1 /*-
2  * Copyright (c) 1982, 1986, 1990, 1993
3  *	The Regents of the University of California.  All rights reserved.
4  *
5  * Redistribution and use in source and binary forms, with or without
6  * modification, are permitted provided that the following conditions
7  * are met:
8  * 1. Redistributions of source code must retain the above copyright
9  *    notice, this list of conditions and the following disclaimer.
10  * 2. Redistributions in binary form must reproduce the above copyright
11  *    notice, this list of conditions and the following disclaimer in the
12  *    documentation and/or other materials provided with the distribution.
13  * 4. Neither the name of the University nor the names of its contributors
14  *    may be used to endorse or promote products derived from this software
15  *    without specific prior written permission.
16  *
17  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
18  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
19  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
20  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
21  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
22  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
23  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
24  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
25  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
26  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
27  * SUCH DAMAGE.
28  *
29  *	@(#)socketvar.h	8.3 (Berkeley) 2/19/95
30  *
31  * $FreeBSD: stable/10/sys/sys/socketvar.h 321021 2017-07-15 17:28:03Z dchagin $
32  */
33 
34 #ifndef _SYS_SOCKETVAR_H_
35 #define _SYS_SOCKETVAR_H_
36 
37 #include <sys/queue.h>			/* for TAILQ macros */
38 #include <sys/selinfo.h>		/* for struct selinfo */
39 #include <sys/_lock.h>
40 #include <sys/_mutex.h>
41 #include <sys/_sx.h>
42 #include <sys/sockbuf.h>
43 #include <sys/sockstate.h>
44 #ifdef _KERNEL
45 #include <sys/caprights.h>
46 #include <sys/sockopt.h>
47 #endif
48 
49 struct vnet;
50 
51 /*
52  * Kernel structure per socket.
53  * Contains send and receive buffer queues,
54  * handle on protocol and pointer to protocol
55  * private data and error information.
56  */
57 typedef	u_quad_t so_gen_t;
58 
59 struct socket;
60 
61 /*-
62  * Locking key to struct socket:
63  * (a) constant after allocation, no locking required.
64  * (b) locked by SOCK_LOCK(so).
65  * (c) locked by SOCKBUF_LOCK(&so->so_rcv).
66  * (d) locked by SOCKBUF_LOCK(&so->so_snd).
67  * (e) locked by ACCEPT_LOCK().
68  * (f) not locked since integer reads/writes are atomic.
69  * (g) used only as a sleep/wakeup address, no value.
70  * (h) locked by global mutex so_global_mtx.
71  */
72 struct socket {
73 	int	so_count;		/* (b) reference count */
74 	short	so_type;		/* (a) generic type, see socket.h */
75 	short	so_options;		/* from socket call, see socket.h */
76 	short	so_linger;		/* time to linger while closing */
77 	short	so_state;		/* (b) internal state flags SS_* */
78 	int	so_qstate;		/* (e) internal state flags SQ_* */
79 	void	*so_pcb;		/* protocol control block */
80 	struct	vnet *so_vnet;		/* (a) network stack instance */
81 	struct	protosw *so_proto;	/* (a) protocol handle */
82 /*
83  * Variables for connection queuing.
84  * Socket where accepts occur is so_head in all subsidiary sockets.
85  * If so_head is 0, socket is not related to an accept.
86  * For head socket so_incomp queues partially completed connections,
87  * while so_comp is a queue of connections ready to be accepted.
88  * If a connection is aborted and it has so_head set, then
89  * it has to be pulled out of either so_incomp or so_comp.
90  * We allow connections to queue up based on current queue lengths
91  * and limit on number of queued connections for this socket.
