1 /*
2  * Copyright (c) 2020, 2021, 2022 Omar Polo <op@omarpolo.com>
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 #ifndef GMID_H
18 #define GMID_H
19 
20 #include "config.h"
21 
22 #include <sys/socket.h>
23 #include <sys/types.h>
24 
25 #include <arpa/inet.h>
26 #include <netinet/in.h>
27 
28 #include <dirent.h>
29 #include <limits.h>
30 #include <netdb.h>
31 #include <signal.h>
32 #include <stdio.h>
33 #include <stdlib.h>
34 #include <time.h>
35 #include <tls.h>
36 #include <unistd.h>
37 
38 #include <openssl/x509.h>
39 
40 #if HAVE_EVENT2
41 # include <event2/event.h>
42 # include <event2/event_compat.h>
43 # include <event2/event_struct.h>
44 # include <event2/buffer.h>
45 # include <event2/buffer_compat.h>
46 # include <event2/bufferevent.h>
47 # include <event2/bufferevent_struct.h>
48 # include <event2/bufferevent_compat.h>
49 #else
50 # include <event.h>
51 #endif
52 
53 #define GMID_STRING	"gmid " VERSION
54 #define GMID_VERSION	"gmid/" VERSION
55 
56 #define GEMINI_URL_LEN (1024+3)	/* URL max len + \r\n + \0 */
57 
58 #define SUCCESS		20
59 #define TEMP_REDIRECT	30
60 #define TEMP_FAILURE	40
61 #define CGI_ERROR	42
62 #define PROXY_ERROR	43
63 #define NOT_FOUND	51
64 #define PROXY_REFUSED	53
65 #define BAD_REQUEST	59
66 #define CLIENT_CERT_REQ	60
67 #define CERT_NOT_AUTH	61
68 
69 /* maximum hostname and label length, +1 for the NUL-terminator */
70 #define DOMAIN_NAME_LEN	(253+1)
71 #define LABEL_LEN	(63+1)
72 
73 #define FCGI_MAX	32
74 #define PROC_MAX	16
75 
76 struct iri {
77 	char		*schema;
78 	char		*host;
79 	char		*port;
80 	uint16_t	 port_no;
81 	char		*path;
82 	char		*query;
83 	char		*fragment;
84 };
85 
86 struct parser {
87 	char		*iri;
88 	struct iri	*parsed;
89 	const char	*err;
90 };
91 
92 struct fcgi {
93 	int		 id;
94 	char		*path;
95 	char		*port;
96 	char		*prog;
97 };
98 extern struct fcgi fcgi[FCGI_MAX];
99 
100 TAILQ_HEAD(proxyhead, proxy);
101 struct proxy {
102 	char		*match_proto;
103 	char		*match_host;
104 	const char	*match_port;
105 
106 	char		*host;
107 	const char	*port;
108 	int		 notls;
109 	uint32_t	 protocols;
110 	int		 noverifyname;
111 	uint8_t		*cert;
112 	size_t		 certlen;
113 	uint8_t		*key;
114 	size_t		 keylen;
115 
116 	TAILQ_ENTRY(proxy) proxies;
117 };
118 
119 TAILQ_HEAD(lochead, location);
120 struct location {
121 	const char	*match;
122 	const char	*lang;
123 	const char	*default_mime;
124 	const char	*index;
125 	int		 auto_index; /* 0 auto, -1 off, 1 on */
126 	int		 block_code;
127 	const char	*block_fmt;
128 	int		 strip;
129 	X509_STORE	*reqca;
130 	int		 disable_log;
131 	int		 fcgi;
132 
133 	const char	*dir;
134 	int		 dirfd;
135 
136 	TAILQ_ENTRY(location) locations;
137 };
138 
139 TAILQ_HEAD(envhead, envlist);
140 struct envlist {
141 	char		*name;
142 	char		*value;
143 	TAILQ_ENTRY(envlist) envs;
144 };
145 
146 TAILQ_HEAD(aliashead, alist);
147 struct alist {
148 	char		*alias;
149 	TAILQ_ENTRY(alist) aliases;
150 };
151 
152 extern TAILQ_HEAD(vhosthead, vhost) hosts;
153 struct vhost {
154 	const char	*domain;
155 	const char	*cert;
156 	const char	*key;
157 	const char	*ocsp;
158 	const char	*cgi;
159 	const char	*entrypoint;
160 
161 	TAILQ_ENTRY(vhost) vhosts;
162 
163 	/*
164 	 * the first location rule is always '*' and holds the default
165 	 * settings for the vhost, then follows the "real" location
166 	 * rules as specified in the configuration.
