1 // SPDX-License-Identifier: GPL-2.0-or-later 2 /* RxRPC recvmsg() implementation 3 * 4 * Copyright (C) 2007 Red Hat, Inc. All Rights Reserved. 5 * Written by David Howells (dhowells@redhat.com) 6 */ 7 8 #define pr_fmt(fmt) KBUILD_MODNAME ": " fmt 9 10 #include <linux/net.h> 11 #include <linux/skbuff.h> 12 #include <linux/export.h> 13 #include <linux/sched/signal.h> 14 15 #include <net/sock.h> 16 #include <net/af_rxrpc.h> 17 #include "ar-internal.h" 18 19 /* 20 * Post a call for attention by the socket or kernel service. Further 21 * notifications are suppressed by putting recvmsg_link on a dummy queue. 22 */ 23 void rxrpc_notify_socket(struct rxrpc_call *call) 24 { 25 struct rxrpc_sock *rx; 26 struct sock *sk; 27 28 _enter("%d", call->debug_id); 29 30 if (!list_empty(&call->recvmsg_link)) 31 return; 32 33 rcu_read_lock(); 34 35 rx = rcu_dereference(call->socket); 36 sk = &rx->sk; 37 if (rx && sk->sk_state < RXRPC_CLOSE) { 38 if (call->notify_rx) { 39 spin_lock_bh(&call->notify_lock); 40 call->notify_rx(sk, call, call->user_call_ID); 41 spin_unlock_bh(&call->notify_lock); 42 } else { 43 write_lock_bh(&rx->recvmsg_lock); 44 if (list_empty(&call->recvmsg_link)) { 45 rxrpc_get_call(call, rxrpc_call_got); 46 list_add_tail(&call->recvmsg_link, &rx->recvmsg_q); 47 } 48 write_unlock_bh(&rx->recvmsg_lock); 49 50 if (!sock_flag(sk, SOCK_DEAD)) { 51 _debug("call %ps", sk->sk_data_ready); 52 sk->sk_data_ready(sk); 53 } 54 } 55 } 56 57 rcu_read_unlock(); 58 _leave(""); 59 } 60 61 /* 62 * Transition a call to the complete state. 63 */ 64 bool __rxrpc_set_call_completion(struct rxrpc_call *call, 65 enum rxrpc_call_completion compl, 66 u32 abort_code, 67 int error) 68 { 69 if (call->state < RXRPC_CALL_COMPLETE) { 70 call->abort_code = abort_code; 71 call->error = error; 72 call->completion = compl, 73 call->state = RXRPC_CALL_COMPLETE; 74 trace_rxrpc_call_complete(call); 75 wake_up(&call->waitq); 76 rxrpc_notify_socket(call); 77 return true; 78 } 79 return false; 80 } 81 82 bool rxrpc_set_call_completion(struct rxrpc_call *call, 83 enum rxrpc_call_completion compl, 84 u32 abort_code, 85 int error) 86 { 87 bool ret = false; 88 89 if (call->state < RXRPC_CALL_COMPLETE) { 90 write_lock_bh(&call->state_lock); 91 ret = __rxrpc_set_call_completion(call, compl, abort_code, error); 92 write_unlock_bh(&call->state_lock); 93 } 94 return ret; 95 } 96 97 /* 98 * Record that a call successfully completed. 99 */ 100 bool __rxrpc_call_completed(struct rxrpc_call *call) 101 { 102 return __rxrpc_set_call_completion(call, RXRPC_CALL_SUCCEEDED, 0, 0); 103 } 104 105 bool rxrpc_call_completed(struct rxrpc_call *call) 106 { 107 bool ret = false; 108 109 if (call->state < RXRPC_CALL_COMPLETE) { 110 write_lock_bh(&call->state_lock); 111 ret = __rxrpc_call_completed(call); 112 write_unlock_bh(&call->state_lock); 113 } 114 return ret; 115 } 116 117 /* 118 * Record that a call is locally aborted. 