1 /* -*- mode: c; c-basic-offset: 8; -*-
2 * vim: noexpandtab sw=8 ts=8 sts=0:
4 * Copyright (C) 2004, 2005 Oracle. All rights reserved.
6 * This program is free software; you can redistribute it and/or
7 * modify it under the terms of the GNU General Public
8 * License as published by the Free Software Foundation; either
9 * version 2 of the License, or (at your option) any later version.
11 * This program is distributed in the hope that it will be useful,
12 * but WITHOUT ANY WARRANTY; without even the implied warranty of
13 * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE. See the GNU
14 * General Public License for more details.
16 * You should have received a copy of the GNU General Public
17 * License along with this program; if not, write to the
18 * Free Software Foundation, Inc., 59 Temple Place - Suite 330,
19 * Boston, MA 021110-1307, USA.
22 #include <linux/kernel.h>
23 #include <linux/sched.h>
24 #include <linux/jiffies.h>
25 #include <linux/module.h>
27 #include <linux/bio.h>
28 #include <linux/blkdev.h>
29 #include <linux/delay.h>
30 #include <linux/file.h>
31 #include <linux/kthread.h>
32 #include <linux/configfs.h>
33 #include <linux/random.h>
34 #include <linux/crc32.h>
35 #include <linux/time.h>
36 #include <linux/debugfs.h>
37 #include <linux/slab.h>
38 #include <linux/bitmap.h>
39 #include <linux/ktime.h>
40 #include "heartbeat.h"
42 #include "nodemanager.h"
49 * The first heartbeat pass had one global thread that would serialize all hb
50 * callback calls. This global serializing sem should only be removed once
51 * we've made sure that all callees can deal with being called concurrently
52 * from multiple hb region threads.
54 static DECLARE_RWSEM(o2hb_callback_sem);
57 * multiple hb threads are watching multiple regions. A node is live
58 * whenever any of the threads sees activity from the node in its region.
60 static DEFINE_SPINLOCK(o2hb_live_lock);
61 static struct list_head o2hb_live_slots[O2NM_MAX_NODES];
62 static unsigned long o2hb_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
63 static LIST_HEAD(o2hb_node_events);
64 static DECLARE_WAIT_QUEUE_HEAD(o2hb_steady_queue);
67 * In global heartbeat, we maintain a series of region bitmaps.
68 * - o2hb_region_bitmap allows us to limit the region number to max region.
69 * - o2hb_live_region_bitmap tracks live regions (seen steady iterations).
70 * - o2hb_quorum_region_bitmap tracks live regions that have seen all nodes
72 * - o2hb_failed_region_bitmap tracks the regions that have seen io timeouts.
74 static unsigned long o2hb_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
75 static unsigned long o2hb_live_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
76 static unsigned long o2hb_quorum_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
77 static unsigned long o2hb_failed_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
79 #define O2HB_DB_TYPE_LIVENODES 0
80 #define O2HB_DB_TYPE_LIVEREGIONS 1
81 #define O2HB_DB_TYPE_QUORUMREGIONS 2
82 #define O2HB_DB_TYPE_FAILEDREGIONS 3
83 #define O2HB_DB_TYPE_REGION_LIVENODES 4
84 #define O2HB_DB_TYPE_REGION_NUMBER 5
85 #define O2HB_DB_TYPE_REGION_ELAPSED_TIME 6
86 #define O2HB_DB_TYPE_REGION_PINNED 7
87 struct o2hb_debug_buf {
94 static struct o2hb_debug_buf *o2hb_db_livenodes;
95 static struct o2hb_debug_buf *o2hb_db_liveregions;
96 static struct o2hb_debug_buf *o2hb_db_quorumregions;
97 static struct o2hb_debug_buf *o2hb_db_failedregions;
99 #define O2HB_DEBUG_DIR "o2hb"
100 #define O2HB_DEBUG_LIVENODES "livenodes"
101 #define O2HB_DEBUG_LIVEREGIONS "live_regions"
102 #define O2HB_DEBUG_QUORUMREGIONS "quorum_regions"
103 #define O2HB_DEBUG_FAILEDREGIONS "failed_regions"
104 #define O2HB_DEBUG_REGION_NUMBER "num"
105 #define O2HB_DEBUG_REGION_ELAPSED_TIME "elapsed_time_in_ms"
106 #define O2HB_DEBUG_REGION_PINNED "pinned"
108 static struct dentry *o2hb_debug_dir;
109 static struct dentry *o2hb_debug_livenodes;
110 static struct dentry *o2hb_debug_liveregions;
111 static struct dentry *o2hb_debug_quorumregions;
112 static struct dentry *o2hb_debug_failedregions;
114 static LIST_HEAD(o2hb_all_regions);
116 static struct o2hb_callback {
117 struct list_head list;
118 } o2hb_callbacks[O2HB_NUM_CB];
120 static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type);
122 #define O2HB_DEFAULT_BLOCK_BITS 9
124 enum o2hb_heartbeat_modes {
125 O2HB_HEARTBEAT_LOCAL = 0,
126 O2HB_HEARTBEAT_GLOBAL,
127 O2HB_HEARTBEAT_NUM_MODES,
130 char *o2hb_heartbeat_mode_desc[O2HB_HEARTBEAT_NUM_MODES] = {
131 "local", /* O2HB_HEARTBEAT_LOCAL */
132 "global", /* O2HB_HEARTBEAT_GLOBAL */
135 unsigned int o2hb_dead_threshold = O2HB_DEFAULT_DEAD_THRESHOLD;
136 unsigned int o2hb_heartbeat_mode = O2HB_HEARTBEAT_LOCAL;
139 * o2hb_dependent_users tracks the number of registered callbacks that depend
140 * on heartbeat. o2net and o2dlm are two entities that register this callback.