92  */
93 	struct	socket *so_head;	/* (e) back pointer to listen socket */
94 	TAILQ_HEAD(, socket) so_incomp;	/* (e) queue of partial unaccepted connections */
95 	TAILQ_HEAD(, socket) so_comp;	/* (e) queue of complete unaccepted connections */
96 	TAILQ_ENTRY(socket) so_list;	/* (e) list of unaccepted connections */
97 	u_short	so_qlen;		/* (e) number of unaccepted connections */
98 	u_short	so_incqlen;		/* (e) number of unaccepted incomplete
99 					   connections */
100 	u_short	so_qlimit;		/* (e) max number queued connections */
101 	short	so_timeo;		/* (g) connection timeout */
102 	u_short	so_error;		/* (f) error affecting connection */
103 	struct	sigio *so_sigio;	/* [sg] information for async I/O or
104 					   out of band data (SIGURG) */
105 	u_long	so_oobmark;		/* (c) chars to oob mark */
106 	TAILQ_HEAD(, aiocblist) so_aiojobq; /* AIO ops waiting on socket */
107 
108 	struct sockbuf so_rcv, so_snd;
109 
110 	struct	ucred *so_cred;		/* (a) user credentials */
111 	struct	label *so_label;	/* (b) MAC label for socket */
112 	struct	label *so_peerlabel;	/* (b) cached MAC label for peer */
113 	/* NB: generation count must not be first. */
114 	so_gen_t so_gencnt;		/* (h) generation count */
115 	void	*so_emuldata;		/* (b) private data for emulators */
116  	struct so_accf {
117 		struct	accept_filter *so_accept_filter;
118 		void	*so_accept_filter_arg;	/* saved filter args */
119 		char	*so_accept_filter_str;	/* saved user args */
120 	} *so_accf;
121 	/*
122 	 * so_fibnum, so_user_cookie and friends can be used to attach
123 	 * some user-specified metadata to a socket, which then can be
124 	 * used by the kernel for various actions.
125 	 * so_user_cookie is used by ipfw/dummynet.
126 	 */
127 	int so_fibnum;		/* routing domain for this socket */
128 	uint32_t so_user_cookie;
129 };
130 
131 /*
132  * Global accept mutex to serialize access to accept queues and
133  * fields associated with multiple sockets.  This allows us to
134  * avoid defining a lock order between listen and accept sockets
135  * until such time as it proves to be a good idea.
136  */
137 extern struct mtx accept_mtx;
138 #define	ACCEPT_LOCK_ASSERT()		mtx_assert(&accept_mtx, MA_OWNED)
139 #define	ACCEPT_UNLOCK_ASSERT()		mtx_assert(&accept_mtx, MA_NOTOWNED)
140 #define	ACCEPT_LOCK()			mtx_lock(&accept_mtx)
141 #define	ACCEPT_UNLOCK()			mtx_unlock(&accept_mtx)
142 
143 /*
144  * Per-socket mutex: we reuse the receive socket buffer mutex for space
145  * efficiency.  This decision should probably be revisited as we optimize
146  * locking for the socket code.
147  */
148 #define	SOCK_MTX(_so)			SOCKBUF_MTX(&(_so)->so_rcv)
149 #define	SOCK_LOCK(_so)			SOCKBUF_LOCK(&(_so)->so_rcv)
150 #define	SOCK_OWNED(_so)			SOCKBUF_OWNED(&(_so)->so_rcv)
151 #define	SOCK_UNLOCK(_so)		SOCKBUF_UNLOCK(&(_so)->so_rcv)
152 #define	SOCK_LOCK_ASSERT(_so)		SOCKBUF_LOCK_ASSERT(&(_so)->so_rcv)
153 
154 /*
155  * Socket state bits stored in so_qstate.
156  */
157 #define	SQ_INCOMP		0x0800	/* unaccepted, incomplete connection */
158 #define	SQ_COMP			0x1000	/* unaccepted, complete connection */
159 
160 /*
161  * Externalized form of struct socket used by the sysctl(3) interface.
162  */
163 struct xsocket {
164 	size_t	xso_len;	/* length of this structure */
165 	struct	socket *xso_so;	/* makes a convenient handle sometimes */
166 	short	so_type;
167 	short	so_options;
168 	short	so_linger;
169 	short	so_state;
170 	caddr_t	so_pcb;		/* another convenient handle */
171 	int	xso_protocol;
172 	int	xso_family;
173 	u_short	so_qlen;
174 	u_short	so_incqlen;
175 	u_short	so_qlimit;
176 	short	so_timeo;
177 	u_short	so_error;
178 	pid_t	so_pgid;
179 	u_long	so_oobmark;
180 	struct xsockbuf so_rcv, so_snd;
181 	uid_t	so_uid;		/* XXX */
182 };
183 
184 #ifdef _KERNEL
185 
186 /*
187  * Macros for sockets and socket buffering.
188  */
189 
190 /*
191  * Flags to sblock().
192  */
193 #define	SBL_WAIT	0x00000001	/* Wait if not immediately available. */
194 #define	SBL_NOINTR	0x00000002	/* Force non-interruptible sleep. */
195 #define	SBL_VALID	(SBL_WAIT | SBL_NOINTR)
196 
197 /*
198  * Do we need to notify the other side when I/O is possible?
