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 <fcntl.h>
29 #include <stdio.h>
30 #include <stdlib.h>
31 #include <string.h>
32 #include <stdint.h>
33 #include <sha1.h>
34 #include <zlib.h>
35 #include <ctype.h>
36 #include <limits.h>
37 #include <imsg.h>
38 #include <time.h>
39 #include <uuid.h>
40 #include <netdb.h>
41 #include <netinet/in.h>
42 
43 #include "got_error.h"
44 #include "got_reference.h"
45 #include "got_repository.h"
46 #include "got_path.h"
47 #include "got_cancel.h"
48 #include "got_worktree.h"
49 #include "got_object.h"
50 #include "got_opentemp.h"
51 
52 #include "got_lib_delta.h"
53 #include "got_lib_inflate.h"
54 #include "got_lib_object.h"
55 #include "got_lib_object_parse.h"
56 #include "got_lib_object_create.h"
57 #include "got_lib_pack.h"
58 #include "got_lib_sha1.h"
59 #include "got_lib_privsep.h"
60 #include "got_lib_object_cache.h"
61 #include "got_lib_repository.h"
62 
63 #define GOT_PROTOMAX	64
64 #define GOT_HOSTMAX	256
65 #define GOT_PATHMAX	512
66 #define GOT_REPOMAX	256
67 #define GOT_PORTMAX	16
68 #define GOT_URIMAX	1024
69 
70 static int
71 mkpath(char *path)
72 {
73 	char *p, namebuf[PATH_MAX];
74 	struct stat sb;
75 	int done;
76 
77 	while (*path == '/')
78 		path++;
79 	if (strlcpy(namebuf, path, sizeof(namebuf)) >= sizeof(namebuf)) {
80 		errno = ENAMETOOLONG;
81 		return -1;
82 	}
83 
84 	p = namebuf;
85 	for (;;) {
86 		p += strspn(p, "/");
87 		p += strcspn(p, "/");
88 		done = (*p == '\0');
89 		*p = '\0';
90 
91 		if (mkdir(namebuf, 0755) != 0) {
92 			int mkdir_errno = errno;
93 			if (stat(path, &sb) == -1) {
94 				/* Not there; use mkdir()s errno */
95 				errno = mkdir_errno;
96 				return -1;
97 			}
98 			if (!S_ISDIR(sb.st_mode)) {
99 				/* Is there, but isn't a directory */
100 				errno = ENOTDIR;
101 				return -1;
102 			}
103 		}
104 
105 		if (done)
106 			break;
107 		*p = '/';
108 	}
109 
110 	return 0;
111 }
112 
113 static int
114 hassuffix(char *base, char *suf)
115 {
116 	int nb, ns;
117 
118 	nb = strlen(base);
119 	ns = strlen(suf);
120 	if (ns <= nb && strcmp(base + (nb - ns), suf) == 0)
121 		return 1;
122 	return 0;
123 }
124 
125 static int
126 grab(char *dst, int n, char *p, char *e)
127 {
128 	int l;
129 
130 	l = e - p;
131 	if (l >= n) {
132 		errno = ENAMETOOLONG;
133 		return -1;
134 	}
135 	return strlcpy(dst, p, l + 1);
136 }
137 
138 static int
139 got_dial_ssh(char *host, char *port, char *path, char *direction)
140 {
141 	int pid, pfd[2];
142 	char cmd[64];
143 
144 	if (pipe(pfd) == -1)
145 		return -1;
146 	pid = fork();
147 	if (pid == -1)
148 		return -1;
149 	if (pid == 0) {
150 		close(pfd[1]);
151 		dup2(pfd[0], 0);
152 		dup2(pfd[0], 1);
153 		snprintf(cmd, sizeof(cmd), "git-%s-pack", direction);
154 		execlp("ssh", "ssh", host, cmd, path, NULL);
155 		abort();
156 	}else{
157 		close(pfd[0]);
158 		return pfd[1];
159 	}
160 }
161 
162 static int
163 got_dial_git(char *host, char *port, char *path, char *direction)
164 {
165 	struct addrinfo hints, *servinfo, *p;
166 	char *cmd, *pkt;
167 	int fd, l, r;
168 
169 	memset(&hints, 0, sizeof hints);
170 	hints.ai_family = AF_UNSPEC;
171 	hints.ai_socktype = SOCK_STREAM;
172 	if (getaddrinfo(host, port, &hints, &servinfo) != 0)
