*fetchfd = fd;
182 	return err;
183 }
184 
185 const struct got_error *
186 got_fetch_parse_uri(char **proto, char **host, char **port,
187     char **server_path, char **repo_name, const char *uri)
188 {
189 	const struct got_error *err = NULL;
190 	char *s, *p, *q;
191 	int n, hasport;
192 
193 	*proto = *host = *port = *server_path = *repo_name = NULL;
194 
195 	p = strstr(uri, "://");
196 	if (!p) {
197 		return got_error(GOT_ERR_PARSE_URI);
198 	}
199 	*proto = strndup(uri, p - uri);
200 	if (proto == NULL) {
201 		err = got_error_from_errno("strndup");
202 		goto done;
203 	}
204 
205 	hasport = (strcmp(*proto, "git") == 0 ||
206 	    strstr(*proto, "http") == *proto);
207 	s = p + 3;
208 	p = NULL;
209 	if (!hasport) {
210 		p = strstr(s, ":");
211 		if (p != NULL)
212 			p++;
213 	}
214 	if (p == NULL)
215 		p = strstr(s, "/");
216 	if (p == NULL || strlen(p) == 1) {
217 		err = got_error(GOT_ERR_PARSE_URI);
218 		goto done;
219 	}
220 
221 	q = memchr(s, ':', p - s);
222 	if (q) {
223 		*host = strndup(s, q - s);
224 		if (*host == NULL) {
225 			err = got_error_from_errno("strndup");
226 			goto done;
227 		}
228 		*port = strndup(q + 1, p - (q + 1));
229 		if (*port == NULL) {
230 			err = got_error_from_errno("strndup");
231 			goto done;
232 		}
233 	} else {
234 		*host = strndup(s, p - s);
235 		if (*host == NULL) {
236 			err = got_error_from_errno("strndup");
237 			goto done;
238 		}
239 		if (asprintf(port, "%u", GOT_DEFAULT_GIT_PORT) == -1) {
240 			err = got_error_from_errno("asprintf");
241 			goto done;
242 		}
243 	}
244 
245 	*server_path = strdup(p);
246 	if (*server_path == NULL) {
247 		err = got_error_from_errno("strdup");
248 		goto done;
249 	}
250 
251 	p = strrchr(p, '/') + 1;
252 	if (!p || strlen(p) == 0) {
253 		//werrstr("missing repository in uri");
254 		err = got_error(GOT_ERR_PARSE_URI);
255 		goto done;
256 	}
257 	n = strlen(p);
258 	if (hassuffix(p, ".git"))
259 		n -= 4;
260 	*repo_name = strndup(p, (p + n) - p);
261 	if (*repo_name == NULL) {
262 		err = got_error_from_errno("strndup");
263 		goto done;
264 	}
265 done:
266 	if (err) {
267 		free(*proto);
268 		*proto = NULL;
269 		free(*host);
270 		*host = NULL;
271 		free(*port);
272 		*port = NULL;
273 		free(*server_path);
274 		*server_path = NULL;
275 		free(*repo_name);
276 		*repo_name = NULL;
277 	}
278 	return err;
279 }
280 
281 const struct got_error*
282 got_fetch(struct got_object_id **pack_hash, struct got_pathlist_head *refs,
283     struct got_pathlist_head *symrefs, const char *proto, const char *host,
284     const char *port, const char *server_path, const char *repo_name,
285     const char *branch_filter, struct got_repository *repo)
286 {
287 	int imsg_fetchfds[2], imsg_idxfds[2], fetchfd = -1;
288 	int packfd = -1, npackfd = -1, idxfd = -1, nidxfd = -1;
289 	int status, done = 0;
290 	const struct got_error *err;
291 	struct imsgbuf ibuf;
292 	pid_t pid;
293 	char *tmppackpath = NULL, *tmpidxpath = NULL;
294 	char *packpath = NULL, *idxpath = NULL, *id_str = NULL;
295 	const char *repo_path = got_repo_get_path(repo);
296 	struct got_pathlist_entry *pe;
297 	char *path;
298 
299 	*pack_hash = NULL;
300 
301 	fetchfd = -1;
302 
303 	if (asprintf(&path, "%s/%s/fetching.pack",
304 	    repo_path, GOT_OBJECTS_PACK_DIR) == -1) {
305 		err = got_error_from_errno("asprintf");
306 		goto done;
307 	}
308 	err = got_opentemp_named_fd(&tmppackpath, &packfd, path);
309 	free(path);
310 	if (err)
311 		goto done;
312 	npackfd = dup(packfd);
313 	if (npackfd == -1) {
314 		err = got_error_from_errno("dup");
315 		goto done;
316 	}
317 	if (asprintf(&path, "%s/%s/fetching.idx",
318 	    repo_path, GOT_OBJECTS_PACK_DIR) == -1) {
319 		err = got_error_from_errno("asprintf");
320 		goto done;
321 	}
322 	err = got_opentemp_named_fd(&tmpidxpath, &idxfd, path);
323 	free(path);
324 	if (err)
325 		goto done;
326 	nidxfd = dup(idxfd);
327 	if (nidxfd == -1) {
328 		err = got_error_from_errno("dup");
329 		goto done;
330 	}
331 
332 	if (strcmp(proto, "ssh") == 0 || strcmp(proto, "git+ssh") == 0)
333 		err = dial_ssh(&fetchfd, host, port, server_path, "upload");
334 	else if (strcmp(proto, "git") == 0)
335 		err = dial_git(&fetchfd, host, port, server_path, "upload");
336 	else if (strcmp(proto, "http") == 0 || strcmp(proto, "git+http") == 0)
