1/* $OpenBSD: socketvar.h,v 1.131 2024/05/17 19:11:14 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_ 49typedef __socklen_t socklen_t; /* length type for network syscalls */ 50#endif 51 52TAILQ_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 */ 60struct 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_WAIT 0x0001 /* someone is waiting for data/space */ 132#define SB_ASYNC 0x0002 /* ASYNC I/O, need signals */ 133#define SB_SPLICE 0x0004 /* buffer is splice source or drain */ 134#define SB_NOINTR 0x0008 /* operations not interruptible */ 135#define SB_MTXLOCK 0x0010 /* sblock() doesn't need solock() */ 136 137 void (*so_upcall)(struct socket *so, caddr_t arg, int waitf); 138 caddr_t so_upcallarg; /* Arg for above */ 139 uid_t so_euid, so_ruid; /* who opened the socket */ 140 gid_t so_egid, so_rgid; 141 pid_t so_cpid; /* pid of process that opened socket */ 142}; 143 144/* 145 * Socket state bits. 146 * 147 * NOTE: The following states should be used with corresponding socket's 148 * buffer `sb_state' only: 149 * 150 * SS_CANTSENDMORE with `so_snd' 151 * SS_ISSENDING with `so_snd' 152 * SS_CANTRCVMORE with `so_rcv' 153 * SS_RCVATMARK with `so_rcv' 154 */ 155 156#define SS_NOFDREF 0x001 /* no file table ref any more */ 157#define SS_ISCONNECTED 0x002 /* socket connected to a peer */ 158#define SS_ISCONNECTING 0x004 /* in process of connecting to peer */ 159#define SS_ISDISCONNECTING 0x008 /* in process of disconnecting */ 160#define SS_CANTSENDMORE 0x010 /* can't send more data to peer */ 161#define SS_CANTRCVMORE 0x020 /* can't receive more data from peer */ 162#define SS_RCVATMARK 0x040 /* at mark on input */ 163#define SS_ISDISCONNECTED 0x800 /* socket disconnected from peer */ 164 165#define SS_PRIV 0x080 /* privileged for broadcast, raw... */ 166#define SS_CONNECTOUT 0x1000 /* connect, not accept, at this end */ 167#define SS_ISSENDING 0x2000 /* hint for lower layer */ 168#define SS_DNS 0x4000 /* created using SOCK_DNS socket(2) */ 169#define SS_YP 0x8000 /* created using ypconnect(2) */ 170 171#ifdef _KERNEL 172 173#include <sys/protosw.h> 174#include <lib/libkern/libkern.h> 175 176void soassertlocked(struct socket *); 177void soassertlocked_readonly(struct socket *); 178 179static inline void 180soref(struct socket *so) 181{ 182 refcnt_take(&so->so_refcnt); 183} 184 185static inline void 186sorele(struct socket *so) 187{ 188 refcnt_rele_wake(&so->so_refcnt); 189} 190 191/* 192 * Macros for sockets and socket buffering. 193 */ 194 195#define isspliced(so) ((so)->so_sp && (so)->so_sp->ssp_socket) 196#define issplicedback(so) ((so)->so_sp && (so)->so_sp->ssp_soback) 197 198static inline void 199sb_mtx_lock(struct sockbuf *sb) 200{ 201 if (sb->sb_flags & SB_MTXLOCK) 202 mtx_enter(&sb->sb_mtx); 203} 204 205static inline void 206sb_mtx_unlock(struct sockbuf *sb) 207{ 208 if (sb->sb_flags & SB_MTXLOCK) 209 mtx_leave(&sb->sb_mtx); 210} 211 212void sbmtxassertlocked(struct socket *so, struct sockbuf *); 213 214/* 215 * Do we need to notify the other side when I/O is possible? 216 */ 217static inline int 218sb_notify(struct socket *so, struct sockbuf *sb) 219{ 220 int rv; 221 222 soassertlocked(so); 223 224 mtx_enter(&sb->sb_mtx); 225 rv = ((sb->sb_flags & (SB_WAIT|SB_ASYNC|SB_SPLICE)) != 0 || 226 !klist_empty(&sb->sb_klist)); 227 mtx_leave(&sb->sb_mtx); 228 229 return rv; 230} 231 232/* 233 * How much space is there in a socket buffer (so->so_snd or so->so_rcv)? 