xref: /openbmc/linux/fs/dlm/lowcomms.c (revision 8e9356c6)
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 **  This copyrighted material is made available to anyone wishing to use,
8 **  modify, copy, or redistribute it subject to the terms and conditions
9 **  of the GNU General Public License v.2.
10 **
11 *******************************************************************************
12 ******************************************************************************/
13 
14 /*
15  * lowcomms.c
16  *
17  * This is the "low-level" comms layer.
18  *
19  * It is responsible for sending/receiving messages
20  * from other nodes in the cluster.
21  *
22  * Cluster nodes are referred to by their nodeids. nodeids are
23  * simply 32 bit numbers to the locking module - if they need to
24  * be expanded for the cluster infrastructure then that is its
25  * responsibility. It is this layer's
26  * responsibility to resolve these into IP address or
27  * whatever it needs for inter-node communication.
28  *
29  * The comms level is two kernel threads that deal mainly with
30  * the receiving of messages from other nodes and passing them
31  * up to the mid-level comms layer (which understands the
32  * message format) for execution by the locking core, and
33  * a send thread which does all the setting up of connections
34  * to remote nodes and the sending of data. Threads are not allowed
35  * to send their own data because it may cause them to wait in times
36  * of high load. Also, this way, the sending thread can collect together
37  * messages bound for one node and send them in one block.
38  *
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
41  * to 0 (default) for TCP or 1 for SCTP. It should be configured using a
42  * cluster-wide mechanism as it must be the same on all nodes of the cluster
43  * for the DLM to function.
44  *
45  */
46 
47 #include <asm/ioctls.h>
48 #include <net/sock.h>
49 #include <net/tcp.h>
50 #include <linux/pagemap.h>
51 #include <linux/file.h>
52 #include <linux/mutex.h>
53 #include <linux/sctp.h>
54 #include <linux/slab.h>
55 #include <net/sctp/sctp.h>
56 #include <net/ipv6.h>
57 
58 #include "dlm_internal.h"
59 #include "lowcomms.h"
60 #include "midcomms.h"
61 #include "config.h"
62 
63 #define NEEDED_RMEM (4*1024*1024)
64 #define CONN_HASH_SIZE 32
65 
66 /* Number of messages to send before rescheduling */
67 #define MAX_SEND_MSG_COUNT 25
68 
69 struct cbuf {
70 	unsigned int base;
71 	unsigned int len;
72 	unsigned int mask;
73 };
74 
75 static void cbuf_add(struct cbuf *cb, int n)
76 {
77 	cb->len += n;
78 }
79 
80 static int cbuf_data(struct cbuf *cb)
81 {
82 	return ((cb->base + cb->len) & cb->mask);
83 }
84 
85 static void cbuf_init(struct cbuf *cb, int size)
86 {
87 	cb->base = cb->len = 0;
88 	cb->mask = size-1;
89 }
90 
91 static void cbuf_eat(struct cbuf *cb, int n)
92 {
93 	cb->len  -= n;
94 	cb->base += n;
95 	cb->base &= cb->mask;
96 }
97 
98 static bool cbuf_empty(struct cbuf *cb)
99 {
100 	return cb->len == 0;
101 }
102 
103 struct connection {
104 	struct socket *sock;	/* NULL if not connected */
105 	uint32_t nodeid;	/* So we know who we are in the list */
106 	struct mutex sock_mutex;
107 	unsigned long flags;
108 #define CF_READ_PENDING 1
109 #define CF_WRITE_PENDING 2
110 #define CF_CONNECT_PENDING 3
111 #define CF_INIT_PENDING 4
112 #define CF_IS_OTHERCON 5
113 #define CF_CLOSE 6
114 #define CF_APP_LIMITED 7
115 	struct list_head writequeue;  /* List of outgoing writequeue_entries */
116 	spinlock_t writequeue_lock;
117 	int (*rx_action) (struct connection *);	/* What to do when active */
118 	void (*connect_action) (struct connection *);	/* What to do to connect */
119 	struct page *rx_page;
120 	struct cbuf cb;
121 	int retries;
122 #define MAX_CONNECT_RETRIES 3
123 	int sctp_assoc;
124 	struct hlist_node list;
125 	struct connection *othercon;
126 	struct work_struct rwork; /* Receive workqueue */
127 	struct work_struct swork; /* Send workqueue */
128 	bool try_new_addr;
129 };
130 #define sock2con(x) ((struct connection *)(x)->sk_user_data)
131 
132 /* An entry waiting to be sent */
133 struct writequeue_entry {
134 	struct list_head list;
135 	struct page *page;
136 	int offset;
137 	int len;
138 	int end;
139 	int users;
140 	struct connection *con;
141 };
142 
143 struct dlm_node_addr {
144 	struct list_head list;
145 	int nodeid;
146 	int addr_count;
147 	int curr_addr_index;
148 	struct sockaddr_storage *addr[DLM_MAX_ADDR_COUNT];
149 };
150 
151 static LIST_HEAD(dlm_node_addrs);
152 static DEFINE_SPINLOCK(dlm_node_addrs_spin);
153 
154 static struct sockaddr_storage *dlm_local_addr[DLM_MAX_ADDR_COUNT];
155 static int dlm_local_count;
156 static int dlm_allow_conn;
157 
158 /* Work queues */
159 static struct workqueue_struct *recv_workqueue;
160 static struct workqueue_struct *send_workqueue;
161 
162 static struct hlist_head connection_hash[CONN_HASH_SIZE];
163 static DEFINE_MUTEX(connections_lock);
164 static struct kmem_cache *con_cache;
165 
166 static void process_recv_sockets(struct work_struct *work);
167 static void process_send_sockets(struct work_struct *work);
168 
169 
170 /* This is deliberately very simple because most clusters have simple
171    sequential nodeids, so we should be able to go straight to a connection
172    struct in the array */
173 static inline int nodeid_hash(int nodeid)
174 {
175 	return nodeid & (CONN_HASH_SIZE-1);
176 }
177 
178 static struct connection *__find_con(int nodeid)
179 {
180 	int r;
181 	struct connection *con;
182 
183 	r = nodeid_hash(nodeid);
184 
185 	hlist_for_each_entry(con, &connection_hash[r], list) {
186 		if (con->nodeid == nodeid)
187 			return con;
188 	}
189 	return NULL;
190 }
191 
192 /*
193  * If 'allocation' is zero then we don't attempt to create a new
194  * connection structure for this node.
