Commit | Line | Data |
---|---|---|
6714d8e8 KH |
1 | /* -*- mode: c; c-basic-offset: 8; -*- |
2 | * vim: noexpandtab sw=8 ts=8 sts=0: | |
3 | * | |
4 | * dlmdomain.c | |
5 | * | |
6 | * defines domain join / leave apis | |
7 | * | |
8 | * Copyright (C) 2004 Oracle. All rights reserved. | |
9 | * | |
10 | * This program is free software; you can redistribute it and/or | |
11 | * modify it under the terms of the GNU General Public | |
12 | * License as published by the Free Software Foundation; either | |
13 | * version 2 of the License, or (at your option) any later version. | |
14 | * | |
15 | * This program is distributed in the hope that it will be useful, | |
16 | * but WITHOUT ANY WARRANTY; without even the implied warranty of | |
17 | * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU | |
18 | * General Public License for more details. | |
19 | * | |
20 | * You should have received a copy of the GNU General Public | |
21 | * License along with this program; if not, write to the | |
22 | * Free Software Foundation, Inc., 59 Temple Place - Suite 330, | |
23 | * Boston, MA 021110-1307, USA. | |
24 | * | |
25 | */ | |
26 | ||
27 | #include <linux/module.h> | |
28 | #include <linux/types.h> | |
29 | #include <linux/slab.h> | |
30 | #include <linux/highmem.h> | |
31 | #include <linux/utsname.h> | |
32 | #include <linux/init.h> | |
33 | #include <linux/spinlock.h> | |
34 | #include <linux/delay.h> | |
35 | #include <linux/err.h> | |
36 | ||
37 | #include "cluster/heartbeat.h" | |
38 | #include "cluster/nodemanager.h" | |
39 | #include "cluster/tcp.h" | |
40 | ||
41 | #include "dlmapi.h" | |
42 | #include "dlmcommon.h" | |
43 | ||
6714d8e8 KH |
44 | #include "dlmdomain.h" |
45 | ||
46 | #include "dlmver.h" | |
47 | ||
48 | #define MLOG_MASK_PREFIX (ML_DLM|ML_DLM_DOMAIN) | |
49 | #include "cluster/masklog.h" | |
50 | ||
03d864c0 DP |
51 | static void dlm_free_pagevec(void **vec, int pages) |
52 | { | |
53 | while (pages--) | |
54 | free_page((unsigned long)vec[pages]); | |
55 | kfree(vec); | |
56 | } | |
57 | ||
58 | static void **dlm_alloc_pagevec(int pages) | |
59 | { | |
60 | void **vec = kmalloc(pages * sizeof(void *), GFP_KERNEL); | |
61 | int i; | |
62 | ||
63 | if (!vec) | |
64 | return NULL; | |
65 | ||
66 | for (i = 0; i < pages; i++) | |
67 | if (!(vec[i] = (void *)__get_free_page(GFP_KERNEL))) | |
68 | goto out_free; | |
c8f33b6e | 69 | |
685f1adb MF |
70 | mlog(0, "Allocated DLM hash pagevec; %d pages (%lu expected), %lu buckets per page\n", |
71 | pages, DLM_HASH_PAGES, (unsigned long)DLM_BUCKETS_PER_PAGE); | |
03d864c0 DP |
72 | return vec; |
73 | out_free: | |
74 | dlm_free_pagevec(vec, i); | |
75 | return NULL; | |
76 | } | |
77 | ||
6714d8e8 KH |
78 | /* |
79 | * | |
80 | * spinlock lock ordering: if multiple locks are needed, obey this ordering: | |
81 | * dlm_domain_lock | |
82 | * struct dlm_ctxt->spinlock | |
83 | * struct dlm_lock_resource->spinlock | |
84 | * struct dlm_ctxt->master_lock | |
85 | * struct dlm_ctxt->ast_lock | |
86 | * dlm_master_list_entry->spinlock | |
87 | * dlm_lock->spinlock | |
88 | * | |
89 | */ | |
90 | ||
34af946a | 91 | DEFINE_SPINLOCK(dlm_domain_lock); |
6714d8e8 KH |
92 | LIST_HEAD(dlm_domains); |
93 | static DECLARE_WAIT_QUEUE_HEAD(dlm_domain_events); | |
94 | ||
95 | #define DLM_DOMAIN_BACKOFF_MS 200 | |
96 | ||
97 | static int dlm_query_join_handler(struct o2net_msg *msg, u32 len, void *data); | |
98 | static int dlm_assert_joined_handler(struct o2net_msg *msg, u32 len, void *data); | |
99 | static int dlm_cancel_join_handler(struct o2net_msg *msg, u32 len, void *data); | |
100 | static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data); | |
101 | ||
102 | static void dlm_unregister_domain_handlers(struct dlm_ctxt *dlm); | |
103 | ||
104 | void __dlm_unhash_lockres(struct dlm_lock_resource *lockres) | |
105 | { | |
81f2094a | 106 | hlist_del_init(&lockres->hash_node); |
6714d8e8 KH |
107 | dlm_lockres_put(lockres); |
108 | } | |
109 | ||
110 | void __dlm_insert_lockres(struct dlm_ctxt *dlm, | |
111 | struct dlm_lock_resource *res) | |
112 | { | |
81f2094a | 113 | struct hlist_head *bucket; |
6714d8e8 KH |
114 | struct qstr *q; |
115 | ||
116 | assert_spin_locked(&dlm->spinlock); | |
117 | ||
118 | q = &res->lockname; | |
03d864c0 | 119 | bucket = dlm_lockres_hash(dlm, q->hash); |
6714d8e8 KH |
120 | |
121 | /* get a reference for our hashtable */ | |
122 | dlm_lockres_get(res); | |
123 | ||
81f2094a | 124 | hlist_add_head(&res->hash_node, bucket); |
6714d8e8 KH |
125 | } |
126 | ||
127 | struct dlm_lock_resource * __dlm_lookup_lockres(struct dlm_ctxt *dlm, | |
a3d33291 MF |
128 | const char *name, |
129 | unsigned int len, | |
130 | unsigned int hash) | |
6714d8e8 | 131 | { |
81f2094a | 132 | struct hlist_head *bucket; |
4198985f | 133 | struct hlist_node *list; |
6714d8e8 KH |
134 | |
135 | mlog_entry("%.*s\n", len, name); | |
136 | ||
137 | assert_spin_locked(&dlm->spinlock); | |
138 | ||
03d864c0 DP |
139 | bucket = dlm_lockres_hash(dlm, hash); |
140 | ||
4198985f DP |
141 | hlist_for_each(list, bucket) { |
142 | struct dlm_lock_resource *res = hlist_entry(list, | |
143 | struct dlm_lock_resource, hash_node); | |
144 | if (res->lockname.name[0] != name[0]) | |
145 | continue; | |
146 | if (unlikely(res->lockname.len != len)) | |
147 | continue; | |
148 | if (memcmp(res->lockname.name + 1, name + 1, len - 1)) | |
149 | continue; | |
150 | dlm_lockres_get(res); | |
151 | return res; | |
6714d8e8 | 152 | } |
4198985f | 153 | return NULL; |
6714d8e8 KH |
154 | } |
155 | ||
156 | struct dlm_lock_resource * dlm_lookup_lockres(struct dlm_ctxt *dlm, | |
157 | const char *name, | |
158 | unsigned int len) | |
159 | { | |
160 | struct dlm_lock_resource *res; | |
a3d33291 | 161 | unsigned int hash = dlm_lockid_hash(name, len); |
6714d8e8 KH |
162 | |
163 | spin_lock(&dlm->spinlock); | |
a3d33291 | 164 | res = __dlm_lookup_lockres(dlm, name, len, hash); |
6714d8e8 KH |
165 | spin_unlock(&dlm->spinlock); |
166 | return res; | |
167 | } | |
168 | ||
169 | static struct dlm_ctxt * __dlm_lookup_domain_full(const char *domain, int len) | |
170 | { | |
171 | struct dlm_ctxt *tmp = NULL; | |
172 | struct list_head *iter; | |
173 | ||
174 | assert_spin_locked(&dlm_domain_lock); | |
175 | ||
176 | /* tmp->name here is always NULL terminated, | |
177 | * but domain may not be! */ | |
178 | list_for_each(iter, &dlm_domains) { | |
179 | tmp = list_entry (iter, struct dlm_ctxt, list); | |
