summaryrefslogtreecommitdiff
path: root/fs/dlm/lowcomms-sctp.c
blob: fe158d7a92853074da36daa33dcbcbabaae00ec7 (plain)
1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
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
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257
258
259
260
261
262
263
264
265
266
267
268
269
270
271
272
273
274
275
276
277
278
279
280
281
282
283
284
285
286
287
288
289
290
291
292
293
294
295
296
297
298
299
300
301
302
303
304
305
306
307
308
309
310
311
312
313
314
315
316
317
318
319
320
321
322
323
324
325
326
327
328
329
330
331
332
333
334
335
336
337
338
339
340
341
342
343
344
345
346
347
348
349
350
351
352
353
354
355
356
357
358
359
360
361
362
363
364
365
366
367
368
369
370
371
372
373
374
375
376
377
378
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
417
418
419
420
421
422
423
424
425
426
427
428
429
430
431
432
433
434
435
436
437
438
439
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
466
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
503
504
505
506
507
508
509
510
511
512
513
514
515
516
517
518
519
520
521
522
523
524
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
566
567
568
569
570
571
572
573
574
575
576
577
578
579
580
581
582
583
584
585
586
587
588
589
590
591
592
593
594
595
596
597
598
599
600
601
602
603
604
605
606
607
608
609
610
611
612
613
614
615
616
617
618
619
620
621
622
623
624
625
626
627
628
629
630
631
632
633
634
635
636
637
638
639
640
641
642
643
644
645
646
647
648
649
650
651
652
653
654
655
656
657
658
659
660
661
662
663
664
665
666
667
668
669
670
671
672
673
674
675
676
677
678
679
680
681
682
683
684
685
686
687
688
689
690
691
692
693
694
695
696
697
698
699
700
701
702
703
704
705
706
707
708
709
710
711
712
713
714
715
716
717
718
719
720
721
722
723
724
725
726
727
728
729
730
731
732
733
734
735
736
737
738
739
740
741
742
743
744
745
746
747
748
749
750
751
752
753
754
755
756
757
758
759
760
761
762
763
764
765
766
767
768
769
770
771
772
773
774
775
776
777
778
779
780
781
782
783
784
785
786
787
788
789
790
791
792
793
794
795
796
797
798
799
800
801
802
803
804
805
806
807
808
809
810
811
812
813
814
815
816
817
818
819
820
821
822
823
824
825
826
827
828
829
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
879
880
881
882
883
884
885
886
887
888
889
890
891
892
893
894
895
896
897
898
899
900
901
902
903
904
905
906
907
908
909
910
911
912
913
914
915
916
917
918
919
920
921
922
923
924
925
926
927
928
929
930
931
932
933
934
935
936
937
938
939
940
941
942
943
944
945
946
947
948
949
950
951
952
953
954
955
956
957
958
959
960
961
962
963
964
965
966
967
968
969
970
971
972
973
974
975
976
977
978
979
980
981
982
983
984
985
986
987
988
989
990
991
992
993
994
995
996
997
998
999
1000
1001
1002
1003
1004
1005
1006
1007
1008
1009
1010
1011
1012
1013
1014
1015
1016
1017
1018
1019
1020
1021
1022
1023
1024
1025
1026
1027
1028
1029
1030
1031
1032
1033
1034
1035
1036
1037
1038
1039
1040
1041
1042
1043
1044
1045
1046
1047
1048
1049
1050
1051
1052
1053
1054
1055
1056
1057
1058
1059
1060
1061
1062
1063
1064
1065
1066
1067
1068
1069
1070
1071
1072
1073
1074
1075
1076
1077
1078
1079
1080
1081
1082
1083
1084
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
1116
1117
1118
1119
1120
1121
1122
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
1163
1164
1165
1166
1167
1168
1169
1170
1171
1172
1173
1174
1175
1176
1177
1178
1179
1180
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
1218
1219
1220
1221
1222
1223
1224
1225
1226
1227
/******************************************************************************
*******************************************************************************
**
**  Copyright (C) Sistina Software, Inc.  1997-2003  All rights reserved.
**  Copyright (C) 2004-2006 Red Hat, Inc.  All rights reserved.
**
**  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
 * be expanded for the cluster infrastructure then that is it's
 * 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.
 *
 * I don't see any problem with the recv thread executing the locking
 * code on behalf of remote processes as the locking code is
 * short, efficient and never (well, hardly ever) waits.
 *
 */

