xref: /openbmc/linux/fs/ocfs2/cluster/heartbeat.c (revision 8dd3cdea)
1 // SPDX-License-Identifier: GPL-2.0-or-later
2 /*
3  * Copyright (C) 2004, 2005 Oracle.  All rights reserved.
4  */
5 
6 #include <linux/kernel.h>
7 #include <linux/sched.h>
8 #include <linux/jiffies.h>
9 #include <linux/module.h>
10 #include <linux/fs.h>
11 #include <linux/bio.h>
12 #include <linux/blkdev.h>
13 #include <linux/delay.h>
14 #include <linux/file.h>
15 #include <linux/kthread.h>
16 #include <linux/configfs.h>
17 #include <linux/random.h>
18 #include <linux/crc32.h>
19 #include <linux/time.h>
20 #include <linux/debugfs.h>
21 #include <linux/slab.h>
22 #include <linux/bitmap.h>
23 #include <linux/ktime.h>
24 #include "heartbeat.h"
25 #include "tcp.h"
26 #include "nodemanager.h"
27 #include "quorum.h"
28 
29 #include "masklog.h"
30 
31 
32 /*
33  * The first heartbeat pass had one global thread that would serialize all hb
34  * callback calls.  This global serializing sem should only be removed once
35  * we've made sure that all callees can deal with being called concurrently
36  * from multiple hb region threads.
37  */
38 static DECLARE_RWSEM(o2hb_callback_sem);
39 
40 /*
41  * multiple hb threads are watching multiple regions.  A node is live
42  * whenever any of the threads sees activity from the node in its region.
43  */
44 static DEFINE_SPINLOCK(o2hb_live_lock);
45 static struct list_head o2hb_live_slots[O2NM_MAX_NODES];
46 static unsigned long o2hb_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
47 static LIST_HEAD(o2hb_node_events);
48 static DECLARE_WAIT_QUEUE_HEAD(o2hb_steady_queue);
49 
50 /*
51  * In global heartbeat, we maintain a series of region bitmaps.
52  * 	- o2hb_region_bitmap allows us to limit the region number to max region.
53  * 	- o2hb_live_region_bitmap tracks live regions (seen steady iterations).
54  * 	- o2hb_quorum_region_bitmap tracks live regions that have seen all nodes
55  * 		heartbeat on it.
56  * 	- o2hb_failed_region_bitmap tracks the regions that have seen io timeouts.
57  */
58 static unsigned long o2hb_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
59 static unsigned long o2hb_live_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
60 static unsigned long o2hb_quorum_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
61 static unsigned long o2hb_failed_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
62 
63 #define O2HB_DB_TYPE_LIVENODES		0
64 #define O2HB_DB_TYPE_LIVEREGIONS	1
65 #define O2HB_DB_TYPE_QUORUMREGIONS	2
66 #define O2HB_DB_TYPE_FAILEDREGIONS	3
67 #define O2HB_DB_TYPE_REGION_LIVENODES	4
68 #define O2HB_DB_TYPE_REGION_NUMBER	5
69 #define O2HB_DB_TYPE_REGION_ELAPSED_TIME	6
70 #define O2HB_DB_TYPE_REGION_PINNED	7
71 struct o2hb_debug_buf {
72 	int db_type;
73 	int db_size;
74 	int db_len;
75 	void *db_data;
76 };
77 
78 static struct o2hb_debug_buf *o2hb_db_livenodes;
79 static struct o2hb_debug_buf *o2hb_db_liveregions;
80 static struct o2hb_debug_buf *o2hb_db_quorumregions;
81 static struct o2hb_debug_buf *o2hb_db_failedregions;
82 
83 #define O2HB_DEBUG_DIR			"o2hb"
84 #define O2HB_DEBUG_LIVENODES		"livenodes"
85 #define O2HB_DEBUG_LIVEREGIONS		"live_regions"
86 #define O2HB_DEBUG_QUORUMREGIONS	"quorum_regions"
87 #define O2HB_DEBUG_FAILEDREGIONS	"failed_regions"
88 #define O2HB_DEBUG_REGION_NUMBER	"num"
89 #define O2HB_DEBUG_REGION_ELAPSED_TIME	"elapsed_time_in_ms"
90 #define O2HB_DEBUG_REGION_PINNED	"pinned"
91 
92 static struct dentry *o2hb_debug_dir;
93 
94 static LIST_HEAD(o2hb_all_regions);
95 
96 static struct o2hb_callback {
97 	struct list_head list;
98 } o2hb_callbacks[O2HB_NUM_CB];
99 
100 static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type);
101 
102 enum o2hb_heartbeat_modes {
103 	O2HB_HEARTBEAT_LOCAL		= 0,
104 	O2HB_HEARTBEAT_GLOBAL,
105 	O2HB_HEARTBEAT_NUM_MODES,
106 };
107 
108 static const char *o2hb_heartbeat_mode_desc[O2HB_HEARTBEAT_NUM_MODES] = {
109 	"local",	/* O2HB_HEARTBEAT_LOCAL */
110 	"global",	/* O2HB_HEARTBEAT_GLOBAL */
111 };
112 
113 unsigned int o2hb_dead_threshold = O2HB_DEFAULT_DEAD_THRESHOLD;
114 static unsigned int o2hb_heartbeat_mode = O2HB_HEARTBEAT_LOCAL;
115 
116 /*
117  * o2hb_dependent_users tracks the number of registered callbacks that depend
118  * on heartbeat. o2net and o2dlm are two entities that register this callback.
119  * However only o2dlm depends on the heartbeat. It does not want the heartbeat
120  * to stop while a dlm domain is still active.
121  */
122 static unsigned int o2hb_dependent_users;
123 
124 /*
125  * In global heartbeat mode, all regions are pinned if there are one or more
126  * dependent users and the quorum region count is <= O2HB_PIN_CUT_OFF. All
127  * regions are unpinned if the region count exceeds the cut off or the number
128  * of dependent users falls to zero.
129  */
130 #define O2HB_PIN_CUT_OFF		3
131 
132 /*
133  * In local heartbeat mode, we assume the dlm domain name to be the same as
134  * region uuid. This is true for domains created for the file system but not
135  * necessarily true for userdlm domains. This is a known limitation.
136  *
137  * In global heartbeat mode, we pin/unpin all o2hb regions. This solution
138  * works for both file system and userdlm domains.
139  */
140 static int o2hb_region_pin(const char *region_uuid);
141 static void o2hb_region_unpin(const char *region_uuid);
142 
143 /* Only sets a new threshold if there are no active regions.
144  *
145  * No locking or otherwise interesting code is required for reading
146  * o2hb_dead_threshold as it can't change once regions are active and
147  * it's not interesting to anyone until then anyway. */
148 static void o2hb_dead_threshold_set(unsigned int threshold)
149 {
150 	if (threshold > O2HB_MIN_DEAD_THRESHOLD) {
151 		spin_lock(&o2hb_live_lock);
152 		if (list_empty(&o2hb_all_regions))
153 			o2hb_dead_threshold = threshold;
154 		spin_unlock(&o2hb_live_lock);
155 	}
156 }
157 
158 static int o2hb_global_heartbeat_mode_set(unsigned int hb_mode)
159 {
160 	int ret = -1;
161 
162 	if (hb_mode < O2HB_HEARTBEAT_NUM_MODES) {
163 		spin_lock(&o2hb_live_lock);
164 		if (list_empty(&o2hb_all_regions)) {
165 			o2hb_heartbeat_mode = hb_mode;
166 			ret = 0;
167 		}
168 		spin_unlock(&o2hb_live_lock);
169 	}
170 
171 	return ret;
172 }
173 
174 struct o2hb_node_event {
175 	struct list_head        hn_item;
176 	enum o2hb_callback_type hn_event_type;
177 	struct o2nm_node        *hn_node;
178 	int                     hn_node_num;
179 };
180 
181 struct o2hb_disk_slot {
182 	struct o2hb_disk_heartbeat_block *ds_raw_block;
183 	u8			ds_node_num;
184 	u64			ds_last_time;
185 	u64			ds_last_generation;
186 	u16			ds_equal_samples;
187 	u16			ds_changed_samples;
188 	struct list_head	ds_live_item;
189 };
190 
191 /* each thread owns a region.. when we're asked to tear down the region
192  * we ask the thread to stop, who cleans up the region */
193 struct o2hb_region {
194 	struct config_item	hr_item;
195 
196 	struct list_head	hr_all_item;
197 	unsigned		hr_unclean_stop:1,
198 				hr_aborted_start:1,
199 				hr_item_pinned:1,
200 				hr_item_dropped:1,
201 				hr_node_deleted:1;
202 
203 	/* protected by the hr_callback_sem */
204 	struct task_struct 	*hr_task;
205 
206 	unsigned int		hr_blocks;
207 	unsigned long long	hr_start_block;
208 
209 	unsigned int		hr_block_bits;
210 	unsigned int		hr_block_bytes;
211 
212 	unsigned int		hr_slots_per_page;
213 	unsigned int		hr_num_pages;
214 
215 	struct page             **hr_slot_data;
216 	struct block_device	*hr_bdev;
217 	struct o2hb_disk_slot	*hr_slots;
218 
219 	/* live node map of this region */
220 	unsigned long		hr_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
221 	unsigned int		hr_region_num;
222 
223 	struct dentry		*hr_debug_dir;
224 	struct o2hb_debug_buf	*hr_db_livenodes;
225 	struct o2hb_debug_buf	*hr_db_regnum;
226 	struct o2hb_debug_buf	*hr_db_elapsed_time;
227 	struct o2hb_debug_buf	*hr_db_pinned;
228 
229 	/* let the person setting up hb wait for it to return until it
230 	 * has reached a 'steady' state.  This will be fixed when we have
231 	 * a more complete api that doesn't lead to this sort of fragility. */
232 	atomic_t		hr_steady_iterations;
233 
234 	/* terminate o2hb thread if it does not reach steady state
235 	 * (hr_steady_iterations == 0) within hr_unsteady_iterations */
236 	atomic_t		hr_unsteady_iterations;
237 
238 	char			hr_dev_name[BDEVNAME_SIZE];
239 
240 	unsigned int		hr_timeout_ms;
241 
242 	/* randomized as the region goes up and down so that a node
243 	 * recognizes a node going up and down in one iteration */
244 	u64			hr_generation;
245 
246 	struct delayed_work	hr_write_timeout_work;
247 	unsigned long		hr_last_timeout_start;
248 
249 	/* negotiate timer, used to negotiate extending hb timeout. */
250 	struct delayed_work	hr_nego_timeout_work;
251 	unsigned long		hr_nego_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
252 
253 	/* Used during o2hb_check_slot to hold a copy of the block
254 	 * being checked because we temporarily have to zero out the
255 	 * crc field. */
256 	struct o2hb_disk_heartbeat_block *hr_tmp_block;
257 
258 	/* Message key for negotiate timeout message. */
259 	unsigned int		hr_key;
260 	struct list_head	hr_handler_list;
261 
262 	/* last hb status, 0 for success, other value for error. */
263 	int			hr_last_hb_status;
264 };
265 
266 struct o2hb_bio_wait_ctxt {
267 	atomic_t          wc_num_reqs;
268 	struct completion wc_io_complete;
269 	int               wc_error;
270 };
271 
272 #define O2HB_NEGO_TIMEOUT_MS (O2HB_MAX_WRITE_TIMEOUT_MS/2)
273 
274 enum {
275 	O2HB_NEGO_TIMEOUT_MSG = 1,
276 	O2HB_NEGO_APPROVE_MSG = 2,
277 };
278 
279 struct o2hb_nego_msg {
280 	u8 node_num;
281 };
282 
283 static void o2hb_write_timeout(struct work_struct *work)
284 {
285 	int failed, quorum;
286 	struct o2hb_region *reg =
287 		container_of(work, struct o2hb_region,
288 			     hr_write_timeout_work.work);
289 
290 	mlog(ML_ERROR, "Heartbeat write timeout to device %s after %u "
291 	     "milliseconds\n", reg->hr_dev_name,
292 	     jiffies_to_msecs(jiffies - reg->hr_last_timeout_start));
293 
294 	if (o2hb_global_heartbeat_active()) {
295 		spin_lock(&o2hb_live_lock);
296 		if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
297 			set_bit(reg->hr_region_num, o2hb_failed_region_bitmap);
298 		failed = bitmap_weight(o2hb_failed_region_bitmap,
299 					O2NM_MAX_REGIONS);
300 		quorum = bitmap_weight(o2hb_quorum_region_bitmap,
301 					O2NM_MAX_REGIONS);
302 		spin_unlock(&o2hb_live_lock);
303 
304 		mlog(ML_HEARTBEAT, "Number of regions %d, failed regions %d\n",
305 		     quorum, failed);
306 
307 		/*
308 		 * Fence if the number of failed regions >= half the number
309 		 * of  quorum regions
310 		 */
311 		if ((failed << 1) < quorum)
312 			return;
313 	}
314 
315 	o2quo_disk_timeout();
316 }
317 
318 static void o2hb_arm_timeout(struct o2hb_region *reg)
319 {
320 	/* Arm writeout only after thread reaches steady state */
321 	if (atomic_read(&reg->hr_steady_iterations) != 0)
322 		return;
323 
324 	mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n",
325 	     O2HB_MAX_WRITE_TIMEOUT_MS);
326 
327 	if (o2hb_global_heartbeat_active()) {
328 		spin_lock(&o2hb_live_lock);
329 		clear_bit(reg->hr_region_num, o2hb_failed_region_bitmap);
330 		spin_unlock(&o2hb_live_lock);
331 	}
332 	cancel_delayed_work(&reg->hr_write_timeout_work);
333 	schedule_delayed_work(&reg->hr_write_timeout_work,
334 			      msecs_to_jiffies(O2HB_MAX_WRITE_TIMEOUT_MS));
335 
336 	cancel_delayed_work(&reg->hr_nego_timeout_work);
337 	/* negotiate timeout must be less than write timeout. */
338 	schedule_delayed_work(&reg->hr_nego_timeout_work,
339 			      msecs_to_jiffies(O2HB_NEGO_TIMEOUT_MS));
340 	memset(reg->hr_nego_node_bitmap, 0, sizeof(reg->hr_nego_node_bitmap));
341 }
342 
343 static void o2hb_disarm_timeout(struct o2hb_region *reg)
344 {
345 	cancel_delayed_work_sync(&reg->hr_write_timeout_work);
346 	cancel_delayed_work_sync(&reg->hr_nego_timeout_work);
347 }
348 
349 static int o2hb_send_nego_msg(int key, int type, u8 target)
350 {
351 	struct o2hb_nego_msg msg;
352 	int status, ret;
353 
354 	msg.node_num = o2nm_this_node();
355 again:
356 	ret = o2net_send_message(type, key, &msg, sizeof(msg),
357 			target, &status);
358 
359 	if (ret == -EAGAIN || ret == -ENOMEM) {
360 		msleep(100);
361 		goto again;
362 	}
363 
364 	return ret;
365 }
366 
367 static void o2hb_nego_timeout(struct work_struct *work)
368 {
369 	unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
370 	int master_node, i, ret;
371 	struct o2hb_region *reg;
372 
373 	reg = container_of(work, struct o2hb_region, hr_nego_timeout_work.work);
374 	/* don't negotiate timeout if last hb failed since it is very
375 	 * possible io failed. Should let write timeout fence self.