180 | if (strlen(tmp->name) == len && | |
181 | memcmp(tmp->name, domain, len)==0) | |
182 | break; | |
183 | tmp = NULL; | |
184 | } | |
185 | ||
186 | return tmp; | |
187 | } | |
188 | ||
189 | /* For null terminated domain strings ONLY */ | |
190 | static struct dlm_ctxt * __dlm_lookup_domain(const char *domain) | |
191 | { | |
192 | assert_spin_locked(&dlm_domain_lock); | |
193 | ||
194 | return __dlm_lookup_domain_full(domain, strlen(domain)); | |
195 | } | |
196 | ||
197 | ||
198 | /* returns true on one of two conditions: | |
199 | * 1) the domain does not exist | |
200 | * 2) the domain exists and it's state is "joined" */ | |
201 | static int dlm_wait_on_domain_helper(const char *domain) | |
202 | { | |
203 | int ret = 0; | |
204 | struct dlm_ctxt *tmp = NULL; | |
205 | ||
206 | spin_lock(&dlm_domain_lock); | |
207 | ||
208 | tmp = __dlm_lookup_domain(domain); | |
209 | if (!tmp) | |
210 | ret = 1; | |
211 | else if (tmp->dlm_state == DLM_CTXT_JOINED) | |
212 | ret = 1; | |
213 | ||
214 | spin_unlock(&dlm_domain_lock); | |
215 | return ret; | |
216 | } | |
217 | ||
218 | static void dlm_free_ctxt_mem(struct dlm_ctxt *dlm) | |
219 | { | |
81f2094a | 220 | if (dlm->lockres_hash) |
03d864c0 | 221 | dlm_free_pagevec((void **)dlm->lockres_hash, DLM_HASH_PAGES); |
6714d8e8 KH |
222 | |
223 | if (dlm->name) | |
224 | kfree(dlm->name); | |
225 | ||
226 | kfree(dlm); | |
227 | } | |
228 | ||
229 | /* A little strange - this function will be called while holding | |
230 | * dlm_domain_lock and is expected to be holding it on the way out. We | |
231 | * will however drop and reacquire it multiple times */ | |
232 | static void dlm_ctxt_release(struct kref *kref) | |
233 | { | |
234 | struct dlm_ctxt *dlm; | |
235 | ||
236 | dlm = container_of(kref, struct dlm_ctxt, dlm_refs); | |
237 | ||
238 | BUG_ON(dlm->num_joins); | |
239 | BUG_ON(dlm->dlm_state == DLM_CTXT_JOINED); | |
240 | ||
241 | /* we may still be in the list if we hit an error during join. */ | |
242 | list_del_init(&dlm->list); | |
243 | ||
244 | spin_unlock(&dlm_domain_lock); | |
245 | ||
246 | mlog(0, "freeing memory from domain %s\n", dlm->name); | |
247 | ||
248 | wake_up(&dlm_domain_events); | |
249 | ||
250 | dlm_free_ctxt_mem(dlm); | |
251 | ||
252 | spin_lock(&dlm_domain_lock); | |
253 | } | |
254 | ||
255 | void dlm_put(struct dlm_ctxt *dlm) | |
256 | { | |
257 | spin_lock(&dlm_domain_lock); | |
258 | kref_put(&dlm->dlm_refs, dlm_ctxt_release); | |
259 | spin_unlock(&dlm_domain_lock); | |
260 | } | |
261 | ||
262 | static void __dlm_get(struct dlm_ctxt *dlm) | |
263 | { | |
264 | kref_get(&dlm->dlm_refs); | |
265 | } | |
266 | ||
267 | /* given a questionable reference to a dlm object, gets a reference if | |
268 | * it can find it in the list, otherwise returns NULL in which case | |
269 | * you shouldn't trust your pointer. */ | |
270 | struct dlm_ctxt *dlm_grab(struct dlm_ctxt *dlm) | |
271 | { | |
272 | struct list_head *iter; | |
273 | struct dlm_ctxt *target = NULL; | |
274 | ||
275 | spin_lock(&dlm_domain_lock); | |
276 | ||
277 | list_for_each(iter, &dlm_domains) { | |
278 | target = list_entry (iter, struct dlm_ctxt, list); | |
279 | ||
280 | if (target == dlm) { | |
281 | __dlm_get(target); | |
282 | break; | |
283 | } | |
284 | ||
285 | target = NULL; | |
286 | } | |
287 | ||
288 | spin_unlock(&dlm_domain_lock); | |
289 | ||
290 | return target; | |
291 | } | |
292 | ||
293 | int dlm_domain_fully_joined(struct dlm_ctxt *dlm) | |
294 | { | |
295 | int ret; | |
296 | ||
297 | spin_lock(&dlm_domain_lock); | |
298 | ret = (dlm->dlm_state == DLM_CTXT_JOINED) || | |
299 | (dlm->dlm_state == DLM_CTXT_IN_SHUTDOWN); | |
300 | spin_unlock(&dlm_domain_lock); | |
301 | ||
302 | return ret; | |
303 | } | |
304 | ||
3156d267 KH |
305 | static void dlm_destroy_dlm_worker(struct dlm_ctxt *dlm) |
306 | { | |
307 | if (dlm->dlm_worker) { | |
308 | flush_workqueue(dlm->dlm_worker); | |
309 | destroy_workqueue(dlm->dlm_worker); | |
310 | dlm->dlm_worker = NULL; | |
311 | } | |
312 | } | |
313 | ||
6714d8e8 KH |
314 | static void dlm_complete_dlm_shutdown(struct dlm_ctxt *dlm) |
315 | { | |
316 | dlm_unregister_domain_handlers(dlm); | |
317 | dlm_complete_thread(dlm); | |
318 | dlm_complete_recovery_thread(dlm); | |
3156d267 | 319 | dlm_destroy_dlm_worker(dlm); |
6714d8e8 KH |
320 | |
321 | /* We've left the domain. Now we can take ourselves out of the | |
322 | * list and allow the kref stuff to help us free the | |
323 | * memory. */ | |
324 | spin_lock(&dlm_domain_lock); | |
325 | list_del_init(&dlm->list); | |
326 | spin_unlock(&dlm_domain_lock); | |
327 | ||
328 | /* Wake up anyone waiting for us to remove this domain */ | |
329 | wake_up(&dlm_domain_events); | |
330 | } | |
331 | ||
332 | static void dlm_migrate_all_locks(struct dlm_ctxt *dlm) | |
333 | { | |
334 | int i; | |
335 | struct dlm_lock_resource *res; | |
336 | ||
337 | mlog(0, "Migrating locks from domain %s\n", dlm->name); | |
338 | restart: | |
339 | spin_lock(&dlm->spinlock); | |
81f2094a | 340 | for (i = 0; i < DLM_HASH_BUCKETS; i++) { |
03d864c0 DP |
341 | while (!hlist_empty(dlm_lockres_hash(dlm, i))) { |
342 | res = hlist_entry(dlm_lockres_hash(dlm, i)->first, | |
81f2094a | 343 | struct dlm_lock_resource, hash_node); |
6714d8e8 KH |
344 | /* need reference when manually grabbing lockres */ |
345 | dlm_lockres_get(res); | |
346 | /* this should unhash the lockres | |
347 | * and exit with dlm->spinlock */ | |
348 | mlog(0, "purging res=%p\n", res); | |
349 | if (dlm_lockres_is_dirty(dlm, res)) { | |
350 | /* HACK! this should absolutely go. | |
351 | * need to figure out why some empty | |
352 | * lockreses are still marked dirty */ | |
353 | mlog(ML_ERROR, "lockres %.*s dirty!\n", | |
354 | res->lockname.len, res->lockname.name); | |
355 | ||
356 | spin_unlock(&dlm->spinlock); | |
357 | dlm_kick_thread(dlm, res); | |
358 | wait_event(dlm->ast_wq, !dlm_lockres_is_dirty(dlm, res)); | |
359 | dlm_lockres_put(res); | |
360 | goto restart; | |
361 | } | |
362 | dlm_purge_lockres(dlm, res); | |
363 | dlm_lockres_put(res); | |
364 | } | |
365 | } | |
366 | spin_unlock(&dlm->spinlock); | |
367 | ||
368 | mlog(0, "DONE Migrating locks from domain %s\n", dlm->name); | |
369 | } | |
370 | ||
371 | static int dlm_no_joining_node(struct dlm_ctxt *dlm) | |
372 | { | |
373 | int ret; | |
374 | ||
375 | spin_lock(&dlm->spinlock); | |
376 | ret = dlm->joining_node == DLM_LOCK_RES_OWNER_UNKNOWN; | |
