Commit | Line | Data |
---|---|---|
fdda387f PC |
1 | /****************************************************************************** |
2 | ******************************************************************************* | |
3 | ** | |
4 | ** Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. | |
5e9ccc37 | 5 | ** Copyright (C) 2004-2009 Red Hat, Inc. All rights reserved. |
fdda387f PC |
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 | |
2cf12c0b | 24 | * be expanded for the cluster infrastructure then that is its |
fdda387f PC |
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 | * | |
2cf12c0b | 39 | * lowcomms will choose to use either TCP or SCTP as its transport layer |
6ed7257b | 40 | * depending on the configuration variable 'protocol'. This should be set |
2cf12c0b | 41 | * to 0 (default) for TCP or 1 for SCTP. It should be configured using a |
6ed7257b PC |
42 | * cluster-wide mechanism as it must be the same on all nodes of the cluster |
43 | * for the DLM to function. | |
fdda387f PC |
44 | * |
45 | */ | |
46 | ||
fdda387f PC |
47 | #include <asm/ioctls.h> |
48 | #include <net/sock.h> | |
49 | #include <net/tcp.h> | |
50 | #include <linux/pagemap.h> | |
6ed7257b | 51 | #include <linux/file.h> |
7a936ce7 | 52 | #include <linux/mutex.h> |
6ed7257b | 53 | #include <linux/sctp.h> |
5a0e3ad6 | 54 | #include <linux/slab.h> |
2f2d76cc | 55 | #include <net/sctp/sctp.h> |
44ad532b | 56 | #include <net/ipv6.h> |
fdda387f PC |
57 | |
58 | #include "dlm_internal.h" | |
59 | #include "lowcomms.h" | |
60 | #include "midcomms.h" | |
61 | #include "config.h" | |
62 | ||
6ed7257b | 63 | #define NEEDED_RMEM (4*1024*1024) |
5e9ccc37 | 64 | #define CONN_HASH_SIZE 32 |
6ed7257b | 65 | |
f92c8dd7 BP |
66 | /* Number of messages to send before rescheduling */ |
67 | #define MAX_SEND_MSG_COUNT 25 | |
68 | ||
fdda387f | 69 | struct cbuf { |
ac33d071 PC |
70 | unsigned int base; |
71 | unsigned int len; | |
72 | unsigned int mask; | |
fdda387f PC |
73 | }; |
74 | ||
ac33d071 PC |
75 | static void cbuf_add(struct cbuf *cb, int n) |
76 | { | |
77 | cb->len += n; | |
78 | } | |
fdda387f | 79 | |
ac33d071 PC |
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 | } | |
fdda387f | 102 | |
fdda387f PC |
103 | struct connection { |
104 | struct socket *sock; /* NULL if not connected */ | |
105 | uint32_t nodeid; /* So we know who we are in the list */ | |
f1f1c1cc | 106 | struct mutex sock_mutex; |
6ed7257b | 107 | unsigned long flags; |
fdda387f PC |
108 | #define CF_READ_PENDING 1 |
109 | #define CF_WRITE_PENDING 2 | |
110 | #define CF_CONNECT_PENDING 3 | |
6ed7257b PC |
111 | #define CF_INIT_PENDING 4 |
112 | #define CF_IS_OTHERCON 5 | |
063c4c99 | 113 | #define CF_CLOSE 6 |
b36930dd | 114 | #define CF_APP_LIMITED 7 |
ac33d071 | 115 | struct list_head writequeue; /* List of outgoing writequeue_entries */ |
fdda387f PC |
116 | spinlock_t writequeue_lock; |
117 | int (*rx_action) (struct connection *); /* What to do when active */ | |
6ed7257b | 118 | void (*connect_action) (struct connection *); /* What to do to connect */ |
fdda387f PC |
119 | struct page *rx_page; |
120 | struct cbuf cb; | |
121 | int retries; | |
fdda387f | 122 | #define MAX_CONNECT_RETRIES 3 |
5e9ccc37 | 123 | struct hlist_node list; |
fdda387f | 124 | struct connection *othercon; |
1d6e8131 PC |
125 | struct work_struct rwork; /* Receive workqueue */ |
126 | struct work_struct swork; /* Send workqueue */ | |
fdda387f PC |
127 | }; |
128 | #define sock2con(x) ((struct connection *)(x)->sk_user_data) | |
129 | ||
130 | /* An entry waiting to be sent */ | |
131 | struct writequeue_entry { | |
132 | struct list_head list; | |
133 | struct page *page; | |
134 | int offset; | |
135 | int len; | |
136 | int end; | |
137 | int users; | |
138 | struct connection *con; | |
139 | }; | |
140 | ||
36b71a8b DT |
141 | struct dlm_node_addr { |
142 | struct list_head list; | |
143 | int nodeid; | |
144 | int addr_count; | |
98e1b60e | 145 | int curr_addr_index; |
36b71a8b DT |
146 | struct sockaddr_storage *addr[DLM_MAX_ADDR_COUNT]; |
147 | }; | |
148 | ||
149 | static LIST_HEAD(dlm_node_addrs); | |
150 | static DEFINE_SPINLOCK(dlm_node_addrs_spin); | |
151 | ||
6ed7257b PC |
152 | static struct sockaddr_storage *dlm_local_addr[DLM_MAX_ADDR_COUNT]; |
153 | static int dlm_local_count; | |
513ef596 | 154 | static int dlm_allow_conn; |
fdda387f | 155 | |
1d6e8131 PC |
156 | /* Work queues */ |
157 | static struct workqueue_struct *recv_workqueue; | |
158 | static struct workqueue_struct *send_workqueue; | |
fdda387f | 159 | |
5e9ccc37 | 160 | static struct hlist_head connection_hash[CONN_HASH_SIZE]; |
7a936ce7 | 161 | static DEFINE_MUTEX(connections_lock); |
c80e7c83 | 162 | static struct kmem_cache *con_cache; |
fdda387f | 163 | |
1d6e8131 PC |
164 | static void process_recv_sockets(struct work_struct *work); |
165 | static void process_send_sockets(struct work_struct *work); | |
fdda387f | 166 | |
5e9ccc37 CC |
167 | |
168 | /* This is deliberately very simple because most clusters have simple | |
169 | sequential nodeids, so we should be able to go straight to a connection | |
170 | struct in the array */ | |
171 | static inline int nodeid_hash(int nodeid) | |
172 | { | |
173 | return nodeid & (CONN_HASH_SIZE-1); | |
174 | } | |
175 | ||
176 | static struct connection *__find_con(int nodeid) | |
177 | { | |
178 | int r; | |
5e9ccc37 CC |
179 | struct connection *con; |
180 | ||
181 | r = nodeid_hash(nodeid); | |
182 | ||
b67bfe0d | 183 | hlist_for_each_entry(con, &connection_hash[r], list) { |
5e9ccc37 CC |
184 | if (con->nodeid == nodeid) |
185 | return con; | |
186 | } | |
187 | return NULL; | |
188 | } | |
189 | ||
6ed7257b PC |
190 | /* |
191 | * If 'allocation' is zero then we don't attempt to create a new | |
192 | * connection structure for this node. | |
193 | */ | |
194 | static struct connection *__nodeid2con(int nodeid, gfp_t alloc) | |
fdda387f PC |
195 | { |
196 | struct connection *con = NULL; | |
6ed7257b | 197 | int r; |
fdda387f | 198 | |
5e9ccc37 | 199 | con = __find_con(nodeid); |
6ed7257b PC |
200 | if (con || !alloc) |
201 | return con; | |
fdda387f | 202 | |
6ed7257b PC |
203 | con = kmem_cache_zalloc(con_cache, alloc); |
204 | if (!con) | |
205 | return NULL; | |
fdda387f | 206 | |
5e9ccc37 CC |
207 | r = nodeid_hash(nodeid); |
208 | hlist_add_head(&con->list, &connection_hash[r]); | |
fdda387f | 209 | |
6ed7257b PC |
210 | con->nodeid = nodeid; |
211 | mutex_init(&con->sock_mutex); | |
212 | INIT_LIST_HEAD(&con->writequeue); | |
213 | spin_lock_init(&con->writequeue_lock); | |
214 | INIT_WORK(&con->swork, process_send_sockets); | |
215 | INIT_WORK(&con->rwork, process_recv_sockets); | |
fdda387f | 216 | |
6ed7257b PC |
217 | /* Setup action pointers for child sockets */ |
218 | if (con->nodeid) { | |
5e9ccc37 | 219 | struct connection *zerocon = __find_con(0); |
fdda387f | 220 | |
6ed7257b PC |
221 | con->connect_action = zerocon->connect_action; |
222 | if (!con->rx_action) | |
223 | con->rx_action = zerocon->rx_action; | |
fdda387f PC |
224 | } |
225 | ||
6ed7257b PC |
226 | return con; |
227 | } | |
228 | ||
5e9ccc37 CC |
229 | /* Loop round all connections */ |
230 | static void foreach_conn(void (*conn_func)(struct connection *c)) | |
231 | { | |
232 | int i; | |
b67bfe0d | 233 | struct hlist_node *n; |
5e9ccc37 CC |
