summaryrefslogtreecommitdiffstats
path: root/net/rds/tcp.c
blob: 9920d2f84eff8e358702ec69f010f0394b0f63fa (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
/*
 * Copyright (c) 2006 Oracle.  All rights reserved.
 *
 * This software is available to you under a choice of one of two
 * licenses.  You may choose to be licensed under the terms of the GNU
 * General Public License (GPL) Version 2, available from the file
 * COPYING in the main directory of this source tree, or the
 * OpenIB.org BSD license below:
 *
 *     Redistribution and use in source and binary forms, with or
 *     without modification, are permitted provided that the following
 *     conditions are met:
 *
 *      - Redistributions of source code must retain the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer.
 *
 *      - Redistributions in binary form must reproduce the above
 *        copyright notice, this list of conditions and the following
 *        disclaimer in the documentation and/or other materials
 *        provided with the distribution.
 *
 * THE SOFTWARE IS PROVIDED "AS IS", WITHOUT WARRANTY OF ANY KIND,
 * EXPRESS OR IMPLIED, INCLUDING BUT NOT LIMITED TO THE WARRANTIES OF
 * MERCHANTABILITY, FITNESS FOR A PARTICULAR PURPOSE AND
 * NONINFRINGEMENT. IN NO EVENT SHALL THE AUTHORS OR COPYRIGHT HOLDERS
 * BE LIABLE FOR ANY CLAIM, DAMAGES OR OTHER LIABILITY, WHETHER IN AN
 * ACTION OF CONTRACT, TORT OR OTHERWISE, ARISING FROM, OUT OF OR IN
 * CONNECTION WITH THE SOFTWARE OR THE USE OR OTHER DEALINGS IN THE
 * SOFTWARE.
 *
 */
#include <linux/kernel.h>
#include <linux/slab.h>
#include <linux/in.h>
#include <linux/module.h>
#include <net/tcp.h>
#include <net/net_namespace.h>
#include <net/netns/generic.h>

#include "rds.h"
#include "tcp.h"

/* only for info exporting */
static DEFINE_SPINLOCK(rds_tcp_tc_list_lock);
static LIST_HEAD(rds_tcp_tc_list);
static unsigned int rds_tcp_tc_count;

/* Track rds_tcp_connection structs so they can be cleaned up */
static DEFINE_SPINLOCK(rds_tcp_conn_lock);
static LIST_HEAD(rds_tcp_conn_list);

static struct kmem_cache *rds_tcp_conn_slab;

static int rds_tcp_skbuf_handler(struct ctl_table *ctl, int write,
				 void __user *buffer, size_t *lenp,
				 loff_t *fpos);

static int rds_tcp_min_sndbuf = SOCK_MIN_SNDBUF;
static int rds_tcp_min_rcvbuf = SOCK_MIN_RCVBUF;

static struct ctl_table rds_tcp_sysctl_table[] = {
#define	RDS_TCP_SNDBUF	0
	{
		.procname       = "rds_tcp_sndbuf",
		/* data is per-net pointer */
		.maxlen         = sizeof(int),
		.mode           = 0644,
		.proc_handler   = rds_tcp_skbuf_handler,
		.extra1		= &rds_tcp_min_sndbuf,
	},
#define	RDS_TCP_RCVBUF	1
	{
		.procname       = "rds_tcp_rcvbuf",
		/* data is per-net pointer */
		.maxlen         = sizeof(int),
		.mode           = 0644,
		.proc_handler   = rds_tcp_skbuf_handler,
		.extra1		= &rds_tcp_min_rcvbuf,
	},
	{ }
};

/* doing it this way avoids calling tcp_sk() */
void rds_tcp_nonagle(struct socket *sock)
{
	int val = 1;

	kernel_setsockopt(sock, SOL_TCP, TCP_NODELAY, (void *)&val,
			      sizeof(val));
}

u32 rds_tcp_write_seq(struct rds_tcp_connection *tc)
{
	/* seq# of the last byte of data in tcp send buffer */
	return tcp_sk(tc->t_sock->sk)->write_seq;
}

u32 rds_tcp_snd_una(struct rds_tcp_connection *tc)
{
	return tcp_sk(tc->t_sock->sk)->snd_una;
}

void rds_tcp_restore_callbacks(struct socket *sock,
			       struct rds_tcp_connection *tc)
{
	rdsdebug("restoring sock %p callbacks from tc %p\n", sock, tc);
	write_lock_bh(&sock->sk->sk_callback_lock);

