1 /* 2 * net/tipc/socket.c: TIPC socket API 3 * 4 * Copyright (c) 2001-2007, Ericsson AB 5 * Copyright (c) 2004-2008, 2010-2011, Wind River Systems 6 * 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 are met: 10 * 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 names of the copyright holders nor the names of its 17 * contributors may be used to endorse or promote products derived from 18 * this software without specific prior written permission. 19 * 20 * Alternatively, this software may be distributed under the terms of the 21 * GNU General Public License ("GPL") version 2 as published by the Free 22 * Software Foundation. 23 * 24 * THIS SOFTWARE IS PROVIDED BY THE COPYRIGHT HOLDERS AND CONTRIBUTORS "AS IS" 25 * AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 26 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 27 * ARE DISCLAIMED. IN NO EVENT SHALL THE COPYRIGHT OWNER OR CONTRIBUTORS BE 28 * LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 29 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 30 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 31 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 32 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 33 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 34 * POSSIBILITY OF SUCH DAMAGE. 35 */ 36 37 #include <net/sock.h> 38 39 #include <linux/tipc.h> 40 #include <linux/tipc_config.h> 41 42 #include "core.h" 43 #include "port.h" 44 45 #define SS_LISTENING -1 /* socket is listening */ 46 #define SS_READY -2 /* socket is connectionless */ 47 48 #define OVERLOAD_LIMIT_BASE 5000 49 #define CONN_TIMEOUT_DEFAULT 8000 /* default connect timeout = 8s */ 50 51 struct tipc_sock { 52 struct sock sk; 53 struct tipc_port *p; 54 struct tipc_portid peer_name; 55 long conn_timeout; 56 }; 57 58 #define tipc_sk(sk) ((struct tipc_sock *)(sk)) 59 #define tipc_sk_port(sk) ((struct tipc_port *)(tipc_sk(sk)->p)) 60 61 #define tipc_rx_ready(sock) (!skb_queue_empty(&sock->sk->sk_receive_queue) || \ 62 (sock->state == SS_DISCONNECTING)) 63 64 static int backlog_rcv(struct sock *sk, struct sk_buff *skb); 65 static u32 dispatch(struct tipc_port *tport, struct sk_buff *buf); 66 static void wakeupdispatch(struct tipc_port *tport); 67 68 static const struct proto_ops packet_ops; 69 static const struct proto_ops stream_ops; 70 static const struct proto_ops msg_ops; 71 72 static struct proto tipc_proto; 73 74 static int sockets_enabled; 75 76 static atomic_t tipc_queue_size = ATOMIC_INIT(0); 77 78 /* 79 * Revised TIPC socket locking policy: 80 * 81 * Most socket operations take the standard socket lock when they start 82 * and hold it until they finish (or until they need to sleep). Acquiring 83 * this lock grants the owner exclusive access to the fields of the socket 84 * data structures, with the exception of the backlog queue. A few socket 85 * operations can be done without taking the socket lock because they only 86 * read socket information that never changes during the life of the socket. 87 * 88 * Socket operations may acquire the lock for the associated TIPC port if they 89 * need to perform an operation on the port. If any routine needs to acquire 90 * both the socket lock and the port lock it must take the socket lock first 91 * to avoid the risk of deadlock. 92 * 93 * The dispatcher handling incoming messages cannot grab the socket lock in 94 * the standard fashion, since invoked it runs at the BH level and cannot block. 95 * Instead, it checks to see if the socket lock is currently owned by someone, 96 * and either handles the message itself or adds it to the socket's backlog 97 * queue; in the latter case the queued message is processed once the process 98 * owning the socket lock releases it. 99 * 100 * NOTE: Releasing the socket lock while an operation is sleeping overcomes 101 * the problem of a blocked socket operation preventing any other operations 102 * from occurring. However, applications must be careful if they have 103 * multiple threads trying to send (or receive) on the same socket, as these 104 * operations might interfere with each other. For example, doing a connect 105 * and a receive at the same time might allow the receive to consume the 106 * ACK message meant for the connect. While additional work could be done 107 * to try and overcome this, it doesn't seem to be worthwhile at the present. 108 * 109 * NOTE: Releasing the socket lock while an operation is sleeping also ensures 110 * that another operation that must be performed in a non-blocking manner is 111 * not delayed for very long because the lock has already been taken. 112 * 113 * NOTE: This code assumes that certain fields of a port/socket pair are 114 * constant over its lifetime; such fields can be examined without taking 115 * the socket lock and/or port lock, and do not need to be re-read even 116 * after resuming processing after waiting. These fields include: 117 * - socket type 118 * - pointer to socket sk structure (aka tipc_sock structure) 119 * - pointer to port structure 120 * - port reference 121 */ 122 123 /** 124 * advance_rx_queue - discard first buffer in socket receive queue 125 * 126 * Caller must hold socket lock 127 */ 128 129 static void advance_rx_queue(struct sock *sk) 130 { 131 buf_discard(__skb_dequeue(&sk->sk_receive_queue)); 132 atomic_dec(&tipc_queue_size); 133 } 134 135 /** 136 * discard_rx_queue - discard all buffers in socket receive queue 137 * 138 * Caller must hold socket lock 139 */ 140 141 static void discard_rx_queue(struct sock *sk) 142 { 143 struct sk_buff *buf; 144 145 while ((buf = __skb_dequeue(&sk->sk_receive_queue))) { 146 atomic_dec(&tipc_queue_size); 147 buf_discard(buf); 148 } 149 } 150 151 /** 152 * reject_rx_queue - reject all buffers in socket receive queue 153 * 154 * Caller must hold socket lock 155 */ 156 157 static void reject_rx_queue(struct sock *sk) 158 { 159 struct sk_buff *buf; 160 161 while ((buf = __skb_dequeue(&sk->sk_receive_queue))) { 162 tipc_reject_msg(buf, TIPC_ERR_NO_PORT); 163 atomic_dec(&tipc_queue_size); 164 } 165 } 166 167 /** 168 * tipc_create - create a TIPC socket 169 * @net: network namespace (must be default network) 170 * @sock: pre-allocated socket structure 171 * @protocol: protocol indicator (must be 0) 172 * @kern: caused by kernel or by userspace? 173 * 174 * This routine creates additional data structures used by the TIPC socket, 175 * initializes them, and links them together. 