xref: /openbmc/linux/fs/ocfs2/dlm/dlmthread.c (revision d7a3d85e)
1 /* -*- mode: c; c-basic-offset: 8; -*-
2  * vim: noexpandtab sw=8 ts=8 sts=0:
3  *
4  * dlmthread.c
5  *
6  * standalone DLM module
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 
28 #include <linux/module.h>
29 #include <linux/fs.h>
30 #include <linux/types.h>
31 #include <linux/highmem.h>
32 #include <linux/init.h>
33 #include <linux/sysctl.h>
34 #include <linux/random.h>
35 #include <linux/blkdev.h>
36 #include <linux/socket.h>
37 #include <linux/inet.h>
38 #include <linux/timer.h>
39 #include <linux/kthread.h>
40 #include <linux/delay.h>
41 
42 
43 #include "cluster/heartbeat.h"
44 #include "cluster/nodemanager.h"
45 #include "cluster/tcp.h"
46 
47 #include "dlmapi.h"
48 #include "dlmcommon.h"
49 #include "dlmdomain.h"
50 
51 #define MLOG_MASK_PREFIX (ML_DLM|ML_DLM_THREAD)
52 #include "cluster/masklog.h"
53 
54 static int dlm_thread(void *data);
55 static void dlm_flush_asts(struct dlm_ctxt *dlm);
56 
57 #define dlm_lock_is_remote(dlm, lock)     ((lock)->ml.node != (dlm)->node_num)
58 
59 /* will exit holding res->spinlock, but may drop in function */
60 /* waits until flags are cleared on res->state */
61 void __dlm_wait_on_lockres_flags(struct dlm_lock_resource *res, int flags)
62 {
63 	DECLARE_WAITQUEUE(wait, current);
64 
65 	assert_spin_locked(&res->spinlock);
66 
67 	add_wait_queue(&res->wq, &wait);
68 repeat:
69 	set_current_state(TASK_UNINTERRUPTIBLE);
70 	if (res->state & flags) {
71 		spin_unlock(&res->spinlock);
72 		schedule();
73 		spin_lock(&res->spinlock);
74 		goto repeat;
75 	}
76 	remove_wait_queue(&res->wq, &wait);
77 	__set_current_state(TASK_RUNNING);
78 }
79 
80 int __dlm_lockres_has_locks(struct dlm_lock_resource *res)
81 {
82 	if (list_empty(&res->granted) &&
83 	    list_empty(&res->converting) &&
84 	    list_empty(&res->blocked))
85 		return 0;
86 	return 1;
87 }
88 
89 /* "unused": the lockres has no locks, is not on the dirty list,
90  * has no inflight locks (in the gap between mastery and acquiring
91  * the first lock), and has no bits in its refmap.
92  * truly ready to be freed. */
93 int __dlm_lockres_unused(struct dlm_lock_resource *res)
94 {
95 	int bit;
96 
97 	assert_spin_locked(&res->spinlock);
98 
99 	if (__dlm_lockres_has_locks(res))
100 		return 0;
101 
102 	/* Locks are in the process of being created */
103 	if (res->inflight_locks)
104 		return 0;
105 
106 	if (!list_empty(&res->dirty) || res->state & DLM_LOCK_RES_DIRTY)
107 		return 0;
108 
109 	if (res->state & DLM_LOCK_RES_RECOVERING)
110 		return 0;
111 
112 	/* Another node has this resource with this node as the master */
113 	bit = find_next_bit(res->refmap, O2NM_MAX_NODES, 0);
114 	if (bit < O2NM_MAX_NODES)
115 		return 0;
116 
117 	return 1;
118 }
119 
120 
121 /* Call whenever you may have added or deleted something from one of
122  * the lockres queue's. This will figure out whether it belongs on the
123  * unused list or not and does the appropriate thing. */
124 void __dlm_lockres_calc_usage(struct dlm_ctxt *dlm,
125 			      struct dlm_lock_resource *res)
126 {
127 	assert_spin_locked(&dlm->spinlock);
128 	assert_spin_locked(&res->spinlock);
129 
130 	if (__dlm_lockres_unused(res)){
131 		if (list_empty(&res->purge)) {
132 			mlog(0, "%s: Adding res %.*s to purge list\n",
