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 
53 #include "got_lib_delta.h"
54 #include "got_lib_inflate.h"
55 #include "got_lib_object.h"
56 #include "got_lib_object_parse.h"
57 #include "got_lib_object_create.h"
58 #include "got_lib_pack.h"
59 #include "got_lib_sha1.h"
60 #include "got_lib_privsep.h"
61 #include "got_lib_object_cache.h"
62 #include "got_lib_repository.h"
63 
64 #define GOT_PROTOMAX	64
65 #define GOT_HOSTMAX	256
66 #define GOT_PATHMAX	512
67 #define GOT_REPOMAX	256
68 #define GOT_PORTMAX	16
69 #define GOT_URIMAX	1024
70 
71 static int
72 mkpath(char *path)
73 {
74 	char *p, namebuf[PATH_MAX];
75 	struct stat sb;
76 	int done;
77 
78 	while (*path == '/')
79 		path++;
80 	if (strlcpy(namebuf, path, sizeof(namebuf)) >= sizeof(namebuf)) {
81 		errno = ENAMETOOLONG;
82 		return -1;
83 	}
84 
85 	p = namebuf;
86 	for (;;) {
87 		p += strspn(p, "/");
88 		p += strcspn(p, "/");
89 		done = (*p == '\0');
90 		*p = '\0';
91 
92 		if (mkdir(namebuf, 0755) != 0) {
93 			int mkdir_errno = errno;
94 			if (stat(path, &sb) == -1) {
95 				/* Not there; use mkdir()s errno */
96 				errno = mkdir_errno;
97 				return -1;
98 			}
99 			if (!S_ISDIR(sb.st_mode)) {
100 				/* Is there, but isn't a directory */
101 				errno = ENOTDIR;
102 				return -1;
103 			}
104 		}
105 
106 		if (done)
107 			break;
108 		*p = '/';
109 	}
110 
111 	return 0;
112 }
113 
114 static int
115 hassuffix(char *base, char *suf)
116 {
117 	int nb, ns;
118 
119 	nb = strlen(base);
120 	ns = strlen(suf);
121 	if (ns <= nb && strcmp(base + (nb - ns), suf) == 0)
122 		return 1;
123 	return 0;
124 }
125 
126 static int
127 grab(char *dst, int n, char *p, char *e)
128 {
129 	int l;
130 
131 	l = e - p;
132 	if (l >= n) {
133 		errno = ENAMETOOLONG;
134 		return -1;
135 	}
136 	return strlcpy(dst, p, l + 1);
137 }
138 
139 static const struct got_error *
140 dial_ssh(int *fetchfd, char *host, char *port, char *path, char *direction)
141 {
142 	const struct got_error *error = NULL;
143 	int pid, pfd[2];
144 	char cmd[64];
145 
146 	*fetchfd = -1;
147 
148 	if (pipe(pfd) == -1)
149 		return got_error_from_errno("pipe");
150 
151 	pid = fork();
152 	if (pid == -1) {
153 		error = got_error_from_errno("fork");
154 		close(pfd[0]);
155 		close(pfd[1]);
156 		return error;
157 	} else if (pid == 0) {
158 		int n;
159 		close(pfd[1]);
160 		dup2(pfd[0], 0);
161 		dup2(pfd[0], 1);
162 		n = snprintf(cmd, sizeof(cmd), "git-%s-pack", direction);
163 		if (n < 0 || n >= sizeof(cmd))
164 			err(1, "snprintf");
165 		if (execlp("ssh", "ssh", host, cmd, path, NULL) == -1)
166 			err(1, "execlp");
167 		abort(); /* not reached */
168 	} else {
169 		close(pfd[0]);
170 		*fetchfd = pfd[1];
171 		return NULL;
172 	}
173 }
174 
175 static const struct got_error *
176 dial_git(int *fetchfd, char *host, char *port, char *path, char *direction)
177 {
178 	const struct got_error *err = NULL;
179 	struct addrinfo hints, *servinfo, *p;
180 	char *cmd = NULL, *pkt = NULL;
181 	int fd = -1, totlen, r, eaicode;
182 
183 	*fetchfd = -1;
184 
185 	memset(&hints, 0, sizeof hints);
186 	hints.ai_family = AF_UNSPEC;
187 	hints.ai_socktype = SOCK_STREAM;
188 	eaicode = getaddrinfo(host, port, &hints, &servinfo);
189 	if (eaicode)
190 		return got_error_msg(GOT_ERR_ADDRINFO, gai_strerror(eaicode));
191 
192 	for (p = servinfo; p != NULL; p = p->ai_next) {
193 		if ((fd = socket(p->ai_family, p->ai_socktype,
194 		    p->ai_protocol)) == -1)
195 			continue;
196 		if (connect(fd, p->ai_addr, p->ai_addrlen) == 0)
197 			break;
198 		err = got_error_from_errno("connect");
199 		close(fd);
200 	}
201 	if (p == NULL)
202 		goto done;
203 
204 	if (asprintf(&cmd, "git-%s-pack %s", direction, path) == -1) {
205 		err = got_error_from_errno("asprintf");