176 * 177 * Returns 0 on success, errno otherwise 178 */ 179 180 static int tipc_create(struct net *net, struct socket *sock, int protocol, 181 int kern) 182 { 183 const struct proto_ops *ops; 184 socket_state state; 185 struct sock *sk; 186 struct tipc_port *tp_ptr; 187 188 /* Validate arguments */ 189 190 if (!net_eq(net, &init_net)) 191 return -EAFNOSUPPORT; 192 193 if (unlikely(protocol != 0)) 194 return -EPROTONOSUPPORT; 195 196 switch (sock->type) { 197 case SOCK_STREAM: 198 ops = &stream_ops; 199 state = SS_UNCONNECTED; 200 break; 201 case SOCK_SEQPACKET: 202 ops = &packet_ops; 203 state = SS_UNCONNECTED; 204 break; 205 case SOCK_DGRAM: 206 case SOCK_RDM: 207 ops = &msg_ops; 208 state = SS_READY; 209 break; 210 default: 211 return -EPROTOTYPE; 212 } 213 214 /* Allocate socket's protocol area */ 215 216 sk = sk_alloc(net, AF_TIPC, GFP_KERNEL, &tipc_proto); 217 if (sk == NULL) 218 return -ENOMEM; 219 220 /* Allocate TIPC port for socket to use */ 221 222 tp_ptr = tipc_createport_raw(sk, &dispatch, &wakeupdispatch, 223 TIPC_LOW_IMPORTANCE); 224 if (unlikely(!tp_ptr)) { 225 sk_free(sk); 226 return -ENOMEM; 227 } 228 229 /* Finish initializing socket data structures */ 230 231 sock->ops = ops; 232 sock->state = state; 233 234 sock_init_data(sock, sk); 235 sk->sk_backlog_rcv = backlog_rcv; 236 tipc_sk(sk)->p = tp_ptr; 237 tipc_sk(sk)->conn_timeout = msecs_to_jiffies(CONN_TIMEOUT_DEFAULT); 238 239 spin_unlock_bh(tp_ptr->lock); 240 241 if (sock->state == SS_READY) { 242 tipc_set_portunreturnable(tp_ptr->ref, 1); 243 if (sock->type == SOCK_DGRAM) 244 tipc_set_portunreliable(tp_ptr->ref, 1); 245 } 246 247 return 0; 248 } 249 250 /** 251 * release - destroy a TIPC socket 252 * @sock: socket to destroy 253 * 254 * This routine cleans up any messages that are still queued on the socket. 255 * For DGRAM and RDM socket types, all queued messages are rejected. 256 * For SEQPACKET and STREAM socket types, the first message is rejected 257 * and any others are discarded. (If the first message on a STREAM socket 258 * is partially-read, it is discarded and the next one is rejected instead.) 259 * 260 * NOTE: Rejected messages are not necessarily returned to the sender! They 261 * are returned or discarded according to the "destination droppable" setting 262 * specified for the message by the sender. 263 * 264 * Returns 0 on success, errno otherwise 265 */ 266 267 static int release(struct socket *sock) 268 { 269 struct sock *sk = sock->sk; 270 struct tipc_port *tport; 271 struct sk_buff *buf; 272 int res; 273 274 /* 275 * Exit if socket isn't fully initialized (occurs when a failed accept() 276 * releases a pre-allocated child socket that was never used) 277 */ 278 279 if (sk == NULL) 280 return 0; 281 282 tport = tipc_sk_port(sk); 283 lock_sock(sk); 284 285 /* 286 * Reject all unreceived messages, except on an active connection 287 * (which disconnects locally & sends a 'FIN+' to peer) 288 */ 289 290 while (sock->state != SS_DISCONNECTING) { 291 buf = __skb_dequeue(&sk->sk_receive_queue); 292 if (buf == NULL) 293 break; 294 atomic_dec(&tipc_queue_size); 295 if (TIPC_SKB_CB(buf)->handle != 0) 296 buf_discard(buf); 297 else { 298 if ((sock->state == SS_CONNECTING) || 299 (sock->state == SS_CONNECTED)) { 300 sock->state = SS_DISCONNECTING; 301 tipc_disconnect(tport->ref); 302 } 303 tipc_reject_msg(buf, TIPC_ERR_NO_PORT); 304 } 305 } 306 307 /* 308 * Delete TIPC port; this ensures no more messages are queued 309 * (also disconnects an active connection & sends a 'FIN-' to peer) 310 */ 311 312 res = tipc_deleteport(tport->ref); 313 314 /* Discard any remaining (connection-based) messages in receive queue */ 315 316 discard_rx_queue(sk); 317 318 /* Reject any messages that accumulated in backlog queue */ 319 320 sock->state = SS_DISCONNECTING; 321 release_sock(sk); 322 323 sock_put(sk); 324 sock->sk = NULL; 325 326 return res; 327 } 328 329 /** 330 * bind - associate or disassocate TIPC name(s) with a socket 331 * @sock: socket structure 332 * @uaddr: socket address describing name(s) and desired operation 333 * @uaddr_len: size of socket address data structure 334 * 335 * Name and name sequence binding is indicated using a positive scope value; 336 * a negative scope value unbinds the specified name. Specifying no name 337 * (i.e. a socket address length of 0) unbinds all names from the socket. 338 * 339 * Returns 0 on success, errno otherwise 340 * 341 * NOTE: This routine doesn't need to take the socket lock since it doesn't 342 * access any non-constant socket information. 343 */ 344 345 static int bind(struct socket *sock, struct sockaddr *uaddr, int uaddr_len) 346 { 347 struct sockaddr_tipc *addr = (struct sockaddr_tipc *)uaddr; 348 u32 portref = tipc_sk_port(sock->sk)->ref; 349 350 if (unlikely(!uaddr_len)) 351 return tipc_withdraw(portref, 0, NULL); 352 353 if (uaddr_len < sizeof(struct sockaddr_tipc)) 354 return -EINVAL; 355 if (addr->family != AF_TIPC) 356 return -EAFNOSUPPORT; 357 358 if (addr->addrtype == TIPC_ADDR_NAME) 359 addr->addr.nameseq.upper = addr->addr.nameseq.lower; 360 else if (addr->addrtype != TIPC_ADDR_NAMESEQ) 361 return -EAFNOSUPPORT; 362 363 return (addr->scope > 0) ? 364 tipc_publish(portref, addr->scope, &addr->addr.nameseq) : 365 tipc_withdraw(portref, -addr->scope, &addr->addr.nameseq); 366 } 367 368 /** 369 * get_name - get port ID of socket or peer socket 370 * @sock: socket structure 371 * @uaddr: area for returned socket address 372 * @uaddr_len: area for returned length of socket address 373 * @peer: 0 = own ID, 1 = current peer ID, 2 = current/former peer ID 374 * 375 * Returns 0 on success, errno otherwise 376 * 377 * NOTE: This routine doesn't need to take the socket lock since it only 378 * accesses socket information that is unchanging (or which changes in 379 * a completely predictable manner). 380 */ 381 382 static int get_name(struct socket *sock, struct sockaddr *uaddr, 383 int *uaddr_len, int peer) 384 { 385 struct sockaddr_tipc *addr = (struct sockaddr_tipc *)uaddr; 386 struct tipc_sock *tsock = tipc_sk(sock->sk); 387 388 memset(addr, 0, sizeof(*addr)); 389 if (peer) { 390 if ((sock->state != SS_CONNECTED) && 391 ((peer != 2) || (sock->state != SS_DISCONNECTING))) 392 return -ENOTCONN; 393 addr->addr.id.ref = tsock->peer_name.ref; 394 addr->addr.id.node = tsock->peer_name.node; 395 } else { 396 addr->addr.id.ref = tsock->p->ref; 397 addr->addr.id.node = tipc_own_addr; 398 } 399 400 *uaddr_len = sizeof(*addr); 401 addr->addrtype = TIPC_ADDR_ID; 402 addr->family = AF_TIPC; 403 addr->scope = 0; 404 addr->addr.name.domain = 0; 405 406 return 0; 407 } 408 409 /** 410 * poll - read and possibly block on pollmask 411 * @file: file structure associated with the socket 412 * @sock: socket for which to calculate the poll bits 413 * @wait: ??? 