#include <asm/ioctls.h>
#include <net/sock.h>
#include <net/tcp.h>
#include <net/sctp/user.h>
#include <linux/pagemap.h>
#include <linux/socket.h>
#include <linux/idr.h>

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

static struct sockaddr_storage *dlm_local_addr[DLM_MAX_ADDR_COUNT];
static int			dlm_local_count;
static int			dlm_local_nodeid;

/* One of these per connected node */

#define NI_INIT_PENDING 1
#define NI_WRITE_PENDING 2

struct nodeinfo {
	spinlock_t		lock;
	sctp_assoc_t		assoc_id;
	unsigned long		flags;
	struct list_head	write_list; /* nodes with pending writes */
	struct list_head	writequeue; /* outgoing writequeue_entries */
	spinlock_t		writequeue_lock;
	int			nodeid;
};

static DEFINE_IDR(nodeinfo_idr);
static DECLARE_RWSEM(nodeinfo_lock);
static int max_nodeid;

struct cbuf {
	unsigned int base;
	unsigned int len;
	unsigned int mask;
};

/* Just the one of these, now. But this struct keeps
   the connection-specific variables together */

#define CF_READ_PENDING 1

struct connection {
	struct socket           *sock;
	unsigned long		flags;
	struct page             *rx_page;
	atomic_t		waiting_requests;
	struct cbuf		cb;
	int                     eagain_flag;
};

/* 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 nodeinfo         *ni;
};

static void cbuf_add(struct cbuf *cb, int n)
{
	cb->len += n;
}

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;
}

/* List of nodes which have writes pending */
static LIST_HEAD(write_nodes);
static DEFINE_SPINLOCK(write_nodes_lock);

/* Maximum number of incoming messages to process before
 * doing a schedule()
 */
#define MAX_RX_MSG_COUNT 25

/* Manage daemons */
static struct task_struct *recv_task;
static struct task_struct *send_task;
static DECLARE_WAIT_QUEUE_HEAD(lowcomms_recv_wait);

/* The SCTP connection */
static struct connection sctp_con;


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;
		memcpy(&ret6->sin6_addr, &in6->sin6_addr,
		       sizeof(in6->sin6_addr));
	}

	return 0;
}

/* If alloc is 0 here we will not attempt to allocate a new
   nodeinfo struct */
static struct nodeinfo *nodeid2nodeinfo(int nodeid, gfp_t alloc)
{
	struct nodeinfo *ni;
	int r;
	int n;

	down_read(&nodeinfo_lock);
	ni = idr_find(&nodeinfo_idr, nodeid);
	up_read(&nodeinfo_lock);

	if (ni || !alloc)
		return ni;

	down_write(&nodeinfo_lock);

	ni = idr_find(&nodeinfo_idr, nodeid);
	if (ni)
		goto out_up;

	r = idr_pre_get(&nodeinfo_idr, alloc);
	if (!r)
		goto out_up;

	ni = kmalloc(sizeof(struct nodeinfo), alloc);
	if (!ni)
		goto out_up;

	r = idr_get_new_above(&nodeinfo_idr, ni, nodeid, &n);
	if (r) {
		kfree(ni);
		ni = NULL;
		goto out_up;
	}
	if (n != nodeid) {
		idr_remove(&nodeinfo_idr, n);
		kfree(ni);
		ni = NULL;
		goto out_up;
	}
	memset(ni, 0, sizeof(struct nodeinfo));
	spin_lock_init(&ni->lock);
	INIT_LIST_HEAD(&ni->writequeue);
	spin_lock_init(&ni->writequeue_lock);
	ni->nodeid = nodeid;

	if (nodeid > max_nodeid)
		max_nodeid = nodeid;
out_up:
	up_write(&nodeinfo_lock);

	return ni;
}

/* Don't call this too often... */
static struct nodeinfo *assoc2nodeinfo(sctp_assoc_t assoc)
{
	int i;
	struct nodeinfo *ni;

	for (i=1; i<=max_nodeid; i++) {
		ni = nodeid2nodeinfo(i, 0);
		if (ni && ni->assoc_id == assoc)
			return ni;
	}
	return NULL;
}

