socketvar.h revision 129958
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 * $FreeBSD: head/sys/sys/socketvar.h 129958 2004-06-01 19:33:06Z rwatson $
31 */
32
33#ifndef _SYS_SOCKETVAR_H_
34#define _SYS_SOCKETVAR_H_
35
36#include <sys/queue.h>			/* for TAILQ macros */
37#include <sys/selinfo.h>		/* for struct selinfo */
38
39/*
40 * Kernel structure per socket.
41 * Contains send and receive buffer queues,
42 * handle on protocol and pointer to protocol
43 * private data and error information.
44 */
45typedef	u_quad_t so_gen_t;
46
47/*-
48 * Locking key to struct socket:
49 * (a) constant after allocation, no locking required.
50 * (b) locked by SOCK_LOCK(so).
51 * (c) locked by SOCKBUF_LOCK(&so->so_rcv).
52 * (d) locked by SOCKBUF_LOCK(&so->so_snd).
53 * (e) locked by ACCEPT_LOCK().
54 * (f) not locked since integer reads/writes are atomic.
55 * (g) used only as a sleep/wakeup address, no value.
56 */
57struct socket {
58	int	so_count;		/* reference count */
59	short	so_type;		/* (a) generic type, see socket.h */
60	short	so_options;		/* from socket call, see socket.h */
61	short	so_linger;		/* time to linger while closing */
62	short	so_state;		/* internal state flags SS_* */
63	int	so_qstate;		/* internal state flags SQ_* */
64	void	*so_pcb;		/* protocol control block */
65	struct	protosw *so_proto;	/* (a) protocol handle */
66/*
67 * Variables for connection queuing.
68 * Socket where accepts occur is so_head in all subsidiary sockets.
69 * If so_head is 0, socket is not related to an accept.
70 * For head socket so_incomp queues partially completed connections,
71 * while so_comp is a queue of connections ready to be accepted.
72 * If a connection is aborted and it has so_head set, then
73 * it has to be pulled out of either so_incomp or so_comp.
74 * We allow connections to queue up based on current queue lengths
75 * and limit on number of queued connections for this socket.
76 */
77	struct	socket *so_head;	/* back pointer to accept socket */
78	TAILQ_HEAD(, socket) so_incomp;	/* queue of partial unaccepted connections */
79	TAILQ_HEAD(, socket) so_comp;	/* queue of complete unaccepted connections */
80	TAILQ_ENTRY(socket) so_list;	/* list of unaccepted connections */
81	short	so_qlen;		/* number of unaccepted connections */
82	short	so_incqlen;		/* number of unaccepted incomplete
83					   connections */
84	short	so_qlimit;		/* max number queued connections */
85	short	so_timeo;		/* connection timeout */
86	u_short	so_error;		/* error affecting connection */
87	struct	sigio *so_sigio;	/* [sg] information for async I/O or
88					   out of band data (SIGURG) */
89	u_long	so_oobmark;		/* chars to oob mark */
90	TAILQ_HEAD(, aiocblist) so_aiojobq; /* AIO ops waiting on socket */
91/*
92 * Variables for socket buffering.
