lowcomms.c 36.8 KB
Newer Older
1 2 3 4
/******************************************************************************
*******************************************************************************
**
**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
5
**  Copyright (C) 2004-2009 Red Hat, Inc.  All rights reserved.
6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23
**
**  This copyrighted material is made available to anyone wishing to use,
**  modify, copy, or redistribute it subject to the terms and conditions
**  of the GNU General Public License v.2.
**
*******************************************************************************
******************************************************************************/

/*
 * lowcomms.c
 *
 * This is the "low-level" comms layer.
 *
 * It is responsible for sending/receiving messages
 * from other nodes in the cluster.
 *
 * Cluster nodes are referred to by their nodeids. nodeids are
 * simply 32 bit numbers to the locking module - if they need to
J
Joe Perches 已提交
24
 * be expanded for the cluster infrastructure then that is its
25 26 27 28 29 30 31 32 33 34 35 36 37 38
 * responsibility. It is this layer's
 * responsibility to resolve these into IP address or
 * whatever it needs for inter-node communication.
 *
 * The comms level is two kernel threads that deal mainly with
 * the receiving of messages from other nodes and passing them
 * up to the mid-level comms layer (which understands the
 * message format) for execution by the locking core, and
 * a send thread which does all the setting up of connections
 * to remote nodes and the sending of data. Threads are not allowed
 * to send their own data because it may cause them to wait in times
 * of high load. Also, this way, the sending thread can collect together
 * messages bound for one node and send them in one block.
 *
J
Joe Perches 已提交
39
 * lowcomms will choose to use either TCP or SCTP as its transport layer
40
 * depending on the configuration variable 'protocol'. This should be set
J
Joe Perches 已提交
41
 * to 0 (default) for TCP or 1 for SCTP. It should be configured using a
42 43
 * cluster-wide mechanism as it must be the same on all nodes of the cluster
 * for the DLM to function.
44 45 46 47 48 49 50
 *
 */

#include <asm/ioctls.h>
#include <net/sock.h>
#include <net/tcp.h>
#include <linux/pagemap.h>
51
#include <linux/file.h>
52
#include <linux/mutex.h>
53 54
#include <linux/sctp.h>
#include <net/sctp/user.h>
J
Joe Perches 已提交
55
#include <net/ipv6.h>
56 57 58 59 60 61

#include "dlm_internal.h"
#include "lowcomms.h"
#include "midcomms.h"
#include "config.h"

62
#define NEEDED_RMEM (4*1024*1024)
63
#define CONN_HASH_SIZE 32
64

65
struct cbuf {
P
Patrick Caulfield 已提交
66 67 68
	unsigned int base;
	unsigned int len;
	unsigned int mask;
69 70
};

P
Patrick Caulfield 已提交
71 72 73 74
static void cbuf_add(struct cbuf *cb, int n)
{
	cb->len += n;
}
75

P
Patrick Caulfield 已提交
76 77 78 79 80 81 82 83 84 85 86 87 88 89 90 91 92 93 94 95 96 97
static int cbuf_data(struct cbuf *cb)
{
	return ((cb->base + cb->len) & cb->mask);
}

static void cbuf_init(struct cbuf *cb, int size)
{
	cb->base = cb->len = 0;
	cb->mask = size-1;
}

static void cbuf_eat(struct cbuf *cb, int n)
{
	cb->len  -= n;
	cb->base += n;
	cb->base &= cb->mask;
}

static bool cbuf_empty(struct cbuf *cb)
{
	return cb->len == 0;
}
98 99 100 101

struct connection {
	struct socket *sock;	/* NULL if not connected */
	uint32_t nodeid;	/* So we know who we are in the list */
102
	struct mutex sock_mutex;
103
	unsigned long flags;
104 105 106
#define CF_READ_PENDING 1
#define CF_WRITE_PENDING 2
#define CF_CONNECT_PENDING 3
107 108
#define CF_INIT_PENDING 4
#define CF_IS_OTHERCON 5
P
Patrick Caulfield 已提交
109
	struct list_head writequeue;  /* List of outgoing writequeue_entries */
110 111
	spinlock_t writequeue_lock;
	int (*rx_action) (struct connection *);	/* What to do when active */
112
	void (*connect_action) (struct connection *);	/* What to do to connect */
113 114 115 116
	struct page *rx_page;
	struct cbuf cb;
	int retries;
#define MAX_CONNECT_RETRIES 3
117
	int sctp_assoc;
118
	struct hlist_node list;
119
	struct connection *othercon;
120 121
	struct work_struct rwork; /* Receive workqueue */
	struct work_struct swork; /* Send workqueue */
122 123 124 125 126 127 128 129 130 131 132 133 134 135
};
#define sock2con(x) ((struct connection *)(x)->sk_user_data)

/* An entry waiting to be sent */
struct writequeue_entry {
	struct list_head list;
	struct page *page;
	int offset;
	int len;
	int end;
	int users;
	struct connection *con;
};

136 137
static struct sockaddr_storage *dlm_local_addr[DLM_MAX_ADDR_COUNT];
static int dlm_local_count;
138

139 140 141
/* Work queues */
static struct workqueue_struct *recv_workqueue;
static struct workqueue_struct *send_workqueue;
142

143
static struct hlist_head connection_hash[CONN_HASH_SIZE];
144
static DEFINE_MUTEX(connections_lock);
P
Patrick Caulfield 已提交
145
static struct kmem_cache *con_cache;
146

147 148
static void process_recv_sockets(struct work_struct *work);
static void process_send_sockets(struct work_struct *work);
149

150 151 152 153 154 155 156 157 158 159 160 161 162 163 164 165 166 167 168 169 170 171 172 173

/* This is deliberately very simple because most clusters have simple
   sequential nodeids, so we should be able to go straight to a connection
   struct in the array */
static inline int nodeid_hash(int nodeid)
{
	return nodeid & (CONN_HASH_SIZE-1);
}

static struct connection *__find_con(int nodeid)
{
	int r;
	struct hlist_node *h;
	struct connection *con;

	r = nodeid_hash(nodeid);

	hlist_for_each_entry(con, h, &connection_hash[r], list) {
		if (con->nodeid == nodeid)
			return con;
	}
	return NULL;
}

174 175 176 177 178
/*
 * If 'allocation' is zero then we don't attempt to create a new
 * connection structure for this node.
 */
static struct connection *__nodeid2con(int nodeid, gfp_t alloc)
179 180
{
	struct connection *con = NULL;
181
	int r;
182

183
	con = __find_con(nodeid);
184 185
	if (con || !alloc)
		return con;
186

187 188 189
	con = kmem_cache_zalloc(con_cache, alloc);
	if (!con)
		return NULL;
190

191 192
	r = nodeid_hash(nodeid);
	hlist_add_head(&con->list, &connection_hash[r]);
193

194 195 196 197 198 199
	con->nodeid = nodeid;
	mutex_init(&con->sock_mutex);
	INIT_LIST_HEAD(&con->writequeue);
	spin_lock_init(&con->writequeue_lock);
	INIT_WORK(&con->swork, process_send_sockets);
	INIT_WORK(&con->rwork, process_recv_sockets);
200

201 202
	/* Setup action pointers for child sockets */
	if (con->nodeid) {
203
		struct connection *zerocon = __find_con(0);
204