	/* done under the callback_lock to serialize with write_space */
	spin_lock(&rds_tcp_tc_list_lock);
	list_del_init(&tc->t_list_item);
	rds_tcp_tc_count--;
	spin_unlock(&rds_tcp_tc_list_lock);

	tc->t_sock = NULL;

	sock->sk->sk_write_space = tc->t_orig_write_space;
	sock->sk->sk_data_ready = tc->t_orig_data_ready;
	sock->sk->sk_state_change = tc->t_orig_state_change;
	sock->sk->sk_user_data = NULL;

	write_unlock_bh(&sock->sk->sk_callback_lock);
}

/*
 * rds_tcp_reset_callbacks() switches the to the new sock and
 * returns the existing tc->t_sock.
 *
 * The only functions that set tc->t_sock are rds_tcp_set_callbacks
 * and rds_tcp_reset_callbacks.  Send and receive trust that
 * it is set.  The absence of RDS_CONN_UP bit protects those paths
 * from being called while it isn't set.
 */
void rds_tcp_reset_callbacks(struct socket *sock,
			     struct rds_conn_path *cp)
{
	struct rds_tcp_connection *tc = cp->cp_transport_data;
	struct socket *osock = tc->t_sock;

	if (!osock)
		goto newsock;

	/* Need to resolve a duelling SYN between peers.
	 * We have an outstanding SYN to this peer, which may
	 * potentially have transitioned to the RDS_CONN_UP state,
	 * so we must quiesce any send threads before resetting
	 * cp_transport_data. We quiesce these threads by setting
	 * cp_state to something other than RDS_CONN_UP, and then
	 * waiting for any existing threads in rds_send_xmit to
	 * complete release_in_xmit(). (Subsequent threads entering
	 * rds_send_xmit() will bail on !rds_conn_up().
	 *
	 * However an incoming syn-ack at this point would end up
	 * marking the conn as RDS_CONN_UP, and would again permit
	 * rds_send_xmi() threads through, so ideally we would
	 * synchronize on RDS_CONN_UP after lock_sock(), but cannot
	 * do that: waiting on !RDS_IN_XMIT after lock_sock() may
	 * end up deadlocking with tcp_sendmsg(), and the RDS_IN_XMIT
	 * would not get set. As a result, we set c_state to
	 * RDS_CONN_RESETTTING, to ensure that rds_tcp_state_change
	 * cannot mark rds_conn_path_up() in the window before lock_sock()
	 */
	atomic_set(&cp->cp_state, RDS_CONN_RESETTING);
	wait_event(cp->cp_waitq, !test_bit(RDS_IN_XMIT, &cp->cp_flags));
	lock_sock(osock->sk);
	/* reset receive side state for rds_tcp_data_recv() for osock  */
	cancel_delayed_work_sync(&cp->cp_send_w);
	cancel_delayed_work_sync(&cp->cp_recv_w);
	if (tc->t_tinc) {
		rds_inc_put(&tc->t_tinc->ti_inc);
		tc->t_tinc = NULL;
	}
	tc->t_tinc_hdr_rem = sizeof(struct rds_header);
	tc->t_tinc_data_rem = 0;
	rds_tcp_restore_callbacks(osock, tc);
	release_sock(osock->sk);
	sock_release(osock);
newsock:
	rds_send_path_reset(cp);
	lock_sock(sock->sk);
	rds_tcp_set_callbacks(sock, cp);
	release_sock(sock->sk);
}

/* Add tc to rds_tcp_tc_list and set tc->t_sock. See comments
 * above rds_tcp_reset_callbacks for notes about synchronization
 * with data path
 */
void rds_tcp_set_callbacks(struct socket *sock, struct rds_conn_path *cp)
{
	struct rds_tcp_connection *tc = cp->cp_transport_data;

	rdsdebug("setting sock %p callbacks to tc %p\n", sock, tc);
	write_lock_bh(&sock->sk->sk_callback_lock);

	/* done under the callback_lock to serialize with write_space */
	spin_lock(&rds_tcp_tc_list_lock);
	list_add_tail(&tc->t_list_item, &rds_tcp_tc_list);
	rds_tcp_tc_count++;
	spin_unlock(&rds_tcp_tc_list_lock);