167 	 */
168 	struct lochead	 locations;
169 
170 	struct envhead	 env;
171 	struct envhead	 params;
172 	struct aliashead aliases;
173 	struct proxyhead proxies;
174 };
175 
176 struct etm {			/* extension to mime */
177 	const char	*mime;
178 	const char	*ext;
179 };
180 
181 struct mime {
182 	struct etm	*t;
183 	size_t		len;
184 	size_t		cap;
185 };
186 
187 struct conf {
188 	/* from command line */
189 	int		 foreground;
190 	int		 verbose;
191 
192 	/* in the config */
193 	int		 port;
194 	int		 ipv6;
195 	uint32_t	 protos;
196 	struct mime	 mime;
197 	char		*chroot;
198 	char		*user;
199 	int		 prefork;
200 };
201 
202 extern const char *config_path;
203 extern struct conf conf;
204 
205 extern struct imsgbuf logibuf, exibuf, servibuf[PROC_MAX];
206 
207 extern int servpipes[PROC_MAX];
208 
209 typedef void (imsg_handlerfn)(struct imsgbuf*, struct imsg*, size_t);
210 
211 enum {
212 	REQUEST_UNDECIDED,
213 	REQUEST_FILE,
214 	REQUEST_DIR,
215 	REQUEST_CGI,
216 	REQUEST_FCGI,
217 	REQUEST_PROXY,
218 	REQUEST_DONE,
219 };
220 
221 #define IS_INTERNAL_REQUEST(x) \
222 	((x) != REQUEST_CGI && \
223 	 (x) != REQUEST_FCGI && \
224 	 (x) != REQUEST_PROXY)
225 
226 struct client {
227 	uint32_t	 id;
228 	struct tls	*ctx;
229 	char		*req;
230 	struct iri	 iri;
231 	char		 domain[DOMAIN_NAME_LEN];
232 
233 	struct bufferevent *bev;
234 
235 	int		 type;
236 
237 	struct bufferevent *cgibev;
238 
239 	struct proxy	*proxy;
240 	struct bufferevent *proxybev;
241 	struct tls	*proxyctx;
242 	struct event	 proxyev;
243 
244 	char		*header;
245 
246 	int		 code;
247 	const char	*meta;
248 	int		 fd, pfd;
249 	struct dirent	**dir;
250 	int		 dirlen, diroff;
251 
252 	/* big enough to store STATUS + SPACE + META + CRLF */
253 	char		 sbuf[1029];
254 	ssize_t		 len, off;
255 
256 	struct sockaddr_storage	 addr;
257 	struct vhost	*host;	/* host they're talking to */
258 	size_t		 loc;	/* location matched */
259 
260 	SPLAY_ENTRY(client) entry;
261 };
262 SPLAY_HEAD(client_tree_id, client);
263 extern struct client_tree_id clients;
264 
265 struct cgireq {
266 	char		buf[GEMINI_URL_LEN];
267 
268 	size_t		iri_schema_off;
269 	size_t		iri_host_off;
270 	size_t		iri_port_off;
271 	size_t		iri_path_off;
272 	size_t		iri_query_off;
273 	size_t		iri_fragment_off;
274 	int		iri_portno;
275 
276 	char		spath[PATH_MAX+1];
277 	char		relpath[PATH_MAX+1];
278 	char		addr[NI_MAXHOST+1];
279 
280 	/* AFAIK there isn't an upper limit for these two fields. */
281 	char		subject[64+1];
282 	char		issuer[64+1];
283 
284 	char		hash[128+1];
285 	char		version[8];
286 	char		cipher[32];
287 	int		cipher_strength;
288 	time_t		notbefore;
289 	time_t		notafter;
290 
291 	size_t		host_off;
292 	size_t		loc_off;
293 };
294 
295 struct connreq {
296 	char	host[NI_MAXHOST];
297 	char	port[NI_MAXSERV];
298 	int	flag;
299 };
300 
301 enum {
302 	FILE_EXISTS,
303 	FILE_EXECUTABLE,
304 	FILE_DIRECTORY,
305 	FILE_MISSING,
306 };
307 
308 enum imsg_type {
309 	IMSG_CGI_REQ,
310 	IMSG_CGI_RES,
311 	IMSG_FCGI_REQ,
312 	IMSG_FCGI_FD,
313 	IMSG_CONN_REQ,
314 	IMSG_CONN_FD,
315 	IMSG_LOG,
316 	IMSG_LOG_REQUEST,
317 	IMSG_LOG_TYPE,
318 	IMSG_QUIT,
319 };
320 
321 /* gmid.c */
322 char		*data_dir(void);
323 void		 load_local_cert(const char*, const char*);
324 void		 load_vhosts(void);
325 int		 make_socket(int, int);
326 void		 setup_tls(void);
327 void		 init_config(void);
328 void		 free_config(void);
329 void		 drop_priv(void);
330 
331 void		 yyerror(const char*, ...);
332 void		 parse_conf(const char*);
333 void		 print_conf(void);
334 int		 cmdline_symset(char *);
335 
336 /* log.c */
337 void		 fatal(const char*, ...)