234 | struct connection *con; |
235 | ||
236 | for (i = 0; i < CONN_HASH_SIZE; i++) { | |
b67bfe0d | 237 | hlist_for_each_entry_safe(con, n, &connection_hash[i], list) |
5e9ccc37 | 238 | conn_func(con); |
5e9ccc37 CC |
239 | } |
240 | } | |
241 | ||
6ed7257b PC |
242 | static struct connection *nodeid2con(int nodeid, gfp_t allocation) |
243 | { | |
244 | struct connection *con; | |
245 | ||
7a936ce7 | 246 | mutex_lock(&connections_lock); |
6ed7257b | 247 | con = __nodeid2con(nodeid, allocation); |
7a936ce7 | 248 | mutex_unlock(&connections_lock); |
6ed7257b | 249 | |
fdda387f PC |
250 | return con; |
251 | } | |
252 | ||
36b71a8b DT |
253 | static struct dlm_node_addr *find_node_addr(int nodeid) |
254 | { | |
255 | struct dlm_node_addr *na; | |
256 | ||
257 | list_for_each_entry(na, &dlm_node_addrs, list) { | |
258 | if (na->nodeid == nodeid) | |
259 | return na; | |
260 | } | |
261 | return NULL; | |
262 | } | |
263 | ||
264 | static int addr_compare(struct sockaddr_storage *x, struct sockaddr_storage *y) | |
6ed7257b | 265 | { |
36b71a8b DT |
266 | switch (x->ss_family) { |
267 | case AF_INET: { | |
268 | struct sockaddr_in *sinx = (struct sockaddr_in *)x; | |
269 | struct sockaddr_in *siny = (struct sockaddr_in *)y; | |
270 | if (sinx->sin_addr.s_addr != siny->sin_addr.s_addr) | |
271 | return 0; | |
272 | if (sinx->sin_port != siny->sin_port) | |
273 | return 0; | |
274 | break; | |
275 | } | |
276 | case AF_INET6: { | |
277 | struct sockaddr_in6 *sinx = (struct sockaddr_in6 *)x; | |
278 | struct sockaddr_in6 *siny = (struct sockaddr_in6 *)y; | |
279 | if (!ipv6_addr_equal(&sinx->sin6_addr, &siny->sin6_addr)) | |
280 | return 0; | |
281 | if (sinx->sin6_port != siny->sin6_port) | |
282 | return 0; | |
283 | break; | |
284 | } | |
285 | default: | |
286 | return 0; | |
287 | } | |
288 | return 1; | |
289 | } | |
290 | ||
291 | static int nodeid_to_addr(int nodeid, struct sockaddr_storage *sas_out, | |
98e1b60e | 292 | struct sockaddr *sa_out, bool try_new_addr) |
36b71a8b DT |
293 | { |
294 | struct sockaddr_storage sas; | |
295 | struct dlm_node_addr *na; | |
6ed7257b PC |
296 | |
297 | if (!dlm_local_count) | |
298 | return -1; | |
299 | ||
36b71a8b DT |
300 | spin_lock(&dlm_node_addrs_spin); |
301 | na = find_node_addr(nodeid); | |
98e1b60e | 302 | if (na && na->addr_count) { |
ee44b4bc MRL |
303 | memcpy(&sas, na->addr[na->curr_addr_index], |
304 | sizeof(struct sockaddr_storage)); | |
305 | ||
98e1b60e MC |
306 | if (try_new_addr) { |
307 | na->curr_addr_index++; | |
308 | if (na->curr_addr_index == na->addr_count) | |
309 | na->curr_addr_index = 0; | |
310 | } | |
98e1b60e | 311 | } |
36b71a8b DT |
312 | spin_unlock(&dlm_node_addrs_spin); |
313 | ||
314 | if (!na) | |
315 | return -EEXIST; | |
316 | ||
317 | if (!na->addr_count) | |
318 | return -ENOENT; | |
319 | ||
320 | if (sas_out) | |
321 | memcpy(sas_out, &sas, sizeof(struct sockaddr_storage)); | |
322 | ||
323 | if (!sa_out) | |
324 | return 0; | |
6ed7257b PC |
325 | |
326 | if (dlm_local_addr[0]->ss_family == AF_INET) { | |
36b71a8b DT |
327 | struct sockaddr_in *in4 = (struct sockaddr_in *) &sas; |
328 | struct sockaddr_in *ret4 = (struct sockaddr_in *) sa_out; | |
6ed7257b PC |
329 | ret4->sin_addr.s_addr = in4->sin_addr.s_addr; |
330 | } else { | |
36b71a8b DT |
331 | struct sockaddr_in6 *in6 = (struct sockaddr_in6 *) &sas; |
332 | struct sockaddr_in6 *ret6 = (struct sockaddr_in6 *) sa_out; | |
4e3fd7a0 | 333 | ret6->sin6_addr = in6->sin6_addr; |
6ed7257b PC |
334 | } |
335 | ||
336 | return 0; | |
337 | } | |
338 | ||
36b71a8b DT |
339 | static int addr_to_nodeid(struct sockaddr_storage *addr, int *nodeid) |
340 | { | |
341 | struct dlm_node_addr *na; | |
342 | int rv = -EEXIST; | |
98e1b60e | 343 | int addr_i; |
36b71a8b DT |
344 | |
345 | spin_lock(&dlm_node_addrs_spin); | |
346 | list_for_each_entry(na, &dlm_node_addrs, list) { | |
347 | if (!na->addr_count) | |
348 | continue; | |
349 | ||
98e1b60e MC |
350 | for (addr_i = 0; addr_i < na->addr_count; addr_i++) { |
351 | if (addr_compare(na->addr[addr_i], addr)) { | |
352 | *nodeid = na->nodeid; | |
353 | rv = 0; | |
354 | goto unlock; | |
355 | } | |
356 | } | |
36b71a8b | 357 | } |
98e1b60e | 358 | unlock: |
36b71a8b DT |
359 | spin_unlock(&dlm_node_addrs_spin); |
360 | return rv; | |
361 | } | |
362 | ||
363 | int dlm_lowcomms_addr(int nodeid, struct sockaddr_storage *addr, int len) | |
364 | { | |
365 | struct sockaddr_storage *new_addr; | |
366 | struct dlm_node_addr *new_node, *na; | |
367 | ||
368 | new_node = kzalloc(sizeof(struct dlm_node_addr), GFP_NOFS); | |
369 | if (!new_node) | |
370 | return -ENOMEM; | |
371 | ||
372 | new_addr = kzalloc(sizeof(struct sockaddr_storage), GFP_NOFS); | |
373 | if (!new_addr) { | |
374 | kfree(new_node); | |
375 | return -ENOMEM; | |
376 | } | |
377 | ||
378 | memcpy(new_addr, addr, len); | |
379 | ||
380 | spin_lock(&dlm_node_addrs_spin); | |
381 | na = find_node_addr(nodeid); | |
382 | if (!na) { | |
383 | new_node->nodeid = nodeid; | |
384 | new_node->addr[0] = new_addr; | |
385 | new_node->addr_count = 1; | |
386 | list_add(&new_node->list, &dlm_node_addrs); | |
387 | spin_unlock(&dlm_node_addrs_spin); | |
388 | return 0; | |
389 | } | |
390 | ||
391 | if (na->addr_count >= DLM_MAX_ADDR_COUNT) { | |
392 | spin_unlock(&dlm_node_addrs_spin); | |
393 | kfree(new_addr); | |
394 | kfree(new_node); | |
395 | return -ENOSPC; | |
396 | } | |
397 | ||
398 | na->addr[na->addr_count++] = new_addr; | |
399 | spin_unlock(&dlm_node_addrs_spin); | |
400 | kfree(new_node); | |
401 | return 0; | |
402 | } | |
403 | ||
fdda387f | 404 | /* Data available on socket or listen socket received a connect */ |
676d2369 | 405 | static void lowcomms_data_ready(struct sock *sk) |
fdda387f PC |
406 | { |
407 | struct connection *con = sock2con(sk); | |
afb853fb | 408 | if (con && !test_and_set_bit(CF_READ_PENDING, &con->flags)) |
1d6e8131 | 409 | queue_work(recv_workqueue, &con->rwork); |
fdda387f PC |
410 | } |
411 | ||
412 | static void lowcomms_write_space(struct sock *sk) | |
413 | { | |
414 | struct connection *con = sock2con(sk); | |
415 | ||
b36930dd DM |
416 | if (!con) |
417 | return; | |
418 | ||
419 | clear_bit(SOCK_NOSPACE, &con->sock->flags); | |
420 | ||
421 | if (test_and_clear_bit(CF_APP_LIMITED, &con->flags)) { | |
422 | con->sock->sk->sk_write_pending--; | |
423 | clear_bit(SOCK_ASYNC_NOSPACE, &con->sock->flags); | |
424 | } | |
425 | ||
426 | if (!test_and_set_bit(CF_WRITE_PENDING, &con->flags)) | |
1d6e8131 | 427 | queue_work(send_workqueue, &con->swork); |
fdda387f PC |
428 | } |
429 | ||
430 | static inline void lowcomms_connect_sock(struct connection *con) | |
431 | { | |
063c4c99 LMB |
432 | if (test_bit(CF_CLOSE, &con->flags)) |
433 | return; | |
1d6e8131 PC |
434 | if (!test_and_set_bit(CF_CONNECT_PENDING, &con->flags)) |
435 | queue_work(send_workqueue, &con->swork); | |
fdda387f PC |
436 | } |
437 | ||
438 | static void lowcomms_state_change(struct sock *sk) | |
439 | { | |
ee44b4bc MRL |
440 | /* SCTP layer is not calling sk_data_ready when the connection |
441 | * is done, so we catch the signal through here. Also, it | |
442 | * doesn't switch socket state when entering shutdown, so we | |
443 | * skip the write in that case. | |
444 | */ | |
445 | if (sk->sk_shutdown) { | |
446 | if (sk->sk_shutdown == RCV_SHUTDOWN) | |
