caif_socket.c 30.3 KB
Newer Older
1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17
/*
 * Copyright (C) ST-Ericsson AB 2010
 * Author:	Sjur Brendeland sjur.brandeland@stericsson.com
 * License terms: GNU General Public License (GPL) version 2
 */

#include <linux/fs.h>
#include <linux/init.h>
#include <linux/module.h>
#include <linux/sched.h>
#include <linux/spinlock.h>
#include <linux/mutex.h>
#include <linux/list.h>
#include <linux/wait.h>
#include <linux/poll.h>
#include <linux/tcp.h>
#include <linux/uaccess.h>
18 19
#include <linux/mutex.h>
#include <linux/debugfs.h>
20
#include <linux/caif/caif_socket.h>
21 22 23
#include <asm/atomic.h>
#include <net/sock.h>
#include <net/tcp_states.h>
24 25 26 27 28
#include <net/caif/caif_layer.h>
#include <net/caif/caif_dev.h>
#include <net/caif/cfpkt.h>

MODULE_LICENSE("GPL");
29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46
MODULE_ALIAS_NETPROTO(AF_CAIF);

#define CAIF_DEF_SNDBUF (CAIF_MAX_PAYLOAD_SIZE*10)
#define CAIF_DEF_RCVBUF (CAIF_MAX_PAYLOAD_SIZE*100)

/*
 * CAIF state is re-using the TCP socket states.
 * caif_states stored in sk_state reflect the state as reported by
 * the CAIF stack, while sk_socket->state is the state of the socket.
 */
enum caif_states {
	CAIF_CONNECTED		= TCP_ESTABLISHED,
	CAIF_CONNECTING	= TCP_SYN_SENT,
	CAIF_DISCONNECTED	= TCP_CLOSE
};

#define TX_FLOW_ON_BIT	1
#define RX_FLOW_ON_BIT	2
47 48 49 50 51

static struct dentry *debugfsdir;

#ifdef CONFIG_DEBUG_FS
struct debug_fs_counter {
52 53 54 55 56
	atomic_t caif_nr_socks;
	atomic_t num_connect_req;
	atomic_t num_connect_resp;
	atomic_t num_connect_fail_resp;
	atomic_t num_disconnect;
57 58 59 60 61 62
	atomic_t num_remote_shutdown_ind;
	atomic_t num_tx_flow_off_ind;
	atomic_t num_tx_flow_on_ind;
	atomic_t num_rx_flow_off;
	atomic_t num_rx_flow_on;
};
63
struct debug_fs_counter cnt;
64 65 66 67 68 69 70 71
#define	dbfs_atomic_inc(v) atomic_inc(v)
#define	dbfs_atomic_dec(v) atomic_dec(v)
#else
#define	dbfs_atomic_inc(v)
#define	dbfs_atomic_dec(v)
#endif

struct caifsock {
72
	struct sock sk; /* must be first member */
73
	struct cflayer layer;
74
	char name[CAIF_LAYER_NAME_SZ]; /* Used for debugging */
75 76
	u32 flow_state;
	struct caif_connect_request conn_req;
77
	struct mutex readlock;
78 79 80
	struct dentry *debugfs_socket_dir;
};

81 82 83 84 85 86 87 88 89 90 91
static int rx_flow_is_on(struct caifsock *cf_sk)
{
	return test_bit(RX_FLOW_ON_BIT,
			(void *) &cf_sk->flow_state);
}

static int tx_flow_is_on(struct caifsock *cf_sk)
{
	return test_bit(TX_FLOW_ON_BIT,
			(void *) &cf_sk->flow_state);
}
92

93
static void set_rx_flow_off(struct caifsock *cf_sk)
94
{
95 96 97
	 clear_bit(RX_FLOW_ON_BIT,
		 (void *) &cf_sk->flow_state);
}
98

99 100 101 102 103
static void set_rx_flow_on(struct caifsock *cf_sk)
{
	 set_bit(RX_FLOW_ON_BIT,
			(void *) &cf_sk->flow_state);
}
104

105 106 107 108 109
static void set_tx_flow_off(struct caifsock *cf_sk)
{
	 clear_bit(TX_FLOW_ON_BIT,
		(void *) &cf_sk->flow_state);
}
110

111 112 113 114 115
static void set_tx_flow_on(struct caifsock *cf_sk)
{
	 set_bit(TX_FLOW_ON_BIT,
		(void *) &cf_sk->flow_state);
}
116

117 118 119 120 121 122
static void caif_read_lock(struct sock *sk)
{
	struct caifsock *cf_sk;
	cf_sk = container_of(sk, struct caifsock, sk);
	mutex_lock(&cf_sk->readlock);
}
123

124 125 126 127 128 129
static void caif_read_unlock(struct sock *sk)
{
	struct caifsock *cf_sk;
	cf_sk = container_of(sk, struct caifsock, sk);
	mutex_unlock(&cf_sk->readlock);
}
130

131 132 133 134 135
int sk_rcvbuf_lowwater(struct caifsock *cf_sk)
{
	/* A quarter of full buffer is used a low water mark */
	return cf_sk->sk.sk_rcvbuf / 4;
}
136

137 138 139 140
void caif_flow_ctrl(struct sock *sk, int mode)
{
	struct caifsock *cf_sk;
	cf_sk = container_of(sk, struct caifsock, sk);
141
	if (cf_sk->layer.dn && cf_sk->layer.dn->modemcmd)
142 143
		cf_sk->layer.dn->modemcmd(cf_sk->layer.dn, mode);
}
144

145 146 147 148 149 150 151 152 153 154 155
/*
 * Copied from sock.c:sock_queue_rcv_skb(), but changed so packets are
 * not dropped, but CAIF is sending flow off instead.
 */
int caif_queue_rcv_skb(struct sock *sk, struct sk_buff *skb)
{
	int err;
	int skb_len;
	unsigned long flags;
	struct sk_buff_head *list = &sk->sk_receive_queue;
	struct caifsock *cf_sk = container_of(sk, struct caifsock, sk);
156