205 206 207
		con->connect_action = zerocon->connect_action;
		if (!con->rx_action)
			con->rx_action = zerocon->rx_action;
208 209
	}

210 211 212
	return con;
}

213 214 215 216 217 218 219 220 221 222 223 224 225 226
/* Loop round all connections */
static void foreach_conn(void (*conn_func)(struct connection *c))
{
	int i;
	struct hlist_node *h, *n;
	struct connection *con;

	for (i = 0; i < CONN_HASH_SIZE; i++) {
		hlist_for_each_entry_safe(con, h, n, &connection_hash[i], list){
			conn_func(con);
		}
	}
}

227 228 229 230
static struct connection *nodeid2con(int nodeid, gfp_t allocation)
{
	struct connection *con;

231
	mutex_lock(&connections_lock);
232
	con = __nodeid2con(nodeid, allocation);
233
	mutex_unlock(&connections_lock);
234

235 236 237
	return con;
}

238 239 240 241
/* This is a bit drastic, but only called when things go wrong */
static struct connection *assoc2con(int assoc_id)
{
	int i;
242
	struct hlist_node *h;
243 244
	struct connection *con;

245
	mutex_lock(&connections_lock);
246 247 248 249 250 251 252

	for (i = 0 ; i < CONN_HASH_SIZE; i++) {
		hlist_for_each_entry(con, h, &connection_hash[i], list) {
			if (con && con->sctp_assoc == assoc_id) {
				mutex_unlock(&connections_lock);
				return con;
			}
253 254
		}
	}
255
	mutex_unlock(&connections_lock);
256 257 258 259 260 261 262 263 264 265 266 267 268 269 270 271 272 273 274 275 276 277
	return NULL;
}

static int nodeid_to_addr(int nodeid, struct sockaddr *retaddr)
{
	struct sockaddr_storage addr;
	int error;

	if (!dlm_local_count)
		return -1;

	error = dlm_nodeid_to_addr(nodeid, &addr);
	if (error)
		return error;

	if (dlm_local_addr[0]->ss_family == AF_INET) {
		struct sockaddr_in *in4  = (struct sockaddr_in *) &addr;
		struct sockaddr_in *ret4 = (struct sockaddr_in *) retaddr;
		ret4->sin_addr.s_addr = in4->sin_addr.s_addr;
	} else {
		struct sockaddr_in6 *in6  = (struct sockaddr_in6 *) &addr;
		struct sockaddr_in6 *ret6 = (struct sockaddr_in6 *) retaddr;
J
Joe Perches 已提交
278
		ipv6_addr_copy(&ret6->sin6_addr, &in6->sin6_addr);
279 280 281 282 283
	}

	return 0;
}

284 285 286 287
/* Data available on socket or listen socket received a connect */
static void lowcomms_data_ready(struct sock *sk, int count_unused)
{
	struct connection *con = sock2con(sk);
P
Patrick Caulfield 已提交
288
	if (con && !test_and_set_bit(CF_READ_PENDING, &con->flags))
289
		queue_work(recv_workqueue, &con->rwork);
290 291 292 293 294 295
}

static void lowcomms_write_space(struct sock *sk)
{
	struct connection *con = sock2con(sk);

P
Patrick Caulfield 已提交
296
	if (con && !test_and_set_bit(CF_WRITE_PENDING, &con->flags))
297
		queue_work(send_workqueue, &con->swork);
298 299 300 301
}

static inline void lowcomms_connect_sock(struct connection *con)
{
302 303
	if (!test_and_set_bit(CF_CONNECT_PENDING, &con->flags))
		queue_work(send_workqueue, &con->swork);
304 305 306 307
}

static void lowcomms_state_change(struct sock *sk)
{
P
Patrick Caulfield 已提交
308
	if (sk->sk_state == TCP_ESTABLISHED)
309 310 311
		lowcomms_write_space(sk);
}

312 313 314 315 316 317 318 319 320 321 322 323 324 325
int dlm_lowcomms_connect_node(int nodeid)
{
	struct connection *con;

	if (nodeid == dlm_our_nodeid())
		return 0;

	con = nodeid2con(nodeid, GFP_NOFS);
	if (!con)
		return -ENOMEM;
	lowcomms_connect_sock(con);
	return 0;
}

326 327 328 329 330 331 332 333 334
/* Make a socket active */
static int add_sock(struct socket *sock, struct connection *con)
{
	con->sock = sock;

	/* Install a data_ready callback */
	con->sock->sk->sk_data_ready = lowcomms_data_ready;
	con->sock->sk->sk_write_space = lowcomms_write_space;
	con->sock->sk->sk_state_change = lowcomms_state_change;
335
	con->sock->sk->sk_user_data = con;
336
	con->sock->sk->sk_allocation = GFP_NOFS;
337 338 339
	return 0;
}

340
/* Add the port number to an IPv6 or 4 sockaddr and return the address
341 342 343 344
   length */
static void make_sockaddr(struct sockaddr_storage *saddr, uint16_t port,
			  int *addr_len)
{
345
	saddr->ss_family =  dlm_local_addr[0]->ss_family;
P
Patrick Caulfield 已提交
346
	if (saddr->ss_family == AF_INET) {
347 348 349
		struct sockaddr_in *in4_addr = (struct sockaddr_in *)saddr;
		in4_addr->sin_port = cpu_to_be16(port);
		*addr_len = sizeof(struct sockaddr_in);
350
		memset(&in4_addr->sin_zero, 0, sizeof(in4_addr->sin_zero));
P
Patrick Caulfield 已提交
351
	} else {
352 353 354 355
		struct sockaddr_in6 *in6_addr = (struct sockaddr_in6 *)saddr;
		in6_addr->sin6_port = cpu_to_be16(port);
		*addr_len = sizeof(struct sockaddr_in6);
	}
356
	memset((char *)saddr + *addr_len, 0, sizeof(struct sockaddr_storage) - *addr_len);
357 358 359
}

/* Close a remote connection and tidy up */
P
Patrick Caulfield 已提交
360
static void close_connection(struct connection *con, bool and_other)
361
{
362
	mutex_lock(&con->sock_mutex);
363 364 365 366 367 368

	if (con->sock) {
		sock_release(con->sock);
		con->sock = NULL;
	}
	if (con->othercon && and_other) {
P
Patrick Caulfield 已提交
369 370
		/* Will only re-enter once. */
		close_connection(con->othercon, false);
371 372 373 374 375
	}
	if (con->rx_page) {
		__free_page(con->rx_page);
		con->rx_page = NULL;
	}
P
Patrick Caulfield 已提交
376