447 | lowcomms_data_ready(sk); | |
448 | } else if (sk->sk_state == TCP_ESTABLISHED) { | |
fdda387f | 449 | lowcomms_write_space(sk); |
ee44b4bc | 450 | } |
fdda387f PC |
451 | } |
452 | ||
391fbdc5 CC |
453 | int dlm_lowcomms_connect_node(int nodeid) |
454 | { | |
455 | struct connection *con; | |
456 | ||
457 | if (nodeid == dlm_our_nodeid()) | |
458 | return 0; | |
459 | ||
460 | con = nodeid2con(nodeid, GFP_NOFS); | |
461 | if (!con) | |
462 | return -ENOMEM; | |
463 | lowcomms_connect_sock(con); | |
464 | return 0; | |
465 | } | |
466 | ||
fdda387f | 467 | /* Make a socket active */ |
4dd40f0c | 468 | static void add_sock(struct socket *sock, struct connection *con) |
fdda387f PC |
469 | { |
470 | con->sock = sock; | |
471 | ||
472 | /* Install a data_ready callback */ | |
473 | con->sock->sk->sk_data_ready = lowcomms_data_ready; | |
474 | con->sock->sk->sk_write_space = lowcomms_write_space; | |
475 | con->sock->sk->sk_state_change = lowcomms_state_change; | |
6ed7257b | 476 | con->sock->sk->sk_user_data = con; |
d6d7b702 | 477 | con->sock->sk->sk_allocation = GFP_NOFS; |
fdda387f PC |
478 | } |
479 | ||
6ed7257b | 480 | /* Add the port number to an IPv6 or 4 sockaddr and return the address |
fdda387f PC |
481 | length */ |
482 | static void make_sockaddr(struct sockaddr_storage *saddr, uint16_t port, | |
483 | int *addr_len) | |
484 | { | |
6ed7257b | 485 | saddr->ss_family = dlm_local_addr[0]->ss_family; |
ac33d071 | 486 | if (saddr->ss_family == AF_INET) { |
fdda387f PC |
487 | struct sockaddr_in *in4_addr = (struct sockaddr_in *)saddr; |
488 | in4_addr->sin_port = cpu_to_be16(port); | |
489 | *addr_len = sizeof(struct sockaddr_in); | |
6ed7257b | 490 | memset(&in4_addr->sin_zero, 0, sizeof(in4_addr->sin_zero)); |
ac33d071 | 491 | } else { |
fdda387f PC |
492 | struct sockaddr_in6 *in6_addr = (struct sockaddr_in6 *)saddr; |
493 | in6_addr->sin6_port = cpu_to_be16(port); | |
494 | *addr_len = sizeof(struct sockaddr_in6); | |
495 | } | |
01c8cab2 | 496 | memset((char *)saddr + *addr_len, 0, sizeof(struct sockaddr_storage) - *addr_len); |
fdda387f PC |
497 | } |
498 | ||
499 | /* Close a remote connection and tidy up */ | |
0d737a8c MRL |
500 | static void close_connection(struct connection *con, bool and_other, |
501 | bool tx, bool rx) | |
fdda387f | 502 | { |
0d737a8c MRL |
503 | clear_bit(CF_CONNECT_PENDING, &con->flags); |
504 | clear_bit(CF_WRITE_PENDING, &con->flags); | |
505 | if (tx && cancel_work_sync(&con->swork)) | |
506 | log_print("canceled swork for node %d", con->nodeid); | |
507 | if (rx && cancel_work_sync(&con->rwork)) | |
508 | log_print("canceled rwork for node %d", con->nodeid); | |
fdda387f | 509 | |
0d737a8c | 510 | mutex_lock(&con->sock_mutex); |
fdda387f PC |
511 | if (con->sock) { |
512 | sock_release(con->sock); | |
513 | con->sock = NULL; | |
514 | } | |
515 | if (con->othercon && and_other) { | |
ac33d071 | 516 | /* Will only re-enter once. */ |
0d737a8c | 517 | close_connection(con->othercon, false, true, true); |
fdda387f PC |
518 | } |
519 | if (con->rx_page) { | |
520 | __free_page(con->rx_page); | |
521 | con->rx_page = NULL; | |
522 | } | |
9e5f2825 | 523 | |
61d96be0 PC |
524 | con->retries = 0; |
525 | mutex_unlock(&con->sock_mutex); | |
fdda387f PC |
526 | } |
527 | ||
528 | /* Data received from remote end */ | |
529 | static int receive_from_sock(struct connection *con) | |
530 | { | |
531 | int ret = 0; | |
58addbff AV |
532 | struct msghdr msg = {}; |
533 | struct kvec iov[2]; | |
fdda387f PC |
534 | unsigned len; |
535 | int r; | |
536 | int call_again_soon = 0; | |
58addbff | 537 | int nvec; |
fdda387f | 538 | |
f1f1c1cc | 539 | mutex_lock(&con->sock_mutex); |
fdda387f | 540 | |
a34fbc63 PC |
541 | if (con->sock == NULL) { |
542 | ret = -EAGAIN; | |
543 | goto out_close; | |
544 | } | |
acee4e52 MRL |
545 | if (con->nodeid == 0) { |
546 | ret = -EINVAL; | |
547 | goto out_close; | |
548 | } | |
a34fbc63 | 549 | |
fdda387f PC |
550 | if (con->rx_page == NULL) { |
551 | /* | |
552 | * This doesn't need to be atomic, but I think it should | |
553 | * improve performance if it is. | |
554 | */ | |
555 | con->rx_page = alloc_page(GFP_ATOMIC); | |
556 | if (con->rx_page == NULL) | |
557 | goto out_resched; | |
ac33d071 | 558 | cbuf_init(&con->cb, PAGE_CACHE_SIZE); |
fdda387f PC |
559 | } |
560 | ||
fdda387f PC |
561 | /* |
562 | * iov[0] is the bit of the circular buffer between the current end | |
563 | * point (cb.base + cb.len) and the end of the buffer. | |
564 | */ | |
ac33d071 PC |
565 | iov[0].iov_len = con->cb.base - cbuf_data(&con->cb); |
566 | iov[0].iov_base = page_address(con->rx_page) + cbuf_data(&con->cb); | |
89adc934 | 567 | iov[1].iov_len = 0; |
58addbff | 568 | nvec = 1; |
fdda387f PC |
569 | |
570 | /* | |
571 | * iov[1] is the bit of the circular buffer between the start of the | |
572 | * buffer and the start of the currently used section (cb.base) | |
573 | */ | |
ac33d071 PC |
574 | if (cbuf_data(&con->cb) >= con->cb.base) { |
575 | iov[0].iov_len = PAGE_CACHE_SIZE - cbuf_data(&con->cb); | |
fdda387f PC |
576 | iov[1].iov_len = con->cb.base; |
577 | iov[1].iov_base = page_address(con->rx_page); | |
58addbff | 578 | nvec = 2; |
fdda387f PC |
579 | } |
580 | len = iov[0].iov_len + iov[1].iov_len; | |
581 | ||
58addbff | 582 | r = ret = kernel_recvmsg(con->sock, &msg, iov, nvec, len, |
fdda387f | 583 | MSG_DONTWAIT | MSG_NOSIGNAL); |
fdda387f PC |
584 | if (ret <= 0) |
585 | goto out_close; | |
ee44b4bc MRL |
586 | else if (ret == len) |
587 | call_again_soon = 1; | |
bd44e2b0 | 588 | |
ac33d071 | 589 | cbuf_add(&con->cb, ret); |
fdda387f PC |
590 | ret = dlm_process_incoming_buffer(con->nodeid, |
591 | page_address(con->rx_page), | |
592 | con->cb.base, con->cb.len, | |
593 | PAGE_CACHE_SIZE); | |
594 | if (ret == -EBADMSG) { | |
ee44b4bc MRL |
595 | log_print("lowcomms: addr=%p, base=%u, len=%u, read=%d", |
596 | page_address(con->rx_page), con->cb.base, | |
597 | con->cb.len, r); | |
fdda387f PC |
598 | } |
599 | if (ret < 0) | |
600 | goto out_close; | |
ac33d071 | 601 | cbuf_eat(&con->cb, ret); |
fdda387f | 602 | |
ac33d071 | 603 | if (cbuf_empty(&con->cb) && !call_again_soon) { |
fdda387f PC |
604 | __free_page(con->rx_page); |
605 | con->rx_page = NULL; | |
606 | } | |
607 | ||
fdda387f PC |
608 | if (call_again_soon) |
609 | goto out_resched; | |
f1f1c1cc | 610 | mutex_unlock(&con->sock_mutex); |
ac33d071 | 611 | return 0; |
fdda387f | 612 | |
ac33d071 | 613 | out_resched: |
1d6e8131 PC |
614 | if (!test_and_set_bit(CF_READ_PENDING, &con->flags)) |
615 | queue_work(recv_workqueue, &con->rwork); | |
f1f1c1cc | 616 | mutex_unlock(&con->sock_mutex); |
bd44e2b0 | 617 | return -EAGAIN; |
fdda387f | 618 | |
ac33d071 | 619 | out_close: |
f1f1c1cc | 620 | mutex_unlock(&con->sock_mutex); |
9e5f2825 | 621 | if (ret != -EAGAIN) { |
0d737a8c | 622 | close_connection(con, false, true, false); |
fdda387f PC |
623 | /* Reconnect when there is something to send */ |
624 | } | |
a34fbc63 PC |
625 | /* Don't return success if we really got EOF */ |
626 | if (ret == 0) | |
627 | ret = -EAGAIN; | |
fdda387f | 628 | |
fdda387f PC |
629 | return ret; |
630 | } | |
631 | ||
632 | /* Listening socket is busy, accept a connection */ | |
6ed7257b | 633 | static int tcp_accept_from_sock(struct connection *con) |
fdda387f PC |
634 | { |
