1 /* $OpenBSD: socketvar.h,v 1.130 2024/05/03 17:43:09 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/event.h>
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/mutex.h>
44 #include <sys/rwlock.h>
45 #include <sys/refcnt.h>
46
47 #ifndef _SOCKLEN_T_DEFINED_
48 #define _SOCKLEN_T_DEFINED_
49 typedef __socklen_t socklen_t; /* length type for network syscalls */
50 #endif
51
52 TAILQ_HEAD(soqhead, socket);
53
54 /*
55 * Kernel structure per socket.
56 * Contains send and receive buffer queues,
57 * handle on protocol and pointer to protocol
58 * private data and error information.
59 */
60 struct socket {
61 const struct protosw *so_proto; /* protocol handle */
62 struct rwlock so_lock; /* this socket lock */
63 struct refcnt so_refcnt; /* references to this socket */
64 void *so_pcb; /* protocol control block */
65 u_int so_state; /* internal state flags SS_*, below */
66 short so_type; /* generic type, see socket.h */
67 short so_options; /* from socket call, see socket.h */
68 short so_linger; /* time to linger while closing */
69 /*
70 * Variables for connection queueing.
71 * Socket where accepts occur is so_head in all subsidiary sockets.
72 * If so_head is 0, socket is not related to an accept.
73 * For head socket so_q0 queues partially completed connections,
74 * while so_q is a queue of connections ready to be accepted.
75 * If a connection is aborted and it has so_head set, then
76 * it has to be pulled out of either so_q0 or so_q.
77 * We allow connections to queue up based on current queue lengths
78 * and limit on number of queued connections for this socket.
79 */
80 struct socket *so_head; /* back pointer to accept socket */
81 struct soqhead *so_onq; /* queue (q or q0) that we're on */
82 struct soqhead so_q0; /* queue of partial connections */
83 struct soqhead so_q; /* queue of incoming connections */
84 struct sigio_ref so_sigio; /* async I/O registration */
85 TAILQ_ENTRY(socket) so_qe; /* our queue entry (q or q0) */
86 short so_q0len; /* partials on so_q0 */
87 short so_qlen; /* number of connections on so_q */
88 short so_qlimit; /* max number queued connections */
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 struct rwlock sb_lock;
109 struct mutex sb_mtx;
110 /* The following fields are all zeroed on flush. */
111 #define sb_startzero sb_cc
112 u_long sb_cc; /* actual chars in buffer */
113 u_long sb_datacc; /* data only chars in buffer */
114 u_long sb_hiwat; /* max actual char count */
115 u_long sb_wat; /* default watermark */
116 u_long sb_mbcnt; /* chars of mbufs used */
117 u_long sb_mbmax; /* max chars of mbufs to use */
118 long sb_lowat; /* low water mark */
119 struct mbuf *sb_mb; /* the mbuf chain */
120 struct mbuf *sb_mbtail; /* the last mbuf in the chain */
121 struct mbuf *sb_lastrecord;/* first mbuf of last record in
122 socket buffer */
123 short sb_flags; /* flags, see below */
124 /* End area that is zeroed on flush. */
125 #define sb_endzero sb_flags
126 short sb_state; /* socket state on sockbuf */
127 uint64_t sb_timeo_nsecs;/* timeout for read/write */
128 struct klist sb_klist; /* process selecting read/write */
129 } so_rcv, so_snd;
130 #define SB_MAX (2*1024*1024) /* default for max chars in sockbuf */
131 #define SB_LOCK 0x0001 /* lock on data queue */
132 #define SB_WANT 0x0002 /* someone is waiting to lock */
133 #define SB_WAIT 0x0004 /* someone is waiting for data/space */
134 #define SB_ASYNC 0x0010 /* ASYNC I/O, need signals */
135 #define SB_SPLICE 0x0020 /* buffer is splice source or drain */
136 #define SB_NOINTR 0x0040 /* operations not interruptible */
137 #define SB_MTXLOCK 0x0080 /* sblock() doesn't need solock() */
138
139 void (*so_upcall)(struct socket *so, caddr_t arg, int waitf);
140 caddr_t so_upcallarg; /* Arg for above */
141 uid_t so_euid, so_ruid; /* who opened the socket */
142 gid_t so_egid, so_rgid;
143 pid_t so_cpid; /* pid of process that opened socket */
144 };
145
146 /*
147 * Socket state bits.
