sendmsg.c 16.4 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
/* AF_RXRPC sendmsg() implementation.
 *
 * Copyright (C) 2007, 2016 Red Hat, Inc. All Rights Reserved.
 * Written by David Howells (dhowells@redhat.com)
 *
 * This program is free software; you can redistribute it and/or
 * modify it under the terms of the GNU General Public Licence
 * as published by the Free Software Foundation; either version
 * 2 of the Licence, or (at your option) any later version.
 */

#define pr_fmt(fmt) KBUILD_MODNAME ": " fmt

#include <linux/net.h>
#include <linux/gfp.h>
#include <linux/skbuff.h>
#include <linux/export.h>
18 19
#include <linux/sched/signal.h>

20 21 22 23
#include <net/sock.h>
#include <net/af_rxrpc.h>
#include "ar-internal.h"

24 25 26 27 28 29 30
enum rxrpc_command {
	RXRPC_CMD_SEND_DATA,		/* send data message */
	RXRPC_CMD_SEND_ABORT,		/* request abort generation */
	RXRPC_CMD_ACCEPT,		/* [server] accept incoming call */
	RXRPC_CMD_REJECT_BUSY,		/* [server] reject a call as busy */
};

31
/*
32 33
 * wait for space to appear in the transmit/ACK window
 * - caller holds the socket locked
34
 */
35 36 37
static int rxrpc_wait_for_tx_window(struct rxrpc_sock *rx,
				    struct rxrpc_call *call,
				    long *timeo)
38
{
39 40
	DECLARE_WAITQUEUE(myself, current);
	int ret;
41

42 43
	_enter(",{%u,%u,%u}",
	       call->tx_hard_ack, call->tx_top, call->tx_winsize);
44

45
	add_wait_queue(&call->waitq, &myself);
46

47 48 49
	for (;;) {
		set_current_state(TASK_INTERRUPTIBLE);
		ret = 0;
D
David Howells 已提交
50 51 52
		if (call->tx_top - call->tx_hard_ack <
		    min_t(unsigned int, call->tx_winsize,
			  call->cong_cwnd + call->cong_extra))
53
			break;
54 55 56 57
		if (call->state >= RXRPC_CALL_COMPLETE) {
			ret = -call->error;
			break;
		}
58 59
		if (signal_pending(current)) {
			ret = sock_intr_errno(*timeo);
60 61
			break;
		}
62

63
		trace_rxrpc_transmit(call, rxrpc_transmit_wait);
64
		mutex_unlock(&call->user_mutex);
65
		*timeo = schedule_timeout(*timeo);
66 67 68 69
		if (mutex_lock_interruptible(&call->user_mutex) < 0) {
			ret = sock_intr_errno(*timeo);
			break;
		}
70 71
	}

72 73 74 75
	remove_wait_queue(&call->waitq, &myself);
	set_current_state(TASK_RUNNING);
	_leave(" = %d", ret);
	return ret;
76 77 78
}

/*
79
 * Schedule an instant Tx resend.
80
 */
81
static inline void rxrpc_instant_resend(struct rxrpc_call *call, int ix)
82
{
83 84 85 86 87
	spin_lock_bh(&call->lock);

	if (call->state < RXRPC_CALL_COMPLETE) {
		call->rxtx_annotations[ix] = RXRPC_TX_ANNO_RETRANS;
		if (!test_and_set_bit(RXRPC_CALL_EV_RESEND, &call->events))
88
			rxrpc_queue_call(call);
89
	}
90 91

	spin_unlock_bh(&call->lock);
92 93 94
}

/*
95 96
 * Queue a DATA packet for transmission, set the resend timeout and send the
 * packet immediately
97
 */
98 99
static void rxrpc_queue_packet(struct rxrpc_call *call, struct sk_buff *skb,
			       bool last)
100
{
101
	struct rxrpc_skb_priv *sp = rxrpc_skb(skb);
102 103
	rxrpc_seq_t seq = sp->hdr.seq;
	int ret, ix;
104
	u8 annotation = RXRPC_TX_ANNO_UNACK;
105 106

	_net("queue skb %p [%d]", skb, seq);
107

108
	ASSERTCMP(seq, ==, call->tx_top + 1);
109

110 111 112
	if (last)
		annotation |= RXRPC_TX_ANNO_LAST;