206 		goto done;
207 	}
208 	totlen = 4 + strlen(cmd) + 1 + strlen("host=") + strlen(host) + 1;
209 	if (asprintf(&pkt, "%04x%s", totlen, cmd) == -1) {
210 		err = got_error_from_errno("asprintf");
211 		goto done;
212 	}
213 	r = write(fd, pkt, strlen(pkt) + 1);
214 	if (r == -1) {
215 		err = got_error_from_errno("write");
216 		goto done;
217 	}
218 	if (asprintf(&pkt, "host=%s", host) == -1) {
219 		err = got_error_from_errno("asprintf");
220 		goto done;
221 	}
222 	r = write(fd, pkt, strlen(pkt) + 1);
223 	if (r == -1) {
224 		err = got_error_from_errno("write");
225 		goto done;
226 	}
227 done:
228 	free(cmd);
229 	free(pkt);
230 	if (err) {
231 		if (fd != -1)
232 			close(fd);
233 	} else
234 		*fetchfd = fd;
235 	return err;
236 }
237 
238 int
239 got_parse_uri(char *uri, char *proto, char *host, char *port, char *path, char *repo)
240 {
241 	char *s, *p, *q;
242 	int n, hasport;
243 
244 	p = strstr(uri, "://");
245 	if (!p) {
246 		//werrstr("missing protocol");
247 		return -1;
248 	}
249 	if (grab(proto, GOT_PROTOMAX, uri, p) == -1)
250 		return -1;
251 	hasport = (strcmp(proto, "git") == 0 || strstr(proto, "http") == proto);
252 	s = p + 3;
253 	p = NULL;
254 	if (!hasport) {
255 		p = strstr(s, ":");
256 		if (p != NULL)
257 			p++;
258 	}
259 	if (p == NULL)
260 		p = strstr(s, "/");
261 	if (p == NULL || strlen(p) == 1) {
262 		//werrstr("missing path");
263 		return -1;
264 	}
265 
266 	q = memchr(s, ':', p - s);
267 	if (q) {
268 		grab(host, GOT_HOSTMAX, s, q);
269 		grab(port, GOT_PORTMAX, q + 1, p);
270 	}else{
271 		grab(host, GOT_HOSTMAX, s, p);
272 		snprintf(port, GOT_PORTMAX, "9418");
273 	}
274 
275 	snprintf(path, GOT_PATHMAX, "%s", p);
276 	p = strrchr(p, '/') + 1;
277 	if (!p || strlen(p) == 0) {
278 		//werrstr("missing repository in uri");
279 		return -1;
280 	}
281 	n = strlen(p);
282 	if (hassuffix(p, ".git"))
283 		n -= 4;
284 	grab(repo, GOT_REPOMAX, p, p + n);
285 	return 0;
286 }
287 
288 const struct got_error*
289 got_fetch(char *uri, char *branch_filter, char *destdir)
290 {
291 	char proto[GOT_PROTOMAX], host[GOT_HOSTMAX], port[GOT_PORTMAX];
292 	char repo[GOT_REPOMAX], path[GOT_PATHMAX];
293 	int imsg_fetchfds[2], imsg_idxfds[2], fetchfd;
294 	int packfd = -1, npackfd, idxfd = -1, nidxfd, status, done = 0;
295 	struct got_object_id *packhash = NULL;
296 	const struct got_error *err;
297 	struct imsgbuf ibuf;
298 	pid_t pid;
299 	char *tmppackpath = NULL, *tmpidxpath = NULL, *default_destdir = NULL;
300 	char *packpath = NULL, *idxpath = NULL, *id_str = NULL;
301 	struct got_pathlist_head symrefs;
302 	struct got_pathlist_entry *pe;
303 
304 	TAILQ_INIT(&symrefs);
305 
306 	fetchfd = -1;
307 	if (got_parse_uri(uri, proto, host, port, path, repo) == -1)
308 		return got_error(GOT_ERR_PARSE_URI);
309 	if (destdir == NULL) {
310 		if (asprintf(&default_destdir, "%s.git", repo) == -1)
311 			return got_error_from_errno("asprintf");
312 	}
313 	err = got_repo_init(destdir ? destdir : default_destdir);
314 	if (err != NULL)
315 		return err;
316 	if (chdir(destdir ? destdir : default_destdir))
317 		return got_error_from_errno("enter new repo");
318 	if (mkpath("objects/pack") == -1)
319 		return got_error_from_errno("mkpath");
320 	err = got_opentemp_named_fd(&tmppackpath, &packfd,
321 	    "objects/pack/fetching.pack");
322 	if (err)
323 		return err;
324 	npackfd = dup(packfd);
325 	if (npackfd == -1)
326 		return got_error_from_errno("dup");
327 	err = got_opentemp_named_fd(&tmpidxpath, &idxfd,
328 	    "objects/pack/fetching.idx");
329 	if (err)
330 		return err;
331 	nidxfd = dup(idxfd);
332 	if (nidxfd == -1)
333 		return got_error_from_errno("dup");
334 
335 	if (strcmp(proto, "ssh") == 0 || strcmp(proto, "git+ssh") == 0)