195  */
196 static struct connection *__nodeid2con(int nodeid, gfp_t alloc)
197 {
198 	struct connection *con = NULL;
199 	int r;
200 
201 	con = __find_con(nodeid);
202 	if (con || !alloc)
203 		return con;
204 
205 	con = kmem_cache_zalloc(con_cache, alloc);
206 	if (!con)
207 		return NULL;
208 
209 	r = nodeid_hash(nodeid);
210 	hlist_add_head(&con->list, &connection_hash[r]);
211 
212 	con->nodeid = nodeid;
213 	mutex_init(&con->sock_mutex);
214 	INIT_LIST_HEAD(&con->writequeue);
215 	spin_lock_init(&con->writequeue_lock);
216 	INIT_WORK(&con->swork, process_send_sockets);
217 	INIT_WORK(&con->rwork, process_recv_sockets);
218 
219 	/* Setup action pointers for child sockets */
220 	if (con->nodeid) {
221 		struct connection *zerocon = __find_con(0);
222 
223 		con->connect_action = zerocon->connect_action;
224 		if (!con->rx_action)
225 			con->rx_action = zerocon->rx_action;
226 	}
227 
228 	return con;
229 }
230 
231 /* Loop round all connections */
232 static void foreach_conn(void (*conn_func)(struct connection *c))
233 {
234 	int i;
235 	struct hlist_node *n;
236 	struct connection *con;
237 
238 	for (i = 0; i < CONN_HASH_SIZE; i++) {
239 		hlist_for_each_entry_safe(con, n, &connection_hash[i], list)
240 			conn_func(con);
241 	}
242 }
243 
244 static struct connection *nodeid2con(int nodeid, gfp_t allocation)
245 {
246 	struct connection *con;
247 
248 	mutex_lock(&connections_lock);
249 	con = __nodeid2con(nodeid, allocation);
250 	mutex_unlock(&connections_lock);
251 
252 	return con;
253 }
254 
255 /* This is a bit drastic, but only called when things go wrong */
256 static struct connection *assoc2con(int assoc_id)
257 {
258 	int i;
259 	struct connection *con;
260 
261 	mutex_lock(&connections_lock);
262 
263 	for (i = 0 ; i < CONN_HASH_SIZE; i++) {
264 		hlist_for_each_entry(con, &connection_hash[i], list) {
265 			if (con->sctp_assoc == assoc_id) {
266 				mutex_unlock(&connections_lock);
267 				return con;
268 			}
269 		}
270 	}
271 	mutex_unlock(&connections_lock);
272 	return NULL;
273 }
274 
275 static struct dlm_node_addr *find_node_addr(int nodeid)
276 {
277 	struct dlm_node_addr *na;
278 
279 	list_for_each_entry(na, &dlm_node_addrs, list) {
280 		if (na->nodeid == nodeid)
281 			return na;
282 	}
283 	return NULL;
284 }
285 
286 static int addr_compare(struct sockaddr_storage *x, struct sockaddr_storage *y)
287 {
288 	switch (x->ss_family) {
289 	case AF_INET: {
290 		struct sockaddr_in *sinx = (struct sockaddr_in *)x;
291 		struct sockaddr_in *siny = (struct sockaddr_in *)y;
292 		if (sinx->sin_addr.s_addr != siny->sin_addr.s_addr)
293 			return 0;
294 		if (sinx->sin_port != siny->sin_port)
295 			return 0;
296 		break;
297 	}
298 	case AF_INET6: {
299 		struct sockaddr_in6 *sinx = (struct sockaddr_in6 *)x;
300 		struct sockaddr_in6 *siny = (struct sockaddr_in6 *)y;
301 		if (!ipv6_addr_equal(&sinx->sin6_addr, &siny->sin6_addr))
302 			return 0;
303 		if (sinx->sin6_port != siny->sin6_port)
304 			return 0;
305 		break;
306 	}
307 	default:
308 		return 0;
309 	}
310 	return 1;
311 }
312 
313 static int nodeid_to_addr(int nodeid, struct sockaddr_storage *sas_out,
314 			  struct sockaddr *sa_out, bool try_new_addr)
315 {
316 	struct sockaddr_storage sas;
317 	struct dlm_node_addr *na;
318 
319 	if (!dlm_local_count)
320 		return -1;
321 
322 	spin_lock(&dlm_node_addrs_spin);
323 	na = find_node_addr(nodeid);
324 	if (na && na->addr_count) {
325 		if (try_new_addr) {
326 			na->curr_addr_index++;
327 			if (na->curr_addr_index == na->addr_count)
328 				na->curr_addr_index = 0;
329 		}
330 
331 		memcpy(&sas, na->addr[na->curr_addr_index ],
332 			sizeof(struct sockaddr_storage));
333 	}
334 	spin_unlock(&dlm_node_addrs_spin);
335 
336 	if (!na)
337 		return -EEXIST;
338 
339 	if (!na->addr_count)
340 		return -ENOENT;
341 
342 	if (sas_out)
343 		memcpy(sas_out, &sas, sizeof(struct sockaddr_storage));
344 
345 	if (!sa_out)
346 		return 0;
347 
348 	if (dlm_local_addr[0]->ss_family == AF_INET) {
349 		struct sockaddr_in *in4  = (struct sockaddr_in *) &sas;
350 		struct sockaddr_in *ret4 = (struct sockaddr_in *) sa_out;
351 		ret4->sin_addr.s_addr = in4->sin_addr.s_addr;
352 	} else {
353 		struct sockaddr_in6 *in6  = (struct sockaddr_in6 *) &sas;
354 		struct sockaddr_in6 *ret6 = (struct sockaddr_in6 *) sa_out;
355 		ret6->sin6_addr = in6->sin6_addr;
356 	}
357 
358 	return 0;
359 }
360 
361 static int addr_to_nodeid(struct sockaddr_storage *addr, int *nodeid)
362 {
363 	struct dlm_node_addr *na;
364 	int rv = -EEXIST;
365 	int addr_i;
366 
367 	spin_lock(&dlm_node_addrs_spin);
368 	list_for_each_entry(na, &dlm_node_addrs, list) {
369 		if (!na->addr_count)
370 			continue;
371 
372 		for (addr_i = 0; addr_i < na->addr_count; addr_i++) {
373 			if (addr_compare(na->addr[addr_i], addr)) {
374 				*nodeid = na->nodeid;
375 				rv = 0;
376 				goto unlock;
377 			}
378 		}
379 	}
380 unlock:
381 	spin_unlock(&dlm_node_addrs_spin);
382 	return rv;
383 }
384 
385 int dlm_lowcomms_addr(int nodeid, struct sockaddr_storage *addr, int len)
386 {
387 	struct sockaddr_storage *new_addr;
388 	struct dlm_node_addr *new_node, *na;
389 
390 	new_node = kzalloc(sizeof(struct dlm_node_addr), GFP_NOFS);
391 	if (!new_node)
392 		return -ENOMEM;
393 
394 	new_addr = kzalloc(sizeof(struct sockaddr_storage), GFP_NOFS);
395 	if (!new_addr) {
396 		kfree(new_node);
397 		return -ENOMEM;
398 	}
399 
400 	memcpy(new_addr, addr, len);
401 
402 	spin_lock(&dlm_node_addrs_spin);
403 	na = find_node_addr(nodeid);
404 	if (!na) {
405 		new_node->nodeid = nodeid;
406 		new_node->addr[0] = new_addr;
407 		new_node->addr_count = 1;
408 		list_add(&new_node->list, &dlm_node_addrs);
409 		spin_unlock(&dlm_node_addrs_spin);
410 		return 0;
411 	}
412 
413 	if (na->addr_count >= DLM_MAX_ADDR_COUNT) {
414 		spin_unlock(&dlm_node_addrs_spin);
415 		kfree(new_addr);
416 		kfree(new_node);
417 		return -ENOSPC;
418 	}
419 
420 	na->addr[na->addr_count++] = new_addr;
421 	spin_unlock(&dlm_node_addrs_spin);
422 	kfree(new_node);
423 	return 0;
424 }
425 
426 /* Data available on socket or listen socket received a connect */
427 static void lowcomms_data_ready(struct sock *sk, int count_unused)
428 {
429 	struct connection *con = sock2con(sk);
430 	if (con && !test_and_set_bit(CF_READ_PENDING, &con->flags))
431 		queue_work(recv_workqueue, &con->rwork);
432 }
433 
434 static void lowcomms_write_space(struct sock *sk)
435 {
436 	struct connection *con = sock2con(sk);
437 
438 	if (!con)
439 		return;
440 
441 	clear_bit(SOCK_NOSPACE, &con->sock->flags);
442 
443 	if (test_and_clear_bit(CF_APP_LIMITED, &con->flags)) {
444 		con->sock->sk->sk_write_pending--;
445 		clear_bit(SOCK_ASYNC_NOSPACE, &con->sock->flags);
446 	}
447 
448 	if (!test_and_set_bit(CF_WRITE_PENDING, &con->flags))
449 		queue_work(send_workqueue, &con->swork);
450 }
451 
452 static inline void lowcomms_connect_sock(struct connection *con)