/* Data or notification available on socket */
static void lowcomms_data_ready(struct sock *sk, int count_unused)
{
	atomic_inc(&sctp_con.waiting_requests);
	if (test_and_set_bit(CF_READ_PENDING, &sctp_con.flags))
		return;

	wake_up_interruptible(&lowcomms_recv_wait);
}


/* Add the port number to an IP6 or 4 sockaddr and return the address length.
   Also padd out the struct with zeros to make comparisons meaningful */

static void make_sockaddr(struct sockaddr_storage *saddr, uint16_t port,
			  int *addr_len)
{
	struct sockaddr_in *local4_addr;
	struct sockaddr_in6 *local6_addr;

	if (!dlm_local_count)
		return;

	if (!port) {
		if (dlm_local_addr[0]->ss_family == AF_INET) {
			local4_addr = (struct sockaddr_in *)dlm_local_addr[0];
			port = be16_to_cpu(local4_addr->sin_port);
		} else {
			local6_addr = (struct sockaddr_in6 *)dlm_local_addr[0];
			port = be16_to_cpu(local6_addr->sin6_port);
		}
	}

	saddr->ss_family = dlm_local_addr[0]->ss_family;
	if (dlm_local_addr[0]->ss_family == AF_INET) {
		struct sockaddr_in *in4_addr = (struct sockaddr_in *)saddr;
		in4_addr->sin_port = cpu_to_be16(port);
		memset(&in4_addr->sin_zero, 0, sizeof(in4_addr->sin_zero));
		memset(in4_addr+1, 0, sizeof(struct sockaddr_storage) -
		       sizeof(struct sockaddr_in));
		*addr_len = sizeof(struct sockaddr_in);
	} else {
		struct sockaddr_in6 *in6_addr = (struct sockaddr_in6 *)saddr;
		in6_addr->sin6_port = cpu_to_be16(port);
		memset(in6_addr+1, 0, sizeof(struct sockaddr_storage) -
		       sizeof(struct sockaddr_in6));
		*addr_len = sizeof(struct sockaddr_in6);
	}
}

/* Close the connection and tidy up */
static void close_connection(void)
{
	if (sctp_con.sock) {
		sock_release(sctp_con.sock);
		sctp_con.sock = NULL;
	}

	if (sctp_con.rx_page) {
		__free_page(sctp_con.rx_page);
		sctp_con.rx_page = NULL;
	}
}

/* We only send shutdown messages to nodes that are not part of the cluster */
static void 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;

	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(sctp_con.sock, &outmessage, NULL, 0, 0);

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


/* INIT failed but we don't know which node...
   restart INIT on all pending nodes */
static void init_failed(void)
{
	int i;
	struct nodeinfo *ni;

	for (i=1; i<=max_nodeid; i++) {
		ni = nodeid2nodeinfo(i, 0);
		if (!ni)
			continue;

		if (test_and_clear_bit(NI_INIT_PENDING, &ni->flags)) {
			ni->assoc_id = 0;
			if (!test_and_set_bit(NI_WRITE_PENDING, &ni->flags)) {
				spin_lock_bh(&write_nodes_lock);
				list_add_tail(&ni->write_list, &write_nodes);
				spin_unlock_bh(&write_nodes_lock);
			}
		}
	}
	wake_up_process(send_task);
}

/* Something happened to an association */
static void process_sctp_notification(struct msghdr *msg, char *buf)
{
	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;
			mm_segment_t fs;
			int nodeid;
			int prim_len, ret;
			int addr_len;
			struct nodeinfo *ni;