157 158 159 160 161 162 163 164
	if (atomic_read(&sk->sk_rmem_alloc) + skb->truesize >=
		(unsigned)sk->sk_rcvbuf && rx_flow_is_on(cf_sk)) {
		trace_printk("CAIF: %s():"
			" sending flow OFF (queue len = %d %d)\n",
			__func__,
			atomic_read(&cf_sk->sk.sk_rmem_alloc),
			sk_rcvbuf_lowwater(cf_sk));
		set_rx_flow_off(cf_sk);
165 166
		dbfs_atomic_inc(&cnt.num_rx_flow_off);
		caif_flow_ctrl(sk, CAIF_MODEMCMD_FLOW_OFF_REQ);
167
	}
168

169 170 171 172 173 174 175 176
	err = sk_filter(sk, skb);
	if (err)
		return err;
	if (!sk_rmem_schedule(sk, skb->truesize) && rx_flow_is_on(cf_sk)) {
		set_rx_flow_off(cf_sk);
		trace_printk("CAIF: %s():"
			" sending flow OFF due to rmem_schedule\n",
			__func__);
177 178
		dbfs_atomic_inc(&cnt.num_rx_flow_off);
		caif_flow_ctrl(sk, CAIF_MODEMCMD_FLOW_OFF_REQ);
179 180 181 182 183 184 185 186 187 188 189 190 191 192 193 194 195 196
	}
	skb->dev = NULL;
	skb_set_owner_r(skb, sk);
	/* Cache the SKB length before we tack it onto the receive
	 * queue. Once it is added it no longer belongs to us and
	 * may be freed by other threads of control pulling packets
	 * from the queue.
	 */
	skb_len = skb->len;
	spin_lock_irqsave(&list->lock, flags);
	if (!sock_flag(sk, SOCK_DEAD))
		__skb_queue_tail(list, skb);
	spin_unlock_irqrestore(&list->lock, flags);

	if (!sock_flag(sk, SOCK_DEAD))
		sk->sk_data_ready(sk, skb_len);
	else
		kfree_skb(skb);
197 198 199
	return 0;
}

200 201
/* Packet Receive Callback function called from CAIF Stack */
static int caif_sktrecv_cb(struct cflayer *layr, struct cfpkt *pkt)
202 203
{
	struct caifsock *cf_sk;
204
	struct sk_buff *skb;
205 206

	cf_sk = container_of(layr, struct caifsock, layer);
207 208 209 210 211 212 213 214 215
	skb = cfpkt_tonative(pkt);

	if (unlikely(cf_sk->sk.sk_state != CAIF_CONNECTED)) {
		cfpkt_destroy(pkt);
		return 0;
	}
	caif_queue_rcv_skb(&cf_sk->sk, skb);
	return 0;
}
216

217 218 219 220 221 222
/* Packet Control Callback function called from CAIF */
static void caif_ctrl_cb(struct cflayer *layr,
				enum caif_ctrlcmd flow,
				int phyid)
{
	struct caifsock *cf_sk = container_of(layr, struct caifsock, layer);
223 224
	switch (flow) {
	case CAIF_CTRLCMD_FLOW_ON_IND:
225
		/* OK from modem to start sending again */
226
		dbfs_atomic_inc(&cnt.num_tx_flow_on_ind);
227 228
		set_tx_flow_on(cf_sk);
		cf_sk->sk.sk_state_change(&cf_sk->sk);
229 230 231
		break;

	case CAIF_CTRLCMD_FLOW_OFF_IND:
232
		/* Modem asks us to shut up */
233
		dbfs_atomic_inc(&cnt.num_tx_flow_off_ind);
234 235
		set_tx_flow_off(cf_sk);
		cf_sk->sk.sk_state_change(&cf_sk->sk);
236 237 238
		break;

	case CAIF_CTRLCMD_INIT_RSP:
239 240 241 242 243
		/* We're now connected */
		dbfs_atomic_inc(&cnt.num_connect_resp);
		cf_sk->sk.sk_state = CAIF_CONNECTED;
		set_tx_flow_on(cf_sk);
		cf_sk->sk.sk_state_change(&cf_sk->sk);
244 245 246
		break;

	case CAIF_CTRLCMD_DEINIT_RSP:
247 248 249 250
		/* We're now disconnected */
		cf_sk->sk.sk_state = CAIF_DISCONNECTED;
		cf_sk->sk.sk_state_change(&cf_sk->sk);
		cfcnfg_release_adap_layer(&cf_sk->layer);
251 252 253
		break;

	case CAIF_CTRLCMD_INIT_FAIL_RSP:
254 255 256 257 258 259 260 261 262 263 264
		/* Connect request failed */
		dbfs_atomic_inc(&cnt.num_connect_fail_resp);
		cf_sk->sk.sk_err = ECONNREFUSED;
		cf_sk->sk.sk_state = CAIF_DISCONNECTED;
		cf_sk->sk.sk_shutdown = SHUTDOWN_MASK;
		/*
		 * Socket "standards" seems to require POLLOUT to
		 * be set at connect failure.
		 */
		set_tx_flow_on(cf_sk);
		cf_sk->sk.sk_state_change(&cf_sk->sk);
265 266 267
		break;

	case CAIF_CTRLCMD_REMOTE_SHUTDOWN_IND:
268
		/* Modem has closed this connection, or device is down. */
269
		dbfs_atomic_inc(&cnt.num_remote_shutdown_ind);
270 271 272 273
		cf_sk->sk.sk_shutdown = SHUTDOWN_MASK;
		cf_sk->sk.sk_err = ECONNRESET;
		set_rx_flow_on(cf_sk);
		cf_sk->sk.sk_error_report(&cf_sk->sk);
274 275 276 277
		break;

	default:
		pr_debug("CAIF: %s(): Unexpected flow command %d\n",
278
				__func__, flow);
279 280 281
	}
}

282
static void caif_check_flow_release(struct sock *sk)
283
{
284
	struct caifsock *cf_sk = container_of(sk, struct caifsock, sk);
285