453 {
454 	if (test_bit(CF_CLOSE, &con->flags))
455 		return;
456 	if (!test_and_set_bit(CF_CONNECT_PENDING, &con->flags))
457 		queue_work(send_workqueue, &con->swork);
458 }
459 
460 static void lowcomms_state_change(struct sock *sk)
461 {
462 	if (sk->sk_state == TCP_ESTABLISHED)
463 		lowcomms_write_space(sk);
464 }
465 
466 int dlm_lowcomms_connect_node(int nodeid)
467 {
468 	struct connection *con;
469 
470 	/* with sctp there's no connecting without sending */
471 	if (dlm_config.ci_protocol != 0)
472 		return 0;
473 
474 	if (nodeid == dlm_our_nodeid())
475 		return 0;
476 
477 	con = nodeid2con(nodeid, GFP_NOFS);
478 	if (!con)
479 		return -ENOMEM;
480 	lowcomms_connect_sock(con);
481 	return 0;
482 }
483 
484 /* Make a socket active */
485 static void add_sock(struct socket *sock, struct connection *con)
486 {
487 	con->sock = sock;
488 
489 	/* Install a data_ready callback */
490 	con->sock->sk->sk_data_ready = lowcomms_data_ready;
491 	con->sock->sk->sk_write_space = lowcomms_write_space;
492 	con->sock->sk->sk_state_change = lowcomms_state_change;
493 	con->sock->sk->sk_user_data = con;
494 	con->sock->sk->sk_allocation = GFP_NOFS;
495 }
496 
497 /* Add the port number to an IPv6 or 4 sockaddr and return the address
498    length */
499 static void make_sockaddr(struct sockaddr_storage *saddr, uint16_t port,
500 			  int *addr_len)
501 {
502 	saddr->ss_family =  dlm_local_addr[0]->ss_family;
503 	if (saddr->ss_family == AF_INET) {
504 		struct sockaddr_in *in4_addr = (struct sockaddr_in *)saddr;
505 		in4_addr->sin_port = cpu_to_be16(port);
506 		*addr_len = sizeof(struct sockaddr_in);
507 		memset(&in4_addr->sin_zero, 0, sizeof(in4_addr->sin_zero));
508 	} else {
509 		struct sockaddr_in6 *in6_addr = (struct sockaddr_in6 *)saddr;
510 		in6_addr->sin6_port = cpu_to_be16(port);
511 		*addr_len = sizeof(struct sockaddr_in6);
512 	}
513 	memset((char *)saddr + *addr_len, 0, sizeof(struct sockaddr_storage) - *addr_len);
514 }
515 
516 /* Close a remote connection and tidy up */
517 static void close_connection(struct connection *con, bool and_other)
518 {
519 	mutex_lock(&con->sock_mutex);
520 
521 	if (con->sock) {
522 		sock_release(con->sock);
523 		con->sock = NULL;
524 	}
525 	if (con->othercon && and_other) {
526 		/* Will only re-enter once. */
527 		close_connection(con->othercon, false);
528 	}
529 	if (con->rx_page) {
530 		__free_page(con->rx_page);
531 		con->rx_page = NULL;
532 	}
533 
534 	con->retries = 0;
535 	mutex_unlock(&con->sock_mutex);
536 }
537 
538 /* We only send shutdown messages to nodes that are not part of the cluster */
539 static void sctp_send_shutdown(sctp_assoc_t associd)
540 {
541 	static char outcmsg[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];
542 	struct msghdr outmessage;
543 	struct cmsghdr *cmsg;
544 	struct sctp_sndrcvinfo *sinfo;
545 	int ret;
546 	struct connection *con;
547 
548 	con = nodeid2con(0,0);
549 	BUG_ON(con == NULL);
550 
551 	outmessage.msg_name = NULL;
552 	outmessage.msg_namelen = 0;
553 	outmessage.msg_control = outcmsg;
554 	outmessage.msg_controllen = sizeof(outcmsg);
555 	outmessage.msg_flags = MSG_EOR;
556 
557 	cmsg = CMSG_FIRSTHDR(&outmessage);
558 	cmsg->cmsg_level = IPPROTO_SCTP;
559 	cmsg->cmsg_type = SCTP_SNDRCV;
560 	cmsg->cmsg_len = CMSG_LEN(sizeof(struct sctp_sndrcvinfo));
561 	outmessage.msg_controllen = cmsg->cmsg_len;
562 	sinfo = CMSG_DATA(cmsg);
563 	memset(sinfo, 0x00, sizeof(struct sctp_sndrcvinfo));
564 
565 	sinfo->sinfo_flags |= MSG_EOF;
566 	sinfo->sinfo_assoc_id = associd;
567 
568 	ret = kernel_sendmsg(con->sock, &outmessage, NULL, 0, 0);
569 
570 	if (ret != 0)
571 		log_print("send EOF to node failed: %d", ret);
572 }
573 
574 static void sctp_init_failed_foreach(struct connection *con)
575 {
576 
577 	/*
578 	 * Don't try to recover base con and handle race where the
579 	 * other node's assoc init creates a assoc and we get that
580 	 * notification, then we get a notification that our attempt
581 	 * failed due. This happens when we are still trying the primary
582 	 * address, but the other node has already tried secondary addrs
583 	 * and found one that worked.
584 	 */
585 	if (!con->nodeid || con->sctp_assoc)
586 		return;
587 
588 	log_print("Retrying SCTP association init for node %d\n", con->nodeid);
589 
590 	con->try_new_addr = true;
591 	con->sctp_assoc = 0;
592 	if (test_and_clear_bit(CF_INIT_PENDING, &con->flags)) {
593 		if (!test_and_set_bit(CF_WRITE_PENDING, &con->flags))
594 			queue_work(send_workqueue, &con->swork);
595 	}
596 }
597 
598 /* INIT failed but we don't know which node...
599    restart INIT on all pending nodes */
600 static void sctp_init_failed(void)
601 {
602 	mutex_lock(&connections_lock);
603 
604 	foreach_conn(sctp_init_failed_foreach);
605 
606 	mutex_unlock(&connections_lock);
607 }
608 
609 static void retry_failed_sctp_send(struct connection *recv_con,
610 				   struct sctp_send_failed *sn_send_failed,
611 				   char *buf)
612 {
613 	int len = sn_send_failed->ssf_length - sizeof(struct sctp_send_failed);
614 	struct dlm_mhandle *mh;
615 	struct connection *con;
616 	char *retry_buf;
617 	int nodeid = sn_send_failed->ssf_info.sinfo_ppid;
618 
619 	log_print("Retry sending %d bytes to node id %d", len, nodeid);
620 
621 	con = nodeid2con(nodeid, 0);
622 	if (!con) {
623 		log_print("Could not look up con for nodeid %d\n",
624 			  nodeid);
625 		return;
626 	}
627 
628 	mh = dlm_lowcomms_get_buffer(nodeid, len, GFP_NOFS, &retry_buf);
629 	if (!mh) {
630 		log_print("Could not allocate buf for retry.");
631 		return;
632 	}
633 	memcpy(retry_buf, buf + sizeof(struct sctp_send_failed), len);
634 	dlm_lowcomms_commit_buffer(mh);
635 
636 	/*
637 	 * If we got a assoc changed event before the send failed event then
638 	 * we only need to retry the send.
639 	 */
640 	if (con->sctp_assoc) {
641 		if (!test_and_set_bit(CF_WRITE_PENDING, &con->flags))
642 			queue_work(send_workqueue, &con->swork);
643 	} else
644 		sctp_init_failed_foreach(con);
645 }
646 
647 /* Something happened to an association */
648 static void process_sctp_notification(struct connection *con,
649 				      struct msghdr *msg, char *buf)
650 {
651 	union sctp_notification *sn = (union sctp_notification *)buf;
652 	struct linger linger;
653 
654 	switch (sn->sn_header.sn_type) {
655 	case SCTP_SEND_FAILED:
656 		retry_failed_sctp_send(con, &sn->sn_send_failed, buf);
657 		break;
658 	case SCTP_ASSOC_CHANGE:
659 		switch (sn->sn_assoc_change.sac_state) {
660 		case SCTP_COMM_UP:
661 		case SCTP_RESTART:
662 		{
663 			/* Check that the new node is in the lockspace */
664 			struct sctp_prim prim;
665 			int nodeid;
666 			int prim_len, ret;
667 			int addr_len;
668 			struct connection *new_con;
669 
670 			/*
671 			 * We get this before any data for an association.
672 			 * We verify that the node is in the cluster and
673 			 * then peel off a socket for it.