148 *
149 * NOTE: The following states should be used with corresponding socket's
150 * buffer `sb_state' only:
151 *
152 * SS_CANTSENDMORE with `so_snd'
153 * SS_ISSENDING with `so_snd'
154 * SS_CANTRCVMORE with `so_rcv'
155 * SS_RCVATMARK with `so_rcv'
156 */
157
158 #define SS_NOFDREF 0x001 /* no file table ref any more */
159 #define SS_ISCONNECTED 0x002 /* socket connected to a peer */
160 #define SS_ISCONNECTING 0x004 /* in process of connecting to peer */
161 #define SS_ISDISCONNECTING 0x008 /* in process of disconnecting */
162 #define SS_CANTSENDMORE 0x010 /* can't send more data to peer */
163 #define SS_CANTRCVMORE 0x020 /* can't receive more data from peer */
164 #define SS_RCVATMARK 0x040 /* at mark on input */
165 #define SS_ISDISCONNECTED 0x800 /* socket disconnected from peer */
166
167 #define SS_PRIV 0x080 /* privileged for broadcast, raw... */
168 #define SS_CONNECTOUT 0x1000 /* connect, not accept, at this end */
169 #define SS_ISSENDING 0x2000 /* hint for lower layer */
170 #define SS_DNS 0x4000 /* created using SOCK_DNS socket(2) */
171 #define SS_YP 0x8000 /* created using ypconnect(2) */
172
173 #ifdef _KERNEL
174
175 #include <sys/protosw.h>
176 #include <lib/libkern/libkern.h>
177
178 void soassertlocked(struct socket *);
179 void soassertlocked_readonly(struct socket *);
180
181 static inline void
soref(struct socket * so)182 soref(struct socket *so)
183 {
184 refcnt_take(&so->so_refcnt);
185 }
186
187 static inline void
sorele(struct socket * so)188 sorele(struct socket *so)
189 {
190 refcnt_rele_wake(&so->so_refcnt);
191 }
192
193 /*
194 * Macros for sockets and socket buffering.
195 */
196
197 #define isspliced(so) ((so)->so_sp && (so)->so_sp->ssp_socket)
198 #define issplicedback(so) ((so)->so_sp && (so)->so_sp->ssp_soback)
199
200 static inline void
sb_mtx_lock(struct sockbuf * sb)201 sb_mtx_lock(struct sockbuf *sb)
202 {
203 if (sb->sb_flags & SB_MTXLOCK)
204 mtx_enter(&sb->sb_mtx);
205 }
206
207 static inline void
sb_mtx_unlock(struct sockbuf * sb)208 sb_mtx_unlock(struct sockbuf *sb)
209 {
210 if (sb->sb_flags & SB_MTXLOCK)
211 mtx_leave(&sb->sb_mtx);
212 }
213
214 void sbmtxassertlocked(struct socket *so, struct sockbuf *);
215
216 /*
217 * Do we need to notify the other side when I/O is possible?
218 */
219 static inline int
sb_notify(struct socket * so,struct sockbuf * sb)220 sb_notify(struct socket *so, struct sockbuf *sb)
221 {
222 int rv;
223
224 soassertlocked(so);
225
226 mtx_enter(&sb->sb_mtx);
227 rv = ((sb->sb_flags & (SB_WAIT|SB_ASYNC|SB_SPLICE)) != 0 ||
228 !klist_empty(&sb->sb_klist));
229 mtx_leave(&sb->sb_mtx);
230
231 return rv;
232 }
233
234 /*
235 * How much space is there in a socket buffer (so->so_snd or so->so_rcv)?
236 * This is problematical if the fields are unsigned, as the space might
237 * still be negative (cc > hiwat or mbcnt > mbmax). Should detect
238 * overflow and return 0.