376 	 */
377 	if (reg->hr_last_hb_status)
378 		return;
379 
380 	o2hb_fill_node_map(live_node_bitmap, sizeof(live_node_bitmap));
381 	/* lowest node as master node to make negotiate decision. */
382 	master_node = find_first_bit(live_node_bitmap, O2NM_MAX_NODES);
383 
384 	if (master_node == o2nm_this_node()) {
385 		if (!test_bit(master_node, reg->hr_nego_node_bitmap)) {
386 			printk(KERN_NOTICE "o2hb: node %d hb write hung for %ds on region %s (%s).\n",
387 				o2nm_this_node(), O2HB_NEGO_TIMEOUT_MS/1000,
388 				config_item_name(&reg->hr_item), reg->hr_dev_name);
389 			set_bit(master_node, reg->hr_nego_node_bitmap);
390 		}
391 		if (memcmp(reg->hr_nego_node_bitmap, live_node_bitmap,
392 				sizeof(reg->hr_nego_node_bitmap))) {
393 			/* check negotiate bitmap every second to do timeout
394 			 * approve decision.
395 			 */
396 			schedule_delayed_work(&reg->hr_nego_timeout_work,
397 				msecs_to_jiffies(1000));
398 
399 			return;
400 		}
401 
402 		printk(KERN_NOTICE "o2hb: all nodes hb write hung, maybe region %s (%s) is down.\n",
403 			config_item_name(&reg->hr_item), reg->hr_dev_name);
404 		/* approve negotiate timeout request. */
405 		o2hb_arm_timeout(reg);
406 
407 		i = -1;
408 		while ((i = find_next_bit(live_node_bitmap,
409 				O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
410 			if (i == master_node)
411 				continue;
412 
413 			mlog(ML_HEARTBEAT, "send NEGO_APPROVE msg to node %d\n", i);
414 			ret = o2hb_send_nego_msg(reg->hr_key,
415 					O2HB_NEGO_APPROVE_MSG, i);
416 			if (ret)
417 				mlog(ML_ERROR, "send NEGO_APPROVE msg to node %d fail %d\n",
418 					i, ret);
419 		}
420 	} else {
421 		/* negotiate timeout with master node. */
422 		printk(KERN_NOTICE "o2hb: node %d hb write hung for %ds on region %s (%s), negotiate timeout with node %d.\n",
423 			o2nm_this_node(), O2HB_NEGO_TIMEOUT_MS/1000, config_item_name(&reg->hr_item),
424 			reg->hr_dev_name, master_node);
425 		ret = o2hb_send_nego_msg(reg->hr_key, O2HB_NEGO_TIMEOUT_MSG,
426 				master_node);
427 		if (ret)
428 			mlog(ML_ERROR, "send NEGO_TIMEOUT msg to node %d fail %d\n",
429 				master_node, ret);
430 	}
431 }
432 
433 static int o2hb_nego_timeout_handler(struct o2net_msg *msg, u32 len, void *data,
434 				void **ret_data)
435 {
436 	struct o2hb_region *reg = data;
437 	struct o2hb_nego_msg *nego_msg;
438 
439 	nego_msg = (struct o2hb_nego_msg *)msg->buf;
440 	printk(KERN_NOTICE "o2hb: receive negotiate timeout message from node %d on region %s (%s).\n",
441 		nego_msg->node_num, config_item_name(&reg->hr_item), reg->hr_dev_name);
442 	if (nego_msg->node_num < O2NM_MAX_NODES)
443 		set_bit(nego_msg->node_num, reg->hr_nego_node_bitmap);
444 	else
445 		mlog(ML_ERROR, "got nego timeout message from bad node.\n");
446 
447 	return 0;
448 }
449 
450 static int o2hb_nego_approve_handler(struct o2net_msg *msg, u32 len, void *data,
451 				void **ret_data)
452 {
453 	struct o2hb_region *reg = data;
454 
455 	printk(KERN_NOTICE "o2hb: negotiate timeout approved by master node on region %s (%s).\n",
456 		config_item_name(&reg->hr_item), reg->hr_dev_name);
457 	o2hb_arm_timeout(reg);
458 	return 0;
459 }
460 
461 static inline void o2hb_bio_wait_init(struct o2hb_bio_wait_ctxt *wc)
462 {
463 	atomic_set(&wc->wc_num_reqs, 1);
464 	init_completion(&wc->wc_io_complete);
465 	wc->wc_error = 0;
466 }
467 
468 /* Used in error paths too */
469 static inline void o2hb_bio_wait_dec(struct o2hb_bio_wait_ctxt *wc,
470 				     unsigned int num)
471 {
472 	/* sadly atomic_sub_and_test() isn't available on all platforms.  The
473 	 * good news is that the fast path only completes one at a time */
474 	while(num--) {
475 		if (atomic_dec_and_test(&wc->wc_num_reqs)) {
476 			BUG_ON(num > 0);
477 			complete(&wc->wc_io_complete);
478 		}
479 	}
480 }
481 
482 static void o2hb_wait_on_io(struct o2hb_bio_wait_ctxt *wc)
483 {
484 	o2hb_bio_wait_dec(wc, 1);
485 	wait_for_completion(&wc->wc_io_complete);
486 }
487 
488 static void o2hb_bio_end_io(struct bio *bio)
489 {
490 	struct o2hb_bio_wait_ctxt *wc = bio->bi_private;
491 
492 	if (bio->bi_status) {
493 		mlog(ML_ERROR, "IO Error %d\n", bio->bi_status);
494 		wc->wc_error = blk_status_to_errno(bio->bi_status);
495 	}
496 
497 	o2hb_bio_wait_dec(wc, 1);
498 	bio_put(bio);
499 }
500 
501 /* Setup a Bio to cover I/O against num_slots slots starting at
502  * start_slot. */
503 static struct bio *o2hb_setup_one_bio(struct o2hb_region *reg,
504 				      struct o2hb_bio_wait_ctxt *wc,
505 				      unsigned int *current_slot,
506 				      unsigned int max_slots, int op,
507 				      int op_flags)
508 {
509 	int len, current_page;
510 	unsigned int vec_len, vec_start;
511 	unsigned int bits = reg->hr_block_bits;
512 	unsigned int spp = reg->hr_slots_per_page;
513 	unsigned int cs = *current_slot;
514 	struct bio *bio;
515 	struct page *page;
516 
517 	/* Testing has shown this allocation to take long enough under
518 	 * GFP_KERNEL that the local node can get fenced. It would be
519 	 * nicest if we could pre-allocate these bios and avoid this
520 	 * all together. */
521 	bio = bio_alloc(reg->hr_bdev, 16, op | op_flags, GFP_ATOMIC);
522 	if (!bio) {
523 		mlog(ML_ERROR, "Could not alloc slots BIO!\n");
524 		bio = ERR_PTR(-ENOMEM);
525 		goto bail;
526 	}
527 
528 	/* Must put everything in 512 byte sectors for the bio... */
529 	bio->bi_iter.bi_sector = (reg->hr_start_block + cs) << (bits - 9);
530 	bio->bi_private = wc;
531 	bio->bi_end_io = o2hb_bio_end_io;
532 
533 	vec_start = (cs << bits) % PAGE_SIZE;
534 	while(cs < max_slots) {
535 		current_page = cs / spp;
536 		page = reg->hr_slot_data[current_page];
537 
538 		vec_len = min(PAGE_SIZE - vec_start,
539 			      (max_slots-cs) * (PAGE_SIZE/spp) );
540 
541 		mlog(ML_HB_BIO, "page %d, vec_len = %u, vec_start = %u\n",
542 		     current_page, vec_len, vec_start);
543 
544 		len = bio_add_page(bio, page, vec_len, vec_start);
545 		if (len != vec_len) break;
546 
547 		cs += vec_len / (PAGE_SIZE/spp);
548 		vec_start = 0;
549 	}
550 
551 bail:
552 	*current_slot = cs;
553 	return bio;
554 }
555 
556 static int o2hb_read_slots(struct o2hb_region *reg,
557 			   unsigned int begin_slot,
558 			   unsigned int max_slots)
559 {
560 	unsigned int current_slot = begin_slot;
561 	int status;
562 	struct o2hb_bio_wait_ctxt wc;
563 	struct bio *bio;
564 
565 	o2hb_bio_wait_init(&wc);
566 
567 	while(current_slot < max_slots) {
568 		bio = o2hb_setup_one_bio(reg, &wc, &current_slot, max_slots,
569 					 REQ_OP_READ, 0);
570 		if (IS_ERR(bio)) {
571 			status = PTR_ERR(bio);
572 			mlog_errno(status);
573 			goto bail_and_wait;
574 		}
575 
576 		atomic_inc(&wc.wc_num_reqs);
577 		submit_bio(bio);
578 	}
579 
580 	status = 0;
581 
582 bail_and_wait:
583 	o2hb_wait_on_io(&wc);
584 	if (wc.wc_error && !status)
585 		status = wc.wc_error;
586 
587 	return status;
588 }
589 
590 static int o2hb_issue_node_write(struct o2hb_region *reg,
591 				 struct o2hb_bio_wait_ctxt *write_wc)
592 {
593 	int status;
594 	unsigned int slot;
595 	struct bio *bio;
596 
597 	o2hb_bio_wait_init(write_wc);
598 
599 	slot = o2nm_this_node();
600 
601 	bio = o2hb_setup_one_bio(reg, write_wc, &slot, slot+1, REQ_OP_WRITE,
602 				 REQ_SYNC);
603 	if (IS_ERR(bio)) {
604 		status = PTR_ERR(bio);
605 		mlog_errno(status);
606 		goto bail;
607 	}
608 
609 	atomic_inc(&write_wc->wc_num_reqs);
610 	submit_bio(bio);
611 
612 	status = 0;
613 bail:
614 	return status;
615 }
616 
617 static u32 o2hb_compute_block_crc_le(struct o2hb_region *reg,
618 				     struct o2hb_disk_heartbeat_block *hb_block)
619 {
620 	__le32 old_cksum;
621 	u32 ret;
622 
623 	/* We want to compute the block crc with a 0 value in the
624 	 * hb_cksum field. Save it off here and replace after the
625 	 * crc. */
626 	old_cksum = hb_block->hb_cksum;
627 	hb_block->hb_cksum = 0;
628 
629 	ret = crc32_le(0, (unsigned char *) hb_block, reg->hr_block_bytes);
630 
631 	hb_block->hb_cksum = old_cksum;
632 
633 	return ret;
634 }
635 
636 static void o2hb_dump_slot(struct o2hb_disk_heartbeat_block *hb_block)
637 {
638 	mlog(ML_ERROR, "Dump slot information: seq = 0x%llx, node = %u, "
639 	     "cksum = 0x%x, generation 0x%llx\n",
640 	     (long long)le64_to_cpu(hb_block->hb_seq),
641 	     hb_block->hb_node, le32_to_cpu(hb_block->hb_cksum),
642 	     (long long)le64_to_cpu(hb_block->hb_generation));
643 }
644 
645 static int o2hb_verify_crc(struct o2hb_region *reg,
646 			   struct o2hb_disk_heartbeat_block *hb_block)
647 {
648 	u32 read, computed;
649 
650 	read = le32_to_cpu(hb_block->hb_cksum);
651 	computed = o2hb_compute_block_crc_le(reg, hb_block);
652 
653 	return read == computed;
654 }
655 
656 /*
657  * Compare the slot data with what we wrote in the last iteration.