674 			 */
675 			if ((int)sn->sn_assoc_change.sac_assoc_id <= 0) {
676 				log_print("COMM_UP for invalid assoc ID %d",
677 					 (int)sn->sn_assoc_change.sac_assoc_id);
678 				sctp_init_failed();
679 				return;
680 			}
681 			memset(&prim, 0, sizeof(struct sctp_prim));
682 			prim_len = sizeof(struct sctp_prim);
683 			prim.ssp_assoc_id = sn->sn_assoc_change.sac_assoc_id;
684 
685 			ret = kernel_getsockopt(con->sock,
686 						IPPROTO_SCTP,
687 						SCTP_PRIMARY_ADDR,
688 						(char*)&prim,
689 						&prim_len);
690 			if (ret < 0) {
691 				log_print("getsockopt/sctp_primary_addr on "
692 					  "new assoc %d failed : %d",
693 					  (int)sn->sn_assoc_change.sac_assoc_id,
694 					  ret);
695 
696 				/* Retry INIT later */
697 				new_con = assoc2con(sn->sn_assoc_change.sac_assoc_id);
698 				if (new_con)
699 					clear_bit(CF_CONNECT_PENDING, &con->flags);
700 				return;
701 			}
702 			make_sockaddr(&prim.ssp_addr, 0, &addr_len);
703 			if (addr_to_nodeid(&prim.ssp_addr, &nodeid)) {
704 				unsigned char *b=(unsigned char *)&prim.ssp_addr;
705 				log_print("reject connect from unknown addr");
706 				print_hex_dump_bytes("ss: ", DUMP_PREFIX_NONE,
707 						     b, sizeof(struct sockaddr_storage));
708 				sctp_send_shutdown(prim.ssp_assoc_id);
709 				return;
710 			}
711 
712 			new_con = nodeid2con(nodeid, GFP_NOFS);
713 			if (!new_con)
714 				return;
715 
716 			/* Peel off a new sock */
717 			lock_sock(con->sock->sk);
718 			ret = sctp_do_peeloff(con->sock->sk,
719 				sn->sn_assoc_change.sac_assoc_id,
720 				&new_con->sock);
721 			release_sock(con->sock->sk);
722 			if (ret < 0) {
723 				log_print("Can't peel off a socket for "
724 					  "connection %d to node %d: err=%d",
725 					  (int)sn->sn_assoc_change.sac_assoc_id,
726 					  nodeid, ret);
727 				return;
728 			}
729 			add_sock(new_con->sock, new_con);
730 
731 			linger.l_onoff = 1;
732 			linger.l_linger = 0;
733 			ret = kernel_setsockopt(new_con->sock, SOL_SOCKET, SO_LINGER,
734 						(char *)&linger, sizeof(linger));
735 			if (ret < 0)
736 				log_print("set socket option SO_LINGER failed");
737 
738 			log_print("connecting to %d sctp association %d",
739 				 nodeid, (int)sn->sn_assoc_change.sac_assoc_id);
740 
741 			new_con->sctp_assoc = sn->sn_assoc_change.sac_assoc_id;
742 			new_con->try_new_addr = false;
743 			/* Send any pending writes */
744 			clear_bit(CF_CONNECT_PENDING, &new_con->flags);
745 			clear_bit(CF_INIT_PENDING, &new_con->flags);
746 			if (!test_and_set_bit(CF_WRITE_PENDING, &new_con->flags)) {
747 				queue_work(send_workqueue, &new_con->swork);
748 			}
749 			if (!test_and_set_bit(CF_READ_PENDING, &new_con->flags))
750 				queue_work(recv_workqueue, &new_con->rwork);
751 		}
752 		break;
753 
754 		case SCTP_COMM_LOST:
755 		case SCTP_SHUTDOWN_COMP:
756 		{
757 			con = assoc2con(sn->sn_assoc_change.sac_assoc_id);
758 			if (con) {
759 				con->sctp_assoc = 0;
760 			}
761 		}
762 		break;
763 
764 		case SCTP_CANT_STR_ASSOC:
765 		{
766 			/* Will retry init when we get the send failed notification */
767 			log_print("Can't start SCTP association - retrying");
768 		}
769 		break;
770 
771 		default:
772 			log_print("unexpected SCTP assoc change id=%d state=%d",
773 				  (int)sn->sn_assoc_change.sac_assoc_id,
774 				  sn->sn_assoc_change.sac_state);
775 		}
776 	default:
777 		; /* fall through */
778 	}
779 }
780 
781 /* Data received from remote end */
782 static int receive_from_sock(struct connection *con)
783 {
784 	int ret = 0;
785 	struct msghdr msg = {};
786 	struct kvec iov[2];
787 	unsigned len;
788 	int r;
789 	int call_again_soon = 0;
790 	int nvec;
791 	char incmsg[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];
792 
793 	mutex_lock(&con->sock_mutex);
794 
795 	if (con->sock == NULL) {
796 		ret = -EAGAIN;
797 		goto out_close;
798 	}
799 
800 	if (con->rx_page == NULL) {
801 		/*
802 		 * This doesn't need to be atomic, but I think it should
803 		 * improve performance if it is.
804 		 */
805 		con->rx_page = alloc_page(GFP_ATOMIC);
806 		if (con->rx_page == NULL)
807 			goto out_resched;
808 		cbuf_init(&con->cb, PAGE_CACHE_SIZE);
809 	}
810 
811 	/* Only SCTP needs these really */
812 	memset(&incmsg, 0, sizeof(incmsg));
813 	msg.msg_control = incmsg;
814 	msg.msg_controllen = sizeof(incmsg);
815 
816 	/*
817 	 * iov[0] is the bit of the circular buffer between the current end
818 	 * point (cb.base + cb.len) and the end of the buffer.
819 	 */
820 	iov[0].iov_len = con->cb.base - cbuf_data(&con->cb);
821 	iov[0].iov_base = page_address(con->rx_page) + cbuf_data(&con->cb);
822 	iov[1].iov_len = 0;
823 	nvec = 1;
824 
825 	/*
826 	 * iov[1] is the bit of the circular buffer between the start of the
827 	 * buffer and the start of the currently used section (cb.base)
828 	 */
829 	if (cbuf_data(&con->cb) >= con->cb.base) {
830 		iov[0].iov_len = PAGE_CACHE_SIZE - cbuf_data(&con->cb);
831 		iov[1].iov_len = con->cb.base;
832 		iov[1].iov_base = page_address(con->rx_page);
833 		nvec = 2;
834 	}
835 	len = iov[0].iov_len + iov[1].iov_len;
836 
837 	r = ret = kernel_recvmsg(con->sock, &msg, iov, nvec, len,
838 			       MSG_DONTWAIT | MSG_NOSIGNAL);
839 	if (ret <= 0)
840 		goto out_close;
841 
842 	/* Process SCTP notifications */
843 	if (msg.msg_flags & MSG_NOTIFICATION) {
844 		msg.msg_control = incmsg;
845 		msg.msg_controllen = sizeof(incmsg);
846 
847 		process_sctp_notification(con, &msg,
848 				page_address(con->rx_page) + con->cb.base);
849 		mutex_unlock(&con->sock_mutex);
850 		return 0;
851 	}
852 	BUG_ON(con->nodeid == 0);
853 
854 	if (ret == len)
855 		call_again_soon = 1;
856 	cbuf_add(&con->cb, ret);
857 	ret = dlm_process_incoming_buffer(con->nodeid,
858 					  page_address(con->rx_page),
859 					  con->cb.base, con->cb.len,
860 					  PAGE_CACHE_SIZE);
861 	if (ret == -EBADMSG) {
862 		log_print("lowcomms: addr=%p, base=%u, len=%u, "
863 			  "iov_len=%u, iov_base[0]=%p, read=%d",
864 			  page_address(con->rx_page), con->cb.base, con->cb.len,
865 			  len, iov[0].iov_base, r);
866 	}
867 	if (ret < 0)
868 		goto out_close;
869 	cbuf_eat(&con->cb, ret);
870 
871 	if (cbuf_empty(&con->cb) && !call_again_soon) {
872 		__free_page(con->rx_page);
873 		con->rx_page = NULL;
874 	}
875 
876 	if (call_again_soon)
877 		goto out_resched;
878 	mutex_unlock(&con->sock_mutex);
879 	return 0;
880 
881 out_resched:
882 	if (!test_and_set_bit(CF_READ_PENDING, &con->flags))
883 		queue_work(recv_workqueue, &con->rwork);
884 	mutex_unlock(&con->sock_mutex);
885 	return -EAGAIN;
886 
887 out_close:
888 	mutex_unlock(&con->sock_mutex);
889 	if (ret != -EAGAIN) {
890 		close_connection(con, false);
891 		/* Reconnect when there is something to send */
892 	}
893 	/* Don't return success if we really got EOF */
894 	if (ret == 0)
895 		ret = -EAGAIN;
896 
897 	return ret;
898 }
899 
900 /* Listening socket is busy, accept a connection */
901 static int tcp_accept_from_sock(struct connection *con)
902 {
903 	int result;
904 	struct sockaddr_storage peeraddr;
905 	struct socket *newsock;
906 	int len;
907 	int nodeid;
908 	struct connection *newcon;
909 	struct connection *addcon;
910 
911 	mutex_lock(&connections_lock);
912 	if (!dlm_allow_conn) {
913 		mutex_unlock(&connections_lock);
914 		return -1;
915 	}
916 	mutex_unlock(&connections_lock);
917 
918 	memset(&peeraddr, 0, sizeof(peeraddr));
919 	result = sock_create_kern(dlm_local_addr[0]->ss_family, SOCK_STREAM,
920 				  IPPROTO_TCP, &newsock);
921 	if (result < 0)
922 		return -ENOMEM;
923 
924 	mutex_lock_nested(&con->sock_mutex, 0);
925 
926 	result = -ENOTCONN;
927 	if (con->sock == NULL)
928 		goto accept_err;
929 
930 	newsock->type = con->sock->type;
931 	newsock->ops = con->sock->ops;
932 
933 	result = con->sock->ops->accept(con->sock, newsock, O_NONBLOCK);
934 	if (result < 0)
935 		goto accept_err;
936 
937 	/* Get the connected socket's peer */
938 	memset(&peeraddr, 0, sizeof(peeraddr));
939 	if (newsock->ops->getname(newsock, (struct sockaddr *)&peeraddr,
940 				  &len, 2)) {
941 		result = -ECONNABORTED;
942 		goto accept_err;
943 	}
944 
945 	/* Get the new node's NODEID */
946 	make_sockaddr(&peeraddr, 0, &len);
947 	if (addr_to_nodeid(&peeraddr, &nodeid)) {
948 		unsigned char *b=(unsigned char *)&peeraddr;
949 		log_print("connect from non cluster node");
950 		print_hex_dump_bytes("ss: ", DUMP_PREFIX_NONE,
951 				     b, sizeof(struct sockaddr_storage));
952 		sock_release(newsock);
953 		mutex_unlock(&con->sock_mutex);
954 		return -1;
955 	}
956 
957 	log_print("got connection from %d", nodeid);
958 
959 	/*  Check to see if we already have a connection to this node. This
960 	 *  could happen if the two nodes initiate a connection at roughly
961 	 *  the same time and the connections cross on the wire.