286 287
	if (rx_flow_is_on(cf_sk))
		return;
288

289 290 291
	if (atomic_read(&sk->sk_rmem_alloc) <= sk_rcvbuf_lowwater(cf_sk)) {
			dbfs_atomic_inc(&cnt.num_rx_flow_on);
			set_rx_flow_on(cf_sk);
292
			caif_flow_ctrl(sk, CAIF_MODEMCMD_FLOW_ON_REQ);
293 294 295 296 297 298 299 300
	}
}
/*
 * Copied from sock.c:sock_queue_rcv_skb(), and added check that user buffer
 * has sufficient size.
 */

static int caif_seqpkt_recvmsg(struct kiocb *iocb, struct socket *sock,
301 302 303 304 305
				struct msghdr *m, size_t buf_len, int flags)

{
	struct sock *sk = sock->sk;
	struct sk_buff *skb;
306 307
	int ret = 0;
	int len;
308 309 310 311

	if (unlikely(!buf_len))
		return -EINVAL;

312 313
	skb = skb_recv_datagram(sk, flags, 0 , &ret);
	if (!skb)
314 315
		goto read_error;

316 317 318 319
	len = skb->len;

	if (skb && skb->len > buf_len && !(flags & MSG_PEEK)) {
		len = buf_len;
320
		/*
321 322 323 324 325
		 * Push skb back on receive queue if buffer too small.
		 * This has a built-in race where multi-threaded receive
		 * may get packet in wrong order, but multiple read does
		 * not really guarantee ordered delivery anyway.
		 * Let's optimize for speed without taking locks.
326 327
		 */

328 329 330
		skb_queue_head(&sk->sk_receive_queue, skb);
		ret = -EMSGSIZE;
		goto read_error;
331 332
	}

333 334
	ret = skb_copy_datagram_iovec(skb, 0, m->msg_iov, len);
	if (ret)
335 336
		goto read_error;

337
	skb_free_datagram(sk, skb);
338

339
	caif_check_flow_release(sk);
340

341
	return len;
342

343 344 345
read_error:
	return ret;
}
346 347


348 349 350 351 352 353 354 355 356 357 358 359 360 361 362 363 364
/* Copied from unix_stream_wait_data, identical except for lock call. */
static long caif_stream_data_wait(struct sock *sk, long timeo)
{
	DEFINE_WAIT(wait);
	lock_sock(sk);

	for (;;) {
		prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);

		if (!skb_queue_empty(&sk->sk_receive_queue) ||
			sk->sk_err ||
			sk->sk_state != CAIF_CONNECTED ||
			sock_flag(sk, SOCK_DEAD) ||
			(sk->sk_shutdown & RCV_SHUTDOWN) ||
			signal_pending(current) ||
			!timeo)
			break;
365

366 367 368 369 370
		set_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
		release_sock(sk);
		timeo = schedule_timeout(timeo);
		lock_sock(sk);
		clear_bit(SOCK_ASYNC_WAITDATA, &sk->sk_socket->flags);
371 372
	}

373 374 375 376
	finish_wait(sk_sleep(sk), &wait);
	release_sock(sk);
	return timeo;
}
377 378


379 380 381 382 383 384 385 386 387 388 389 390 391
/*
 * Copied from unix_stream_recvmsg, but removed credit checks,
 * changed locking calls, changed address handling.
 */
static int caif_stream_recvmsg(struct kiocb *iocb, struct socket *sock,
				struct msghdr *msg, size_t size,
				int flags)
{
	struct sock *sk = sock->sk;
	int copied = 0;
	int target;
	int err = 0;
	long timeo;
392

393 394 395
	err = -EOPNOTSUPP;
	if (flags&MSG_OOB)
		goto out;
396

397
	msg->msg_namelen = 0;
398

399 400 401 402 403 404 405
	/*
	 * Lock the socket to prevent queue disordering
	 * while sleeps in memcpy_tomsg
	 */
	err = -EAGAIN;
	if (sk->sk_state == CAIF_CONNECTING)
		goto out;
406

407 408 409
	caif_read_lock(sk);
	target = sock_rcvlowat(sk, flags&MSG_WAITALL, size);
	timeo = sock_rcvtimeo(sk, flags&MSG_DONTWAIT);
410

411 412 413
	do {
		int chunk;
		struct sk_buff *skb;
414

415 416 417
		lock_sock(sk);
		skb = skb_dequeue(&sk->sk_receive_queue);
		caif_check_flow_release(sk);
418

419 420 421 422 423 424 425 426 427 428 429 430
		if (skb == NULL) {
			if (copied >= target)
				goto unlock;
			/*
			 *	POSIX 1003.1g mandates this order.
			 */
			err = sock_error(sk);
			if (err)
				goto unlock;
			err = -ECONNRESET;
			if (sk->sk_shutdown & RCV_SHUTDOWN)
				goto unlock;
431

432 433 434 435 436
			err = -EPIPE;
			if (sk->sk_state != CAIF_CONNECTED)
				goto unlock;
			if (sock_flag(sk, SOCK_DEAD))
				goto unlock;
437

438
			release_sock(sk);
439

440 441 442
			err = -EAGAIN;
			if (!timeo)
				break;
443

444
			caif_read_unlock(sk);
445

446
			timeo = caif_stream_data_wait(sk, timeo);
447

448 449 450 451 452 453 454 455 456
			if (signal_pending(current)) {
				err = sock_intr_errno(timeo);
				goto out;
			}
			caif_read_lock(sk);
			continue;
unlock:
			release_sock(sk);
			break;
457
		}
458 459 460 461 462 463 464 465 466 467
		release_sock(sk);
		chunk = min_t(unsigned int, skb->len, size);
		if (memcpy_toiovec(msg->msg_iov, skb->data, chunk)) {
			skb_queue_head(&sk->sk_receive_queue, skb);
			if (copied == 0)
				copied = -EFAULT;
			break;
		}
		copied += chunk;
		size -= chunk;
468