414 * 415 * Returns pollmask value 416 * 417 * COMMENTARY: 418 * It appears that the usual socket locking mechanisms are not useful here 419 * since the pollmask info is potentially out-of-date the moment this routine 420 * exits. TCP and other protocols seem to rely on higher level poll routines 421 * to handle any preventable race conditions, so TIPC will do the same ... 422 * 423 * TIPC sets the returned events as follows: 424 * 425 * socket state flags set 426 * ------------ --------- 427 * unconnected no read flags 428 * no write flags 429 * 430 * connecting POLLIN/POLLRDNORM if ACK/NACK in rx queue 431 * no write flags 432 * 433 * connected POLLIN/POLLRDNORM if data in rx queue 434 * POLLOUT if port is not congested 435 * 436 * disconnecting POLLIN/POLLRDNORM/POLLHUP 437 * no write flags 438 * 439 * listening POLLIN if SYN in rx queue 440 * no write flags 441 * 442 * ready POLLIN/POLLRDNORM if data in rx queue 443 * [connectionless] POLLOUT (since port cannot be congested) 444 * 445 * IMPORTANT: The fact that a read or write operation is indicated does NOT 446 * imply that the operation will succeed, merely that it should be performed 447 * and will not block. 448 */ 449 450 static unsigned int poll(struct file *file, struct socket *sock, 451 poll_table *wait) 452 { 453 struct sock *sk = sock->sk; 454 u32 mask = 0; 455 456 poll_wait(file, sk_sleep(sk), wait); 457 458 switch ((int)sock->state) { 459 case SS_READY: 460 case SS_CONNECTED: 461 if (!tipc_sk_port(sk)->congested) 462 mask |= POLLOUT; 463 /* fall thru' */ 464 case SS_CONNECTING: 465 case SS_LISTENING: 466 if (!skb_queue_empty(&sk->sk_receive_queue)) 467 mask |= (POLLIN | POLLRDNORM); 468 break; 469 case SS_DISCONNECTING: 470 mask = (POLLIN | POLLRDNORM | POLLHUP); 471 break; 472 } 473 474 return mask; 475 } 476 477 /** 478 * dest_name_check - verify user is permitted to send to specified port name 479 * @dest: destination address 480 * @m: descriptor for message to be sent 481 * 482 * Prevents restricted configuration commands from being issued by 483 * unauthorized users. 484 * 485 * Returns 0 if permission is granted, otherwise errno 486 */ 487 488 static int dest_name_check(struct sockaddr_tipc *dest, struct msghdr *m) 489 { 490 struct tipc_cfg_msg_hdr hdr; 491 492 if (likely(dest->addr.name.name.type >= TIPC_RESERVED_TYPES)) 493 return 0; 494 if (likely(dest->addr.name.name.type == TIPC_TOP_SRV)) 495 return 0; 496 if (likely(dest->addr.name.name.type != TIPC_CFG_SRV)) 497 return -EACCES; 498 499 if (!m->msg_iovlen || (m->msg_iov[0].iov_len < sizeof(hdr))) 500 return -EMSGSIZE; 501 if (copy_from_user(&hdr, m->msg_iov[0].iov_base, sizeof(hdr))) 502 return -EFAULT; 503 if ((ntohs(hdr.tcm_type) & 0xC000) && (!capable(CAP_NET_ADMIN))) 504 return -EACCES; 505 506 return 0; 507 } 508 509 /** 510 * send_msg - send message in connectionless manner 511 * @iocb: if NULL, indicates that socket lock is already held 512 * @sock: socket structure 513 * @m: message to send 514 * @total_len: length of message 515 * 516 * Message must have an destination specified explicitly. 517 * Used for SOCK_RDM and SOCK_DGRAM messages, 518 * and for 'SYN' messages on SOCK_SEQPACKET and SOCK_STREAM connections. 519 * (Note: 'SYN+' is prohibited on SOCK_STREAM.) 520 * 521 * Returns the number of bytes sent on success, or errno otherwise 522 */ 523 524 static int send_msg(struct kiocb *iocb, struct socket *sock, 525 struct msghdr *m, size_t total_len) 526 { 527 struct sock *sk = sock->sk; 528 struct tipc_port *tport = tipc_sk_port(sk); 529 struct sockaddr_tipc *dest = (struct sockaddr_tipc *)m->msg_name; 530 int needs_conn; 531 int res = -EINVAL; 532 533 if (unlikely(!dest)) 534 return -EDESTADDRREQ; 535 if (unlikely((m->msg_namelen < sizeof(*dest)) || 536 (dest->family != AF_TIPC))) 537 return -EINVAL; 538 if ((total_len > TIPC_MAX_USER_MSG_SIZE) || 539 (m->msg_iovlen > (unsigned)INT_MAX)) 540 return -EMSGSIZE; 541 542 if (iocb) 543 lock_sock(sk); 544 545 needs_conn = (sock->state != SS_READY); 546 if (unlikely(needs_conn)) { 547 if (sock->state == SS_LISTENING) { 548 res = -EPIPE; 549 goto exit; 550 } 551 if (sock->state != SS_UNCONNECTED) { 552 res = -EISCONN; 553 goto exit; 554 } 555 if ((tport->published) || 556 ((sock->type == SOCK_STREAM) && (total_len != 0))) { 557 res = -EOPNOTSUPP; 558 goto exit; 559 } 560 if (dest->addrtype == TIPC_ADDR_NAME) { 561 tport->conn_type = dest->addr.name.name.type; 562 tport->conn_instance = dest->addr.name.name.instance; 563 } 564 565 /* Abort any pending connection attempts (very unlikely) */ 566 567 reject_rx_queue(sk); 568 } 569 570 do { 571 if (dest->addrtype == TIPC_ADDR_NAME) { 572 res = dest_name_check(dest, m); 573 if (res) 574 break; 575 res = tipc_send2name(tport->ref, 576 &dest->addr.name.name, 577 dest->addr.name.domain, 578 m->msg_iovlen, 579 m->msg_iov, 580 total_len); 581 } else if (dest->addrtype == TIPC_ADDR_ID) { 582 res = tipc_send2port(tport->ref, 583 &dest->addr.id, 584 m->msg_iovlen, 585 m->msg_iov, 586 total_len); 587 } else if (dest->addrtype == TIPC_ADDR_MCAST) { 588 if (needs_conn) { 589 res = -EOPNOTSUPP; 590 break; 591 } 592 res = dest_name_check(dest, m); 593 if (res) 594 break; 595 res = tipc_multicast(tport->ref, 596 &dest->addr.nameseq, 597 m->msg_iovlen, 598 m->msg_iov, 599 total_len); 600 } 601 if (likely(res != -ELINKCONG)) { 602 if (needs_conn && (res >= 0)) 603 sock->state = SS_CONNECTING; 604 break; 605 } 606 if (m->msg_flags & MSG_DONTWAIT) { 607 res = -EWOULDBLOCK; 608 break; 609 } 610 release_sock(sk); 611 res = wait_event_interruptible(*sk_sleep(sk), 612 !tport->congested); 613 lock_sock(sk); 614 if (res) 615 break; 616 } while (1); 617 618 exit: 619 if (iocb) 620 release_sock(sk); 621 return res; 622 } 623 624 /** 625 * send_packet - send a connection-oriented message 626 * @iocb: if NULL, indicates that socket lock is already held 627 * @sock: socket structure 628 * @m: message to send 629 * @total_len: length of message 630 * 631 * Used for SOCK_SEQPACKET messages and SOCK_STREAM data. 632 * 633 * Returns the number of bytes sent on success, or errno otherwise 634 */ 635 636 static int send_packet(struct kiocb *iocb, struct socket *sock, 637 struct msghdr *m, size_t total_len) 638 { 639 struct sock *sk = sock->sk; 640 struct