234 * This is problematical if the fields are unsigned, as the space might 235 * still be negative (cc > hiwat or mbcnt > mbmax). Should detect 236 * overflow and return 0. 237 */ 238 239static inline long 240sbspace(struct socket *so, struct sockbuf *sb) 241{ 242 if (sb->sb_flags & SB_MTXLOCK) 243 sbmtxassertlocked(so, sb); 244 else 245 soassertlocked_readonly(so); 246 247 return lmin(sb->sb_hiwat - sb->sb_cc, sb->sb_mbmax - sb->sb_mbcnt); 248} 249 250/* do we have to send all at once on a socket? */ 251#define sosendallatonce(so) \ 252 ((so)->so_proto->pr_flags & PR_ATOMIC) 253 254/* are we sending on this socket? */ 255#define soissending(so) \ 256 ((so)->so_snd.sb_state & SS_ISSENDING) 257 258/* can we read something from so? */ 259static inline int 260soreadable(struct socket *so) 261{ 262 soassertlocked_readonly(so); 263 if (isspliced(so)) 264 return 0; 265 return (so->so_rcv.sb_state & SS_CANTRCVMORE) || so->so_qlen || 266 so->so_error || so->so_rcv.sb_cc >= so->so_rcv.sb_lowat; 267} 268 269/* can we write something to so? */ 270static inline int 271sowriteable(struct socket *so) 272{ 273 soassertlocked_readonly(so); 274 return ((sbspace(so, &so->so_snd) >= so->so_snd.sb_lowat && 275 ((so->so_state & SS_ISCONNECTED) || 276 (so->so_proto->pr_flags & PR_CONNREQUIRED)==0)) || 277 (so->so_snd.sb_state & SS_CANTSENDMORE) || so->so_error); 278} 279 280/* adjust counters in sb reflecting allocation of m */ 281static inline void 282sballoc(struct socket *so, struct sockbuf *sb, struct mbuf *m) 283{ 284 sb->sb_cc += m->m_len; 285 if (m->m_type != MT_CONTROL && m->m_type != MT_SONAME) 286 sb->sb_datacc += m->m_len; 287 sb->sb_mbcnt += MSIZE; 288 if (m->m_flags & M_EXT) 289 sb->sb_mbcnt += m->m_ext.ext_size; 290} 291 292/* adjust counters in sb reflecting freeing of m */ 293static inline void 294sbfree(struct socket *so, struct sockbuf *sb, struct mbuf *m) 295{ 296 sb->sb_cc -= m->m_len; 297 if (m->m_type != MT_CONTROL && m->m_type != MT_SONAME) 298 sb->sb_datacc -= m->m_len; 299 sb->sb_mbcnt -= MSIZE; 300 if (m->m_flags & M_EXT) 301 sb->sb_mbcnt -= m->m_ext.ext_size; 302} 303 304/* 305 * Flags to sblock() 306 */ 307#define SBL_WAIT 0x01 /* Wait if lock not immediately available. */ 308#define SBL_NOINTR 0x02 /* Enforce non-interruptible sleep. */ 309 310/* 311 * Set lock on sockbuf sb; sleep if lock is already held. 312 * Unless SB_NOINTR is set on sockbuf or SBL_NOINTR passed, 313 * sleep is interruptible. Returns error without lock if 314 * sleep is interrupted. 315 */ 316int sblock(struct sockbuf *, int); 317 318/* release lock on sockbuf sb */ 319void sbunlock(struct sockbuf *); 320 321#define SB_EMPTY_FIXUP(sb) do { \ 322 if ((sb)->sb_mb == NULL) { \ 323 (sb)->sb_mbtail = NULL; \ 324 (sb)->sb_lastrecord = NULL; \ 325 } \ 326} while (/*CONSTCOND*/0) 327 328extern u_long sb_max; 329 330extern struct pool socket_pool; 331 332struct mbuf; 333struct sockaddr; 334struct proc; 335struct msghdr; 336struct stat; 337struct knote; 338 339/* 340 * File operations on sockets. 341 */ 342int soo_read(struct file *, struct uio *, int); 343int soo_write(struct file *, struct uio *, int); 344int soo_ioctl(struct file *, u_long, caddr_t, struct proc *); 345int soo_kqfilter(struct file *, struct knote *); 346int soo_close(struct file *, struct proc *); 347int soo_stat(struct file *, struct stat *, struct proc *); 348void sbappend(struct socket *, struct sockbuf *, struct mbuf *); 349void sbappendstream(struct socket *, struct sockbuf *, struct mbuf *); 350int sbappendaddr(struct socket *, struct sockbuf *, 351 const struct sockaddr *, struct mbuf *, struct