469 470 471
		/* Mark read part of skb as used */
		if (!(flags & MSG_PEEK)) {
			skb_pull(skb, chunk);
472

473 474 475 476 477 478
			/* put the skb back if we didn't use it up. */
			if (skb->len) {
				skb_queue_head(&sk->sk_receive_queue, skb);
				break;
			}
			kfree_skb(skb);
479 480

		} else {
481 482 483 484 485 486
			/*
			 * It is questionable, see note in unix_dgram_recvmsg.
			 */
			/* put message back and return */
			skb_queue_head(&sk->sk_receive_queue, skb);
			break;
487
		}
488 489
	} while (size);
	caif_read_unlock(sk);
490

491 492 493
out:
	return copied ? : err;
}
494

495 496 497 498 499 500 501 502 503 504 505 506 507 508 509 510 511 512 513 514 515 516 517 518 519 520 521 522 523 524 525
/*
 * Copied from sock.c:sock_wait_for_wmem, but change to wait for
 * CAIF flow-on and sock_writable.
 */
static long caif_wait_for_flow_on(struct caifsock *cf_sk,
				int wait_writeable, long timeo, int *err)
{
	struct sock *sk = &cf_sk->sk;
	DEFINE_WAIT(wait);
	for (;;) {
		*err = 0;
		if (tx_flow_is_on(cf_sk) &&
			(!wait_writeable || sock_writeable(&cf_sk->sk)))
			break;
		*err = -ETIMEDOUT;
		if (!timeo)
			break;
		*err = -ERESTARTSYS;
		if (signal_pending(current))
			break;
		prepare_to_wait(sk_sleep(sk), &wait, TASK_INTERRUPTIBLE);
		*err = -ECONNRESET;
		if (sk->sk_shutdown & SHUTDOWN_MASK)
			break;
		*err = -sk->sk_err;
		if (sk->sk_err)
			break;
		*err = -EPIPE;
		if (cf_sk->sk.sk_state != CAIF_CONNECTED)
			break;
		timeo = schedule_timeout(timeo);
526
	}
527 528 529
	finish_wait(sk_sleep(sk), &wait);
	return timeo;
}
530

531 532 533 534 535 536 537 538 539
/*
 * Transmit a SKB. The device may temporarily request re-transmission
 * by returning EAGAIN.
 */
static int transmit_skb(struct sk_buff *skb, struct caifsock *cf_sk,
			int noblock, long timeo)
{
	struct cfpkt *pkt;
	int ret, loopcnt = 0;
540

541 542 543
	pkt = cfpkt_fromnative(CAIF_DIR_OUT, skb);
	memset(cfpkt_info(pkt), 0, sizeof(struct caif_payload_info));
	do {
544

545
		ret = -ETIMEDOUT;
546

547 548 549 550 551
		/* Slight paranoia, probably not needed. */
		if (unlikely(loopcnt++ > 1000)) {
			pr_warning("CAIF: %s(): transmit retries failed,"
				" error = %d\n", __func__, ret);
			break;
552 553
		}

554 555 556 557 558 559 560 561 562 563 564
		if (cf_sk->layer.dn != NULL)
			ret = cf_sk->layer.dn->transmit(cf_sk->layer.dn, pkt);
		if (likely(ret >= 0))
			break;
		/* if transmit return -EAGAIN, then retry */
		if (noblock && ret == -EAGAIN)
			break;
		timeo = caif_wait_for_flow_on(cf_sk, 0, timeo, &ret);
		if (signal_pending(current)) {
			ret = sock_intr_errno(timeo);
			break;
565
		}
566 567 568 569 570 571 572 573
		if (ret)
			break;
		if (cf_sk->sk.sk_state != CAIF_CONNECTED ||
			sock_flag(&cf_sk->sk, SOCK_DEAD) ||
			(cf_sk->sk.sk_shutdown & RCV_SHUTDOWN)) {
			ret = -EPIPE;
			cf_sk->sk.sk_err = EPIPE;
			break;
574
		}
575 576 577
	} while (ret == -EAGAIN);
	return ret;
}
578

579 580 581 582 583 584 585 586 587 588 589 590 591 592 593 594 595 596 597 598 599 600 601 602 603 604 605 606 607 608 609 610 611 612 613 614 615 616 617 618 619 620 621 622 623 624 625 626 627 628 629 630 631 632 633 634 635 636 637 638 639 640 641
/* Copied from af_unix:unix_dgram_sendmsg, and adapted to CAIF */
static int caif_seqpkt_sendmsg(struct kiocb *kiocb, struct socket *sock,
			struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
	struct caifsock *cf_sk = container_of(sk, struct caifsock, sk);
	int buffer_size;
	int ret = 0;
	struct sk_buff *skb = NULL;
	int noblock;
	long timeo;
	caif_assert(cf_sk);
	ret = sock_error(sk);
	if (ret)
		goto err;

	ret = -EOPNOTSUPP;
	if (msg->msg_flags&MSG_OOB)
		goto err;

	ret = -EOPNOTSUPP;
	if (msg->msg_namelen)
		goto err;

	ret = -EINVAL;
	if (unlikely(msg->msg_iov->iov_base == NULL))
		goto err;
	noblock = msg->msg_flags & MSG_DONTWAIT;

	buffer_size = len + CAIF_NEEDED_HEADROOM + CAIF_NEEDED_TAILROOM;

	ret = -EMSGSIZE;
	if (buffer_size > CAIF_MAX_PAYLOAD_SIZE)
		goto err;

	timeo = sock_sndtimeo(sk, noblock);
	timeo = caif_wait_for_flow_on(container_of(sk, struct caifsock, sk),
				1, timeo, &ret);

	ret = -EPIPE;
	if (cf_sk->sk.sk_state != CAIF_CONNECTED ||
		sock_flag(sk, SOCK_DEAD) ||
		(sk->sk_shutdown & RCV_SHUTDOWN))
		goto err;

	ret = -ENOMEM;
	skb = sock_alloc_send_skb(sk, buffer_size, noblock, &ret);
	if (!skb)
		goto err;
	skb_reserve(skb, CAIF_NEEDED_HEADROOM);

	ret = memcpy_fromiovec(skb_put(skb, len), msg->msg_iov, len);

	if (ret)
		goto err;
	ret = transmit_skb(skb, cf_sk, noblock, timeo);
	if (ret < 0)
		goto err;
	return len;
err:
	kfree_skb(skb);
	return ret;
}
642