tipc_port *tport = tipc_sk_port(sk); 641 struct sockaddr_tipc *dest = (struct sockaddr_tipc *)m->msg_name; 642 int res; 643 644 /* Handle implied connection establishment */ 645 646 if (unlikely(dest)) 647 return send_msg(iocb, sock, m, total_len); 648 649 if ((total_len > TIPC_MAX_USER_MSG_SIZE) || 650 (m->msg_iovlen > (unsigned)INT_MAX)) 651 return -EMSGSIZE; 652 653 if (iocb) 654 lock_sock(sk); 655 656 do { 657 if (unlikely(sock->state != SS_CONNECTED)) { 658 if (sock->state == SS_DISCONNECTING) 659 res = -EPIPE; 660 else 661 res = -ENOTCONN; 662 break; 663 } 664 665 res = tipc_send(tport->ref, m->msg_iovlen, m->msg_iov, 666 total_len); 667 if (likely(res != -ELINKCONG)) 668 break; 669 if (m->msg_flags & MSG_DONTWAIT) { 670 res = -EWOULDBLOCK; 671 break; 672 } 673 release_sock(sk); 674 res = wait_event_interruptible(*sk_sleep(sk), 675 (!tport->congested || !tport->connected)); 676 lock_sock(sk); 677 if (res) 678 break; 679 } while (1); 680 681 if (iocb) 682 release_sock(sk); 683 return res; 684 } 685 686 /** 687 * send_stream - send stream-oriented data 688 * @iocb: (unused) 689 * @sock: socket structure 690 * @m: data to send 691 * @total_len: total length of data to be sent 692 * 693 * Used for SOCK_STREAM data. 694 * 695 * Returns the number of bytes sent on success (or partial success), 696 * or errno if no data sent 697 */ 698 699 static int send_stream(struct kiocb *iocb, struct socket *sock, 700 struct msghdr *m, size_t total_len) 701 { 702 struct sock *sk = sock->sk; 703 struct tipc_port *tport = tipc_sk_port(sk); 704 struct msghdr my_msg; 705 struct iovec my_iov; 706 struct iovec *curr_iov; 707 int curr_iovlen; 708 char __user *curr_start; 709 u32 hdr_size; 710 int curr_left; 711 int bytes_to_send; 712 int bytes_sent; 713 int res; 714 715 lock_sock(sk); 716 717 /* Handle special cases where there is no connection */ 718 719 if (unlikely(sock->state != SS_CONNECTED)) { 720 if (sock->state == SS_UNCONNECTED) { 721 res = send_packet(NULL, sock, m, total_len); 722 goto exit; 723 } else if (sock->state == SS_DISCONNECTING) { 724 res = -EPIPE; 725 goto exit; 726 } else { 727 res = -ENOTCONN; 728 goto exit; 729 } 730 } 731 732 if (unlikely(m->msg_name)) { 733 res = -EISCONN; 734 goto exit; 735 } 736 737 if ((total_len > (unsigned)INT_MAX) || 738 (m->msg_iovlen > (unsigned)INT_MAX)) { 739 res = -EMSGSIZE; 740 goto exit; 741 } 742 743 /* 744 * Send each iovec entry using one or more messages 745 * 746 * Note: This algorithm is good for the most likely case 747 * (i.e. one large iovec entry), but could be improved to pass sets 748 * of small iovec entries into send_packet(). 749 */ 750 751 curr_iov = m->msg_iov; 752 curr_iovlen = m->msg_iovlen; 753 my_msg.msg_iov = &my_iov; 754 my_msg.msg_iovlen = 1; 755 my_msg.msg_flags = m->msg_flags; 756 my_msg.msg_name = NULL; 757 bytes_sent = 0; 758 759 hdr_size = msg_hdr_sz(&tport->phdr); 760 761 while (curr_iovlen--) { 762 curr_start = curr_iov->iov_base; 763 curr_left = curr_iov->iov_len; 764 765 while (curr_left) { 766 bytes_to_send = tport->max_pkt - hdr_size; 767 if (bytes_to_send > TIPC_MAX_USER_MSG_SIZE) 768 bytes_to_send = TIPC_MAX_USER_MSG_SIZE; 769 if (curr_left < bytes_to_send) 770 bytes_to_send = curr_left; 771 my_iov.iov_base = curr_start; 772 my_iov.iov_len = bytes_to_send; 773 res = send_packet(NULL, sock, &my_msg, bytes_to_send); 774 if (res < 0) { 775 if (bytes_sent) 776 res = bytes_sent; 777 goto exit; 778 } 779 curr_left -= bytes_to_send; 780 curr_start += bytes_to_send; 781 bytes_sent += bytes_to_send; 782 } 783 784 curr_iov++; 785 } 786 res = bytes_sent; 787 exit: 788 release_sock(sk); 789 return res; 790 } 791 792 /** 793 * auto_connect - complete connection setup to a remote port 794 * @sock: socket structure 795 * @msg: peer's response message 796 * 797 * Returns 0 on success, errno otherwise 798 */ 799 800 static int auto_connect(struct socket *sock, struct tipc_msg *msg) 801 { 802 struct tipc_sock *tsock = tipc_sk(sock->sk); 803 804 if (msg_errcode(msg)) { 805 sock->state = SS_DISCONNECTING; 806 return -ECONNREFUSED; 807 } 808 809 tsock->peer_name.ref = msg_origport(msg); 810 tsock->peer_name.node = msg_orignode(msg); 811 tipc_connect2port(tsock->p->ref, &tsock->peer_name); 812 tipc_set_portimportance(tsock->p->ref, msg_importance(msg)); 813 sock->state = SS_CONNECTED; 814 return 0; 815 } 816 817 /** 818 * set_orig_addr - capture sender's address for received message 819 * @m: descriptor for message info 820 * @msg: received message header 821 * 822 * Note: Address is not captured if not requested by receiver. 823 */ 824 825 static void set_orig_addr(struct msghdr *m, struct tipc_msg *msg) 826 { 827 struct sockaddr_tipc *addr = (struct sockaddr_tipc *)m->msg_name; 828 829 if (addr) { 830 addr->family = AF_TIPC; 831 addr->addrtype = TIPC_ADDR_ID; 832 addr->addr.id.ref = msg_origport(msg); 833 addr->addr.id.node = msg_orignode(msg); 834 addr->addr.name.domain = 0; /* could leave uninitialized */ 835 addr->scope = 0; /* could leave uninitialized */ 836 m->msg_namelen = sizeof(struct sockaddr_tipc); 837 } 838 } 839 840 /** 841 * anc_data_recv - optionally capture ancillary data for received message 842 * @m: descriptor for message info 843 * @msg: received message header 844 * @tport: TIPC port associated with message 845 * 846 * Note: Ancillary data is not captured if not requested by receiver. 847 * 848 * Returns 0 if successful, otherwise errno 849 */ 850 851 static int anc_data_recv(struct msghdr *m, struct tipc_msg *msg, 852 struct tipc_port *tport) 853 { 854 u32 anc_data[3]; 855 u32 err; 856 u32 dest_type; 857 int has_name; 858 int res; 859 860 if (likely(m->msg_controllen == 0)) 861 return 0; 862 863 /* Optionally capture errored message object(s) */ 864 865 err = msg ? msg_errcode(msg) : 0; 866 if (unlikely(err)) { 867 anc_data[0] = err; 868 anc_data[1] = msg_data_sz(msg); 869 res = put_cmsg(m, SOL_TIPC, TIPC_ERRINFO, 8, anc_data); 870 if (res) 871 return res; 872 if (anc_data[1]) { 873 res = put_cmsg(m, SOL_TIPC, TIPC_RETDATA, anc_data[1], 874 msg_data(msg)); 875 if (res) 876 return res; 877 } 878 } 879 880 /* Optionally capture message destination object */ 881 882 dest_type = msg ? msg_type(msg) : TIPC_DIRECT_MSG; 883 switch (dest_type) { 884 case TIPC_NAMED_MSG: 885 has_name = 1; 886 anc_data[0] = msg_nametype(msg); 887 anc_data[1] = msg_namelower(msg); 888 anc_data[2] = msg_namelower(msg); 889 break; 890 case TIPC_MCAST_MSG: 891 has_name = 1; 892 anc_data[0] = msg_nametype(msg); 893 anc_data[1] = msg_namelower(msg); 894 anc_data[2] = msg_nameupper(msg); 895 break; 896 case TIPC_CONN_MSG: 897 has_name = (tport->conn_type != 0); 898 anc_data[0] = tport->conn_type; 899 anc_data[1] = tport->conn_instance; 900 anc_data[2] = tport->conn_instance; 901 break; 902 default: 903 has_name = 0; 904 } 905 if (has_name) { 906 res = put_cmsg(m, SOL_TIPC, TIPC_DESTNAME, 12, anc_data); 907 if (res) 908 return res; 909 } 910 911 return 0; 912 } 913 914 /** 915 * recv_msg - receive packet-oriented message 916 * @iocb: (unused) 917 * @m: descriptor for message info 918 * @buf_len: total size of user buffer area 919 * @flags: receive flags 920 * 921 * Used for SOCK_DGRAM, SOCK_RDM, and SOCK_SEQPACKET messages. 