962 	 *  In this case we store the incoming one in "othercon"
963 	 */
964 	newcon = nodeid2con(nodeid, GFP_NOFS);
965 	if (!newcon) {
966 		result = -ENOMEM;
967 		goto accept_err;
968 	}
969 	mutex_lock_nested(&newcon->sock_mutex, 1);
970 	if (newcon->sock) {
971 		struct connection *othercon = newcon->othercon;
972 
973 		if (!othercon) {
974 			othercon = kmem_cache_zalloc(con_cache, GFP_NOFS);
975 			if (!othercon) {
976 				log_print("failed to allocate incoming socket");
977 				mutex_unlock(&newcon->sock_mutex);
978 				result = -ENOMEM;
979 				goto accept_err;
980 			}
981 			othercon->nodeid = nodeid;
982 			othercon->rx_action = receive_from_sock;
983 			mutex_init(&othercon->sock_mutex);
984 			INIT_WORK(&othercon->swork, process_send_sockets);
985 			INIT_WORK(&othercon->rwork, process_recv_sockets);
986 			set_bit(CF_IS_OTHERCON, &othercon->flags);
987 		}
988 		if (!othercon->sock) {
989 			newcon->othercon = othercon;
990 			othercon->sock = newsock;
991 			newsock->sk->sk_user_data = othercon;
992 			add_sock(newsock, othercon);
993 			addcon = othercon;
994 		}
995 		else {
996 			printk("Extra connection from node %d attempted\n", nodeid);
997 			result = -EAGAIN;
998 			mutex_unlock(&newcon->sock_mutex);
999 			goto accept_err;
1000 		}
1001 	}
1002 	else {
1003 		newsock->sk->sk_user_data = newcon;
1004 		newcon->rx_action = receive_from_sock;
1005 		add_sock(newsock, newcon);
1006 		addcon = newcon;
1007 	}
1008 
1009 	mutex_unlock(&newcon->sock_mutex);
1010 
1011 	/*
1012 	 * Add it to the active queue in case we got data
1013 	 * between processing the accept adding the socket
1014 	 * to the read_sockets list
1015 	 */
1016 	if (!test_and_set_bit(CF_READ_PENDING, &addcon->flags))
1017 		queue_work(recv_workqueue, &addcon->rwork);
1018 	mutex_unlock(&con->sock_mutex);
1019 
1020 	return 0;
1021 
1022 accept_err:
1023 	mutex_unlock(&con->sock_mutex);
1024 	sock_release(newsock);
1025 
1026 	if (result != -EAGAIN)
1027 		log_print("error accepting connection from node: %d", result);
1028 	return result;
1029 }
1030 
1031 static void free_entry(struct writequeue_entry *e)
1032 {
1033 	__free_page(e->page);
1034 	kfree(e);
1035 }
1036 
1037 /*
1038  * writequeue_entry_complete - try to delete and free write queue entry
1039  * @e: write queue entry to try to delete
1040  * @completed: bytes completed
1041  *
1042  * writequeue_lock must be held.
1043  */
1044 static void writequeue_entry_complete(struct writequeue_entry *e, int completed)
1045 {
1046 	e->offset += completed;
1047 	e->len -= completed;
1048 
1049 	if (e->len == 0 && e->users == 0) {
1050 		list_del(&e->list);
1051 		free_entry(e);
1052 	}
1053 }
1054 
1055 /* Initiate an SCTP association.
1056    This is a special case of send_to_sock() in that we don't yet have a
1057    peeled-off socket for this association, so we use the listening socket
1058    and add the primary IP address of the remote node.
1059  */
1060 static void sctp_init_assoc(struct connection *con)
1061 {
1062 	struct sockaddr_storage rem_addr;
1063 	char outcmsg[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];
1064 	struct msghdr outmessage;
1065 	struct cmsghdr *cmsg;
1066 	struct sctp_sndrcvinfo *sinfo;
1067 	struct connection *base_con;
1068 	struct writequeue_entry *e;
1069 	int len, offset;
1070 	int ret;
1071 	int addrlen;
1072 	struct kvec iov[1];
1073 
1074 	mutex_lock(&con->sock_mutex);
1075 	if (test_and_set_bit(CF_INIT_PENDING, &con->flags))
1076 		goto unlock;
1077 
1078 	if (nodeid_to_addr(con->nodeid, NULL, (struct sockaddr *)&rem_addr,
1079 			   con->try_new_addr)) {
1080 		log_print("no address for nodeid %d", con->nodeid);
1081 		goto unlock;
1082 	}
1083 	base_con = nodeid2con(0, 0);
1084 	BUG_ON(base_con == NULL);
1085 
1086 	make_sockaddr(&rem_addr, dlm_config.ci_tcp_port, &addrlen);
1087 
1088 	outmessage.msg_name = &rem_addr;
1089 	outmessage.msg_namelen = addrlen;
1090 	outmessage.msg_control = outcmsg;
1091 	outmessage.msg_controllen = sizeof(outcmsg);
1092 	outmessage.msg_flags = MSG_EOR;
1093 
1094 	spin_lock(&con->writequeue_lock);
1095 
1096 	if (list_empty(&con->writequeue)) {
1097 		spin_unlock(&con->writequeue_lock);
1098 		log_print("writequeue empty for nodeid %d", con->nodeid);
1099 		goto unlock;
1100 	}
1101 
1102 	e = list_first_entry(&con->writequeue, struct writequeue_entry, list);
1103 	len = e->len;
1104 	offset = e->offset;
1105 
1106 	/* Send the first block off the write queue */
1107 	iov[0].iov_base = page_address(e->page)+offset;
1108 	iov[0].iov_len = len;
1109 	spin_unlock(&con->writequeue_lock);
1110 
1111 	if (rem_addr.ss_family == AF_INET) {
1112 		struct sockaddr_in *sin = (struct sockaddr_in *)&rem_addr;
1113 		log_print("Trying to connect to %pI4", &sin->sin_addr.s_addr);
1114 	} else {
1115 		struct sockaddr_in6 *sin6 = (struct sockaddr_in6 *)&rem_addr;
1116 		log_print("Trying to connect to %pI6", &sin6->sin6_addr);
1117 	}
1118 
1119 	cmsg = CMSG_FIRSTHDR(&outmessage);
1120 	cmsg->cmsg_level = IPPROTO_SCTP;
1121 	cmsg->cmsg_type = SCTP_SNDRCV;
1122 	cmsg->cmsg_len = CMSG_LEN(sizeof(struct sctp_sndrcvinfo));
1123 	sinfo = CMSG_DATA(cmsg);
1124 	memset(sinfo, 0x00, sizeof(struct sctp_sndrcvinfo));
1125 	sinfo->sinfo_ppid = cpu_to_le32(con->nodeid);
1126 	outmessage.msg_controllen = cmsg->cmsg_len;
1127 	sinfo->sinfo_flags |= SCTP_ADDR_OVER;
1128 
1129 	ret = kernel_sendmsg(base_con->sock, &outmessage, iov, 1, len);
1130 	if (ret < 0) {
1131 		log_print("Send first packet to node %d failed: %d",
1132 			  con->nodeid, ret);
1133 
1134 		/* Try again later */
1135 		clear_bit(CF_CONNECT_PENDING, &con->flags);
1136 		clear_bit(CF_INIT_PENDING, &con->flags);
1137 	}
1138 	else {
1139 		spin_lock(&con->writequeue_lock);
1140 		writequeue_entry_complete(e, ret);
1141 		spin_unlock(&con->writequeue_lock);
1142 	}
1143 
1144 unlock:
1145 	mutex_unlock(&con->sock_mutex);
1146 }
1147 
1148 /* Connect a new socket to its peer */
1149 static void tcp_connect_to_sock(struct connection *con)
1150 {
1151 	struct sockaddr_storage saddr, src_addr;
1152 	int addr_len;
1153 	struct socket *sock = NULL;
1154 	int one = 1;
1155 	int result;
1156 
1157 	if (con->nodeid == 0) {