	/* accepted sockets need our listen data ready undone */
	if (sock->sk->sk_data_ready == rds_tcp_listen_data_ready)
		sock->sk->sk_data_ready = sock->sk->sk_user_data;

	tc->t_sock = sock;
	tc->t_cpath = cp;
	tc->t_orig_data_ready = sock->sk->sk_data_ready;
	tc->t_orig_write_space = sock->sk->sk_write_space;
	tc->t_orig_state_change = sock->sk->sk_state_change;

	sock->sk->sk_user_data = cp;
	sock->sk->sk_data_ready = rds_tcp_data_ready;
	sock->sk->sk_write_space = rds_tcp_write_space;
	sock->sk->sk_state_change = rds_tcp_state_change;

	write_unlock_bh(&sock->sk->sk_callback_lock);
}

static void rds_tcp_tc_info(struct socket *rds_sock, unsigned int len,
			    struct rds_info_iterator *iter,
			    struct rds_info_lengths *lens)
{
	struct rds_info_tcp_socket tsinfo;
	struct rds_tcp_connection *tc;
	unsigned long flags;
	struct sockaddr_in sin;
	int sinlen;
	struct socket *sock;

	spin_lock_irqsave(&rds_tcp_tc_list_lock, flags);

	if (len / sizeof(tsinfo) < rds_tcp_tc_count)
		goto out;

	list_for_each_entry(tc, &rds_tcp_tc_list, t_list_item) {

		sock = tc->t_sock;
		if (sock) {
			sock->ops->getname(sock, (struct sockaddr *)&sin,
					   &sinlen, 0);
			tsinfo.local_addr = sin.sin_addr.s_addr;
			tsinfo.local_port = sin.sin_port;
			sock->ops->getname(sock, (struct sockaddr *)&sin,
					   &sinlen, 1);
			tsinfo.peer_addr = sin.sin_addr.s_addr;
			tsinfo.peer_port = sin.sin_port;
		}

		tsinfo.hdr_rem = tc->t_tinc_hdr_rem;
		tsinfo.data_rem = tc->t_tinc_data_rem;
		tsinfo.last_sent_nxt = tc->t_last_sent_nxt;
		tsinfo.last_expected_una = tc->t_last_expected_una;
		tsinfo.last_seen_una = tc->t_last_seen_una;

		rds_info_copy(iter, &tsinfo, sizeof(tsinfo));
	}

out:
	lens->nr = rds_tcp_tc_count;
	lens->each = sizeof(tsinfo);

	spin_unlock_irqrestore(&rds_tcp_tc_list_lock, flags);
}

static int rds_tcp_laddr_check(struct net *net, __be32 addr)
{
	if (inet_addr_type(net, addr) == RTN_LOCAL)
		return 0;
	return -EADDRNOTAVAIL;
}

static void rds_tcp_conn_free(void *arg)
{
	struct rds_tcp_connection *tc = arg;
	unsigned long flags;

	rdsdebug("freeing tc %p\n", tc);

	spin_lock_irqsave(&rds_tcp_conn_lock, flags);
	if (!tc->t_tcp_node_detached)
		list_del(&tc->t_tcp_node);
	spin_unlock_irqrestore(&rds_tcp_conn_lock, flags);

	kmem_cache_free(rds_tcp_conn_slab, tc);
}

static int rds_tcp_conn_alloc(struct rds_connection *conn, gfp_t gfp)
{
	struct rds_tcp_connection *tc;
	int i, j;
	int ret = 0;

	for (i = 0; i < RDS_MPATH_WORKERS; i++) {
		tc = kmem_cache_alloc(rds_tcp_conn_slab, gfp);
		if (!tc) {
			ret = -ENOMEM;
			break;
		}
		mutex_init(&tc->t_conn_path_lock);
		tc->t_sock = NULL;
		tc->t_tinc = NULL;
		tc->t_tinc_hdr_rem = sizeof(struct rds_header);
		tc->t_tinc_data_rem = 0;

		conn->c_path[i].cp_transport_data = tc;
		tc->t_cpath = &conn->c_path[i];

		spin_lock_irq(&rds_tcp_conn_lock);
		tc->t_tcp_node_detached = false;
		list_add_tail(&tc->t_tcp_node, &rds_tcp_conn_list);
		spin_unlock_irq(&rds_tcp_conn_lock);
		rdsdebug("rds_conn_path [%d] tc %p\n", i,
			 conn->c_path[i].cp_transport_data);
	}
	if (ret) {
		for (j = 0; j < i; j++)
			rds_tcp_conn_free(conn->c_path[j].cp_transport_data);
	}
	return ret;
}

static bool list_has_conn(struct list_head *list, struct rds_connection *conn)
{
	struct rds_tcp_connection *tc, *_tc;

	list_for_each_entry_safe(tc, _tc, list, t_tcp_node) {
		if (tc->t_cpath->cp_conn == conn)
			return true;
	}
	return false;
}

static void rds_tcp_destroy_conns(void)
{
	struct rds_tcp_connection *tc, *_tc;
	LIST_HEAD(tmp_list);