			/* This seems to happen when we received a connection
			 * too early... or something...  anyway, it happens but
			 * we always seem to get a real message too, see
			 * receive_from_sock */

			if ((int)sn->sn_assoc_change.sac_assoc_id <= 0) {
				log_print("COMM_UP for invalid assoc ID %d",
					  (int)sn->sn_assoc_change.sac_assoc_id);
				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;

			fs = get_fs();
			set_fs(get_ds());
			ret = sctp_con.sock->ops->getsockopt(sctp_con.sock,
							     IPPROTO_SCTP,
							     SCTP_PRIMARY_ADDR,
							     (char*)&prim,
							     &prim_len);
			set_fs(fs);
			if (ret < 0) {
				struct nodeinfo *ni;

				log_print("getsockopt/sctp_primary_addr on "
					  "new assoc %d failed : %d",
					  (int)sn->sn_assoc_change.sac_assoc_id,
					  ret);

				/* Retry INIT later */
				ni = assoc2nodeinfo(sn->sn_assoc_change.sac_assoc_id);
				if (ni)
					clear_bit(NI_INIT_PENDING, &ni->flags);
				return;
			}
			make_sockaddr(&prim.ssp_addr, 0, &addr_len);
			if (dlm_addr_to_nodeid(&prim.ssp_addr, &nodeid)) {
				log_print("reject connect from unknown addr");
				send_shutdown(prim.ssp_assoc_id);
				return;
			}

			ni = nodeid2nodeinfo(nodeid, GFP_KERNEL);
			if (!ni)
				return;

			/* Save the assoc ID */
			ni->assoc_id = sn->sn_assoc_change.sac_assoc_id;

			log_print("got new/restarted association %d nodeid %d",
				  (int)sn->sn_assoc_change.sac_assoc_id, nodeid);

			/* Send any pending writes */
			clear_bit(NI_INIT_PENDING, &ni->flags);
			if (!test_and_set_bit(NI_WRITE_PENDING, &ni->flags)) {
				spin_lock_bh(&write_nodes_lock);
				list_add_tail(&ni->write_list, &write_nodes);
				spin_unlock_bh(&write_nodes_lock);
			}
			wake_up_process(send_task);
		}
		break;

		case SCTP_COMM_LOST:
		case SCTP_SHUTDOWN_COMP:
		{
			struct nodeinfo *ni;

			ni = assoc2nodeinfo(sn->sn_assoc_change.sac_assoc_id);
			if (ni) {
				spin_lock(&ni->lock);
				ni->assoc_id = 0;
				spin_unlock(&ni->lock);
			}
		}
		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");
			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);
		}
	}
}

/* Data received from remote end */
static int receive_from_sock(void)
{
	int ret = 0;
	struct msghdr msg;
	struct kvec iov[2];
	unsigned len;
	int r;
	struct sctp_sndrcvinfo *sinfo;
	struct cmsghdr *cmsg;
	struct nodeinfo *ni;

	/* These two are marginally too big for stack allocation, but this
	 * function is (currently) only called by dlm_recvd so static should be
	 * OK.
	 */
	static struct sockaddr_storage msgname;
	static char incmsg[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];

	if (sctp_con.sock == NULL)
		goto out;

	if (sctp_con.rx_page == NULL) {
		/*
		 * This doesn't need to be atomic, but I think it should
		 * improve performance if it is.
		 */
		sctp_con.rx_page = alloc_page(GFP_ATOMIC);
		if (sctp_con.rx_page == NULL)
			goto out_resched;
		cbuf_init(&sctp_con.cb, PAGE_CACHE_SIZE);
	}

	memset(&incmsg, 0, sizeof(incmsg));
	memset(&msgname, 0, sizeof(msgname));

	msg.msg_name = &msgname;
	msg.msg_namelen = sizeof(msgname);
	msg.msg_flags = 0;
	msg.msg_control = incmsg;
	msg.msg_controllen = sizeof(incmsg);
	msg.msg_iovlen = 1;

	/* I don't see why this circular buffer stuff is necessary for SCTP
	 * which is a packet-based protocol, but the whole thing breaks under
	 * load without it! The overhead is minimal (and is in the TCP lowcomms
	 * anyway, of course) so I'll leave it in until I can figure out what's
	 * really happening.
	 */

	/*
	 * iov[0] is the bit of the circular buffer between the current end
	 * point (cb.base + cb.len) and the end of the buffer.
	 */
	iov[0].iov_len = sctp_con.cb.base - cbuf_data(&sctp_con.cb);
	iov[0].iov_base = page_address(sctp_con.rx_page) +
		cbuf_data(&sctp_con.cb);
	iov[1].iov_len = 0;