199  */
200 #define	sb_notify(sb)	(((sb)->sb_flags & (SB_WAIT | SB_SEL | SB_ASYNC | \
201     SB_UPCALL | SB_AIO | SB_KNOTE)) != 0)
202 
203 /* do we have to send all at once on a socket? */
204 #define	sosendallatonce(so) \
205     ((so)->so_proto->pr_flags & PR_ATOMIC)
206 
207 /* can we read something from so? */
208 #define	soreadabledata(so) \
209     ((so)->so_rcv.sb_cc >= (so)->so_rcv.sb_lowat || \
210 	!TAILQ_EMPTY(&(so)->so_comp) || (so)->so_error)
211 #define	soreadable(so) \
212 	(soreadabledata(so) || ((so)->so_rcv.sb_state & SBS_CANTRCVMORE))
213 
214 /* can we write something to so? */
215 #define	sowriteable(so) \
216     ((sbspace(&(so)->so_snd) >= (so)->so_snd.sb_lowat && \
217 	(((so)->so_state&SS_ISCONNECTED) || \
218 	  ((so)->so_proto->pr_flags&PR_CONNREQUIRED)==0)) || \
219      ((so)->so_snd.sb_state & SBS_CANTSENDMORE) || \
220      (so)->so_error)
221 
222 /*
223  * soref()/sorele() ref-count the socket structure.  Note that you must
224  * still explicitly close the socket, but the last ref count will free
225  * the structure.
226  */
227 #define	soref(so) do {							\
228 	SOCK_LOCK_ASSERT(so);						\
229 	++(so)->so_count;						\
230 } while (0)
231 
232 #define	sorele(so) do {							\
233 	ACCEPT_LOCK_ASSERT();						\
234 	SOCK_LOCK_ASSERT(so);						\
235 	if ((so)->so_count <= 0)					\
236 		panic("sorele");					\
237 	if (--(so)->so_count == 0)					\
238 		sofree(so);						\
239 	else {								\
240 		SOCK_UNLOCK(so);					\
241 		ACCEPT_UNLOCK();					\
242 	}								\
243 } while (0)
244 
245 /*
246  * In sorwakeup() and sowwakeup(), acquire the socket buffer lock to
247  * avoid a non-atomic test-and-wakeup.  However, sowakeup is
248  * responsible for releasing the lock if it is called.  We unlock only
249  * if we don't call into sowakeup.  If any code is introduced that
250  * directly invokes the underlying sowakeup() primitives, it must
251  * maintain the same semantics.
252  */
253 #define	sorwakeup_locked(so) do {					\
254 	SOCKBUF_LOCK_ASSERT(&(so)->so_rcv);				\
255 	if (sb_notify(&(so)->so_rcv))					\
256 		sowakeup((so), &(so)->so_rcv);	 			\
257 	else								\
258 		SOCKBUF_UNLOCK(&(so)->so_rcv);				\
259 } while (0)
260 
261 #define	sorwakeup(so) do {						\
262 	SOCKBUF_LOCK(&(so)->so_rcv);					\
263 	sorwakeup_locked(so);						\
264 } while (0)
265 
266 #define	sowwakeup_locked(so) do {					\
267 	SOCKBUF_LOCK_ASSERT(&(so)->so_snd);				\
268 	if (sb_notify(&(so)->so_snd))					\
269 		sowakeup((so), &(so)->so_snd); 				\
270 	else								\
271 		SOCKBUF_UNLOCK(&(so)->so_snd);				\
272 } while (0)
273 
274 #define	sowwakeup(so) do {						\
275 	SOCKBUF_LOCK(&(so)->so_snd);					\
276 	sowwakeup_locked(so);						\
277 } while (0)
278 
279 struct accept_filter {
280 	char	accf_name[16];
281 	int	(*accf_callback)
282 		(struct socket *so, void *arg, int waitflag);
283 	void *	(*accf_create)
284 		(struct socket *so, char *arg);
285 	void	(*accf_destroy)
286 		(struct socket *so);
287 	SLIST_ENTRY(accept_filter) accf_next;
288 };
289 
290 #ifdef MALLOC_DECLARE
291 MALLOC_DECLARE(M_ACCF);
292 MALLOC_DECLARE(M_PCB);
293 MALLOC_DECLARE(M_SONAME);
294 #endif
295 
296 extern int	maxsockets;
297 extern u_long	sb_max;
298 extern so_gen_t so_gencnt;
299 
300 struct file;
301 struct filedesc;
302 struct mbuf;
303 struct sockaddr;
304 struct ucred;
305 struct uio;
306 
307 /* 'which' values for socket upcalls. */
308 #define	SO_RCV		1
309 #define	SO_SND		2
310 
311 /* Return values for socket upcalls. */
312 #define	SU_OK		0
313 #define	SU_ISCONNECTED	1
314 
315 /*
316  * From uipc_socket and friends
317  */
318 int	sockargs(struct mbuf **mp, caddr_t buf, int buflen, int type);