377 | spin_unlock(&dlm->spinlock); | |
378 | ||
379 | return ret; | |
380 | } | |
381 | ||
382 | static void dlm_mark_domain_leaving(struct dlm_ctxt *dlm) | |
383 | { | |
384 | /* Yikes, a double spinlock! I need domain_lock for the dlm | |
385 | * state and the dlm spinlock for join state... Sorry! */ | |
386 | again: | |
387 | spin_lock(&dlm_domain_lock); | |
388 | spin_lock(&dlm->spinlock); | |
389 | ||
390 | if (dlm->joining_node != DLM_LOCK_RES_OWNER_UNKNOWN) { | |
391 | mlog(0, "Node %d is joining, we wait on it.\n", | |
392 | dlm->joining_node); | |
393 | spin_unlock(&dlm->spinlock); | |
394 | spin_unlock(&dlm_domain_lock); | |
395 | ||
396 | wait_event(dlm->dlm_join_events, dlm_no_joining_node(dlm)); | |
397 | goto again; | |
398 | } | |
399 | ||
400 | dlm->dlm_state = DLM_CTXT_LEAVING; | |
401 | spin_unlock(&dlm->spinlock); | |
402 | spin_unlock(&dlm_domain_lock); | |
403 | } | |
404 | ||
405 | static void __dlm_print_nodes(struct dlm_ctxt *dlm) | |
406 | { | |
407 | int node = -1; | |
408 | ||
409 | assert_spin_locked(&dlm->spinlock); | |
410 | ||
781ee3e2 | 411 | printk(KERN_INFO "ocfs2_dlm: Nodes in domain (\"%s\"): ", dlm->name); |
6714d8e8 KH |
412 | |
413 | while ((node = find_next_bit(dlm->domain_map, O2NM_MAX_NODES, | |
414 | node + 1)) < O2NM_MAX_NODES) { | |
781ee3e2 | 415 | printk("%d ", node); |
6714d8e8 | 416 | } |
781ee3e2 | 417 | printk("\n"); |
6714d8e8 KH |
418 | } |
419 | ||
420 | static int dlm_exit_domain_handler(struct o2net_msg *msg, u32 len, void *data) | |
421 | { | |
422 | struct dlm_ctxt *dlm = data; | |
423 | unsigned int node; | |
424 | struct dlm_exit_domain *exit_msg = (struct dlm_exit_domain *) msg->buf; | |
425 | ||
426 | mlog_entry("%p %u %p", msg, len, data); | |
427 | ||
428 | if (!dlm_grab(dlm)) | |
429 | return 0; | |
430 | ||
431 | node = exit_msg->node_idx; | |
432 | ||
781ee3e2 | 433 | printk(KERN_INFO "ocfs2_dlm: Node %u leaves domain %s\n", node, dlm->name); |
6714d8e8 KH |
434 | |
435 | spin_lock(&dlm->spinlock); | |
436 | clear_bit(node, dlm->domain_map); | |
437 | __dlm_print_nodes(dlm); | |
438 | ||
439 | /* notify anything attached to the heartbeat events */ | |
440 | dlm_hb_event_notify_attached(dlm, node, 0); | |
441 | ||
442 | spin_unlock(&dlm->spinlock); | |
443 | ||
444 | dlm_put(dlm); | |
445 | ||
446 | return 0; | |
447 | } | |
448 | ||
449 | static int dlm_send_one_domain_exit(struct dlm_ctxt *dlm, | |
450 | unsigned int node) | |
451 | { | |
452 | int status; | |
453 | struct dlm_exit_domain leave_msg; | |
454 | ||
455 | mlog(0, "Asking node %u if we can leave the domain %s me = %u\n", | |
456 | node, dlm->name, dlm->node_num); | |
457 | ||
458 | memset(&leave_msg, 0, sizeof(leave_msg)); | |
459 | leave_msg.node_idx = dlm->node_num; | |
460 | ||
461 | status = o2net_send_message(DLM_EXIT_DOMAIN_MSG, dlm->key, | |
462 | &leave_msg, sizeof(leave_msg), node, | |
463 | NULL); | |
464 | ||
465 | mlog(0, "status return %d from o2net_send_message\n", status); | |
466 | ||
467 | return status; | |
468 | } | |
469 | ||
470 | ||
471 | static void dlm_leave_domain(struct dlm_ctxt *dlm) | |
472 | { | |
473 | int node, clear_node, status; | |
474 | ||
475 | /* At this point we've migrated away all our locks and won't | |
476 | * accept mastership of new ones. The dlm is responsible for | |
477 | * almost nothing now. We make sure not to confuse any joining | |
478 | * nodes and then commence shutdown procedure. */ | |
479 | ||
480 | spin_lock(&dlm->spinlock); | |
481 | /* Clear ourselves from the domain map */ | |
482 | clear_bit(dlm->node_num, dlm->domain_map); | |
483 | while ((node = find_next_bit(dlm->domain_map, O2NM_MAX_NODES, | |
484 | 0)) < O2NM_MAX_NODES) { | |
485 | /* Drop the dlm spinlock. This is safe wrt the domain_map. | |
486 | * -nodes cannot be added now as the | |
487 | * query_join_handlers knows to respond with OK_NO_MAP | |
488 | * -we catch the right network errors if a node is | |
489 | * removed from the map while we're sending him the | |
490 | * exit message. */ | |
491 | spin_unlock(&dlm->spinlock); | |
492 | ||
493 | clear_node = 1; | |
494 | ||
495 | status = dlm_send_one_domain_exit(dlm, node); | |
496 | if (status < 0 && | |
497 | status != -ENOPROTOOPT && | |
498 | status != -ENOTCONN) { | |
499 | mlog(ML_NOTICE, "Error %d sending domain exit message " | |
500 | "to node %d\n", status, node); | |
501 | ||
502 | /* Not sure what to do here but lets sleep for | |
503 | * a bit in case this was a transient | |
504 | * error... */ | |
505 | msleep(DLM_DOMAIN_BACKOFF_MS); | |
506 | clear_node = 0; | |
507 | } | |
508 | ||
509 | spin_lock(&dlm->spinlock); | |
510 | /* If we're not clearing the node bit then we intend | |
511 | * to loop back around to try again. */ | |
512 | if (clear_node) | |
513 | clear_bit(node, dlm->domain_map); | |
514 | } | |
515 | spin_unlock(&dlm->spinlock); | |
516 | } | |
517 | ||
518 | int dlm_joined(struct dlm_ctxt *dlm) | |
519 | { | |
520 | int ret = 0; | |
521 | ||
522 | spin_lock(&dlm_domain_lock); | |
523 | ||
524 | if (dlm->dlm_state == DLM_CTXT_JOINED) | |
525 | ret = 1; | |
526 | ||
527 | spin_unlock(&dlm_domain_lock); | |
528 | ||
529 | return ret; | |
530 | } | |
531 | ||
532 | int dlm_shutting_down(struct dlm_ctxt *dlm) | |
533 | { | |
534 | int ret = 0; | |
535 | ||
536 | spin_lock(&dlm_domain_lock); | |
537 | ||
538 | if (dlm->dlm_state == DLM_CTXT_IN_SHUTDOWN) | |
539 | ret = 1; | |
540 | ||
541 | spin_unlock(&dlm_domain_lock); | |
542 | ||
543 | return ret; | |
544 | } | |
545 | ||
546 | void dlm_unregister_domain(struct dlm_ctxt *dlm) | |
547 | { | |
548 | int leave = 0; | |
549 | ||
550 | spin_lock(&dlm_domain_lock); | |
551 | BUG_ON(dlm->dlm_state != DLM_CTXT_JOINED); | |
552 | BUG_ON(!dlm->num_joins); | |
553 | ||
554 | dlm->num_joins--; | |
555 | if (!dlm->num_joins) { | |
556 | /* We mark it "in shutdown" now so new register | |
557 | * requests wait until we've completely left the | |
558 | * domain. Don't use DLM_CTXT_LEAVING yet as we still | |
559 | * want new domain joins to communicate with us at | |
560 | * least until we've completed migration of our | |
561 | * resources. */ | |
562 | dlm->dlm_state = DLM_CTXT_IN_SHUTDOWN; | |
563 | leave = 1; | |
564 | } | |
565 | spin_unlock(&dlm_domain_lock); | |
566 | ||
567 | if (leave) { | |
568 | mlog(0, "shutting down domain %s\n", dlm->name); | |
569 | ||
570 | /* We changed dlm state, notify the thread */ | |
571 | dlm_kick_thread(dlm, NULL); | |