173 		return -1;
174 
175 	for (p = servinfo; p != NULL; p = p->ai_next) {
176 		if ((fd = socket(p->ai_family, p->ai_socktype,
177 		    p->ai_protocol)) == -1)
178 			continue;
179 		if (connect(fd, p->ai_addr, p->ai_addrlen) == 0)
180 			break;
181 		close(fd);
182 	}
183 	if (p == NULL)
184 		return -1;
185 
186 	if ((l = asprintf(&cmd, "git-%s-pack %s\n", direction, path)) == -1)
187 		return -1;
188 	if ((l = asprintf(&pkt, "%04x%s", l+4, cmd)) == -1)
189 		return -1;
190 	r = write(fd, pkt, l);
191 	free(cmd);
192 	free(pkt);
193 	if (r == -1) {
194 		close(fd);
195 		return -1;
196 	}
197 	return fd;
198 }
199 
200 int
201 got_parse_uri(char *uri, char *proto, char *host, char *port, char *path, char *repo)
202 {
203 	char *s, *p, *q;
204 	int n, hasport;
205 
206 	p = strstr(uri, "://");
207 	if (!p) {
208 		//werrstr("missing protocol");
209 		return -1;
210 	}
211 	if (grab(proto, GOT_PROTOMAX, uri, p) == -1)
212 		return -1;
213 	hasport = (strcmp(proto, "git") == 0 || strstr(proto, "http") == proto);
214 	s = p + 3;
215 	p = NULL;
216 	if (!hasport) {
217 		p = strstr(s, ":");
218 		if (p != NULL)
219 			p++;
220 	}
221 	if (p == NULL)
222 		p = strstr(s, "/");
223 	if (p == NULL || strlen(p) == 1) {
224 		//werrstr("missing path");
225 		return -1;
226 	}
227 
228 	q = memchr(s, ':', p - s);
229 	if (q) {
230 		grab(host, GOT_HOSTMAX, s, q);
231 		grab(port, GOT_PORTMAX, q + 1, p);
232 	}else{
233 		grab(host, GOT_HOSTMAX, s, p);
234 		snprintf(port, GOT_PORTMAX, "9418");
235 	}
236 
237 	snprintf(path, GOT_PATHMAX, "%s", p);
238 	p = strrchr(p, '/') + 1;
239 	if (!p || strlen(p) == 0) {
240 		//werrstr("missing repository in uri");
241 		return -1;
242 	}
243 	n = strlen(p);
244 	if (hassuffix(p, ".git"))
245 		n -= 4;
246 	grab(repo, GOT_REPOMAX, p, p + n);
247 	return 0;
248 }
249 
250 const struct got_error*
251 got_fetch(char *uri, char *branch_filter, char *destdir)
252 {
253 	char proto[GOT_PROTOMAX], host[GOT_HOSTMAX], port[GOT_PORTMAX];
254 	char repo[GOT_REPOMAX], path[GOT_PATHMAX];
255 	int imsg_fetchfds[2], imsg_idxfds[2], fetchfd;
256 	int packfd = -1, npackfd, idxfd = -1, nidxfd, status;
257 	struct got_object_id packhash;
258 	const struct got_error *err;
259 	struct imsgbuf ibuf;
260 	pid_t pid;
261 	char *tmppackpath = NULL, *tmpidxpath = NULL, *default_destdir = NULL;
262 	char *packpath = NULL, *idxpath = NULL, *id_str = NULL;
263 
264 	fetchfd = -1;
265 	if (got_parse_uri(uri, proto, host, port, path, repo) == -1)
266 		return got_error(GOT_ERR_PARSE_URI);
267 	if (destdir == NULL) {
268 		if (asprintf(&default_destdir, "%s.git", repo) == -1)
269 			return got_error_from_errno("asprintf");
270 	}
271 	err = got_repo_init(destdir ? destdir : default_destdir);
272 	if (err != NULL)
273 		return err;
274 	if (chdir(destdir ? destdir : default_destdir))
275 		return got_error_from_errno("enter new repo");
276 	if (mkpath("objects/pack") == -1)
277 		return got_error_from_errno("mkpath");
278 	err = got_opentemp_named_fd(&tmppackpath, &packfd,
279 	    "objects/pack/fetching.pack");
280 	if (err)
281 		return err;
282 	npackfd = dup(packfd);
283 	if (npackfd == -1)
284 		return got_error_from_errno("dup");