643 644 645 646 647 648 649 650 651 652 653 654 655 656
/*
 * Copied from unix_stream_sendmsg and adapted to CAIF:
 * Changed removed permission handling and added waiting for flow on
 * and other minor adaptations.
 */
static int caif_stream_sendmsg(struct kiocb *kiocb, struct socket *sock,
				struct msghdr *msg, size_t len)
{
	struct sock *sk = sock->sk;
	struct caifsock *cf_sk = container_of(sk, struct caifsock, sk);
	int err, size;
	struct sk_buff *skb;
	int sent = 0;
	long timeo;
657

658
	err = -EOPNOTSUPP;
659

660 661
	if (unlikely(msg->msg_flags&MSG_OOB))
		goto out_err;
662

663 664
	if (unlikely(msg->msg_namelen))
		goto out_err;
665

666 667
	timeo = sock_sndtimeo(sk, msg->msg_flags & MSG_DONTWAIT);
	timeo = caif_wait_for_flow_on(cf_sk, 1, timeo, &err);
668

669 670
	if (unlikely(sk->sk_shutdown & SEND_SHUTDOWN))
		goto pipe_err;
671

672
	while (sent < len) {
673

674
		size = len-sent;
675

676 677
		if (size > CAIF_MAX_PAYLOAD_SIZE)
			size = CAIF_MAX_PAYLOAD_SIZE;
678

679 680 681
		/* If size is more than half of sndbuf, chop up message */
		if (size > ((sk->sk_sndbuf >> 1) - 64))
			size = (sk->sk_sndbuf >> 1) - 64;
682

683 684
		if (size > SKB_MAX_ALLOC)
			size = SKB_MAX_ALLOC;
685

686 687 688 689 690 691 692
		skb = sock_alloc_send_skb(sk,
					size + CAIF_NEEDED_HEADROOM
					+ CAIF_NEEDED_TAILROOM,
					msg->msg_flags&MSG_DONTWAIT,
					&err);
		if (skb == NULL)
			goto out_err;
693

694 695 696 697 698 699 700 701 702
		skb_reserve(skb, CAIF_NEEDED_HEADROOM);
		/*
		 *	If you pass two values to the sock_alloc_send_skb
		 *	it tries to grab the large buffer with GFP_NOFS
		 *	(which can fail easily), and if it fails grab the
		 *	fallback size buffer which is under a page and will
		 *	succeed. [Alan]
		 */
		size = min_t(int, size, skb_tailroom(skb));
703

704 705 706 707 708 709 710 711 712 713 714 715
		err = memcpy_fromiovec(skb_put(skb, size), msg->msg_iov, size);
		if (err) {
			kfree_skb(skb);
			goto out_err;
		}
		err = transmit_skb(skb, cf_sk,
				msg->msg_flags&MSG_DONTWAIT, timeo);
		if (err < 0) {
			kfree_skb(skb);
			goto pipe_err;
		}
		sent += size;
716 717
	}

718
	return sent;
719

720 721 722 723 724 725
pipe_err:
	if (sent == 0 && !(msg->msg_flags&MSG_NOSIGNAL))
		send_sig(SIGPIPE, current, 0);
	err = -EPIPE;
out_err:
	return sent ? : err;
726 727 728 729 730 731 732 733 734 735
}

static int setsockopt(struct socket *sock,
			int lvl, int opt, char __user *ov, unsigned int ol)
{
	struct sock *sk = sock->sk;
	struct caifsock *cf_sk = container_of(sk, struct caifsock, sk);
	int prio, linksel;
	struct ifreq ifreq;

736
	if (cf_sk->sk.sk_socket->state != SS_UNCONNECTED)
737
		return -ENOPROTOOPT;
738

739 740
	switch (opt) {
	case CAIFSO_LINK_SELECT:
741
		if (ol < sizeof(int))
742 743 744 745 746 747 748 749 750 751 752 753 754
			return -EINVAL;
		if (lvl != SOL_CAIF)
			goto bad_sol;
		if (copy_from_user(&linksel, ov, sizeof(int)))
			return -EINVAL;
		lock_sock(&(cf_sk->sk));
		cf_sk->conn_req.link_selector = linksel;
		release_sock(&cf_sk->sk);
		return 0;

	case SO_PRIORITY:
		if (lvl != SOL_SOCKET)
			goto bad_sol;
755
		if (ol < sizeof(int))
756 757 758 759 760 761 762 763 764 765 766
			return -EINVAL;
		if (copy_from_user(&prio, ov, sizeof(int)))
			return -EINVAL;
		lock_sock(&(cf_sk->sk));
		cf_sk->conn_req.priority = prio;
		release_sock(&cf_sk->sk);
		return 0;

	case SO_BINDTODEVICE:
		if (lvl != SOL_SOCKET)
			goto bad_sol;
767
		if (ol < sizeof(struct ifreq))
768 769 770 771 772 773 774 775 776 777 778 779 780 781 782 783 784 785
			return -EINVAL;
		if (copy_from_user(&ifreq, ov, sizeof(ifreq)))
			return -EFAULT;
		lock_sock(&(cf_sk->sk));
		strncpy(cf_sk->conn_req.link_name, ifreq.ifr_name,
			sizeof(cf_sk->conn_req.link_name));
		cf_sk->conn_req.link_name
			[sizeof(cf_sk->conn_req.link_name)-1] = 0;
		release_sock(&cf_sk->sk);
		return 0;

	case CAIFSO_REQ_PARAM:
		if (lvl != SOL_CAIF)
			goto bad_sol;
		if (cf_sk->sk.sk_protocol != CAIFPROTO_UTIL)
			return -ENOPROTOOPT;
		lock_sock(&(cf_sk->sk));
		cf_sk->conn_req.param.size = ol;
786 787
		if (ol > sizeof(cf_sk->conn_req.param.data) ||
			copy_from_user(&cf_sk->conn_req.param.data, ov, ol)) {
788 789 790 791 792 793 794
			release_sock(&cf_sk->sk);
			return -EINVAL;
		}
		release_sock(&cf_sk->sk);
		return 0;

	default:
795
		return -ENOPROTOOPT;
796 797 798 799 800 801 802 803
	}

	return 0;
bad_sol:
	return -ENOPROTOOPT;