mbuf *); 352int sbappendcontrol(struct socket *, struct sockbuf *, struct mbuf *, 353 struct mbuf *); 354void sbappendrecord(struct socket *, struct sockbuf *, struct mbuf *); 355void sbcompress(struct socket *, struct sockbuf *, struct mbuf *, 356 struct mbuf *); 357struct mbuf * 358 sbcreatecontrol(const void *, size_t, int, int); 359void sbdrop(struct socket *, struct sockbuf *, int); 360void sbdroprecord(struct socket *, struct sockbuf *); 361void sbflush(struct socket *, struct sockbuf *); 362void sbrelease(struct socket *, struct sockbuf *); 363int sbcheckreserve(u_long, u_long); 364int sbchecklowmem(void); 365int sbreserve(struct socket *, struct sockbuf *, u_long); 366int sbwait(struct socket *, struct sockbuf *); 367void soinit(void); 368void soabort(struct socket *); 369int soaccept(struct socket *, struct mbuf *); 370int sobind(struct socket *, struct mbuf *, struct proc *); 371void socantrcvmore(struct socket *); 372void socantsendmore(struct socket *); 373int soclose(struct socket *, int); 374int soconnect(struct socket *, struct mbuf *); 375int soconnect2(struct socket *, struct socket *); 376int socreate(int, struct socket **, int, int); 377int sodisconnect(struct socket *); 378struct socket *soalloc(const struct protosw *, int); 379void sofree(struct socket *, int); 380int sogetopt(struct socket *, int, int, struct mbuf *); 381void sohasoutofband(struct socket *); 382void soisconnected(struct socket *); 383void soisconnecting(struct socket *); 384void soisdisconnected(struct socket *); 385void soisdisconnecting(struct socket *); 386int solisten(struct socket *, int); 387struct socket *sonewconn(struct socket *, int, int); 388void soqinsque(struct socket *, struct socket *, int); 389int soqremque(struct socket *, int); 390int soreceive(struct socket *, struct mbuf **, struct uio *, 391 struct mbuf **, struct mbuf **, int *, socklen_t); 392int soreserve(struct socket *, u_long, u_long); 393int sosend(struct socket *, struct mbuf *, struct uio *, 394 struct mbuf *, struct mbuf *, int); 395int sosetopt(struct socket *, int, int, struct mbuf *); 396int soshutdown(struct socket *, int); 397void sowakeup(struct socket *, struct sockbuf *); 398void sorwakeup(struct socket *); 399void sowwakeup(struct socket *); 400int sockargs(struct mbuf **, const void *, size_t, int); 401 402int sosleep_nsec(struct socket *, void *, int, const char *, uint64_t); 403void solock(struct socket *); 404void solock_shared(struct socket *); 405int solock_persocket(struct socket *); 406void solock_pair(struct socket *, struct socket *); 407void sounlock(struct socket *); 408void sounlock_shared(struct socket *); 409 410int sendit(struct proc *, int, struct msghdr *, int, register_t *); 411int recvit(struct proc *, int, struct msghdr *, caddr_t, register_t *); 412int doaccept(struct proc *, int, struct sockaddr *, socklen_t *, int, 413 register_t *); 414 415#ifdef SOCKBUF_DEBUG 416void sblastrecordchk(struct sockbuf *, const char *); 417#define SBLASTRECORDCHK(sb, where) sblastrecordchk((sb), (where)) 418 419void sblastmbufchk(struct sockbuf *, const char *); 420#define SBLASTMBUFCHK(sb, where) sblastmbufchk((sb), (where)) 421void sbcheck(struct socket *, struct sockbuf *); 422#define SBCHECK(so, sb) sbcheck((so), (sb)) 423#else 424#define SBLASTRECORDCHK(sb, where) /* nothing */ 425#define SBLASTMBUFCHK(sb, where) /* nothing */ 426#define SBCHECK(so, sb) /* nothing */ 427#endif /* SOCKBUF_DEBUG */ 428 429#endif /* _KERNEL */ 430 431#endif /* _SYS_SOCKETVAR_H_ */ 432