133 			     dlm->name, res->lockname.len, res->lockname.name);
134 
135 			res->last_used = jiffies;
136 			dlm_lockres_get(res);
137 			list_add_tail(&res->purge, &dlm->purge_list);
138 			dlm->purge_count++;
139 		}
140 	} else if (!list_empty(&res->purge)) {
141 		mlog(0, "%s: Removing res %.*s from purge list\n",
142 		     dlm->name, res->lockname.len, res->lockname.name);
143 
144 		list_del_init(&res->purge);
145 		dlm_lockres_put(res);
146 		dlm->purge_count--;
147 	}
148 }
149 
150 void dlm_lockres_calc_usage(struct dlm_ctxt *dlm,
151 			    struct dlm_lock_resource *res)
152 {
153 	spin_lock(&dlm->spinlock);
154 	spin_lock(&res->spinlock);
155 
156 	__dlm_lockres_calc_usage(dlm, res);
157 
158 	spin_unlock(&res->spinlock);
159 	spin_unlock(&dlm->spinlock);
160 }
161 
162 static void dlm_purge_lockres(struct dlm_ctxt *dlm,
163 			     struct dlm_lock_resource *res)
164 {
165 	int master;
166 	int ret = 0;
167 
168 	assert_spin_locked(&dlm->spinlock);
169 	assert_spin_locked(&res->spinlock);
170 
171 	master = (res->owner == dlm->node_num);
172 
173 	mlog(0, "%s: Purging res %.*s, master %d\n", dlm->name,
174 	     res->lockname.len, res->lockname.name, master);
175 
176 	if (!master) {
177 		res->state |= DLM_LOCK_RES_DROPPING_REF;
178 		/* drop spinlock...  retake below */
179 		spin_unlock(&res->spinlock);
180 		spin_unlock(&dlm->spinlock);
181 
182 		spin_lock(&res->spinlock);
183 		/* This ensures that clear refmap is sent after the set */
184 		__dlm_wait_on_lockres_flags(res, DLM_LOCK_RES_SETREF_INPROG);
185 		spin_unlock(&res->spinlock);
186 
187 		/* clear our bit from the master's refmap, ignore errors */
188 		ret = dlm_drop_lockres_ref(dlm, res);
189 		if (ret < 0) {
190 			if (!dlm_is_host_down(ret))
191 				BUG();
192 		}
193 		spin_lock(&dlm->spinlock);
194 		spin_lock(&res->spinlock);
195 	}
196 
197 	if (!list_empty(&res->purge)) {
198 		mlog(0, "%s: Removing res %.*s from purgelist, master %d\n",
199 		     dlm->name, res->lockname.len, res->lockname.name, master);
200 		list_del_init(&res->purge);
201 		dlm_lockres_put(res);
202 		dlm->purge_count--;
203 	}
204 
205 	if (!__dlm_lockres_unused(res)) {
206 		mlog(ML_ERROR, "%s: res %.*s in use after deref\n",
207 		     dlm->name, res->lockname.len, res->lockname.name);
208 		__dlm_print_one_lock_resource(res);
209 		BUG();
210 	}
211 
212 	__dlm_unhash_lockres(dlm, res);
213 
214 	/* lockres is not in the hash now.  drop the flag and wake up
215 	 * any processes waiting in dlm_get_lock_resource. */
216 	if (!master) {
217 		res->state &= ~DLM_LOCK_RES_DROPPING_REF;
218 		spin_unlock(&res->spinlock);
219 		wake_up(&res->wq);
220 	} else
221 		spin_unlock(&res->spinlock);
222 }
223 
224 static void dlm_run_purge_list(struct dlm_ctxt *dlm,
225 			       int purge_now)
226 {
227 	unsigned int run_max, unused;
228 	unsigned long purge_jiffies;
229 	struct dlm_lock_resource *lockres;
230 
231 	spin_lock(&dlm->spinlock);
232 	run_max = dlm->purge_count;
233 
234 	while(run_max && !list_empty(&dlm->purge_list)) {
235 		run_max--;
236 
237 		lockres = list_entry(dlm->purge_list.next,
238 				     struct dlm_lock_resource, purge);
239 
240 		spin_lock(&lockres->spinlock);
241 
242 		purge_jiffies = lockres->last_used +
243 			msecs_to_jiffies(DLM_PURGE_INTERVAL_MS);
244 
245 		/* Make sure that we want to be processing this guy at
246 		 * this time. */
