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 * $FreeBSD$ 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/osd.h> 42 #include <sys/_sx.h> 43 #include <sys/sockbuf.h> 44 #include <sys/sockstate.h> 45 #ifdef _KERNEL 46 #include <sys/caprights.h> 47 #include <sys/sockopt.h> 48 #endif 49 50 struct vnet; 51 52 /* 53 * Kernel structure per socket. 54 * Contains send and receive buffer queues, 55 * handle on protocol and pointer to protocol 56 * private data and error information. 57 */ 58 typedef u_quad_t so_gen_t; 59 60 struct socket; 61 62 /*- 63 * Locking key to struct socket: 64 * (a) constant after allocation, no locking required. 65 * (b) locked by SOCK_LOCK(so). 66 * (c) locked by SOCKBUF_LOCK(&so->so_rcv). 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_int so_qlen; /* (e) number of unaccepted connections */ 98 u_int so_incqlen; /* (e) number of unaccepted incomplete 99 connections */ 100 u_int 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 */ 106 107 struct sockbuf so_rcv, so_snd; 108 109 struct ucred *so_cred; /* (a) user credentials */ 110 struct label *so_label; /* (b) MAC label for socket */ 111 struct label *so_peerlabel; /* (b) cached MAC label for peer */ 112 /* NB: generation count must not be first. */ 113 so_gen_t so_gencnt; /* (h) generation count */ 114 void *so_emuldata; /* (b) private data for emulators */ 115 struct so_accf { 116 struct accept_filter *so_accept_filter; 117 void *so_accept_filter_arg; /* saved filter args */ 118 char *so_accept_filter_str; /* saved user args */ 119 } *so_accf; 120 struct osd osd; /* Object Specific extensions */ 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 void *so_pspare[2]; /* packet pacing / general use */ 131 int so_ispare[2]; /* packet pacing / general use */ 132 }; 133 134 /* 135 * Global accept mutex to serialize access to accept queues and 136 * fields associated with multiple sockets. This allows us to 137 * avoid defining a lock order between listen and accept sockets 138 * until such time as it proves to be a good idea. 139 */ 140 extern struct mtx accept_mtx; 141 #define ACCEPT_LOCK_ASSERT() mtx_assert(&accept_mtx, MA_OWNED) 142 #define ACCEPT_UNLOCK_ASSERT() mtx_assert(&accept_mtx, MA_NOTOWNED) 143 #define ACCEPT_LOCK() mtx_lock(&accept_mtx) 144 #define ACCEPT_UNLOCK() mtx_unlock(&accept_mtx) 145 146 /* 147 * Per-socket mutex: we reuse the receive socket buffer mutex for space 148 * efficiency. This decision should probably be revisited as we optimize 149 * locking for the socket code. 150 */ 151 #define SOCK_MTX(_so) SOCKBUF_MTX(&(_so)->so_rcv) 152 #define SOCK_LOCK(_so) SOCKBUF_LOCK(&(_so)->so_rcv) 153 #define SOCK_OWNED(_so) SOCKBUF_OWNED(&(_so)->so_rcv) 154 #define SOCK_UNLOCK(_so) SOCKBUF_UNLOCK(&(_so)->so_rcv) 155 #define SOCK_LOCK_ASSERT(_so) SOCKBUF_LOCK_ASSERT(&(_so)->so_rcv) 156 157 /* 158 * Socket state bits stored in so_qstate. 159 */ 160 #define SQ_INCOMP 0x0800 /* unaccepted, incomplete connection */ 161 #define SQ_COMP 0x1000 /* unaccepted, complete connection */ 162 163 /* 164 * Externalized form of struct socket used by the sysctl(3) interface. 165 */ 166 struct xsocket { 167 size_t xso_len; /* length of this structure */ 168 struct socket *xso_so; /* makes a convenient handle sometimes */ 169 short so_type; 170 short so_options; 171 short so_linger; 172 short so_state; 173 caddr_t so_pcb; /* another convenient handle */ 174 int xso_protocol; 175 int xso_family; 176 u_int so_qlen; 177 u_int so_incqlen; 178 u_int so_qlimit; 179 short so_timeo; 180 u_short so_error; 181 pid_t so_pgid; 182 u_long so_oobmark; 183 struct xsockbuf so_rcv, so_snd; 184 uid_t so_uid; /* XXX */ 185 }; 186 187 #ifdef _KERNEL 188 189 /* 190 * Macros for sockets and socket buffering. 191 */ 192 193 /* 194 * Flags to sblock(). 