113 114 115 116 117
	/* We have to set the timestamp before queueing as the retransmit
	 * algorithm can see the packet as soon as we queue it.
	 */
	skb->tstamp = ktime_get_real();

118
	ix = seq & RXRPC_RXTX_BUFF_MASK;
D
David Howells 已提交
119
	rxrpc_get_skb(skb, rxrpc_skb_tx_got);
120
	call->rxtx_annotations[ix] = annotation;
121
	smp_wmb();
122 123
	call->rxtx_buffer[ix] = skb;
	call->tx_top = seq;
124
	if (last)
125
		trace_rxrpc_transmit(call, rxrpc_transmit_queue_last);
126
	else
127
		trace_rxrpc_transmit(call, rxrpc_transmit_queue);
128

129 130 131 132 133 134 135 136 137
	if (last || call->state == RXRPC_CALL_SERVER_ACK_REQUEST) {
		_debug("________awaiting reply/ACK__________");
		write_lock_bh(&call->state_lock);
		switch (call->state) {
		case RXRPC_CALL_CLIENT_SEND_REQUEST:
			call->state = RXRPC_CALL_CLIENT_AWAIT_REPLY;
			break;
		case RXRPC_CALL_SERVER_ACK_REQUEST:
			call->state = RXRPC_CALL_SERVER_SEND_REPLY;
138 139 140 141 142
			call->ack_at = call->expire_at;
			if (call->ackr_reason == RXRPC_ACK_DELAY)
				call->ackr_reason = 0;
			__rxrpc_set_timer(call, rxrpc_timer_init_for_send_reply,
					  ktime_get_real());
143 144 145 146 147 148 149 150 151 152
			if (!last)
				break;
		case RXRPC_CALL_SERVER_SEND_REPLY:
			call->state = RXRPC_CALL_SERVER_AWAIT_ACK;
			break;
		default:
			break;
		}
		write_unlock_bh(&call->state_lock);
	}
153

154 155
	if (seq == 1 && rxrpc_is_client_call(call))
		rxrpc_expose_client_call(call);
156

157
	ret = rxrpc_send_data_packet(call, skb, false);
158 159
	if (ret < 0) {
		_debug("need instant resend %d", ret);
160
		rxrpc_instant_resend(call, ix);
161
	} else {
162
		ktime_t now = ktime_get_real(), resend_at;
163

164
		resend_at = ktime_add_ms(now, rxrpc_resend_timeout);
165

166
		if (ktime_before(resend_at, call->resend_at)) {
167
			call->resend_at = resend_at;
168
			rxrpc_set_timer(call, rxrpc_timer_set_for_send, now);
169
		}
170 171
	}

D
David Howells 已提交
172
	rxrpc_free_skb(skb, rxrpc_skb_tx_freed);
173
	_leave("");
174 175
}

176 177 178
/*
 * send data through a socket
 * - must be called in process context
179
 * - The caller holds the call user access mutex, but not the socket lock.
180
 */
181 182 183
static int rxrpc_send_data(struct rxrpc_sock *rx,
			   struct rxrpc_call *call,
			   struct msghdr *msg, size_t len)
184
{
185 186 187 188 189 190
	struct rxrpc_skb_priv *sp;
	struct sk_buff *skb;
	struct sock *sk = &rx->sk;
	long timeo;
	bool more;
	int ret, copied;
191

192
	timeo = sock_sndtimeo(sk, msg->msg_flags & MSG_DONTWAIT);
193

194 195
	/* this should be in poll */
	sk_clear_bit(SOCKWQ_ASYNC_NOSPACE, sk);
196

197 198
	if (sk->sk_err || (sk->sk_shutdown & SEND_SHUTDOWN))
		return -EPIPE;
199

200
	more = msg->msg_flags & MSG_MORE;
201

202 203
	skb = call->tx_pending;
	call->tx_pending = NULL;
D
David Howells 已提交
204
	rxrpc_see_skb(skb, rxrpc_skb_tx_seen);
205

206 207
	copied = 0;
	do {
208 209
		/* Check to see if there's a ping ACK to reply to. */
		if (call->ackr_reason == RXRPC_ACK_PING_RESPONSE)
210
			rxrpc_send_ack_packet(call, false);
211