}

804 805 806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829
/*
 * caif_connect() - Connect a CAIF Socket
 * Copied and modified af_irda.c:irda_connect().
 *
 * Note : by consulting "errno", the user space caller may learn the cause
 * of the failure. Most of them are visible in the function, others may come
 * from subroutines called and are listed here :
 *  o -EAFNOSUPPORT: bad socket family or type.
 *  o -ESOCKTNOSUPPORT: bad socket type or protocol
 *  o -EINVAL: bad socket address, or CAIF link type
 *  o -ECONNREFUSED: remote end refused the connection.
 *  o -EINPROGRESS: connect request sent but timed out (or non-blocking)
 *  o -EISCONN: already connected.
 *  o -ETIMEDOUT: Connection timed out (send timeout)
 *  o -ENODEV: No link layer to send request
 *  o -ECONNRESET: Received Shutdown indication or lost link layer
 *  o -ENOMEM: Out of memory
 *
 *  State Strategy:
 *  o sk_state: holds the CAIF_* protocol state, it's updated by
 *	caif_ctrl_cb.
 *  o sock->state: holds the SS_* socket state and is updated by connect and
 *	disconnect.
 */
static int caif_connect(struct socket *sock, struct sockaddr *uaddr,
			int addr_len, int flags)
830 831
{
	struct sock *sk = sock->sk;
832 833 834 835
	struct caifsock *cf_sk = container_of(sk, struct caifsock, sk);
	long timeo;
	int err;
	lock_sock(sk);
836

837 838
	err = -EAFNOSUPPORT;
	if (uaddr->sa_family != AF_CAIF)
839 840
		goto out;

841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878 879 880 881
	err = -ESOCKTNOSUPPORT;
	if (unlikely(!(sk->sk_type == SOCK_STREAM &&
		       cf_sk->sk.sk_protocol == CAIFPROTO_AT) &&
		       sk->sk_type != SOCK_SEQPACKET))
		goto out;
	switch (sock->state) {
	case SS_UNCONNECTED:
		/* Normal case, a fresh connect */
		caif_assert(sk->sk_state == CAIF_DISCONNECTED);
		break;
	case SS_CONNECTING:
		switch (sk->sk_state) {
		case CAIF_CONNECTED:
			sock->state = SS_CONNECTED;
			err = -EISCONN;
			goto out;
		case CAIF_DISCONNECTED:
			/* Reconnect allowed */
			break;
		case CAIF_CONNECTING:
			err = -EALREADY;
			if (flags & O_NONBLOCK)
				goto out;
			goto wait_connect;
		}
		break;
	case SS_CONNECTED:
		caif_assert(sk->sk_state == CAIF_CONNECTED ||
				sk->sk_state == CAIF_DISCONNECTED);
		if (sk->sk_shutdown & SHUTDOWN_MASK) {
			/* Allow re-connect after SHUTDOWN_IND */
			caif_disconnect_client(&cf_sk->layer);
			break;
		}
		/* No reconnect on a seqpacket socket */
		err = -EISCONN;
		goto out;
	case SS_DISCONNECTING:
	case SS_FREE:
		caif_assert(1); /*Should never happen */
		break;
882
	}
883 884 885
	sk->sk_state = CAIF_DISCONNECTED;
	sock->state = SS_UNCONNECTED;
	sk_stream_kill_queues(&cf_sk->sk);
886

887 888 889 890
	err = -EINVAL;
	if (addr_len != sizeof(struct sockaddr_caif) ||
		!uaddr)
		goto out;
891

892
	memcpy(&cf_sk->conn_req.sockaddr, uaddr,
893 894
		sizeof(struct sockaddr_caif));

895 896 897 898 899 900 901 902 903 904 905 906
	/* Move to connecting socket, start sending Connect Requests */
	sock->state = SS_CONNECTING;
	sk->sk_state = CAIF_CONNECTING;

	dbfs_atomic_inc(&cnt.num_connect_req);
	cf_sk->layer.receive = caif_sktrecv_cb;
	err = caif_connect_client(&cf_sk->conn_req,
				&cf_sk->layer);
	if (err < 0) {
		cf_sk->sk.sk_socket->state = SS_UNCONNECTED;
		cf_sk->sk.sk_state = CAIF_DISCONNECTED;
		goto out;
907 908
	}

909 910
	err = -EINPROGRESS;
wait_connect:
911

912 913
	if (sk->sk_state != CAIF_CONNECTED && (flags & O_NONBLOCK))
		goto out;
914

915 916 917
	timeo = sock_sndtimeo(sk, flags & O_NONBLOCK);

	release_sock(sk);
918 919
	err = -ERESTARTSYS;
	timeo = wait_event_interruptible_timeout(*sk_sleep(sk),
920 921 922
			sk->sk_state != CAIF_CONNECTING,
			timeo);
	lock_sock(sk);
923
	if (timeo < 0)
924
		goto out; /* -ERESTARTSYS */
925

926 927 928
	err = -ETIMEDOUT;
	if (timeo == 0 && sk->sk_state != CAIF_CONNECTED)
		goto out;
929 930 931 932 933 934 935 936 937
	if (sk->sk_state != CAIF_CONNECTED) {
		sock->state = SS_UNCONNECTED;
		err = sock_error(sk);
		if (!err)
			err = -ECONNREFUSED;
		goto out;
	}
	sock->state = SS_CONNECTED;
	err = 0;
938
out:
939 940
	release_sock(sk);
	return err;
941 942
}