922 * If the complete message doesn't fit in user area, truncate it. 923 * 924 * Returns size of returned message data, errno otherwise 925 */ 926 927 static int recv_msg(struct kiocb *iocb, struct socket *sock, 928 struct msghdr *m, size_t buf_len, int flags) 929 { 930 struct sock *sk = sock->sk; 931 struct tipc_port *tport = tipc_sk_port(sk); 932 struct sk_buff *buf; 933 struct tipc_msg *msg; 934 long timeout; 935 unsigned int sz; 936 u32 err; 937 int res; 938 939 /* Catch invalid receive requests */ 940 941 if (unlikely(!buf_len)) 942 return -EINVAL; 943 944 lock_sock(sk); 945 946 if (unlikely(sock->state == SS_UNCONNECTED)) { 947 res = -ENOTCONN; 948 goto exit; 949 } 950 951 timeout = sock_rcvtimeo(sk, flags & MSG_DONTWAIT); 952 restart: 953 954 /* Look for a message in receive queue; wait if necessary */ 955 956 while (skb_queue_empty(&sk->sk_receive_queue)) { 957 if (sock->state == SS_DISCONNECTING) { 958 res = -ENOTCONN; 959 goto exit; 960 } 961 if (timeout <= 0L) { 962 res = timeout ? timeout : -EWOULDBLOCK; 963 goto exit; 964 } 965 release_sock(sk); 966 timeout = wait_event_interruptible_timeout(*sk_sleep(sk), 967 tipc_rx_ready(sock), 968 timeout); 969 lock_sock(sk); 970 } 971 972 /* Look at first message in receive queue */ 973 974 buf = skb_peek(&sk->sk_receive_queue); 975 msg = buf_msg(buf); 976 sz = msg_data_sz(msg); 977 err = msg_errcode(msg); 978 979 /* Complete connection setup for an implied connect */ 980 981 if (unlikely(sock->state == SS_CONNECTING)) { 982 res = auto_connect(sock, msg); 983 if (res) 984 goto exit; 985 } 986 987 /* Discard an empty non-errored message & try again */ 988 989 if ((!sz) && (!err)) { 990 advance_rx_queue(sk); 991 goto restart; 992 } 993 994 /* Capture sender's address (optional) */ 995 996 set_orig_addr(m, msg); 997 998 /* Capture ancillary data (optional) */ 999 1000 res = anc_data_recv(m, msg, tport); 1001 if (res) 1002 goto exit; 1003 1004 /* Capture message data (if valid) & compute return value (always) */ 1005 1006 if (!err) { 1007 if (unlikely(buf_len < sz)) { 1008 sz = buf_len; 1009 m->msg_flags |= MSG_TRUNC; 1010 } 1011 res = skb_copy_datagram_iovec(buf, msg_hdr_sz(msg), 1012 m->msg_iov, sz); 1013 if (res) 1014 goto exit; 1015 res = sz; 1016 } else { 1017 if ((sock->state == SS_READY) || 1018 ((err == TIPC_CONN_SHUTDOWN) || m->msg_control)) 1019 res = 0; 1020 else 1021 res = -ECONNRESET; 1022 } 1023 1024 /* Consume received message (optional) */ 1025 1026 if (likely(!(flags & MSG_PEEK))) { 1027 if ((sock->state != SS_READY) && 1028 (++tport->conn_unacked >= TIPC_FLOW_CONTROL_WIN)) 1029 tipc_acknowledge(tport->ref, tport->conn_unacked); 1030 advance_rx_queue(sk); 1031 } 1032 exit: 1033 release_sock(sk); 1034 return res; 1035 } 1036 1037 /** 1038 * recv_stream - receive stream-oriented data 1039 * @iocb: (unused) 1040 * @m: descriptor for message info 1041 * @buf_len: total size of user buffer area 1042 * @flags: receive flags 1043 * 1044 * Used for SOCK_STREAM messages only. If not enough data is available 1045 * will optionally wait for more; never truncates data. 1046 * 1047 * Returns size of returned message data, errno otherwise 1048 */ 1049 1050 static int recv_stream(struct kiocb *iocb, struct socket *sock, 1051 struct msghdr *m, size_t buf_len, int flags) 1052 { 1053 struct sock *sk = sock->sk; 1054 struct tipc_port *tport = tipc_sk_port(sk); 1055 struct sk_buff *buf; 1056 struct tipc_msg *msg; 1057 long timeout; 1058 unsigned int sz; 1059 int sz_to_copy, target, needed; 1060 int sz_copied = 0; 1061 u32 err; 1062 int res = 0; 1063 1064 /* Catch invalid receive attempts */ 1065 1066 if (unlikely(!buf_len)) 1067 return -EINVAL; 1068 1069 lock_sock(sk); 1070 1071 if (unlikely((sock->state == SS_UNCONNECTED) || 1072 (sock->state == SS_CONNECTING))) { 1073 res = -ENOTCONN; 1074 goto exit; 1075 } 1076 1077 target = sock_rcvlowat(sk, flags & MSG_WAITALL, buf_len); 1078 timeout = sock_rcvtimeo(sk, flags & MSG_DONTWAIT); 1079 restart: 1080 1081 /* Look for a message in receive queue; wait if necessary */ 1082 1083 while (skb_queue_empty(&sk->sk_receive_queue)) { 1084 if (sock->state == SS_DISCONNECTING) { 1085 res = -ENOTCONN; 1086 goto exit; 1087 } 1088 if (timeout <= 0L) { 1089 res = timeout ? timeout : -EWOULDBLOCK; 1090 goto exit; 1091 } 1092 release_sock(sk); 1093 timeout = wait_event_interruptible_timeout(*sk_sleep(sk), 1094 tipc_rx_ready(sock), 1095 timeout); 1096 lock_sock(sk); 1097 } 1098 1099 /* Look at first message in receive queue */ 1100 1101 buf = skb_peek(&sk->sk_receive_queue); 1102 msg = buf_msg(buf); 1103 sz = msg_data_sz(msg); 1104 err = msg_errcode(msg); 1105 1106 /* Discard an empty non-errored message & try again */ 1107 1108 if ((!sz) && (!err)) { 1109 advance_rx_queue(sk); 1110 goto restart; 1111 } 1112 1113 /* Optionally capture sender's address & ancillary data of first msg */ 1114 1115 if (sz_copied == 0) { 1116 set_orig_addr(m, msg); 1117 res = anc_data_recv(m, msg, tport); 1118 if (res) 1119 goto exit; 1120 } 1121 1122 /* Capture message data (if valid) & compute return value (always) */ 1123 1124 if (!err) { 1125 u32 offset = (u32)(unsigned long)(TIPC_SKB_CB(buf)->handle); 1126 1127 sz -= offset; 1128 needed = (buf_len - sz_copied); 1129 sz_to_copy = (sz <= needed) ? sz : needed; 1130 1131 res = skb_copy_datagram_iovec(buf, msg_hdr_sz(msg) + offset, 1132 m->msg_iov, sz_to_copy); 1133 if (res) 1134 goto exit; 1135 1136 sz_copied += sz_to_copy; 1137 1138 if (sz_to_copy < sz) { 1139 if (!(flags & MSG_PEEK)) 1140 TIPC_SKB_CB(buf)->handle = 1141 (void *)(unsigned long)(offset + sz_to_copy); 1142 goto exit; 1143 } 1144 } else { 1145 if (sz_copied != 0) 1146 goto exit; /* can't add error msg to valid data */ 1147 1148 if ((err == TIPC_CONN_SHUTDOWN) || m->msg_control) 1149 res = 0; 1150 else 1151 res = -ECONNRESET; 1152 } 1153 1154 /* Consume received message (optional) */ 1155 1156 if (likely(!