247 		if (!purge_now && time_after(purge_jiffies, jiffies)) {
248 			/* Since resources are added to the purge list
249 			 * in tail order, we can stop at the first
250 			 * unpurgable resource -- anyone added after
251 			 * him will have a greater last_used value */
252 			spin_unlock(&lockres->spinlock);
253 			break;
254 		}
255 
256 		/* Status of the lockres *might* change so double
257 		 * check. If the lockres is unused, holding the dlm
258 		 * spinlock will prevent people from getting and more
259 		 * refs on it. */
260 		unused = __dlm_lockres_unused(lockres);
261 		if (!unused ||
262 		    (lockres->state & DLM_LOCK_RES_MIGRATING) ||
263 		    (lockres->inflight_assert_workers != 0)) {
264 			mlog(0, "%s: res %.*s is in use or being remastered, "
265 			     "used %d, state %d, assert master workers %u\n",
266 			     dlm->name, lockres->lockname.len,
267 			     lockres->lockname.name,
268 			     !unused, lockres->state,
269 			     lockres->inflight_assert_workers);
270 			list_move_tail(&lockres->purge, &dlm->purge_list);
271 			spin_unlock(&lockres->spinlock);
272 			continue;
273 		}
274 
275 		dlm_lockres_get(lockres);
276 
277 		dlm_purge_lockres(dlm, lockres);
278 
279 		dlm_lockres_put(lockres);
280 
281 		/* Avoid adding any scheduling latencies */
282 		cond_resched_lock(&dlm->spinlock);
283 	}
284 
285 	spin_unlock(&dlm->spinlock);
286 }
287 
288 static void dlm_shuffle_lists(struct dlm_ctxt *dlm,
289 			      struct dlm_lock_resource *res)
290 {
291 	struct dlm_lock *lock, *target;
292 	int can_grant = 1;
293 
294 	/*
295 	 * Because this function is called with the lockres
296 	 * spinlock, and because we know that it is not migrating/
297 	 * recovering/in-progress, it is fine to reserve asts and
298 	 * basts right before queueing them all throughout
299 	 */
300 	assert_spin_locked(&dlm->ast_lock);
301 	assert_spin_locked(&res->spinlock);
302 	BUG_ON((res->state & (DLM_LOCK_RES_MIGRATING|
303 			      DLM_LOCK_RES_RECOVERING|
304 			      DLM_LOCK_RES_IN_PROGRESS)));
305 
306 converting:
307 	if (list_empty(&res->converting))
308 		goto blocked;
309 	mlog(0, "%s: res %.*s has locks on the convert queue\n", dlm->name,
310 	     res->lockname.len, res->lockname.name);
311 
312 	target = list_entry(res->converting.next, struct dlm_lock, list);
313 	if (target->ml.convert_type == LKM_IVMODE) {
314 		mlog(ML_ERROR, "%s: res %.*s converting lock to invalid mode\n",
315 		     dlm->name, res->lockname.len, res->lockname.name);
316 		BUG();
317 	}
318 	list_for_each_entry(lock, &res->granted, list) {
319 		if (lock==target)
320 			continue;
321 		if (!dlm_lock_compatible(lock->ml.type,
322 					 target->ml.convert_type)) {
323 			can_grant = 0;
324 			/* queue the BAST if not already */
325 			if (lock->ml.highest_blocked == LKM_IVMODE) {
326 				__dlm_lockres_reserve_ast(res);
327 				__dlm_queue_bast(dlm, lock);
328 			}
329 			/* update the highest_blocked if needed */
330 			if (lock->ml.highest_blocked < target->ml.convert_type)
331 				lock->ml.highest_blocked =
332 					target->ml.convert_type;
333 		}
334 	}
335 
336 	list_for_each_entry(lock, &res->converting, list) {
337 		if (lock==target)
338 			continue;
339 		if (!dlm_lock_compatible(lock->ml.type,
340 					 target->ml.convert_type)) {
341 			can_grant = 0;
342 			if (lock->ml.highest_blocked == LKM_IVMODE) {