377 378
	con->retries = 0;
	mutex_unlock(&con->sock_mutex);
379 380
}

381 382 383 384 385 386 387 388 389 390 391 392 393 394 395 396 397 398 399 400 401 402 403 404 405 406 407 408 409 410 411 412 413 414 415 416
/* We only send shutdown messages to nodes that are not part of the cluster */
static void sctp_send_shutdown(sctp_assoc_t associd)
{
	static char outcmsg[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];
	struct msghdr outmessage;
	struct cmsghdr *cmsg;
	struct sctp_sndrcvinfo *sinfo;
	int ret;
	struct connection *con;

	con = nodeid2con(0,0);
	BUG_ON(con == NULL);

	outmessage.msg_name = NULL;
	outmessage.msg_namelen = 0;
	outmessage.msg_control = outcmsg;
	outmessage.msg_controllen = sizeof(outcmsg);
	outmessage.msg_flags = MSG_EOR;

	cmsg = CMSG_FIRSTHDR(&outmessage);
	cmsg->cmsg_level = IPPROTO_SCTP;
	cmsg->cmsg_type = SCTP_SNDRCV;
	cmsg->cmsg_len = CMSG_LEN(sizeof(struct sctp_sndrcvinfo));
	outmessage.msg_controllen = cmsg->cmsg_len;
	sinfo = CMSG_DATA(cmsg);
	memset(sinfo, 0x00, sizeof(struct sctp_sndrcvinfo));

	sinfo->sinfo_flags |= MSG_EOF;
	sinfo->sinfo_assoc_id = associd;

	ret = kernel_sendmsg(con->sock, &outmessage, NULL, 0, 0);

	if (ret != 0)
		log_print("send EOF to node failed: %d", ret);
}

417 418 419 420 421 422 423 424 425
static void sctp_init_failed_foreach(struct connection *con)
{
	con->sctp_assoc = 0;
	if (test_and_clear_bit(CF_CONNECT_PENDING, &con->flags)) {
		if (!test_and_set_bit(CF_WRITE_PENDING, &con->flags))
			queue_work(send_workqueue, &con->swork);
	}
}

426 427 428 429
/* INIT failed but we don't know which node...
   restart INIT on all pending nodes */
static void sctp_init_failed(void)
{
430
	mutex_lock(&connections_lock);
431 432 433

	foreach_conn(sctp_init_failed_foreach);

434
	mutex_unlock(&connections_lock);
435 436 437
}

/* Something happened to an association */
D
David Teigland 已提交
438 439
static void process_sctp_notification(struct connection *con,
				      struct msghdr *msg, char *buf)
440 441 442 443 444 445 446 447 448 449 450 451 452 453 454 455 456 457 458 459 460 461 462 463 464 465
{
	union sctp_notification *sn = (union sctp_notification *)buf;

	if (sn->sn_header.sn_type == SCTP_ASSOC_CHANGE) {
		switch (sn->sn_assoc_change.sac_state) {

		case SCTP_COMM_UP:
		case SCTP_RESTART:
		{
			/* Check that the new node is in the lockspace */
			struct sctp_prim prim;
			int nodeid;
			int prim_len, ret;
			int addr_len;
			struct connection *new_con;
			struct file *file;
			sctp_peeloff_arg_t parg;
			int parglen = sizeof(parg);

			/*
			 * We get this before any data for an association.
			 * We verify that the node is in the cluster and
			 * then peel off a socket for it.
			 */
			if ((int)sn->sn_assoc_change.sac_assoc_id <= 0) {
				log_print("COMM_UP for invalid assoc ID %d",
D
David Teigland 已提交
466
					 (int)sn->sn_assoc_change.sac_assoc_id);
467 468 469 470 471 472 473 474 475 476 477 478 479 480 481 482 483 484 485 486 487 488 489 490 491 492 493 494 495 496 497 498 499 500 501 502
				sctp_init_failed();
				return;
			}
			memset(&prim, 0, sizeof(struct sctp_prim));
			prim_len = sizeof(struct sctp_prim);
			prim.ssp_assoc_id = sn->sn_assoc_change.sac_assoc_id;

			ret = kernel_getsockopt(con->sock,
						IPPROTO_SCTP,
						SCTP_PRIMARY_ADDR,
						(char*)&prim,
						&prim_len);
			if (ret < 0) {
				log_print("getsockopt/sctp_primary_addr on "
					  "new assoc %d failed : %d",
					  (int)sn->sn_assoc_change.sac_assoc_id,
					  ret);

				/* Retry INIT later */
				new_con = assoc2con(sn->sn_assoc_change.sac_assoc_id);
				if (new_con)
					clear_bit(CF_CONNECT_PENDING, &con->flags);
				return;
			}
			make_sockaddr(&prim.ssp_addr, 0, &addr_len);
			if (dlm_addr_to_nodeid(&prim.ssp_addr, &nodeid)) {
				int i;
				unsigned char *b=(unsigned char *)&prim.ssp_addr;
				log_print("reject connect from unknown addr");
				for (i=0; i<sizeof(struct sockaddr_storage);i++)
					printk("%02x ", b[i]);
				printk("\n");
				sctp_send_shutdown(prim.ssp_assoc_id);
				return;
			}

D
David Teigland 已提交
503
			new_con = nodeid2con(nodeid, GFP_NOFS);
504 505 506 507 508
			if (!new_con)
				return;

			/* Peel off a new sock */
			parg.associd = sn->sn_assoc_change.sac_assoc_id;
D
David Teigland 已提交
509 510
			ret = kernel_getsockopt(con->sock, IPPROTO_SCTP,
						SCTP_SOCKOPT_PEELOFF,
511
						(void *)&parg, &parglen);
512
			if (ret) {
D
David Teigland 已提交
513 514
				log_print("Can't peel off a socket for "
					  "connection %d to node %d: err=%d\n",
515 516 517 518 519 520 521 522 523
					  parg.associd, nodeid, ret);
			}
			file = fget(parg.sd);
			new_con->sock = SOCKET_I(file->f_dentry->d_inode);
			add_sock(new_con->sock, new_con);
			fput(file);
			put_unused_fd(parg.sd);

			log_print("got new/restarted association %d nodeid %d",
D
David Teigland 已提交
524
				 (int)sn->sn_assoc_change.sac_assoc_id, nodeid);
525 526 527 528 529 530 531 532 533 534 535 536 537 538 539 540 541 542 543 544 545 546 547 548 549 550 551 552 553 554 555 556 557 558 559 560 561 562 563 564 565

			/* Send any pending writes */
			clear_bit(CF_CONNECT_PENDING, &new_con->flags);
			clear_bit(CF_INIT_PENDING, &con->flags);
			if (!test_and_set_bit(CF_WRITE_PENDING, &new_con->flags)) {
				queue_work(send_workqueue, &new_con->swork);
			}
			if (!test_and_set_bit(CF_READ_PENDING, &new_con->flags))
				queue_work(recv_workqueue, &new_con->rwork);
		}
		break;

		case SCTP_COMM_LOST:
		case SCTP_SHUTDOWN_COMP:
		{
			con = assoc2con(sn->sn_assoc_change.sac_assoc_id);
			if (con) {
				con->sctp_assoc = 0;
			}
		}
		break;

		/* We don't know which INIT failed, so clear the PENDING flags
		 * on them all.  if assoc_id is zero then it will then try
		 * again */

		case SCTP_CANT_STR_ASSOC:
		{
			log_print("Can't start SCTP association - retrying");
			sctp_init_failed();
		}
		break;

		default:
			log_print("unexpected SCTP assoc change id=%d state=%d",
				  (int)sn->sn_assoc_change.sac_assoc_id,
				  sn->sn_assoc_change.sac_state);
		}
	}
}

566 567 568 569
/* Data received from remote end */
static int receive_from_sock(struct connection *con)
{
	int ret = 0;
A
Al Viro 已提交
570 571
	struct msghdr msg = {};
	struct kvec iov[2];
572 573 574
	unsigned len;
	int r;
	int call_again_soon = 0;
A
Al Viro 已提交
575
	int nvec;
576
	char incmsg[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];
577