285 	err = got_opentemp_named_fd(&tmpidxpath, &idxfd,
286 	    "objects/pack/fetching.idx");
287 	if (err)
288 		return err;
289 	nidxfd = dup(idxfd);
290 	if (nidxfd == -1)
291 		return got_error_from_errno("dup");
292 
293 	if (strcmp(proto, "ssh") == 0 || strcmp(proto, "git+ssh") == 0)
294 		fetchfd = got_dial_ssh(host, port, path, "upload");
295 	else if (strcmp(proto, "git") == 0)
296 		fetchfd = got_dial_git(host, port, path, "upload");
297 	else if (strcmp(proto, "http") == 0 || strcmp(proto, "git+http") == 0)
298 		err = got_error(GOT_ERR_BAD_PROTO);
299 	else
300 		err = got_error(GOT_ERR_BAD_PROTO);
301 
302 	if (fetchfd == -1)
303 		err = got_error_from_errno("dial uri");
304 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_fetchfds) == -1)
305 		return got_error_from_errno("socketpair");
306 
307 	pid = fork();
308 	if (pid == -1)
309 		return got_error_from_errno("fork");
310 	else if (pid == 0){
311 		got_privsep_exec_child(imsg_fetchfds, GOT_PATH_PROG_FETCH_PACK, ".");
312 	}
313 
314 	if (close(imsg_fetchfds[1]) != 0)
315 		return got_error_from_errno("close");
316 	imsg_init(&ibuf, imsg_fetchfds[0]);
317 	err = got_privsep_send_fetch_req(&ibuf, fetchfd);
318 	if (err != NULL)
319 		return err;
320 	err = got_privsep_send_tmpfd(&ibuf, npackfd);
321 	if (err != NULL)
322 		return err;
323 	npackfd = dup(packfd);
324 	if (npackfd == -1)
325 		return got_error_from_errno("dup");
326 	err = got_privsep_wait_fetch_done(&ibuf, &packhash);
327 	if (err != NULL)
328 		return err;
329 	if (waitpid(pid, &status, 0) == -1)
330 		return got_error_from_errno("child exit");
331 
332 	if (socketpair(AF_UNIX, SOCK_STREAM, PF_UNSPEC, imsg_idxfds) == -1)
333 		return got_error_from_errno("socketpair");
334 	pid = fork();
335 	if (pid == -1)
336 		return got_error_from_errno("fork");
337 	else if (pid == 0)
338 		got_privsep_exec_child(imsg_idxfds, GOT_PATH_PROG_INDEX_PACK, ".");
339 	if (close(imsg_idxfds[1]) != 0)
340 		return got_error_from_errno("close");
341 	imsg_init(&ibuf, imsg_idxfds[0]);
342 
343 	err = got_privsep_send_index_pack_req(&ibuf, npackfd, &packhash);
344 	if (err != NULL)
345 		return err;
346 	err = got_privsep_send_tmpfd(&ibuf, nidxfd);
347 	if (err != NULL)
348 		return err;
349 	err = got_privsep_wait_index_pack_done(&ibuf);
350 	if (err != NULL)
351 		return err;
352 	imsg_clear(&ibuf);
353 	if (close(imsg_idxfds[0]) == -1)
354 		return got_error_from_errno("close child");
355 	if (waitpid(pid, &status, 0) == -1)
356 		return got_error_from_errno("child exit");
357 
358 	err = got_object_id_str(&id_str, &packhash);
359 	if (err)
360 		return err;
361 	if (asprintf(&packpath, "objects/pack/pack-%s.pack", id_str) == -1)
362 		return got_error_from_errno("asprintf");
363 
364 	if (asprintf(&idxpath, "objects/pack/pack-%s.idx", id_str) == -1)
365 		return got_error_from_errno("asprintf");
366 
367 	if (rename(tmppackpath, packpath) == -1)
368 		return got_error_from_errno3("rename", tmppackpath, packpath);
369 	if (rename(tmpidxpath, idxpath) == -1)
370 		return got_error_from_errno3("rename", tmpidxpath, idxpath);
371 
372 	free(tmppackpath);
373 	free(tmpidxpath);
374 	free(idxpath);
375 	free(packpath);
376 	free(default_destdir);
377 
378 	return NULL;
379 }