	/*
	 * 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)
	 */
	if (cbuf_data(&sctp_con.cb) >= sctp_con.cb.base) {
		iov[0].iov_len = PAGE_CACHE_SIZE - cbuf_data(&sctp_con.cb);
		iov[1].iov_len = sctp_con.cb.base;
		iov[1].iov_base = page_address(sctp_con.rx_page);
		msg.msg_iovlen = 2;
	}
	len = iov[0].iov_len + iov[1].iov_len;

	r = ret = kernel_recvmsg(sctp_con.sock, &msg, iov, msg.msg_iovlen, len,
				 MSG_NOSIGNAL | MSG_DONTWAIT);
	if (ret <= 0)
		goto out_close;

	msg.msg_control = incmsg;
	msg.msg_controllen = sizeof(incmsg);
	cmsg = CMSG_FIRSTHDR(&msg);
	sinfo = CMSG_DATA(cmsg);

	if (msg.msg_flags & MSG_NOTIFICATION) {
		process_sctp_notification(&msg, page_address(sctp_con.rx_page));
		return 0;
	}

	/* Is this a new association ? */
	ni = nodeid2nodeinfo(le32_to_cpu(sinfo->sinfo_ppid), GFP_KERNEL);
	if (ni) {
		ni->assoc_id = sinfo->sinfo_assoc_id;
		if (test_and_clear_bit(NI_INIT_PENDING, &ni->flags)) {

			if (!test_and_set_bit(NI_WRITE_PENDING, &ni->flags)) {
				spin_lock_bh(&write_nodes_lock);
				list_add_tail(&ni->write_list, &write_nodes);
				spin_unlock_bh(&write_nodes_lock);
			}
			wake_up_process(send_task);
		}
	}

	/* INIT sends a message with length of 1 - ignore it */
	if (r == 1)
		return 0;

	cbuf_add(&sctp_con.cb, ret);
	ret = dlm_process_incoming_buffer(cpu_to_le32(sinfo->sinfo_ppid),
					  page_address(sctp_con.rx_page),
					  sctp_con.cb.base, sctp_con.cb.len,
					  PAGE_CACHE_SIZE);
	if (ret < 0)
		goto out_close;
	cbuf_eat(&sctp_con.cb, ret);

out:
	ret = 0;
	goto out_ret;

out_resched:
	lowcomms_data_ready(sctp_con.sock->sk, 0);
	ret = 0;
	cond_resched();
	goto out_ret;

out_close:
	if (ret != -EAGAIN)
		log_print("error reading from sctp socket: %d", ret);
out_ret:
	return ret;
}

/* Bind to an IP address. SCTP allows multiple address so it can do multi-homing */
static int add_bind_addr(struct sockaddr_storage *addr, int addr_len, int num)
{
	mm_segment_t fs;
	int result = 0;

	fs = get_fs();
	set_fs(get_ds());
	if (num == 1)
		result = sctp_con.sock->ops->bind(sctp_con.sock,
						  (struct sockaddr *) addr,
						  addr_len);
	else
		result = sctp_con.sock->ops->setsockopt(sctp_con.sock, SOL_SCTP,
							SCTP_SOCKOPT_BINDX_ADD,
							(char *)addr, addr_len);
	set_fs(fs);

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

	return result;
}

static void init_local(void)
{
	struct sockaddr_storage sas, *addr;
	int i;

	dlm_local_nodeid = dlm_our_nodeid();

	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;
	}
}

/* Initialise SCTP socket and bind to all interfaces */
static int init_sock(void)
{
	mm_segment_t fs;
	struct socket *sock = NULL;
	struct sockaddr_storage localaddr;
	struct sctp_event_subscribe subscribe;
	int result = -EINVAL, num = 1, i, addr_len;

	if (!dlm_local_count) {
		init_local();
		if (!dlm_local_count) {
			log_print("no local IP address has been set");
			goto out;
		}
	}