(flags & MSG_PEEK))) { 1157 if (unlikely(++tport->conn_unacked >= TIPC_FLOW_CONTROL_WIN)) 1158 tipc_acknowledge(tport->ref, tport->conn_unacked); 1159 advance_rx_queue(sk); 1160 } 1161 1162 /* Loop around if more data is required */ 1163 1164 if ((sz_copied < buf_len) && /* didn't get all requested data */ 1165 (!skb_queue_empty(&sk->sk_receive_queue) || 1166 (sz_copied < target)) && /* and more is ready or required */ 1167 (!(flags & MSG_PEEK)) && /* and aren't just peeking at data */ 1168 (!err)) /* and haven't reached a FIN */ 1169 goto restart; 1170 1171 exit: 1172 release_sock(sk); 1173 return sz_copied ? sz_copied : res; 1174 } 1175 1176 /** 1177 * rx_queue_full - determine if receive queue can accept another message 1178 * @msg: message to be added to queue 1179 * @queue_size: current size of queue 1180 * @base: nominal maximum size of queue 1181 * 1182 * Returns 1 if queue is unable to accept message, 0 otherwise 1183 */ 1184 1185 static int rx_queue_full(struct tipc_msg *msg, u32 queue_size, u32 base) 1186 { 1187 u32 threshold; 1188 u32 imp = msg_importance(msg); 1189 1190 if (imp == TIPC_LOW_IMPORTANCE) 1191 threshold = base; 1192 else if (imp == TIPC_MEDIUM_IMPORTANCE) 1193 threshold = base * 2; 1194 else if (imp == TIPC_HIGH_IMPORTANCE) 1195 threshold = base * 100; 1196 else 1197 return 0; 1198 1199 if (msg_connected(msg)) 1200 threshold *= 4; 1201 1202 return queue_size >= threshold; 1203 } 1204 1205 /** 1206 * filter_rcv - validate incoming message 1207 * @sk: socket 1208 * @buf: message 1209 * 1210 * Enqueues message on receive queue if acceptable; optionally handles 1211 * disconnect indication for a connected socket. 1212 * 1213 * Called with socket lock already taken; port lock may also be taken. 1214 * 1215 * Returns TIPC error status code (TIPC_OK if message is not to be rejected) 1216 */ 1217 1218 static u32 filter_rcv(struct sock *sk, struct sk_buff *buf) 1219 { 1220 struct socket *sock = sk->sk_socket; 1221 struct tipc_msg *msg = buf_msg(buf); 1222 u32 recv_q_len; 1223 1224 /* Reject message if it is wrong sort of message for socket */ 1225 1226 /* 1227 * WOULD IT BE BETTER TO JUST DISCARD THESE MESSAGES INSTEAD? 1228 * "NO PORT" ISN'T REALLY THE RIGHT ERROR CODE, AND THERE MAY 1229 * BE SECURITY IMPLICATIONS INHERENT IN REJECTING INVALID TRAFFIC 1230 */ 1231 1232 if (sock->state == SS_READY) { 1233 if (msg_connected(msg)) 1234 return TIPC_ERR_NO_PORT; 1235 } else { 1236 if (msg_mcast(msg)) 1237 return TIPC_ERR_NO_PORT; 1238 if (sock->state == SS_CONNECTED) { 1239 if (!msg_connected(msg)) 1240 return TIPC_ERR_NO_PORT; 1241 } else if (sock->state == SS_CONNECTING) { 1242 if (!msg_connected(msg) && (msg_errcode(msg) == 0)) 1243 return TIPC_ERR_NO_PORT; 1244 } else if (sock->state == SS_LISTENING) { 1245 if (msg_connected(msg) || msg_errcode(msg)) 1246 return TIPC_ERR_NO_PORT; 1247 } else if (sock->state == SS_DISCONNECTING) { 1248 return TIPC_ERR_NO_PORT; 1249 } else /* (sock->state == SS_UNCONNECTED) */ { 1250 if (msg_connected(msg) || msg_errcode(msg)) 1251 return TIPC_ERR_NO_PORT; 1252 } 1253 } 1254 1255 /* Reject message if there isn't room to queue it */ 1256 1257 recv_q_len = (u32)atomic_read(&tipc_queue_size); 1258 if (unlikely(recv_q_len >= OVERLOAD_LIMIT_BASE)) { 1259 if (rx_queue_full(msg, recv_q_len, OVERLOAD_LIMIT_BASE)) 1260 return TIPC_ERR_OVERLOAD; 1261 } 1262 recv_q_len = skb_queue_len(&sk->sk_receive_queue); 1263 if (unlikely(recv_q_len >= (OVERLOAD_LIMIT_BASE / 2))) { 1264 if (rx_queue_full(msg, recv_q_len, OVERLOAD_LIMIT_BASE / 2)) 1265 return TIPC_ERR_OVERLOAD; 1266 } 1267 1268 /* Enqueue message (finally!) */ 1269 1270 TIPC_SKB_CB(buf)->handle = 0; 1271 atomic_inc(&tipc_queue_size); 1272 __skb_queue_tail(&sk->sk_receive_queue, buf); 1273 1274 /* Initiate connection termination for an incoming 'FIN' */ 1275 1276 if (unlikely(msg_errcode(msg) && (sock->state == SS_CONNECTED))) { 1277 sock->state = SS_DISCONNECTING; 1278 tipc_disconnect_port(tipc_sk_port(sk)); 1279 } 1280 1281 if (waitqueue_active(sk_sleep(sk))) 1282 wake_up_interruptible(sk_sleep(sk)); 1283 return TIPC_OK; 1284 } 1285 1286 /** 1287 * backlog_rcv - handle incoming message from backlog queue 1288 * @sk: socket 1289 * @buf: message 1290 * 1291 * Caller must hold socket lock, but not port lock. 1292 * 1293 * Returns 0 1294 */ 1295 1296 static int backlog_rcv(struct sock *sk, struct sk_buff *buf) 1297 { 1298 u32 res; 1299 1300 res = filter_rcv(sk, buf); 1301 if (res) 1302 tipc_reject_msg(buf, res); 1303 return 0; 1304 } 1305 1306 /** 1307 * dispatch - handle incoming message 1308 * @tport: TIPC port that received message 1309 * @buf: message 1310 * 1311 * Called with port lock already taken. 1312 * 1313 * Returns TIPC error status code (TIPC_OK if message is not to be rejected) 1314 */ 1315 1316 static u32 dispatch(struct tipc_port *tport, struct sk_buff *buf) 1317 { 1318 struct sock *sk = (struct sock *)tport->usr_handle; 1319 u32 res; 1320 1321 /* 1322 * Process message if socket is unlocked; otherwise add to backlog queue 1323 * 1324 * This code is based on sk_receive_skb(), but must be distinct from it 1325 * since a TIPC-specific filter/reject mechanism is utilized 1326 */ 1327 1328 bh_lock_sock(sk); 1329 if (!sock_owned_by_user(sk)) { 1330 res = filter_rcv(sk, buf); 1331 } else { 1332 if (sk_add_backlog(sk, buf)) 1333 res = TIPC_ERR_OVERLOAD; 1334 else 1335 res = TIPC_OK; 1336 } 1337 bh_unlock_sock(sk); 1338 1339 return res; 1340 } 1341 1342 /** 1343 * wakeupdispatch - wake up port after congestion 1344 * @tport: port to wakeup 1345 * 1346 * Called with port lock already taken. 