343 				__dlm_lockres_reserve_ast(res);
344 				__dlm_queue_bast(dlm, lock);
345 			}
346 			if (lock->ml.highest_blocked < target->ml.convert_type)
347 				lock->ml.highest_blocked =
348 					target->ml.convert_type;
349 		}
350 	}
351 
352 	/* we can convert the lock */
353 	if (can_grant) {
354 		spin_lock(&target->spinlock);
355 		BUG_ON(target->ml.highest_blocked != LKM_IVMODE);
356 
357 		mlog(0, "%s: res %.*s, AST for Converting lock %u:%llu, type "
358 		     "%d => %d, node %u\n", dlm->name, res->lockname.len,
359 		     res->lockname.name,
360 		     dlm_get_lock_cookie_node(be64_to_cpu(target->ml.cookie)),
361 		     dlm_get_lock_cookie_seq(be64_to_cpu(target->ml.cookie)),
362 		     target->ml.type,
363 		     target->ml.convert_type, target->ml.node);
364 
365 		target->ml.type = target->ml.convert_type;
366 		target->ml.convert_type = LKM_IVMODE;
367 		list_move_tail(&target->list, &res->granted);
368 
369 		BUG_ON(!target->lksb);
370 		target->lksb->status = DLM_NORMAL;
371 
372 		spin_unlock(&target->spinlock);
373 
374 		__dlm_lockres_reserve_ast(res);
375 		__dlm_queue_ast(dlm, target);
376 		/* go back and check for more */
377 		goto converting;
378 	}
379 
380 blocked:
381 	if (list_empty(&res->blocked))
382 		goto leave;
383 	target = list_entry(res->blocked.next, struct dlm_lock, list);
384 
385 	list_for_each_entry(lock, &res->granted, list) {
386 		if (lock==target)
387 			continue;
388 		if (!dlm_lock_compatible(lock->ml.type, target->ml.type)) {
389 			can_grant = 0;
390 			if (lock->ml.highest_blocked == LKM_IVMODE) {
391 				__dlm_lockres_reserve_ast(res);
392 				__dlm_queue_bast(dlm, lock);
393 			}
394 			if (lock->ml.highest_blocked < target->ml.type)
395 				lock->ml.highest_blocked = target->ml.type;
396 		}
397 	}
398 
399 	list_for_each_entry(lock, &res->converting, list) {
400 		if (lock==target)
401 			continue;
402 		if (!dlm_lock_compatible(lock->ml.type, target->ml.type)) {
403 			can_grant = 0;
404 			if (lock->ml.highest_blocked == LKM_IVMODE) {
405 				__dlm_lockres_reserve_ast(res);
406 				__dlm_queue_bast(dlm, lock);
407 			}
408 			if (lock->ml.highest_blocked < target->ml.type)
409 				lock->ml.highest_blocked = target->ml.type;
410 		}
411 	}
412 
413 	/* we can grant the blocked lock (only
414 	 * possible if converting list empty) */
415 	if (can_grant) {
416 		spin_lock(&target->spinlock);
417 		BUG_ON(target->ml.highest_blocked != LKM_IVMODE);
418 
419 		mlog(0, "%s: res %.*s, AST for Blocked lock %u:%llu, type %d, "
420 		     "node %u\n", dlm->name, res->lockname.len,
421 		     res->lockname.name,
422 		     dlm_get_lock_cookie_node(be64_to_cpu(target->ml.cookie)),
423 		     dlm_get_lock_cookie_seq(be64_to_cpu(target->ml.cookie)),
424 		     target->ml.type, target->ml.node);
425 
426 		/* target->ml.type is already correct */
427 		list_move_tail(&target->list, &res->granted);
428 
429 		BUG_ON(!target->lksb);
430 		target->lksb->status = DLM_NORMAL;
431 
432 		spin_unlock(&target->spinlock);
433 
434 		__dlm_lockres_reserve_ast(res);
435 		__dlm_queue_ast(dlm, target);
436 		/* go back and check for more */
437 		goto converting;
438 	}
439 
440 leave:
441 	return;
442 }
443 
444 /* must have NO locks when calling this with res !=NULL * */
445 void dlm_kick_thread(struct dlm_ctxt *dlm, struct dlm_lock_resource *res)
