xref: /openbsd/sys/sys/socketvar.h (revision d452b926)
1 /*	$OpenBSD: socketvar.h,v 1.112 2022/11/26 17:52:35 mvs Exp $	*/
2 /*	$NetBSD: socketvar.h,v 1.18 1996/02/09 18:25:38 christos Exp $	*/
3 
4 /*-
5  * Copyright (c) 1982, 1986, 1990, 1993
6  *	The Regents of the University of California.  All rights reserved.
7  *
8  * Redistribution and use in source and binary forms, with or without
9  * modification, are permitted provided that the following conditions
10  * are met:
11  * 1. Redistributions of source code must retain the above copyright
12  *    notice, this list of conditions and the following disclaimer.
13  * 2. Redistributions in binary form must reproduce the above copyright
14  *    notice, this list of conditions and the following disclaimer in the
15  *    documentation and/or other materials provided with the distribution.
16  * 3. Neither the name of the University nor the names of its contributors
17  *    may be used to endorse or promote products derived from this software
18  *    without specific prior written permission.
19  *
20  * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND
21  * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
22  * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
23  * ARE DISCLAIMED.  IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE
24  * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
25  * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
26  * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
27  * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
28  * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
29  * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
30  * SUCH DAMAGE.
31  *
32  *	@(#)socketvar.h	8.1 (Berkeley) 6/2/93
33  */
34 
35 #ifndef _SYS_SOCKETVAR_H_
36 #define _SYS_SOCKETVAR_H_
37 
38 #include <sys/selinfo.h>			/* for struct selinfo */
39 #include <sys/queue.h>
40 #include <sys/sigio.h>				/* for struct sigio_ref */
41 #include <sys/task.h>
42 #include <sys/timeout.h>
43 #include <sys/rwlock.h>
44 #include <sys/refcnt.h>
45 
46 #ifndef	_SOCKLEN_T_DEFINED_
47 #define	_SOCKLEN_T_DEFINED_
48 typedef	__socklen_t	socklen_t;	/* length type for network syscalls */
49 #endif
50 
51 TAILQ_HEAD(soqhead, socket);
52 
53 /*
54  * Kernel structure per socket.
55  * Contains send and receive buffer queues,
56  * handle on protocol and pointer to protocol
57  * private data and error information.
58  */
59 struct socket {
60 	const struct protosw *so_proto;	/* protocol handle */
61 	struct rwlock so_lock;		/* this socket lock */
62 	struct refcnt so_refcnt;	/* references to this socket */
63 	void	*so_pcb;		/* protocol control block */
64 	u_int	so_state;		/* internal state flags SS_*, below */
65 	short	so_type;		/* generic type, see socket.h */
66 	short	so_options;		/* from socket call, see socket.h */
67 	short	so_linger;		/* time to linger while closing */
68 /*
69  * Variables for connection queueing.
70  * Socket where accepts occur is so_head in all subsidiary sockets.
71  * If so_head is 0, socket is not related to an accept.
72  * For head socket so_q0 queues partially completed connections,
73  * while so_q is a queue of connections ready to be accepted.
74  * If a connection is aborted and it has so_head set, then
75  * it has to be pulled out of either so_q0 or so_q.
76  * We allow connections to queue up based on current queue lengths
77  * and limit on number of queued connections for this socket.
78  */
79 	struct	socket	*so_head;	/* back pointer to accept socket */
80 	struct	soqhead	*so_onq;	/* queue (q or q0) that we're on */
81 	struct	soqhead	so_q0;		/* queue of partial connections */
82 	struct	soqhead	so_q;		/* queue of incoming connections */
83 	struct	sigio_ref so_sigio;	/* async I/O registration */
84 	TAILQ_ENTRY(socket) so_qe;	/* our queue entry (q or q0) */
85 	short	so_q0len;		/* partials on so_q0 */
86 	short	so_qlen;		/* number of connections on so_q */
87 	short	so_qlimit;		/* max number queued connections */
88 	u_long	so_newconn;		/* # of pending sonewconn() threads */
89 	short	so_timeo;		/* connection timeout */
90 	u_long	so_oobmark;		/* chars to oob mark */
91 	u_int	so_error;		/* error affecting connection */
92 /*
93  * Variables for socket splicing, allocated only when needed.
94  */
95 	struct sosplice {
96 		struct	socket *ssp_socket;	/* send data to drain socket */
97 		struct	socket *ssp_soback;	/* back ref to source socket */
98 		off_t	ssp_len;		/* number of bytes spliced */
99 		off_t	ssp_max;		/* maximum number of bytes */
100 		struct	timeval ssp_idletv;	/* idle timeout */
101 		struct	timeout ssp_idleto;
102 		struct	task ssp_task;		/* task for somove */
103 	} *so_sp;
104 /*
105  * Variables for socket buffering.