658  * If the match fails, print an appropriate error message. This is to
659  * detect errors like... another node hearting on the same slot,
660  * flaky device that is losing writes, etc.
661  * Returns 1 if check succeeds, 0 otherwise.
662  */
663 static int o2hb_check_own_slot(struct o2hb_region *reg)
664 {
665 	struct o2hb_disk_slot *slot;
666 	struct o2hb_disk_heartbeat_block *hb_block;
667 	char *errstr;
668 
669 	slot = &reg->hr_slots[o2nm_this_node()];
670 	/* Don't check on our 1st timestamp */
671 	if (!slot->ds_last_time)
672 		return 0;
673 
674 	hb_block = slot->ds_raw_block;
675 	if (le64_to_cpu(hb_block->hb_seq) == slot->ds_last_time &&
676 	    le64_to_cpu(hb_block->hb_generation) == slot->ds_last_generation &&
677 	    hb_block->hb_node == slot->ds_node_num)
678 		return 1;
679 
680 #define ERRSTR1		"Another node is heartbeating on device"
681 #define ERRSTR2		"Heartbeat generation mismatch on device"
682 #define ERRSTR3		"Heartbeat sequence mismatch on device"
683 
684 	if (hb_block->hb_node != slot->ds_node_num)
685 		errstr = ERRSTR1;
686 	else if (le64_to_cpu(hb_block->hb_generation) !=
687 		 slot->ds_last_generation)
688 		errstr = ERRSTR2;
689 	else
690 		errstr = ERRSTR3;
691 
692 	mlog(ML_ERROR, "%s (%s): expected(%u:0x%llx, 0x%llx), "
693 	     "ondisk(%u:0x%llx, 0x%llx)\n", errstr, reg->hr_dev_name,
694 	     slot->ds_node_num, (unsigned long long)slot->ds_last_generation,
695 	     (unsigned long long)slot->ds_last_time, hb_block->hb_node,
696 	     (unsigned long long)le64_to_cpu(hb_block->hb_generation),
697 	     (unsigned long long)le64_to_cpu(hb_block->hb_seq));
698 
699 	return 0;
700 }
701 
702 static inline void o2hb_prepare_block(struct o2hb_region *reg,
703 				      u64 generation)
704 {
705 	int node_num;
706 	u64 cputime;
707 	struct o2hb_disk_slot *slot;
708 	struct o2hb_disk_heartbeat_block *hb_block;
709 
710 	node_num = o2nm_this_node();
711 	slot = &reg->hr_slots[node_num];
712 
713 	hb_block = (struct o2hb_disk_heartbeat_block *)slot->ds_raw_block;
714 	memset(hb_block, 0, reg->hr_block_bytes);
715 	/* TODO: time stuff */
716 	cputime = ktime_get_real_seconds();
717 	if (!cputime)
718 		cputime = 1;
719 
720 	hb_block->hb_seq = cpu_to_le64(cputime);
721 	hb_block->hb_node = node_num;
722 	hb_block->hb_generation = cpu_to_le64(generation);
723 	hb_block->hb_dead_ms = cpu_to_le32(o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS);
724 
725 	/* This step must always happen last! */
726 	hb_block->hb_cksum = cpu_to_le32(o2hb_compute_block_crc_le(reg,
727 								   hb_block));
728 
729 	mlog(ML_HB_BIO, "our node generation = 0x%llx, cksum = 0x%x\n",
730 	     (long long)generation,
731 	     le32_to_cpu(hb_block->hb_cksum));
732 }
733 
734 static void o2hb_fire_callbacks(struct o2hb_callback *hbcall,
735 				struct o2nm_node *node,
736 				int idx)
737 {
738 	struct o2hb_callback_func *f;
739 
740 	list_for_each_entry(f, &hbcall->list, hc_item) {
741 		mlog(ML_HEARTBEAT, "calling funcs %p\n", f);
742 		(f->hc_func)(node, idx, f->hc_data);
743 	}
744 }
745 
746 /* Will run the list in order until we process the passed event */
747 static void o2hb_run_event_list(struct o2hb_node_event *queued_event)
748 {
749 	struct o2hb_callback *hbcall;
750 	struct o2hb_node_event *event;
751 
752 	/* Holding callback sem assures we don't alter the callback
753 	 * lists when doing this, and serializes ourselves with other
754 	 * processes wanting callbacks. */
755 	down_write(&o2hb_callback_sem);
756 
757 	spin_lock(&o2hb_live_lock);
758 	while (!list_empty(&o2hb_node_events)
759 	       && !list_empty(&queued_event->hn_item)) {
760 		event = list_entry(o2hb_node_events.next,
761 				   struct o2hb_node_event,
762 				   hn_item);
763 		list_del_init(&event->hn_item);
764 		spin_unlock(&o2hb_live_lock);
765 
766 		mlog(ML_HEARTBEAT, "Node %s event for %d\n",
767 		     event->hn_event_type == O2HB_NODE_UP_CB ? "UP" : "DOWN",
768 		     event->hn_node_num);
769 
770 		hbcall = hbcall_from_type(event->hn_event_type);
771 
772 		/* We should *never* have gotten on to the list with a
773 		 * bad type... This isn't something that we should try
774 		 * to recover from. */
775 		BUG_ON(IS_ERR(hbcall));
776 
777 		o2hb_fire_callbacks(hbcall, event->hn_node, event->hn_node_num);
778 
779 		spin_lock(&o2hb_live_lock);
780 	}
781 	spin_unlock(&o2hb_live_lock);
782 
783 	up_write(&o2hb_callback_sem);
784 }
785 
786 static void o2hb_queue_node_event(struct o2hb_node_event *event,
787 				  enum o2hb_callback_type type,
788 				  struct o2nm_node *node,
789 				  int node_num)
790 {
791 	assert_spin_locked(&o2hb_live_lock);
792 
793 	BUG_ON((!node) && (type != O2HB_NODE_DOWN_CB));
794 
795 	event->hn_event_type = type;
796 	event->hn_node = node;
797 	event->hn_node_num = node_num;
798 
799 	mlog(ML_HEARTBEAT, "Queue node %s event for node %d\n",
800 	     type == O2HB_NODE_UP_CB ? "UP" : "DOWN", node_num);
801 
802 	list_add_tail(&event->hn_item, &o2hb_node_events);
803 }
804 
805 static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot)
806 {
807 	struct o2hb_node_event event =
808 		{ .hn_item = LIST_HEAD_INIT(event.hn_item), };
809 	struct o2nm_node *node;
810 	int queued = 0;
811 
812 	node = o2nm_get_node_by_num(slot->ds_node_num);
813 	if (!node)
814 		return;
815 
816 	spin_lock(&o2hb_live_lock);
817 	if (!list_empty(&slot->ds_live_item)) {
818 		mlog(ML_HEARTBEAT, "Shutdown, node %d leaves region\n",
819 		     slot->ds_node_num);
820 
821 		list_del_init(&slot->ds_live_item);
822 
823 		if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
824 			clear_bit(slot->ds_node_num, o2hb_live_node_bitmap);
825 
826 			o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB, node,
827 					      slot->ds_node_num);
828 			queued = 1;
829 		}
830 	}
831 	spin_unlock(&o2hb_live_lock);
832 
833 	if (queued)
834 		o2hb_run_event_list(&event);
835 
836 	o2nm_node_put(node);
837 }
838 
839 static void o2hb_set_quorum_device(struct o2hb_region *reg)
840 {
841 	if (!o2hb_global_heartbeat_active())
842 		return;
843 
844 	/* Prevent race with o2hb_heartbeat_group_drop_item() */
845 	if (kthread_should_stop())
846 		return;
847 
848 	/* Tag region as quorum only after thread reaches steady state */
849 	if (atomic_read(&reg->hr_steady_iterations) != 0)
850 		return;
851 
852 	spin_lock(&o2hb_live_lock);
853 
854 	if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
855 		goto unlock;
856 
857 	/*
858 	 * A region can be added to the quorum only when it sees all
859 	 * live nodes heartbeat on it. In other words, the region has been
860 	 * added to all nodes.
861 	 */
862 	if (memcmp(reg->hr_live_node_bitmap, o2hb_live_node_bitmap,
863 		   sizeof(o2hb_live_node_bitmap)))
864 		goto unlock;
865 
866 	printk(KERN_NOTICE "o2hb: Region %s (%s) is now a quorum device\n",
867 	       config_item_name(&reg->hr_item), reg->hr_dev_name);
868 
869 	set_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
870 
871 	/*
872 	 * If global heartbeat active, unpin all regions if the
873 	 * region count > CUT_OFF
874 	 */
875 	if (bitmap_weight(o2hb_quorum_region_bitmap,
876 			   O2NM_MAX_REGIONS) > O2HB_PIN_CUT_OFF)
877 		o2hb_region_unpin(NULL);
878 unlock:
879 	spin_unlock(&o2hb_live_lock);
880 }
881 
882 static int o2hb_check_slot(struct o2hb_region *reg,
883 			   struct o2hb_disk_slot *slot)
884 {
885 	int changed = 0, gen_changed = 0;
886 	struct o2hb_node_event event =
887 		{ .hn_item = LIST_HEAD_INIT(event.hn_item), };
888 	struct o2nm_node *node;
889 	struct o2hb_disk_heartbeat_block *hb_block = reg->hr_tmp_block;
890 	u64 cputime;
891 	unsigned int dead_ms = o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS;
892 	unsigned int slot_dead_ms;
893 	int tmp;
894 	int queued = 0;
895 
896 	memcpy(hb_block, slot->ds_raw_block, reg->hr_block_bytes);
897 
898 	/*
899 	 * If a node is no longer configured but is still in the livemap, we
900 	 * may need to clear that bit from the livemap.