239 */
240
241 static inline long
sbspace(struct socket * so,struct sockbuf * sb)242 sbspace(struct socket *so, struct sockbuf *sb)
243 {
244 if (sb->sb_flags & SB_MTXLOCK)
245 sbmtxassertlocked(so, sb);
246 else
247 soassertlocked_readonly(so);
248
249 return lmin(sb->sb_hiwat - sb->sb_cc, sb->sb_mbmax - sb->sb_mbcnt);
250 }
251
252 /* do we have to send all at once on a socket? */
253 #define sosendallatonce(so) \
254 ((so)->so_proto->pr_flags & PR_ATOMIC)
255
256 /* are we sending on this socket? */
257 #define soissending(so) \
258 ((so)->so_snd.sb_state & SS_ISSENDING)
259
260 /* can we read something from so? */
261 static inline int
soreadable(struct socket * so)262 soreadable(struct socket *so)
263 {
264 soassertlocked_readonly(so);
265 if (isspliced(so))
266 return 0;
267 return (so->so_rcv.sb_state & SS_CANTRCVMORE) || so->so_qlen ||
268 so->so_error || so->so_rcv.sb_cc >= so->so_rcv.sb_lowat;
269 }
270
271 /* can we write something to so? */
272 static inline int
sowriteable(struct socket * so)273 sowriteable(struct socket *so)
274 {
275 soassertlocked_readonly(so);
276 return ((sbspace(so, &so->so_snd) >= so->so_snd.sb_lowat &&
277 ((so->so_state & SS_ISCONNECTED) ||
278 (so->so_proto->pr_flags & PR_CONNREQUIRED)==0)) ||
279 (so->so_snd.sb_state & SS_CANTSENDMORE) || so->so_error);
280 }
281
282 /* adjust counters in sb reflecting allocation of m */
283 static inline void
sballoc(struct socket * so,struct sockbuf * sb,struct mbuf * m)284 sballoc(struct socket *so, struct sockbuf *sb, struct mbuf *m)
285 {
286 sb->sb_cc += m->m_len;
287 if (m->m_type != MT_CONTROL && m->m_type != MT_SONAME)
288 sb->sb_datacc += m->m_len;
289 sb->sb_mbcnt += MSIZE;
290 if (m->m_flags & M_EXT)
291 sb->sb_mbcnt += m->m_ext.ext_size;
292 }
293
294 /* adjust counters in sb reflecting freeing of m */
295 static inline void
sbfree(struct socket * so,struct sockbuf * sb,struct mbuf * m)296 sbfree(struct socket *so, struct sockbuf *sb, struct mbuf *m)
297 {
298 sb->sb_cc -= m->m_len;
299 if (m->m_type != MT_CONTROL && m->m_type != MT_SONAME)
300 sb->sb_datacc -= m->m_len;
301 sb->sb_mbcnt -= MSIZE;
302 if (m->m_flags & M_EXT)
303 sb->sb_mbcnt -= m->m_ext.ext_size;
304 }
305
306 /*
307 * Flags to sblock()
308 */
309 #define SBL_WAIT 0x01 /* Wait if lock not immediately available. */
310 #define SBL_NOINTR 0x02 /* Enforce non-interruptible sleep. */
311
312 /*
313 * Set lock on sockbuf sb; sleep if lock is already held.
314 * Unless SB_NOINTR is set on sockbuf or SBL_NOINTR passed,
315 * sleep is interruptible. Returns error without lock if
316 * sleep is interrupted.
317 */
318 int sblock(struct socket *, struct sockbuf *, int);
319
320 /* release lock on sockbuf sb */
321 void sbunlock(struct socket *, struct sockbuf *);
322 void sbunlock_locked(struct socket *, struct sockbuf *);
323
324 #define SB_EMPTY_FIXUP(sb) do { \
325 if ((sb)->sb_mb == NULL) { \
326 (sb)->sb_mbtail = NULL; \
327 (sb)->sb_lastrecord = NULL; \
328 } \
329 } while (/*CONSTCOND*/0)
330
331 extern u_long sb_max;
332
333 extern struct pool socket_pool;
334
335 struct mbuf;
336 struct sockaddr;
337 struct proc;
338 struct msghdr;
339 struct stat;
340 struct knote;
341
342 /*
343 * File operations on sockets.