1347 */ 1348 1349 static void wakeupdispatch(struct tipc_port *tport) 1350 { 1351 struct sock *sk = (struct sock *)tport->usr_handle; 1352 1353 if (waitqueue_active(sk_sleep(sk))) 1354 wake_up_interruptible(sk_sleep(sk)); 1355 } 1356 1357 /** 1358 * connect - establish a connection to another TIPC port 1359 * @sock: socket structure 1360 * @dest: socket address for destination port 1361 * @destlen: size of socket address data structure 1362 * @flags: file-related flags associated with socket 1363 * 1364 * Returns 0 on success, errno otherwise 1365 */ 1366 1367 static int connect(struct socket *sock, struct sockaddr *dest, int destlen, 1368 int flags) 1369 { 1370 struct sock *sk = sock->sk; 1371 struct sockaddr_tipc *dst = (struct sockaddr_tipc *)dest; 1372 struct msghdr m = {NULL,}; 1373 struct sk_buff *buf; 1374 struct tipc_msg *msg; 1375 long timeout; 1376 int res; 1377 1378 lock_sock(sk); 1379 1380 /* For now, TIPC does not allow use of connect() with DGRAM/RDM types */ 1381 1382 if (sock->state == SS_READY) { 1383 res = -EOPNOTSUPP; 1384 goto exit; 1385 } 1386 1387 /* For now, TIPC does not support the non-blocking form of connect() */ 1388 1389 if (flags & O_NONBLOCK) { 1390 res = -EOPNOTSUPP; 1391 goto exit; 1392 } 1393 1394 /* Issue Posix-compliant error code if socket is in the wrong state */ 1395 1396 if (sock->state == SS_LISTENING) { 1397 res = -EOPNOTSUPP; 1398 goto exit; 1399 } 1400 if (sock->state == SS_CONNECTING) { 1401 res = -EALREADY; 1402 goto exit; 1403 } 1404 if (sock->state != SS_UNCONNECTED) { 1405 res = -EISCONN; 1406 goto exit; 1407 } 1408 1409 /* 1410 * Reject connection attempt using multicast address 1411 * 1412 * Note: send_msg() validates the rest of the address fields, 1413 * so there's no need to do it here 1414 */ 1415 1416 if (dst->addrtype == TIPC_ADDR_MCAST) { 1417 res = -EINVAL; 1418 goto exit; 1419 } 1420 1421 /* Reject any messages already in receive queue (very unlikely) */ 1422 1423 reject_rx_queue(sk); 1424 1425 /* Send a 'SYN-' to destination */ 1426 1427 m.msg_name = dest; 1428 m.msg_namelen = destlen; 1429 res = send_msg(NULL, sock, &m, 0); 1430 if (res < 0) 1431 goto exit; 1432 1433 /* Wait until an 'ACK' or 'RST' arrives, or a timeout occurs */ 1434 1435 timeout = tipc_sk(sk)->conn_timeout; 1436 release_sock(sk); 1437 res = wait_event_interruptible_timeout(*sk_sleep(sk), 1438 (!skb_queue_empty(&sk->sk_receive_queue) || 1439 (sock->state != SS_CONNECTING)), 1440 timeout ? timeout : MAX_SCHEDULE_TIMEOUT); 1441 lock_sock(sk); 1442 1443 if (res > 0) { 1444 buf = skb_peek(&sk->sk_receive_queue); 1445 if (buf != NULL) { 1446 msg = buf_msg(buf); 1447 res = auto_connect(sock, msg); 1448 if (!res) { 1449 if (!msg_data_sz(msg)) 1450 advance_rx_queue(sk); 1451 } 1452 } else { 1453 if (sock->state == SS_CONNECTED) 1454 res = -EISCONN; 1455 else 1456 res = -ECONNREFUSED; 1457 } 1458 } else { 1459 if (res == 0) 1460 res = -ETIMEDOUT; 1461 else 1462 ; /* leave "res" unchanged */ 1463 sock->state = SS_DISCONNECTING; 1464 } 1465 1466 exit: 1467 release_sock(sk); 1468 return res; 1469 } 1470 1471 /** 1472 * listen - allow socket to listen for incoming connections 1473 * @sock: socket structure 1474 * @len: (unused) 1475 * 1476 * Returns 0 on success, errno otherwise 1477 */ 1478 1479 static int listen(struct socket *sock, int len) 1480 { 1481 struct sock *sk = sock->sk; 1482 int res; 1483 1484 lock_sock(sk); 1485 1486 if (sock->state == SS_READY) 1487 res = -EOPNOTSUPP; 1488 else if (sock->state != SS_UNCONNECTED) 1489 res = -EINVAL; 1490 else { 1491 sock->state = SS_LISTENING; 1492 res = 0; 1493 } 1494 1495 release_sock(sk); 1496 return res; 1497 } 1498 1499 /** 1500 * accept - wait for connection request 1501 * @sock: listening socket 1502 * @newsock: new socket that is to be connected 1503 * @flags: file-related flags associated with socket 1504 * 1505 * Returns 0 on success, errno otherwise 1506 */ 1507 1508 static int accept(struct socket *sock, struct socket *new_sock, int flags) 1509 { 1510 struct sock *sk = sock->sk; 1511 struct sk_buff *buf; 1512 int res; 1513 1514 lock_sock(sk); 1515 1516 if (sock->state == SS_READY) { 1517 res = -EOPNOTSUPP; 1518 goto exit; 1519 } 1520 if (sock->state != SS_LISTENING) { 1521 res = -EINVAL; 1522 goto exit; 1523 } 1524 1525 while (skb_queue_empty(&sk->sk_receive_queue)) { 1526 if (flags & O_NONBLOCK) { 1527 res = -EWOULDBLOCK; 1528 goto exit; 1529 } 1530 release_sock(sk); 1531 res = wait_event_interruptible(*sk_sleep(sk), 1532 (!skb_queue_empty(&sk->sk_receive_queue))); 1533 lock_sock(sk); 1534 if (res) 1535 goto exit; 1536 } 1537 1538 buf = skb_peek(&sk->sk_receive_queue); 1539 1540 res = tipc_create(sock_net(sock->sk), new_sock, 0, 0); 1541 if (!res) { 1542 struct sock *new_sk = new_sock->sk; 1543 struct tipc_sock *new_tsock = tipc_sk(new_sk); 1544 struct tipc_port *new_tport = new_tsock->p; 1545 u32 new_ref = new_tport->ref; 1546 struct tipc_msg *msg = buf_msg(buf); 1547 1548 lock_sock(new_sk); 1549 1550 /* 1551 * Reject any stray messages received by new socket 1552 * before the socket lock was taken (very, very unlikely) 1553 */ 1554 1555 reject_rx_queue(new_sk); 1556 1557 /* Connect new socket to it's peer */ 1558 1559 new_tsock->peer_name.ref = msg_origport(msg); 1560 new_tsock->peer_name.node = msg_orignode(msg); 1561 tipc_connect2port(new_ref, &new_tsock->peer_name); 1562 new_sock->state = SS_CONNECTED; 1563 1564 tipc_set_portimportance(new_ref, msg_importance(msg)); 1565 if (msg_named(msg)) { 1566 new_tport->conn_type = msg_nametype(msg); 1567 new_tport->conn_instance = msg_nameinst(msg); 1568 } 1569 1570 /* 1571 * Respond to 'SYN-' by discarding it & returning 'ACK'-. 1572 * Respond to 'SYN+' by queuing it on new socket. 1573 */ 1574 1575 if (!msg_data_sz(msg)) { 1576 struct msghdr m = {NULL,}; 1577 1578 advance_rx_queue(sk); 1579 send_packet(NULL, new_sock, &m, 0); 1580 } else { 1581 __skb_dequeue(&sk->sk_receive_queue); 1582 __skb_queue_head(&new_sk->sk_receive_queue, buf); 1583 } 1584 release_sock(new_sk); 1585 } 1586 exit: 1587 release_sock(sk); 1588 return res; 1589 } 1590 1591 /** 1592 * shutdown - shutdown socket connection 1593 * @sock: socket structure 1594 * @how: direction to close (must be SHUT_RDWR) 1595 * 1596 * Terminates connection (if necessary), then purges socket's receive queue. 1597 * 1598 * Returns 0 on success, errno otherwise 1599 */ 1600 1601 static int shutdown(struct socket *sock, int how) 1602 { 1603 struct sock *sk = sock->sk; 1604 struct tipc_port *tport = tipc_sk_port(sk); 1605 struct sk_buff *buf; 1606 int res; 1607 1608 if (how != SHUT_RDWR) 1609 return -EINVAL; 1610 1611 lock_sock(sk); 1612 1613 switch (sock->state) { 1614 case SS_CONNECTING: 1615 case SS_CONNECTED: 1616 1617 /* Disconnect and send a 'FIN+' or 'FIN-' message to peer */ 1618 restart: 1619 buf = __skb_dequeue(&sk->sk_receive_queue); 1620 if (buf) { 1621 atomic_dec(&tipc_queue_size); 1622 if (TIPC_SKB_CB(buf)->handle != 0) { 1623 buf_discard(buf); 1624 goto restart; 1625 } 1626 tipc_disconnect(tport->ref); 1627 tipc_reject_msg(buf, TIPC_CONN_SHUTDOWN); 1628 } else { 1629 tipc_shutdown(tport->ref); 1630 } 1631 1632 sock->state = SS_DISCONNECTING; 1633 1634 /* fall through */ 1635 1636 case SS_DISCONNECTING: 1637 1638 /* Discard any unreceived messages; wake up sleeping tasks */ 1639 1640 discard_rx_queue(sk); 1641 if (waitqueue_active(sk_sleep(sk))) 1642 wake_up_interruptible(sk_sleep(sk)); 1643 res = 0; 1644 break; 1645 1646 default: 1647 res = -ENOTCONN; 1648 } 1649 1650 release_sock(sk); 1651 return res; 1652 } 1653 1654 /** 1655 * setsockopt - set socket option 1656 * @sock: socket structure 1657 * @lvl: option level 1658 * @opt: option identifier 1659 * @ov: pointer to new option value 1660 * @ol: length of option value 1661 * 1662 * For stream sockets only, accepts and ignores all IPPROTO_TCP options 1663 * (to ease compatibility). 