901 	 */
902 	node = o2nm_get_node_by_num(slot->ds_node_num);
903 	if (!node) {
904 		spin_lock(&o2hb_live_lock);
905 		tmp = test_bit(slot->ds_node_num, o2hb_live_node_bitmap);
906 		spin_unlock(&o2hb_live_lock);
907 		if (!tmp)
908 			return 0;
909 	}
910 
911 	if (!o2hb_verify_crc(reg, hb_block)) {
912 		/* all paths from here will drop o2hb_live_lock for
913 		 * us. */
914 		spin_lock(&o2hb_live_lock);
915 
916 		/* Don't print an error on the console in this case -
917 		 * a freshly formatted heartbeat area will not have a
918 		 * crc set on it. */
919 		if (list_empty(&slot->ds_live_item))
920 			goto out;
921 
922 		/* The node is live but pushed out a bad crc. We
923 		 * consider it a transient miss but don't populate any
924 		 * other values as they may be junk. */
925 		mlog(ML_ERROR, "Node %d has written a bad crc to %s\n",
926 		     slot->ds_node_num, reg->hr_dev_name);
927 		o2hb_dump_slot(hb_block);
928 
929 		slot->ds_equal_samples++;
930 		goto fire_callbacks;
931 	}
932 
933 	/* we don't care if these wrap.. the state transitions below
934 	 * clear at the right places */
935 	cputime = le64_to_cpu(hb_block->hb_seq);
936 	if (slot->ds_last_time != cputime)
937 		slot->ds_changed_samples++;
938 	else
939 		slot->ds_equal_samples++;
940 	slot->ds_last_time = cputime;
941 
942 	/* The node changed heartbeat generations. We assume this to
943 	 * mean it dropped off but came back before we timed out. We
944 	 * want to consider it down for the time being but don't want
945 	 * to lose any changed_samples state we might build up to
946 	 * considering it live again. */
947 	if (slot->ds_last_generation != le64_to_cpu(hb_block->hb_generation)) {
948 		gen_changed = 1;
949 		slot->ds_equal_samples = 0;
950 		mlog(ML_HEARTBEAT, "Node %d changed generation (0x%llx "
951 		     "to 0x%llx)\n", slot->ds_node_num,
952 		     (long long)slot->ds_last_generation,
953 		     (long long)le64_to_cpu(hb_block->hb_generation));
954 	}
955 
956 	slot->ds_last_generation = le64_to_cpu(hb_block->hb_generation);
957 
958 	mlog(ML_HEARTBEAT, "Slot %d gen 0x%llx cksum 0x%x "
959 	     "seq %llu last %llu changed %u equal %u\n",
960 	     slot->ds_node_num, (long long)slot->ds_last_generation,
961 	     le32_to_cpu(hb_block->hb_cksum),
962 	     (unsigned long long)le64_to_cpu(hb_block->hb_seq),
963 	     (unsigned long long)slot->ds_last_time, slot->ds_changed_samples,
964 	     slot->ds_equal_samples);
965 
966 	spin_lock(&o2hb_live_lock);
967 
968 fire_callbacks:
969 	/* dead nodes only come to life after some number of
970 	 * changes at any time during their dead time */
971 	if (list_empty(&slot->ds_live_item) &&
972 	    slot->ds_changed_samples >= O2HB_LIVE_THRESHOLD) {
973 		mlog(ML_HEARTBEAT, "Node %d (id 0x%llx) joined my region\n",
974 		     slot->ds_node_num, (long long)slot->ds_last_generation);
975 
976 		set_bit(slot->ds_node_num, reg->hr_live_node_bitmap);
977 
978 		/* first on the list generates a callback */
979 		if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
980 			mlog(ML_HEARTBEAT, "o2hb: Add node %d to live nodes "
981 			     "bitmap\n", slot->ds_node_num);
982 			set_bit(slot->ds_node_num, o2hb_live_node_bitmap);
983 
984 			o2hb_queue_node_event(&event, O2HB_NODE_UP_CB, node,
985 					      slot->ds_node_num);
986 
987 			changed = 1;
988 			queued = 1;
989 		}
990 
991 		list_add_tail(&slot->ds_live_item,
992 			      &o2hb_live_slots[slot->ds_node_num]);
993 
994 		slot->ds_equal_samples = 0;
995 
996 		/* We want to be sure that all nodes agree on the
997 		 * number of milliseconds before a node will be
998 		 * considered dead. The self-fencing timeout is
999 		 * computed from this value, and a discrepancy might
1000 		 * result in heartbeat calling a node dead when it
1001 		 * hasn't self-fenced yet. */
1002 		slot_dead_ms = le32_to_cpu(hb_block->hb_dead_ms);
1003 		if (slot_dead_ms && slot_dead_ms != dead_ms) {
1004 			/* TODO: Perhaps we can fail the region here. */
1005 			mlog(ML_ERROR, "Node %d on device %s has a dead count "
1006 			     "of %u ms, but our count is %u ms.\n"
1007 			     "Please double check your configuration values "
1008 			     "for 'O2CB_HEARTBEAT_THRESHOLD'\n",
1009 			     slot->ds_node_num, reg->hr_dev_name, slot_dead_ms,
1010 			     dead_ms);
1011 		}
1012 		goto out;
1013 	}
1014 
1015 	/* if the list is dead, we're done.. */
1016 	if (list_empty(&slot->ds_live_item))
1017 		goto out;
1018 
1019 	/* live nodes only go dead after enough consequtive missed
1020 	 * samples..  reset the missed counter whenever we see
1021 	 * activity */
1022 	if (slot->ds_equal_samples >= o2hb_dead_threshold || gen_changed) {
1023 		mlog(ML_HEARTBEAT, "Node %d left my region\n",
1024 		     slot->ds_node_num);
1025 
1026 		clear_bit(slot->ds_node_num, reg->hr_live_node_bitmap);
1027 
1028 		/* last off the live_slot generates a callback */
1029 		list_del_init(&slot->ds_live_item);
1030 		if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
1031 			mlog(ML_HEARTBEAT, "o2hb: Remove node %d from live "
1032 			     "nodes bitmap\n", slot->ds_node_num);
1033 			clear_bit(slot->ds_node_num, o2hb_live_node_bitmap);
1034 
1035 			/* node can be null */
1036 			o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB,
1037 					      node, slot->ds_node_num);
1038 
1039 			changed = 1;
1040 			queued = 1;
1041 		}
1042 
1043 		/* We don't clear this because the node is still
1044 		 * actually writing new blocks. */
1045 		if (!gen_changed)
1046 			slot->ds_changed_samples = 0;
1047 		goto out;
1048 	}
1049 	if (slot->ds_changed_samples) {
1050 		slot->ds_changed_samples = 0;
1051 		slot->ds_equal_samples = 0;
1052 	}
1053 out:
1054 	spin_unlock(&o2hb_live_lock);
1055 
1056 	if (queued)
1057 		o2hb_run_event_list(&event);
1058 
1059 	if (node)
1060 		o2nm_node_put(node);
1061 	return changed;
1062 }
1063 
1064 static int o2hb_highest_node(unsigned long *nodes, int numbits)
1065 {
1066 	return find_last_bit(nodes, numbits);
1067 }
1068 
1069 static int o2hb_lowest_node(unsigned long *nodes, int numbits)
1070 {
1071 	return find_first_bit(nodes, numbits);
1072 }
1073 
1074 static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
1075 {
1076 	int i, ret, highest_node, lowest_node;
1077 	int membership_change = 0, own_slot_ok = 0;
1078 	unsigned long configured_nodes[BITS_TO_LONGS(O2NM_MAX_NODES)];
1079 	unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
1080 	struct o2hb_bio_wait_ctxt write_wc;
1081 
1082 	ret = o2nm_configured_node_map(configured_nodes,
1083 				       sizeof(configured_nodes));
1084 	if (ret) {
1085 		mlog_errno(ret);
1086 		goto bail;
1087 	}
1088 
1089 	/*
1090 	 * If a node is not configured but is in the livemap, we still need
1091 	 * to read the slot so as to be able to remove it from the livemap.
1092 	 */
1093 	o2hb_fill_node_map(live_node_bitmap, sizeof(live_node_bitmap));
1094 	i = -1;
1095 	while ((i = find_next_bit(live_node_bitmap,
1096 				  O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
1097 		set_bit(i, configured_nodes);
1098 	}
1099 
1100 	highest_node = o2hb_highest_node(configured_nodes, O2NM_MAX_NODES);
1101 	lowest_node = o2hb_lowest_node(configured_nodes, O2NM_MAX_NODES);
1102 	if (highest_node >= O2NM_MAX_NODES || lowest_node >= O2NM_MAX_NODES) {
1103 		mlog(ML_NOTICE, "o2hb: No configured nodes found!\n");
1104 		ret = -EINVAL;
1105 		goto bail;
1106 	}
1107 
1108 	/* No sense in reading the slots of nodes that don't exist
1109 	 * yet. Of course, if the node definitions have holes in them
1110 	 * then we're reading an empty slot anyway... Consider this
1111 	 * best-effort. */
1112 	ret = o2hb_read_slots(reg, lowest_node, highest_node + 1);
1113 	if (ret < 0) {
1114 		mlog_errno(ret);
1115 		goto bail;
1116 	}
1117 
1118 	/* With an up to date view of the slots, we can check that no
1119 	 * other node has been improperly configured to heartbeat in
1120 	 * our slot. */
1121 	own_slot_ok = o2hb_check_own_slot(reg);
1122 
1123 	/* fill in the proper info for our next heartbeat */
1124 	o2hb_prepare_block(reg, reg->hr_generation);
1125 
1126 	ret = o2hb_issue_node_write(reg, &write_wc);
1127 	if (ret < 0) {
1128 		mlog_errno(ret);
1129 		goto bail;
1130 	}
1131 
1132 	i = -1;
1133 	while((i = find_next_bit(configured_nodes,
1134 				 O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
1135 		membership_change |= o2hb_check_slot(reg, &reg->hr_slots[i]);
1136 	}
1137 
1138 	/*
1139 	 * We have to be sure we've advertised ourselves on disk
1140 	 * before we can go to steady state.  This ensures that
1141 	 * people we find in our steady state have seen us.
1142 	 */
1143 	o2hb_wait_on_io(&write_wc);
1144 	if (write_wc.wc_error) {
1145 		/* Do not re-arm the write timeout on I/O error - we
1146 		 * can't be sure that the new block ever made it to
1147 		 * disk */
1148 		mlog(ML_ERROR, "Write error %d on device \"%s\"\n",
1149 		     write_wc.wc_error, reg->hr_dev_name);
1150 		ret = write_wc.wc_error;
1151 		goto bail;
1152 	}
1153 
1154 	/* Skip disarming the timeout if own slot has stale/bad data */
1155 	if (own_slot_ok) {
1156 		o2hb_set_quorum_device(reg);
1157 		o2hb_arm_timeout(reg);
1158 		reg->hr_last_timeout_start = jiffies;
1159 	}
1160 
1161 bail:
1162 	/* let the person who launched us know when things are steady */
1163 	if (atomic_read(&reg->hr_steady_iterations) != 0) {
1164 		if (!ret && own_slot_ok && !membership_change) {
1165 			if (atomic_dec_and_test(&reg->hr_steady_iterations))
1166 				wake_up(&o2hb_steady_queue);
1167 		}
1168 	}
1169 
1170 	if (atomic_read(&reg->hr_steady_iterations) != 0) {
1171 		if (atomic_dec_and_test(&reg->hr_unsteady_iterations)) {
1172 			printk(KERN_NOTICE "o2hb: Unable to stabilize "
1173 			       "heartbeat on region %s (%s)\n",
1174 			       config_item_name(&reg->hr_item),
1175 			       reg->hr_dev_name);
1176 			atomic_set(&reg->hr_steady_iterations, 0);
1177 			reg->hr_aborted_start = 1;
1178 			wake_up(&o2hb_steady_queue);
1179 			ret = -EIO;
1180 		}
1181 	}
1182 
1183 	return ret;
1184 }
1185 
1186 /*
1187  * we ride the region ref that the region dir holds.  before the region
1188  * dir is removed and drops it ref it will wait to tear down this
1189  * thread.
1190  */
1191 static int o2hb_thread(void *data)
1192 {
1193 	int i, ret;
1194 	struct o2hb_region *reg = data;
1195 	struct o2hb_bio_wait_ctxt write_wc;
1196 	ktime_t before_hb, after_hb;
1197 	unsigned int elapsed_msec;
1198 
1199 	mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread running\n");
1200 
1201 	set_user_nice(current, MIN_NICE);
1202 
1203 	/* Pin node */
1204 	ret = o2nm_depend_this_node();
1205 	if (ret) {
1206 		mlog(ML_ERROR, "Node has been deleted, ret = %d\n", ret);
1207 		reg->hr_node_deleted = 1;
1208 		wake_up(&o2hb_steady_queue);
1209 		return 0;
1210 	}
1211 
1212 	while (!kthread_should_stop() &&
1213 	       !reg->hr_unclean_stop && !reg->hr_aborted_start) {
1214 		/* We track the time spent inside
1215 		 * o2hb_do_disk_heartbeat so that we avoid more than
1216 		 * hr_timeout_ms between disk writes. On busy systems
1217 		 * this should result in a heartbeat which is less
1218 		 * likely to time itself out. */
1219 		before_hb = ktime_get_real();
1220 
1221 		ret = o2hb_do_disk_heartbeat(reg);
1222 		reg->hr_last_hb_status = ret;
1223 
1224 		after_hb = ktime_get_real();
1225 
1226 		elapsed_msec = (unsigned int)
1227 				ktime_ms_delta(after_hb, before_hb);
1228 
1229 		mlog(ML_HEARTBEAT,
1230 		     "start = %lld, end = %lld, msec = %u, ret = %d\n",
1231 		     before_hb, after_hb, elapsed_msec, ret);
1232 
1233 		if (!kthread_should_stop() &&
1234 		    elapsed_msec < reg->hr_timeout_ms) {
1235 			/* the kthread api has blocked signals for us so no
1236 			 * need to record the return value. */
1237 			msleep_interruptible(reg->hr_timeout_ms - elapsed_msec);
1238 		}
1239 	}
1240 
1241 	o2hb_disarm_timeout(reg);
1242 
1243 	/* unclean stop is only used in very bad situation */
1244 	for(i = 0; !reg->hr_unclean_stop && i < reg->hr_blocks; i++)
1245 		o2hb_shutdown_slot(&reg->hr_slots[i]);
1246 
1247 	/* Explicit down notification - avoid forcing the other nodes
1248 	 * to timeout on this region when we could just as easily
1249 	 * write a clear generation - thus indicating to them that
1250 	 * this node has left this region.