943 944 945 946 947
/*
 * caif_release() - Disconnect a CAIF Socket
 * Copied and modified af_irda.c:irda_release().
 */
static int caif_release(struct socket *sock)
948 949
{
	struct sock *sk = sock->sk;
950 951
	struct caifsock *cf_sk = container_of(sk, struct caifsock, sk);
	int res = 0;
952

953 954
	if (!sk)
		return 0;
955

956
	set_tx_flow_off(cf_sk);
957 958

	/*
959 960 961
	 * Ensure that packets are not queued after this point in time.
	 * caif_queue_rcv_skb checks SOCK_DEAD holding the queue lock,
	 * this ensures no packets when sock is dead.
962
	 */
963 964 965 966
	spin_lock(&sk->sk_receive_queue.lock);
	sock_set_flag(sk, SOCK_DEAD);
	spin_unlock(&sk->sk_receive_queue.lock);
	sock->sk = NULL;
967

968
	dbfs_atomic_inc(&cnt.num_disconnect);
969 970 971 972 973

	if (cf_sk->debugfs_socket_dir != NULL)
		debugfs_remove_recursive(cf_sk->debugfs_socket_dir);

	lock_sock(&(cf_sk->sk));
974 975
	sk->sk_state = CAIF_DISCONNECTED;
	sk->sk_shutdown = SHUTDOWN_MASK;
976

977 978 979
	if (cf_sk->sk.sk_socket->state == SS_CONNECTED ||
		cf_sk->sk.sk_socket->state == SS_CONNECTING)
		res = caif_disconnect_client(&cf_sk->layer);
980

981 982
	cf_sk->sk.sk_socket->state = SS_DISCONNECTING;
	wake_up_interruptible_poll(sk_sleep(sk), POLLERR|POLLHUP);
983

984 985 986 987 988 989 990
	sock_orphan(sk);
	cf_sk->layer.dn = NULL;
	sk_stream_kill_queues(&cf_sk->sk);
	release_sock(sk);
	sock_put(sk);
	return res;
}
991

992 993 994 995 996 997 998
/* Copied from af_unix.c:unix_poll(), added CAIF tx_flow handling */
static unsigned int caif_poll(struct file *file,
				struct socket *sock, poll_table *wait)
{
	struct sock *sk = sock->sk;
	unsigned int mask;
	struct caifsock *cf_sk = container_of(sk, struct caifsock, sk);
999

1000 1001
	sock_poll_wait(file, sk_sleep(sk), wait);
	mask = 0;
1002

1003 1004 1005 1006 1007 1008 1009
	/* exceptional events? */
	if (sk->sk_err)
		mask |= POLLERR;
	if (sk->sk_shutdown == SHUTDOWN_MASK)
		mask |= POLLHUP;
	if (sk->sk_shutdown & RCV_SHUTDOWN)
		mask |= POLLRDHUP;
1010

1011 1012 1013 1014
	/* readable? */
	if (!skb_queue_empty(&sk->sk_receive_queue) ||
		(sk->sk_shutdown & RCV_SHUTDOWN))
		mask |= POLLIN | POLLRDNORM;
1015 1016

	/*
1017 1018
	 * we set writable also when the other side has shut down the
	 * connection. This prevents stuck sockets.
1019
	 */
1020 1021 1022 1023
	if (sock_writeable(sk) && tx_flow_is_on(cf_sk))
		mask |= POLLOUT | POLLWRNORM | POLLWRBAND;

	return mask;
1024 1025
}

1026 1027 1028 1029 1030 1031 1032 1033 1034 1035 1036 1037 1038 1039 1040 1041 1042 1043 1044 1045 1046 1047
static const struct proto_ops caif_seqpacket_ops = {
	.family = PF_CAIF,
	.owner = THIS_MODULE,
	.release = caif_release,
	.bind = sock_no_bind,
	.connect = caif_connect,
	.socketpair = sock_no_socketpair,
	.accept = sock_no_accept,
	.getname = sock_no_getname,
	.poll = caif_poll,
	.ioctl = sock_no_ioctl,
	.listen = sock_no_listen,
	.shutdown = sock_no_shutdown,
	.setsockopt = setsockopt,
	.getsockopt = sock_no_getsockopt,
	.sendmsg = caif_seqpkt_sendmsg,
	.recvmsg = caif_seqpkt_recvmsg,
	.mmap = sock_no_mmap,
	.sendpage = sock_no_sendpage,
};

static const struct proto_ops caif_stream_ops = {
1048 1049 1050 1051 1052 1053 1054 1055 1056 1057 1058
	.family = PF_CAIF,
	.owner = THIS_MODULE,
	.release = caif_release,
	.bind = sock_no_bind,
	.connect = caif_connect,
	.socketpair = sock_no_socketpair,
	.accept = sock_no_accept,
	.getname = sock_no_getname,
	.poll = caif_poll,
	.ioctl = sock_no_ioctl,
	.listen = sock_no_listen,
1059
	.shutdown = sock_no_shutdown,
1060 1061
	.setsockopt = setsockopt,
	.getsockopt = sock_no_getsockopt,
1062 1063
	.sendmsg = caif_stream_sendmsg,
	.recvmsg = caif_stream_recvmsg,
1064
	.mmap = sock_no_mmap,
1065
	.sendpage = sock_no_sendpage,
1066 1067 1068 1069 1070
};

/* This function is called when a socket is finally destroyed. */
static void caif_sock_destructor(struct sock *sk)
{
1071
	struct caifsock *cf_sk = container_of(sk, struct caifsock, sk);
1072 1073 1074 1075
	caif_assert(!atomic_read(&sk->sk_wmem_alloc));
	caif_assert(sk_unhashed(sk));
	caif_assert(!sk->sk_socket);
	if (!sock_flag(sk, SOCK_DEAD)) {
1076
		pr_info("Attempt to release alive CAIF socket: %p\n", sk);
1077 1078
		return;
	}
1079 1080
	sk_stream_kill_queues(&cf_sk->sk);
	dbfs_atomic_dec(&cnt.caif_nr_socks);
1081 1082 1083
}

static int caif_create(struct net *net, struct socket *sock, int protocol,
1084
			int kern)
1085 1086 1087 1088 1089 1090 1091 1092
{
	struct sock *sk = NULL;
	struct caifsock *cf_sk = NULL;
	static struct proto prot = {.name = "PF_CAIF",
		.owner = THIS_MODULE,
		.obj_size = sizeof(struct caifsock),
	};