319 int	getsockaddr(struct sockaddr **namp, caddr_t uaddr, size_t len);
320 int	getsock_cap(struct thread *td, int fd, cap_rights_t *rightsp,
321 	    struct file **fpp, u_int *fflagp);
322 void	soabort(struct socket *so);
323 int	soaccept(struct socket *so, struct sockaddr **nam);
324 int	socheckuid(struct socket *so, uid_t uid);
325 int	sobind(struct socket *so, struct sockaddr *nam, struct thread *td);
326 int	sobindat(int fd, struct socket *so, struct sockaddr *nam,
327 	    struct thread *td);
328 int	soclose(struct socket *so);
329 int	soconnect(struct socket *so, struct sockaddr *nam, struct thread *td);
330 int	soconnectat(int fd, struct socket *so, struct sockaddr *nam,
331 	    struct thread *td);
332 int	soconnect2(struct socket *so1, struct socket *so2);
333 int	socreate(int dom, struct socket **aso, int type, int proto,
334 	    struct ucred *cred, struct thread *td);
335 int	sodisconnect(struct socket *so);
336 struct	sockaddr *sodupsockaddr(const struct sockaddr *sa, int mflags);
337 void	sofree(struct socket *so);
338 void	sohasoutofband(struct socket *so);
339 int	solisten(struct socket *so, int backlog, struct thread *td);
340 void	solisten_proto(struct socket *so, int backlog);
341 int	solisten_proto_check(struct socket *so);
342 struct socket *
343 	sonewconn(struct socket *head, int connstatus);
344 
345 
346 int	sopoll(struct socket *so, int events, struct ucred *active_cred,
347 	    struct thread *td);
348 int	sopoll_generic(struct socket *so, int events,
349 	    struct ucred *active_cred, struct thread *td);
350 int	soreceive(struct socket *so, struct sockaddr **paddr, struct uio *uio,
351 	    struct mbuf **mp0, struct mbuf **controlp, int *flagsp);
352 int	soreceive_stream(struct socket *so, struct sockaddr **paddr,
353 	    struct uio *uio, struct mbuf **mp0, struct mbuf **controlp,
354 	    int *flagsp);
355 int	soreceive_dgram(struct socket *so, struct sockaddr **paddr,
356 	    struct uio *uio, struct mbuf **mp0, struct mbuf **controlp,
357 	    int *flagsp);
358 int	soreceive_generic(struct socket *so, struct sockaddr **paddr,
359 	    struct uio *uio, struct mbuf **mp0, struct mbuf **controlp,
360 	    int *flagsp);
361 int	soreserve(struct socket *so, u_long sndcc, u_long rcvcc);
362 void	sorflush(struct socket *so);
363 int	sosend(struct socket *so, struct sockaddr *addr, struct uio *uio,
364 	    struct mbuf *top, struct mbuf *control, int flags,
365 	    struct thread *td);
366 int	sosend_dgram(struct socket *so, struct sockaddr *addr,
367 	    struct uio *uio, struct mbuf *top, struct mbuf *control,
368 	    int flags, struct thread *td);
369 int	sosend_generic(struct socket *so, struct sockaddr *addr,
370 	    struct uio *uio, struct mbuf *top, struct mbuf *control,
371 	    int flags, struct thread *td);
372 int	soshutdown(struct socket *so, int how);
373 void	sotoxsocket(struct socket *so, struct xsocket *xso);
374 void	soupcall_clear(struct socket *so, int which);
375 void	soupcall_set(struct socket *so, int which,
376 	    int (*func)(struct socket *, void *, int), void *arg);
377 void	sowakeup(struct socket *so, struct sockbuf *sb);
378 int	selsocket(struct socket *so, int events, struct timeval *tv,
379 	    struct thread *td);
380 
381 /*
382  * Accept filter functions (duh).
383  */
384 int	accept_filt_add(struct accept_filter *filt);
385 int	accept_filt_del(char *name);
386 struct	accept_filter *accept_filt_get(char *name);
387 #ifdef ACCEPT_FILTER_MOD
388 #ifdef SYSCTL_DECL
389 SYSCTL_DECL(_net_inet_accf);
390 #endif
391 int	accept_filt_generic_mod_event(module_t mod, int event, void *data);
392 #endif
393 
394 #endif /* _KERNEL */
395 
396 #endif /* !_SYS_SOCKETVAR_H_ */
397