578
	mutex_lock(&con->sock_mutex);
579

580 581 582 583 584
	if (con->sock == NULL) {
		ret = -EAGAIN;
		goto out_close;
	}

585 586 587 588 589 590 591 592
	if (con->rx_page == NULL) {
		/*
		 * This doesn't need to be atomic, but I think it should
		 * improve performance if it is.
		 */
		con->rx_page = alloc_page(GFP_ATOMIC);
		if (con->rx_page == NULL)
			goto out_resched;
P
Patrick Caulfield 已提交
593
		cbuf_init(&con->cb, PAGE_CACHE_SIZE);
594 595
	}

596 597 598 599 600
	/* Only SCTP needs these really */
	memset(&incmsg, 0, sizeof(incmsg));
	msg.msg_control = incmsg;
	msg.msg_controllen = sizeof(incmsg);

601 602 603 604
	/*
	 * iov[0] is the bit of the circular buffer between the current end
	 * point (cb.base + cb.len) and the end of the buffer.
	 */
P
Patrick Caulfield 已提交
605 606
	iov[0].iov_len = con->cb.base - cbuf_data(&con->cb);
	iov[0].iov_base = page_address(con->rx_page) + cbuf_data(&con->cb);
607
	iov[1].iov_len = 0;
A
Al Viro 已提交
608
	nvec = 1;
609 610 611 612 613

	/*
	 * iov[1] is the bit of the circular buffer between the start of the
	 * buffer and the start of the currently used section (cb.base)
	 */
P
Patrick Caulfield 已提交
614 615
	if (cbuf_data(&con->cb) >= con->cb.base) {
		iov[0].iov_len = PAGE_CACHE_SIZE - cbuf_data(&con->cb);
616 617
		iov[1].iov_len = con->cb.base;
		iov[1].iov_base = page_address(con->rx_page);
A
Al Viro 已提交
618
		nvec = 2;
619 620 621
	}
	len = iov[0].iov_len + iov[1].iov_len;

A
Al Viro 已提交
622
	r = ret = kernel_recvmsg(con->sock, &msg, iov, nvec, len,
623 624 625
			       MSG_DONTWAIT | MSG_NOSIGNAL);
	if (ret <= 0)
		goto out_close;
P
Patrick Caulfield 已提交
626

627 628 629 630 631 632
	/* Process SCTP notifications */
	if (msg.msg_flags & MSG_NOTIFICATION) {
		msg.msg_control = incmsg;
		msg.msg_controllen = sizeof(incmsg);

		process_sctp_notification(con, &msg,
D
David Teigland 已提交
633
				page_address(con->rx_page) + con->cb.base);
634 635 636 637 638
		mutex_unlock(&con->sock_mutex);
		return 0;
	}
	BUG_ON(con->nodeid == 0);

639 640
	if (ret == len)
		call_again_soon = 1;
P
Patrick Caulfield 已提交
641
	cbuf_add(&con->cb, ret);
642 643 644 645 646
	ret = dlm_process_incoming_buffer(con->nodeid,
					  page_address(con->rx_page),
					  con->cb.base, con->cb.len,
					  PAGE_CACHE_SIZE);
	if (ret == -EBADMSG) {
D
David Teigland 已提交
647 648 649 650
		log_print("lowcomms: addr=%p, base=%u, len=%u, "
			  "iov_len=%u, iov_base[0]=%p, read=%d",
			  page_address(con->rx_page), con->cb.base, con->cb.len,
			  len, iov[0].iov_base, r);
651 652 653
	}
	if (ret < 0)
		goto out_close;
P
Patrick Caulfield 已提交
654
	cbuf_eat(&con->cb, ret);
655

P
Patrick Caulfield 已提交
656
	if (cbuf_empty(&con->cb) && !call_again_soon) {
657 658 659 660 661 662
		__free_page(con->rx_page);
		con->rx_page = NULL;
	}

	if (call_again_soon)
		goto out_resched;
663
	mutex_unlock(&con->sock_mutex);
P
Patrick Caulfield 已提交
664
	return 0;
665

P
Patrick Caulfield 已提交
666
out_resched:
667 668
	if (!test_and_set_bit(CF_READ_PENDING, &con->flags))
		queue_work(recv_workqueue, &con->rwork);
669
	mutex_unlock(&con->sock_mutex);
P
Patrick Caulfield 已提交
670
	return -EAGAIN;
671

P
Patrick Caulfield 已提交
672
out_close:
673
	mutex_unlock(&con->sock_mutex);
P
Patrick Caulfield 已提交
674
	if (ret != -EAGAIN) {
P
Patrick Caulfield 已提交
675
		close_connection(con, false);
676 677
		/* Reconnect when there is something to send */
	}
678 679 680
	/* Don't return success if we really got EOF */
	if (ret == 0)
		ret = -EAGAIN;
681 682 683 684 685

	return ret;
}

/* Listening socket is busy, accept a connection */
686
static int tcp_accept_from_sock(struct connection *con)
687 688 689 690 691 692 693
{
	int result;
	struct sockaddr_storage peeraddr;
	struct socket *newsock;
	int len;
	int nodeid;
	struct connection *newcon;
P
Patrick Caulfield 已提交
694
	struct connection *addcon;
695 696

	memset(&peeraddr, 0, sizeof(peeraddr));
697
	result = sock_create_kern(dlm_local_addr[0]->ss_family, SOCK_STREAM,
P
Patrick Caulfield 已提交
698
				  IPPROTO_TCP, &newsock);
699 700 701
	if (result < 0)
		return -ENOMEM;

702
	mutex_lock_nested(&con->sock_mutex, 0);
703 704 705 706 707 708 709 710 711 712 713 714 715 716 717 718 719 720 721 722 723 724 725

	result = -ENOTCONN;
	if (con->sock == NULL)
		goto accept_err;

	newsock->type = con->sock->type;
	newsock->ops = con->sock->ops;

	result = con->sock->ops->accept(con->sock, newsock, O_NONBLOCK);
	if (result < 0)
		goto accept_err;

	/* Get the connected socket's peer */
	memset(&peeraddr, 0, sizeof(peeraddr));
	if (newsock->ops->getname(newsock, (struct sockaddr *)&peeraddr,
				  &len, 2)) {
		result = -ECONNABORTED;
		goto accept_err;
	}

	/* Get the new node's NODEID */
	make_sockaddr(&peeraddr, 0, &len);
	if (dlm_addr_to_nodeid(&peeraddr, &nodeid)) {
D
David Teigland 已提交
726
		log_print("connect from non cluster node");
727
		sock_release(newsock);
728
		mutex_unlock(&con->sock_mutex);
729 730 731 732 733 734 735 736 737 738
		return -1;
	}

	log_print("got connection from %d", nodeid);