1158 		log_print("attempt to connect sock 0 foiled");
1159 		return;
1160 	}
1161 
1162 	mutex_lock(&con->sock_mutex);
1163 	if (con->retries++ > MAX_CONNECT_RETRIES)
1164 		goto out;
1165 
1166 	/* Some odd races can cause double-connects, ignore them */
1167 	if (con->sock)
1168 		goto out;
1169 
1170 	/* Create a socket to communicate with */
1171 	result = sock_create_kern(dlm_local_addr[0]->ss_family, SOCK_STREAM,
1172 				  IPPROTO_TCP, &sock);
1173 	if (result < 0)
1174 		goto out_err;
1175 
1176 	memset(&saddr, 0, sizeof(saddr));
1177 	result = nodeid_to_addr(con->nodeid, &saddr, NULL, false);
1178 	if (result < 0) {
1179 		log_print("no address for nodeid %d", con->nodeid);
1180 		goto out_err;
1181 	}
1182 
1183 	sock->sk->sk_user_data = con;
1184 	con->rx_action = receive_from_sock;
1185 	con->connect_action = tcp_connect_to_sock;
1186 	add_sock(sock, con);
1187 
1188 	/* Bind to our cluster-known address connecting to avoid
1189 	   routing problems */
1190 	memcpy(&src_addr, dlm_local_addr[0], sizeof(src_addr));
1191 	make_sockaddr(&src_addr, 0, &addr_len);
1192 	result = sock->ops->bind(sock, (struct sockaddr *) &src_addr,
1193 				 addr_len);
1194 	if (result < 0) {
1195 		log_print("could not bind for connect: %d", result);
1196 		/* This *may* not indicate a critical error */
1197 	}
1198 
1199 	make_sockaddr(&saddr, dlm_config.ci_tcp_port, &addr_len);
1200 
1201 	log_print("connecting to %d", con->nodeid);
1202 
1203 	/* Turn off Nagle's algorithm */
1204 	kernel_setsockopt(sock, SOL_TCP, TCP_NODELAY, (char *)&one,
1205 			  sizeof(one));
1206 
1207 	result = sock->ops->connect(sock, (struct sockaddr *)&saddr, addr_len,
1208 				   O_NONBLOCK);
1209 	if (result == -EINPROGRESS)
1210 		result = 0;
1211 	if (result == 0)
1212 		goto out;
1213 
1214 out_err:
1215 	if (con->sock) {
1216 		sock_release(con->sock);
1217 		con->sock = NULL;
1218 	} else if (sock) {
1219 		sock_release(sock);
1220 	}
1221 	/*
1222 	 * Some errors are fatal and this list might need adjusting. For other
1223 	 * errors we try again until the max number of retries is reached.
1224 	 */
1225 	if (result != -EHOSTUNREACH &&
1226 	    result != -ENETUNREACH &&
1227 	    result != -ENETDOWN &&
1228 	    result != -EINVAL &&
1229 	    result != -EPROTONOSUPPORT) {
1230 		log_print("connect %d try %d error %d", con->nodeid,
1231 			  con->retries, result);
1232 		mutex_unlock(&con->sock_mutex);
1233 		msleep(1000);
1234 		lowcomms_connect_sock(con);
1235 		return;
1236 	}
1237 out:
1238 	mutex_unlock(&con->sock_mutex);
1239 	return;
1240 }
1241 
1242 static struct socket *tcp_create_listen_sock(struct connection *con,
1243 					     struct sockaddr_storage *saddr)
1244 {
1245 	struct socket *sock = NULL;
1246 	int result = 0;
1247 	int one = 1;
1248 	int addr_len;
1249 
1250 	if (dlm_local_addr[0]->ss_family == AF_INET)
1251 		addr_len = sizeof(struct sockaddr_in);
1252 	else
1253 		addr_len = sizeof(struct sockaddr_in6);
1254 
1255 	/* Create a socket to communicate with */
1256 	result = sock_create_kern(dlm_local_addr[0]->ss_family, SOCK_STREAM,
1257 				  IPPROTO_TCP, &sock);
1258 	if (result < 0) {
1259 		log_print("Can't create listening comms socket");
1260 		goto create_out;
1261 	}
1262 
1263 	/* Turn off Nagle's algorithm */
1264 	kernel_setsockopt(sock, SOL_TCP, TCP_NODELAY, (char *)&one,
1265 			  sizeof(one));
1266 
1267 	result = kernel_setsockopt(sock, SOL_SOCKET, SO_REUSEADDR,
1268 				   (char *)&one, sizeof(one));
1269 
1270 	if (result < 0) {
1271 		log_print("Failed to set SO_REUSEADDR on socket: %d", result);
1272 	}
1273 	con->rx_action = tcp_accept_from_sock;
1274 	con->connect_action = tcp_connect_to_sock;
1275 
1276 	/* Bind to our port */
1277 	make_sockaddr(saddr, dlm_config.ci_tcp_port, &addr_len);
1278 	result = sock->ops->bind(sock, (struct sockaddr *) saddr, addr_len);
1279 	if (result < 0) {
1280 		log_print("Can't bind to port %d", dlm_config.ci_tcp_port);
1281 		sock_release(sock);
1282 		sock = NULL;
1283 		con->sock = NULL;
1284 		goto create_out;
1285 	}
1286 	result = kernel_setsockopt(sock, SOL_SOCKET, SO_KEEPALIVE,
1287 				 (char *)&one, sizeof(one));
1288 	if (result < 0) {
1289 		log_print("Set keepalive failed: %d", result);
1290 	}
1291 
1292 	result = sock->ops->listen(sock, 5);
1293 	if (result < 0) {
1294 		log_print("Can't listen on port %d", dlm_config.ci_tcp_port);
1295 		sock_release(sock);
1296 		sock = NULL;
1297 		goto create_out;
1298 	}
1299 
1300 create_out:
1301 	return sock;
1302 }
1303 
1304 /* Get local addresses */
1305 static void init_local(void)
1306 {
1307 	struct sockaddr_storage sas, *addr;
1308 	int i;
1309 
1310 	dlm_local_count = 0;
1311 	for (i = 0; i < DLM_MAX_ADDR_COUNT; i++) {
1312 		if (dlm_our_addr(&sas, i))
1313 			break;
1314 
1315 		addr = kmalloc(sizeof(*addr), GFP_NOFS);
1316 		if (!addr)
1317 			break;
1318 		memcpy(addr, &sas, sizeof(*addr));
1319 		dlm_local_addr[dlm_local_count++] = addr;
1320 	}
1321 }
1322 
1323 /* Bind to an IP address. SCTP allows multiple address so it can do
1324    multi-homing */
1325 static int add_sctp_bind_addr(struct connection *sctp_con,
1326 			      struct sockaddr_storage *addr,
1327 			      int addr_len, int num)
1328 {
1329 	int result = 0;
1330 
1331 	if (num == 1)
1332 		result = kernel_bind(sctp_con->sock,
1333 				     (struct sockaddr *) addr,
1334 				     addr_len);
1335 	else
1336 		result = kernel_setsockopt(sctp_con->sock, SOL_SCTP,
1337 					   SCTP_SOCKOPT_BINDX_ADD,
1338 					   (char *)addr, addr_len);
1339 
1340 	if (result < 0)
1341 		log_print("Can't bind to port %d addr number %d",
1342 			  dlm_config.ci_tcp_port, num);
1343 
1344 	return result;
1345 }
1346 
1347 /* Initialise SCTP socket and bind to all interfaces */
1348 static int sctp_listen_for_all(void)
1349 {
1350 	struct socket *sock = NULL;
1351 	struct sockaddr_storage localaddr;
1352 	struct sctp_event_subscribe subscribe;
1353 	int result = -EINVAL, num = 1, i, addr_len;
1354 	struct connection *con = nodeid2con(0, GFP_NOFS);
1355 	int bufsize = NEEDED_RMEM;
1356 	int one = 1;
1357 
1358 	if (!con)
1359 		return -ENOMEM;
1360 
1361 	log_print("Using SCTP for communications");