93 */
94	struct sockbuf {
95		struct	selinfo sb_sel;	/* process selecting read/write */
96#define	sb_startzero	sb_mb
97		struct	mbuf *sb_mb;	/* the mbuf chain */
98		struct	mbuf *sb_mbtail; /* the last mbuf in the chain */
99		struct	mbuf *sb_lastrecord;	/* first mbuf of last record in
100						 * socket buffer */
101		u_int	sb_cc;		/* actual chars in buffer */
102		u_int	sb_hiwat;	/* max actual char count */
103		u_int	sb_mbcnt;	/* chars of mbufs used */
104		u_int	sb_mbmax;	/* max chars of mbufs to use */
105		u_int	sb_ctl;		/* non-data chars in buffer */
106		int	sb_lowat;	/* low water mark */
107		int	sb_timeo;	/* timeout for read/write */
108		short	sb_flags;	/* flags, see below */
109	} so_rcv, so_snd;
110#define	SB_MAX		(256*1024)	/* default for max chars in sockbuf */
111#define	SB_LOCK		0x01		/* lock on data queue */
112#define	SB_WANT		0x02		/* someone is waiting to lock */
113#define	SB_WAIT		0x04		/* someone is waiting for data/space */
114#define	SB_SEL		0x08		/* someone is selecting */
115#define	SB_ASYNC	0x10		/* ASYNC I/O, need signals */
116#define	SB_UPCALL	0x20		/* someone wants an upcall */
117#define	SB_NOINTR	0x40		/* operations not interruptible */
118#define SB_AIO		0x80		/* AIO operations queued */
119#define SB_KNOTE	0x100		/* kernel note attached */
120
121	void	(*so_upcall)(struct socket *, void *, int);
122	void	*so_upcallarg;
123	struct	ucred *so_cred;		/* user credentials */
124	struct	label *so_label;	/* MAC label for socket */
125	struct	label *so_peerlabel;	/* cached MAC label for socket peer */
126	/* NB: generation count must not be first; easiest to make it last. */
127	so_gen_t so_gencnt;		/* generation count */
128	void	*so_emuldata;		/* private data for emulators */
129 	struct so_accf {
130		struct	accept_filter *so_accept_filter;
131		void	*so_accept_filter_arg;	/* saved filter args */
132		char	*so_accept_filter_str;	/* saved user args */
133	} *so_accf;
134};
135
136#define SB_EMPTY_FIXUP(sb) do {						\
137	if ((sb)->sb_mb == NULL) {					\
138		(sb)->sb_mbtail = NULL;					\
139		(sb)->sb_lastrecord = NULL;				\
140	}								\
141} while (/*CONSTCOND*/0)
142
143/*
144 * Socket state bits.
145 */
146#define	SS_NOFDREF		0x0001	/* no file table ref any more */
147#define	SS_ISCONNECTED		0x0002	/* socket connected to a peer */
148#define	SS_ISCONNECTING		0x0004	/* in process of connecting to peer */
149#define	SS_ISDISCONNECTING	0x0008	/* in process of disconnecting */
150#define	SS_CANTSENDMORE		0x0010	/* can't send more data to peer */
151#define	SS_CANTRCVMORE		0x0020	/* can't receive more data from peer */
152#define	SS_RCVATMARK		0x0040	/* at mark on input */
153
154#define	SS_NBIO			0x0100	/* non-blocking ops */
155#define	SS_ASYNC		0x0200	/* async i/o notify */
156#define	SS_ISCONFIRMING		0x0400	/* deciding to accept connection req */
157
158#define	SS_ISDISCONNECTED	0x2000	/* socket disconnected from peer */
159
160/*
161 * Socket state bits stored in so_qstate.
162 */
163#define	SQ_INCOMP		0x0800	/* unaccepted, incomplete connection */
164#define	SQ_COMP			0x1000	/* unaccepted, complete connection */
165
166/*
167 * Externalized form of struct socket used by the sysctl(3) interface.
168 */
169struct xsocket {
170	size_t	xso_len;	/* length of this structure */
171	struct	socket *xso_so;	/* makes a convenient handle sometimes */
172	short	so_type;
173	short	so_options;
174	short	so_linger;
175	short	so_state;
176	caddr_t	so_pcb;		/* another convenient handle */
177	int	xso_protocol;
178	int	xso_family;
179	short	so_qlen;
180	short	so_incqlen;
181	short	so_qlimit;
182	short	so_timeo;
183	u_short	so_error;
184	pid_t	so_pgid;
185	u_long	so_oobmark;
186	struct xsockbuf {
187		u_int	sb_cc;
188		u_int	sb_hiwat;
189		u_int	sb_mbcnt;
190		u_int	sb_mbmax;
191		int	sb_lowat;
192		int	sb_timeo;
193		short	sb_flags;
194	} so_rcv, so_snd;
195	uid_t	so_uid;		/* XXX */
196};
197
198/*
199 * Macros for sockets and socket buffering.