572 | ||
573 | dlm_migrate_all_locks(dlm); | |
574 | dlm_mark_domain_leaving(dlm); | |
575 | dlm_leave_domain(dlm); | |
576 | dlm_complete_dlm_shutdown(dlm); | |
577 | } | |
578 | dlm_put(dlm); | |
579 | } | |
580 | EXPORT_SYMBOL_GPL(dlm_unregister_domain); | |
581 | ||
582 | static int dlm_query_join_handler(struct o2net_msg *msg, u32 len, void *data) | |
583 | { | |
584 | struct dlm_query_join_request *query; | |
585 | enum dlm_query_join_response response; | |
586 | struct dlm_ctxt *dlm = NULL; | |
587 | ||
588 | query = (struct dlm_query_join_request *) msg->buf; | |
589 | ||
590 | mlog(0, "node %u wants to join domain %s\n", query->node_idx, | |
591 | query->domain); | |
592 | ||
593 | /* | |
594 | * If heartbeat doesn't consider the node live, tell it | |
595 | * to back off and try again. This gives heartbeat a chance | |
596 | * to catch up. | |
597 | */ | |
598 | if (!o2hb_check_node_heartbeating(query->node_idx)) { | |
599 | mlog(0, "node %u is not in our live map yet\n", | |
600 | query->node_idx); | |
601 | ||
602 | response = JOIN_DISALLOW; | |
603 | goto respond; | |
604 | } | |
605 | ||
606 | response = JOIN_OK_NO_MAP; | |
607 | ||
608 | spin_lock(&dlm_domain_lock); | |
609 | dlm = __dlm_lookup_domain_full(query->domain, query->name_len); | |
610 | /* Once the dlm ctxt is marked as leaving then we don't want | |
e2faea4c KH |
611 | * to be put in someone's domain map. |
612 | * Also, explicitly disallow joining at certain troublesome | |
613 | * times (ie. during recovery). */ | |
6714d8e8 | 614 | if (dlm && dlm->dlm_state != DLM_CTXT_LEAVING) { |
e2faea4c | 615 | int bit = query->node_idx; |
6714d8e8 KH |
616 | spin_lock(&dlm->spinlock); |
617 | ||
618 | if (dlm->dlm_state == DLM_CTXT_NEW && | |
619 | dlm->joining_node == DLM_LOCK_RES_OWNER_UNKNOWN) { | |
620 | /*If this is a brand new context and we | |
621 | * haven't started our join process yet, then | |
622 | * the other node won the race. */ | |
623 | response = JOIN_OK_NO_MAP; | |
624 | } else if (dlm->joining_node != DLM_LOCK_RES_OWNER_UNKNOWN) { | |
625 | /* Disallow parallel joins. */ | |
626 | response = JOIN_DISALLOW; | |
e2faea4c KH |
627 | } else if (dlm->reco.state & DLM_RECO_STATE_ACTIVE) { |
628 | mlog(ML_NOTICE, "node %u trying to join, but recovery " | |
629 | "is ongoing.\n", bit); | |
630 | response = JOIN_DISALLOW; | |
631 | } else if (test_bit(bit, dlm->recovery_map)) { | |
632 | mlog(ML_NOTICE, "node %u trying to join, but it " | |
633 | "still needs recovery.\n", bit); | |
634 | response = JOIN_DISALLOW; | |
635 | } else if (test_bit(bit, dlm->domain_map)) { | |
636 | mlog(ML_NOTICE, "node %u trying to join, but it " | |
637 | "is still in the domain! needs recovery?\n", | |
638 | bit); | |
639 | response = JOIN_DISALLOW; | |
6714d8e8 KH |
640 | } else { |
641 | /* Alright we're fully a part of this domain | |
642 | * so we keep some state as to who's joining | |
643 | * and indicate to him that needs to be fixed | |
644 | * up. */ | |
645 | response = JOIN_OK; | |
646 | __dlm_set_joining_node(dlm, query->node_idx); | |
647 | } | |
648 | ||
649 | spin_unlock(&dlm->spinlock); | |
650 | } | |
651 | spin_unlock(&dlm_domain_lock); | |
652 | ||
653 | respond: | |
654 | mlog(0, "We respond with %u\n", response); | |
655 | ||
656 | return response; | |
657 | } | |
658 | ||
659 | static int dlm_assert_joined_handler(struct o2net_msg *msg, u32 len, void *data) | |
660 | { | |
661 | struct dlm_assert_joined *assert; | |
662 | struct dlm_ctxt *dlm = NULL; | |
663 | ||
664 | assert = (struct dlm_assert_joined *) msg->buf; | |
665 | ||
666 | mlog(0, "node %u asserts join on domain %s\n", assert->node_idx, | |
667 | assert->domain); | |
668 | ||
669 | spin_lock(&dlm_domain_lock); | |
670 | dlm = __dlm_lookup_domain_full(assert->domain, assert->name_len); | |
671 | /* XXX should we consider no dlm ctxt an error? */ | |
672 | if (dlm) { | |
673 | spin_lock(&dlm->spinlock); | |
674 | ||
675 | /* Alright, this node has officially joined our | |
676 | * domain. Set him in the map and clean up our | |
677 | * leftover join state. */ | |
678 | BUG_ON(dlm->joining_node != assert->node_idx); | |
679 | set_bit(assert->node_idx, dlm->domain_map); | |
680 | __dlm_set_joining_node(dlm, DLM_LOCK_RES_OWNER_UNKNOWN); | |
681 | ||
781ee3e2 SM |
682 | printk(KERN_INFO "ocfs2_dlm: Node %u joins domain %s\n", |
683 | assert->node_idx, dlm->name); | |
6714d8e8 KH |
684 | __dlm_print_nodes(dlm); |
685 | ||
686 | /* notify anything attached to the heartbeat events */ | |
687 | dlm_hb_event_notify_attached(dlm, assert->node_idx, 1); | |
688 | ||
689 | spin_unlock(&dlm->spinlock); | |
690 | } | |
691 | spin_unlock(&dlm_domain_lock); | |
692 | ||
693 | return 0; | |
694 | } | |
695 | ||
696 | static int dlm_cancel_join_handler(struct o2net_msg *msg, u32 len, void *data) | |
697 | { | |
698 | struct dlm_cancel_join *cancel; | |
699 | struct dlm_ctxt *dlm = NULL; | |
700 | ||
701 | cancel = (struct dlm_cancel_join *) msg->buf; | |
702 | ||
703 | mlog(0, "node %u cancels join on domain %s\n", cancel->node_idx, | |
704 | cancel->domain); | |
705 | ||
706 | spin_lock(&dlm_domain_lock); | |
707 | dlm = __dlm_lookup_domain_full(cancel->domain, cancel->name_len); | |
708 | ||
709 | if (dlm) { | |
710 | spin_lock(&dlm->spinlock); | |
711 | ||
712 | /* Yikes, this guy wants to cancel his join. No | |
713 | * problem, we simply cleanup our join state. */ | |
714 | BUG_ON(dlm->joining_node != cancel->node_idx); | |
715 | __dlm_set_joining_node(dlm, DLM_LOCK_RES_OWNER_UNKNOWN); | |
716 | ||
717 | spin_unlock(&dlm->spinlock); | |
718 | } | |
719 | spin_unlock(&dlm_domain_lock); | |
720 | ||
721 | return 0; | |
722 | } | |
723 | ||
724 | static int dlm_send_one_join_cancel(struct dlm_ctxt *dlm, | |
725 | unsigned int node) | |
726 | { | |
727 | int status; | |
728 | struct dlm_cancel_join cancel_msg; | |
729 | ||
730 | memset(&cancel_msg, 0, sizeof(cancel_msg)); | |
731 | cancel_msg.node_idx = dlm->node_num; | |
732 | cancel_msg.name_len = strlen(dlm->name); | |
733 | memcpy(cancel_msg.domain, dlm->name, cancel_msg.name_len); | |
734 | ||
735 | status = o2net_send_message(DLM_CANCEL_JOIN_MSG, DLM_MOD_KEY, | |
736 | &cancel_msg, sizeof(cancel_msg), node, | |
737 | NULL); | |
738 | if (status < 0) { | |
739 | mlog_errno(status); | |
740 | goto bail; | |
741 | } | |
742 | ||
743 | bail: | |
744 | return status; | |
745 | } | |
746 | ||
747 | /* map_size should be in bytes. */ | |