338 	__attribute__((format (printf, 1, 2)))
339 	__attribute__((__noreturn__));
340 
341 #define LOG_ATTR_FMT __attribute__((format (printf, 2, 3)))
342 void		 log_err(struct client*, const char*, ...)	LOG_ATTR_FMT;
343 void		 log_warn(struct client*, const char*, ...)	LOG_ATTR_FMT;
344 void		 log_notice(struct client*, const char*, ...)	LOG_ATTR_FMT;
345 void		 log_info(struct client*, const char*, ...)	LOG_ATTR_FMT;
346 void		 log_debug(struct client*, const char*, ...)	LOG_ATTR_FMT;
347 void		 log_request(struct client*, char*, size_t);
348 int		 logger_main(int, struct imsgbuf*);
349 
350 /* mime.c */
351 void		 init_mime(struct mime*);
352 void		 add_mime(struct mime*, const char*, const char*);
353 void		 load_default_mime(struct mime*);
354 const char	*mime(struct vhost*, const char*);
355 
356 /* server.c */
357 extern int	shutting_down;
358 const char	*vhost_lang(struct vhost*, const char*);
359 const char	*vhost_default_mime(struct vhost*, const char*);
360 const char	*vhost_index(struct vhost*, const char*);
361 int		 vhost_auto_index(struct vhost*, const char*);
362 int		 vhost_block_return(struct vhost*, const char*, int*, const char**);
363 int		 vhost_fastcgi(struct vhost*, const char*);
364 int		 vhost_dirfd(struct vhost*, const char*, size_t*);
365 int		 vhost_strip(struct vhost*, const char*);
366 X509_STORE	*vhost_require_ca(struct vhost*, const char*);
367 int		 vhost_disable_log(struct vhost*, const char*);
368 
369 void		 mark_nonblock(int);
370 void		 client_write(struct bufferevent *, void *);
371 void		 start_reply(struct client*, int, const char*);
372 void		 client_close(struct client *);
373 struct client	*try_client_by_id(int);
374 void		 loop(struct tls*, int, int, struct imsgbuf*);
375 
376 int		 client_tree_cmp(struct client *, struct client *);
377 SPLAY_PROTOTYPE(client_tree_id, client, entry, client_tree_cmp);
378 
379 /* dirs.c */
380 int		 scandir_fd(int, struct dirent***, int(*)(const struct dirent*),
381 		    int(*)(const struct dirent**, const struct dirent**));
382 int		 select_non_dot(const struct dirent*);
383 int		 select_non_dotdot(const struct dirent*);
384 
385 /* ex.c */
386 int		 send_string(int, const char*);
387 int		 recv_string(int, char**);
388 int		 send_iri(int, struct iri*);
389 int		 recv_iri(int, struct iri*);
390 void		 free_recvd_iri(struct iri*);
391 int		 send_vhost(int, struct vhost*);
392 int		 recv_vhost(int, struct vhost**);
393 int		 send_time(int, time_t);
394 int		 recv_time(int, time_t*);
395 int		 send_fd(int, int);
396 int		 recv_fd(int);
397 int		 executor_main(struct imsgbuf*);
398 
399 /* fcgi.c */
400 void		 fcgi_read(struct bufferevent *, void *);
401 void		 fcgi_write(struct bufferevent *, void *);
402 void		 fcgi_error(struct bufferevent *, short, void *);
403 void		 fcgi_req(struct client *);
404 
405 /* sandbox.c */
406 void		 sandbox_server_process(void);
407 void		 sandbox_executor_process(void);
408 void		 sandbox_logger_process(void);
409 
410 /* utf8.c */
411 int		 valid_multibyte_utf8(struct parser*);
412 char		*utf8_nth(char*, size_t);
413 
414 /* iri.c */
415 int		 parse_iri(char*, struct iri*, const char**);
416 int		 serialize_iri(struct iri*, char*, size_t);
417 char		*pct_decode_str(char *);
418 
419 /* proxy.c */
420 int		 proxy_init(struct client *);
421 
422 /* puny.c */
423 int		 puny_decode(const char*, char*, size_t, const char**);
424 
425 /* utils.c */
426 void		 block_signals(void);
427 void		 unblock_signals(void);
428 int		 starts_with(const char*, const char*);
429 int		 ends_with(const char*, const char*);
430 ssize_t		 filesize(int);
431 char		*absolutify_path(const char*);
432 char		*xstrdup(const char*);
433 void		*xcalloc(size_t, size_t);
434 void		 gen_certificate(const char*, const char*, const char*);
435 X509_STORE	*load_ca(const char*);
436 int		 validate_against_ca(X509_STORE*, const uint8_t*, size_t);
437 void		 dispatch_imsg(struct imsgbuf*, imsg_handlerfn**, size_t);
438 
439 #endif