635 | int result; | |
636 | struct sockaddr_storage peeraddr; | |
637 | struct socket *newsock; | |
638 | int len; | |
639 | int nodeid; | |
640 | struct connection *newcon; | |
bd44e2b0 | 641 | struct connection *addcon; |
fdda387f | 642 | |
513ef596 DT |
643 | mutex_lock(&connections_lock); |
644 | if (!dlm_allow_conn) { | |
645 | mutex_unlock(&connections_lock); | |
646 | return -1; | |
647 | } | |
648 | mutex_unlock(&connections_lock); | |
649 | ||
fdda387f | 650 | memset(&peeraddr, 0, sizeof(peeraddr)); |
eeb1bd5c EB |
651 | result = sock_create_kern(&init_net, dlm_local_addr[0]->ss_family, |
652 | SOCK_STREAM, IPPROTO_TCP, &newsock); | |
fdda387f PC |
653 | if (result < 0) |
654 | return -ENOMEM; | |
655 | ||
f1f1c1cc | 656 | mutex_lock_nested(&con->sock_mutex, 0); |
fdda387f PC |
657 | |
658 | result = -ENOTCONN; | |
659 | if (con->sock == NULL) | |
660 | goto accept_err; | |
661 | ||
662 | newsock->type = con->sock->type; | |
663 | newsock->ops = con->sock->ops; | |
664 | ||
665 | result = con->sock->ops->accept(con->sock, newsock, O_NONBLOCK); | |
666 | if (result < 0) | |
667 | goto accept_err; | |
668 | ||
669 | /* Get the connected socket's peer */ | |
670 | memset(&peeraddr, 0, sizeof(peeraddr)); | |
671 | if (newsock->ops->getname(newsock, (struct sockaddr *)&peeraddr, | |
672 | &len, 2)) { | |
673 | result = -ECONNABORTED; | |
674 | goto accept_err; | |
675 | } | |
676 | ||
677 | /* Get the new node's NODEID */ | |
678 | make_sockaddr(&peeraddr, 0, &len); | |
36b71a8b | 679 | if (addr_to_nodeid(&peeraddr, &nodeid)) { |
bcaadf5c | 680 | unsigned char *b=(unsigned char *)&peeraddr; |
617e82e1 | 681 | log_print("connect from non cluster node"); |
bcaadf5c MY |
682 | print_hex_dump_bytes("ss: ", DUMP_PREFIX_NONE, |
683 | b, sizeof(struct sockaddr_storage)); | |
fdda387f | 684 | sock_release(newsock); |
f1f1c1cc | 685 | mutex_unlock(&con->sock_mutex); |
fdda387f PC |
686 | return -1; |
687 | } | |
688 | ||
689 | log_print("got connection from %d", nodeid); | |
690 | ||
691 | /* Check to see if we already have a connection to this node. This | |
692 | * could happen if the two nodes initiate a connection at roughly | |
693 | * the same time and the connections cross on the wire. | |
fdda387f PC |
694 | * In this case we store the incoming one in "othercon" |
695 | */ | |
748285cc | 696 | newcon = nodeid2con(nodeid, GFP_NOFS); |
fdda387f PC |
697 | if (!newcon) { |
698 | result = -ENOMEM; | |
699 | goto accept_err; | |
700 | } | |
f1f1c1cc | 701 | mutex_lock_nested(&newcon->sock_mutex, 1); |
fdda387f | 702 | if (newcon->sock) { |
ac33d071 | 703 | struct connection *othercon = newcon->othercon; |
fdda387f PC |
704 | |
705 | if (!othercon) { | |
748285cc | 706 | othercon = kmem_cache_zalloc(con_cache, GFP_NOFS); |
fdda387f | 707 | if (!othercon) { |
617e82e1 | 708 | log_print("failed to allocate incoming socket"); |
f1f1c1cc | 709 | mutex_unlock(&newcon->sock_mutex); |
fdda387f PC |
710 | result = -ENOMEM; |
711 | goto accept_err; | |
712 | } | |
fdda387f PC |
713 | othercon->nodeid = nodeid; |
714 | othercon->rx_action = receive_from_sock; | |
f1f1c1cc | 715 | mutex_init(&othercon->sock_mutex); |
1d6e8131 PC |
716 | INIT_WORK(&othercon->swork, process_send_sockets); |
717 | INIT_WORK(&othercon->rwork, process_recv_sockets); | |
fdda387f | 718 | set_bit(CF_IS_OTHERCON, &othercon->flags); |
61d96be0 PC |
719 | } |
720 | if (!othercon->sock) { | |
fdda387f | 721 | newcon->othercon = othercon; |
97d84836 PC |
722 | othercon->sock = newsock; |
723 | newsock->sk->sk_user_data = othercon; | |
724 | add_sock(newsock, othercon); | |
725 | addcon = othercon; | |
726 | } | |
727 | else { | |
728 | printk("Extra connection from node %d attempted\n", nodeid); | |
729 | result = -EAGAIN; | |
f4fadb23 | 730 | mutex_unlock(&newcon->sock_mutex); |
97d84836 | 731 | goto accept_err; |
fdda387f | 732 | } |
fdda387f PC |
733 | } |
734 | else { | |
735 | newsock->sk->sk_user_data = newcon; | |
736 | newcon->rx_action = receive_from_sock; | |
737 | add_sock(newsock, newcon); | |
bd44e2b0 | 738 | addcon = newcon; |
fdda387f PC |
739 | } |
740 | ||
f1f1c1cc | 741 | mutex_unlock(&newcon->sock_mutex); |
fdda387f PC |
742 | |
743 | /* | |
744 | * Add it to the active queue in case we got data | |
25985edc | 745 | * between processing the accept adding the socket |
fdda387f PC |
746 | * to the read_sockets list |
747 | */ | |
bd44e2b0 PC |
748 | if (!test_and_set_bit(CF_READ_PENDING, &addcon->flags)) |
749 | queue_work(recv_workqueue, &addcon->rwork); | |
f1f1c1cc | 750 | mutex_unlock(&con->sock_mutex); |
fdda387f PC |
751 | |
752 | return 0; | |
753 | ||
ac33d071 | 754 | accept_err: |
f1f1c1cc | 755 | mutex_unlock(&con->sock_mutex); |
fdda387f PC |
756 | sock_release(newsock); |
757 | ||
758 | if (result != -EAGAIN) | |
617e82e1 | 759 | log_print("error accepting connection from node: %d", result); |
fdda387f PC |
760 | return result; |
761 | } | |
762 | ||
18df8a87 | 763 | static int sctp_accept_from_sock(struct connection *con) |
ee44b4bc MRL |
764 | { |
765 | /* Check that the new node is in the lockspace */ | |
766 | struct sctp_prim prim; | |
767 | int nodeid; | |
768 | int prim_len, ret; | |
769 | int addr_len; | |
770 | struct connection *newcon; | |
771 | struct connection *addcon; | |
772 | struct socket *newsock; | |
773 | ||
774 | mutex_lock(&connections_lock); | |
775 | if (!dlm_allow_conn) { | |
776 | mutex_unlock(&connections_lock); | |
777 | return -1; | |
778 | } | |
779 | mutex_unlock(&connections_lock); | |
780 | ||
781 | mutex_lock_nested(&con->sock_mutex, 0); | |
782 | ||
783 | ret = kernel_accept(con->sock, &newsock, O_NONBLOCK); | |
784 | if (ret < 0) | |
785 | goto accept_err; | |
786 | ||
787 | memset(&prim, 0, sizeof(struct sctp_prim)); | |
788 | prim_len = sizeof(struct sctp_prim); | |
789 | ||
790 | ret = kernel_getsockopt(newsock, IPPROTO_SCTP, SCTP_PRIMARY_ADDR, | |
791 | (char *)&prim, &prim_len); | |
792 | if (ret < 0) { | |
793 | log_print("getsockopt/sctp_primary_addr failed: %d", ret); | |
794 | goto accept_err; | |
795 | } | |
796 | ||
797 | make_sockaddr(&prim.ssp_addr, 0, &addr_len); | |
798 | if (addr_to_nodeid(&prim.ssp_addr, &nodeid)) { | |
799 | unsigned char *b = (unsigned char *)&prim.ssp_addr; | |
800 | ||
801 | log_print("reject connect from unknown addr"); | |
802 | print_hex_dump_bytes("ss: ", DUMP_PREFIX_NONE, | |
803 | b, sizeof(struct sockaddr_storage)); | |
804 | goto accept_err; | |
805 | } | |
806 | ||
807 | newcon = nodeid2con(nodeid, GFP_NOFS); | |
808 | if (!newcon) { | |
809 | ret = -ENOMEM; | |
810 | goto accept_err; | |
811 | } | |
812 | ||
813 | mutex_lock_nested(&newcon->sock_mutex, 1); | |
814 | ||
815 | if (newcon->sock) { | |
816 | struct connection *othercon = newcon->othercon; | |
817 | ||
818 | if (!othercon) { | |
819 | othercon = kmem_cache_zalloc(con_cache, GFP_NOFS); | |
820 | if (!othercon) { | |
821 | log_print("failed to allocate incoming socket"); | |
822 | mutex_unlock(&newcon->sock_mutex); | |
823 | ret = -ENOMEM; | |
824 | goto accept_err; | |
825 | } | |
826 | othercon->nodeid = nodeid; | |
827 | othercon->rx_action = receive_from_sock; | |
828 | mutex_init(&othercon->sock_mutex); | |
829 | INIT_WORK(&othercon->swork, process_send_sockets); | |
830 | INIT_WORK(&othercon->rwork, process_recv_sockets); | |
831 | set_bit(CF_IS_OTHERCON, &othercon->flags); | |