748 | static int dlm_send_join_cancels(struct dlm_ctxt *dlm, | |
749 | unsigned long *node_map, | |
750 | unsigned int map_size) | |
751 | { | |
752 | int status, tmpstat; | |
753 | unsigned int node; | |
754 | ||
755 | if (map_size != (BITS_TO_LONGS(O2NM_MAX_NODES) * | |
756 | sizeof(unsigned long))) { | |
757 | mlog(ML_ERROR, | |
758 | "map_size %u != BITS_TO_LONGS(O2NM_MAX_NODES) %u\n", | |
759 | map_size, BITS_TO_LONGS(O2NM_MAX_NODES)); | |
760 | return -EINVAL; | |
761 | } | |
762 | ||
763 | status = 0; | |
764 | node = -1; | |
765 | while ((node = find_next_bit(node_map, O2NM_MAX_NODES, | |
766 | node + 1)) < O2NM_MAX_NODES) { | |
767 | if (node == dlm->node_num) | |
768 | continue; | |
769 | ||
770 | tmpstat = dlm_send_one_join_cancel(dlm, node); | |
771 | if (tmpstat) { | |
772 | mlog(ML_ERROR, "Error return %d cancelling join on " | |
773 | "node %d\n", tmpstat, node); | |
774 | if (!status) | |
775 | status = tmpstat; | |
776 | } | |
777 | } | |
778 | ||
779 | if (status) | |
780 | mlog_errno(status); | |
781 | return status; | |
782 | } | |
783 | ||
784 | static int dlm_request_join(struct dlm_ctxt *dlm, | |
785 | int node, | |
786 | enum dlm_query_join_response *response) | |
787 | { | |
788 | int status, retval; | |
789 | struct dlm_query_join_request join_msg; | |
790 | ||
791 | mlog(0, "querying node %d\n", node); | |
792 | ||
793 | memset(&join_msg, 0, sizeof(join_msg)); | |
794 | join_msg.node_idx = dlm->node_num; | |
795 | join_msg.name_len = strlen(dlm->name); | |
796 | memcpy(join_msg.domain, dlm->name, join_msg.name_len); | |
797 | ||
798 | status = o2net_send_message(DLM_QUERY_JOIN_MSG, DLM_MOD_KEY, &join_msg, | |
799 | sizeof(join_msg), node, &retval); | |
800 | if (status < 0 && status != -ENOPROTOOPT) { | |
801 | mlog_errno(status); | |
802 | goto bail; | |
803 | } | |
804 | ||
805 | /* -ENOPROTOOPT from the net code means the other side isn't | |
806 | listening for our message type -- that's fine, it means | |
807 | his dlm isn't up, so we can consider him a 'yes' but not | |
808 | joined into the domain. */ | |
809 | if (status == -ENOPROTOOPT) { | |
810 | status = 0; | |
811 | *response = JOIN_OK_NO_MAP; | |
812 | } else if (retval == JOIN_DISALLOW || | |
813 | retval == JOIN_OK || | |
814 | retval == JOIN_OK_NO_MAP) { | |
815 | *response = retval; | |
816 | } else { | |
817 | status = -EINVAL; | |
818 | mlog(ML_ERROR, "invalid response %d from node %u\n", retval, | |
819 | node); | |
820 | } | |
821 | ||
822 | mlog(0, "status %d, node %d response is %d\n", status, node, | |
823 | *response); | |
824 | ||
825 | bail: | |
826 | return status; | |
827 | } | |
828 | ||
829 | static int dlm_send_one_join_assert(struct dlm_ctxt *dlm, | |
830 | unsigned int node) | |
831 | { | |
832 | int status; | |
833 | struct dlm_assert_joined assert_msg; | |
834 | ||
835 | mlog(0, "Sending join assert to node %u\n", node); | |
836 | ||
837 | memset(&assert_msg, 0, sizeof(assert_msg)); | |
838 | assert_msg.node_idx = dlm->node_num; | |
839 | assert_msg.name_len = strlen(dlm->name); | |
840 | memcpy(assert_msg.domain, dlm->name, assert_msg.name_len); | |
841 | ||
842 | status = o2net_send_message(DLM_ASSERT_JOINED_MSG, DLM_MOD_KEY, | |
843 | &assert_msg, sizeof(assert_msg), node, | |
844 | NULL); | |
845 | if (status < 0) | |
846 | mlog_errno(status); | |
847 | ||
848 | return status; | |
849 | } | |
850 | ||
851 | static void dlm_send_join_asserts(struct dlm_ctxt *dlm, | |
852 | unsigned long *node_map) | |
853 | { | |
854 | int status, node, live; | |
855 | ||
856 | status = 0; | |
857 | node = -1; | |
858 | while ((node = find_next_bit(node_map, O2NM_MAX_NODES, | |
859 | node + 1)) < O2NM_MAX_NODES) { | |
860 | if (node == dlm->node_num) | |
861 | continue; | |
862 | ||
863 | do { | |
864 | /* It is very important that this message be | |
865 | * received so we spin until either the node | |
866 | * has died or it gets the message. */ | |
867 | status = dlm_send_one_join_assert(dlm, node); | |
868 | ||
869 | spin_lock(&dlm->spinlock); | |
870 | live = test_bit(node, dlm->live_nodes_map); | |
871 | spin_unlock(&dlm->spinlock); | |
872 | ||
873 | if (status) { | |
874 | mlog(ML_ERROR, "Error return %d asserting " | |
875 | "join on node %d\n", status, node); | |
876 | ||
877 | /* give us some time between errors... */ | |
878 | if (live) | |
879 | msleep(DLM_DOMAIN_BACKOFF_MS); | |
880 | } | |
881 | } while (status && live); | |
882 | } | |
883 | } | |
884 | ||
885 | struct domain_join_ctxt { | |
886 | unsigned long live_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; | |
887 | unsigned long yes_resp_map[BITS_TO_LONGS(O2NM_MAX_NODES)]; | |
888 | }; | |
889 | ||
890 | static int dlm_should_restart_join(struct dlm_ctxt *dlm, | |
891 | struct domain_join_ctxt *ctxt, | |
892 | enum dlm_query_join_response response) | |
893 | { | |
894 | int ret; | |
895 | ||
896 | if (response == JOIN_DISALLOW) { | |
897 | mlog(0, "Latest response of disallow -- should restart\n"); | |
898 | return 1; | |
899 | } | |
900 | ||
901 | spin_lock(&dlm->spinlock); | |
902 | /* For now, we restart the process if the node maps have | |
903 | * changed at all */ | |
904 | ret = memcmp(ctxt->live_map, dlm->live_nodes_map, | |
905 | sizeof(dlm->live_nodes_map)); | |
906 | spin_unlock(&dlm->spinlock); | |
907 | ||
908 | if (ret) | |
909 | mlog(0, "Node maps changed -- should restart\n"); | |
910 | ||
911 | return ret; | |
912 | } | |
913 | ||
914 | static int dlm_try_to_join_domain(struct dlm_ctxt *dlm) | |
915 | { | |
916 | int status = 0, tmpstat, node; | |
917 | struct domain_join_ctxt *ctxt; | |
918 | enum dlm_query_join_response response; | |
919 | ||
920 | mlog_entry("%p", dlm); | |
921 | ||
922 | ctxt = kcalloc(1, sizeof(*ctxt), GFP_KERNEL); | |
923 | if (!ctxt) { | |
924 | status = -ENOMEM; | |
925 | mlog_errno(status); | |
926 | goto bail; | |
927 | } | |
928 | ||
929 | /* group sem locking should work for us here -- we're already | |
930 | * registered for heartbeat events so filling this should be | |
931 | * atomic wrt getting those handlers called. */ | |
932 | o2hb_fill_node_map(dlm->live_nodes_map, sizeof(dlm->live_nodes_map)); | |
933 | ||
934 | spin_lock(&dlm->spinlock); | |
935 | memcpy(ctxt->live_map, dlm->live_nodes_map, sizeof(ctxt->live_map)); | |
936 | ||
937 | __dlm_set_joining_node(dlm, dlm->node_num); | |
938 | ||
939 | spin_unlock(&dlm->spinlock); | |
940 | ||
941 | node = -1; | |
942 | while ((node = find_next_bit(ctxt->live_map, O2NM_MAX_NODES, | |