446 {
447 	if (res) {
448 		spin_lock(&dlm->spinlock);
449 		spin_lock(&res->spinlock);
450 		__dlm_dirty_lockres(dlm, res);
451 		spin_unlock(&res->spinlock);
452 		spin_unlock(&dlm->spinlock);
453 	}
454 	wake_up(&dlm->dlm_thread_wq);
455 }
456 
457 void __dlm_dirty_lockres(struct dlm_ctxt *dlm, struct dlm_lock_resource *res)
458 {
459 	assert_spin_locked(&dlm->spinlock);
460 	assert_spin_locked(&res->spinlock);
461 
462 	/* don't shuffle secondary queues */
463 	if ((res->owner == dlm->node_num)) {
464 		if (res->state & (DLM_LOCK_RES_MIGRATING |
465 				  DLM_LOCK_RES_BLOCK_DIRTY))
466 		    return;
467 
468 		if (list_empty(&res->dirty)) {
469 			/* ref for dirty_list */
470 			dlm_lockres_get(res);
471 			list_add_tail(&res->dirty, &dlm->dirty_list);
472 			res->state |= DLM_LOCK_RES_DIRTY;
473 		}
474 	}
475 
476 	mlog(0, "%s: res %.*s\n", dlm->name, res->lockname.len,
477 	     res->lockname.name);
478 }
479 
480 
481 /* Launch the NM thread for the mounted volume */
482 int dlm_launch_thread(struct dlm_ctxt *dlm)
483 {
484 	mlog(0, "Starting dlm_thread...\n");
485 
486 	dlm->dlm_thread_task = kthread_run(dlm_thread, dlm, "dlm_thread");
487 	if (IS_ERR(dlm->dlm_thread_task)) {
488 		mlog_errno(PTR_ERR(dlm->dlm_thread_task));
489 		dlm->dlm_thread_task = NULL;
490 		return -EINVAL;
491 	}
492 
493 	return 0;
494 }
495 
496 void dlm_complete_thread(struct dlm_ctxt *dlm)
497 {
498 	if (dlm->dlm_thread_task) {
499 		mlog(ML_KTHREAD, "Waiting for dlm thread to exit\n");
500 		kthread_stop(dlm->dlm_thread_task);
501 		dlm->dlm_thread_task = NULL;
502 	}
503 }
504 
505 static int dlm_dirty_list_empty(struct dlm_ctxt *dlm)
506 {
507 	int empty;
508 
509 	spin_lock(&dlm->spinlock);
510 	empty = list_empty(&dlm->dirty_list);
511 	spin_unlock(&dlm->spinlock);
512 
513 	return empty;
514 }
515 
516 static void dlm_flush_asts(struct dlm_ctxt *dlm)
517 {
518 	int ret;
519 	struct dlm_lock *lock;
520 	struct dlm_lock_resource *res;
521 	u8 hi;
522 
523 	spin_lock(&dlm->ast_lock);
524 	while (!list_empty(&dlm->pending_asts)) {
525 		lock = list_entry(dlm->pending_asts.next,
526 				  struct dlm_lock, ast_list);
527 		/* get an extra ref on lock */
528 		dlm_lock_get(lock);
529 		res = lock->lockres;
530 		mlog(0, "%s: res %.*s, Flush AST for lock %u:%llu, type %d, "
531 		     "node %u\n", dlm->name, res->lockname.len,
532 		     res->lockname.name,
533 		     dlm_get_lock_cookie_node(be64_to_cpu(lock->ml.cookie)),
534 		     dlm_get_lock_cookie_seq(be64_to_cpu(lock->ml.cookie)),
535 		     lock->ml.type, lock->ml.node);
536 
537 		BUG_ON(!lock->ast_pending);
538 
539 		/* remove from list (including ref) */
540 		list_del_init(&lock->ast_list);
541 		dlm_lock_put(lock);
542 		spin_unlock(&dlm->ast_lock);
543 
544 		if (lock->ml.node != dlm->node_num) {
545 			ret = dlm_do_remote_ast(dlm, res, lock);
546 			if (ret < 0)
547 				mlog_errno(ret);
548 		} else
549 			dlm_do_local_ast(dlm, res, lock);
550 
551 		spin_lock(&dlm->ast_lock);
552 
553 		/* possible that another ast was queued while
554 		 * we were delivering the last one */
555 		if (!list_empty(&lock->ast_list)) {
556 			mlog(0, "%s: res %.*s, AST queued while flushing last "
557 			     "one\n", dlm->name, res->lockname.len,
558 			     res->lockname.name);
559 		} else
560 			lock->ast_pending = 0;
561 
562 		/* drop the extra ref.