	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;

	fs = get_fs();
	set_fs(get_ds());
	result = sock->ops->setsockopt(sock, SOL_SCTP, SCTP_EVENTS,
				       (char *)&subscribe, sizeof(subscribe));
	set_fs(fs);

	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 = &sctp_con;
	sctp_con.sock = sock;
	sctp_con.sock->sk->sk_data_ready = lowcomms_data_ready;

	/* 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.tcp_port, &addr_len);

		result = add_bind_addr(&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);
	sctp_con.sock = NULL;
out:
	return result;
}


static struct writequeue_entry *new_writequeue_entry(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;

	return entry;
}

void *dlm_lowcomms_get_buffer(int nodeid, int len, gfp_t allocation, char **ppc)
{
	struct writequeue_entry *e;
	int offset = 0;
	int users = 0;
	struct nodeinfo *ni;

	ni = nodeid2nodeinfo(nodeid, allocation);
	if (!ni)
		return NULL;

	spin_lock(&ni->writequeue_lock);
	e = list_entry(ni->writequeue.prev, struct writequeue_entry, list);
	if ((&e->list == &ni->writequeue) ||
	    (PAGE_CACHE_SIZE - e->end < len)) {
		e = NULL;
	} else {
		offset = e->end;
		e->end += len;
		users = e->users++;
	}
	spin_unlock(&ni->writequeue_lock);

	if (e) {
	got_one:
		if (users == 0)
			kmap(e->page);
		*ppc = page_address(e->page) + offset;
		return e;
	}

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

void dlm_lowcomms_commit_buffer(void *arg)
{
	struct writequeue_entry *e = (struct writequeue_entry *) arg;
	int users;
	struct nodeinfo *ni = e->ni;

	spin_lock(&ni->writequeue_lock);
	users = --e->users;
	if (users)
		goto out;
	e->len = e->end - e->offset;
	kunmap(e->page);
	spin_unlock(&ni->writequeue_lock);

	if (!test_and_set_bit(NI_WRITE_PENDING, &ni->flags)) {
		spin_lock_bh(&write_nodes_lock);
		list_add_tail(&ni->write_list, &write_nodes);
		spin_unlock_bh(&write_nodes_lock);
		wake_up_process(send_task);
	}
	return;

out:
	spin_unlock(&ni->writequeue_lock);
	return;
}

static void free_entry(struct writequeue_entry *e)
{
	__free_page(e->page);
	kfree(e);
}

/* Initiate an SCTP association. In theory we could just use sendmsg() on
   the first IP address and it should work, but this allows us to set up the
   association before sending any valuable data that we can't afford to lose.
   It also keeps the send path clean as it can now always use the association ID */
static void initiate_association(int nodeid)
{
	struct sockaddr_storage rem_addr;
	static char outcmsg[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];
	struct msghdr outmessage;
	struct cmsghdr *cmsg;
	struct sctp_sndrcvinfo *sinfo;
	int ret;
	int addrlen;
	char buf[1];
	struct kvec iov[1];
	struct nodeinfo *ni;

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

	ni = nodeid2nodeinfo(nodeid, GFP_KERNEL);
	if (!ni)
		return;

	if (nodeid_to_addr(nodeid, (struct sockaddr *)&rem_addr)) {
		log_print("no address for nodeid %d", nodeid);
		return;
	}

	make_sockaddr(&rem_addr, dlm_config.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;

	iov[0].iov_base = buf;
	iov[0].iov_len = 1;

	/* Real INIT messages seem to cause trouble. Just send a 1 byte message
	   we can afford to lose */
	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_local_nodeid);

	outmessage.msg_controllen = cmsg->cmsg_len;
	ret = kernel_sendmsg(sctp_con.sock, &outmessage, iov, 1, 1);
	if (ret < 0) {
		log_print("send INIT to node failed: %d", ret);
		/* Try again later */
		clear_bit(NI_INIT_PENDING, &ni->flags);
	}
}

/* Send a message */
static void send_to_sock(struct nodeinfo *ni)
{
	int ret = 0;
	struct writequeue_entry *e;
	int len, offset;
	struct msghdr outmsg;
	static char outcmsg[CMSG_SPACE(sizeof(struct sctp_sndrcvinfo))];
	struct cmsghdr *cmsg;
	struct sctp_sndrcvinfo *sinfo;
	struct kvec iov;