1362 
1363 	result = sock_create_kern(dlm_local_addr[0]->ss_family, SOCK_SEQPACKET,
1364 				  IPPROTO_SCTP, &sock);
1365 	if (result < 0) {
1366 		log_print("Can't create comms socket, check SCTP is loaded");
1367 		goto out;
1368 	}
1369 
1370 	/* Listen for events */
1371 	memset(&subscribe, 0, sizeof(subscribe));
1372 	subscribe.sctp_data_io_event = 1;
1373 	subscribe.sctp_association_event = 1;
1374 	subscribe.sctp_send_failure_event = 1;
1375 	subscribe.sctp_shutdown_event = 1;
1376 	subscribe.sctp_partial_delivery_event = 1;
1377 
1378 	result = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUFFORCE,
1379 				 (char *)&bufsize, sizeof(bufsize));
1380 	if (result)
1381 		log_print("Error increasing buffer space on socket %d", result);
1382 
1383 	result = kernel_setsockopt(sock, SOL_SCTP, SCTP_EVENTS,
1384 				   (char *)&subscribe, sizeof(subscribe));
1385 	if (result < 0) {
1386 		log_print("Failed to set SCTP_EVENTS on socket: result=%d",
1387 			  result);
1388 		goto create_delsock;
1389 	}
1390 
1391 	result = kernel_setsockopt(sock, SOL_SCTP, SCTP_NODELAY, (char *)&one,
1392 				   sizeof(one));
1393 	if (result < 0)
1394 		log_print("Could not set SCTP NODELAY error %d\n", result);
1395 
1396 	/* Init con struct */
1397 	sock->sk->sk_user_data = con;
1398 	con->sock = sock;
1399 	con->sock->sk->sk_data_ready = lowcomms_data_ready;
1400 	con->rx_action = receive_from_sock;
1401 	con->connect_action = sctp_init_assoc;
1402 
1403 	/* Bind to all interfaces. */
1404 	for (i = 0; i < dlm_local_count; i++) {
1405 		memcpy(&localaddr, dlm_local_addr[i], sizeof(localaddr));
1406 		make_sockaddr(&localaddr, dlm_config.ci_tcp_port, &addr_len);
1407 
1408 		result = add_sctp_bind_addr(con, &localaddr, addr_len, num);
1409 		if (result)
1410 			goto create_delsock;
1411 		++num;
1412 	}
1413 
1414 	result = sock->ops->listen(sock, 5);
1415 	if (result < 0) {
1416 		log_print("Can't set socket listening");
1417 		goto create_delsock;
1418 	}
1419 
1420 	return 0;
1421 
1422 create_delsock:
1423 	sock_release(sock);
1424 	con->sock = NULL;
1425 out:
1426 	return result;
1427 }
1428 
1429 static int tcp_listen_for_all(void)
1430 {
1431 	struct socket *sock = NULL;
1432 	struct connection *con = nodeid2con(0, GFP_NOFS);
1433 	int result = -EINVAL;
1434 
1435 	if (!con)
1436 		return -ENOMEM;
1437 
1438 	/* We don't support multi-homed hosts */
1439 	if (dlm_local_addr[1] != NULL) {
1440 		log_print("TCP protocol can't handle multi-homed hosts, "
1441 			  "try SCTP");
1442 		return -EINVAL;
1443 	}
1444 
1445 	log_print("Using TCP for communications");
1446 
1447 	sock = tcp_create_listen_sock(con, dlm_local_addr[0]);
1448 	if (sock) {
1449 		add_sock(sock, con);
1450 		result = 0;
1451 	}
1452 	else {
1453 		result = -EADDRINUSE;
1454 	}
1455 
1456 	return result;
1457 }
1458 
1459 
1460 
1461 static struct writequeue_entry *new_writequeue_entry(struct connection *con,
1462 						     gfp_t allocation)
1463 {
1464 	struct writequeue_entry *entry;
1465 
1466 	entry = kmalloc(sizeof(struct writequeue_entry), allocation);
1467 	if (!entry)
1468 		return NULL;
1469 
1470 	entry->page = alloc_page(allocation);
1471 	if (!entry->page) {
1472 		kfree(entry);
1473 		return NULL;
1474 	}
1475 
1476 	entry->offset = 0;
1477 	entry->len = 0;
1478 	entry->end = 0;
1479 	entry->users = 0;
1480 	entry->con = con;
1481 
1482 	return entry;
1483 }
1484 
1485 void *dlm_lowcomms_get_buffer(int nodeid, int len, gfp_t allocation, char **ppc)
1486 {
1487 	struct connection *con;
1488 	struct writequeue_entry *e;
1489 	int offset = 0;
1490 
1491 	con = nodeid2con(nodeid, allocation);
1492 	if (!con)
1493 		return NULL;
1494 
1495 	spin_lock(&con->writequeue_lock);
1496 	e = list_entry(con->writequeue.prev, struct writequeue_entry, list);
1497 	if ((&e->list == &con->writequeue) ||
1498 	    (PAGE_CACHE_SIZE - e->end < len)) {
1499 		e = NULL;
1500 	} else {
1501 		offset = e->end;
1502 		e->end += len;
1503 		e->users++;
1504 	}
1505 	spin_unlock(&con->writequeue_lock);
1506 
1507 	if (e) {
1508 	got_one:
1509 		*ppc = page_address(e->page) + offset;
1510 		return e;
1511 	}
1512 
1513 	e = new_writequeue_entry(con, allocation);
1514 	if (e) {
1515 		spin_lock(&con->writequeue_lock);
1516 		offset = e->end;
1517 		e->end += len;
1518 		e->users++;
1519 		list_add_tail(&e->list, &con->writequeue);
1520 		spin_unlock(&con->writequeue_lock);
1521 		goto got_one;
1522 	}
1523 	return NULL;
1524 }
1525 
1526 void dlm_lowcomms_commit_buffer(void *mh)
1527 {
1528 	struct writequeue_entry *e = (struct writequeue_entry *)mh;
1529 	struct connection *con = e->con;
1530 	int users;
1531 
1532 	spin_lock(&con->writequeue_lock);
1533 	users = --e->users;
1534 	if (users)
1535 		goto out;
1536 	e->len = e->end - e->offset;
1537 	spin_unlock(&con->writequeue_lock);
1538 
1539 	if (!test_and_set_bit(CF_WRITE_PENDING, &con->flags)) {
1540 		queue_work(send_workqueue, &con->swork);
1541 	}
1542 	return;
1543 
1544 out:
1545 	spin_unlock(&con->writequeue_lock);
1546 	return;
1547 }
1548 
1549 /* Send a message */
1550 static void send_to_sock(struct connection *con)
1551 {
1552 	int ret = 0;
1553 	const int msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL;
1554 	struct writequeue_entry *e;
1555 	int len, offset;
1556 	int count = 0;
1557 
1558 	mutex_lock(&con->sock_mutex);
1559 	if (con->sock == NULL)
1560 		goto out_connect;
1561 
1562 	spin_lock(&con->writequeue_lock);
1563 	for (;;) {
1564 		e = list_entry(con->writequeue.next, struct writequeue_entry,
1565 			       list);
1566 		if ((struct list_head *) e == &con->writequeue)
1567 			break;
1568 
1569 		len = e->len;
1570 		offset = e->offset;
1571 		BUG_ON(len == 0 && e->users == 0);
1572 		spin_unlock(&con->writequeue_lock);
1573 
1574 		ret = 0;
1575 		if (len) {
1576 			ret = kernel_sendpage(con->sock, e->page, offset, len,
1577 					      msg_flags);
1578 			if (ret == -EAGAIN || ret == 0) {
1579 				if (ret == -EAGAIN &&
1580 				    test_bit(SOCK_ASYNC_NOSPACE, &con->sock->flags) &&
1581 				    !test_and_set_bit(CF_APP_LIMITED, &con->flags)) {
1582 					/* Notify TCP that we're limited by the
1583 					 * application window size.