119 */ 120 bool __rxrpc_abort_call(const char *why, struct rxrpc_call *call, 121 rxrpc_seq_t seq, u32 abort_code, int error) 122 { 123 trace_rxrpc_abort(call->debug_id, why, call->cid, call->call_id, seq, 124 abort_code, error); 125 return __rxrpc_set_call_completion(call, RXRPC_CALL_LOCALLY_ABORTED, 126 abort_code, error); 127 } 128 129 bool rxrpc_abort_call(const char *why, struct rxrpc_call *call, 130 rxrpc_seq_t seq, u32 abort_code, int error) 131 { 132 bool ret; 133 134 write_lock_bh(&call->state_lock); 135 ret = __rxrpc_abort_call(why, call, seq, abort_code, error); 136 write_unlock_bh(&call->state_lock); 137 return ret; 138 } 139 140 /* 141 * Pass a call terminating message to userspace. 142 */ 143 static int rxrpc_recvmsg_term(struct rxrpc_call *call, struct msghdr *msg) 144 { 145 u32 tmp = 0; 146 int ret; 147 148 switch (call->completion) { 149 case RXRPC_CALL_SUCCEEDED: 150 ret = 0; 151 if (rxrpc_is_service_call(call)) 152 ret = put_cmsg(msg, SOL_RXRPC, RXRPC_ACK, 0, &tmp); 153 break; 154 case RXRPC_CALL_REMOTELY_ABORTED: 155 tmp = call->abort_code; 156 ret = put_cmsg(msg, SOL_RXRPC, RXRPC_ABORT, 4, &tmp); 157 break; 158 case RXRPC_CALL_LOCALLY_ABORTED: 159 tmp = call->abort_code; 160 ret = put_cmsg(msg, SOL_RXRPC, RXRPC_ABORT, 4, &tmp); 161 break; 162 case RXRPC_CALL_NETWORK_ERROR: 163 tmp = -call->error; 164 ret = put_cmsg(msg, SOL_RXRPC, RXRPC_NET_ERROR, 4, &tmp); 165 break; 166 case RXRPC_CALL_LOCAL_ERROR: 167 tmp = -call->error; 168 ret = put_cmsg(msg, SOL_RXRPC, RXRPC_LOCAL_ERROR, 4, &tmp); 169 break; 170 default: 171 pr_err("Invalid terminal call state %u\n", call->state); 172 BUG(); 173 break; 174 } 175 176 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_terminal, call->rx_hard_ack, 177 call->rx_pkt_offset, call->rx_pkt_len, ret); 178 return ret; 179 } 180 181 /* 182 * Pass back notification of a new call. The call is added to the 183 * to-be-accepted list. This means that the next call to be accepted might not 184 * be the last call seen awaiting acceptance, but unless we leave this on the 185 * front of the queue and block all other messages until someone gives us a 186 * user_ID for it, there's not a lot we can do. 187 */ 188 static int rxrpc_recvmsg_new_call(struct rxrpc_sock *rx, 189 struct rxrpc_call *call, 190 struct msghdr *msg, int flags) 191 { 192 int tmp = 0, ret; 193 194 ret = put_cmsg(msg, SOL_RXRPC, RXRPC_NEW_CALL, 0, &tmp); 195 196 if (ret == 0 && !(flags & MSG_PEEK)) { 197 _debug("to be accepted"); 198 write_lock_bh(&rx->recvmsg_lock); 199 list_del_init(&call->recvmsg_link); 200 write_unlock_bh(&rx->recvmsg_lock); 201 202 rxrpc_get_call(call, rxrpc_call_got); 203 write_lock(&rx->call_lock); 204 list_add_tail(&call->accept_link, &rx->to_be_accepted); 205 write_unlock(&rx->call_lock); 206 } 207 208 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_to_be_accepted, 1, 0, 0, ret); 209 return ret; 210 } 211 212 /* 213 * End the packet reception phase. 