212 213
		if (!skb) {
			size_t size, chunk, max, space;
214

215
			_debug("alloc");
216

217
			if (call->tx_top - call->tx_hard_ack >=
D
David Howells 已提交
218 219
			    min_t(unsigned int, call->tx_winsize,
				  call->cong_cwnd + call->cong_extra)) {
220 221 222 223 224 225 226 227
				ret = -EAGAIN;
				if (msg->msg_flags & MSG_DONTWAIT)
					goto maybe_error;
				ret = rxrpc_wait_for_tx_window(rx, call,
							       &timeo);
				if (ret < 0)
					goto maybe_error;
			}
228

229
			max = RXRPC_JUMBO_DATALEN;
230 231
			max -= call->conn->security_size;
			max &= ~(call->conn->size_align - 1UL);
232

233 234 235
			chunk = max;
			if (chunk > msg_data_left(msg) && !more)
				chunk = msg_data_left(msg);
236

237 238
			space = chunk + call->conn->size_align;
			space &= ~(call->conn->size_align - 1UL);
239

240
			size = space + call->conn->security_size;
241

242
			_debug("SIZE: %zu/%zu/%zu", chunk, space, size);
243

244 245 246 247 248
			/* create a buffer that we can retain until it's ACK'd */
			skb = sock_alloc_send_skb(
				sk, size, msg->msg_flags & MSG_DONTWAIT, &ret);
			if (!skb)
				goto maybe_error;
249

D
David Howells 已提交
250
			rxrpc_new_skb(skb, rxrpc_skb_tx_new);
251

252
			_debug("ALLOC SEND %p", skb);
253

254
			ASSERTCMP(skb->mark, ==, 0);
255

256 257 258
			_debug("HS: %u", call->conn->security_size);
			skb_reserve(skb, call->conn->security_size);
			skb->len += call->conn->security_size;
259

260 261 262 263
			sp = rxrpc_skb(skb);
			sp->remain = chunk;
			if (sp->remain > skb_tailroom(skb))
				sp->remain = skb_tailroom(skb);
264

265 266 267 268 269
			_net("skb: hr %d, tr %d, hl %d, rm %d",
			       skb_headroom(skb),
			       skb_tailroom(skb),
			       skb_headlen(skb),
			       sp->remain);
270

271 272
			skb->ip_summed = CHECKSUM_UNNECESSARY;
		}
273

274 275
		_debug("append");
		sp = rxrpc_skb(skb);
276

277 278 279 280 281 282 283 284
		/* append next segment of data to the current buffer */
		if (msg_data_left(msg) > 0) {
			int copy = skb_tailroom(skb);
			ASSERTCMP(copy, >, 0);
			if (copy > msg_data_left(msg))
				copy = msg_data_left(msg);
			if (copy > sp->remain)
				copy = sp->remain;
285

286 287 288 289 290 291 292 293
			_debug("add");
			ret = skb_add_data(skb, &msg->msg_iter, copy);
			_debug("added");
			if (ret < 0)
				goto efault;
			sp->remain -= copy;
			skb->mark += copy;
			copied += copy;
294 295
		}

296 297 298 299
		/* check for the far side aborting the call or a network error
		 * occurring */
		if (call->state == RXRPC_CALL_COMPLETE)
			goto call_terminated;
300

301 302 303 304 305 306
		/* add the packet to the send queue if it's now full */
		if (sp->remain <= 0 ||
		    (msg_data_left(msg) == 0 && !more)) {
			struct rxrpc_connection *conn = call->conn;
			uint32_t seq;
			size_t pad;
307

308 309 310 311 312 313 314 315 316
			/* pad out if we're using security */
			if (conn->security_ix) {
				pad = conn->security_size + skb->mark;
				pad = conn->size_align - pad;
				pad &= conn->size_align - 1;
				_debug("pad %zu", pad);
				if (pad)
					memset(skb_put(skb, pad), 0, pad);
			}
317

318
			seq = call->tx_top + 1;
319

320 321
			sp->hdr.seq	= seq;
			sp->hdr._rsvd	= 0;
322
			sp->hdr.flags	= conn->out_clientflag;
323