141 * However only o2dlm depends on the heartbeat. It does not want the heartbeat
142 * to stop while a dlm domain is still active.
144 unsigned int o2hb_dependent_users;
147 * In global heartbeat mode, all regions are pinned if there are one or more
148 * dependent users and the quorum region count is <= O2HB_PIN_CUT_OFF. All
149 * regions are unpinned if the region count exceeds the cut off or the number
150 * of dependent users falls to zero.
152 #define O2HB_PIN_CUT_OFF 3
155 * In local heartbeat mode, we assume the dlm domain name to be the same as
156 * region uuid. This is true for domains created for the file system but not
157 * necessarily true for userdlm domains. This is a known limitation.
159 * In global heartbeat mode, we pin/unpin all o2hb regions. This solution
160 * works for both file system and userdlm domains.
162 static int o2hb_region_pin(const char *region_uuid);
163 static void o2hb_region_unpin(const char *region_uuid);
165 /* Only sets a new threshold if there are no active regions.
167 * No locking or otherwise interesting code is required for reading
168 * o2hb_dead_threshold as it can't change once regions are active and
169 * it's not interesting to anyone until then anyway. */
170 static void o2hb_dead_threshold_set(unsigned int threshold)
172 if (threshold > O2HB_MIN_DEAD_THRESHOLD) {
173 spin_lock(&o2hb_live_lock);
174 if (list_empty(&o2hb_all_regions))
175 o2hb_dead_threshold = threshold;
176 spin_unlock(&o2hb_live_lock);
180 static int o2hb_global_heartbeat_mode_set(unsigned int hb_mode)
184 if (hb_mode < O2HB_HEARTBEAT_NUM_MODES) {
185 spin_lock(&o2hb_live_lock);
186 if (list_empty(&o2hb_all_regions)) {
187 o2hb_heartbeat_mode = hb_mode;
190 spin_unlock(&o2hb_live_lock);
196 struct o2hb_node_event {
197 struct list_head hn_item;
198 enum o2hb_callback_type hn_event_type;
199 struct o2nm_node *hn_node;
203 struct o2hb_disk_slot {
204 struct o2hb_disk_heartbeat_block *ds_raw_block;
207 u64 ds_last_generation;
208 u16 ds_equal_samples;
209 u16 ds_changed_samples;
210 struct list_head ds_live_item;
213 /* each thread owns a region.. when we're asked to tear down the region
214 * we ask the thread to stop, who cleans up the region */
216 struct config_item hr_item;
218 struct list_head hr_all_item;
219 unsigned hr_unclean_stop:1,
225 /* protected by the hr_callback_sem */
226 struct task_struct *hr_task;
228 unsigned int hr_blocks;
229 unsigned long long hr_start_block;
231 unsigned int hr_block_bits;
232 unsigned int hr_block_bytes;
234 unsigned int hr_slots_per_page;
235 unsigned int hr_num_pages;
237 struct page **hr_slot_data;
238 struct block_device *hr_bdev;
239 struct o2hb_disk_slot *hr_slots;
241 /* live node map of this region */
242 unsigned long hr_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
243 unsigned int hr_region_num;
245 struct dentry *hr_debug_dir;
246 struct dentry *hr_debug_livenodes;
247 struct dentry *hr_debug_regnum;
248 struct dentry *hr_debug_elapsed_time;
249 struct dentry *hr_debug_pinned;
250 struct o2hb_debug_buf *hr_db_livenodes;
251 struct o2hb_debug_buf *hr_db_regnum;
252 struct o2hb_debug_buf *hr_db_elapsed_time;
253 struct o2hb_debug_buf *hr_db_pinned;
255 /* let the person setting up hb wait for it to return until it
256 * has reached a 'steady' state. This will be fixed when we have
257 * a more complete api that doesn't lead to this sort of fragility. */
258 atomic_t hr_steady_iterations;
260 /* terminate o2hb thread if it does not reach steady state
261 * (hr_steady_iterations == 0) within hr_unsteady_iterations */
262 atomic_t hr_unsteady_iterations;
264 char hr_dev_name[BDEVNAME_SIZE];
266 unsigned int hr_timeout_ms;
268 /* randomized as the region goes up and down so that a node
269 * recognizes a node going up and down in one iteration */
272 struct delayed_work hr_write_timeout_work;
273 unsigned long hr_last_timeout_start;
275 /* negotiate timer, used to negotiate extending hb timeout. */