	/* avoid calling conn_destroy with irqs off */
	spin_lock_irq(&rds_tcp_conn_lock);
	list_for_each_entry_safe(tc, _tc, &rds_tcp_conn_list, t_tcp_node) {
		if (!list_has_conn(&tmp_list, tc->t_cpath->cp_conn))
			list_move_tail(&tc->t_tcp_node, &tmp_list);
	}
	spin_unlock_irq(&rds_tcp_conn_lock);

	list_for_each_entry_safe(tc, _tc, &tmp_list, t_tcp_node)
		rds_conn_destroy(tc->t_cpath->cp_conn);
}

static void rds_tcp_exit(void);

struct rds_transport rds_tcp_transport = {
	.laddr_check		= rds_tcp_laddr_check,
	.xmit_path_prepare	= rds_tcp_xmit_path_prepare,
	.xmit_path_complete	= rds_tcp_xmit_path_complete,
	.xmit			= rds_tcp_xmit,
	.recv_path		= rds_tcp_recv_path,
	.conn_alloc		= rds_tcp_conn_alloc,
	.conn_free		= rds_tcp_conn_free,
	.conn_path_connect	= rds_tcp_conn_path_connect,
	.conn_path_shutdown	= rds_tcp_conn_path_shutdown,
	.inc_copy_to_user	= rds_tcp_inc_copy_to_user,
	.inc_free		= rds_tcp_inc_free,
	.stats_info_copy	= rds_tcp_stats_info_copy,
	.exit			= rds_tcp_exit,
	.t_owner		= THIS_MODULE,
	.t_name			= "tcp",
	.t_type			= RDS_TRANS_TCP,
	.t_prefer_loopback	= 1,
	.t_mp_capable		= 1,
};

static unsigned int rds_tcp_netid;

/* per-network namespace private data for this module */
struct rds_tcp_net {
	struct socket *rds_tcp_listen_sock;
	struct work_struct rds_tcp_accept_w;
	struct ctl_table_header *rds_tcp_sysctl;
	struct ctl_table *ctl_table;
	int sndbuf_size;
	int rcvbuf_size;
};

/* All module specific customizations to the RDS-TCP socket should be done in
 * rds_tcp_tune() and applied after socket creation.
 */
void rds_tcp_tune(struct socket *sock)
{
	struct sock *sk = sock->sk;
	struct net *net = sock_net(sk);
	struct rds_tcp_net *rtn = net_generic(net, rds_tcp_netid);

	rds_tcp_nonagle(sock);
	lock_sock(sk);
	if (rtn->sndbuf_size > 0) {
		sk->sk_sndbuf = rtn->sndbuf_size;
		sk->sk_userlocks |= SOCK_SNDBUF_LOCK;
	}
	if (rtn->rcvbuf_size > 0) {
		sk->sk_sndbuf = rtn->rcvbuf_size;
		sk->sk_userlocks |= SOCK_RCVBUF_LOCK;
	}
	release_sock(sk);
}

static void rds_tcp_accept_worker(struct work_struct *work)
{
	struct rds_tcp_net *rtn = container_of(work,
					       struct rds_tcp_net,
					       rds_tcp_accept_w);

	while (rds_tcp_accept_one(rtn->rds_tcp_listen_sock) == 0)
		cond_resched();
}

void rds_tcp_accept_work(struct sock *sk)
{
	struct net *net = sock_net(sk);
	struct rds_tcp_net *rtn = net_generic(net, rds_tcp_netid);

	queue_work(rds_wq, &rtn->rds_tcp_accept_w);
}

static __net_init int rds_tcp_init_net(struct net *net)
{
	struct rds_tcp_net *rtn = net_generic(net, rds_tcp_netid);
	struct ctl_table *tbl;
	int err = 0;

	memset(rtn, 0, sizeof(*rtn));