	/*  Check to see if we already have a connection to this node. This
	 *  could happen if the two nodes initiate a connection at roughly
	 *  the same time and the connections cross on the wire.
	 *  In this case we store the incoming one in "othercon"
	 */
D
David Teigland 已提交
739
	newcon = nodeid2con(nodeid, GFP_NOFS);
740 741 742 743
	if (!newcon) {
		result = -ENOMEM;
		goto accept_err;
	}
744
	mutex_lock_nested(&newcon->sock_mutex, 1);
745
	if (newcon->sock) {
P
Patrick Caulfield 已提交
746
		struct connection *othercon = newcon->othercon;
747 748

		if (!othercon) {
D
David Teigland 已提交
749
			othercon = kmem_cache_zalloc(con_cache, GFP_NOFS);
750
			if (!othercon) {
D
David Teigland 已提交
751
				log_print("failed to allocate incoming socket");
752
				mutex_unlock(&newcon->sock_mutex);
753 754 755 756 757
				result = -ENOMEM;
				goto accept_err;
			}
			othercon->nodeid = nodeid;
			othercon->rx_action = receive_from_sock;
758
			mutex_init(&othercon->sock_mutex);
759 760
			INIT_WORK(&othercon->swork, process_send_sockets);
			INIT_WORK(&othercon->rwork, process_recv_sockets);
761
			set_bit(CF_IS_OTHERCON, &othercon->flags);
762 763
		}
		if (!othercon->sock) {
764
			newcon->othercon = othercon;
765 766 767 768 769 770 771 772
			othercon->sock = newsock;
			newsock->sk->sk_user_data = othercon;
			add_sock(newsock, othercon);
			addcon = othercon;
		}
		else {
			printk("Extra connection from node %d attempted\n", nodeid);
			result = -EAGAIN;
773
			mutex_unlock(&newcon->sock_mutex);
774
			goto accept_err;
775 776 777 778 779 780
		}
	}
	else {
		newsock->sk->sk_user_data = newcon;
		newcon->rx_action = receive_from_sock;
		add_sock(newsock, newcon);
P
Patrick Caulfield 已提交
781
		addcon = newcon;
782 783
	}

784
	mutex_unlock(&newcon->sock_mutex);
785 786 787 788 789 790

	/*
	 * Add it to the active queue in case we got data
	 * beween processing the accept adding the socket
	 * to the read_sockets list
	 */
P
Patrick Caulfield 已提交
791 792
	if (!test_and_set_bit(CF_READ_PENDING, &addcon->flags))
		queue_work(recv_workqueue, &addcon->rwork);
793
	mutex_unlock(&con->sock_mutex);
794 795 796

	return 0;

P
Patrick Caulfield 已提交
797
accept_err:
798
	mutex_unlock(&con->sock_mutex);
799 800 801
	sock_release(newsock);

	if (result != -EAGAIN)
D
David Teigland 已提交
802
		log_print("error accepting connection from node: %d", result);
803 804 805
	return result;
}

806 807 808 809 810 811 812 813 814 815 816 817 818 819 820 821 822 823 824 825 826 827 828 829 830 831 832 833 834 835 836 837 838 839 840 841 842 843 844 845 846 847 848 849 850 851 852 853 854 855 856 857 858 859 860 861 862 863 864 865 866 867 868 869 870 871 872 873 874 875 876 877 878
static void free_entry(struct writequeue_entry *e)
{
	__free_page(e->page);
	kfree(e);
}

/* Initiate an SCTP association.
   This is a special case of send_to_sock() in that we don't yet have a
   peeled-off socket for this association, so we use the listening socket
   and add the primary IP address of the remote node.
 */
static void sctp_init_assoc(struct connection *con)
{
	struct sockaddr_storage rem_addr;
	char outcmsg[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];
	struct msghdr outmessage;
	struct cmsghdr *cmsg;
	struct sctp_sndrcvinfo *sinfo;
	struct connection *base_con;
	struct writequeue_entry *e;
	int len, offset;
	int ret;
	int addrlen;
	struct kvec iov[1];

	if (test_and_set_bit(CF_INIT_PENDING, &con->flags))
		return;

	if (con->retries++ > MAX_CONNECT_RETRIES)
		return;

	log_print("Initiating association with node %d", con->nodeid);

	if (nodeid_to_addr(con->nodeid, (struct sockaddr *)&rem_addr)) {
		log_print("no address for nodeid %d", con->nodeid);
		return;
	}
	base_con = nodeid2con(0, 0);
	BUG_ON(base_con == NULL);

	make_sockaddr(&rem_addr, dlm_config.ci_tcp_port, &addrlen);

	outmessage.msg_name = &rem_addr;
	outmessage.msg_namelen = addrlen;
	outmessage.msg_control = outcmsg;
	outmessage.msg_controllen = sizeof(outcmsg);
	outmessage.msg_flags = MSG_EOR;

	spin_lock(&con->writequeue_lock);
	e = list_entry(con->writequeue.next, struct writequeue_entry,
		       list);

	BUG_ON((struct list_head *) e == &con->writequeue);

	len = e->len;
	offset = e->offset;
	spin_unlock(&con->writequeue_lock);

	/* Send the first block off the write queue */
	iov[0].iov_base = page_address(e->page)+offset;
	iov[0].iov_len = len;

	cmsg = CMSG_FIRSTHDR(&outmessage);
	cmsg->cmsg_level = IPPROTO_SCTP;
	cmsg->cmsg_type = SCTP_SNDRCV;
	cmsg->cmsg_len = CMSG_LEN(sizeof(struct sctp_sndrcvinfo));
	sinfo = CMSG_DATA(cmsg);
	memset(sinfo, 0x00, sizeof(struct sctp_sndrcvinfo));
	sinfo->sinfo_ppid = cpu_to_le32(dlm_our_nodeid());
	outmessage.msg_controllen = cmsg->cmsg_len;

	ret = kernel_sendmsg(base_con->sock, &outmessage, iov, 1, len);
	if (ret < 0) {
D
David Teigland 已提交
879 880
		log_print("Send first packet to node %d failed: %d",
			  con->nodeid, ret);
881 882 883 884 885 886 887 888 889 890 891 892 893 894 895 896 897 898

		/* Try again later */
		clear_bit(CF_CONNECT_PENDING, &con->flags);
		clear_bit(CF_INIT_PENDING, &con->flags);
	}
	else {
		spin_lock(&con->writequeue_lock);
		e->offset += ret;
		e->len -= ret;

		if (e->len == 0 && e->users == 0) {
			list_del(&e->list);
			free_entry(e);
		}
		spin_unlock(&con->writequeue_lock);
	}
}

899
/* Connect a new socket to its peer */
900
static void tcp_connect_to_sock(struct connection *con)
901 902
{
	int result = -EHOSTUNREACH;
903
	struct sockaddr_storage saddr, src_addr;
904
	int addr_len;
905
	struct socket *sock = NULL;
906 907 908

	if (con->nodeid == 0) {
		log_print("attempt to connect sock 0 foiled");
P
Patrick Caulfield 已提交
909
		return;
910 911
	}

912
	mutex_lock(&con->sock_mutex);
913 914 915 916 917 918 919 920 921 922
	if (con->retries++ > MAX_CONNECT_RETRIES)
		goto out;

	/* Some odd races can cause double-connects, ignore them */
	if (con->sock) {
		result = 0;
		goto out;
	}

	/* Create a socket to communicate with */
923
	result = sock_create_kern(dlm_local_addr[0]->ss_family, SOCK_STREAM,
P
Patrick Caulfield 已提交
924
				  IPPROTO_TCP, &sock);
925 926 927 928
	if (result < 0)
		goto out_err;

	memset(&saddr, 0, sizeof(saddr));
M
Masatake YAMATO 已提交
929 930
	if (dlm_nodeid_to_addr(con->nodeid, &saddr)) {
		sock_release(sock);
P
Patrick Caulfield 已提交
931
		goto out_err;
M
Masatake YAMATO 已提交
932
	}
933 934 935

	sock->sk->sk_user_data = con;
	con->rx_action = receive_from_sock;
936 937
	con->connect_action = tcp_connect_to_sock;
	add_sock(sock, con);
938