943 | node + 1)) < O2NM_MAX_NODES) { | |
944 | if (node == dlm->node_num) | |
945 | continue; | |
946 | ||
947 | status = dlm_request_join(dlm, node, &response); | |
948 | if (status < 0) { | |
949 | mlog_errno(status); | |
950 | goto bail; | |
951 | } | |
952 | ||
953 | /* Ok, either we got a response or the node doesn't have a | |
954 | * dlm up. */ | |
955 | if (response == JOIN_OK) | |
956 | set_bit(node, ctxt->yes_resp_map); | |
957 | ||
958 | if (dlm_should_restart_join(dlm, ctxt, response)) { | |
959 | status = -EAGAIN; | |
960 | goto bail; | |
961 | } | |
962 | } | |
963 | ||
964 | mlog(0, "Yay, done querying nodes!\n"); | |
965 | ||
966 | /* Yay, everyone agree's we can join the domain. My domain is | |
967 | * comprised of all nodes who were put in the | |
968 | * yes_resp_map. Copy that into our domain map and send a join | |
969 | * assert message to clean up everyone elses state. */ | |
970 | spin_lock(&dlm->spinlock); | |
971 | memcpy(dlm->domain_map, ctxt->yes_resp_map, | |
972 | sizeof(ctxt->yes_resp_map)); | |
973 | set_bit(dlm->node_num, dlm->domain_map); | |
974 | spin_unlock(&dlm->spinlock); | |
975 | ||
976 | dlm_send_join_asserts(dlm, ctxt->yes_resp_map); | |
977 | ||
978 | /* Joined state *must* be set before the joining node | |
979 | * information, otherwise the query_join handler may read no | |
980 | * current joiner but a state of NEW and tell joining nodes | |
981 | * we're not in the domain. */ | |
982 | spin_lock(&dlm_domain_lock); | |
983 | dlm->dlm_state = DLM_CTXT_JOINED; | |
984 | dlm->num_joins++; | |
985 | spin_unlock(&dlm_domain_lock); | |
986 | ||
987 | bail: | |
988 | spin_lock(&dlm->spinlock); | |
989 | __dlm_set_joining_node(dlm, DLM_LOCK_RES_OWNER_UNKNOWN); | |
990 | if (!status) | |
991 | __dlm_print_nodes(dlm); | |
992 | spin_unlock(&dlm->spinlock); | |
993 | ||
994 | if (ctxt) { | |
995 | /* Do we need to send a cancel message to any nodes? */ | |
996 | if (status < 0) { | |
997 | tmpstat = dlm_send_join_cancels(dlm, | |
998 | ctxt->yes_resp_map, | |
999 | sizeof(ctxt->yes_resp_map)); | |
1000 | if (tmpstat < 0) | |
1001 | mlog_errno(tmpstat); | |
1002 | } | |
1003 | kfree(ctxt); | |
1004 | } | |
1005 | ||
1006 | mlog(0, "returning %d\n", status); | |
1007 | return status; | |
1008 | } | |
1009 | ||
1010 | static void dlm_unregister_domain_handlers(struct dlm_ctxt *dlm) | |
1011 | { | |
1012 | o2hb_unregister_callback(&dlm->dlm_hb_up); | |
1013 | o2hb_unregister_callback(&dlm->dlm_hb_down); | |
1014 | o2net_unregister_handler_list(&dlm->dlm_domain_handlers); | |
1015 | } | |
1016 | ||
1017 | static int dlm_register_domain_handlers(struct dlm_ctxt *dlm) | |
1018 | { | |
1019 | int status; | |
1020 | ||
1021 | mlog(0, "registering handlers.\n"); | |
1022 | ||
1023 | o2hb_setup_callback(&dlm->dlm_hb_down, O2HB_NODE_DOWN_CB, | |
1024 | dlm_hb_node_down_cb, dlm, DLM_HB_NODE_DOWN_PRI); | |
1025 | status = o2hb_register_callback(&dlm->dlm_hb_down); | |
1026 | if (status) | |
1027 | goto bail; | |
1028 | ||
1029 | o2hb_setup_callback(&dlm->dlm_hb_up, O2HB_NODE_UP_CB, | |
1030 | dlm_hb_node_up_cb, dlm, DLM_HB_NODE_UP_PRI); | |
1031 | status = o2hb_register_callback(&dlm->dlm_hb_up); | |
1032 | if (status) | |
1033 | goto bail; | |
1034 | ||
1035 | status = o2net_register_handler(DLM_MASTER_REQUEST_MSG, dlm->key, | |
1036 | sizeof(struct dlm_master_request), | |
1037 | dlm_master_request_handler, | |
1038 | dlm, &dlm->dlm_domain_handlers); | |
1039 | if (status) | |
1040 | goto bail; | |
1041 | ||
1042 | status = o2net_register_handler(DLM_ASSERT_MASTER_MSG, dlm->key, | |
1043 | sizeof(struct dlm_assert_master), | |
1044 | dlm_assert_master_handler, | |
1045 | dlm, &dlm->dlm_domain_handlers); | |
1046 | if (status) | |
1047 | goto bail; | |
1048 | ||
1049 | status = o2net_register_handler(DLM_CREATE_LOCK_MSG, dlm->key, | |
1050 | sizeof(struct dlm_create_lock), | |
1051 | dlm_create_lock_handler, | |
1052 | dlm, &dlm->dlm_domain_handlers); | |
1053 | if (status) | |
1054 | goto bail; | |
1055 | ||
1056 | status = o2net_register_handler(DLM_CONVERT_LOCK_MSG, dlm->key, | |
1057 | DLM_CONVERT_LOCK_MAX_LEN, | |
1058 | dlm_convert_lock_handler, | |
1059 | dlm, &dlm->dlm_domain_handlers); | |
1060 | if (status) | |
1061 | goto bail; | |
1062 | ||
1063 | status = o2net_register_handler(DLM_UNLOCK_LOCK_MSG, dlm->key, | |
1064 | DLM_UNLOCK_LOCK_MAX_LEN, | |
1065 | dlm_unlock_lock_handler, | |
1066 | dlm, &dlm->dlm_domain_handlers); | |
1067 | if (status) | |
1068 | goto bail; | |
1069 | ||
1070 | status = o2net_register_handler(DLM_PROXY_AST_MSG, dlm->key, | |
1071 | DLM_PROXY_AST_MAX_LEN, | |
1072 | dlm_proxy_ast_handler, | |
1073 | dlm, &dlm->dlm_domain_handlers); | |
1074 | if (status) | |
1075 | goto bail; | |
1076 | ||
1077 | status = o2net_register_handler(DLM_EXIT_DOMAIN_MSG, dlm->key, | |
1078 | sizeof(struct dlm_exit_domain), | |
1079 | dlm_exit_domain_handler, | |
1080 | dlm, &dlm->dlm_domain_handlers); | |
1081 | if (status) | |
1082 | goto bail; | |
1083 | ||
1084 | status = o2net_register_handler(DLM_MIGRATE_REQUEST_MSG, dlm->key, | |
1085 | sizeof(struct dlm_migrate_request), | |
1086 | dlm_migrate_request_handler, | |
1087 | dlm, &dlm->dlm_domain_handlers); | |
1088 | if (status) | |
1089 | goto bail; | |
1090 | ||
1091 | status = o2net_register_handler(DLM_MIG_LOCKRES_MSG, dlm->key, | |
1092 | DLM_MIG_LOCKRES_MAX_LEN, | |
1093 | dlm_mig_lockres_handler, | |
1094 | dlm, &dlm->dlm_domain_handlers); | |
1095 | if (status) | |
1096 | goto bail; | |
1097 | ||
1098 | status = o2net_register_handler(DLM_MASTER_REQUERY_MSG, dlm->key, | |
1099 | sizeof(struct dlm_master_requery), | |
1100 | dlm_master_requery_handler, | |
1101 | dlm, &dlm->dlm_domain_handlers); | |
1102 | if (status) | |
1103 | goto bail; | |
1104 | ||
1105 | status = o2net_register_handler(DLM_LOCK_REQUEST_MSG, dlm->key, | |
1106 | sizeof(struct dlm_lock_request), | |
1107 | dlm_request_all_locks_handler, | |
1108 | dlm, &dlm->dlm_domain_handlers); | |
1109 | if (status) | |
1110 | goto bail; | |
1111 | ||
1112 | status = o2net_register_handler(DLM_RECO_DATA_DONE_MSG, dlm->key, | |
1113 | sizeof(struct dlm_reco_data_done), | |
1114 | dlm_reco_data_done_handler, | |
1115 | dlm, &dlm->dlm_domain_handlers); | |
1116 | if (status) | |
1117 | goto bail; | |
1118 | ||
1119 | status = o2net_register_handler(DLM_BEGIN_RECO_MSG, dlm->key, | |
1120 | sizeof(struct dlm_begin_reco), | |
