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, l, 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 ((l = asprintf(&cmd, "git-%s-pack %s\n", direction, path)) == -1) {
205 		err = got_error_from_errno("asprintf");
206 		goto done;
207 	}
208 	if ((l = asprintf(&pkt, "%04x%s", l + 4, cmd)) == -1) {
209 		err = got_error_from_errno("asprintf");
210 		goto done;
211 	}
212 	r = write(fd, pkt, l);
213 	if (r == -1)
214 		err = got_error_from_errno("write");
215 done:
216 	free(cmd);
217 	free(pkt);
218 	if (err) {
219 		if (fd != -1)
220 			close(fd);
221 	} else
222 		*fetchfd = fd;
223 	return err;
224 }
225 
226 int
227 got_parse_uri(char *uri, char *proto, char *host, char *port, char *path, char *repo)
228 {
229 	char *s, *p, *q;
230 	int n, hasport;
231 
232 	p = strstr(uri, "://");
233 	if (!p) {
234 		//werrstr("missing protocol");
235 		return -1;
236 	}
237 	if (grab(proto, GOT_PROTOMAX, uri, p) == -1)
238 		return -1;
239 	hasport = (strcmp(proto, "git") == 0 || strstr(proto, "http") == proto);
240 	s = p + 3;
241 	p = NULL;
242 	if (!hasport) {
243 		p = strstr(s, ":");
244 		if (p != NULL)
245 			p++;
246 	}
247 	if (p == NULL)
248 		p = strstr(s, "/");
249 	if (p == NULL || strlen(p) == 1) {
250 		//werrstr("missing path");
251 		return -1;
252 	}
253 
254 	q = memchr(s, ':', p - s);
255 	if (q) {
256 		grab(host, GOT_HOSTMAX, s, q);
257 		grab(port, GOT_PORTMAX, q + 1, p);
258 	}else{
259 		grab(host, GOT_HOSTMAX, s, p);
260 		snprintf(port, GOT_PORTMAX, "9418");
261 	}
262 
263 	snprintf(path, GOT_PATHMAX, "%s", p);
264 	p = strrchr(p, '/') + 1;
265 	if (!p || strlen(p) == 0) {
266 		//werrstr("missing repository in uri");
267 		return -1;
268 	}
269 	n = strlen(p);
270 	if (hassuffix(p, ".git"))
271 		n -= 4;
272 	grab(repo, GOT_REPOMAX, p, p + n);
273 	return 0;
274 }
275 
276 const struct got_error*
277 got_fetch(char *uri, char *branch_filter, char *destdir)
278 {
279 	char proto[GOT_PROTOMAX], host[GOT_HOSTMAX], port[GOT_PORTMAX];
280 	char repo[GOT_REPOMAX], path[GOT_PATHMAX];
281 	int imsg_fetchfds[2], imsg_idxfds[2], fetchfd;
282 	int packfd = -1, npackfd, idxfd = -1, nidxfd, status, done = 0;
283 	struct got_object_id *packhash = NULL;
284 	const struct got_error *err;
285 	struct imsgbuf ibuf;
286 	pid_t pid;
287 	char *tmppackpath = NULL, *tmpidxpath = NULL, *default_destdir = NULL;
288 	char *packpath = NULL, *idxpath = NULL, *id_str = NULL;
289 
290 	fetchfd = -1;
291 	if (got_parse_uri(uri, proto, host, port, path, repo) == -1)
292 		return got_error(GOT_ERR_PARSE_URI);
293 	if (destdir == NULL) {
294 		if (asprintf(&default_destdir, "%s.git", repo) == -1)
295 			return got_error_from_errno("asprintf");
296 	}
297 	err = got_repo_init(destdir ? destdir : default_destdir);
298 	if (err != NULL)
299 		return err;
300 	if (chdir(destdir ? destdir : default_destdir))
301 		return got_error_from_errno("enter new repo");
302 	if (mkpath("objects/pack") == -1)
303 		return got_error_from_errno("mkpath");
304 	err = got_opentemp_named_fd(&tmppackpath, &packfd,
305 	    "objects/pack/fetching.pack");
306 	if (err)
307 		return err;
308 	npackfd = dup(packfd);
309 	if (npackfd == -1)
310 		return got_error_from_errno("dup");
311 	err = got_opentemp_named_fd(&tmpidxpath, &idxfd,
312 	    "objects/pack/fetching.idx");
313 	if (err)
314 		return err;
315 	nidxfd = dup(idxfd);
316 	if (nidxfd == -1)
317 		return got_error_from_errno("dup");
318 
319 	if (strcmp(proto, "ssh") == 0 || strcmp(proto, "git+ssh") == 0)