214 */ 215 static void rxrpc_end_rx_phase(struct rxrpc_call *call, rxrpc_serial_t serial) 216 { 217 _enter("%d,%s", call->debug_id, rxrpc_call_states[call->state]); 218 219 trace_rxrpc_receive(call, rxrpc_receive_end, 0, call->rx_top); 220 ASSERTCMP(call->rx_hard_ack, ==, call->rx_top); 221 222 if (call->state == RXRPC_CALL_CLIENT_RECV_REPLY) { 223 rxrpc_propose_ACK(call, RXRPC_ACK_IDLE, serial, false, true, 224 rxrpc_propose_ack_terminal_ack); 225 //rxrpc_send_ack_packet(call, false, NULL); 226 } 227 228 write_lock_bh(&call->state_lock); 229 230 switch (call->state) { 231 case RXRPC_CALL_CLIENT_RECV_REPLY: 232 __rxrpc_call_completed(call); 233 write_unlock_bh(&call->state_lock); 234 break; 235 236 case RXRPC_CALL_SERVER_RECV_REQUEST: 237 call->tx_phase = true; 238 call->state = RXRPC_CALL_SERVER_ACK_REQUEST; 239 call->expect_req_by = jiffies + MAX_JIFFY_OFFSET; 240 write_unlock_bh(&call->state_lock); 241 rxrpc_propose_ACK(call, RXRPC_ACK_DELAY, serial, false, true, 242 rxrpc_propose_ack_processing_op); 243 break; 244 default: 245 write_unlock_bh(&call->state_lock); 246 break; 247 } 248 } 249 250 /* 251 * Discard a packet we've used up and advance the Rx window by one. 252 */ 253 static void rxrpc_rotate_rx_window(struct rxrpc_call *call) 254 { 255 struct rxrpc_skb_priv *sp; 256 struct sk_buff *skb; 257 rxrpc_serial_t serial; 258 rxrpc_seq_t hard_ack, top; 259 bool last = false; 260 u8 subpacket; 261 int ix; 262 263 _enter("%d", call->debug_id); 264 265 hard_ack = call->rx_hard_ack; 266 top = smp_load_acquire(&call->rx_top); 267 ASSERT(before(hard_ack, top)); 268 269 hard_ack++; 270 ix = hard_ack & RXRPC_RXTX_BUFF_MASK; 271 skb = call->rxtx_buffer[ix]; 272 rxrpc_see_skb(skb, rxrpc_skb_rotated); 273 sp = rxrpc_skb(skb); 274 275 subpacket = call->rxtx_annotations[ix] & RXRPC_RX_ANNO_SUBPACKET; 276 serial = sp->hdr.serial + subpacket; 277 278 if (subpacket == sp->nr_subpackets - 1 && 279 sp->rx_flags & RXRPC_SKB_INCL_LAST) 280 last = true; 281 282 call->rxtx_buffer[ix] = NULL; 283 call->rxtx_annotations[ix] = 0; 284 /* Barrier against rxrpc_input_data(). */ 285 smp_store_release(&call->rx_hard_ack, hard_ack); 286 287 rxrpc_free_skb(skb, rxrpc_skb_freed); 288 289 trace_rxrpc_receive(call, rxrpc_receive_rotate, serial, hard_ack); 290 if (last) { 291 rxrpc_end_rx_phase(call, serial); 292 } else { 293 /* Check to see if there's an ACK that needs sending. */ 294 if (after_eq(hard_ack, call->ackr_consumed + 2) || 295 after_eq(top, call->ackr_seen + 2) || 296 (hard_ack == top && after(hard_ack, call->ackr_consumed))) 297 rxrpc_propose_ACK(call, RXRPC_ACK_DELAY, serial, 298 true, true, 299 rxrpc_propose_ack_rotate_rx); 300 if (call->ackr_reason && call->ackr_reason != RXRPC_ACK_DELAY) 301 rxrpc_send_ack_packet(call, false, NULL); 302 } 303 } 304 305 /* 306 * Decrypt and verify a (sub)packet. The packet's length may be changed due to 307 * padding, but if this is the case, the packet length will be resident in the 308 * socket buffer. Note that we can't modify the master skb info as the skb may 309 * be the home to multiple subpackets. 