832 | } | |
833 | if (!othercon->sock) { | |
834 | newcon->othercon = othercon; | |
835 | othercon->sock = newsock; | |
836 | newsock->sk->sk_user_data = othercon; | |
837 | add_sock(newsock, othercon); | |
838 | addcon = othercon; | |
839 | } else { | |
840 | printk("Extra connection from node %d attempted\n", nodeid); | |
841 | ret = -EAGAIN; | |
842 | mutex_unlock(&newcon->sock_mutex); | |
843 | goto accept_err; | |
844 | } | |
845 | } else { | |
846 | newsock->sk->sk_user_data = newcon; | |
847 | newcon->rx_action = receive_from_sock; | |
848 | add_sock(newsock, newcon); | |
849 | addcon = newcon; | |
850 | } | |
851 | ||
852 | log_print("connected to %d", nodeid); | |
853 | ||
854 | mutex_unlock(&newcon->sock_mutex); | |
855 | ||
856 | /* | |
857 | * Add it to the active queue in case we got data | |
858 | * between processing the accept adding the socket | |
859 | * to the read_sockets list | |
860 | */ | |
861 | if (!test_and_set_bit(CF_READ_PENDING, &addcon->flags)) | |
862 | queue_work(recv_workqueue, &addcon->rwork); | |
863 | mutex_unlock(&con->sock_mutex); | |
864 | ||
865 | return 0; | |
866 | ||
867 | accept_err: | |
868 | mutex_unlock(&con->sock_mutex); | |
869 | if (newsock) | |
870 | sock_release(newsock); | |
871 | if (ret != -EAGAIN) | |
872 | log_print("error accepting connection from node: %d", ret); | |
873 | ||
874 | return ret; | |
875 | } | |
876 | ||
6ed7257b PC |
877 | static void free_entry(struct writequeue_entry *e) |
878 | { | |
879 | __free_page(e->page); | |
880 | kfree(e); | |
881 | } | |
882 | ||
5d689871 MC |
883 | /* |
884 | * writequeue_entry_complete - try to delete and free write queue entry | |
885 | * @e: write queue entry to try to delete | |
886 | * @completed: bytes completed | |
887 | * | |
888 | * writequeue_lock must be held. | |
889 | */ | |
890 | static void writequeue_entry_complete(struct writequeue_entry *e, int completed) | |
891 | { | |
892 | e->offset += completed; | |
893 | e->len -= completed; | |
894 | ||
895 | if (e->len == 0 && e->users == 0) { | |
896 | list_del(&e->list); | |
897 | free_entry(e); | |
898 | } | |
899 | } | |
900 | ||
ee44b4bc MRL |
901 | /* |
902 | * sctp_bind_addrs - bind a SCTP socket to all our addresses | |
903 | */ | |
904 | static int sctp_bind_addrs(struct connection *con, uint16_t port) | |
905 | { | |
906 | struct sockaddr_storage localaddr; | |
907 | int i, addr_len, result = 0; | |
908 | ||
909 | for (i = 0; i < dlm_local_count; i++) { | |
910 | memcpy(&localaddr, dlm_local_addr[i], sizeof(localaddr)); | |
911 | make_sockaddr(&localaddr, port, &addr_len); | |
912 | ||
913 | if (!i) | |
914 | result = kernel_bind(con->sock, | |
915 | (struct sockaddr *)&localaddr, | |
916 | addr_len); | |
917 | else | |
918 | result = kernel_setsockopt(con->sock, SOL_SCTP, | |
919 | SCTP_SOCKOPT_BINDX_ADD, | |
920 | (char *)&localaddr, addr_len); | |
921 | ||
922 | if (result < 0) { | |
923 | log_print("Can't bind to %d addr number %d, %d.\n", | |
924 | port, i + 1, result); | |
925 | break; | |
926 | } | |
927 | } | |
928 | return result; | |
929 | } | |
930 | ||
6ed7257b PC |
931 | /* Initiate an SCTP association. |
932 | This is a special case of send_to_sock() in that we don't yet have a | |
933 | peeled-off socket for this association, so we use the listening socket | |
934 | and add the primary IP address of the remote node. | |
935 | */ | |
ee44b4bc | 936 | static void sctp_connect_to_sock(struct connection *con) |
6ed7257b | 937 | { |
ee44b4bc MRL |
938 | struct sockaddr_storage daddr; |
939 | int one = 1; | |
940 | int result; | |
941 | int addr_len; | |
942 | struct socket *sock; | |
943 | ||
944 | if (con->nodeid == 0) { | |
945 | log_print("attempt to connect sock 0 foiled"); | |
946 | return; | |
947 | } | |
6ed7257b | 948 | |
5d689871 | 949 | mutex_lock(&con->sock_mutex); |
6ed7257b | 950 | |
ee44b4bc MRL |
951 | /* Some odd races can cause double-connects, ignore them */ |
952 | if (con->retries++ > MAX_CONNECT_RETRIES) | |
953 | goto out; | |
954 | ||
955 | if (con->sock) { | |
956 | log_print("node %d already connected.", con->nodeid); | |
957 | goto out; | |
958 | } | |
959 | ||
960 | memset(&daddr, 0, sizeof(daddr)); | |
961 | result = nodeid_to_addr(con->nodeid, &daddr, NULL, true); | |
962 | if (result < 0) { | |
6ed7257b | 963 | log_print("no address for nodeid %d", con->nodeid); |
ee44b4bc | 964 | goto out; |
6ed7257b | 965 | } |
6ed7257b | 966 | |
ee44b4bc MRL |
967 | /* Create a socket to communicate with */ |
968 | result = sock_create_kern(&init_net, dlm_local_addr[0]->ss_family, | |
969 | SOCK_STREAM, IPPROTO_SCTP, &sock); | |
970 | if (result < 0) | |
971 | goto socket_err; | |
6ed7257b | 972 | |
ee44b4bc MRL |
973 | sock->sk->sk_user_data = con; |
974 | con->rx_action = receive_from_sock; | |
975 | con->connect_action = sctp_connect_to_sock; | |
976 | add_sock(sock, con); | |
6ed7257b | 977 | |
ee44b4bc MRL |
978 | /* Bind to all addresses. */ |
979 | if (sctp_bind_addrs(con, 0)) | |
980 | goto bind_err; | |
6ed7257b | 981 | |
ee44b4bc | 982 | make_sockaddr(&daddr, dlm_config.ci_tcp_port, &addr_len); |
6ed7257b | 983 | |
ee44b4bc | 984 | log_print("connecting to %d", con->nodeid); |
6ed7257b | 985 | |
ee44b4bc MRL |
986 | /* Turn off Nagle's algorithm */ |
987 | kernel_setsockopt(sock, SOL_TCP, TCP_NODELAY, (char *)&one, | |
988 | sizeof(one)); | |
6ed7257b | 989 | |
ee44b4bc MRL |
990 | result = sock->ops->connect(sock, (struct sockaddr *)&daddr, addr_len, |
991 | O_NONBLOCK); | |
992 | if (result == -EINPROGRESS) | |
993 | result = 0; | |
994 | if (result == 0) | |
995 | goto out; | |
98e1b60e | 996 | |
6ed7257b | 997 | |
ee44b4bc MRL |
998 | bind_err: |
999 | con->sock = NULL; | |
1000 | sock_release(sock); | |
6ed7257b | 1001 | |
ee44b4bc MRL |
1002 | socket_err: |
1003 | /* | |
1004 | * Some errors are fatal and this list might need adjusting. For other | |
1005 | * errors we try again until the max number of retries is reached. | |
1006 | */ | |
1007 | if (result != -EHOSTUNREACH && | |
1008 | result != -ENETUNREACH && | |
1009 | result != -ENETDOWN && | |
1010 | result != -EINVAL && | |
1011 | result != -EPROTONOSUPPORT) { | |
1012 | log_print("connect %d try %d error %d", con->nodeid, | |
1013 | con->retries, result); | |
1014 | mutex_unlock(&con->sock_mutex); | |
1015 | msleep(1000); | |
6ed7257b | 1016 | clear_bit(CF_CONNECT_PENDING, &con->flags); |
ee44b4bc MRL |
1017 | lowcomms_connect_sock(con); |
1018 | return; | |
6ed7257b | 1019 | } |
5d689871 | 1020 | |
ee44b4bc | 1021 | out: |
5d689871 | 1022 | mutex_unlock(&con->sock_mutex); |
00dcffae | 1023 | set_bit(CF_WRITE_PENDING, &con->flags); |
6ed7257b PC |
1024 | } |
1025 | ||
fdda387f | 1026 | /* Connect a new socket to its peer */ |
6ed7257b | 1027 | static void tcp_connect_to_sock(struct connection *con) |
fdda387f | 1028 | { |
6bd8feda | 1029 | struct sockaddr_storage saddr, src_addr; |
fdda387f | 1030 | int addr_len; |
a89d63a1 | 1031 | struct socket *sock = NULL; |
cb2d45da | 1032 | int one = 1; |
36b71a8b | 1033 | int result; |
fdda387f PC |
1034 | |
1035 | if (con->nodeid == 0) { | |
1036 | log_print("attempt to connect sock 0 foiled"); | |
ac33d071 | 1037 | return; |
fdda387f PC |
1038 | } |
1039 | ||
f1f1c1cc | 1040 | mutex_lock(&con->sock_mutex); |
fdda387f PC |
1041 | if (con->retries++ > MAX_CONNECT_RETRIES) |