106  */
107 	struct	sockbuf {
108 /* The following fields are all zeroed on flush. */
109 #define	sb_startzero	sb_cc
110 		u_long	sb_cc;		/* actual chars in buffer */
111 		u_long	sb_datacc;	/* data only chars in buffer */
112 		u_long	sb_hiwat;	/* max actual char count */
113 		u_long  sb_wat;		/* default watermark */
114 		u_long	sb_mbcnt;	/* chars of mbufs used */
115 		u_long	sb_mbmax;	/* max chars of mbufs to use */
116 		long	sb_lowat;	/* low water mark */
117 		struct mbuf *sb_mb;	/* the mbuf chain */
118 		struct mbuf *sb_mbtail;	/* the last mbuf in the chain */
119 		struct mbuf *sb_lastrecord;/* first mbuf of last record in
120 					      socket buffer */
121 		short	sb_flags;	/* flags, see below */
122 /* End area that is zeroed on flush. */
123 #define	sb_endzero	sb_flags
124 		uint64_t sb_timeo_nsecs;/* timeout for read/write */
125 		struct	selinfo sb_sel;	/* process selecting read/write */
126 	} so_rcv, so_snd;
127 #define	SB_MAX		(2*1024*1024)	/* default for max chars in sockbuf */
128 #define	SB_LOCK		0x01		/* lock on data queue */
129 #define	SB_WANT		0x02		/* someone is waiting to lock */
130 #define	SB_WAIT		0x04		/* someone is waiting for data/space */
131 #define	SB_ASYNC	0x10		/* ASYNC I/O, need signals */
132 #define	SB_SPLICE	0x20		/* buffer is splice source or drain */
133 #define	SB_NOINTR	0x40		/* operations not interruptible */
134 
135 	void	(*so_upcall)(struct socket *so, caddr_t arg, int waitf);
136 	caddr_t	so_upcallarg;		/* Arg for above */
137 	uid_t	so_euid, so_ruid;	/* who opened the socket */
138 	gid_t	so_egid, so_rgid;
139 	pid_t	so_cpid;		/* pid of process that opened socket */
140 };
141 
142 /*
143  * Socket state bits.
144  */
145 #define	SS_NOFDREF		0x001	/* no file table ref any more */
146 #define	SS_ISCONNECTED		0x002	/* socket connected to a peer */
147 #define	SS_ISCONNECTING		0x004	/* in process of connecting to peer */
148 #define	SS_ISDISCONNECTING	0x008	/* in process of disconnecting */
149 #define	SS_CANTSENDMORE		0x010	/* can't send more data to peer */
150 #define	SS_CANTRCVMORE		0x020	/* can't receive more data from peer */
151 #define	SS_RCVATMARK		0x040	/* at mark on input */
152 #define	SS_ISDISCONNECTED	0x800	/* socket disconnected from peer */
153 
154 #define	SS_PRIV			0x080	/* privileged for broadcast, raw... */
155 #define	SS_CONNECTOUT		0x1000	/* connect, not accept, at this end */
156 #define	SS_ISSENDING		0x2000	/* hint for lower layer */
157 #define	SS_DNS			0x4000	/* created using SOCK_DNS socket(2) */
158 #define	SS_NEWCONN_WAIT		0x8000	/* waiting sonewconn() relock */
159 #define	SS_YP			0x10000	/* created using ypconnect(2) */
160 
161 #ifdef _KERNEL
162 
163 #include <sys/protosw.h>
164 #include <lib/libkern/libkern.h>
165 
166 void	soassertlocked(struct socket *);
167 
168 static inline void
169 soref(struct socket *so)
170 {
171 	refcnt_take(&so->so_refcnt);
172 }
173 
174 static inline void
175 sorele(struct socket *so)
176 {
177 	refcnt_rele_wake(&so->so_refcnt);
178 }
179 
180 /*
181  * Macros for sockets and socket buffering.
182  */
183 
184 #define isspliced(so)		((so)->so_sp && (so)->so_sp->ssp_socket)
185 #define issplicedback(so)	((so)->so_sp && (so)->so_sp->ssp_soback)
186 
187 /*
188  * Do we need to notify the other side when I/O is possible?