1251 	 */
1252 	if (!reg->hr_unclean_stop && !reg->hr_aborted_start) {
1253 		o2hb_prepare_block(reg, 0);
1254 		ret = o2hb_issue_node_write(reg, &write_wc);
1255 		if (ret == 0)
1256 			o2hb_wait_on_io(&write_wc);
1257 		else
1258 			mlog_errno(ret);
1259 	}
1260 
1261 	/* Unpin node */
1262 	o2nm_undepend_this_node();
1263 
1264 	mlog(ML_HEARTBEAT|ML_KTHREAD, "o2hb thread exiting\n");
1265 
1266 	return 0;
1267 }
1268 
1269 #ifdef CONFIG_DEBUG_FS
1270 static int o2hb_debug_open(struct inode *inode, struct file *file)
1271 {
1272 	struct o2hb_debug_buf *db = inode->i_private;
1273 	struct o2hb_region *reg;
1274 	unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)];
1275 	unsigned long lts;
1276 	char *buf = NULL;
1277 	int i = -1;
1278 	int out = 0;
1279 
1280 	/* max_nodes should be the largest bitmap we pass here */
1281 	BUG_ON(sizeof(map) < db->db_size);
1282 
1283 	buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
1284 	if (!buf)
1285 		goto bail;
1286 
1287 	switch (db->db_type) {
1288 	case O2HB_DB_TYPE_LIVENODES:
1289 	case O2HB_DB_TYPE_LIVEREGIONS:
1290 	case O2HB_DB_TYPE_QUORUMREGIONS:
1291 	case O2HB_DB_TYPE_FAILEDREGIONS:
1292 		spin_lock(&o2hb_live_lock);
1293 		memcpy(map, db->db_data, db->db_size);
1294 		spin_unlock(&o2hb_live_lock);
1295 		break;
1296 
1297 	case O2HB_DB_TYPE_REGION_LIVENODES:
1298 		spin_lock(&o2hb_live_lock);
1299 		reg = (struct o2hb_region *)db->db_data;
1300 		memcpy(map, reg->hr_live_node_bitmap, db->db_size);
1301 		spin_unlock(&o2hb_live_lock);
1302 		break;
1303 
1304 	case O2HB_DB_TYPE_REGION_NUMBER:
1305 		reg = (struct o2hb_region *)db->db_data;
1306 		out += scnprintf(buf + out, PAGE_SIZE - out, "%d\n",
1307 				reg->hr_region_num);
1308 		goto done;
1309 
1310 	case O2HB_DB_TYPE_REGION_ELAPSED_TIME:
1311 		reg = (struct o2hb_region *)db->db_data;
1312 		lts = reg->hr_last_timeout_start;
1313 		/* If 0, it has never been set before */
1314 		if (lts)
1315 			lts = jiffies_to_msecs(jiffies - lts);
1316 		out += scnprintf(buf + out, PAGE_SIZE - out, "%lu\n", lts);
1317 		goto done;
1318 
1319 	case O2HB_DB_TYPE_REGION_PINNED:
1320 		reg = (struct o2hb_region *)db->db_data;
1321 		out += scnprintf(buf + out, PAGE_SIZE - out, "%u\n",
1322 				!!reg->hr_item_pinned);
1323 		goto done;
1324 
1325 	default:
1326 		goto done;
1327 	}
1328 
1329 	while ((i = find_next_bit(map, db->db_len, i + 1)) < db->db_len)
1330 		out += scnprintf(buf + out, PAGE_SIZE - out, "%d ", i);
1331 	out += scnprintf(buf + out, PAGE_SIZE - out, "\n");
1332 
1333 done:
1334 	i_size_write(inode, out);
1335 
1336 	file->private_data = buf;
1337 
1338 	return 0;
1339 bail:
1340 	return -ENOMEM;
1341 }
1342 
1343 static int o2hb_debug_release(struct inode *inode, struct file *file)
1344 {
1345 	kfree(file->private_data);
1346 	return 0;
1347 }
1348 
1349 static ssize_t o2hb_debug_read(struct file *file, char __user *buf,
1350 				 size_t nbytes, loff_t *ppos)
1351 {
1352 	return simple_read_from_buffer(buf, nbytes, ppos, file->private_data,
1353 				       i_size_read(file->f_mapping->host));
1354 }
1355 #else
1356 static int o2hb_debug_open(struct inode *inode, struct file *file)
1357 {
1358 	return 0;
1359 }
1360 static int o2hb_debug_release(struct inode *inode, struct file *file)
1361 {
1362 	return 0;
1363 }
1364 static ssize_t o2hb_debug_read(struct file *file, char __user *buf,
1365 			       size_t nbytes, loff_t *ppos)
1366 {
1367 	return 0;
1368 }
1369 #endif  /* CONFIG_DEBUG_FS */
1370 
1371 static const struct file_operations o2hb_debug_fops = {
1372 	.open =		o2hb_debug_open,
1373 	.release =	o2hb_debug_release,
1374 	.read =		o2hb_debug_read,
1375 	.llseek =	generic_file_llseek,
1376 };
1377 
1378 void o2hb_exit(void)
1379 {
1380 	debugfs_remove_recursive(o2hb_debug_dir);
1381 	kfree(o2hb_db_livenodes);
1382 	kfree(o2hb_db_liveregions);
1383 	kfree(o2hb_db_quorumregions);
1384 	kfree(o2hb_db_failedregions);
1385 }
1386 
1387 static void o2hb_debug_create(const char *name, struct dentry *dir,
1388 			      struct o2hb_debug_buf **db, int db_len, int type,
1389 			      int size, int len, void *data)
1390 {
1391 	*db = kmalloc(db_len, GFP_KERNEL);
1392 	if (!*db)
1393 		return;
1394 
1395 	(*db)->db_type = type;
1396 	(*db)->db_size = size;
1397 	(*db)->db_len = len;
1398 	(*db)->db_data = data;
1399 
1400 	debugfs_create_file(name, S_IFREG|S_IRUSR, dir, *db, &o2hb_debug_fops);
1401 }
1402 
1403 static void o2hb_debug_init(void)
1404 {
1405 	o2hb_debug_dir = debugfs_create_dir(O2HB_DEBUG_DIR, NULL);
1406 
1407 	o2hb_debug_create(O2HB_DEBUG_LIVENODES, o2hb_debug_dir,
1408 			  &o2hb_db_livenodes, sizeof(*o2hb_db_livenodes),
1409 			  O2HB_DB_TYPE_LIVENODES, sizeof(o2hb_live_node_bitmap),
1410 			  O2NM_MAX_NODES, o2hb_live_node_bitmap);
1411 
1412 	o2hb_debug_create(O2HB_DEBUG_LIVEREGIONS, o2hb_debug_dir,
1413 			  &o2hb_db_liveregions, sizeof(*o2hb_db_liveregions),
1414 			  O2HB_DB_TYPE_LIVEREGIONS,
1415 			  sizeof(o2hb_live_region_bitmap), O2NM_MAX_REGIONS,
1416 			  o2hb_live_region_bitmap);
1417 
1418 	o2hb_debug_create(O2HB_DEBUG_QUORUMREGIONS, o2hb_debug_dir,
1419 			  &o2hb_db_quorumregions,
1420 			  sizeof(*o2hb_db_quorumregions),
1421 			  O2HB_DB_TYPE_QUORUMREGIONS,
1422 			  sizeof(o2hb_quorum_region_bitmap), O2NM_MAX_REGIONS,
1423 			  o2hb_quorum_region_bitmap);
1424 
1425 	o2hb_debug_create(O2HB_DEBUG_FAILEDREGIONS, o2hb_debug_dir,
1426 			  &o2hb_db_failedregions,
1427 			  sizeof(*o2hb_db_failedregions),
1428 			  O2HB_DB_TYPE_FAILEDREGIONS,
1429 			  sizeof(o2hb_failed_region_bitmap), O2NM_MAX_REGIONS,
1430 			  o2hb_failed_region_bitmap);
1431 }
1432 
1433 void o2hb_init(void)
1434 {
1435 	int i;
1436 
1437 	for (i = 0; i < ARRAY_SIZE(o2hb_callbacks); i++)
1438 		INIT_LIST_HEAD(&o2hb_callbacks[i].list);
1439 
1440 	for (i = 0; i < ARRAY_SIZE(o2hb_live_slots); i++)
1441 		INIT_LIST_HEAD(&o2hb_live_slots[i]);
1442 
1443 	memset(o2hb_live_node_bitmap, 0, sizeof(o2hb_live_node_bitmap));
1444 	memset(o2hb_region_bitmap, 0, sizeof(o2hb_region_bitmap));
1445 	memset(o2hb_live_region_bitmap, 0, sizeof(o2hb_live_region_bitmap));
1446 	memset(o2hb_quorum_region_bitmap, 0, sizeof(o2hb_quorum_region_bitmap));
1447 	memset(o2hb_failed_region_bitmap, 0, sizeof(o2hb_failed_region_bitmap));
1448 
1449 	o2hb_dependent_users = 0;
1450 
1451 	o2hb_debug_init();
1452 }
1453 
1454 /* if we're already in a callback then we're already serialized by the sem */
1455 static void o2hb_fill_node_map_from_callback(unsigned long *map,
1456 					     unsigned bytes)
1457 {
1458 	BUG_ON(bytes < (BITS_TO_LONGS(O2NM_MAX_NODES) * sizeof(unsigned long)));
1459 
1460 	memcpy(map, &o2hb_live_node_bitmap, bytes);
1461 }
1462 
1463 /*
1464  * get a map of all nodes that are heartbeating in any regions
1465  */
1466 void o2hb_fill_node_map(unsigned long *map, unsigned bytes)
1467 {
1468 	/* callers want to serialize this map and callbacks so that they
1469 	 * can trust that they don't miss nodes coming to the party */
1470 	down_read(&o2hb_callback_sem);
1471 	spin_lock(&o2hb_live_lock);
1472 	o2hb_fill_node_map_from_callback(map, bytes);
1473 	spin_unlock(&o2hb_live_lock);
1474 	up_read(&o2hb_callback_sem);
1475 }
1476 EXPORT_SYMBOL_GPL(o2hb_fill_node_map);
1477 
1478 /*
1479  * heartbeat configfs bits.  The heartbeat set is a default set under
1480  * the cluster set in nodemanager.c.