	/* See if we need to init an association before we start
	   sending precious messages */
	spin_lock(&ni->lock);
	if (!ni->assoc_id && !test_and_set_bit(NI_INIT_PENDING, &ni->flags)) {
		spin_unlock(&ni->lock);
		initiate_association(ni->nodeid);
		return;
	}
	spin_unlock(&ni->lock);

	outmsg.msg_name = NULL; /* We use assoc_id */
	outmsg.msg_namelen = 0;
	outmsg.msg_control = outcmsg;
	outmsg.msg_controllen = sizeof(outcmsg);
	outmsg.msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL | MSG_EOR;

	cmsg = CMSG_FIRSTHDR(&outmsg);
	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_local_nodeid);
	sinfo->sinfo_assoc_id = ni->assoc_id;
	outmsg.msg_controllen = cmsg->cmsg_len;

	spin_lock(&ni->writequeue_lock);
	for (;;) {
		if (list_empty(&ni->writequeue))
			break;
		e = list_entry(ni->writequeue.next, struct writequeue_entry,
			       list);
		len = e->len;
		offset = e->offset;
		BUG_ON(len == 0 && e->users == 0);
		spin_unlock(&ni->writequeue_lock);
		kmap(e->page);

		ret = 0;
		if (len) {
			iov.iov_base = page_address(e->page)+offset;
			iov.iov_len = len;

			ret = kernel_sendmsg(sctp_con.sock, &outmsg, &iov, 1,
					     len);
			if (ret == -EAGAIN) {
				sctp_con.eagain_flag = 1;
				goto out;
			} else if (ret < 0)
				goto send_error;
		} else {
			/* Don't starve people filling buffers */
			cond_resched();
		}

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

		if (e->len == 0 && e->users == 0) {
			list_del(&e->list);
			kunmap(e->page);
			free_entry(e);
			continue;
		}
	}
	spin_unlock(&ni->writequeue_lock);
out:
	return;

send_error:
	log_print("Error sending to node %d %d", ni->nodeid, ret);
	spin_lock(&ni->lock);
	if (!test_and_set_bit(NI_INIT_PENDING, &ni->flags)) {
		ni->assoc_id = 0;
		spin_unlock(&ni->lock);
		initiate_association(ni->nodeid);
	} else
		spin_unlock(&ni->lock);

	return;
}

/* Try to send any messages that are pending */
static void process_output_queue(void)
{
	struct list_head *list;
	struct list_head *temp;

	spin_lock_bh(&write_nodes_lock);
	list_for_each_safe(list, temp, &write_nodes) {
		struct nodeinfo *ni =
			list_entry(list, struct nodeinfo, write_list);
		clear_bit(NI_WRITE_PENDING, &ni->flags);
		list_del(&ni->write_list);

		spin_unlock_bh(&write_nodes_lock);

		send_to_sock(ni);
		spin_lock_bh(&write_nodes_lock);
	}
	spin_unlock_bh(&write_nodes_lock);
}

/* Called after we've had -EAGAIN and been woken up */
static void refill_write_queue(void)
{
	int i;

	for (i=1; i<=max_nodeid; i++) {
		struct nodeinfo *ni = nodeid2nodeinfo(i, 0);

		if (ni) {
			if (!test_and_set_bit(NI_WRITE_PENDING, &ni->flags)) {
				spin_lock_bh(&write_nodes_lock);
				list_add_tail(&ni->write_list, &write_nodes);
				spin_unlock_bh(&write_nodes_lock);
			}
		}
	}
}

static void clean_one_writequeue(struct nodeinfo *ni)
{
	struct list_head *list;
	struct list_head *temp;

	spin_lock(&ni->writequeue_lock);
	list_for_each_safe(list, temp, &ni->writequeue) {
		struct writequeue_entry *e =
			list_entry(list, struct writequeue_entry, list);
		list_del(&e->list);
		free_entry(e);
	}
	spin_unlock(&ni->writequeue_lock);
}

static void clean_writequeues(void)
{
	int i;