1042 | goto out; | |
1043 | ||
1044 | /* Some odd races can cause double-connects, ignore them */ | |
36b71a8b | 1045 | if (con->sock) |
fdda387f | 1046 | goto out; |
fdda387f PC |
1047 | |
1048 | /* Create a socket to communicate with */ | |
eeb1bd5c EB |
1049 | result = sock_create_kern(&init_net, dlm_local_addr[0]->ss_family, |
1050 | SOCK_STREAM, IPPROTO_TCP, &sock); | |
fdda387f PC |
1051 | if (result < 0) |
1052 | goto out_err; | |
1053 | ||
1054 | memset(&saddr, 0, sizeof(saddr)); | |
98e1b60e | 1055 | result = nodeid_to_addr(con->nodeid, &saddr, NULL, false); |
36b71a8b DT |
1056 | if (result < 0) { |
1057 | log_print("no address for nodeid %d", con->nodeid); | |
ac33d071 | 1058 | goto out_err; |
36b71a8b | 1059 | } |
fdda387f PC |
1060 | |
1061 | sock->sk->sk_user_data = con; | |
1062 | con->rx_action = receive_from_sock; | |
6ed7257b PC |
1063 | con->connect_action = tcp_connect_to_sock; |
1064 | add_sock(sock, con); | |
fdda387f | 1065 | |
6bd8feda LH |
1066 | /* Bind to our cluster-known address connecting to avoid |
1067 | routing problems */ | |
1068 | memcpy(&src_addr, dlm_local_addr[0], sizeof(src_addr)); | |
1069 | make_sockaddr(&src_addr, 0, &addr_len); | |
1070 | result = sock->ops->bind(sock, (struct sockaddr *) &src_addr, | |
1071 | addr_len); | |
1072 | if (result < 0) { | |
1073 | log_print("could not bind for connect: %d", result); | |
1074 | /* This *may* not indicate a critical error */ | |
1075 | } | |
1076 | ||
68c817a1 | 1077 | make_sockaddr(&saddr, dlm_config.ci_tcp_port, &addr_len); |
fdda387f | 1078 | |
fdda387f | 1079 | log_print("connecting to %d", con->nodeid); |
cb2d45da DT |
1080 | |
1081 | /* Turn off Nagle's algorithm */ | |
1082 | kernel_setsockopt(sock, SOL_TCP, TCP_NODELAY, (char *)&one, | |
1083 | sizeof(one)); | |
1084 | ||
36b71a8b | 1085 | result = sock->ops->connect(sock, (struct sockaddr *)&saddr, addr_len, |
ac33d071 | 1086 | O_NONBLOCK); |
fdda387f PC |
1087 | if (result == -EINPROGRESS) |
1088 | result = 0; | |
ac33d071 PC |
1089 | if (result == 0) |
1090 | goto out; | |
fdda387f | 1091 | |
ac33d071 | 1092 | out_err: |
fdda387f PC |
1093 | if (con->sock) { |
1094 | sock_release(con->sock); | |
1095 | con->sock = NULL; | |
a89d63a1 CD |
1096 | } else if (sock) { |
1097 | sock_release(sock); | |
fdda387f PC |
1098 | } |
1099 | /* | |
1100 | * Some errors are fatal and this list might need adjusting. For other | |
1101 | * errors we try again until the max number of retries is reached. | |
1102 | */ | |
36b71a8b DT |
1103 | if (result != -EHOSTUNREACH && |
1104 | result != -ENETUNREACH && | |
1105 | result != -ENETDOWN && | |
1106 | result != -EINVAL && | |
1107 | result != -EPROTONOSUPPORT) { | |
1108 | log_print("connect %d try %d error %d", con->nodeid, | |
1109 | con->retries, result); | |
1110 | mutex_unlock(&con->sock_mutex); | |
1111 | msleep(1000); | |
356344c4 | 1112 | clear_bit(CF_CONNECT_PENDING, &con->flags); |
fdda387f | 1113 | lowcomms_connect_sock(con); |
36b71a8b | 1114 | return; |
fdda387f | 1115 | } |
ac33d071 | 1116 | out: |
f1f1c1cc | 1117 | mutex_unlock(&con->sock_mutex); |
00dcffae | 1118 | set_bit(CF_WRITE_PENDING, &con->flags); |
ac33d071 | 1119 | return; |
fdda387f PC |
1120 | } |
1121 | ||
6ed7257b PC |
1122 | static struct socket *tcp_create_listen_sock(struct connection *con, |
1123 | struct sockaddr_storage *saddr) | |
fdda387f | 1124 | { |
ac33d071 | 1125 | struct socket *sock = NULL; |
fdda387f PC |
1126 | int result = 0; |
1127 | int one = 1; | |
1128 | int addr_len; | |
1129 | ||
6ed7257b | 1130 | if (dlm_local_addr[0]->ss_family == AF_INET) |
fdda387f PC |
1131 | addr_len = sizeof(struct sockaddr_in); |
1132 | else | |
1133 | addr_len = sizeof(struct sockaddr_in6); | |
1134 | ||
1135 | /* Create a socket to communicate with */ | |
eeb1bd5c EB |
1136 | result = sock_create_kern(&init_net, dlm_local_addr[0]->ss_family, |
1137 | SOCK_STREAM, IPPROTO_TCP, &sock); | |
fdda387f | 1138 | if (result < 0) { |
617e82e1 | 1139 | log_print("Can't create listening comms socket"); |
fdda387f PC |
1140 | goto create_out; |
1141 | } | |
1142 | ||
cb2d45da DT |
1143 | /* Turn off Nagle's algorithm */ |
1144 | kernel_setsockopt(sock, SOL_TCP, TCP_NODELAY, (char *)&one, | |
1145 | sizeof(one)); | |
1146 | ||
6ed7257b PC |
1147 | result = kernel_setsockopt(sock, SOL_SOCKET, SO_REUSEADDR, |
1148 | (char *)&one, sizeof(one)); | |
1149 | ||
fdda387f | 1150 | if (result < 0) { |
617e82e1 | 1151 | log_print("Failed to set SO_REUSEADDR on socket: %d", result); |
fdda387f | 1152 | } |
6ed7257b PC |
1153 | con->rx_action = tcp_accept_from_sock; |
1154 | con->connect_action = tcp_connect_to_sock; | |
fdda387f PC |
1155 | |
1156 | /* Bind to our port */ | |
68c817a1 | 1157 | make_sockaddr(saddr, dlm_config.ci_tcp_port, &addr_len); |
fdda387f PC |
1158 | result = sock->ops->bind(sock, (struct sockaddr *) saddr, addr_len); |
1159 | if (result < 0) { | |
617e82e1 | 1160 | log_print("Can't bind to port %d", dlm_config.ci_tcp_port); |
fdda387f PC |
1161 | sock_release(sock); |
1162 | sock = NULL; | |
1163 | con->sock = NULL; | |
1164 | goto create_out; | |
1165 | } | |
6ed7257b | 1166 | result = kernel_setsockopt(sock, SOL_SOCKET, SO_KEEPALIVE, |
ac33d071 | 1167 | (char *)&one, sizeof(one)); |
fdda387f | 1168 | if (result < 0) { |
617e82e1 | 1169 | log_print("Set keepalive failed: %d", result); |
fdda387f PC |
1170 | } |
1171 | ||
1172 | result = sock->ops->listen(sock, 5); | |
1173 | if (result < 0) { | |
617e82e1 | 1174 | log_print("Can't listen on port %d", dlm_config.ci_tcp_port); |
fdda387f PC |
1175 | sock_release(sock); |
1176 | sock = NULL; | |
1177 | goto create_out; | |
1178 | } | |
1179 | ||
ac33d071 | 1180 | create_out: |
fdda387f PC |
1181 | return sock; |
1182 | } | |
1183 | ||
6ed7257b PC |
1184 | /* Get local addresses */ |
1185 | static void init_local(void) | |
1186 | { | |
1187 | struct sockaddr_storage sas, *addr; | |
1188 | int i; | |
1189 | ||
30d3a237 | 1190 | dlm_local_count = 0; |
1b189b88 | 1191 | for (i = 0; i < DLM_MAX_ADDR_COUNT; i++) { |
6ed7257b PC |
1192 | if (dlm_our_addr(&sas, i)) |
1193 | break; | |
1194 | ||
573c24c4 | 1195 | addr = kmalloc(sizeof(*addr), GFP_NOFS); |
6ed7257b PC |
1196 | if (!addr) |
1197 | break; | |
1198 | memcpy(addr, &sas, sizeof(*addr)); | |
1199 | dlm_local_addr[dlm_local_count++] = addr; | |
1200 | } | |
1201 | } | |
1202 | ||
6ed7257b PC |
1203 | /* Initialise SCTP socket and bind to all interfaces */ |
1204 | static int sctp_listen_for_all(void) | |
1205 | { | |
1206 | struct socket *sock = NULL; | |
ee44b4bc | 1207 | int result = -EINVAL; |
573c24c4 | 1208 | struct connection *con = nodeid2con(0, GFP_NOFS); |
6ed7257b | 1209 | int bufsize = NEEDED_RMEM; |
86e92ad2 | 1210 | int one = 1; |
6ed7257b PC |
1211 | |
1212 | if (!con) | |
1213 | return -ENOMEM; | |
1214 | ||
1215 | log_print("Using SCTP for communications"); | |
1216 | ||
eeb1bd5c | 1217 | result = sock_create_kern(&init_net, dlm_local_addr[0]->ss_family, |
ee44b4bc | 1218 | SOCK_STREAM, IPPROTO_SCTP, &sock); |
6ed7257b PC |
1219 | if (result < 0) { |
1220 | log_print("Can't create comms socket, check SCTP is loaded"); | |
1221 | goto out; | |
1222 | } | |
1223 | ||
df61c952 | 1224 | result = kernel_setsockopt(sock, SOL_SOCKET, SO_RCVBUFFORCE, |
6ed7257b PC |
1225 | (char *)&bufsize, sizeof(bufsize)); |