1481  */
1482 
1483 static struct o2hb_region *to_o2hb_region(struct config_item *item)
1484 {
1485 	return item ? container_of(item, struct o2hb_region, hr_item) : NULL;
1486 }
1487 
1488 /* drop_item only drops its ref after killing the thread, nothing should
1489  * be using the region anymore.  this has to clean up any state that
1490  * attributes might have built up. */
1491 static void o2hb_region_release(struct config_item *item)
1492 {
1493 	int i;
1494 	struct page *page;
1495 	struct o2hb_region *reg = to_o2hb_region(item);
1496 
1497 	mlog(ML_HEARTBEAT, "hb region release (%s)\n", reg->hr_dev_name);
1498 
1499 	kfree(reg->hr_tmp_block);
1500 
1501 	if (reg->hr_slot_data) {
1502 		for (i = 0; i < reg->hr_num_pages; i++) {
1503 			page = reg->hr_slot_data[i];
1504 			if (page)
1505 				__free_page(page);
1506 		}
1507 		kfree(reg->hr_slot_data);
1508 	}
1509 
1510 	if (reg->hr_bdev)
1511 		blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE);
1512 
1513 	kfree(reg->hr_slots);
1514 
1515 	debugfs_remove_recursive(reg->hr_debug_dir);
1516 	kfree(reg->hr_db_livenodes);
1517 	kfree(reg->hr_db_regnum);
1518 	kfree(reg->hr_db_elapsed_time);
1519 	kfree(reg->hr_db_pinned);
1520 
1521 	spin_lock(&o2hb_live_lock);
1522 	list_del(&reg->hr_all_item);
1523 	spin_unlock(&o2hb_live_lock);
1524 
1525 	o2net_unregister_handler_list(&reg->hr_handler_list);
1526 	kfree(reg);
1527 }
1528 
1529 static int o2hb_read_block_input(struct o2hb_region *reg,
1530 				 const char *page,
1531 				 unsigned long *ret_bytes,
1532 				 unsigned int *ret_bits)
1533 {
1534 	unsigned long bytes;
1535 	char *p = (char *)page;
1536 
1537 	bytes = simple_strtoul(p, &p, 0);
1538 	if (!p || (*p && (*p != '\n')))
1539 		return -EINVAL;
1540 
1541 	/* Heartbeat and fs min / max block sizes are the same. */
1542 	if (bytes > 4096 || bytes < 512)
1543 		return -ERANGE;
1544 	if (hweight16(bytes) != 1)
1545 		return -EINVAL;
1546 
1547 	if (ret_bytes)
1548 		*ret_bytes = bytes;
1549 	if (ret_bits)
1550 		*ret_bits = ffs(bytes) - 1;
1551 
1552 	return 0;
1553 }
1554 
1555 static ssize_t o2hb_region_block_bytes_show(struct config_item *item,
1556 					    char *page)
1557 {
1558 	return sprintf(page, "%u\n", to_o2hb_region(item)->hr_block_bytes);
1559 }
1560 
1561 static ssize_t o2hb_region_block_bytes_store(struct config_item *item,
1562 					     const char *page,
1563 					     size_t count)
1564 {
1565 	struct o2hb_region *reg = to_o2hb_region(item);
1566 	int status;
1567 	unsigned long block_bytes;
1568 	unsigned int block_bits;
1569 
1570 	if (reg->hr_bdev)
1571 		return -EINVAL;
1572 
1573 	status = o2hb_read_block_input(reg, page, &block_bytes,
1574 				       &block_bits);
1575 	if (status)
1576 		return status;
1577 
1578 	reg->hr_block_bytes = (unsigned int)block_bytes;
1579 	reg->hr_block_bits = block_bits;
1580 
1581 	return count;
1582 }
1583 
1584 static ssize_t o2hb_region_start_block_show(struct config_item *item,
1585 					    char *page)
1586 {
1587 	return sprintf(page, "%llu\n", to_o2hb_region(item)->hr_start_block);
1588 }
1589 
1590 static ssize_t o2hb_region_start_block_store(struct config_item *item,
1591 					     const char *page,
1592 					     size_t count)
1593 {
1594 	struct o2hb_region *reg = to_o2hb_region(item);
1595 	unsigned long long tmp;
1596 	char *p = (char *)page;
1597 	ssize_t ret;
1598 
1599 	if (reg->hr_bdev)
1600 		return -EINVAL;
1601 
1602 	ret = kstrtoull(p, 0, &tmp);
1603 	if (ret)
1604 		return -EINVAL;
1605 
1606 	reg->hr_start_block = tmp;
1607 
1608 	return count;
1609 }
1610 
1611 static ssize_t o2hb_region_blocks_show(struct config_item *item, char *page)
1612 {
1613 	return sprintf(page, "%d\n", to_o2hb_region(item)->hr_blocks);
1614 }
1615 
1616 static ssize_t o2hb_region_blocks_store(struct config_item *item,
1617 					const char *page,
1618 					size_t count)
1619 {
1620 	struct o2hb_region *reg = to_o2hb_region(item);
1621 	unsigned long tmp;
1622 	char *p = (char *)page;
1623 
1624 	if (reg->hr_bdev)
1625 		return -EINVAL;
1626 
1627 	tmp = simple_strtoul(p, &p, 0);
1628 	if (!p || (*p && (*p != '\n')))
1629 		return -EINVAL;
1630 
1631 	if (tmp > O2NM_MAX_NODES || tmp == 0)
1632 		return -ERANGE;
1633 
1634 	reg->hr_blocks = (unsigned int)tmp;
1635 
1636 	return count;
1637 }
1638 
1639 static ssize_t o2hb_region_dev_show(struct config_item *item, char *page)
1640 {
1641 	unsigned int ret = 0;
1642 
1643 	if (to_o2hb_region(item)->hr_bdev)
1644 		ret = sprintf(page, "%s\n", to_o2hb_region(item)->hr_dev_name);
1645 
1646 	return ret;
1647 }
1648 
1649 static void o2hb_init_region_params(struct o2hb_region *reg)
1650 {
1651 	reg->hr_slots_per_page = PAGE_SIZE >> reg->hr_block_bits;
1652 	reg->hr_timeout_ms = O2HB_REGION_TIMEOUT_MS;
1653 
1654 	mlog(ML_HEARTBEAT, "hr_start_block = %llu, hr_blocks = %u\n",
1655 	     reg->hr_start_block, reg->hr_blocks);
1656 	mlog(ML_HEARTBEAT, "hr_block_bytes = %u, hr_block_bits = %u\n",
1657 	     reg->hr_block_bytes, reg->hr_block_bits);
1658 	mlog(ML_HEARTBEAT, "hr_timeout_ms = %u\n", reg->hr_timeout_ms);
1659 	mlog(ML_HEARTBEAT, "dead threshold = %u\n", o2hb_dead_threshold);
1660 }
1661 
1662 static int o2hb_map_slot_data(struct o2hb_region *reg)
1663 {
1664 	int i, j;
1665 	unsigned int last_slot;
1666 	unsigned int spp = reg->hr_slots_per_page;
1667 	struct page *page;
1668 	char *raw;
1669 	struct o2hb_disk_slot *slot;
1670 
1671 	reg->hr_tmp_block = kmalloc(reg->hr_block_bytes, GFP_KERNEL);
1672 	if (reg->hr_tmp_block == NULL)
1673 		return -ENOMEM;
1674 
1675 	reg->hr_slots = kcalloc(reg->hr_blocks,
1676 				sizeof(struct o2hb_disk_slot), GFP_KERNEL);
1677 	if (reg->hr_slots == NULL)
1678 		return -ENOMEM;
1679 
1680 	for(i = 0; i < reg->hr_blocks; i++) {
1681 		slot = &reg->hr_slots[i];
1682 		slot->ds_node_num = i;
1683 		INIT_LIST_HEAD(&slot->ds_live_item);
1684 		slot->ds_raw_block = NULL;
1685 	}
1686 
1687 	reg->hr_num_pages = (reg->hr_blocks + spp - 1) / spp;
1688 	mlog(ML_HEARTBEAT, "Going to require %u pages to cover %u blocks "
1689 			   "at %u blocks per page\n",
1690 	     reg->hr_num_pages, reg->hr_blocks, spp);
1691 
1692 	reg->hr_slot_data = kcalloc(reg->hr_num_pages, sizeof(struct page *),
1693 				    GFP_KERNEL);
1694 	if (!reg->hr_slot_data)
1695 		return -ENOMEM;
1696 
1697 	for(i = 0; i < reg->hr_num_pages; i++) {
1698 		page = alloc_page(GFP_KERNEL);
1699 		if (!page)
1700 			return -ENOMEM;
1701 
1702 		reg->hr_slot_data[i] = page;
1703 
1704 		last_slot = i * spp;
1705 		raw = page_address(page);
1706 		for (j = 0;
1707 		     (j < spp) && ((j + last_slot) < reg->hr_blocks);
1708 		     j++) {
1709 			BUG_ON((j + last_slot) >= reg->hr_blocks);
1710 
1711 			slot = &reg->hr_slots[j + last_slot];
1712 			slot->ds_raw_block =
1713 				(struct o2hb_disk_heartbeat_block *) raw;
1714 
1715 			raw += reg->hr_block_bytes;
1716 		}
1717 	}
1718 
1719 	return 0;
1720 }
1721 
1722 /* Read in all the slots available and populate the tracking
1723  * structures so that we can start with a baseline idea of what's
1724  * there. */
1725 static int o2hb_populate_slot_data(struct o2hb_region *reg)
1726 {
1727 	int ret, i;
1728 	struct o2hb_disk_slot *slot;
1729 	struct o2hb_disk_heartbeat_block *hb_block;
1730 
1731 	ret = o2hb_read_slots(reg, 0, reg->hr_blocks);
1732 	if (ret)
1733 		goto out;
1734 
1735 	/* We only want to get an idea of the values initially in each
1736 	 * slot, so we do no verification - o2hb_check_slot will
1737 	 * actually determine if each configured slot is valid and
1738 	 * whether any values have changed. */
1739 	for(i = 0; i < reg->hr_blocks; i++) {
1740 		slot = &reg->hr_slots[i];
1741 		hb_block = (struct o2hb_disk_heartbeat_block *) slot->ds_raw_block;
1742 
1743 		/* Only fill the values that o2hb_check_slot uses to
1744 		 * determine changing slots */
1745 		slot->ds_last_time = le64_to_cpu(hb_block->hb_seq);
1746 		slot->ds_last_generation = le64_to_cpu(hb_block->hb_generation);
1747 	}
1748 
1749 out:
1750 	return ret;
1751 }
1752 
1753 /* this is acting as commit; we set up all of hr_bdev and hr_task or nothing */
1754 static ssize_t o2hb_region_dev_store(struct config_item *item,
1755 				     const char *page,
1756 				     size_t count)
1757 {
1758 	struct o2hb_region *reg = to_o2hb_region(item);
1759 	struct task_struct *hb_task;
1760 	long fd;
1761 	int sectsize;
1762 	char *p = (char *)page;
1763 	struct fd f;
1764 	ssize_t ret = -EINVAL;
1765 	int live_threshold;
1766 
1767 	if (reg->hr_bdev)
1768 		goto out;
1769 
1770 	/* We can't heartbeat without having had our node number
1771 	 * configured yet. */
1772 	if (o2nm_this_node() == O2NM_MAX_NODES)
1773 		goto out;
1774 
1775 	fd = simple_strtol(p, &p, 0);
1776 	if (!p || (*p && (*p != '\n')))
1777 		goto out;
1778 
1779 	if (fd < 0 || fd >= INT_MAX)
1780 		goto out;
1781 
1782 	f = fdget(fd);
1783 	if (f.file == NULL)
1784 		goto out;
1785 
1786 	if (reg->hr_blocks == 0 || reg->hr_start_block == 0 ||
1787 	    reg->hr_block_bytes == 0)
1788 		goto out2;
1789 
1790 	if (!S_ISBLK(f.file->f_mapping->host->i_mode))
1791 		goto out2;
1792 
1793 	reg->hr_bdev = blkdev_get_by_dev(f.file->f_mapping->host->i_rdev,
1794 					 FMODE_WRITE | FMODE_READ, NULL);
1795 	if (IS_ERR(reg->hr_bdev)) {
1796 		ret = PTR_ERR(reg->hr_bdev);
1797 		reg->hr_bdev = NULL;
1798 		goto out2;
1799 	}
1800 
1801 	bdevname(reg->hr_bdev, reg->hr_dev_name);
1802 
1803 	sectsize = bdev_logical_block_size(reg->hr_bdev);
1804 	if (sectsize != reg->hr_block_bytes) {
1805 		mlog(ML_ERROR,
1806 		     "blocksize %u incorrect for device, expected %d",
1807 		     reg->hr_block_bytes, sectsize);
1808 		ret = -EINVAL;
1809 		goto out3;
1810 	}
1811 
1812 	o2hb_init_region_params(reg);
1813 
1814 	/* Generation of zero is invalid */
1815 	do {
1816 		get_random_bytes(&reg->hr_generation,
1817 				 sizeof(reg->hr_generation));
1818 	} while (reg->hr_generation == 0);
1819 
1820 	ret = o2hb_map_slot_data(reg);
1821 	if (ret) {
1822 		mlog_errno(ret);
1823 		goto out3;
1824 	}
1825 
1826 	ret = o2hb_populate_slot_data(reg);
1827 	if (ret) {
1828 		mlog_errno(ret);
1829 		goto out3;
1830 	}
1831 
1832 	INIT_DELAYED_WORK(&reg->hr_write_timeout_work, o2hb_write_timeout);
1833 	INIT_DELAYED_WORK(&reg->hr_nego_timeout_work, o2hb_nego_timeout);
1834 
1835 	/*
1836 	 * A node is considered live after it has beat LIVE_THRESHOLD
1837 	 * times.  We're not steady until we've given them a chance
1838 	 * _after_ our first read.
1839 	 * The default threshold is bare minimum so as to limit the delay
1840 	 * during mounts. For global heartbeat, the threshold doubled for the
1841 	 * first region.