195 */ 196 #define SBL_WAIT 0x00000001 /* Wait if not immediately available. */ 197 #define SBL_NOINTR 0x00000002 /* Force non-interruptible sleep. */ 198 #define SBL_VALID (SBL_WAIT | SBL_NOINTR) 199 200 /* 201 * Do we need to notify the other side when I/O is possible? 202 */ 203 #define sb_notify(sb) (((sb)->sb_flags & (SB_WAIT | SB_SEL | SB_ASYNC | \ 204 SB_UPCALL | SB_AIO | SB_KNOTE)) != 0) 205 206 /* do we have to send all at once on a socket? */ 207 #define sosendallatonce(so) \ 208 ((so)->so_proto->pr_flags & PR_ATOMIC) 209 210 /* can we read something from so? */ 211 #define soreadabledata(so) \ 212 (sbavail(&(so)->so_rcv) >= (so)->so_rcv.sb_lowat || \ 213 !TAILQ_EMPTY(&(so)->so_comp) || (so)->so_error) 214 #define soreadable(so) \ 215 (soreadabledata(so) || ((so)->so_rcv.sb_state & SBS_CANTRCVMORE)) 216 217 /* can we write something to so? */ 218 #define sowriteable(so) \ 219 ((sbspace(&(so)->so_snd) >= (so)->so_snd.sb_lowat && \ 220 (((so)->so_state&SS_ISCONNECTED) || \ 221 ((so)->so_proto->pr_flags&PR_CONNREQUIRED)==0)) || \ 222 ((so)->so_snd.sb_state & SBS_CANTSENDMORE) || \ 223 (so)->so_error) 224 225 /* 226 * soref()/sorele() ref-count the socket structure. Note that you must 227 * still explicitly close the socket, but the last ref count will free 228 * the structure. 229 */ 230 #define soref(so) do { \ 231 SOCK_LOCK_ASSERT(so); \ 232 ++(so)->so_count; \ 233 } while (0) 234 235 #define sorele(so) do { \ 236 ACCEPT_LOCK_ASSERT(); \ 237 SOCK_LOCK_ASSERT(so); \ 238 if ((so)->so_count <= 0) \ 239 panic("sorele"); \ 240 if (--(so)->so_count == 0) \ 241 sofree(so); \ 242 else { \ 243 SOCK_UNLOCK(so); \ 244 ACCEPT_UNLOCK(); \ 245 } \ 246 } while (0) 247 248 /* 249 * In sorwakeup() and sowwakeup(), acquire the socket buffer lock to 250 * avoid a non-atomic test-and-wakeup. However, sowakeup is 251 * responsible for releasing the lock if it is called. We unlock only 252 * if we don't call into sowakeup. If any code is introduced that 253 * directly invokes the underlying sowakeup() primitives, it must 254 * maintain the same semantics. 255 */ 256 #define sorwakeup_locked(so) do { \ 257 SOCKBUF_LOCK_ASSERT(&(so)->so_rcv); \ 258 if (sb_notify(&(so)->so_rcv)) \ 259 sowakeup((so), &(so)->so_rcv); \ 260 else \ 261 SOCKBUF_UNLOCK(&(so)->so_rcv); \ 262 } while (0) 263 264 #define sorwakeup(so) do { \ 265 SOCKBUF_LOCK(&(so)->so_rcv); \ 266 sorwakeup_locked(so); \ 267 } while (0) 268 269 #define sowwakeup_locked(so) do { \ 270 SOCKBUF_LOCK_ASSERT(&(so)->so_snd); \ 271 if (sb_notify(&(so)->so_snd)) \ 272 sowakeup((so), &(so)->so_snd); \ 273 else \ 274 SOCKBUF_UNLOCK(&(so)->so_snd); \ 275 } while (0) 276 277 #define sowwakeup(so) do { \ 278 SOCKBUF_LOCK(&(so)->so_snd); \ 279 sowwakeup_locked(so); \ 280 } while (0) 281 282 struct accept_filter { 283 char accf_name[16]; 284 int (*accf_callback) 285 (struct socket *so, void *arg, int waitflag); 286 void * (*accf_create) 287 (struct socket *so, char *arg); 288 void (*accf_destroy) 289 (struct socket *so); 290 SLIST_ENTRY(accept_filter) accf_next; 291 }; 292 293 #ifdef MALLOC_DECLARE 294 MALLOC_DECLARE(M_ACCF); 295 MALLOC_DECLARE(M_PCB); 296 MALLOC_DECLARE(M_SONAME); 297 #endif 298 299 /* 300 * Socket specific helper hook point identifiers 301 * Do not leave holes in the sequence, hook registration is a loop. 302 */ 303 #define HHOOK_SOCKET_OPT 0 304 #define HHOOK_SOCKET_CREATE 1 305 #define HHOOK_SOCKET_RCV 2 306 #define HHOOK_SOCKET_SND 3 307 #define HHOOK_FILT_SOREAD 4 308 #define HHOOK_FILT_SOWRITE 5 309 #define HHOOK_SOCKET_CLOSE 6 310 #define HHOOK_SOCKET_LAST HHOOK_SOCKET_CLOSE 311 312 struct socket_hhook_data { 313 struct socket *so; 314 struct mbuf *m; 315 void *hctx; /* hook point specific data*/ 316 int status; 317 }; 318 319 extern int maxsockets; 320 extern u_long sb_max; 321 extern so_gen_t so_gencnt; 322 323 struct file; 324 struct filedesc; 325 struct mbuf; 326 struct sockaddr; 327 struct ucred; 328 struct uio; 329 330 /* 'which' values for socket upcalls. */ 331 #define SO_RCV 1 332 #define SO_SND 2 333 334 /* Return