320 		err = dial_ssh(&fetchfd, host, port, path, "upload");
321 	else if (strcmp(proto, "git") == 0)
322 		err = dial_git(&fetchfd, host, port, path, "upload");
323 	else if (strcmp(proto, "http") == 0 || strcmp(proto, "git+http") == 0)
324 		err = got_error(GOT_ERR_BAD_PROTO);
325 	else
326 		err = got_error(GOT_ERR_BAD_PROTO);
327 
328 	if (fetchfd == -1)
329 		err = got_error_from_errno("dial uri");
330 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_fetchfds) == -1)
331 		return got_error_from_errno("socketpair");
332 
333 	pid = fork();
334 	if (pid == -1)
335 		return got_error_from_errno("fork");
336 	else if (pid == 0){
337 		got_privsep_exec_child(imsg_fetchfds, GOT_PATH_PROG_FETCH_PACK, ".");
338 	}
339 
340 	if (close(imsg_fetchfds[1]) != 0)
341 		return got_error_from_errno("close");
342 	imsg_init(&ibuf, imsg_fetchfds[0]);
343 	err = got_privsep_send_fetch_req(&ibuf, fetchfd);
344 	if (err != NULL)
345 		return err;
346 	err = got_privsep_send_tmpfd(&ibuf, npackfd);
347 	if (err != NULL)
348 		return err;
349 	npackfd = dup(packfd);
350 	if (npackfd == -1)
351 		return got_error_from_errno("dup");
352 	while (!done) {
353 		struct got_object_id *id;
354 		char *refname;
355 		err = got_privsep_recv_fetch_progress(&done,
356 		    &id, &refname, &ibuf);
357 		if (err != NULL)
358 			return err;
359 		if (done) {
360 			packhash = got_object_id_dup(id);
361 			if (packhash == NULL)
362 				return got_error_from_errno(
363 				    "got_object_id_dup");
364 		} else {
365 			char *id_str;
366 			/* TODO Use a progress callback */
367 			err = got_object_id_str(&id_str, id);
368 			if (err)
369 				return err;
370 			printf( "%.12s %s\n", id_str, refname);
371 		}
372 	}
373 	if (waitpid(pid, &status, 0) == -1)
374 		return got_error_from_errno("child exit");
375 
376 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_idxfds) == -1)
377 		return got_error_from_errno("socketpair");
378 	pid = fork();
379 	if (pid == -1)
380 		return got_error_from_errno("fork");
381 	else if (pid == 0)
382 		got_privsep_exec_child(imsg_idxfds, GOT_PATH_PROG_INDEX_PACK, ".");
383 	if (close(imsg_idxfds[1]) != 0)
384 		return got_error_from_errno("close");
385 	imsg_init(&ibuf, imsg_idxfds[0]);
386 
387 	err = got_privsep_send_index_pack_req(&ibuf, npackfd, packhash);
388 	if (err != NULL)
389 		return err;
390 	err = got_privsep_send_tmpfd(&ibuf, nidxfd);
391 	if (err != NULL)
392 		return err;
393 	err = got_privsep_wait_index_pack_done(&ibuf);
394 	if (err != NULL)
395 		return err;
396 	imsg_clear(&ibuf);
397 	if (close(imsg_idxfds[0]) == -1)
398 		return got_error_from_errno("close child");
399 	if (waitpid(pid, &status, 0) == -1)
400 		return got_error_from_errno("child exit");
401 
402 	err = got_object_id_str(&id_str, packhash);
403 	if (err)
404 		return err;
405 	if (asprintf(&packpath, "objects/pack/pack-%s.pack", id_str) == -1)
406 		return got_error_from_errno("asprintf");
407 
408 	if (asprintf(&idxpath, "objects/pack/pack-%s.idx", id_str) == -1)
409 		return got_error_from_errno("asprintf");
410 
411 	if (rename(tmppackpath, packpath) == -1)
412 		return got_error_from_errno3("rename", tmppackpath, packpath);
413 	if (rename(tmpidxpath, idxpath) == -1)
414 		return got_error_from_errno3("rename", tmpidxpath, idxpath);
415 
416 	free(tmppackpath);
417 	free(tmpidxpath);
418 	free(idxpath);
419 	free(packpath);
420 	free(default_destdir);
421 	free(packhash);
422 
423 	return NULL;
424 }