939 940 941 942 943 944 945 946 947 948 949
	/* Bind to our cluster-known address connecting to avoid
	   routing problems */
	memcpy(&src_addr, dlm_local_addr[0], sizeof(src_addr));
	make_sockaddr(&src_addr, 0, &addr_len);
	result = sock->ops->bind(sock, (struct sockaddr *) &src_addr,
				 addr_len);
	if (result < 0) {
		log_print("could not bind for connect: %d", result);
		/* This *may* not indicate a critical error */
	}

950
	make_sockaddr(&saddr, dlm_config.ci_tcp_port, &addr_len);
951 952 953 954

	log_print("connecting to %d", con->nodeid);
	result =
		sock->ops->connect(sock, (struct sockaddr *)&saddr, addr_len,
P
Patrick Caulfield 已提交
955
				   O_NONBLOCK);
956 957
	if (result == -EINPROGRESS)
		result = 0;
P
Patrick Caulfield 已提交
958 959
	if (result == 0)
		goto out;
960

P
Patrick Caulfield 已提交
961
out_err:
962 963 964
	if (con->sock) {
		sock_release(con->sock);
		con->sock = NULL;
965 966
	} else if (sock) {
		sock_release(sock);
967 968 969 970 971 972
	}
	/*
	 * Some errors are fatal and this list might need adjusting. For other
	 * errors we try again until the max number of retries is reached.
	 */
	if (result != -EHOSTUNREACH && result != -ENETUNREACH &&
973
	    result != -ENETDOWN && result != -EINVAL
974 975 976 977
	    && result != -EPROTONOSUPPORT) {
		lowcomms_connect_sock(con);
		result = 0;
	}
P
Patrick Caulfield 已提交
978
out:
979
	mutex_unlock(&con->sock_mutex);
P
Patrick Caulfield 已提交
980
	return;
981 982
}

983 984
static struct socket *tcp_create_listen_sock(struct connection *con,
					     struct sockaddr_storage *saddr)
985
{
P
Patrick Caulfield 已提交
986
	struct socket *sock = NULL;
987 988 989 990
	int result = 0;
	int one = 1;
	int addr_len;

991
	if (dlm_local_addr[0]->ss_family == AF_INET)
992 993 994 995 996
		addr_len = sizeof(struct sockaddr_in);
	else
		addr_len = sizeof(struct sockaddr_in6);

	/* Create a socket to communicate with */
D
David Teigland 已提交
997 998
	result = sock_create_kern(dlm_local_addr[0]->ss_family, SOCK_STREAM,
				  IPPROTO_TCP, &sock);
999
	if (result < 0) {
D
David Teigland 已提交
1000
		log_print("Can't create listening comms socket");
1001 1002 1003
		goto create_out;
	}

1004 1005 1006
	result = kernel_setsockopt(sock, SOL_SOCKET, SO_REUSEADDR,
				   (char *)&one, sizeof(one));

1007
	if (result < 0) {
D
David Teigland 已提交
1008
		log_print("Failed to set SO_REUSEADDR on socket: %d", result);
1009 1010
	}
	sock->sk->sk_user_data = con;
1011 1012
	con->rx_action = tcp_accept_from_sock;
	con->connect_action = tcp_connect_to_sock;
1013 1014 1015
	con->sock = sock;

	/* Bind to our port */
1016
	make_sockaddr(saddr, dlm_config.ci_tcp_port, &addr_len);
1017 1018
	result = sock->ops->bind(sock, (struct sockaddr *) saddr, addr_len);
	if (result < 0) {
D
David Teigland 已提交
1019
		log_print("Can't bind to port %d", dlm_config.ci_tcp_port);
1020 1021 1022 1023 1024
		sock_release(sock);
		sock = NULL;
		con->sock = NULL;
		goto create_out;
	}
1025
	result = kernel_setsockopt(sock, SOL_SOCKET, SO_KEEPALIVE,
P
Patrick Caulfield 已提交
1026
				 (char *)&one, sizeof(one));
1027
	if (result < 0) {
D
David Teigland 已提交
1028
		log_print("Set keepalive failed: %d", result);
1029 1030 1031 1032
	}

	result = sock->ops->listen(sock, 5);
	if (result < 0) {
D
David Teigland 已提交
1033
		log_print("Can't listen on port %d", dlm_config.ci_tcp_port);
1034 1035 1036 1037 1038
		sock_release(sock);
		sock = NULL;
		goto create_out;
	}

P
Patrick Caulfield 已提交
1039
create_out:
1040 1041 1042
	return sock;
}

1043 1044 1045 1046 1047 1048
/* Get local addresses */
static void init_local(void)
{
	struct sockaddr_storage sas, *addr;
	int i;

1049
	dlm_local_count = 0;
1050 1051 1052 1053 1054 1055 1056 1057 1058 1059 1060 1061
	for (i = 0; i < DLM_MAX_ADDR_COUNT - 1; i++) {
		if (dlm_our_addr(&sas, i))
			break;

		addr = kmalloc(sizeof(*addr), GFP_KERNEL);
		if (!addr)
			break;
		memcpy(addr, &sas, sizeof(*addr));
		dlm_local_addr[dlm_local_count++] = addr;
	}
}

D
David Teigland 已提交
1062 1063 1064 1065 1066
/* Bind to an IP address. SCTP allows multiple address so it can do
   multi-homing */
static int add_sctp_bind_addr(struct connection *sctp_con,
			      struct sockaddr_storage *addr,
			      int addr_len, int num)
1067 1068 1069 1070 1071 1072 1073 1074 1075 1076 1077 1078 1079 1080 1081 1082 1083 1084
{
	int result = 0;

	if (num == 1)
		result = kernel_bind(sctp_con->sock,
				     (struct sockaddr *) addr,
				     addr_len);
	else
		result = kernel_setsockopt(sctp_con->sock, SOL_SCTP,
					   SCTP_SOCKOPT_BINDX_ADD,
					   (char *)addr, addr_len);

	if (result < 0)
		log_print("Can't bind to port %d addr number %d",
			  dlm_config.ci_tcp_port, num);

	return result;
}
1085

1086 1087 1088 1089 1090 1091 1092 1093 1094 1095 1096 1097 1098 1099 1100 1101 1102 1103 1104 1105 1106 1107 1108 1109 1110 1111 1112 1113 1114 1115
/* Initialise SCTP socket and bind to all interfaces */
static int sctp_listen_for_all(void)
{
	struct socket *sock = NULL;
	struct sockaddr_storage localaddr;
	struct sctp_event_subscribe subscribe;
	int result = -EINVAL, num = 1, i, addr_len;
	struct connection *con = nodeid2con(0, GFP_KERNEL);
	int bufsize = NEEDED_RMEM;

	if (!con)
		return -ENOMEM;

	log_print("Using SCTP for communications");

	result = sock_create_kern(dlm_local_addr[0]->ss_family, SOCK_SEQPACKET,
				  IPPROTO_SCTP, &sock);
	if (result < 0) {
		log_print("Can't create comms socket, check SCTP is loaded");
		goto out;
	}