	/* {snd, rcv}buf_size default to 0, which implies we let the
	 * stack pick the value, and permit auto-tuning of buffer size.
	 */
	if (net == &init_net) {
		tbl = rds_tcp_sysctl_table;
	} else {
		tbl = kmemdup(rds_tcp_sysctl_table,
			      sizeof(rds_tcp_sysctl_table), GFP_KERNEL);
		if (!tbl) {
			pr_warn("could not set allocate syctl table\n");
			return -ENOMEM;
		}
		rtn->ctl_table = tbl;
	}
	tbl[RDS_TCP_SNDBUF].data = &rtn->sndbuf_size;
	tbl[RDS_TCP_RCVBUF].data = &rtn->rcvbuf_size;
	rtn->rds_tcp_sysctl = register_net_sysctl(net, "net/rds/tcp", tbl);
	if (!rtn->rds_tcp_sysctl) {
		pr_warn("could not register sysctl\n");
		err = -ENOMEM;
		goto fail;
	}
	rtn->rds_tcp_listen_sock = rds_tcp_listen_init(net);
	if (!rtn->rds_tcp_listen_sock) {
		pr_warn("could not set up listen sock\n");
		unregister_net_sysctl_table(rtn->rds_tcp_sysctl);
		rtn->rds_tcp_sysctl = NULL;
		err = -EAFNOSUPPORT;
		goto fail;
	}
	INIT_WORK(&rtn->rds_tcp_accept_w, rds_tcp_accept_worker);
	return 0;

fail:
	if (net != &init_net)
		kfree(tbl);
	return err;
}

static void __net_exit rds_tcp_exit_net(struct net *net)
{
	struct rds_tcp_net *rtn = net_generic(net, rds_tcp_netid);

	if (rtn->rds_tcp_sysctl)
		unregister_net_sysctl_table(rtn->rds_tcp_sysctl);

	if (net != &init_net && rtn->ctl_table)
		kfree(rtn->ctl_table);

	/* If rds_tcp_exit_net() is called as a result of netns deletion,
	 * the rds_tcp_kill_sock() device notifier would already have cleaned
	 * up the listen socket, thus there is no work to do in this function.
	 *
	 * If rds_tcp_exit_net() is called as a result of module unload,
	 * i.e., due to rds_tcp_exit() -> unregister_pernet_subsys(), then
	 * we do need to clean up the listen socket here.
	 */
	if (rtn->rds_tcp_listen_sock) {
		struct socket *lsock = rtn->rds_tcp_listen_sock;

		rtn->rds_tcp_listen_sock = NULL;
		rds_tcp_listen_stop(lsock, &rtn->rds_tcp_accept_w);
	}
}

static struct pernet_operations rds_tcp_net_ops = {
	.init = rds_tcp_init_net,
	.exit = rds_tcp_exit_net,
	.id = &rds_tcp_netid,
	.size = sizeof(struct rds_tcp_net),
};

static void rds_tcp_kill_sock(struct net *net)
{
	struct rds_tcp_connection *tc, *_tc;
	LIST_HEAD(tmp_list);
	struct rds_tcp_net *rtn = net_generic(net, rds_tcp_netid);
	struct socket *lsock = rtn->rds_tcp_listen_sock;

	rtn->rds_tcp_listen_sock = NULL;
	rds_tcp_listen_stop(lsock, &rtn->rds_tcp_accept_w);
	spin_lock_irq(&rds_tcp_conn_lock);
	list_for_each_entry_safe(tc, _tc, &rds_tcp_conn_list, t_tcp_node) {
		struct net *c_net = read_pnet(&tc->t_cpath->cp_conn->c_net);

		if (net != c_net || !tc->t_sock)
			continue;
		if (!list_has_conn(&tmp_list, tc->t_cpath->cp_conn)) {
			list_move_tail(&tc->t_tcp_node, &tmp_list);
		} else {
			list_del(&tc->t_tcp_node);
			tc->t_tcp_node_detached = true;
		}
	}
	spin_unlock_irq(&rds_tcp_conn_lock);
	list_for_each_entry_safe(tc, _tc, &tmp_list, t_tcp_node)
		rds_conn_destroy(tc->t_cpath->cp_conn);
}

void *rds_tcp_listen_sock_def_readable(struct net *net)
{
	struct rds_tcp_net *rtn = net_generic(net, rds_tcp_netid);
	struct socket *lsock = rtn->rds_tcp_listen_sock;

	if (!lsock)
		return NULL;

	return lsock->sk->sk_user_data;
}

static int rds_tcp_dev_event(struct notifier_block *this,
			     unsigned long event, void *ptr)
{
	struct net_device *dev = netdev_notifier_info_to_dev(ptr);