324 325
			if (msg_data_left(msg) == 0 && !more)
				sp->hdr.flags |= RXRPC_LAST_PACKET;
326 327
			else if (call->tx_top - call->tx_hard_ack <
				 call->tx_winsize)
328
				sp->hdr.flags |= RXRPC_MORE_PACKETS;
329

330
			ret = conn->security->secure_packet(
331
				call, skb, skb->mark, skb->head);
332 333
			if (ret < 0)
				goto out;
334

335 336 337 338
			rxrpc_queue_packet(call, skb, !msg_data_left(msg) && !more);
			skb = NULL;
		}
	} while (msg_data_left(msg) > 0);
339

340 341 342 343 344 345
success:
	ret = copied;
out:
	call->tx_pending = skb;
	_leave(" = %d", ret);
	return ret;
346

347
call_terminated:
D
David Howells 已提交
348
	rxrpc_free_skb(skb, rxrpc_skb_tx_freed);
349
	_leave(" = %d", -call->error);
350
	return -call->error;
351

352 353 354 355
maybe_error:
	if (copied)
		goto success;
	goto out;
356

357 358 359
efault:
	ret = -EFAULT;
	goto out;
360 361 362
}

/*
363
 * extract control messages from the sendmsg() control buffer
364
 */
365 366 367 368 369
static int rxrpc_sendmsg_cmsg(struct msghdr *msg,
			      unsigned long *user_call_ID,
			      enum rxrpc_command *command,
			      u32 *abort_code,
			      bool *_exclusive)
370
{
371 372 373
	struct cmsghdr *cmsg;
	bool got_user_ID = false;
	int len;
374

375
	*command = RXRPC_CMD_SEND_DATA;
376

377 378
	if (msg->msg_controllen == 0)
		return -EINVAL;
379

380 381 382
	for_each_cmsghdr(cmsg, msg) {
		if (!CMSG_OK(msg, cmsg))
			return -EINVAL;
383

384
		len = cmsg->cmsg_len - sizeof(struct cmsghdr);
385 386
		_debug("CMSG %d, %d, %d",
		       cmsg->cmsg_level, cmsg->cmsg_type, len);
387

388 389
		if (cmsg->cmsg_level != SOL_RXRPC)
			continue;
390

391 392 393 394 395 396 397 398 399 400 401 402 403 404 405
		switch (cmsg->cmsg_type) {
		case RXRPC_USER_CALL_ID:
			if (msg->msg_flags & MSG_CMSG_COMPAT) {
				if (len != sizeof(u32))
					return -EINVAL;
				*user_call_ID = *(u32 *) CMSG_DATA(cmsg);
			} else {
				if (len != sizeof(unsigned long))
					return -EINVAL;
				*user_call_ID = *(unsigned long *)
					CMSG_DATA(cmsg);
			}
			_debug("User Call ID %lx", *user_call_ID);
			got_user_ID = true;
			break;
406

407 408 409 410 411 412 413 414 415 416 417
		case RXRPC_ABORT:
			if (*command != RXRPC_CMD_SEND_DATA)
				return -EINVAL;
			*command = RXRPC_CMD_SEND_ABORT;
			if (len != sizeof(*abort_code))
				return -EINVAL;
			*abort_code = *(unsigned int *) CMSG_DATA(cmsg);
			_debug("Abort %x", *abort_code);
			if (*abort_code == 0)
				return -EINVAL;
			break;
418

419 420 421 422 423 424 425
		case RXRPC_ACCEPT:
			if (*command != RXRPC_CMD_SEND_DATA)
				return -EINVAL;
			*command = RXRPC_CMD_ACCEPT;
			if (len != 0)
				return -EINVAL;
			break;
426

427 428 429 430 431 432 433 434 435
		case RXRPC_EXCLUSIVE_CALL:
			*_exclusive = true;
			if (len != 0)
				return -EINVAL;
			break;
		default:
			return -EINVAL;
		}
	}
436

437 438 439 440 441
	if (!got_user_ID)
		return -EINVAL;
	_leave(" = 0");
	return 0;
}
442