344 */
345 int soo_read(struct file *, struct uio *, int);
346 int soo_write(struct file *, struct uio *, int);
347 int soo_ioctl(struct file *, u_long, caddr_t, struct proc *);
348 int soo_kqfilter(struct file *, struct knote *);
349 int soo_close(struct file *, struct proc *);
350 int soo_stat(struct file *, struct stat *, struct proc *);
351 void sbappend(struct socket *, struct sockbuf *, struct mbuf *);
352 void sbappendstream(struct socket *, struct sockbuf *, struct mbuf *);
353 int sbappendaddr(struct socket *, struct sockbuf *,
354 const struct sockaddr *, struct mbuf *, struct mbuf *);
355 int sbappendcontrol(struct socket *, struct sockbuf *, struct mbuf *,
356 struct mbuf *);
357 void sbappendrecord(struct socket *, struct sockbuf *, struct mbuf *);
358 void sbcompress(struct socket *, struct sockbuf *, struct mbuf *,
359 struct mbuf *);
360 struct mbuf *
361 sbcreatecontrol(const void *, size_t, int, int);
362 void sbdrop(struct socket *, struct sockbuf *, int);
363 void sbdroprecord(struct socket *, struct sockbuf *);
364 void sbflush(struct socket *, struct sockbuf *);
365 void sbrelease(struct socket *, struct sockbuf *);
366 int sbcheckreserve(u_long, u_long);
367 int sbchecklowmem(void);
368 int sbreserve(struct socket *, struct sockbuf *, u_long);
369 int sbwait(struct socket *, struct sockbuf *);
370 int sbwait_locked(struct socket *, struct sockbuf *);
371 void soinit(void);
372 void soabort(struct socket *);
373 int soaccept(struct socket *, struct mbuf *);
374 int sobind(struct socket *, struct mbuf *, struct proc *);
375 void socantrcvmore(struct socket *);
376 void socantsendmore(struct socket *);
377 int soclose(struct socket *, int);
378 int soconnect(struct socket *, struct mbuf *);
379 int soconnect2(struct socket *, struct socket *);
380 int socreate(int, struct socket **, int, int);
381 int sodisconnect(struct socket *);
382 struct socket *soalloc(const struct protosw *, int);
383 void sofree(struct socket *, int);
384 int sogetopt(struct socket *, int, int, struct mbuf *);
385 void sohasoutofband(struct socket *);
386 void soisconnected(struct socket *);
387 void soisconnecting(struct socket *);
388 void soisdisconnected(struct socket *);
389 void soisdisconnecting(struct socket *);
390 int solisten(struct socket *, int);
391 struct socket *sonewconn(struct socket *, int, int);
392 void soqinsque(struct socket *, struct socket *, int);
393 int soqremque(struct socket *, int);
394 int soreceive(struct socket *, struct mbuf **, struct uio *,
395 struct mbuf **, struct mbuf **, int *, socklen_t);
396 int soreserve(struct socket *, u_long, u_long);
397 int sosend(struct socket *, struct mbuf *, struct uio *,
398 struct mbuf *, struct mbuf *, int);
399 int sosetopt(struct socket *, int, int, struct mbuf *);
400 int soshutdown(struct socket *, int);
401 void sowakeup(struct socket *, struct sockbuf *);
402 void sorwakeup(struct socket *);
403 void sowwakeup(struct socket *);
404 int sockargs(struct mbuf **, const void *, size_t, int);
405
406 int sosleep_nsec(struct socket *, void *, int, const char *, uint64_t);
407 void solock(struct socket *);
408 void solock_shared(struct socket *);
409 int solock_persocket(struct socket *);
410 void solock_pair(struct socket *, struct socket *);
411 void sounlock(struct socket *);
412 void sounlock_shared(struct socket *);
413
414 int sendit(struct proc *, int, struct msghdr *, int, register_t *);
415 int recvit(struct proc *, int, struct msghdr *, caddr_t, register_t *);
416 int doaccept(struct proc *, int, struct sockaddr *, socklen_t *, int,
417 register_t *);
418
419 #ifdef SOCKBUF_DEBUG
420 void sblastrecordchk(struct sockbuf *, const char *);
421 #define SBLASTRECORDCHK(sb, where) sblastrecordchk((sb), (where))
422
423 void sblastmbufchk(struct sockbuf *, const char *);
424 #define SBLASTMBUFCHK(sb, where) sblastmbufchk((sb), (where))
425 void sbcheck(struct socket *, struct sockbuf *);
426 #define SBCHECK(so, sb) sbcheck((so), (sb))
427 #else
428 #define SBLASTRECORDCHK(sb, where) /* nothing */
429 #define SBLASTMBUFCHK(sb, where) /* nothing */
430 #define SBCHECK(so, sb) /* nothing */
431 #endif /* SOCKBUF_DEBUG */
432
433 #endif /* _KERNEL */
434
435 #endif /* _SYS_SOCKETVAR_H_ */
436