1121 | dlm_begin_reco_handler, | |
1122 | dlm, &dlm->dlm_domain_handlers); | |
1123 | if (status) | |
1124 | goto bail; | |
1125 | ||
1126 | status = o2net_register_handler(DLM_FINALIZE_RECO_MSG, dlm->key, | |
1127 | sizeof(struct dlm_finalize_reco), | |
1128 | dlm_finalize_reco_handler, | |
1129 | dlm, &dlm->dlm_domain_handlers); | |
1130 | if (status) | |
1131 | goto bail; | |
1132 | ||
1133 | bail: | |
1134 | if (status) | |
1135 | dlm_unregister_domain_handlers(dlm); | |
1136 | ||
1137 | return status; | |
1138 | } | |
1139 | ||
1140 | static int dlm_join_domain(struct dlm_ctxt *dlm) | |
1141 | { | |
1142 | int status; | |
1143 | ||
1144 | BUG_ON(!dlm); | |
1145 | ||
1146 | mlog(0, "Join domain %s\n", dlm->name); | |
1147 | ||
1148 | status = dlm_register_domain_handlers(dlm); | |
1149 | if (status) { | |
1150 | mlog_errno(status); | |
1151 | goto bail; | |
1152 | } | |
1153 | ||
1154 | status = dlm_launch_thread(dlm); | |
1155 | if (status < 0) { | |
1156 | mlog_errno(status); | |
1157 | goto bail; | |
1158 | } | |
1159 | ||
1160 | status = dlm_launch_recovery_thread(dlm); | |
1161 | if (status < 0) { | |
1162 | mlog_errno(status); | |
1163 | goto bail; | |
1164 | } | |
1165 | ||
3156d267 KH |
1166 | dlm->dlm_worker = create_singlethread_workqueue("dlm_wq"); |
1167 | if (!dlm->dlm_worker) { | |
1168 | status = -ENOMEM; | |
1169 | mlog_errno(status); | |
1170 | goto bail; | |
1171 | } | |
1172 | ||
6714d8e8 KH |
1173 | do { |
1174 | unsigned int backoff; | |
1175 | status = dlm_try_to_join_domain(dlm); | |
1176 | ||
1177 | /* If we're racing another node to the join, then we | |
1178 | * need to back off temporarily and let them | |
1179 | * complete. */ | |
1180 | if (status == -EAGAIN) { | |
1181 | if (signal_pending(current)) { | |
1182 | status = -ERESTARTSYS; | |
1183 | goto bail; | |
1184 | } | |
1185 | ||
1186 | /* | |
1187 | * <chip> After you! | |
1188 | * <dale> No, after you! | |
1189 | * <chip> I insist! | |
1190 | * <dale> But you first! | |
1191 | * ... | |
1192 | */ | |
1193 | backoff = (unsigned int)(jiffies & 0x3); | |
1194 | backoff *= DLM_DOMAIN_BACKOFF_MS; | |
1195 | mlog(0, "backoff %d\n", backoff); | |
1196 | msleep(backoff); | |
1197 | } | |
1198 | } while (status == -EAGAIN); | |
1199 | ||
1200 | if (status < 0) { | |
1201 | mlog_errno(status); | |
1202 | goto bail; | |
1203 | } | |
1204 | ||
1205 | status = 0; | |
1206 | bail: | |
1207 | wake_up(&dlm_domain_events); | |
1208 | ||
1209 | if (status) { | |
1210 | dlm_unregister_domain_handlers(dlm); | |
1211 | dlm_complete_thread(dlm); | |
1212 | dlm_complete_recovery_thread(dlm); | |
3156d267 | 1213 | dlm_destroy_dlm_worker(dlm); |
6714d8e8 KH |
1214 | } |
1215 | ||
1216 | return status; | |
1217 | } | |
1218 | ||
1219 | static struct dlm_ctxt *dlm_alloc_ctxt(const char *domain, | |
1220 | u32 key) | |
1221 | { | |
1222 | int i; | |
1223 | struct dlm_ctxt *dlm = NULL; | |
1224 | ||
1225 | dlm = kcalloc(1, sizeof(*dlm), GFP_KERNEL); | |
1226 | if (!dlm) { | |
1227 | mlog_errno(-ENOMEM); | |
1228 | goto leave; | |
1229 | } | |
1230 | ||
1231 | dlm->name = kmalloc(strlen(domain) + 1, GFP_KERNEL); | |
1232 | if (dlm->name == NULL) { | |
1233 | mlog_errno(-ENOMEM); | |
1234 | kfree(dlm); | |
1235 | dlm = NULL; | |
1236 | goto leave; | |
1237 | } | |
1238 | ||
03d864c0 | 1239 | dlm->lockres_hash = (struct hlist_head **)dlm_alloc_pagevec(DLM_HASH_PAGES); |
81f2094a | 1240 | if (!dlm->lockres_hash) { |
6714d8e8 KH |
1241 | mlog_errno(-ENOMEM); |
1242 | kfree(dlm->name); | |
1243 | kfree(dlm); | |
1244 | dlm = NULL; | |
1245 | goto leave; | |
1246 | } | |
6714d8e8 | 1247 | |
03d864c0 DP |
1248 | for (i = 0; i < DLM_HASH_BUCKETS; i++) |
1249 | INIT_HLIST_HEAD(dlm_lockres_hash(dlm, i)); | |
6714d8e8 KH |
1250 | |
1251 | strcpy(dlm->name, domain); | |
1252 | dlm->key = key; | |
1253 | dlm->node_num = o2nm_this_node(); | |
1254 | ||
1255 | spin_lock_init(&dlm->spinlock); | |
1256 | spin_lock_init(&dlm->master_lock); | |
1257 | spin_lock_init(&dlm->ast_lock); | |
1258 | INIT_LIST_HEAD(&dlm->list); | |
1259 | INIT_LIST_HEAD(&dlm->dirty_list); | |
1260 | INIT_LIST_HEAD(&dlm->reco.resources); | |
1261 | INIT_LIST_HEAD(&dlm->reco.received); | |
1262 | INIT_LIST_HEAD(&dlm->reco.node_data); | |
1263 | INIT_LIST_HEAD(&dlm->purge_list); | |
1264 | INIT_LIST_HEAD(&dlm->dlm_domain_handlers); | |
1265 | dlm->reco.state = 0; | |
1266 | ||
1267 | INIT_LIST_HEAD(&dlm->pending_asts); | |
1268 | INIT_LIST_HEAD(&dlm->pending_basts); | |
1269 | ||
1270 | mlog(0, "dlm->recovery_map=%p, &(dlm->recovery_map[0])=%p\n", | |
1271 | dlm->recovery_map, &(dlm->recovery_map[0])); | |
1272 | ||
1273 | memset(dlm->recovery_map, 0, sizeof(dlm->recovery_map)); | |
1274 | memset(dlm->live_nodes_map, 0, sizeof(dlm->live_nodes_map)); | |
1275 | memset(dlm->domain_map, 0, sizeof(dlm->domain_map)); | |
1276 | ||
1277 | dlm->dlm_thread_task = NULL; | |
1278 | dlm->dlm_reco_thread_task = NULL; | |
3156d267 | 1279 | dlm->dlm_worker = NULL; |
6714d8e8 KH |
1280 | init_waitqueue_head(&dlm->dlm_thread_wq); |
1281 | init_waitqueue_head(&dlm->dlm_reco_thread_wq); | |
1282 | init_waitqueue_head(&dlm->reco.event); | |
1283 | init_waitqueue_head(&dlm->ast_wq); | |
1284 | init_waitqueue_head(&dlm->migration_wq); | |
1285 | INIT_LIST_HEAD(&dlm->master_list); | |
1286 | INIT_LIST_HEAD(&dlm->mle_hb_events); | |
1287 | ||
1288 | dlm->joining_node = DLM_LOCK_RES_OWNER_UNKNOWN; | |
1289 | init_waitqueue_head(&dlm->dlm_join_events); | |
1290 | ||
1291 | dlm->reco.new_master = O2NM_INVALID_NODE_NUM; | |
1292 | dlm->reco.dead_node = O2NM_INVALID_NODE_NUM; | |
1293 | atomic_set(&dlm->local_resources, 0); | |
1294 | atomic_set(&dlm->remote_resources, 0); | |
1295 | atomic_set(&dlm->unknown_resources, 0); | |
1296 | ||
1297 | spin_lock_init(&dlm->work_lock); | |
1298 | INIT_LIST_HEAD(&dlm->work_list); | |
c4028958 | 1299 | INIT_WORK(&dlm->dispatched_work, dlm_dispatch_work); |
6714d8e8 KH |
1300 | |
1301 | kref_init(&dlm->dlm_refs); | |
1302 | dlm->dlm_state = DLM_CTXT_NEW; | |
1303 | ||
1304 | INIT_LIST_HEAD(&dlm->dlm_eviction_callbacks); | |
1305 | ||
1306 | mlog(0, "context init: refcount %u\n", | |
1307 | atomic_read(&dlm->dlm_refs.refcount)); | |
1308 | ||
1309 | leave: | |
1310 | return dlm; | |
1311 | } | |
1312 | ||
1313 | /* | |
1314 | * dlm_register_domain: one-time setup per "domain" | |
1315 | */ | |
1316 | struct dlm_ctxt * dlm_register_domain(const char *domain, | |
1317 | u32 key) | |
1318 | { | |
1319 | int ret; | |