563 		 * this may drop it completely. */
564 		dlm_lock_put(lock);
565 		dlm_lockres_release_ast(dlm, res);
566 	}
567 
568 	while (!list_empty(&dlm->pending_basts)) {
569 		lock = list_entry(dlm->pending_basts.next,
570 				  struct dlm_lock, bast_list);
571 		/* get an extra ref on lock */
572 		dlm_lock_get(lock);
573 		res = lock->lockres;
574 
575 		BUG_ON(!lock->bast_pending);
576 
577 		/* get the highest blocked lock, and reset */
578 		spin_lock(&lock->spinlock);
579 		BUG_ON(lock->ml.highest_blocked <= LKM_IVMODE);
580 		hi = lock->ml.highest_blocked;
581 		lock->ml.highest_blocked = LKM_IVMODE;
582 		spin_unlock(&lock->spinlock);
583 
584 		/* remove from list (including ref) */
585 		list_del_init(&lock->bast_list);
586 		dlm_lock_put(lock);
587 		spin_unlock(&dlm->ast_lock);
588 
589 		mlog(0, "%s: res %.*s, Flush BAST for lock %u:%llu, "
590 		     "blocked %d, node %u\n",
591 		     dlm->name, res->lockname.len, res->lockname.name,
592 		     dlm_get_lock_cookie_node(be64_to_cpu(lock->ml.cookie)),
593 		     dlm_get_lock_cookie_seq(be64_to_cpu(lock->ml.cookie)),
594 		     hi, lock->ml.node);
595 
596 		if (lock->ml.node != dlm->node_num) {
597 			ret = dlm_send_proxy_bast(dlm, res, lock, hi);
598 			if (ret < 0)
599 				mlog_errno(ret);
600 		} else
601 			dlm_do_local_bast(dlm, res, lock, hi);
602 
603 		spin_lock(&dlm->ast_lock);
604 
605 		/* possible that another bast was queued while
606 		 * we were delivering the last one */
607 		if (!list_empty(&lock->bast_list)) {
608 			mlog(0, "%s: res %.*s, BAST queued while flushing last "
609 			     "one\n", dlm->name, res->lockname.len,
610 			     res->lockname.name);
611 		} else
612 			lock->bast_pending = 0;
613 
614 		/* drop the extra ref.
615 		 * this may drop it completely. */
616 		dlm_lock_put(lock);
617 		dlm_lockres_release_ast(dlm, res);
618 	}
619 	wake_up(&dlm->ast_wq);
620 	spin_unlock(&dlm->ast_lock);
621 }
622 
623 
624 #define DLM_THREAD_TIMEOUT_MS (4 * 1000)
625 #define DLM_THREAD_MAX_DIRTY  100
626 #define DLM_THREAD_MAX_ASTS   10
627 
628 static int dlm_thread(void *data)
629 {
630 	struct dlm_lock_resource *res;
631 	struct dlm_ctxt *dlm = data;
632 	unsigned long timeout = msecs_to_jiffies(DLM_THREAD_TIMEOUT_MS);
633 
634 	mlog(0, "dlm thread running for %s...\n", dlm->name);
635 
636 	while (!kthread_should_stop()) {
637 		int n = DLM_THREAD_MAX_DIRTY;
638 
639 		/* dlm_shutting_down is very point-in-time, but that
640 		 * doesn't matter as we'll just loop back around if we
641 		 * get false on the leading edge of a state
642 		 * transition. */
643 		dlm_run_purge_list(dlm, dlm_shutting_down(dlm));
644 
645 		/* We really don't want to hold dlm->spinlock while
646 		 * calling dlm_shuffle_lists on each lockres that
647 		 * needs to have its queues adjusted and AST/BASTs
648 		 * run.  So let's pull each entry off the dirty_list
649 		 * and drop dlm->spinlock ASAP.  Once off the list,
650 		 * res->spinlock needs to be taken again to protect
651 		 * the queues while calling dlm_shuffle_lists.  */
652 		spin_lock(&dlm->spinlock);
653 		while (!list_empty(&dlm->dirty_list)) {
654 			int delay = 0;
655 			res = list_entry(dlm->dirty_list.next,
656 					 struct dlm_lock_resource, dirty);