337 		err = got_error(GOT_ERR_BAD_PROTO);
338 	else
339 		err = got_error(GOT_ERR_BAD_PROTO);
340 	if (err)
341 		goto done;
342 
343 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_fetchfds) == -1) {
344 		err = got_error_from_errno("socketpair");
345 		goto done;
346 	}
347 
348 	pid = fork();
349 	if (pid == -1) {
350 		err = got_error_from_errno("fork");
351 		goto done;
352 	} else if (pid == 0){
353 		got_privsep_exec_child(imsg_fetchfds,
354 		    GOT_PATH_PROG_FETCH_PACK, ".");
355 	}
356 
357 	if (close(imsg_fetchfds[1]) != 0) {
358 		err = got_error_from_errno("close");
359 		goto done;
360 	}
361 	imsg_init(&ibuf, imsg_fetchfds[0]);
362 	err = got_privsep_send_fetch_req(&ibuf, fetchfd);
363 	if (err != NULL)
364 		goto done;
365 	fetchfd = -1;
366 	err = got_privsep_send_tmpfd(&ibuf, npackfd);
367 	if (err != NULL)
368 		goto done;
369 	npackfd = dup(packfd);
370 	if (npackfd == -1) {
371 		err = got_error_from_errno("dup");
372 		goto done;
373 	}
374 
375 	while (!done) {
376 		struct got_object_id *id = NULL;
377 		char *refname = NULL;
378 
379 		err = got_privsep_recv_fetch_progress(&done,
380 		    &id, &refname, symrefs, &ibuf);
381 		if (err != NULL)
382 			goto done;
383 		if (done)
384 			*pack_hash = id;
385 		else if (refname && id) {
386 			err = got_pathlist_append(refs, refname, id);
387 			if (err)
388 				goto done;
389 		}
390 		/* TODO remote status / download progress callback */
391 	}
392 	if (waitpid(pid, &status, 0) == -1) {
393 		err = got_error_from_errno("waitpid");
394 		goto done;
395 	}
396 
397 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_idxfds) == -1) {
398 		err = got_error_from_errno("socketpair");
399 		goto done;
400 	}
401 	pid = fork();
402 	if (pid == -1) {
403 		err= got_error_from_errno("fork");
404 		goto done;
405 	} else if (pid == 0)
406 		got_privsep_exec_child(imsg_idxfds,
407 		    GOT_PATH_PROG_INDEX_PACK, ".");
408 	if (close(imsg_idxfds[1]) != 0) {
409 		err = got_error_from_errno("close");
410 		goto done;
411 	}
412 	imsg_init(&ibuf, imsg_idxfds[0]);
413 
414 	err = got_privsep_send_index_pack_req(&ibuf, npackfd, *pack_hash);
415 	if (err != NULL)
416 		goto done;
417 	npackfd = -1;
418 	err = got_privsep_send_tmpfd(&ibuf, nidxfd);
419 	if (err != NULL)
420 		goto done;
421 	nidxfd = -1;
422 	err = got_privsep_wait_index_pack_done(&ibuf);
423 	if (err != NULL)
424 		goto done;
425 	imsg_clear(&ibuf);
426 	if (close(imsg_idxfds[0]) == -1) {
427 		err = got_error_from_errno("close");
428 		goto done;
429 	}
430 	if (waitpid(pid, &status, 0) == -1) {
431 		err = got_error_from_errno("waitpid");
432 		goto done;
433 	}
434 
435 	err = got_object_id_str(&id_str, *pack_hash);
436 	if (err)
437 		goto done;
438 	if (asprintf(&packpath, "%s/%s/pack-%s.pack",
439 	    repo_path, GOT_OBJECTS_PACK_DIR, id_str) == -1) {
440 		err = got_error_from_errno("asprintf");
441 		goto done;
442 	}
443 
444 	if (asprintf(&idxpath, "%s/%s/pack-%s.idx",
445 	    repo_path, GOT_OBJECTS_PACK_DIR, id_str) == -1) {
446 		err = got_error_from_errno("asprintf");
447 		goto done;
448 	}
449 
450 	if (rename(tmppackpath, packpath) == -1) {
451 		err = got_error_from_errno3("rename", tmppackpath, packpath);
452 		goto done;
453 	}
454 	if (rename(tmpidxpath, idxpath) == -1) {
455 		err = got_error_from_errno3("rename", tmpidxpath, idxpath);
456 		goto done;
457 	}
458 
459 done:
460 	if (fetchfd != -1 && close(fetchfd) == -1 && err == NULL)
461 		err = got_error_from_errno("close");
462 	if (npackfd != -1 && close(npackfd) == -1 && err == NULL)
463 		err = got_error_from_errno("close");
464 	if (packfd != -1 && close(packfd) == -1 && err == NULL)
465 		err = got_error_from_errno("close");
466 	if (idxfd != -1 && close(idxfd) == -1 && err == NULL)
467 		err = got_error_from_errno("close");
468 	free(tmppackpath);
469 	free(tmpidxpath);
470 	free(idxpath);
471 	free(packpath);
472 
473 	if (err) {
474 		free(*pack_hash);
475 		*pack_hash = NULL;
476 		TAILQ_FOREACH(pe, refs, entry) {
477 			free((void *)pe->path);
478 			free(pe->data);
479 		}
480 		got_pathlist_free(refs);
481 		TAILQ_FOREACH(pe, symrefs, entry) {
482 			free((void *)pe->path);
483 			free(pe->data);
484 		}
485 		got_pathlist_free(symrefs);
486 	}
487 	return err;
488 }