310 */ 311 static int rxrpc_verify_packet(struct rxrpc_call *call, struct sk_buff *skb, 312 u8 annotation, 313 unsigned int offset, unsigned int len) 314 { 315 struct rxrpc_skb_priv *sp = rxrpc_skb(skb); 316 rxrpc_seq_t seq = sp->hdr.seq; 317 u16 cksum = sp->hdr.cksum; 318 u8 subpacket = annotation & RXRPC_RX_ANNO_SUBPACKET; 319 320 _enter(""); 321 322 /* For all but the head jumbo subpacket, the security checksum is in a 323 * jumbo header immediately prior to the data. 324 */ 325 if (subpacket > 0) { 326 __be16 tmp; 327 if (skb_copy_bits(skb, offset - 2, &tmp, 2) < 0) 328 BUG(); 329 cksum = ntohs(tmp); 330 seq += subpacket; 331 } 332 333 return call->security->verify_packet(call, skb, offset, len, 334 seq, cksum); 335 } 336 337 /* 338 * Locate the data within a packet. This is complicated by: 339 * 340 * (1) An skb may contain a jumbo packet - so we have to find the appropriate 341 * subpacket. 342 * 343 * (2) The (sub)packets may be encrypted and, if so, the encrypted portion 344 * contains an extra header which includes the true length of the data, 345 * excluding any encrypted padding. 346 */ 347 static int rxrpc_locate_data(struct rxrpc_call *call, struct sk_buff *skb, 348 u8 *_annotation, 349 unsigned int *_offset, unsigned int *_len, 350 bool *_last) 351 { 352 struct rxrpc_skb_priv *sp = rxrpc_skb(skb); 353 unsigned int offset = sizeof(struct rxrpc_wire_header); 354 unsigned int len; 355 bool last = false; 356 int ret; 357 u8 annotation = *_annotation; 358 u8 subpacket = annotation & RXRPC_RX_ANNO_SUBPACKET; 359 360 /* Locate the subpacket */ 361 offset += subpacket * RXRPC_JUMBO_SUBPKTLEN; 362 len = skb->len - offset; 363 if (subpacket < sp->nr_subpackets - 1) 364 len = RXRPC_JUMBO_DATALEN; 365 else if (sp->rx_flags & RXRPC_SKB_INCL_LAST) 366 last = true; 367 368 if (!(annotation & RXRPC_RX_ANNO_VERIFIED)) { 369 ret = rxrpc_verify_packet(call, skb, annotation, offset, len); 370 if (ret < 0) 371 return ret; 372 *_annotation |= RXRPC_RX_ANNO_VERIFIED; 373 } 374 375 *_offset = offset; 376 *_len = len; 377 *_last = last; 378 call->security->locate_data(call, skb, _offset, _len); 379 return 0; 380 } 381 382 /* 383 * Deliver messages to a call. This keeps processing packets until the buffer 384 * is filled and we find either more DATA (returns 0) or the end of the DATA 385 * (returns 1). If more packets are required, it returns -EAGAIN. 386 */ 387 static int rxrpc_recvmsg_data(struct socket *sock, struct rxrpc_call *call, 388 struct msghdr *msg, struct iov_iter *iter, 389 size_t len, int flags, size_t *_offset) 390 { 391 struct rxrpc_skb_priv *sp; 392 struct sk_buff *skb; 393 rxrpc_serial_t serial; 394 rxrpc_seq_t hard_ack, top, seq; 395 size_t remain; 396 bool rx_pkt_last; 397 unsigned int rx_pkt_offset, rx_pkt_len; 398 int ix, copy, ret = -EAGAIN, ret2; 399 400 if (test_and_clear_bit(RXRPC_CALL_RX_UNDERRUN, &call->flags) && 401 call->ackr_reason) 402 rxrpc_send_ack_packet(call, false, NULL); 403 404 rx_pkt_offset = call->rx_pkt_offset; 405 rx_pkt_len = call->rx_pkt_len; 406 rx_pkt_last = call->rx_pkt_last; 407 408 if (call->state >= RXRPC_CALL_SERVER_ACK_REQUEST) { 409 seq = call->rx_hard_ack; 410 ret = 1; 411 goto done; 412 } 413 414 /* Barriers against rxrpc_input_data(). */ 415 hard_ack = call->rx_hard_ack; 416 seq = hard_ack + 1; 417 418 while (top = smp_load_acquire(&call->rx_top), 419 before_eq(seq, top) 420 ) { 421 ix = seq & RXRPC_RXTX_BUFF_MASK; 422 skb = call->rxtx_buffer[ix]; 423 if (!skb) { 424 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_hole, seq, 425 rx_pkt_offset, rx_pkt_len, 0); 426 break; 427 } 428 smp_rmb(); 429 rxrpc_see_skb(skb, rxrpc_skb_seen); 430 sp = rxrpc_skb(skb); 431 432 if (!(flags & MSG_PEEK)) { 433 serial = sp->hdr.serial; 434 serial += call->rxtx_annotations[ix] & RXRPC_RX_ANNO_SUBPACKET; 435 trace_rxrpc_receive(call, rxrpc_receive_front, 436 serial, seq); 437 } 438 439 if (msg) 440 sock_recv_timestamp(msg, sock->sk, skb); 441 442 if (rx_pkt_offset == 0) { 443 ret2 = rxrpc_locate_data(call, skb, 444 &call->rxtx_annotations[ix], 445 &rx_pkt_offset, &rx_pkt_len, 446 &rx_pkt_last); 447 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_next, seq, 448 rx_pkt_offset, rx_pkt_len, ret2); 449 if (ret2 < 0) { 450 ret = ret2; 451 goto out; 452 } 453 } else { 454 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_cont, seq, 455 rx_pkt_offset, rx_pkt_len, 0); 456 } 457 458 /* We have to handle short, empty and used-up DATA packets. */ 459 remain = len - *_offset; 460 copy = rx_pkt_len; 461 if (copy > remain) 462 copy = remain; 463 if (copy > 0) { 464 ret2 = skb_copy_datagram_iter(skb, rx_pkt_offset, iter, 465 copy); 466 if (ret2 < 0) { 467 ret = ret2; 468 goto out; 469 } 470 471 /* handle piecemeal consumption of data packets */ 472 rx_pkt_offset += copy; 473 rx_pkt_len -= copy; 474 *_offset += copy; 475 } 476 477 if (rx_pkt_len > 0) { 478 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_full, seq, 479 rx_pkt_offset, rx_pkt_len, 0); 480 ASSERTCMP(*_offset, ==, len); 481 ret = 0; 482 break; 483 } 484 485 /* The whole packet has been transferred. */ 486 if (!(flags & MSG_PEEK)) 487 rxrpc_rotate_rx_window(call); 488 rx_pkt_offset = 0; 489 rx_pkt_len = 0; 490 491 if (rx_pkt_last) { 492 ASSERTCMP(seq, ==, READ_ONCE(call->rx_top)); 493 ret = 1; 494 goto out; 495 } 496 497 seq++; 498 } 499 500 out: 501 if (!(flags & MSG_PEEK)) { 502 call->rx_pkt_offset = rx_pkt_offset; 503 call->rx_pkt_len = rx_pkt_len; 504 call->rx_pkt_last = rx_pkt_last; 505 } 506 done: 507 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_data_return, seq, 508 rx_pkt_offset, rx_pkt_len, ret); 509 if (ret == -EAGAIN) 510 set_bit(RXRPC_CALL_RX_UNDERRUN, &call->flags); 511 return ret; 512 } 513 514 /* 515 * Receive a message from an RxRPC socket 516 * - we need to be careful about two or more threads calling recvmsg 517 * simultaneously 518 */ 519 int rxrpc_recvmsg(struct socket *sock, struct msghdr *msg, size_t len, 520 int flags) 521 { 522 struct rxrpc_call *call; 523 struct rxrpc_sock *rx = rxrpc_sk(sock->sk); 524 struct