200 */
201
202/*
203 * Do we need to notify the other side when I/O is possible?
204 */
205#define	sb_notify(sb)	(((sb)->sb_flags & (SB_WAIT | SB_SEL | SB_ASYNC | \
206    SB_UPCALL | SB_AIO | SB_KNOTE)) != 0)
207
208/*
209 * How much space is there in a socket buffer (so->so_snd or so->so_rcv)?
210 * This is problematical if the fields are unsigned, as the space might
211 * still be negative (cc > hiwat or mbcnt > mbmax).  Should detect
212 * overflow and return 0.  Should use "lmin" but it doesn't exist now.
213 */
214#define	sbspace(sb) \
215    ((long) imin((int)((sb)->sb_hiwat - (sb)->sb_cc), \
216	 (int)((sb)->sb_mbmax - (sb)->sb_mbcnt)))
217
218/* do we have to send all at once on a socket? */
219#define	sosendallatonce(so) \
220    ((so)->so_proto->pr_flags & PR_ATOMIC)
221
222/* can we read something from so? */
223#define	soreadable(so) \
224    ((so)->so_rcv.sb_cc >= (so)->so_rcv.sb_lowat || \
225	((so)->so_state & SS_CANTRCVMORE) || \
226	!TAILQ_EMPTY(&(so)->so_comp) || (so)->so_error)
227
228/* can we write something to so? */
229#define	sowriteable(so) \
230    ((sbspace(&(so)->so_snd) >= (so)->so_snd.sb_lowat && \
231	(((so)->so_state&SS_ISCONNECTED) || \
232	  ((so)->so_proto->pr_flags&PR_CONNREQUIRED)==0)) || \
233     ((so)->so_state & SS_CANTSENDMORE) || \
234     (so)->so_error)
235
236/* adjust counters in sb reflecting allocation of m */
237#define	sballoc(sb, m) { \
238	(sb)->sb_cc += (m)->m_len; \
239	if ((m)->m_type != MT_DATA && (m)->m_type != MT_HEADER && \
240	    (m)->m_type != MT_OOBDATA) \
241		(sb)->sb_ctl += (m)->m_len; \
242	(sb)->sb_mbcnt += MSIZE; \
243	if ((m)->m_flags & M_EXT) \
244		(sb)->sb_mbcnt += (m)->m_ext.ext_size; \
245}
246
247/* adjust counters in sb reflecting freeing of m */
248#define	sbfree(sb, m) { \
249	(sb)->sb_cc -= (m)->m_len; \
250	if ((m)->m_type != MT_DATA && (m)->m_type != MT_HEADER && \
251	    (m)->m_type != MT_OOBDATA) \
252		(sb)->sb_ctl -= (m)->m_len; \
253	(sb)->sb_mbcnt -= MSIZE; \
254	if ((m)->m_flags & M_EXT) \
255		(sb)->sb_mbcnt -= (m)->m_ext.ext_size; \
256}
257
258/*
259 * Set lock on sockbuf sb; sleep if lock is already held.
260 * Unless SB_NOINTR is set on sockbuf, sleep is interruptible.
261 * Returns error without lock if sleep is interrupted.
262 */
263#define sblock(sb, wf) ((sb)->sb_flags & SB_LOCK ? \
264		(((wf) == M_WAITOK) ? sb_lock(sb) : EWOULDBLOCK) : \
265		((sb)->sb_flags |= SB_LOCK), 0)
266
267/* release lock on sockbuf sb */
268#define	sbunlock(sb) { \
269	(sb)->sb_flags &= ~SB_LOCK; \
270	if ((sb)->sb_flags & SB_WANT) { \
271		(sb)->sb_flags &= ~SB_WANT; \
272		wakeup(&(sb)->sb_flags); \
273	} \
274}
275
276/*
277 * soref()/sorele() ref-count the socket structure.  Note that you must
278 * still explicitly close the socket, but the last ref count will free
279 * the structure.