1320 | struct dlm_ctxt *dlm = NULL; | |
1321 | struct dlm_ctxt *new_ctxt = NULL; | |
1322 | ||
1323 | if (strlen(domain) > O2NM_MAX_NAME_LEN) { | |
1324 | ret = -ENAMETOOLONG; | |
1325 | mlog(ML_ERROR, "domain name length too long\n"); | |
1326 | goto leave; | |
1327 | } | |
1328 | ||
1329 | if (!o2hb_check_local_node_heartbeating()) { | |
1330 | mlog(ML_ERROR, "the local node has not been configured, or is " | |
1331 | "not heartbeating\n"); | |
1332 | ret = -EPROTO; | |
1333 | goto leave; | |
1334 | } | |
1335 | ||
1336 | mlog(0, "register called for domain \"%s\"\n", domain); | |
1337 | ||
1338 | retry: | |
1339 | dlm = NULL; | |
1340 | if (signal_pending(current)) { | |
1341 | ret = -ERESTARTSYS; | |
1342 | mlog_errno(ret); | |
1343 | goto leave; | |
1344 | } | |
1345 | ||
1346 | spin_lock(&dlm_domain_lock); | |
1347 | ||
1348 | dlm = __dlm_lookup_domain(domain); | |
1349 | if (dlm) { | |
1350 | if (dlm->dlm_state != DLM_CTXT_JOINED) { | |
1351 | spin_unlock(&dlm_domain_lock); | |
1352 | ||
1353 | mlog(0, "This ctxt is not joined yet!\n"); | |
1354 | wait_event_interruptible(dlm_domain_events, | |
1355 | dlm_wait_on_domain_helper( | |
1356 | domain)); | |
1357 | goto retry; | |
1358 | } | |
1359 | ||
1360 | __dlm_get(dlm); | |
1361 | dlm->num_joins++; | |
1362 | ||
1363 | spin_unlock(&dlm_domain_lock); | |
1364 | ||
1365 | ret = 0; | |
1366 | goto leave; | |
1367 | } | |
1368 | ||
1369 | /* doesn't exist */ | |
1370 | if (!new_ctxt) { | |
1371 | spin_unlock(&dlm_domain_lock); | |
1372 | ||
1373 | new_ctxt = dlm_alloc_ctxt(domain, key); | |
1374 | if (new_ctxt) | |
1375 | goto retry; | |
1376 | ||
1377 | ret = -ENOMEM; | |
1378 | mlog_errno(ret); | |
1379 | goto leave; | |
1380 | } | |
1381 | ||
1382 | /* a little variable switch-a-roo here... */ | |
1383 | dlm = new_ctxt; | |
1384 | new_ctxt = NULL; | |
1385 | ||
1386 | /* add the new domain */ | |
1387 | list_add_tail(&dlm->list, &dlm_domains); | |
1388 | spin_unlock(&dlm_domain_lock); | |
1389 | ||
1390 | ret = dlm_join_domain(dlm); | |
1391 | if (ret) { | |
1392 | mlog_errno(ret); | |
1393 | dlm_put(dlm); | |
1394 | goto leave; | |
1395 | } | |
1396 | ||
1397 | ret = 0; | |
1398 | leave: | |
1399 | if (new_ctxt) | |
1400 | dlm_free_ctxt_mem(new_ctxt); | |
1401 | ||
1402 | if (ret < 0) | |
1403 | dlm = ERR_PTR(ret); | |
1404 | ||
1405 | return dlm; | |
1406 | } | |
1407 | EXPORT_SYMBOL_GPL(dlm_register_domain); | |
1408 | ||
1409 | static LIST_HEAD(dlm_join_handlers); | |
1410 | ||
1411 | static void dlm_unregister_net_handlers(void) | |
1412 | { | |
1413 | o2net_unregister_handler_list(&dlm_join_handlers); | |
1414 | } | |
1415 | ||
1416 | static int dlm_register_net_handlers(void) | |
1417 | { | |
1418 | int status = 0; | |
1419 | ||
1420 | status = o2net_register_handler(DLM_QUERY_JOIN_MSG, DLM_MOD_KEY, | |
1421 | sizeof(struct dlm_query_join_request), | |
1422 | dlm_query_join_handler, | |
1423 | NULL, &dlm_join_handlers); | |
1424 | if (status) | |
1425 | goto bail; | |
1426 | ||
1427 | status = o2net_register_handler(DLM_ASSERT_JOINED_MSG, DLM_MOD_KEY, | |
1428 | sizeof(struct dlm_assert_joined), | |
1429 | dlm_assert_joined_handler, | |
1430 | NULL, &dlm_join_handlers); | |
1431 | if (status) | |
1432 | goto bail; | |
1433 | ||
1434 | status = o2net_register_handler(DLM_CANCEL_JOIN_MSG, DLM_MOD_KEY, | |
1435 | sizeof(struct dlm_cancel_join), | |
1436 | dlm_cancel_join_handler, | |
1437 | NULL, &dlm_join_handlers); | |
1438 | ||
1439 | bail: | |
1440 | if (status < 0) | |
1441 | dlm_unregister_net_handlers(); | |
1442 | ||
1443 | return status; | |
1444 | } | |
1445 | ||
1446 | /* Domain eviction callback handling. | |
1447 | * | |
1448 | * The file system requires notification of node death *before* the | |
1449 | * dlm completes it's recovery work, otherwise it may be able to | |
1450 | * acquire locks on resources requiring recovery. Since the dlm can | |
1451 | * evict a node from it's domain *before* heartbeat fires, a similar | |
1452 | * mechanism is required. */ | |
1453 | ||
1454 | /* Eviction is not expected to happen often, so a per-domain lock is | |
1455 | * not necessary. Eviction callbacks are allowed to sleep for short | |
1456 | * periods of time. */ | |
1457 | static DECLARE_RWSEM(dlm_callback_sem); | |
1458 | ||
1459 | void dlm_fire_domain_eviction_callbacks(struct dlm_ctxt *dlm, | |
1460 | int node_num) | |
1461 | { | |
1462 | struct list_head *iter; | |
1463 | struct dlm_eviction_cb *cb; | |
1464 | ||
1465 | down_read(&dlm_callback_sem); | |
1466 | list_for_each(iter, &dlm->dlm_eviction_callbacks) { | |
1467 | cb = list_entry(iter, struct dlm_eviction_cb, ec_item); | |
1468 | ||
1469 | cb->ec_func(node_num, cb->ec_data); | |
1470 | } | |
1471 | up_read(&dlm_callback_sem); | |
1472 | } | |
1473 | ||
1474 | void dlm_setup_eviction_cb(struct dlm_eviction_cb *cb, | |
1475 | dlm_eviction_func *f, | |
1476 | void *data) | |
1477 | { | |
1478 | INIT_LIST_HEAD(&cb->ec_item); | |
1479 | cb->ec_func = f; | |
1480 | cb->ec_data = data; | |
1481 | } | |
1482 | EXPORT_SYMBOL_GPL(dlm_setup_eviction_cb); | |
1483 | ||
1484 | void dlm_register_eviction_cb(struct dlm_ctxt *dlm, | |
1485 | struct dlm_eviction_cb *cb) | |
1486 | { | |
1487 | down_write(&dlm_callback_sem); | |
1488 | list_add_tail(&cb->ec_item, &dlm->dlm_eviction_callbacks); | |
1489 | up_write(&dlm_callback_sem); | |
1490 | } | |
1491 | EXPORT_SYMBOL_GPL(dlm_register_eviction_cb); | |
1492 | ||
1493 | void dlm_unregister_eviction_cb(struct dlm_eviction_cb *cb) | |
1494 | { | |
1495 | down_write(&dlm_callback_sem); | |
1496 | list_del_init(&cb->ec_item); | |
1497 | up_write(&dlm_callback_sem); | |
1498 | } | |
1499 | EXPORT_SYMBOL_GPL(dlm_unregister_eviction_cb); | |
1500 | ||
1501 | static int __init dlm_init(void) | |
1502 | { | |
1503 | int status; | |
1504 | ||
1505 | dlm_print_version(); | |
1506 | ||
1507 | status = dlm_init_mle_cache(); | |
1508 | if (status) | |
1509 | return -1; | |
1510 | ||
1511 | status = dlm_register_net_handlers(); | |
1512 | if (status) { | |
1513 | dlm_destroy_mle_cache(); | |
1514 | return -1; | |
1515 | } | |
1516 | ||
1517 | return 0; | |
1518 | } | |
1519 | ||
1520 | static void __exit dlm_exit (void) | |
1521 | { | |
1522 | dlm_unregister_net_handlers(); | |
1523 | dlm_destroy_mle_cache(); | |
1524 | } | |
1525 | ||
1526 | MODULE_AUTHOR("Oracle"); | |
1527 | MODULE_LICENSE("GPL"); | |
1528 | ||
1529 | module_init(dlm_init); | |
1530 | module_exit(dlm_exit); |