	for (i=1; i<=max_nodeid; i++) {
		struct nodeinfo *ni = nodeid2nodeinfo(i, 0);
		if (ni)
			clean_one_writequeue(ni);
	}
}


static void dealloc_nodeinfo(void)
{
	int i;

	for (i=1; i<=max_nodeid; i++) {
		struct nodeinfo *ni = nodeid2nodeinfo(i, 0);
		if (ni) {
			idr_remove(&nodeinfo_idr, i);
			kfree(ni);
		}
	}
}

int dlm_lowcomms_close(int nodeid)
{
	struct nodeinfo *ni;

	ni = nodeid2nodeinfo(nodeid, 0);
	if (!ni)
		return -1;

	spin_lock(&ni->lock);
	if (ni->assoc_id) {
		ni->assoc_id = 0;
		/* Don't send shutdown here, sctp will just queue it
		   till the node comes back up! */
	}
	spin_unlock(&ni->lock);

	clean_one_writequeue(ni);
	clear_bit(NI_INIT_PENDING, &ni->flags);
	return 0;
}

static int write_list_empty(void)
{
	int status;

	spin_lock_bh(&write_nodes_lock);
	status = list_empty(&write_nodes);
	spin_unlock_bh(&write_nodes_lock);

	return status;
}

static int dlm_recvd(void *data)
{
	DECLARE_WAITQUEUE(wait, current);

	while (!kthread_should_stop()) {
		int count = 0;

		set_current_state(TASK_INTERRUPTIBLE);
		add_wait_queue(&lowcomms_recv_wait, &wait);
		if (!test_bit(CF_READ_PENDING, &sctp_con.flags))
			cond_resched();
		remove_wait_queue(&lowcomms_recv_wait, &wait);
		set_current_state(TASK_RUNNING);

		if (test_and_clear_bit(CF_READ_PENDING, &sctp_con.flags)) {
			int ret;

			do {
				ret = receive_from_sock();

				/* Don't starve out everyone else */
				if (++count >= MAX_RX_MSG_COUNT) {
					cond_resched();
					count = 0;
				}
			} while (!kthread_should_stop() && ret >=0);
		}
		cond_resched();
	}

	return 0;
}

static int dlm_sendd(void *data)
{
	DECLARE_WAITQUEUE(wait, current);

	add_wait_queue(sctp_con.sock->sk->sk_sleep, &wait);

	while (!kthread_should_stop()) {
		set_current_state(TASK_INTERRUPTIBLE);
		if (write_list_empty())
			cond_resched();
		set_current_state(TASK_RUNNING);

		if (sctp_con.eagain_flag) {
			sctp_con.eagain_flag = 0;
			refill_write_queue();
		}
		process_output_queue();
	}

	remove_wait_queue(sctp_con.sock->sk->sk_sleep, &wait);

	return 0;
}

static void daemons_stop(void)
{
	kthread_stop(recv_task);
	kthread_stop(send_task);
}

static int daemons_start(void)
{
	struct task_struct *p;
	int error;

	p = kthread_run(dlm_recvd, NULL, "dlm_recvd");
	error = IS_ERR(p);
	if (error) {
		log_print("can't start dlm_recvd %d", error);
		return error;
	}
	recv_task = p;

	p = kthread_run(dlm_sendd, NULL, "dlm_sendd");
	error = IS_ERR(p);
	if (error) {
		log_print("can't start dlm_sendd %d", error);
		kthread_stop(recv_task);
		return error;
	}
	send_task = p;

	return 0;
}

/*
 * This is quite likely to sleep...
 */
int dlm_lowcomms_start(void)
{
	int error;

	error = init_sock();
	if (error)
		goto fail_sock;
	error = daemons_start();
	if (error)
		goto fail_sock;
	return 0;

fail_sock:
	close_connection();
	return error;
}

void dlm_lowcomms_stop(void)
{
	int i;

	sctp_con.flags = 0x7;
	daemons_stop();
	clean_writequeues();
	close_connection();
	dealloc_nodeinfo();
	max_nodeid = 0;

	dlm_local_count = 0;
	dlm_local_nodeid = 0;

	for (i = 0; i < dlm_local_count; i++)
		kfree(dlm_local_addr[i]);
}