1093 1094
	if (!capable(CAP_SYS_ADMIN) && !capable(CAP_NET_ADMIN))
		return -EPERM;
1095 1096
	/*
	 * The sock->type specifies the socket type to use.
1097 1098 1099
	 * The CAIF socket is a packet stream in the sense
	 * that it is packet based. CAIF trusts the reliability
	 * of the link, no resending is implemented.
1100
	 */
1101 1102 1103 1104 1105
	if (sock->type == SOCK_SEQPACKET)
		sock->ops = &caif_seqpacket_ops;
	else if (sock->type == SOCK_STREAM)
		sock->ops = &caif_stream_ops;
	else
1106 1107 1108 1109 1110
		return -ESOCKTNOSUPPORT;

	if (protocol < 0 || protocol >= CAIFPROTO_MAX)
		return -EPROTONOSUPPORT;
	/*
1111 1112
	 * Set the socket state to unconnected.	 The socket state
	 * is really not used at all in the net/core or socket.c but the
1113 1114 1115 1116 1117 1118 1119 1120 1121 1122 1123
	 * initialization makes sure that sock->state is not uninitialized.
	 */
	sk = sk_alloc(net, PF_CAIF, GFP_KERNEL, &prot);
	if (!sk)
		return -ENOMEM;

	cf_sk = container_of(sk, struct caifsock, sk);

	/* Store the protocol */
	sk->sk_protocol = (unsigned char) protocol;

1124 1125 1126
	/* Sendbuf dictates the amount of outbound packets not yet sent */
	sk->sk_sndbuf = CAIF_DEF_SNDBUF;
	sk->sk_rcvbuf = CAIF_DEF_RCVBUF;
1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137

	/*
	 * Lock in order to try to stop someone from opening the socket
	 * too early.
	 */
	lock_sock(&(cf_sk->sk));

	/* Initialize the nozero default sock structure data. */
	sock_init_data(sock, sk);
	sk->sk_destruct = caif_sock_destructor;

1138 1139 1140 1141
	mutex_init(&cf_sk->readlock); /* single task reading lock */
	cf_sk->layer.ctrlcmd = caif_ctrl_cb;
	cf_sk->sk.sk_socket->state = SS_UNCONNECTED;
	cf_sk->sk.sk_state = CAIF_DISCONNECTED;
1142

1143 1144
	set_tx_flow_off(cf_sk);
	set_rx_flow_on(cf_sk);
1145 1146 1147

	/* Set default options on configuration */
	cf_sk->conn_req.priority = CAIF_PRIO_NORMAL;
1148
	cf_sk->conn_req.link_selector = CAIF_LINK_LOW_LATENCY;
1149 1150
	cf_sk->conn_req.protocol = protocol;
	/* Increase the number of sockets created. */
1151 1152
	dbfs_atomic_inc(&cnt.caif_nr_socks);
#ifdef CONFIG_DEBUG_FS
1153
	if (!IS_ERR(debugfsdir)) {
1154 1155 1156 1157
		/* Fill in some information concerning the misc socket. */
		snprintf(cf_sk->name, sizeof(cf_sk->name), "cfsk%d",
				atomic_read(&cnt.caif_nr_socks));

1158 1159
		cf_sk->debugfs_socket_dir =
			debugfs_create_dir(cf_sk->name, debugfsdir);
1160 1161 1162
		debugfs_create_u32("sk_state", S_IRUSR | S_IWUSR,
				cf_sk->debugfs_socket_dir,
				(u32 *) &cf_sk->sk.sk_state);
1163 1164
		debugfs_create_u32("flow_state", S_IRUSR | S_IWUSR,
				cf_sk->debugfs_socket_dir, &cf_sk->flow_state);
1165 1166 1167 1168
		debugfs_create_u32("sk_rmem_alloc", S_IRUSR | S_IWUSR,
				cf_sk->debugfs_socket_dir,
				(u32 *) &cf_sk->sk.sk_rmem_alloc);
		debugfs_create_u32("sk_wmem_alloc", S_IRUSR | S_IWUSR,
1169
				cf_sk->debugfs_socket_dir,
1170
				(u32 *) &cf_sk->sk.sk_wmem_alloc);
1171 1172 1173 1174
		debugfs_create_u32("identity", S_IRUSR | S_IWUSR,
				cf_sk->debugfs_socket_dir,
				(u32 *) &cf_sk->layer.id);
	}
1175
#endif
1176 1177 1178 1179
	release_sock(&cf_sk->sk);
	return 0;
}

1180

1181 1182 1183 1184 1185 1186
static struct net_proto_family caif_family_ops = {
	.family = PF_CAIF,
	.create = caif_create,
	.owner = THIS_MODULE,
};

1187
int af_caif_init(void)
1188
{
1189
	int err = sock_register(&caif_family_ops);
1190 1191 1192 1193 1194 1195 1196 1197
	if (!err)
		return err;
	return 0;
}

static int __init caif_sktinit_module(void)
{
#ifdef CONFIG_DEBUG_FS
1198
	debugfsdir = debugfs_create_dir("caif_sk", NULL);
1199 1200 1201
	if (!IS_ERR(debugfsdir)) {
		debugfs_create_u32("num_sockets", S_IRUSR | S_IWUSR,
				debugfsdir,
1202 1203
				(u32 *) &cnt.caif_nr_socks);
		debugfs_create_u32("num_connect_req", S_IRUSR | S_IWUSR,
1204
				debugfsdir,
1205