1584 					 */
1585 					set_bit(SOCK_NOSPACE, &con->sock->flags);
1586 					con->sock->sk->sk_write_pending++;
1587 				}
1588 				cond_resched();
1589 				goto out;
1590 			} else if (ret < 0)
1591 				goto send_error;
1592 		}
1593 
1594 		/* Don't starve people filling buffers */
1595 		if (++count >= MAX_SEND_MSG_COUNT) {
1596 			cond_resched();
1597 			count = 0;
1598 		}
1599 
1600 		spin_lock(&con->writequeue_lock);
1601 		writequeue_entry_complete(e, ret);
1602 	}
1603 	spin_unlock(&con->writequeue_lock);
1604 out:
1605 	mutex_unlock(&con->sock_mutex);
1606 	return;
1607 
1608 send_error:
1609 	mutex_unlock(&con->sock_mutex);
1610 	close_connection(con, false);
1611 	lowcomms_connect_sock(con);
1612 	return;
1613 
1614 out_connect:
1615 	mutex_unlock(&con->sock_mutex);
1616 	if (!test_bit(CF_INIT_PENDING, &con->flags))
1617 		lowcomms_connect_sock(con);
1618 }
1619 
1620 static void clean_one_writequeue(struct connection *con)
1621 {
1622 	struct writequeue_entry *e, *safe;
1623 
1624 	spin_lock(&con->writequeue_lock);
1625 	list_for_each_entry_safe(e, safe, &con->writequeue, list) {
1626 		list_del(&e->list);
1627 		free_entry(e);
1628 	}
1629 	spin_unlock(&con->writequeue_lock);
1630 }
1631 
1632 /* Called from recovery when it knows that a node has
1633    left the cluster */
1634 int dlm_lowcomms_close(int nodeid)
1635 {
1636 	struct connection *con;
1637 	struct dlm_node_addr *na;
1638 
1639 	log_print("closing connection to node %d", nodeid);
1640 	con = nodeid2con(nodeid, 0);
1641 	if (con) {
1642 		clear_bit(CF_CONNECT_PENDING, &con->flags);
1643 		clear_bit(CF_WRITE_PENDING, &con->flags);
1644 		set_bit(CF_CLOSE, &con->flags);
1645 		if (cancel_work_sync(&con->swork))
1646 			log_print("canceled swork for node %d", nodeid);
1647 		if (cancel_work_sync(&con->rwork))
1648 			log_print("canceled rwork for node %d", nodeid);
1649 		clean_one_writequeue(con);
1650 		close_connection(con, true);
1651 	}
1652 
1653 	spin_lock(&dlm_node_addrs_spin);
1654 	na = find_node_addr(nodeid);
1655 	if (na) {
1656 		list_del(&na->list);
1657 		while (na->addr_count--)
1658 			kfree(na->addr[na->addr_count]);
1659 		kfree(na);
1660 	}
1661 	spin_unlock(&dlm_node_addrs_spin);
1662 
1663 	return 0;
1664 }
1665 
1666 /* Receive workqueue function */
1667 static void process_recv_sockets(struct work_struct *work)
1668 {
1669 	struct connection *con = container_of(work, struct connection, rwork);
1670 	int err;
1671 
1672 	clear_bit(CF_READ_PENDING, &con->flags);
1673 	do {
1674 		err = con->rx_action(con);
1675 	} while (!err);
1676 }
1677 
1678 /* Send workqueue function */
1679 static void process_send_sockets(struct work_struct *work)
1680 {
1681 	struct connection *con = container_of(work, struct connection, swork);
1682 
1683 	if (test_and_clear_bit(CF_CONNECT_PENDING, &con->flags)) {
1684 		con->connect_action(con);
1685 		set_bit(CF_WRITE_PENDING, &con->flags);
1686 	}
1687 	if (test_and_clear_bit(CF_WRITE_PENDING, &con->flags))
1688 		send_to_sock(con);
1689 }
1690 
1691 
1692 /* Discard all entries on the write queues */
1693 static void clean_writequeues(void)
1694 {
1695 	foreach_conn(clean_one_writequeue);
1696 }
1697 
1698 static void work_stop(void)
1699 {
1700 	destroy_workqueue(recv_workqueue);
1701 	destroy_workqueue(send_workqueue);
1702 }
1703 
1704 static int work_start(void)
1705 {
1706 	recv_workqueue = alloc_workqueue("dlm_recv",
1707 					 WQ_UNBOUND | WQ_MEM_RECLAIM, 1);
1708 	if (!recv_workqueue) {
1709 		log_print("can't start dlm_recv");
1710 		return -ENOMEM;
1711 	}
1712 
1713 	send_workqueue = alloc_workqueue("dlm_send",
1714 					 WQ_UNBOUND | WQ_MEM_RECLAIM, 1);
1715 	if (!send_workqueue) {
1716 		log_print("can't start dlm_send");
1717 		destroy_workqueue(recv_workqueue);
1718 		return -ENOMEM;
1719 	}
1720 
1721 	return 0;
1722 }
1723 
1724 static void stop_conn(struct connection *con)
1725 {
1726 	con->flags |= 0x0F;
1727 	if (con->sock && con->sock->sk)
1728 		con->sock->sk->sk_user_data = NULL;
1729 }
1730 
1731 static void free_conn(struct connection *con)
1732 {
1733 	close_connection(con, true);
1734 	if (con->othercon)
1735 		kmem_cache_free(con_cache, con->othercon);
1736 	hlist_del(&con->list);
1737 	kmem_cache_free(con_cache, con);
1738 }
1739 
1740 void dlm_lowcomms_stop(void)
1741 {
1742 	/* Set all the flags to prevent any
1743 	   socket activity.
1744 	*/
1745 	mutex_lock(&connections_lock);
1746 	dlm_allow_conn = 0;
1747 	foreach_conn(stop_conn);
1748 	mutex_unlock(&connections_lock);
1749 
1750 	work_stop();
1751 
1752 	mutex_lock(&connections_lock);
1753 	clean_writequeues();
1754 
1755 	foreach_conn(free_conn);
1756 
1757 	mutex_unlock(&connections_lock);
1758 	kmem_cache_destroy(con_cache);
1759 }
1760 
1761 int dlm_lowcomms_start(void)
1762 {
1763 	int error = -EINVAL;
1764 	struct connection *con;
1765 	int i;
1766 
1767 	for (i = 0; i < CONN_HASH_SIZE; i++)
1768 		INIT_HLIST_HEAD(&connection_hash[i]);
1769 
1770 	init_local();
1771 	if (!dlm_local_count) {
1772 		error = -ENOTCONN;
1773 		log_print("no local IP address has been set");
1774 		goto fail;
1775 	}
1776 
1777 	error = -ENOMEM;
1778 	con_cache = kmem_cache_create("dlm_conn", sizeof(struct connection),
1779 				      __alignof__(struct connection), 0,
1780 				      NULL);
1781 	if (!con_cache)
1782 		goto fail;
1783 
1784 	error = work_start();
1785 	if (error)
1786 		goto fail_destroy;
1787 
1788 	dlm_allow_conn = 1;
1789 
1790 	/* Start listening */
1791 	if (dlm_config.ci_protocol == 0)
1792 		error = tcp_listen_for_all();
1793 	else
1794 		error = sctp_listen_for_all();
1795 	if (error)
1796 		goto fail_unlisten;
1797 
1798 	return 0;
1799 
1800 fail_unlisten:
1801 	dlm_allow_conn = 0;
1802 	con = nodeid2con(0,0);
1803 	if (con) {
1804 		close_connection(con, false);
1805 		kmem_cache_free(con_cache, con);
1806 	}
1807 fail_destroy:
1808 	kmem_cache_destroy(con_cache);
1809 fail:
1810 	return error;
1811 }
1812 
1813 void dlm_lowcomms_exit(void)
1814 {
1815 	struct dlm_node_addr *na, *safe;
1816 
1817 	spin_lock(&dlm_node_addrs_spin);
1818 	list_for_each_entry_safe(na, safe, &dlm_node_addrs, list) {
1819 		list_del(&na->list);
1820 		while (na->addr_count--)
1821 			kfree(na->addr[na->addr_count]);
1822 		kfree(na);
1823 	}
1824 	spin_unlock(&dlm_node_addrs_spin);
1825 }
1826