189  */
190 static inline int
191 sb_notify(struct socket *so, struct sockbuf *sb)
192 {
193 	KASSERT(sb == &so->so_rcv || sb == &so->so_snd);
194 	soassertlocked(so);
195 	return ((sb->sb_flags & (SB_WAIT|SB_ASYNC|SB_SPLICE)) != 0 ||
196 	    !klist_empty(&sb->sb_sel.si_note));
197 }
198 
199 /*
200  * How much space is there in a socket buffer (so->so_snd or so->so_rcv)?
201  * This is problematical if the fields are unsigned, as the space might
202  * still be negative (cc > hiwat or mbcnt > mbmax).  Should detect
203  * overflow and return 0.
204  */
205 static inline long
206 sbspace(struct socket *so, struct sockbuf *sb)
207 {
208 	KASSERT(sb == &so->so_rcv || sb == &so->so_snd);
209 	soassertlocked(so);
210 	return lmin(sb->sb_hiwat - sb->sb_cc, sb->sb_mbmax - sb->sb_mbcnt);
211 }
212 
213 /* do we have to send all at once on a socket? */
214 #define	sosendallatonce(so) \
215     ((so)->so_proto->pr_flags & PR_ATOMIC)
216 
217 /* are we sending on this socket? */
218 #define	soissending(so) \
219     ((so)->so_state & SS_ISSENDING)
220 
221 /* can we read something from so? */
222 static inline int
223 soreadable(struct socket *so)
224 {
225 	soassertlocked(so);
226 	if (isspliced(so))
227 		return 0;
228 	return (so->so_state & SS_CANTRCVMORE) || so->so_qlen || so->so_error ||
229 	    so->so_rcv.sb_cc >= so->so_rcv.sb_lowat;
230 }
231 
232 /* can we write something to so? */
233 static inline int
234 sowriteable(struct socket *so)
235 {
236 	soassertlocked(so);
237 	return ((sbspace(so, &so->so_snd) >= so->so_snd.sb_lowat &&
238 	    ((so->so_state & SS_ISCONNECTED) ||
239 	    (so->so_proto->pr_flags & PR_CONNREQUIRED)==0)) ||
240 	    (so->so_state & SS_CANTSENDMORE) || so->so_error);
241 }
242 
243 /* adjust counters in sb reflecting allocation of m */
244 static inline void
245 sballoc(struct socket *so, struct sockbuf *sb, struct mbuf *m)
246 {
247 	sb->sb_cc += m->m_len;
248 	if (m->m_type != MT_CONTROL && m->m_type != MT_SONAME)
249 		sb->sb_datacc += m->m_len;
250 	sb->sb_mbcnt += MSIZE;
251 	if (m->m_flags & M_EXT)
252 		sb->sb_mbcnt += m->m_ext.ext_size;
253 }
254 
255 /* adjust counters in sb reflecting freeing of m */
256 static inline void
257 sbfree(struct socket *so, struct sockbuf *sb, struct mbuf *m)
258 {
259 	sb->sb_cc -= m->m_len;
260 	if (m->m_type != MT_CONTROL && m->m_type != MT_SONAME)
261 		sb->sb_datacc -= m->m_len;
262 	sb->sb_mbcnt -= MSIZE;
263 	if (m->m_flags & M_EXT)
264 		sb->sb_mbcnt -= m->m_ext.ext_size;
265 }
266 
267 /*
268  * Set lock on sockbuf sb; sleep if lock is already held.
269  * Unless SB_NOINTR is set on sockbuf, sleep is interruptible.
270  * Returns error without lock if sleep is interrupted.
271  */
272 int sblock(struct socket *, struct sockbuf *, int);
273 
274 /* release lock on sockbuf sb */
275 void sbunlock(struct socket *, struct sockbuf *);
276 
277 #define	SB_EMPTY_FIXUP(sb) do {						\
278 	if ((sb)->sb_mb == NULL) {					\
279 		(sb)->sb_mbtail = NULL;					\
280 		(sb)->sb_lastrecord = NULL;				\
281 	}								\
282 } while (/*CONSTCOND*/0)
283 
284 extern u_long sb_max;
285 
286 extern struct pool	socket_pool;
287 
288 struct mbuf;
289 struct sockaddr;
290 struct proc;
291 struct msghdr;
292 struct stat;
293 struct knote;
294 
295 /*
296  * File operations on sockets.