list_head *l; 525 size_t copied = 0; 526 long timeo; 527 int ret; 528 529 DEFINE_WAIT(wait); 530 531 trace_rxrpc_recvmsg(NULL, rxrpc_recvmsg_enter, 0, 0, 0, 0); 532 533 if (flags & (MSG_OOB | MSG_TRUNC)) 534 return -EOPNOTSUPP; 535 536 timeo = sock_rcvtimeo(&rx->sk, flags & MSG_DONTWAIT); 537 538 try_again: 539 lock_sock(&rx->sk); 540 541 /* Return immediately if a client socket has no outstanding calls */ 542 if (RB_EMPTY_ROOT(&rx->calls) && 543 list_empty(&rx->recvmsg_q) && 544 rx->sk.sk_state != RXRPC_SERVER_LISTENING) { 545 release_sock(&rx->sk); 546 return -ENODATA; 547 } 548 549 if (list_empty(&rx->recvmsg_q)) { 550 ret = -EWOULDBLOCK; 551 if (timeo == 0) { 552 call = NULL; 553 goto error_no_call; 554 } 555 556 release_sock(&rx->sk); 557 558 /* Wait for something to happen */ 559 prepare_to_wait_exclusive(sk_sleep(&rx->sk), &wait, 560 TASK_INTERRUPTIBLE); 561 ret = sock_error(&rx->sk); 562 if (ret) 563 goto wait_error; 564 565 if (list_empty(&rx->recvmsg_q)) { 566 if (signal_pending(current)) 567 goto wait_interrupted; 568 trace_rxrpc_recvmsg(NULL, rxrpc_recvmsg_wait, 569 0, 0, 0, 0); 570 timeo = schedule_timeout(timeo); 571 } 572 finish_wait(sk_sleep(&rx->sk), &wait); 573 goto try_again; 574 } 575 576 /* Find the next call and dequeue it if we're not just peeking. If we 577 * do dequeue it, that comes with a ref that we will need to release. 578 */ 579 write_lock_bh(&rx->recvmsg_lock); 580 l = rx->recvmsg_q.next; 581 call = list_entry(l, struct rxrpc_call, recvmsg_link); 582 if (!(flags & MSG_PEEK)) 583 list_del_init(&call->recvmsg_link); 584 else 585 rxrpc_get_call(call, rxrpc_call_got); 586 write_unlock_bh(&rx->recvmsg_lock); 587 588 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_dequeue, 0, 0, 0, 0); 589 590 /* We're going to drop the socket lock, so we need to lock the call 591 * against interference by sendmsg. 592 */ 593 if (!mutex_trylock(&call->user_mutex)) { 594 ret = -EWOULDBLOCK; 595 if (flags & MSG_DONTWAIT) 596 goto error_requeue_call; 597 ret = -ERESTARTSYS; 598 if (mutex_lock_interruptible(&call->user_mutex) < 0) 599 goto error_requeue_call; 600 } 601 602 release_sock(&rx->sk); 603 604 if (test_bit(RXRPC_CALL_RELEASED, &call->flags)) 605 BUG(); 606 607 if (test_bit(RXRPC_CALL_HAS_USERID, &call->flags)) { 608 if (flags & MSG_CMSG_COMPAT) { 609 unsigned int id32 = call->user_call_ID; 610 611 ret = put_cmsg(msg, SOL_RXRPC, RXRPC_USER_CALL_ID, 612 sizeof(unsigned int), &id32); 613 } else { 614 unsigned long idl = call->user_call_ID; 615 616 ret = put_cmsg(msg, SOL_RXRPC, RXRPC_USER_CALL_ID, 617 sizeof(unsigned long), &idl); 618 } 619 if (ret < 0) 620 goto error_unlock_call; 621 } 622 623 if (msg->msg_name) { 624 struct sockaddr_rxrpc *srx = msg->msg_name; 625 size_t len = sizeof(call->peer->srx); 626 627 memcpy(msg->msg_name, &call->peer->srx, len); 628 srx->srx_service = call->service_id; 629 msg->msg_namelen = len; 630 } 631 632 switch (READ_ONCE(call->state)) { 633 case RXRPC_CALL_SERVER_ACCEPTING: 634 ret = rxrpc_recvmsg_new_call(rx, call, msg, flags); 635 break; 636 case RXRPC_CALL_CLIENT_RECV_REPLY: 637 case RXRPC_CALL_SERVER_RECV_REQUEST: 638 case RXRPC_CALL_SERVER_ACK_REQUEST: 639 ret = rxrpc_recvmsg_data(sock, call, msg, &msg->msg_iter, len, 640 flags, &copied); 641 if (ret == -EAGAIN) 642 ret = 0; 643 644 if (after(call->rx_top, call->rx_hard_ack) && 645 call->rxtx_buffer[(call->rx_hard_ack + 1) & RXRPC_RXTX_BUFF_MASK]) 646 rxrpc_notify_socket(call); 647 break; 648 default: 649 ret = 0; 650 break; 651 } 652 653 if (ret < 0) 654 goto error_unlock_call; 655 656 if (call->state == RXRPC_CALL_COMPLETE) { 657 ret = rxrpc_recvmsg_term(call, msg); 658 if (ret < 0) 659 goto error_unlock_call; 660 if (!(flags & MSG_PEEK)) 661 rxrpc_release_call(rx, call); 662 msg->msg_flags |= MSG_EOR; 663 ret = 1; 664 } 665 666 if (ret == 0) 667 msg->msg_flags |= MSG_MORE; 668 else 669 msg->msg_flags &= ~MSG_MORE; 670 ret = copied; 671 672 error_unlock_call: 673 mutex_unlock(&call->user_mutex); 674 rxrpc_put_call(call, rxrpc_call_put); 675 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_return, 0, 0, 0, ret); 676 return ret; 677 678 error_requeue_call: 679 if (!(flags & MSG_PEEK)) { 680 write_lock_bh(&rx->recvmsg_lock); 681 list_add(&call->recvmsg_link, &rx->recvmsg_q); 682 write_unlock_bh(&rx->recvmsg_lock); 683 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_requeue, 0, 0, 0, 0); 684 } else { 685 rxrpc_put_call(call, rxrpc_call_put); 686 } 687 error_no_call: 688 release_sock(&rx->sk); 689 error_trace: 690 trace_rxrpc_recvmsg(call, rxrpc_recvmsg_return, 0, 0, 0, ret); 691 return ret; 692 693 wait_interrupted: 694 ret = sock_intr_errno(timeo); 695 wait_error: 696 finish_wait(sk_sleep(&rx->sk), &wait); 697 call = NULL; 698 goto error_trace; 699 } 700 701 /** 702 * rxrpc_kernel_recv_data - Allow a kernel service to receive data/info 703 * @sock: The socket that the call exists on 704 * @call: The call to send data through 705 * @iter: The buffer to receive into 706 * @want_more: True if more data is expected to be read 707 * @_abort: Where the abort code is stored if -ECONNABORTED is returned 708 * @_service: Where to store the actual service ID (may be upgraded) 709 * 710 * Allow a kernel service to receive data and pick up information about the 711 * state of a call. Returns 0 if got what was asked for and there's more 712 * available, 1 if we got what was asked for and we're at the end of the data 713 * and -EAGAIN if we need more data. 714 * 715 * Note that we may return -EAGAIN to drain empty packets at the end of the 716 * data, even if we've already copied over the requested data. 717 * 718 * *_abort should also be initialised to 0. 719 */ 720 int rxrpc_kernel_recv_data(struct socket *sock, struct rxrpc_call *call, 721 struct iov_iter *iter, 722 bool want_more, u32 *_abort, u16 *_service) 723 { 724 size_t offset = 0; 725 int ret; 726 727 _enter("{%d,%s},%zu,%d", 728 call->debug_id, rxrpc_call_states[call->state], 729 iov_iter_count(iter), want_more); 730 731 ASSERTCMP(call->state, !