336 		err = dial_ssh(&fetchfd, host, port, path, "upload");
337 	else if (strcmp(proto, "git") == 0)
338 		err = dial_git(&fetchfd, host, port, path, "upload");
339 	else if (strcmp(proto, "http") == 0 || strcmp(proto, "git+http") == 0)
340 		err = got_error(GOT_ERR_BAD_PROTO);
341 	else
342 		err = got_error(GOT_ERR_BAD_PROTO);
343 	if (err)
344 		return err;
345 
346 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_fetchfds) == -1)
347 		return got_error_from_errno("socketpair");
348 
349 	pid = fork();
350 	if (pid == -1)
351 		return got_error_from_errno("fork");
352 	else if (pid == 0){
353 		got_privsep_exec_child(imsg_fetchfds, GOT_PATH_PROG_FETCH_PACK, ".");
354 	}
355 
356 	if (close(imsg_fetchfds[1]) != 0)
357 		return got_error_from_errno("close");
358 	imsg_init(&ibuf, imsg_fetchfds[0]);
359 	err = got_privsep_send_fetch_req(&ibuf, fetchfd);
360 	if (err != NULL)
361 		return err;
362 	err = got_privsep_send_tmpfd(&ibuf, npackfd);
363 	if (err != NULL)
364 		return err;
365 	npackfd = dup(packfd);
366 	if (npackfd == -1)
367 		return got_error_from_errno("dup");
368 	while (!done) {
369 		struct got_object_id *id;
370 		char *refname;
371 		err = got_privsep_recv_fetch_progress(&done,
372 		    &id, &refname, &symrefs, &ibuf);
373 		if (err != NULL)
374 			return err;
375 		if (done) {
376 			packhash = got_object_id_dup(id);
377 			if (packhash == NULL)
378 				return got_error_from_errno(
379 				    "got_object_id_dup");
380 			printf("symrefs:");
381 			TAILQ_FOREACH(pe, &symrefs, entry) {
382 				printf(" %s:%s", pe->path,
383 				    (const char *)pe->data);
384 			}
385 			printf("\n");
386 		} else if (id) {
387 			char *id_str;
388 			/* TODO Use a progress callback */
389 			err = got_object_id_str(&id_str, id);
390 			if (err)
391 				return err;
392 			printf( "%.12s %s\n", id_str, refname);
393 		}
394 	}
395 	if (waitpid(pid, &status, 0) == -1)
396 		return got_error_from_errno("child exit");
397 
398 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_idxfds) == -1)
399 		return got_error_from_errno("socketpair");
400 	pid = fork();
401 	if (pid == -1)
402 		return got_error_from_errno("fork");
403 	else if (pid == 0)
404 		got_privsep_exec_child(imsg_idxfds, GOT_PATH_PROG_INDEX_PACK, ".");
405 	if (close(imsg_idxfds[1]) != 0)
406 		return got_error_from_errno("close");
407 	imsg_init(&ibuf, imsg_idxfds[0]);
408 
409 	err = got_privsep_send_index_pack_req(&ibuf, npackfd, packhash);
410 	if (err != NULL)
411 		return err;
412 	err = got_privsep_send_tmpfd(&ibuf, nidxfd);
413 	if (err != NULL)
414 		return err;
415 	err = got_privsep_wait_index_pack_done(&ibuf);
416 	if (err != NULL)
417 		return err;
418 	imsg_clear(&ibuf);
419 	if (close(imsg_idxfds[0]) == -1)
420 		return got_error_from_errno("close child");
421 	if (waitpid(pid, &status, 0) == -1)
422 		return got_error_from_errno("child exit");
423 
424 	err = got_object_id_str(&id_str, packhash);
425 	if (err)
426 		return err;
427 	if (asprintf(&packpath, "objects/pack/pack-%s.pack", id_str) == -1)
428 		return got_error_from_errno("asprintf");
429 
430 	if (asprintf(&idxpath, "objects/pack/pack-%s.idx", id_str) == -1)
431 		return got_error_from_errno("asprintf");
432 
433 	if (rename(tmppackpath, packpath) == -1)
434 		return got_error_from_errno3("rename", tmppackpath, packpath);
435 	if (rename(tmpidxpath, idxpath) == -1)
436 		return got_error_from_errno3("rename", tmpidxpath, idxpath);
437 
438 	free(tmppackpath);
439 	free(tmpidxpath);
440 	free(idxpath);
441 	free(packpath);
442 	free(default_destdir);
443 	free(packhash);
444 	TAILQ_FOREACH(pe, &symrefs, entry) {
445 		free((void *)pe->path);
446 		free(pe->data);
447 	}
448 	got_pathlist_free(&symrefs);
449 
450 	return NULL;
451 }