280 */
281#define soref(so)	do {			\
282				++(so)->so_count; \
283			} while (0)
284
285#define sorele(so)	do {				\
286				if ((so)->so_count <= 0)	\
287					panic("sorele");\
288				if (--(so)->so_count == 0)\
289					sofree(so);	\
290			} while (0)
291
292#define sotryfree(so)	do {				\
293				if ((so)->so_count == 0)	\
294					sofree(so);	\
295			} while(0)
296
297#define	sorwakeup(so)	do {					\
298				if (sb_notify(&(so)->so_rcv))	\
299					sowakeup((so), &(so)->so_rcv); \
300			} while (0)
301
302#define	sowwakeup(so)	do {					\
303				if (sb_notify(&(so)->so_snd))	\
304					sowakeup((so), &(so)->so_snd); \
305			} while (0)
306
307#ifdef _KERNEL
308
309/*
310 * Argument structure for sosetopt et seq.  This is in the KERNEL
311 * section because it will never be visible to user code.
312 */
313enum sopt_dir { SOPT_GET, SOPT_SET };
314struct sockopt {
315	enum	sopt_dir sopt_dir; /* is this a get or a set? */
316	int	sopt_level;	/* second arg of [gs]etsockopt */
317	int	sopt_name;	/* third arg of [gs]etsockopt */
318	void   *sopt_val;	/* fourth arg of [gs]etsockopt */
319	size_t	sopt_valsize;	/* (almost) fifth arg of [gs]etsockopt */
320	struct	thread *sopt_td; /* calling thread or null if kernel */
321};
322
323struct accept_filter {
324	char	accf_name[16];
325	void	(*accf_callback)
326		(struct socket *so, void *arg, int waitflag);
327	void *	(*accf_create)
328		(struct socket *so, char *arg);
329	void	(*accf_destroy)
330		(struct socket *so);
331	SLIST_ENTRY(accept_filter) accf_next;
332};
333
334#ifdef MALLOC_DECLARE
335MALLOC_DECLARE(M_ACCF);
336MALLOC_DECLARE(M_PCB);
337MALLOC_DECLARE(M_SONAME);
338#endif
339
340extern int	maxsockets;
341extern u_long	sb_max;
342extern struct uma_zone *socket_zone;
343extern so_gen_t so_gencnt;
344
345struct mbuf;
346struct sockaddr;
347struct ucred;
348struct uio;
349
350/*
351 * From uipc_socket and friends
352 */
353int	sockargs(struct mbuf **mp, caddr_t buf, int buflen, int type);
354int	getsockaddr(struct sockaddr **namp, caddr_t uaddr, size_t len);
355void	sbappend(struct sockbuf *sb, struct mbuf *m);
356void	sbappendstream(struct sockbuf *sb, struct mbuf *m);
357int	sbappendaddr(struct sockbuf *sb, struct sockaddr *asa,
358	    struct mbuf *m0, struct mbuf *control);
359int	sbappendcontrol(struct sockbuf *sb, struct mbuf *m0,
360	    struct mbuf *control);
361void	sbappendrecord(struct sockbuf *sb, struct mbuf *m0);
362void	sbcheck(struct sockbuf *sb);
363void	sbcompress(struct sockbuf *sb, struct mbuf *m, struct mbuf *n);
364struct mbuf *
365	sbcreatecontrol(caddr_t p, int size, int type, int level);
366void	sbdrop(struct sockbuf *sb, int len);
367void	sbdroprecord(struct sockbuf *sb);
368void	sbflush(struct sockbuf *sb);
369void	sbinsertoob(struct sockbuf *sb, struct mbuf *m0);
370void	sbrelease(struct sockbuf *sb, struct socket *so);
371int	sbreserve(struct sockbuf *sb, u_long cc, struct socket *so,
372	    struct thread *td);
373void	sbtoxsockbuf(struct sockbuf *sb, struct xsockbuf *xsb);
374int	sbwait(struct sockbuf *sb);
375int	sb_lock(struct sockbuf *sb);
376int	soabort(struct socket *so);
377int	soaccept(struct socket *so, struct sockaddr **nam);