	/* Listen for events */
	memset(&subscribe, 0, sizeof(subscribe));
	subscribe.sctp_data_io_event = 1;
	subscribe.sctp_association_event = 1;
	subscribe.sctp_send_failure_event = 1;
	subscribe.sctp_shutdown_event = 1;
	subscribe.sctp_partial_delivery_event = 1;

1116
	result = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUFFORCE,
1117 1118
				 (char *)&bufsize, sizeof(bufsize));
	if (result)
D
David Teigland 已提交
1119
		log_print("Error increasing buffer space on socket %d", result);
1120 1121

	result = kernel_setsockopt(sock, SOL_SCTP, SCTP_EVENTS,
D
David Teigland 已提交
1122
				   (char *)&subscribe, sizeof(subscribe));
1123 1124 1125 1126 1127 1128 1129 1130 1131 1132 1133 1134 1135 1136 1137 1138 1139 1140 1141 1142 1143 1144 1145 1146 1147 1148 1149 1150 1151 1152 1153 1154 1155 1156 1157 1158 1159 1160 1161 1162
	if (result < 0) {
		log_print("Failed to set SCTP_EVENTS on socket: result=%d",
			  result);
		goto create_delsock;
	}

	/* Init con struct */
	sock->sk->sk_user_data = con;
	con->sock = sock;
	con->sock->sk->sk_data_ready = lowcomms_data_ready;
	con->rx_action = receive_from_sock;
	con->connect_action = sctp_init_assoc;

	/* Bind to all interfaces. */
	for (i = 0; i < dlm_local_count; i++) {
		memcpy(&localaddr, dlm_local_addr[i], sizeof(localaddr));
		make_sockaddr(&localaddr, dlm_config.ci_tcp_port, &addr_len);

		result = add_sctp_bind_addr(con, &localaddr, addr_len, num);
		if (result)
			goto create_delsock;
		++num;
	}

	result = sock->ops->listen(sock, 5);
	if (result < 0) {
		log_print("Can't set socket listening");
		goto create_delsock;
	}

	return 0;

create_delsock:
	sock_release(sock);
	con->sock = NULL;
out:
	return result;
}

static int tcp_listen_for_all(void)
1163 1164 1165 1166 1167
{
	struct socket *sock = NULL;
	struct connection *con = nodeid2con(0, GFP_KERNEL);
	int result = -EINVAL;

1168 1169 1170
	if (!con)
		return -ENOMEM;

1171
	/* We don't support multi-homed hosts */
1172
	if (dlm_local_addr[1] != NULL) {
D
David Teigland 已提交
1173 1174
		log_print("TCP protocol can't handle multi-homed hosts, "
			  "try SCTP");
1175 1176 1177 1178 1179 1180
		return -EINVAL;
	}

	log_print("Using TCP for communications");

	sock = tcp_create_listen_sock(con, dlm_local_addr[0]);
1181 1182 1183 1184 1185 1186 1187 1188 1189 1190 1191 1192 1193 1194 1195 1196 1197 1198 1199 1200 1201 1202 1203 1204 1205 1206 1207 1208 1209 1210 1211 1212 1213 1214 1215 1216 1217
	if (sock) {
		add_sock(sock, con);
		result = 0;
	}
	else {
		result = -EADDRINUSE;
	}

	return result;
}



static struct writequeue_entry *new_writequeue_entry(struct connection *con,
						     gfp_t allocation)
{
	struct writequeue_entry *entry;

	entry = kmalloc(sizeof(struct writequeue_entry), allocation);
	if (!entry)
		return NULL;

	entry->page = alloc_page(allocation);
	if (!entry->page) {
		kfree(entry);
		return NULL;
	}

	entry->offset = 0;
	entry->len = 0;
	entry->end = 0;
	entry->users = 0;
	entry->con = con;

	return entry;
}

D
David Teigland 已提交
1218
void *dlm_lowcomms_get_buffer(int nodeid, int len, gfp_t allocation, char **ppc)
1219 1220 1221 1222 1223 1224 1225 1226 1227 1228
{
	struct connection *con;
	struct writequeue_entry *e;
	int offset = 0;
	int users = 0;

	con = nodeid2con(nodeid, allocation);
	if (!con)
		return NULL;

1229
	spin_lock(&con->writequeue_lock);
1230
	e = list_entry(con->writequeue.prev, struct writequeue_entry, list);
P
Patrick Caulfield 已提交
1231
	if ((&e->list == &con->writequeue) ||
1232 1233 1234 1235 1236 1237 1238 1239 1240 1241
	    (PAGE_CACHE_SIZE - e->end < len)) {
		e = NULL;
	} else {
		offset = e->end;
		e->end += len;
		users = e->users++;
	}
	spin_unlock(&con->writequeue_lock);

	if (e) {
P
Patrick Caulfield 已提交
1242
	got_one:
1243 1244 1245 1246 1247 1248 1249 1250 1251 1252 1253 1254 1255 1256 1257 1258 1259 1260 1261 1262 1263 1264 1265
		*ppc = page_address(e->page) + offset;
		return e;
	}

	e = new_writequeue_entry(con, allocation);
	if (e) {
		spin_lock(&con->writequeue_lock);
		offset = e->end;
		e->end += len;
		users = e->users++;
		list_add_tail(&e->list, &con->writequeue);
		spin_unlock(&con->writequeue_lock);
		goto got_one;
	}
	return NULL;
}

void dlm_lowcomms_commit_buffer(void *mh)
{
	struct writequeue_entry *e = (struct writequeue_entry *)mh;
	struct connection *con = e->con;
	int users;

1266
	spin_lock(&con->writequeue_lock);
1267 1268 1269 1270 1271 1272
	users = --e->users;
	if (users)
		goto out;
	e->len = e->end - e->offset;
	spin_unlock(&con->writequeue_lock);

1273 1274
	if (!test_and_set_bit(CF_WRITE_PENDING, &con->flags)) {
		queue_work(send_workqueue, &con->swork);
1275 1276 1277
	}
	return;

P
Patrick Caulfield 已提交
1278
out:
1279 1280 1281 1282 1283
	spin_unlock(&con->writequeue_lock);
	return;
}

/* Send a message */
P
Patrick Caulfield 已提交
1284
static void send_to_sock(struct connection *con)
1285 1286 1287 1288 1289 1290 1291
{
	int ret = 0;
	ssize_t(*sendpage) (struct socket *, struct page *, int, size_t, int);
	const int msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL;
	struct writequeue_entry *e;
	int len, offset;