1226 | if (result) | |
617e82e1 | 1227 | log_print("Error increasing buffer space on socket %d", result); |
6ed7257b | 1228 | |
86e92ad2 MC |
1229 | result = kernel_setsockopt(sock, SOL_SCTP, SCTP_NODELAY, (char *)&one, |
1230 | sizeof(one)); | |
1231 | if (result < 0) | |
1232 | log_print("Could not set SCTP NODELAY error %d\n", result); | |
1233 | ||
6ed7257b PC |
1234 | /* Init con struct */ |
1235 | sock->sk->sk_user_data = con; | |
1236 | con->sock = sock; | |
1237 | con->sock->sk->sk_data_ready = lowcomms_data_ready; | |
ee44b4bc MRL |
1238 | con->rx_action = sctp_accept_from_sock; |
1239 | con->connect_action = sctp_connect_to_sock; | |
6ed7257b | 1240 | |
ee44b4bc MRL |
1241 | /* Bind to all addresses. */ |
1242 | if (sctp_bind_addrs(con, dlm_config.ci_tcp_port)) | |
1243 | goto create_delsock; | |
6ed7257b PC |
1244 | |
1245 | result = sock->ops->listen(sock, 5); | |
1246 | if (result < 0) { | |
1247 | log_print("Can't set socket listening"); | |
1248 | goto create_delsock; | |
1249 | } | |
1250 | ||
1251 | return 0; | |
1252 | ||
1253 | create_delsock: | |
1254 | sock_release(sock); | |
1255 | con->sock = NULL; | |
1256 | out: | |
1257 | return result; | |
1258 | } | |
1259 | ||
1260 | static int tcp_listen_for_all(void) | |
fdda387f PC |
1261 | { |
1262 | struct socket *sock = NULL; | |
573c24c4 | 1263 | struct connection *con = nodeid2con(0, GFP_NOFS); |
fdda387f PC |
1264 | int result = -EINVAL; |
1265 | ||
6ed7257b PC |
1266 | if (!con) |
1267 | return -ENOMEM; | |
1268 | ||
fdda387f | 1269 | /* We don't support multi-homed hosts */ |
6ed7257b | 1270 | if (dlm_local_addr[1] != NULL) { |
617e82e1 DT |
1271 | log_print("TCP protocol can't handle multi-homed hosts, " |
1272 | "try SCTP"); | |
6ed7257b PC |
1273 | return -EINVAL; |
1274 | } | |
1275 | ||
1276 | log_print("Using TCP for communications"); | |
1277 | ||
6ed7257b | 1278 | sock = tcp_create_listen_sock(con, dlm_local_addr[0]); |
fdda387f PC |
1279 | if (sock) { |
1280 | add_sock(sock, con); | |
1281 | result = 0; | |
1282 | } | |
1283 | else { | |
1284 | result = -EADDRINUSE; | |
1285 | } | |
1286 | ||
1287 | return result; | |
1288 | } | |
1289 | ||
1290 | ||
1291 | ||
1292 | static struct writequeue_entry *new_writequeue_entry(struct connection *con, | |
1293 | gfp_t allocation) | |
1294 | { | |
1295 | struct writequeue_entry *entry; | |
1296 | ||
1297 | entry = kmalloc(sizeof(struct writequeue_entry), allocation); | |
1298 | if (!entry) | |
1299 | return NULL; | |
1300 | ||
1301 | entry->page = alloc_page(allocation); | |
1302 | if (!entry->page) { | |
1303 | kfree(entry); | |
1304 | return NULL; | |
1305 | } | |
1306 | ||
1307 | entry->offset = 0; | |
1308 | entry->len = 0; | |
1309 | entry->end = 0; | |
1310 | entry->users = 0; | |
1311 | entry->con = con; | |
1312 | ||
1313 | return entry; | |
1314 | } | |
1315 | ||
617e82e1 | 1316 | void *dlm_lowcomms_get_buffer(int nodeid, int len, gfp_t allocation, char **ppc) |
fdda387f PC |
1317 | { |
1318 | struct connection *con; | |
1319 | struct writequeue_entry *e; | |
1320 | int offset = 0; | |
fdda387f | 1321 | |
fdda387f PC |
1322 | con = nodeid2con(nodeid, allocation); |
1323 | if (!con) | |
1324 | return NULL; | |
1325 | ||
4edde74e | 1326 | spin_lock(&con->writequeue_lock); |
fdda387f | 1327 | e = list_entry(con->writequeue.prev, struct writequeue_entry, list); |
ac33d071 | 1328 | if ((&e->list == &con->writequeue) || |
fdda387f PC |
1329 | (PAGE_CACHE_SIZE - e->end < len)) { |
1330 | e = NULL; | |
1331 | } else { | |
1332 | offset = e->end; | |
1333 | e->end += len; | |
eeee2b5f | 1334 | e->users++; |
fdda387f PC |
1335 | } |
1336 | spin_unlock(&con->writequeue_lock); | |
1337 | ||
1338 | if (e) { | |
ac33d071 | 1339 | got_one: |
fdda387f PC |
1340 | *ppc = page_address(e->page) + offset; |
1341 | return e; | |
1342 | } | |
1343 | ||
1344 | e = new_writequeue_entry(con, allocation); | |
1345 | if (e) { | |
1346 | spin_lock(&con->writequeue_lock); | |
1347 | offset = e->end; | |
1348 | e->end += len; | |
eeee2b5f | 1349 | e->users++; |
fdda387f PC |
1350 | list_add_tail(&e->list, &con->writequeue); |
1351 | spin_unlock(&con->writequeue_lock); | |
1352 | goto got_one; | |
1353 | } | |
1354 | return NULL; | |
1355 | } | |
1356 | ||
1357 | void dlm_lowcomms_commit_buffer(void *mh) | |
1358 | { | |
1359 | struct writequeue_entry *e = (struct writequeue_entry *)mh; | |
1360 | struct connection *con = e->con; | |
1361 | int users; | |
1362 | ||
4edde74e | 1363 | spin_lock(&con->writequeue_lock); |
fdda387f PC |
1364 | users = --e->users; |
1365 | if (users) | |
1366 | goto out; | |
1367 | e->len = e->end - e->offset; | |
fdda387f PC |
1368 | spin_unlock(&con->writequeue_lock); |
1369 | ||
1d6e8131 PC |
1370 | if (!test_and_set_bit(CF_WRITE_PENDING, &con->flags)) { |
1371 | queue_work(send_workqueue, &con->swork); | |
fdda387f PC |
1372 | } |
1373 | return; | |
1374 | ||
ac33d071 | 1375 | out: |
fdda387f PC |
1376 | spin_unlock(&con->writequeue_lock); |
1377 | return; | |
1378 | } | |
1379 | ||
fdda387f | 1380 | /* Send a message */ |
ac33d071 | 1381 | static void send_to_sock(struct connection *con) |
fdda387f PC |
1382 | { |
1383 | int ret = 0; | |
fdda387f PC |
1384 | const int msg_flags = MSG_DONTWAIT | MSG_NOSIGNAL; |
1385 | struct writequeue_entry *e; | |
1386 | int len, offset; | |
f92c8dd7 | 1387 | int count = 0; |
fdda387f | 1388 | |
f1f1c1cc | 1389 | mutex_lock(&con->sock_mutex); |
fdda387f PC |
1390 | if (con->sock == NULL) |
1391 | goto out_connect; | |
1392 | ||
fdda387f PC |
1393 | spin_lock(&con->writequeue_lock); |
1394 | for (;;) { | |
1395 | e = list_entry(con->writequeue.next, struct writequeue_entry, | |
1396 | list); | |
1397 | if ((struct list_head *) e == &con->writequeue) | |
1398 | break; | |
1399 | ||
1400 | len = e->len; | |
1401 | offset = e->offset; | |
1402 | BUG_ON(len == 0 && e->users == 0); | |
1403 | spin_unlock(&con->writequeue_lock); | |
1404 | ||
1405 | ret = 0; | |
1406 | if (len) { | |
1329e3f2 PB |
1407 | ret = kernel_sendpage(con->sock, e->page, offset, len, |
1408 | msg_flags); | |
d66f8277 | 1409 | if (ret == -EAGAIN || ret == 0) { |
b36930dd DM |
1410 | if (ret == -EAGAIN && |
1411 | test_bit(SOCK_ASYNC_NOSPACE, &con->sock->flags) && | |
1412 | !test_and_set_bit(CF_APP_LIMITED, &con->flags)) { | |
1413 | /* Notify TCP that we're limited by the | |
1414 | * application window size. | |
1415 | */ | |
1416 | set_bit(SOCK_NOSPACE, &con->sock->flags); | |
1417 | con->sock->sk->sk_write_pending++; | |
1418 | } | |
d66f8277 | 1419 | cond_resched(); |
fdda387f | 1420 | goto out; |
9c5bef58 | 1421 | } else if (ret < 0) |
fdda387f | 1422 | goto send_error; |
d66f8277 | 1423 | } |
f92c8dd7 BP |
1424 | |
1425 | /* Don't starve people filling buffers */ | |
1426 | if (++count >= MAX_SEND_MSG_COUNT) { | |
ac33d071 | 1427 | cond_resched(); |
f92c8dd7 BP |
1428 | count = 0; |
1429 | } | |
fdda387f PC |
1430 | |
1431 | spin_lock(&con->writequeue_lock); | |
5d689871 | 1432 | writequeue_entry_complete(e, ret); |
fdda387f PC |
1433 | } |
1434 | spin_unlock(&con->writequeue_lock); | |
ac33d071 | 1435 | out: |
f1f1c1cc | 1436 | mutex_unlock(&con->sock_mutex); |
ac33d071 | 1437 | return; |
fdda387f | 1438 | |
ac33d071 | 1439 | send_error: |
f1f1c1cc | 1440 | mutex_unlock(&con->sock_mutex); |
0d737a8c | 1441 | close_connection(con, false, false, true); |