297  */
298 int	soo_read(struct file *, struct uio *, int);
299 int	soo_write(struct file *, struct uio *, int);
300 int	soo_ioctl(struct file *, u_long, caddr_t, struct proc *);
301 int	soo_kqfilter(struct file *, struct knote *);
302 int 	soo_close(struct file *, struct proc *);
303 int	soo_stat(struct file *, struct stat *, struct proc *);
304 void	sbappend(struct socket *, struct sockbuf *, struct mbuf *);
305 void	sbappendstream(struct socket *, struct sockbuf *, struct mbuf *);
306 int	sbappendaddr(struct socket *, struct sockbuf *,
307 	    const struct sockaddr *, struct mbuf *, struct mbuf *);
308 int	sbappendcontrol(struct socket *, struct sockbuf *, struct mbuf *,
309 	    struct mbuf *);
310 void	sbappendrecord(struct socket *, struct sockbuf *, struct mbuf *);
311 void	sbcompress(struct socket *, struct sockbuf *, struct mbuf *,
312 	    struct mbuf *);
313 struct mbuf *
314 	sbcreatecontrol(const void *, size_t, int, int);
315 void	sbdrop(struct socket *, struct sockbuf *, int);
316 void	sbdroprecord(struct socket *, struct sockbuf *);
317 void	sbflush(struct socket *, struct sockbuf *);
318 void	sbrelease(struct socket *, struct sockbuf *);
319 int	sbcheckreserve(u_long, u_long);
320 int	sbchecklowmem(void);
321 int	sbreserve(struct socket *, struct sockbuf *, u_long);
322 int	sbwait(struct socket *, struct sockbuf *);
323 void	soinit(void);
324 void	soabort(struct socket *);
325 int	soaccept(struct socket *, struct mbuf *);
326 int	sobind(struct socket *, struct mbuf *, struct proc *);
327 void	socantrcvmore(struct socket *);
328 void	socantsendmore(struct socket *);
329 int	soclose(struct socket *, int);
330 int	soconnect(struct socket *, struct mbuf *);
331 int	soconnect2(struct socket *, struct socket *);
332 int	socreate(int, struct socket **, int, int);
333 int	sodisconnect(struct socket *);
334 struct socket *soalloc(int);
335 void	sofree(struct socket *, int);
336 int	sogetopt(struct socket *, int, int, struct mbuf *);
337 void	sohasoutofband(struct socket *);
338 void	soisconnected(struct socket *);
339 void	soisconnecting(struct socket *);
340 void	soisdisconnected(struct socket *);
341 void	soisdisconnecting(struct socket *);
342 int	solisten(struct socket *, int);
343 struct socket *sonewconn(struct socket *, int, int);
344 void	soqinsque(struct socket *, struct socket *, int);
345 int	soqremque(struct socket *, int);
346 int	soreceive(struct socket *, struct mbuf **, struct uio *,
347 	    struct mbuf **, struct mbuf **, int *, socklen_t);
348 int	soreserve(struct socket *, u_long, u_long);
349 int	sosend(struct socket *, struct mbuf *, struct uio *,
350 	    struct mbuf *, struct mbuf *, int);
351 int	sosetopt(struct socket *, int, int, struct mbuf *);
352 int	soshutdown(struct socket *, int);
353 void	sowakeup(struct socket *, struct sockbuf *);
354 void	sorwakeup(struct socket *);
355 void	sowwakeup(struct socket *);
356 int	sockargs(struct mbuf **, const void *, size_t, int);
357 
358 int	sosleep_nsec(struct socket *, void *, int, const char *, uint64_t);
359 void	solock(struct socket *);
360 void	solock_shared(struct socket *);
361 int	solock_persocket(struct socket *);
362 void	solock_pair(struct socket *, struct socket *);
363 void	sounlock(struct socket *);
364 void	sounlock_shared(struct socket *);
365 
366 int	sendit(struct proc *, int, struct msghdr *, int, register_t *);
367 int	recvit(struct proc *, int, struct msghdr *, caddr_t, register_t *);
368 int	doaccept(struct proc *, int, struct sockaddr *, socklen_t *, int,
369 	    register_t *);
370 
371 #ifdef SOCKBUF_DEBUG
372 void	sblastrecordchk(struct sockbuf *, const char *);
373 #define	SBLASTRECORDCHK(sb, where)	sblastrecordchk((sb), (where))
374 
375 void	sblastmbufchk(struct sockbuf *, const char *);
376 #define	SBLASTMBUFCHK(sb, where)	sblastmbufchk((sb), (where))
377 void	sbcheck(struct socket *, struct sockbuf *);
378 #define	SBCHECK(so, sb)			sbcheck((so), (sb))
379 #else
380 #define	SBLASTRECORDCHK(sb, where)	/* nothing */
381 #define	SBLASTMBUFCHK(sb, where)	/* nothing */
382 #define	SBCHECK(so, sb)			/* nothing */
383 #endif /* SOCKBUF_DEBUG */
384 
385 #endif /* _KERNEL */
386 
387 #endif /* _SYS_SOCKETVAR_H_ */
388