657 
658 			/* peel a lockres off, remove it from the list,
659 			 * unset the dirty flag and drop the dlm lock */
660 			BUG_ON(!res);
661 			dlm_lockres_get(res);
662 
663 			spin_lock(&res->spinlock);
664 			/* We clear the DLM_LOCK_RES_DIRTY state once we shuffle lists below */
665 			list_del_init(&res->dirty);
666 			spin_unlock(&res->spinlock);
667 			spin_unlock(&dlm->spinlock);
668 			/* Drop dirty_list ref */
669 			dlm_lockres_put(res);
670 
671 		 	/* lockres can be re-dirtied/re-added to the
672 			 * dirty_list in this gap, but that is ok */
673 
674 			spin_lock(&dlm->ast_lock);
675 			spin_lock(&res->spinlock);
676 			if (res->owner != dlm->node_num) {
677 				__dlm_print_one_lock_resource(res);
678 				mlog(ML_ERROR, "%s: inprog %d, mig %d, reco %d,"
679 				     " dirty %d\n", dlm->name,
680 				     !!(res->state & DLM_LOCK_RES_IN_PROGRESS),
681 				     !!(res->state & DLM_LOCK_RES_MIGRATING),
682 				     !!(res->state & DLM_LOCK_RES_RECOVERING),
683 				     !!(res->state & DLM_LOCK_RES_DIRTY));
684 			}
685 			BUG_ON(res->owner != dlm->node_num);
686 
687 			/* it is now ok to move lockreses in these states
688 			 * to the dirty list, assuming that they will only be
689 			 * dirty for a short while. */
690 			BUG_ON(res->state & DLM_LOCK_RES_MIGRATING);
691 			if (res->state & (DLM_LOCK_RES_IN_PROGRESS |
692 					  DLM_LOCK_RES_RECOVERING)) {
693 				/* move it to the tail and keep going */
694 				res->state &= ~DLM_LOCK_RES_DIRTY;
695 				spin_unlock(&res->spinlock);
696 				spin_unlock(&dlm->ast_lock);
697 				mlog(0, "%s: res %.*s, inprogress, delay list "
698 				     "shuffle, state %d\n", dlm->name,
699 				     res->lockname.len, res->lockname.name,
700 				     res->state);
701 				delay = 1;
702 				goto in_progress;
703 			}
704 
705 			/* at this point the lockres is not migrating/
706 			 * recovering/in-progress.  we have the lockres
707 			 * spinlock and do NOT have the dlm lock.
708 			 * safe to reserve/queue asts and run the lists. */
709 
710 			/* called while holding lockres lock */
711 			dlm_shuffle_lists(dlm, res);
712 			res->state &= ~DLM_LOCK_RES_DIRTY;
713 			spin_unlock(&res->spinlock);
714 			spin_unlock(&dlm->ast_lock);
715 
716 			dlm_lockres_calc_usage(dlm, res);
717 
718 in_progress:
719 
720 			spin_lock(&dlm->spinlock);
721 			/* if the lock was in-progress, stick
722 			 * it on the back of the list */
723 			if (delay) {
724 				spin_lock(&res->spinlock);
725 				__dlm_dirty_lockres(dlm, res);
726 				spin_unlock(&res->spinlock);
727 			}
728 			dlm_lockres_put(res);
729 
730 			/* unlikely, but we may need to give time to
731 			 * other tasks */
732 			if (!--n) {
733 				mlog(0, "%s: Throttling dlm thread\n",
734 				     dlm->name);
735 				break;
736 			}
737 		}
738 
739 		spin_unlock(&dlm->spinlock);
740 		dlm_flush_asts(dlm);
741 
742 		/* yield and continue right away if there is more work to do */
743 		if (!n) {
744 			cond_resched();
745 			continue;
746 		}
747 
748 		wait_event_interruptible_timeout(dlm->dlm_thread_wq,
749 						 !dlm_dirty_list_empty(dlm) ||
750 						 kthread_should_stop(),
751 						 timeout);
752 	}
753 
754 	mlog(0, "quitting DLM thread\n");
755 	return 0;
756 }
757