=, RXRPC_CALL_SERVER_ACCEPTING); 732 733 mutex_lock(&call->user_mutex); 734 735 switch (READ_ONCE(call->state)) { 736 case RXRPC_CALL_CLIENT_RECV_REPLY: 737 case RXRPC_CALL_SERVER_RECV_REQUEST: 738 case RXRPC_CALL_SERVER_ACK_REQUEST: 739 ret = rxrpc_recvmsg_data(sock, call, NULL, iter, 740 iov_iter_count(iter), 0, 741 &offset); 742 if (ret < 0) 743 goto out; 744 745 /* We can only reach here with a partially full buffer if we 746 * have reached the end of the data. We must otherwise have a 747 * full buffer or have been given -EAGAIN. 748 */ 749 if (ret == 1) { 750 if (iov_iter_count(iter) > 0) 751 goto short_data; 752 if (!want_more) 753 goto read_phase_complete; 754 ret = 0; 755 goto out; 756 } 757 758 if (!want_more) 759 goto excess_data; 760 goto out; 761 762 case RXRPC_CALL_COMPLETE: 763 goto call_complete; 764 765 default: 766 ret = -EINPROGRESS; 767 goto out; 768 } 769 770 read_phase_complete: 771 ret = 1; 772 out: 773 switch (call->ackr_reason) { 774 case RXRPC_ACK_IDLE: 775 break; 776 case RXRPC_ACK_DELAY: 777 if (ret != -EAGAIN) 778 break; 779 /* Fall through */ 780 default: 781 rxrpc_send_ack_packet(call, false, NULL); 782 } 783 784 if (_service) 785 *_service = call->service_id; 786 mutex_unlock(&call->user_mutex); 787 _leave(" = %d [%zu,%d]", ret, iov_iter_count(iter), *_abort); 788 return ret; 789 790 short_data: 791 trace_rxrpc_rx_eproto(call, 0, tracepoint_string("short_data")); 792 ret = -EBADMSG; 793 goto out; 794 excess_data: 795 trace_rxrpc_rx_eproto(call, 0, tracepoint_string("excess_data")); 796 ret = -EMSGSIZE; 797 goto out; 798 call_complete: 799 *_abort = call->abort_code; 800 ret = call->error; 801 if (call->completion == RXRPC_CALL_SUCCEEDED) { 802 ret = 1; 803 if (iov_iter_count(iter) > 0) 804 ret = -ECONNRESET; 805 } 806 goto out; 807 } 808 EXPORT_SYMBOL(rxrpc_kernel_recv_data); 809 810 /** 811 * rxrpc_kernel_get_reply_time - Get timestamp on first reply packet 812 * @sock: The socket that the call exists on 813 * @call: The call to query 814 * @_ts: Where to put the timestamp 815 * 816 * Retrieve the timestamp from the first DATA packet of the reply if it is 817 * in the ring. Returns true if successful, false if not. 818 */ 819 bool rxrpc_kernel_get_reply_time(struct socket *sock, struct rxrpc_call *call, 820 ktime_t *_ts) 821 { 822 struct sk_buff *skb; 823 rxrpc_seq_t hard_ack, top, seq; 824 bool success = false; 825 826 mutex_lock(&call->user_mutex); 827 828 if (READ_ONCE(call->state) != RXRPC_CALL_CLIENT_RECV_REPLY) 829 goto out; 830 831 hard_ack = call->rx_hard_ack; 832 if (hard_ack != 0) 833 goto out; 834 835 seq = hard_ack + 1; 836 top = smp_load_acquire(&call->rx_top); 837 if (after(seq, top)) 838 goto out; 839 840 skb = call->rxtx_buffer[seq & RXRPC_RXTX_BUFF_MASK]; 841 if (!skb) 842 goto out; 843 844 *_ts = skb_get_ktime(skb); 845 success = true; 846 847 out: 848 mutex_unlock(&call->user_mutex); 849 return success; 850 } 851 EXPORT_SYMBOL(rxrpc_kernel_get_reply_time); 852