443 444
/*
 * Create a new client call for sendmsg().
445 446
 * - Called with the socket lock held, which it must release.
 * - If it returns a call, the call's lock will need releasing by the caller.
447 448 449 450
 */
static struct rxrpc_call *
rxrpc_new_client_call_for_sendmsg(struct rxrpc_sock *rx, struct msghdr *msg,
				  unsigned long user_call_ID, bool exclusive)
451
	__releases(&rx->sk.sk_lock.slock)
452 453 454 455
{
	struct rxrpc_conn_parameters cp;
	struct rxrpc_call *call;
	struct key *key;
456

457
	DECLARE_SOCKADDR(struct sockaddr_rxrpc *, srx, msg->msg_name);
458

459
	_enter("");
460

461 462
	if (!msg->msg_name) {
		release_sock(&rx->sk);
463
		return ERR_PTR(-EDESTADDRREQ);
464
	}
465

466 467 468
	key = rx->key;
	if (key && !rx->key->payload.data[0])
		key = NULL;
469

470 471 472 473 474 475 476
	memset(&cp, 0, sizeof(cp));
	cp.local		= rx->local;
	cp.key			= rx->key;
	cp.security_level	= rx->min_sec_level;
	cp.exclusive		= rx->exclusive | exclusive;
	cp.service_id		= srx->srx_service;
	call = rxrpc_new_client_call(rx, &cp, srx, user_call_ID, GFP_KERNEL);
477
	/* The socket is now unlocked */
478

479 480 481
	_leave(" = %p\n", call);
	return call;
}
482

483 484 485 486 487 488
/*
 * send a message forming part of a client call through an RxRPC socket
 * - caller holds the socket locked
 * - the socket may be either a client socket or a server socket
 */
int rxrpc_do_sendmsg(struct rxrpc_sock *rx, struct msghdr *msg, size_t len)
489
	__releases(&rx->sk.sk_lock.slock)
490 491 492 493 494 495 496
{
	enum rxrpc_command cmd;
	struct rxrpc_call *call;
	unsigned long user_call_ID = 0;
	bool exclusive = false;
	u32 abort_code = 0;
	int ret;
497

498
	_enter("");
499

500 501 502
	ret = rxrpc_sendmsg_cmsg(msg, &user_call_ID, &cmd, &abort_code,
				 &exclusive);
	if (ret < 0)
503
		goto error_release_sock;
504

505
	if (cmd == RXRPC_CMD_ACCEPT) {
506
		ret = -EINVAL;
507
		if (rx->sk.sk_state != RXRPC_SERVER_LISTENING)
508
			goto error_release_sock;
509
		call = rxrpc_accept_call(rx, user_call_ID, NULL);
510
		/* The socket is now unlocked. */
511 512
		if (IS_ERR(call))
			return PTR_ERR(call);
513
		rxrpc_put_call(call, rxrpc_call_put);
514 515
		return 0;
	}
516

517 518
	call = rxrpc_find_call_by_user_ID(rx, user_call_ID);
	if (!call) {
519
		ret = -EBADSLT;
520
		if (cmd != RXRPC_CMD_SEND_DATA)
521
			goto error_release_sock;
522 523
		call = rxrpc_new_client_call_for_sendmsg(rx, msg, user_call_ID,
							 exclusive);
524
		/* The socket is now unlocked... */
525 526
		if (IS_ERR(call))
			return PTR_ERR(call);
527 528
		/* ... and we have the call lock. */
	} else {
529 530 531 532 533 534 535 536
		ret = -EBUSY;
		if (call->state == RXRPC_CALL_UNINITIALISED ||
		    call->state == RXRPC_CALL_CLIENT_AWAIT_CONN ||
		    call->state == RXRPC_CALL_SERVER_PREALLOC ||
		    call->state == RXRPC_CALL_SERVER_SECURING ||
		    call->state == RXRPC_CALL_SERVER_ACCEPTING)
			goto error_release_sock;

537 538 539 540 541 542
		ret = mutex_lock_interruptible(&call->user_mutex);
		release_sock(&rx->sk);
		if (ret < 0) {
			ret = -ERESTARTSYS;
			goto error_put;
		}
543
	}
544

545 546
	_debug("CALL %d USR %lx ST %d on CONN %p",
	       call->debug_id, call->user_call_ID, call->state, call->conn);
547