378struct	socket *soalloc(int mflags);
379int	socheckuid(struct socket *so, uid_t uid);
380int	sobind(struct socket *so, struct sockaddr *nam, struct thread *td);
381void	socantrcvmore(struct socket *so);
382void	socantsendmore(struct socket *so);
383int	soclose(struct socket *so);
384int	soconnect(struct socket *so, struct sockaddr *nam, struct thread *td);
385int	soconnect2(struct socket *so1, struct socket *so2);
386int	socow_setup(struct mbuf *m0, struct uio *uio);
387int	socreate(int dom, struct socket **aso, int type, int proto,
388	    struct ucred *cred, struct thread *td);
389void	sodealloc(struct socket *so);
390int	sodisconnect(struct socket *so);
391struct	sockaddr *sodupsockaddr(const struct sockaddr *sa, int mflags);
392void	sofree(struct socket *so);
393int	sogetopt(struct socket *so, struct sockopt *sopt);
394void	sohasoutofband(struct socket *so);
395void	soisconnected(struct socket *so);
396void	soisconnecting(struct socket *so);
397void	soisdisconnected(struct socket *so);
398void	soisdisconnecting(struct socket *so);
399int	solisten(struct socket *so, int backlog, struct thread *td);
400struct socket *
401	sonewconn(struct socket *head, int connstatus);
402int	sooptcopyin(struct sockopt *sopt, void *buf, size_t len, size_t minlen);
403int	sooptcopyout(struct sockopt *sopt, const void *buf, size_t len);
404
405/* XXX; prepare mbuf for (__FreeBSD__ < 3) routines. */
406int	soopt_getm(struct sockopt *sopt, struct mbuf **mp);
407int	soopt_mcopyin(struct sockopt *sopt, struct mbuf *m);
408int	soopt_mcopyout(struct sockopt *sopt, struct mbuf *m);
409
410int	sopoll(struct socket *so, int events, struct ucred *active_cred,
411	    struct thread *td);
412int	soreceive(struct socket *so, struct sockaddr **paddr, struct uio *uio,
413	    struct mbuf **mp0, struct mbuf **controlp, int *flagsp);
414int	soreserve(struct socket *so, u_long sndcc, u_long rcvcc);
415void	sorflush(struct socket *so);
416int	sosend(struct socket *so, struct sockaddr *addr, struct uio *uio,
417	    struct mbuf *top, struct mbuf *control, int flags,
418	    struct thread *td);
419int	sosetopt(struct socket *so, struct sockopt *sopt);
420int	soshutdown(struct socket *so, int how);
421void	sotoxsocket(struct socket *so, struct xsocket *xso);
422void	sowakeup(struct socket *so, struct sockbuf *sb);
423
424#ifdef SOCKBUF_DEBUG
425void	sblastrecordchk(struct sockbuf *, const char *, int);
426#define	SBLASTRECORDCHK(sb)	sblastrecordchk((sb), __FILE__, __LINE__)
427
428void	sblastmbufchk(struct sockbuf *, const char *, int);
429#define	SBLASTMBUFCHK(sb)	sblastmbufchk((sb), __FILE__, __LINE__)
430#else
431#define	SBLASTRECORDCHK(sb)      /* nothing */
432#define	SBLASTMBUFCHK(sb)        /* nothing */
433#endif /* SOCKBUF_DEBUG */
434
435/*
436 * Accept filter functions (duh).
437 */
438int	accept_filt_add(struct accept_filter *filt);
439int	accept_filt_del(char *name);
440struct	accept_filter *accept_filt_get(char *name);
441#ifdef ACCEPT_FILTER_MOD
442#ifdef SYSCTL_DECL
443SYSCTL_DECL(_net_inet_accf);
444#endif
445int	accept_filt_generic_mod_event(module_t mod, int event, void *data);
446#endif
447
448#endif /* _KERNEL */
449
450#endif /* !_SYS_SOCKETVAR_H_ */
451