1842 	 */
1843 	live_threshold = O2HB_LIVE_THRESHOLD;
1844 	if (o2hb_global_heartbeat_active()) {
1845 		spin_lock(&o2hb_live_lock);
1846 		if (bitmap_weight(o2hb_region_bitmap, O2NM_MAX_REGIONS) == 1)
1847 			live_threshold <<= 1;
1848 		spin_unlock(&o2hb_live_lock);
1849 	}
1850 	++live_threshold;
1851 	atomic_set(&reg->hr_steady_iterations, live_threshold);
1852 	/* unsteady_iterations is triple the steady_iterations */
1853 	atomic_set(&reg->hr_unsteady_iterations, (live_threshold * 3));
1854 
1855 	hb_task = kthread_run(o2hb_thread, reg, "o2hb-%s",
1856 			      reg->hr_item.ci_name);
1857 	if (IS_ERR(hb_task)) {
1858 		ret = PTR_ERR(hb_task);
1859 		mlog_errno(ret);
1860 		goto out3;
1861 	}
1862 
1863 	spin_lock(&o2hb_live_lock);
1864 	reg->hr_task = hb_task;
1865 	spin_unlock(&o2hb_live_lock);
1866 
1867 	ret = wait_event_interruptible(o2hb_steady_queue,
1868 				atomic_read(&reg->hr_steady_iterations) == 0 ||
1869 				reg->hr_node_deleted);
1870 	if (ret) {
1871 		atomic_set(&reg->hr_steady_iterations, 0);
1872 		reg->hr_aborted_start = 1;
1873 	}
1874 
1875 	if (reg->hr_aborted_start) {
1876 		ret = -EIO;
1877 		goto out3;
1878 	}
1879 
1880 	if (reg->hr_node_deleted) {
1881 		ret = -EINVAL;
1882 		goto out3;
1883 	}
1884 
1885 	/* Ok, we were woken.  Make sure it wasn't by drop_item() */
1886 	spin_lock(&o2hb_live_lock);
1887 	hb_task = reg->hr_task;
1888 	if (o2hb_global_heartbeat_active())
1889 		set_bit(reg->hr_region_num, o2hb_live_region_bitmap);
1890 	spin_unlock(&o2hb_live_lock);
1891 
1892 	if (hb_task)
1893 		ret = count;
1894 	else
1895 		ret = -EIO;
1896 
1897 	if (hb_task && o2hb_global_heartbeat_active())
1898 		printk(KERN_NOTICE "o2hb: Heartbeat started on region %s (%s)\n",
1899 		       config_item_name(&reg->hr_item), reg->hr_dev_name);
1900 
1901 out3:
1902 	if (ret < 0) {
1903 		blkdev_put(reg->hr_bdev, FMODE_READ | FMODE_WRITE);
1904 		reg->hr_bdev = NULL;
1905 	}
1906 out2:
1907 	fdput(f);
1908 out:
1909 	return ret;
1910 }
1911 
1912 static ssize_t o2hb_region_pid_show(struct config_item *item, char *page)
1913 {
1914 	struct o2hb_region *reg = to_o2hb_region(item);
1915 	pid_t pid = 0;
1916 
1917 	spin_lock(&o2hb_live_lock);
1918 	if (reg->hr_task)
1919 		pid = task_pid_nr(reg->hr_task);
1920 	spin_unlock(&o2hb_live_lock);
1921 
1922 	if (!pid)
1923 		return 0;
1924 
1925 	return sprintf(page, "%u\n", pid);
1926 }
1927 
1928 CONFIGFS_ATTR(o2hb_region_, block_bytes);
1929 CONFIGFS_ATTR(o2hb_region_, start_block);
1930 CONFIGFS_ATTR(o2hb_region_, blocks);
1931 CONFIGFS_ATTR(o2hb_region_, dev);
1932 CONFIGFS_ATTR_RO(o2hb_region_, pid);
1933 
1934 static struct configfs_attribute *o2hb_region_attrs[] = {
1935 	&o2hb_region_attr_block_bytes,
1936 	&o2hb_region_attr_start_block,
1937 	&o2hb_region_attr_blocks,
1938 	&o2hb_region_attr_dev,
1939 	&o2hb_region_attr_pid,
1940 	NULL,
1941 };
1942 
1943 static struct configfs_item_operations o2hb_region_item_ops = {
1944 	.release		= o2hb_region_release,
1945 };
1946 
1947 static const struct config_item_type o2hb_region_type = {
1948 	.ct_item_ops	= &o2hb_region_item_ops,
1949 	.ct_attrs	= o2hb_region_attrs,
1950 	.ct_owner	= THIS_MODULE,
1951 };
1952 
1953 /* heartbeat set */
1954 
1955 struct o2hb_heartbeat_group {
1956 	struct config_group hs_group;
1957 	/* some stuff? */
1958 };
1959 
1960 static struct o2hb_heartbeat_group *to_o2hb_heartbeat_group(struct config_group *group)
1961 {
1962 	return group ?
1963 		container_of(group, struct o2hb_heartbeat_group, hs_group)
1964 		: NULL;
1965 }
1966 
1967 static void o2hb_debug_region_init(struct o2hb_region *reg,
1968 				   struct dentry *parent)
1969 {
1970 	struct dentry *dir;
1971 
1972 	dir = debugfs_create_dir(config_item_name(&reg->hr_item), parent);
1973 	reg->hr_debug_dir = dir;
1974 
1975 	o2hb_debug_create(O2HB_DEBUG_LIVENODES, dir, &(reg->hr_db_livenodes),
1976 			  sizeof(*(reg->hr_db_livenodes)),
1977 			  O2HB_DB_TYPE_REGION_LIVENODES,
1978 			  sizeof(reg->hr_live_node_bitmap), O2NM_MAX_NODES,
1979 			  reg);
1980 
1981 	o2hb_debug_create(O2HB_DEBUG_REGION_NUMBER, dir, &(reg->hr_db_regnum),
1982 			  sizeof(*(reg->hr_db_regnum)),
1983 			  O2HB_DB_TYPE_REGION_NUMBER, 0, O2NM_MAX_NODES, reg);
1984 
1985 	o2hb_debug_create(O2HB_DEBUG_REGION_ELAPSED_TIME, dir,
1986 			  &(reg->hr_db_elapsed_time),
1987 			  sizeof(*(reg->hr_db_elapsed_time)),
1988 			  O2HB_DB_TYPE_REGION_ELAPSED_TIME, 0, 0, reg);
1989 
1990 	o2hb_debug_create(O2HB_DEBUG_REGION_PINNED, dir, &(reg->hr_db_pinned),
1991 			  sizeof(*(reg->hr_db_pinned)),
1992 			  O2HB_DB_TYPE_REGION_PINNED, 0, 0, reg);
1993 
1994 }
1995 
1996 static struct config_item *o2hb_heartbeat_group_make_item(struct config_group *group,
1997 							  const char *name)
1998 {
1999 	struct o2hb_region *reg = NULL;
2000 	int ret;
2001 
2002 	reg = kzalloc(sizeof(struct o2hb_region), GFP_KERNEL);
2003 	if (reg == NULL)
2004 		return ERR_PTR(-ENOMEM);
2005 
2006 	if (strlen(name) > O2HB_MAX_REGION_NAME_LEN) {
2007 		ret = -ENAMETOOLONG;
2008 		goto free;
2009 	}
2010 
2011 	spin_lock(&o2hb_live_lock);
2012 	reg->hr_region_num = 0;
2013 	if (o2hb_global_heartbeat_active()) {
2014 		reg->hr_region_num = find_first_zero_bit(o2hb_region_bitmap,
2015 							 O2NM_MAX_REGIONS);
2016 		if (reg->hr_region_num >= O2NM_MAX_REGIONS) {
2017 			spin_unlock(&o2hb_live_lock);
2018 			ret = -EFBIG;
2019 			goto free;
2020 		}
2021 		set_bit(reg->hr_region_num, o2hb_region_bitmap);
2022 	}
2023 	list_add_tail(&reg->hr_all_item, &o2hb_all_regions);
2024 	spin_unlock(&o2hb_live_lock);
2025 
2026 	config_item_init_type_name(&reg->hr_item, name, &o2hb_region_type);
2027 
2028 	/* this is the same way to generate msg key as dlm, for local heartbeat,
2029 	 * name is also the same, so make initial crc value different to avoid
2030 	 * message key conflict.
2031 	 */
2032 	reg->hr_key = crc32_le(reg->hr_region_num + O2NM_MAX_REGIONS,
2033 		name, strlen(name));
2034 	INIT_LIST_HEAD(&reg->hr_handler_list);
2035 	ret = o2net_register_handler(O2HB_NEGO_TIMEOUT_MSG, reg->hr_key,
2036 			sizeof(struct o2hb_nego_msg),
2037 			o2hb_nego_timeout_handler,
2038 			reg, NULL, &reg->hr_handler_list);
2039 	if (ret)
2040 		goto remove_item;
2041 
2042 	ret = o2net_register_handler(O2HB_NEGO_APPROVE_MSG, reg->hr_key,
2043 			sizeof(struct o2hb_nego_msg),
2044 			o2hb_nego_approve_handler,
2045 			reg, NULL, &reg->hr_handler_list);
2046 	if (ret)
2047 		goto unregister_handler;
2048 
2049 	o2hb_debug_region_init(reg, o2hb_debug_dir);
2050 
2051 	return &reg->hr_item;
2052 
2053 unregister_handler:
2054 	o2net_unregister_handler_list(&reg->hr_handler_list);
2055 remove_item:
2056 	spin_lock(&o2hb_live_lock);
2057 	list_del(&reg->hr_all_item);
2058 	if (o2hb_global_heartbeat_active())
2059 		clear_bit(reg->hr_region_num, o2hb_region_bitmap);
2060 	spin_unlock(&o2hb_live_lock);
2061 free:
2062 	kfree(reg);
2063 	return ERR_PTR(ret);
2064 }
2065 
2066 static void o2hb_heartbeat_group_drop_item(struct config_group *group,
2067 					   struct config_item *item)
2068 {
2069 	struct task_struct *hb_task;
2070 	struct o2hb_region *reg = to_o2hb_region(item);
2071 	int quorum_region = 0;
2072 
2073 	/* stop the thread when the user removes the region dir */
2074 	spin_lock(&o2hb_live_lock);
2075 	hb_task = reg->hr_task;
2076 	reg->hr_task = NULL;
2077 	reg->hr_item_dropped = 1;
2078 	spin_unlock(&o2hb_live_lock);
2079 
2080 	if (hb_task)
2081 		kthread_stop(hb_task);
2082 
2083 	if (o2hb_global_heartbeat_active()) {
2084 		spin_lock(&o2hb_live_lock);
2085 		clear_bit(reg->hr_region_num, o2hb_region_bitmap);
2086 		clear_bit(reg->hr_region_num, o2hb_live_region_bitmap);
2087 		if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
2088 			quorum_region = 1;
2089 		clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
2090 		spin_unlock(&o2hb_live_lock);
2091 		printk(KERN_NOTICE "o2hb: Heartbeat %s on region %s (%s)\n",
2092 		       ((atomic_read(&reg->hr_steady_iterations) == 0) ?
2093 			"stopped" : "start aborted"), config_item_name(item),
2094 		       reg->hr_dev_name);
2095 	}
2096 
2097 	/*
2098 	 * If we're racing a dev_write(), we need to wake them.  They will
2099 	 * check reg->hr_task
2100 	 */
2101 	if (atomic_read(&reg->hr_steady_iterations) != 0) {
2102 		reg->hr_aborted_start = 1;
2103 		atomic_set(&reg->hr_steady_iterations, 0);
2104 		wake_up(&o2hb_steady_queue);
2105 	}
2106 
2107 	config_item_put(item);
2108 
2109 	if (!o2hb_global_heartbeat_active() || !quorum_region)
2110 		return;
2111 
2112 	/*
2113 	 * If global heartbeat active and there are dependent users,
2114 	 * pin all regions if quorum region count <= CUT_OFF
2115 	 */
2116 	spin_lock(&o2hb_live_lock);
2117 
2118 	if (!o2hb_dependent_users)
2119 		goto unlock;
2120 
2121 	if (bitmap_weight(o2hb_quorum_region_bitmap,
2122 			   O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
2123 		o2hb_region_pin(NULL);
2124 
2125 unlock:
2126 	spin_unlock(&o2hb_live_lock);
2127 }
2128 
2129 static ssize_t o2hb_heartbeat_group_dead_threshold_show(struct config_item *item,
2130 		char *page)
2131 {
2132 	return sprintf(page, "%u\n", o2hb_dead_threshold);
2133 }
2134 
2135 static ssize_t o2hb_heartbeat_group_dead_threshold_store(struct config_item *item,
2136 		const char *page, size_t count)
2137 {
2138 	unsigned long tmp;
2139 	char *p = (char *)page;
2140 
2141 	tmp = simple_strtoul(p, &p, 10);
2142 	if (!p || (*p && (*p != '\n')))
2143                 return -EINVAL;
2144 
2145 	/* this will validate ranges for us. */
2146 	o2hb_dead_threshold_set((unsigned int) tmp);
2147 
2148 	return count;
2149 }
2150 
2151 static ssize_t o2hb_heartbeat_group_mode_show(struct config_item *item,
2152 		char *page)
2153 {
2154 	return sprintf(page, "%s\n",
2155 		       o2hb_heartbeat_mode_desc[o2hb_heartbeat_mode]);
2156 }
2157 
2158 static ssize_t o2hb_heartbeat_group_mode_store(struct config_item *item,
2159 		const char *page, size_t count)
2160 {
2161 	unsigned int i;
2162 	int ret;
2163 	size_t len;
2164 
2165 	len = (page[count - 1] == '\n') ? count - 1 : count;
2166 	if (!len)
2167 		return -EINVAL;
2168 
2169 	for (i = 0; i < O2HB_HEARTBEAT_NUM_MODES; ++i) {
2170 		if (strncasecmp(page, o2hb_heartbeat_mode_desc[i], len))
2171 			continue;
2172 
2173 		ret = o2hb_global_heartbeat_mode_set(i);
2174 		if (!ret)
2175 			printk(KERN_NOTICE "o2hb: Heartbeat mode set to %s\n",
2176 			       o2hb_heartbeat_mode_desc[i]);
2177 		return count;
2178 	}
2179 
2180 	return -EINVAL;
2181 
2182 }
2183 
2184 CONFIGFS_ATTR(o2hb_heartbeat_group_, dead_threshold);
2185 CONFIGFS_ATTR(o2hb_heartbeat_group_, mode);
2186 
2187 static struct configfs_attribute *o2hb_heartbeat_group_attrs[] = {
2188 	&o2hb_heartbeat_group_attr_dead_threshold,
2189 	&o2hb_heartbeat_group_attr_mode,
2190 	NULL,
2191 };
2192 
2193 static struct configfs_group_operations o2hb_heartbeat_group_group_ops = {
2194 	.make_item	= o2hb_heartbeat_group_make_item,
2195 	.drop_item	= o2hb_heartbeat_group_drop_item,
2196 };
2197 
2198 static const struct config_item_type o2hb_heartbeat_group_type = {
2199 	.ct_group_ops	= &o2hb_heartbeat_group_group_ops,
2200 	.ct_attrs	= o2hb_heartbeat_group_attrs,
2201 	.ct_owner	= THIS_MODULE,
2202 };
2203 
2204 /* this is just here to avoid touching group in heartbeat.h which the
2205  * entire damn world #includes */
2206 struct config_group *o2hb_alloc_hb_set(void)
2207 {
2208 	struct o2hb_heartbeat_group *hs = NULL;
2209 	struct config_group *ret = NULL;
2210 
2211 	hs = kzalloc(sizeof(struct o2hb_heartbeat_group), GFP_KERNEL);
2212 	if (hs == NULL)
2213 		goto out;
2214 
2215 	config_group_init_type_name(&hs->hs_group, "heartbeat",
2216 				    &o2hb_heartbeat_group_type);
2217 
2218 	ret = &hs->hs_group;
2219 out:
2220 	if (ret == NULL)
2221 		kfree(hs);
2222 	return ret;
2223 }
2224 
2225 void o2hb_free_hb_set(struct config_group *group)
2226 {
2227 	struct o2hb_heartbeat_group *hs = to_o2hb_heartbeat_group(group);
2228 	kfree(hs);
2229 }
2230 
2231 /* hb callback registration and issuing */
2232 
2233 static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type)
2234 {
2235 	if (type == O2HB_NUM_CB)
2236 		return ERR_PTR(-EINVAL);
2237 
2238 	return &o2hb_callbacks[type];
2239 }
2240 
2241 void o2hb_setup_callback(struct o2hb_callback_func *hc,
2242 			 enum o2hb_callback_type type,
2243 			 o2hb_cb_func *func,
2244 			 void *data,
2245 			 int priority)
2246 {
2247 	INIT_LIST_HEAD(&hc->hc_item);
2248 	hc->hc_func = func;
2249 	hc->hc_data = data;
2250 	hc->hc_priority = priority;
2251 	hc->hc_type = type;
2252 	hc->hc_magic = O2HB_CB_MAGIC;
2253 }
2254 EXPORT_SYMBOL_GPL(o2hb_setup_callback);
2255 
2256 /*
2257  * In local heartbeat mode, region_uuid passed matches the dlm domain name.