1664 * 1665 * Returns 0 on success, errno otherwise 1666 */ 1667 1668 static int setsockopt(struct socket *sock, 1669 int lvl, int opt, char __user *ov, unsigned int ol) 1670 { 1671 struct sock *sk = sock->sk; 1672 struct tipc_port *tport = tipc_sk_port(sk); 1673 u32 value; 1674 int res; 1675 1676 if ((lvl == IPPROTO_TCP) && (sock->type == SOCK_STREAM)) 1677 return 0; 1678 if (lvl != SOL_TIPC) 1679 return -ENOPROTOOPT; 1680 if (ol < sizeof(value)) 1681 return -EINVAL; 1682 res = get_user(value, (u32 __user *)ov); 1683 if (res) 1684 return res; 1685 1686 lock_sock(sk); 1687 1688 switch (opt) { 1689 case TIPC_IMPORTANCE: 1690 res = tipc_set_portimportance(tport->ref, value); 1691 break; 1692 case TIPC_SRC_DROPPABLE: 1693 if (sock->type != SOCK_STREAM) 1694 res = tipc_set_portunreliable(tport->ref, value); 1695 else 1696 res = -ENOPROTOOPT; 1697 break; 1698 case TIPC_DEST_DROPPABLE: 1699 res = tipc_set_portunreturnable(tport->ref, value); 1700 break; 1701 case TIPC_CONN_TIMEOUT: 1702 tipc_sk(sk)->conn_timeout = msecs_to_jiffies(value); 1703 /* no need to set "res", since already 0 at this point */ 1704 break; 1705 default: 1706 res = -EINVAL; 1707 } 1708 1709 release_sock(sk); 1710 1711 return res; 1712 } 1713 1714 /** 1715 * getsockopt - get socket option 1716 * @sock: socket structure 1717 * @lvl: option level 1718 * @opt: option identifier 1719 * @ov: receptacle for option value 1720 * @ol: receptacle for length of option value 1721 * 1722 * For stream sockets only, returns 0 length result for all IPPROTO_TCP options 1723 * (to ease compatibility). 1724 * 1725 * Returns 0 on success, errno otherwise 1726 */ 1727 1728 static int getsockopt(struct socket *sock, 1729 int lvl, int opt, char __user *ov, int __user *ol) 1730 { 1731 struct sock *sk = sock->sk; 1732 struct tipc_port *tport = tipc_sk_port(sk); 1733 int len; 1734 u32 value; 1735 int res; 1736 1737 if ((lvl == IPPROTO_TCP) && (sock->type == SOCK_STREAM)) 1738 return put_user(0, ol); 1739 if (lvl != SOL_TIPC) 1740 return -ENOPROTOOPT; 1741 res = get_user(len, ol); 1742 if (res) 1743 return res; 1744 1745 lock_sock(sk); 1746 1747 switch (opt) { 1748 case TIPC_IMPORTANCE: 1749 res = tipc_portimportance(tport->ref, &value); 1750 break; 1751 case TIPC_SRC_DROPPABLE: 1752 res = tipc_portunreliable(tport->ref, &value); 1753 break; 1754 case TIPC_DEST_DROPPABLE: 1755 res = tipc_portunreturnable(tport->ref, &value); 1756 break; 1757 case TIPC_CONN_TIMEOUT: 1758 value = jiffies_to_msecs(tipc_sk(sk)->conn_timeout); 1759 /* no need to set "res", since already 0 at this point */ 1760 break; 1761 case TIPC_NODE_RECVQ_DEPTH: 1762 value = (u32)atomic_read(&tipc_queue_size); 1763 break; 1764 case TIPC_SOCK_RECVQ_DEPTH: 1765 value = skb_queue_len(&sk->sk_receive_queue); 1766 break; 1767 default: 1768 res = -EINVAL; 1769 } 1770 1771 release_sock(sk); 1772 1773 if (res) 1774 return res; /* "get" failed */ 1775 1776 if (len < sizeof(value)) 1777 return -EINVAL; 1778 1779 if (copy_to_user(ov, &value, sizeof(value))) 1780 return -EFAULT; 1781 1782 return put_user(sizeof(value), ol); 1783 } 1784 1785 /** 1786 * Protocol switches for the various types of TIPC sockets 1787 */ 1788 1789 static const struct proto_ops msg_ops = { 1790 .owner = THIS_MODULE, 1791 .family = AF_TIPC, 1792 .release = release, 1793 .bind = bind, 1794 .connect = connect, 1795 .socketpair = sock_no_socketpair, 1796 .accept = accept, 1797 .getname = get_name, 1798 .poll = poll, 1799 .ioctl = sock_no_ioctl, 1800 .listen = listen, 1801 .shutdown = shutdown, 1802 .setsockopt = setsockopt, 1803 .getsockopt = getsockopt, 1804 .sendmsg = send_msg, 1805 .recvmsg = recv_msg, 1806 .mmap = sock_no_mmap, 1807 .sendpage = sock_no_sendpage 1808 }; 1809 1810 static const struct proto_ops packet_ops = { 1811 .owner = THIS_MODULE, 1812 .family = AF_TIPC, 1813 .release = release, 1814 .bind = bind, 1815 .connect = connect, 1816 .socketpair = sock_no_socketpair, 1817 .accept = accept, 1818 .getname = get_name, 1819 .poll = poll, 1820 .ioctl = sock_no_ioctl, 1821 .listen = listen, 1822 .shutdown = shutdown, 1823 .setsockopt = setsockopt, 1824 .getsockopt = getsockopt, 1825 .sendmsg = send_packet, 1826 .recvmsg = recv_msg, 1827 .mmap = sock_no_mmap, 1828 .sendpage = sock_no_sendpage 1829 }; 1830 1831 static const struct proto_ops stream_ops = { 1832 .owner = THIS_MODULE, 1833 .family = AF_TIPC, 1834 .release = release, 1835 .bind = bind, 1836 .connect = connect, 1837 .socketpair = sock_no_socketpair, 1838 .accept = accept, 1839 .getname = get_name, 1840 .poll = poll, 1841 .ioctl = sock_no_ioctl, 1842 .listen = listen, 1843 .shutdown = shutdown, 1844 .setsockopt = setsockopt, 1845 .getsockopt = getsockopt, 1846 .sendmsg = send_stream, 1847 .recvmsg = recv_stream, 1848 .mmap = sock_no_mmap, 1849 .sendpage = sock_no_sendpage 1850 }; 1851 1852 static const struct net_proto_family tipc_family_ops = { 1853 .owner = THIS_MODULE, 1854 .family = AF_TIPC, 1855 .create = tipc_create 1856 }; 1857 1858 static struct proto tipc_proto = { 1859 .name = "TIPC", 1860 .owner = THIS_MODULE, 1861 .obj_size = sizeof(struct tipc_sock) 1862 }; 1863 1864 /** 1865 * tipc_socket_init - initialize TIPC socket interface 1866 * 1867 * Returns 0 on success, errno otherwise 1868 */ 1869 int tipc_socket_init(void) 1870 { 1871 int res; 1872 1873 res = proto_register(&tipc_proto, 1); 1874 if (res) { 1875 err("Failed to register TIPC protocol type\n"); 1876 goto out; 1877 } 1878 1879 res = sock_register(&tipc_family_ops); 1880 if (res) { 1881 err("Failed to register TIPC socket type\n"); 1882 proto_unregister(&tipc_proto); 1883 goto out; 1884 } 1885 1886 sockets_enabled = 1; 1887 out: 1888 return res; 1889 } 1890 1891 /** 1892 * tipc_socket_stop - stop TIPC socket interface 1893 */ 1894 1895 void tipc_socket_stop(void) 1896 { 1897 if (!sockets_enabled) 1898 return; 1899 1900 sockets_enabled = 0; 1901 sock_unregister(tipc_family_ops.family); 1902 proto_unregister(&tipc_proto); 1903 } 1904 1905