values for socket upcalls. */ 335 #define SU_OK 0 336 #define SU_ISCONNECTED 1 337 338 /* 339 * From uipc_socket and friends 340 */ 341 int getsockaddr(struct sockaddr **namp, caddr_t uaddr, size_t len); 342 int getsock_cap(struct thread *td, int fd, cap_rights_t *rightsp, 343 struct file **fpp, u_int *fflagp); 344 void soabort(struct socket *so); 345 int soaccept(struct socket *so, struct sockaddr **nam); 346 void soaio_enqueue(struct task *task); 347 void soaio_rcv(void *context, int pending); 348 void soaio_snd(void *context, int pending); 349 int socheckuid(struct socket *so, uid_t uid); 350 int sobind(struct socket *so, struct sockaddr *nam, struct thread *td); 351 int sobindat(int fd, struct socket *so, struct sockaddr *nam, 352 struct thread *td); 353 int soclose(struct socket *so); 354 int soconnect(struct socket *so, struct sockaddr *nam, struct thread *td); 355 int soconnectat(int fd, struct socket *so, struct sockaddr *nam, 356 struct thread *td); 357 int soconnect2(struct socket *so1, struct socket *so2); 358 int socreate(int dom, struct socket **aso, int type, int proto, 359 struct ucred *cred, struct thread *td); 360 int sodisconnect(struct socket *so); 361 struct sockaddr *sodupsockaddr(const struct sockaddr *sa, int mflags); 362 void sofree(struct socket *so); 363 void sohasoutofband(struct socket *so); 364 int solisten(struct socket *so, int backlog, struct thread *td); 365 void solisten_proto(struct socket *so, int backlog); 366 int solisten_proto_check(struct socket *so); 367 struct socket * 368 sonewconn(struct socket *head, int connstatus); 369 370 371 int sopoll(struct socket *so, int events, struct ucred *active_cred, 372 struct thread *td); 373 int sopoll_generic(struct socket *so, int events, 374 struct ucred *active_cred, struct thread *td); 375 int soreceive(struct socket *so, struct sockaddr **paddr, struct uio *uio, 376 struct mbuf **mp0, struct mbuf **controlp, int *flagsp); 377 int soreceive_stream(struct socket *so, struct sockaddr **paddr, 378 struct uio *uio, struct mbuf **mp0, struct mbuf **controlp, 379 int *flagsp); 380 int soreceive_dgram(struct socket *so, struct sockaddr **paddr, 381 struct uio *uio, struct mbuf **mp0, struct mbuf **controlp, 382 int *flagsp); 383 int soreceive_generic(struct socket *so, struct sockaddr **paddr, 384 struct uio *uio, struct mbuf **mp0, struct mbuf **controlp, 385 int *flagsp); 386 int soreserve(struct socket *so, u_long sndcc, u_long rcvcc); 387 void sorflush(struct socket *so); 388 int sosend(struct socket *so, struct sockaddr *addr, struct uio *uio, 389 struct mbuf *top, struct mbuf *control, int flags, 390 struct thread *td); 391 int sosend_dgram(struct socket *so, struct sockaddr *addr, 392 struct uio *uio, struct mbuf *top, struct mbuf *control, 393 int flags, struct thread *td); 394 int sosend_generic(struct socket *so, struct sockaddr *addr, 395 struct uio *uio, struct mbuf *top, struct mbuf *control, 396 int flags, struct thread *td); 397 int soshutdown(struct socket *so, int how); 398 void sotoxsocket(struct socket *so, struct xsocket *xso); 399 void soupcall_clear(struct socket *so, int which); 400 void soupcall_set(struct socket *so, int which, 401 int (*func)(struct socket *, void *, int), void *arg); 402 void sowakeup(struct socket *so, struct sockbuf *sb); 403 void sowakeup_aio(struct socket *so, struct sockbuf *sb); 404 int selsocket(struct socket *so, int events, struct timeval *tv, 405 struct thread *td); 406 407 /* 408 * Accept filter functions (duh). 409 */ 410 int accept_filt_add(struct accept_filter *filt); 411 int accept_filt_del(char *name); 412 struct accept_filter *accept_filt_get(char *name); 413 #ifdef ACCEPT_FILTER_MOD 414 #ifdef SYSCTL_DECL 415 SYSCTL_DECL(_net_inet_accf); 416 #endif 417 int accept_filt_generic_mod_event(module_t mod, int event, void *data); 418 #endif 419 420 #endif /* _KERNEL */ 421 422 #endif /* !_SYS_SOCKETVAR_H_ */ 423