	/* rds-tcp registers as a pernet subys, so the ->exit will only
	 * get invoked after network acitivity has quiesced. We need to
	 * clean up all sockets  to quiesce network activity, and use
	 * the unregistration of the per-net loopback device as a trigger
	 * to start that cleanup.
	 */
	if (event == NETDEV_UNREGISTER_FINAL &&
	    dev->ifindex == LOOPBACK_IFINDEX)
		rds_tcp_kill_sock(dev_net(dev));

	return NOTIFY_DONE;
}

static struct notifier_block rds_tcp_dev_notifier = {
	.notifier_call        = rds_tcp_dev_event,
	.priority = -10, /* must be called after other network notifiers */
};

/* when sysctl is used to modify some kernel socket parameters,this
 * function  resets the RDS connections in that netns  so that we can
 * restart with new parameters.  The assumption is that such reset
 * events are few and far-between.
 */
static void rds_tcp_sysctl_reset(struct net *net)
{
	struct rds_tcp_connection *tc, *_tc;

	spin_lock_irq(&rds_tcp_conn_lock);
	list_for_each_entry_safe(tc, _tc, &rds_tcp_conn_list, t_tcp_node) {
		struct net *c_net = read_pnet(&tc->t_cpath->cp_conn->c_net);

		if (net != c_net || !tc->t_sock)
			continue;

		/* reconnect with new parameters */
		rds_conn_path_drop(tc->t_cpath, false);
	}
	spin_unlock_irq(&rds_tcp_conn_lock);
}

static int rds_tcp_skbuf_handler(struct ctl_table *ctl, int write,
				 void __user *buffer, size_t *lenp,
				 loff_t *fpos)
{
	struct net *net = current->nsproxy->net_ns;
	int err;

	err = proc_dointvec_minmax(ctl, write, buffer, lenp, fpos);
	if (err < 0) {
		pr_warn("Invalid input. Must be >= %d\n",
			*(int *)(ctl->extra1));
		return err;
	}
	if (write)
		rds_tcp_sysctl_reset(net);
	return 0;
}

static void rds_tcp_exit(void)
{
	rds_info_deregister_func(RDS_INFO_TCP_SOCKETS, rds_tcp_tc_info);
	unregister_pernet_subsys(&rds_tcp_net_ops);
	if (unregister_netdevice_notifier(&rds_tcp_dev_notifier))
		pr_warn("could not unregister rds_tcp_dev_notifier\n");
	rds_tcp_destroy_conns();
	rds_trans_unregister(&rds_tcp_transport);
	rds_tcp_recv_exit();
	kmem_cache_destroy(rds_tcp_conn_slab);
}
module_exit(rds_tcp_exit);

static int rds_tcp_init(void)
{
	int ret;

	rds_tcp_conn_slab = kmem_cache_create("rds_tcp_connection",
					      sizeof(struct rds_tcp_connection),
					      0, 0, NULL);
	if (!rds_tcp_conn_slab) {
		ret = -ENOMEM;
		goto out;
	}

	ret = rds_tcp_recv_init();
	if (ret)
		goto out_slab;

	ret = register_pernet_subsys(&rds_tcp_net_ops);
	if (ret)
		goto out_recv;

	ret = register_netdevice_notifier(&rds_tcp_dev_notifier);
	if (ret) {
		pr_warn("could not register rds_tcp_dev_notifier\n");
		goto out_pernet;
	}

	rds_trans_register(&rds_tcp_transport);

	rds_info_register_func(RDS_INFO_TCP_SOCKETS, rds_tcp_tc_info);

	goto out;

out_pernet:
	unregister_pernet_subsys(&rds_tcp_net_ops);
out_recv:
	rds_tcp_recv_exit();
out_slab:
	kmem_cache_destroy(rds_tcp_conn_slab);
out:
	return ret;
}
module_init(rds_tcp_init);

MODULE_AUTHOR("Oracle Corporation <rds-devel@oss.oracle.com>");
MODULE_DESCRIPTION("RDS: TCP transport");
MODULE_LICENSE("Dual BSD/GPL");