fdda387f | 1442 | lowcomms_connect_sock(con); |
ac33d071 | 1443 | return; |
fdda387f | 1444 | |
ac33d071 | 1445 | out_connect: |
f1f1c1cc | 1446 | mutex_unlock(&con->sock_mutex); |
ee44b4bc | 1447 | lowcomms_connect_sock(con); |
fdda387f PC |
1448 | } |
1449 | ||
1450 | static void clean_one_writequeue(struct connection *con) | |
1451 | { | |
5e9ccc37 | 1452 | struct writequeue_entry *e, *safe; |
fdda387f PC |
1453 | |
1454 | spin_lock(&con->writequeue_lock); | |
5e9ccc37 | 1455 | list_for_each_entry_safe(e, safe, &con->writequeue, list) { |
fdda387f PC |
1456 | list_del(&e->list); |
1457 | free_entry(e); | |
1458 | } | |
1459 | spin_unlock(&con->writequeue_lock); | |
1460 | } | |
1461 | ||
1462 | /* Called from recovery when it knows that a node has | |
1463 | left the cluster */ | |
1464 | int dlm_lowcomms_close(int nodeid) | |
1465 | { | |
1466 | struct connection *con; | |
36b71a8b | 1467 | struct dlm_node_addr *na; |
fdda387f | 1468 | |
fdda387f PC |
1469 | log_print("closing connection to node %d", nodeid); |
1470 | con = nodeid2con(nodeid, 0); | |
1471 | if (con) { | |
063c4c99 | 1472 | set_bit(CF_CLOSE, &con->flags); |
0d737a8c | 1473 | close_connection(con, true, true, true); |
fdda387f | 1474 | clean_one_writequeue(con); |
fdda387f | 1475 | } |
36b71a8b DT |
1476 | |
1477 | spin_lock(&dlm_node_addrs_spin); | |
1478 | na = find_node_addr(nodeid); | |
1479 | if (na) { | |
1480 | list_del(&na->list); | |
1481 | while (na->addr_count--) | |
1482 | kfree(na->addr[na->addr_count]); | |
1483 | kfree(na); | |
1484 | } | |
1485 | spin_unlock(&dlm_node_addrs_spin); | |
1486 | ||
fdda387f | 1487 | return 0; |
fdda387f PC |
1488 | } |
1489 | ||
6ed7257b | 1490 | /* Receive workqueue function */ |
1d6e8131 | 1491 | static void process_recv_sockets(struct work_struct *work) |
fdda387f | 1492 | { |
1d6e8131 PC |
1493 | struct connection *con = container_of(work, struct connection, rwork); |
1494 | int err; | |
fdda387f | 1495 | |
1d6e8131 PC |
1496 | clear_bit(CF_READ_PENDING, &con->flags); |
1497 | do { | |
1498 | err = con->rx_action(con); | |
1499 | } while (!err); | |
fdda387f PC |
1500 | } |
1501 | ||
6ed7257b | 1502 | /* Send workqueue function */ |
1d6e8131 | 1503 | static void process_send_sockets(struct work_struct *work) |
fdda387f | 1504 | { |
1d6e8131 | 1505 | struct connection *con = container_of(work, struct connection, swork); |
fdda387f | 1506 | |
00dcffae | 1507 | if (test_and_clear_bit(CF_CONNECT_PENDING, &con->flags)) |
6ed7257b | 1508 | con->connect_action(con); |
063c4c99 LMB |
1509 | if (test_and_clear_bit(CF_WRITE_PENDING, &con->flags)) |
1510 | send_to_sock(con); | |
fdda387f PC |
1511 | } |
1512 | ||
1513 | ||
1514 | /* Discard all entries on the write queues */ | |
1515 | static void clean_writequeues(void) | |
1516 | { | |
5e9ccc37 | 1517 | foreach_conn(clean_one_writequeue); |
fdda387f PC |
1518 | } |
1519 | ||
1d6e8131 | 1520 | static void work_stop(void) |
fdda387f | 1521 | { |
1d6e8131 PC |
1522 | destroy_workqueue(recv_workqueue); |
1523 | destroy_workqueue(send_workqueue); | |
fdda387f PC |
1524 | } |
1525 | ||
1d6e8131 | 1526 | static int work_start(void) |
fdda387f | 1527 | { |
e43f055a DT |
1528 | recv_workqueue = alloc_workqueue("dlm_recv", |
1529 | WQ_UNBOUND | WQ_MEM_RECLAIM, 1); | |
b9d41052 NK |
1530 | if (!recv_workqueue) { |
1531 | log_print("can't start dlm_recv"); | |
1532 | return -ENOMEM; | |
fdda387f | 1533 | } |
fdda387f | 1534 | |
e43f055a DT |
1535 | send_workqueue = alloc_workqueue("dlm_send", |
1536 | WQ_UNBOUND | WQ_MEM_RECLAIM, 1); | |
b9d41052 NK |
1537 | if (!send_workqueue) { |
1538 | log_print("can't start dlm_send"); | |
1d6e8131 | 1539 | destroy_workqueue(recv_workqueue); |
b9d41052 | 1540 | return -ENOMEM; |
fdda387f | 1541 | } |
fdda387f PC |
1542 | |
1543 | return 0; | |
1544 | } | |
1545 | ||
5e9ccc37 | 1546 | static void stop_conn(struct connection *con) |
fdda387f | 1547 | { |
5e9ccc37 | 1548 | con->flags |= 0x0F; |
391fbdc5 | 1549 | if (con->sock && con->sock->sk) |
5e9ccc37 CC |
1550 | con->sock->sk->sk_user_data = NULL; |
1551 | } | |
fdda387f | 1552 | |
5e9ccc37 CC |
1553 | static void free_conn(struct connection *con) |
1554 | { | |
0d737a8c | 1555 | close_connection(con, true, true, true); |
5e9ccc37 CC |
1556 | if (con->othercon) |
1557 | kmem_cache_free(con_cache, con->othercon); | |
1558 | hlist_del(&con->list); | |
1559 | kmem_cache_free(con_cache, con); | |
1560 | } | |
1561 | ||
1562 | void dlm_lowcomms_stop(void) | |
1563 | { | |
ac33d071 | 1564 | /* Set all the flags to prevent any |
fdda387f PC |
1565 | socket activity. |
1566 | */ | |
7a936ce7 | 1567 | mutex_lock(&connections_lock); |
513ef596 | 1568 | dlm_allow_conn = 0; |
5e9ccc37 | 1569 | foreach_conn(stop_conn); |
7a936ce7 | 1570 | mutex_unlock(&connections_lock); |
ac33d071 | 1571 | |
1d6e8131 | 1572 | work_stop(); |
6ed7257b | 1573 | |
7a936ce7 | 1574 | mutex_lock(&connections_lock); |
fdda387f PC |
1575 | clean_writequeues(); |
1576 | ||
5e9ccc37 CC |
1577 | foreach_conn(free_conn); |
1578 | ||
7a936ce7 | 1579 | mutex_unlock(&connections_lock); |
fdda387f PC |
1580 | kmem_cache_destroy(con_cache); |
1581 | } | |
1582 | ||
fdda387f PC |
1583 | int dlm_lowcomms_start(void) |
1584 | { | |
6ed7257b PC |
1585 | int error = -EINVAL; |
1586 | struct connection *con; | |
5e9ccc37 CC |
1587 | int i; |
1588 | ||
1589 | for (i = 0; i < CONN_HASH_SIZE; i++) | |
1590 | INIT_HLIST_HEAD(&connection_hash[i]); | |
fdda387f | 1591 | |
6ed7257b PC |
1592 | init_local(); |
1593 | if (!dlm_local_count) { | |
617e82e1 | 1594 | error = -ENOTCONN; |
fdda387f | 1595 | log_print("no local IP address has been set"); |
513ef596 | 1596 | goto fail; |
fdda387f PC |
1597 | } |
1598 | ||
6ed7257b | 1599 | error = -ENOMEM; |
fdda387f | 1600 | con_cache = kmem_cache_create("dlm_conn", sizeof(struct connection), |
ac33d071 | 1601 | __alignof__(struct connection), 0, |
20c2df83 | 1602 | NULL); |
fdda387f | 1603 | if (!con_cache) |
513ef596 DT |
1604 | goto fail; |
1605 | ||
1606 | error = work_start(); | |
1607 | if (error) | |
1608 | goto fail_destroy; | |
1609 | ||
1610 | dlm_allow_conn = 1; | |
fdda387f | 1611 | |
fdda387f | 1612 | /* Start listening */ |
6ed7257b PC |
1613 | if (dlm_config.ci_protocol == 0) |
1614 | error = tcp_listen_for_all(); | |
1615 | else | |
1616 | error = sctp_listen_for_all(); | |
fdda387f PC |
1617 | if (error) |
1618 | goto fail_unlisten; | |
1619 | ||
fdda387f PC |
1620 | return 0; |
1621 | ||
ac33d071 | 1622 | fail_unlisten: |
513ef596 | 1623 | dlm_allow_conn = 0; |
6ed7257b PC |
1624 | con = nodeid2con(0,0); |
1625 | if (con) { | |
0d737a8c | 1626 | close_connection(con, false, true, true); |
6ed7257b PC |
1627 | kmem_cache_free(con_cache, con); |
1628 | } | |
513ef596 | 1629 | fail_destroy: |
fdda387f | 1630 | kmem_cache_destroy(con_cache); |
513ef596 | 1631 | fail: |
fdda387f PC |
1632 | return error; |
1633 | } | |
36b71a8b DT |
1634 | |
1635 | void dlm_lowcomms_exit(void) | |
1636 | { | |
1637 | struct dlm_node_addr *na, *safe; | |
1638 | ||
1639 | spin_lock(&dlm_node_addrs_spin); | |
1640 | list_for_each_entry_safe(na, safe, &dlm_node_addrs, list) { | |
1641 | list_del(&na->list); | |
1642 | while (na->addr_count--) | |
1643 | kfree(na->addr[na->addr_count]); | |
1644 | kfree(na); | |
1645 | } | |
1646 | spin_unlock(&dlm_node_addrs_spin); | |
1647 | } |