1292
	mutex_lock(&con->sock_mutex);
1293 1294 1295 1296 1297 1298 1299 1300 1301 1302 1303 1304 1305 1306 1307 1308 1309 1310 1311 1312 1313
	if (con->sock == NULL)
		goto out_connect;

	sendpage = con->sock->ops->sendpage;

	spin_lock(&con->writequeue_lock);
	for (;;) {
		e = list_entry(con->writequeue.next, struct writequeue_entry,
			       list);
		if ((struct list_head *) e == &con->writequeue)
			break;

		len = e->len;
		offset = e->offset;
		BUG_ON(len == 0 && e->users == 0);
		spin_unlock(&con->writequeue_lock);

		ret = 0;
		if (len) {
			ret = sendpage(con->sock, e->page, offset, len,
				       msg_flags);
1314 1315
			if (ret == -EAGAIN || ret == 0) {
				cond_resched();
1316
				goto out;
1317
			}
1318 1319
			if (ret <= 0)
				goto send_error;
1320
		}
1321
			/* Don't starve people filling buffers */
P
Patrick Caulfield 已提交
1322
			cond_resched();
1323 1324 1325 1326 1327 1328 1329 1330 1331 1332 1333 1334

		spin_lock(&con->writequeue_lock);
		e->offset += ret;
		e->len -= ret;

		if (e->len == 0 && e->users == 0) {
			list_del(&e->list);
			free_entry(e);
			continue;
		}
	}
	spin_unlock(&con->writequeue_lock);
P
Patrick Caulfield 已提交
1335
out:
1336
	mutex_unlock(&con->sock_mutex);
P
Patrick Caulfield 已提交
1337
	return;
1338

P
Patrick Caulfield 已提交
1339
send_error:
1340
	mutex_unlock(&con->sock_mutex);
P
Patrick Caulfield 已提交
1341
	close_connection(con, false);
1342
	lowcomms_connect_sock(con);
P
Patrick Caulfield 已提交
1343
	return;
1344

P
Patrick Caulfield 已提交
1345
out_connect:
1346
	mutex_unlock(&con->sock_mutex);
1347 1348
	if (!test_bit(CF_INIT_PENDING, &con->flags))
		lowcomms_connect_sock(con);
P
Patrick Caulfield 已提交
1349
	return;
1350 1351 1352 1353
}

static void clean_one_writequeue(struct connection *con)
{
1354
	struct writequeue_entry *e, *safe;
1355 1356

	spin_lock(&con->writequeue_lock);
1357
	list_for_each_entry_safe(e, safe, &con->writequeue, list) {
1358 1359 1360 1361 1362 1363 1364 1365 1366 1367 1368 1369 1370 1371 1372 1373
		list_del(&e->list);
		free_entry(e);
	}
	spin_unlock(&con->writequeue_lock);
}

/* Called from recovery when it knows that a node has
   left the cluster */
int dlm_lowcomms_close(int nodeid)
{
	struct connection *con;

	log_print("closing connection to node %d", nodeid);
	con = nodeid2con(nodeid, 0);
	if (con) {
		clean_one_writequeue(con);
P
Patrick Caulfield 已提交
1374
		close_connection(con, true);
1375 1376 1377 1378
	}
	return 0;
}

1379
/* Receive workqueue function */
1380
static void process_recv_sockets(struct work_struct *work)
1381
{
1382 1383
	struct connection *con = container_of(work, struct connection, rwork);
	int err;
1384

1385 1386 1387 1388
	clear_bit(CF_READ_PENDING, &con->flags);
	do {
		err = con->rx_action(con);
	} while (!err);
1389 1390
}

1391
/* Send workqueue function */
1392
static void process_send_sockets(struct work_struct *work)
1393
{
1394
	struct connection *con = container_of(work, struct connection, swork);
1395

1396
	if (test_and_clear_bit(CF_CONNECT_PENDING, &con->flags)) {
1397
		con->connect_action(con);
1398
	}
P
Patrick Caulfield 已提交
1399 1400
	clear_bit(CF_WRITE_PENDING, &con->flags);
	send_to_sock(con);
1401 1402 1403 1404 1405 1406
}


/* Discard all entries on the write queues */
static void clean_writequeues(void)
{
1407
	foreach_conn(clean_one_writequeue);
1408 1409
}

1410
static void work_stop(void)
1411
{
1412 1413
	destroy_workqueue(recv_workqueue);
	destroy_workqueue(send_workqueue);
1414 1415
}

1416
static int work_start(void)
1417 1418
{
	int error;
1419 1420
	recv_workqueue = create_workqueue("dlm_recv");
	error = IS_ERR(recv_workqueue);
P
Patrick Caulfield 已提交
1421
	if (error) {
1422
		log_print("can't start dlm_recv %d", error);
1423 1424 1425
		return error;
	}

1426 1427
	send_workqueue = create_singlethread_workqueue("dlm_send");
	error = IS_ERR(send_workqueue);
P
Patrick Caulfield 已提交
1428
	if (error) {
1429 1430
		log_print("can't start dlm_send %d", error);
		destroy_workqueue(recv_workqueue);
1431 1432 1433 1434 1435 1436
		return error;
	}

	return 0;
}

1437
static void stop_conn(struct connection *con)
1438
{
1439
	con->flags |= 0x0F;
1440
	if (con->sock && con->sock->sk)
1441 1442
		con->sock->sk->sk_user_data = NULL;
}
1443

1444 1445 1446 1447 1448 1449 1450 1451 1452 1453 1454
static void free_conn(struct connection *con)
{
	close_connection(con, true);
	if (con->othercon)
		kmem_cache_free(con_cache, con->othercon);
	hlist_del(&con->list);
	kmem_cache_free(con_cache, con);
}

void dlm_lowcomms_stop(void)
{
P
Patrick Caulfield 已提交
1455
	/* Set all the flags to prevent any
1456 1457
	   socket activity.
	*/
1458
	mutex_lock(&connections_lock);
1459
	foreach_conn(stop_conn);
1460
	mutex_unlock(&connections_lock);
P
Patrick Caulfield 已提交
1461

1462
	work_stop();
1463

1464
	mutex_lock(&connections_lock);
1465 1466
	clean_writequeues();

1467 1468
	foreach_conn(free_conn);

1469
	mutex_unlock(&connections_lock);
1470 1471 1472 1473 1474
	kmem_cache_destroy(con_cache);
}

int dlm_lowcomms_start(void)
{
1475 1476
	int error = -EINVAL;
	struct connection *con;
1477 1478 1479 1480
	int i;

	for (i = 0; i < CONN_HASH_SIZE; i++)
		INIT_HLIST_HEAD(&connection_hash[i]);
1481

1482 1483
	init_local();
	if (!dlm_local_count) {
D
David Teigland 已提交
1484
		error = -ENOTCONN;
1485
		log_print("no local IP address has been set");
1486
		goto out;
1487 1488
	}

1489
	error = -ENOMEM;
1490
	con_cache = kmem_cache_create("dlm_conn", sizeof(struct connection),
P
Patrick Caulfield 已提交
1491
				      __alignof__(struct connection), 0,
1492
				      NULL);
1493
	if (!con_cache)
1494
		goto out;
1495 1496

	/* Start listening */
1497 1498 1499 1500
	if (dlm_config.ci_protocol == 0)
		error = tcp_listen_for_all();
	else
		error = sctp_listen_for_all();
1501 1502 1503
	if (error)
		goto fail_unlisten;

1504
	error = work_start();
1505 1506 1507 1508 1509
	if (error)
		goto fail_unlisten;

	return 0;

P
Patrick Caulfield 已提交
1510
fail_unlisten:
1511 1512 1513 1514 1515
	con = nodeid2con(0,0);
	if (con) {
		close_connection(con, false);
		kmem_cache_free(con_cache, con);
	}
1516 1517
	kmem_cache_destroy(con_cache);

P
Patrick Caulfield 已提交
1518
out:
1519 1520
	return error;
}