2258  * In global heartbeat mode, region_uuid passed is NULL.
2259  *
2260  * In local, we only pin the matching region. In global we pin all the active
2261  * regions.
2262  */
2263 static int o2hb_region_pin(const char *region_uuid)
2264 {
2265 	int ret = 0, found = 0;
2266 	struct o2hb_region *reg;
2267 	char *uuid;
2268 
2269 	assert_spin_locked(&o2hb_live_lock);
2270 
2271 	list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2272 		if (reg->hr_item_dropped)
2273 			continue;
2274 
2275 		uuid = config_item_name(&reg->hr_item);
2276 
2277 		/* local heartbeat */
2278 		if (region_uuid) {
2279 			if (strcmp(region_uuid, uuid))
2280 				continue;
2281 			found = 1;
2282 		}
2283 
2284 		if (reg->hr_item_pinned || reg->hr_item_dropped)
2285 			goto skip_pin;
2286 
2287 		/* Ignore ENOENT only for local hb (userdlm domain) */
2288 		ret = o2nm_depend_item(&reg->hr_item);
2289 		if (!ret) {
2290 			mlog(ML_CLUSTER, "Pin region %s\n", uuid);
2291 			reg->hr_item_pinned = 1;
2292 		} else {
2293 			if (ret == -ENOENT && found)
2294 				ret = 0;
2295 			else {
2296 				mlog(ML_ERROR, "Pin region %s fails with %d\n",
2297 				     uuid, ret);
2298 				break;
2299 			}
2300 		}
2301 skip_pin:
2302 		if (found)
2303 			break;
2304 	}
2305 
2306 	return ret;
2307 }
2308 
2309 /*
2310  * In local heartbeat mode, region_uuid passed matches the dlm domain name.
2311  * In global heartbeat mode, region_uuid passed is NULL.
2312  *
2313  * In local, we only unpin the matching region. In global we unpin all the
2314  * active regions.
2315  */
2316 static void o2hb_region_unpin(const char *region_uuid)
2317 {
2318 	struct o2hb_region *reg;
2319 	char *uuid;
2320 	int found = 0;
2321 
2322 	assert_spin_locked(&o2hb_live_lock);
2323 
2324 	list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2325 		if (reg->hr_item_dropped)
2326 			continue;
2327 
2328 		uuid = config_item_name(&reg->hr_item);
2329 		if (region_uuid) {
2330 			if (strcmp(region_uuid, uuid))
2331 				continue;
2332 			found = 1;
2333 		}
2334 
2335 		if (reg->hr_item_pinned) {
2336 			mlog(ML_CLUSTER, "Unpin region %s\n", uuid);
2337 			o2nm_undepend_item(&reg->hr_item);
2338 			reg->hr_item_pinned = 0;
2339 		}
2340 		if (found)
2341 			break;
2342 	}
2343 }
2344 
2345 static int o2hb_region_inc_user(const char *region_uuid)
2346 {
2347 	int ret = 0;
2348 
2349 	spin_lock(&o2hb_live_lock);
2350 
2351 	/* local heartbeat */
2352 	if (!o2hb_global_heartbeat_active()) {
2353 	    ret = o2hb_region_pin(region_uuid);
2354 	    goto unlock;
2355 	}
2356 
2357 	/*
2358 	 * if global heartbeat active and this is the first dependent user,
2359 	 * pin all regions if quorum region count <= CUT_OFF
2360 	 */
2361 	o2hb_dependent_users++;
2362 	if (o2hb_dependent_users > 1)
2363 		goto unlock;
2364 
2365 	if (bitmap_weight(o2hb_quorum_region_bitmap,
2366 			   O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
2367 		ret = o2hb_region_pin(NULL);
2368 
2369 unlock:
2370 	spin_unlock(&o2hb_live_lock);
2371 	return ret;
2372 }
2373 
2374 static void o2hb_region_dec_user(const char *region_uuid)
2375 {
2376 	spin_lock(&o2hb_live_lock);
2377 
2378 	/* local heartbeat */
2379 	if (!o2hb_global_heartbeat_active()) {
2380 	    o2hb_region_unpin(region_uuid);
2381 	    goto unlock;
2382 	}
2383 
2384 	/*
2385 	 * if global heartbeat active and there are no dependent users,
2386 	 * unpin all quorum regions
2387 	 */
2388 	o2hb_dependent_users--;
2389 	if (!o2hb_dependent_users)
2390 		o2hb_region_unpin(NULL);
2391 
2392 unlock:
2393 	spin_unlock(&o2hb_live_lock);
2394 }
2395 
2396 int o2hb_register_callback(const char *region_uuid,
2397 			   struct o2hb_callback_func *hc)
2398 {
2399 	struct o2hb_callback_func *f;
2400 	struct o2hb_callback *hbcall;
2401 	int ret;
2402 
2403 	BUG_ON(hc->hc_magic != O2HB_CB_MAGIC);
2404 	BUG_ON(!list_empty(&hc->hc_item));
2405 
2406 	hbcall = hbcall_from_type(hc->hc_type);
2407 	if (IS_ERR(hbcall)) {
2408 		ret = PTR_ERR(hbcall);
2409 		goto out;
2410 	}
2411 
2412 	if (region_uuid) {
2413 		ret = o2hb_region_inc_user(region_uuid);
2414 		if (ret) {
2415 			mlog_errno(ret);
2416 			goto out;
2417 		}
2418 	}
2419 
2420 	down_write(&o2hb_callback_sem);
2421 
2422 	list_for_each_entry(f, &hbcall->list, hc_item) {
2423 		if (hc->hc_priority < f->hc_priority) {
2424 			list_add_tail(&hc->hc_item, &f->hc_item);
2425 			break;
2426 		}
2427 	}
2428 	if (list_empty(&hc->hc_item))
2429 		list_add_tail(&hc->hc_item, &hbcall->list);
2430 
2431 	up_write(&o2hb_callback_sem);
2432 	ret = 0;
2433 out:
2434 	mlog(ML_CLUSTER, "returning %d on behalf of %p for funcs %p\n",
2435 	     ret, __builtin_return_address(0), hc);
2436 	return ret;
2437 }
2438 EXPORT_SYMBOL_GPL(o2hb_register_callback);
2439 
2440 void o2hb_unregister_callback(const char *region_uuid,
2441 			      struct o2hb_callback_func *hc)
2442 {
2443 	BUG_ON(hc->hc_magic != O2HB_CB_MAGIC);
2444 
2445 	mlog(ML_CLUSTER, "on behalf of %p for funcs %p\n",
2446 	     __builtin_return_address(0), hc);
2447 
2448 	/* XXX Can this happen _with_ a region reference? */
2449 	if (list_empty(&hc->hc_item))
2450 		return;
2451 
2452 	if (region_uuid)
2453 		o2hb_region_dec_user(region_uuid);
2454 
2455 	down_write(&o2hb_callback_sem);
2456 
2457 	list_del_init(&hc->hc_item);
2458 
2459 	up_write(&o2hb_callback_sem);
2460 }
2461 EXPORT_SYMBOL_GPL(o2hb_unregister_callback);
2462 
2463 int o2hb_check_node_heartbeating_no_sem(u8 node_num)
2464 {
2465 	unsigned long testing_map[BITS_TO_LONGS(O2NM_MAX_NODES)];
2466 
2467 	spin_lock(&o2hb_live_lock);
2468 	o2hb_fill_node_map_from_callback(testing_map, sizeof(testing_map));
2469 	spin_unlock(&o2hb_live_lock);
2470 	if (!test_bit(node_num, testing_map)) {
2471 		mlog(ML_HEARTBEAT,
2472 		     "node (%u) does not have heartbeating enabled.\n",
2473 		     node_num);
2474 		return 0;
2475 	}
2476 
2477 	return 1;
2478 }
2479 EXPORT_SYMBOL_GPL(o2hb_check_node_heartbeating_no_sem);
2480 
2481 int o2hb_check_node_heartbeating_from_callback(u8 node_num)
2482 {
2483 	unsigned long testing_map[BITS_TO_LONGS(O2NM_MAX_NODES)];
2484 
2485 	o2hb_fill_node_map_from_callback(testing_map, sizeof(testing_map));
2486 	if (!test_bit(node_num, testing_map)) {
2487 		mlog(ML_HEARTBEAT,
2488 		     "node (%u) does not have heartbeating enabled.\n",
2489 		     node_num);
2490 		return 0;
2491 	}
2492 
2493 	return 1;
2494 }
2495 EXPORT_SYMBOL_GPL(o2hb_check_node_heartbeating_from_callback);
2496 
2497 /*
2498  * this is just a hack until we get the plumbing which flips file systems
2499  * read only and drops the hb ref instead of killing the node dead.
2500  */
2501 void o2hb_stop_all_regions(void)
2502 {
2503 	struct o2hb_region *reg;
2504 
2505 	mlog(ML_ERROR, "stopping heartbeat on all active regions.\n");
2506 
2507 	spin_lock(&o2hb_live_lock);
2508 
2509 	list_for_each_entry(reg, &o2hb_all_regions, hr_all_item)
2510 		reg->hr_unclean_stop = 1;
2511 
2512 	spin_unlock(&o2hb_live_lock);
2513 }
2514 EXPORT_SYMBOL_GPL(o2hb_stop_all_regions);
2515 
2516 int o2hb_get_all_regions(char *region_uuids, u8 max_regions)
2517 {
2518 	struct o2hb_region *reg;
2519 	int numregs = 0;
2520 	char *p;
2521 
2522 	spin_lock(&o2hb_live_lock);
2523 
2524 	p = region_uuids;
2525 	list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2526 		if (reg->hr_item_dropped)
2527 			continue;
2528 
2529 		mlog(0, "Region: %s\n", config_item_name(&reg->hr_item));
2530 		if (numregs < max_regions) {
2531 			memcpy(p, config_item_name(&reg->hr_item),
2532 			       O2HB_MAX_REGION_NAME_LEN);
2533 			p += O2HB_MAX_REGION_NAME_LEN;
2534 		}
2535 		numregs++;
2536 	}
2537 
2538 	spin_unlock(&o2hb_live_lock);
2539 
2540 	return numregs;
2541 }
2542 EXPORT_SYMBOL_GPL(o2hb_get_all_regions);
2543 
2544 int o2hb_global_heartbeat_active(void)
2545 {
2546 	return (o2hb_heartbeat_mode == O2HB_HEARTBEAT_GLOBAL);
2547 }
2548 EXPORT_SYMBOL(o2hb_global_heartbeat_active);
2549