548 549 550 551 552
	if (call->state >= RXRPC_CALL_COMPLETE) {
		/* it's too late for this call */
		ret = -ESHUTDOWN;
	} else if (cmd == RXRPC_CMD_SEND_ABORT) {
		ret = 0;
553
		if (rxrpc_abort_call("CMD", call, 0, abort_code, ECONNABORTED))
554
			ret = rxrpc_send_abort_packet(call);
555 556 557 558 559 560 561 562 563 564 565 566 567 568
	} else if (cmd != RXRPC_CMD_SEND_DATA) {
		ret = -EINVAL;
	} else if (rxrpc_is_client_call(call) &&
		   call->state != RXRPC_CALL_CLIENT_SEND_REQUEST) {
		/* request phase complete for this client call */
		ret = -EPROTO;
	} else if (rxrpc_is_service_call(call) &&
		   call->state != RXRPC_CALL_SERVER_ACK_REQUEST &&
		   call->state != RXRPC_CALL_SERVER_SEND_REPLY) {
		/* Reply phase not begun or not complete for service call. */
		ret = -EPROTO;
	} else {
		ret = rxrpc_send_data(rx, call, msg, len);
	}
569

570 571
	mutex_unlock(&call->user_mutex);
error_put:
572
	rxrpc_put_call(call, rxrpc_call_put);
573 574
	_leave(" = %d", ret);
	return ret;
575 576 577 578

error_release_sock:
	release_sock(&rx->sk);
	return ret;
579
}
580

581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596
/**
 * rxrpc_kernel_send_data - Allow a kernel service to send data on a call
 * @sock: The socket the call is on
 * @call: The call to send data through
 * @msg: The data to send
 * @len: The amount of data to send
 *
 * Allow a kernel service to send data on a call.  The call must be in an state
 * appropriate to sending data.  No control data should be supplied in @msg,
 * nor should an address be supplied.  MSG_MORE should be flagged if there's
 * more data to come, otherwise this data will end the transmission phase.
 */
int rxrpc_kernel_send_data(struct socket *sock, struct rxrpc_call *call,
			   struct msghdr *msg, size_t len)
{
	int ret;
597

598
	_enter("{%d,%s},", call->debug_id, rxrpc_call_states[call->state]);
599

600 601
	ASSERTCMP(msg->msg_name, ==, NULL);
	ASSERTCMP(msg->msg_control, ==, NULL);
602

603
	mutex_lock(&call->user_mutex);
604

605 606
	_debug("CALL %d USR %lx ST %d on CONN %p",
	       call->debug_id, call->user_call_ID, call->state, call->conn);
607

608 609 610 611 612 613 614 615 616 617
	if (call->state >= RXRPC_CALL_COMPLETE) {
		ret = -ESHUTDOWN; /* it's too late for this call */
	} else if (call->state != RXRPC_CALL_CLIENT_SEND_REQUEST &&
		   call->state != RXRPC_CALL_SERVER_ACK_REQUEST &&
		   call->state != RXRPC_CALL_SERVER_SEND_REPLY) {
		ret = -EPROTO; /* request phase complete for this client call */
	} else {
		ret = rxrpc_send_data(rxrpc_sk(sock->sk), call, msg, len);
	}

618
	mutex_unlock(&call->user_mutex);
619 620
	_leave(" = %d", ret);
	return ret;
621 622
}
EXPORT_SYMBOL(rxrpc_kernel_send_data);
623

624 625 626 627 628
/**
 * rxrpc_kernel_abort_call - Allow a kernel service to abort a call
 * @sock: The socket the call is on
 * @call: The call to be aborted
 * @abort_code: The abort code to stick into the ABORT packet
629 630
 * @error: Local error value
 * @why: 3-char string indicating why.
631 632 633 634
 *
 * Allow a kernel service to abort a call, if it's still in an abortable state.
 */
void rxrpc_kernel_abort_call(struct socket *sock, struct rxrpc_call *call,
635
			     u32 abort_code, int error, const char *why)
636
{
637
	_enter("{%d},%d,%d,%s", call->debug_id, abort_code, error, why);
638

639
	mutex_lock(&call->user_mutex);
640

641
	if (rxrpc_abort_call(why, call, 0, abort_code, error))
642
		rxrpc_send_abort_packet(call);
643

644
	mutex_unlock(&call->user_mutex);
645
	_leave("");
646
}
647 648

EXPORT_SYMBOL(rxrpc_kernel_abort_call);