276 struct delayed_work hr_nego_timeout_work;
277 unsigned long hr_nego_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
279 /* Used during o2hb_check_slot to hold a copy of the block
280 * being checked because we temporarily have to zero out the
282 struct o2hb_disk_heartbeat_block *hr_tmp_block;
284 /* Message key for negotiate timeout message. */
286 struct list_head hr_handler_list;
288 /* last hb status, 0 for success, other value for error. */
289 int hr_last_hb_status;
292 struct o2hb_bio_wait_ctxt {
293 atomic_t wc_num_reqs;
294 struct completion wc_io_complete;
298 #define O2HB_NEGO_TIMEOUT_MS (O2HB_MAX_WRITE_TIMEOUT_MS/2)
301 O2HB_NEGO_TIMEOUT_MSG = 1,
302 O2HB_NEGO_APPROVE_MSG = 2,
305 struct o2hb_nego_msg {
309 static void o2hb_write_timeout(struct work_struct *work)
312 struct o2hb_region *reg =
313 container_of(work, struct o2hb_region,
314 hr_write_timeout_work.work);
316 mlog(ML_ERROR, "Heartbeat write timeout to device %s after %u "
317 "milliseconds\n", reg->hr_dev_name,
318 jiffies_to_msecs(jiffies - reg->hr_last_timeout_start));
320 if (o2hb_global_heartbeat_active()) {
321 spin_lock(&o2hb_live_lock);
322 if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
323 set_bit(reg->hr_region_num, o2hb_failed_region_bitmap);
324 failed = bitmap_weight(o2hb_failed_region_bitmap,
326 quorum = bitmap_weight(o2hb_quorum_region_bitmap,
328 spin_unlock(&o2hb_live_lock);
330 mlog(ML_HEARTBEAT, "Number of regions %d, failed regions %d\n",
334 * Fence if the number of failed regions >= half the number
337 if ((failed << 1) < quorum)
341 o2quo_disk_timeout();
344 static void o2hb_arm_timeout(struct o2hb_region *reg)
346 /* Arm writeout only after thread reaches steady state */
347 if (atomic_read(®->hr_steady_iterations) != 0)
350 mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n",
351 O2HB_MAX_WRITE_TIMEOUT_MS);
353 if (o2hb_global_heartbeat_active()) {
354 spin_lock(&o2hb_live_lock);
355 clear_bit(reg->hr_region_num, o2hb_failed_region_bitmap);
356 spin_unlock(&o2hb_live_lock);
358 cancel_delayed_work(®->hr_write_timeout_work);
359 schedule_delayed_work(®->hr_write_timeout_work,
360 msecs_to_jiffies(O2HB_MAX_WRITE_TIMEOUT_MS));
362 cancel_delayed_work(®->hr_nego_timeout_work);
363 /* negotiate timeout must be less than write timeout. */
364 schedule_delayed_work(®->hr_nego_timeout_work,
365 msecs_to_jiffies(O2HB_NEGO_TIMEOUT_MS));
366 memset(reg->hr_nego_node_bitmap, 0, sizeof(reg->hr_nego_node_bitmap));
369 static void o2hb_disarm_timeout(struct o2hb_region *reg)
371 cancel_delayed_work_sync(®->hr_write_timeout_work);
372 cancel_delayed_work_sync(®->hr_nego_timeout_work);
375 static int o2hb_send_nego_msg(int key, int type, u8 target)
377 struct o2hb_nego_msg msg;
380 msg.node_num = o2nm_this_node();
382 ret = o2net_send_message(type, key, &msg, sizeof(msg),
385 if (ret == -EAGAIN || ret == -ENOMEM) {
393 static void o2hb_nego_timeout(struct work_struct *work)
395 unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
396 int master_node, i, ret;
397 struct o2hb_region *reg;
399 reg = container_of(work, struct o2hb_region, hr_nego_timeout_work.work);
400 /* don't negotiate timeout if last hb failed since it is very
401 * possible io failed. Should let write timeout fence self.
403 if (reg->hr_last_hb_status)
406 o2hb_fill_node_map(live_node_bitmap, sizeof(live_node_bitmap));
407 /* lowest node as master node to make negotiate decision. */
408 master_node = find_next_bit(live_node_bitmap, O2NM_MAX_NODES, 0);
410 if (master_node == o2nm_this_node()) {
411 if (!test_bit(master_node, reg->hr_nego_node_bitmap)) {
412 printk(KERN_NOTICE "o2hb: node %d hb write hung for %ds on region %s (%s).\n",
413 o2nm_this_node(), O2HB_NEGO_TIMEOUT_MS/1000,
414 config_item_name(®->hr_item), reg->hr_dev_name);
415 set_bit(master_node, reg->hr_nego_node_bitmap);
417 if (memcmp(reg->hr_nego_node_bitmap, live_node_bitmap,
418 sizeof(reg->hr_nego_node_bitmap))) {
419 /* check negotiate bitmap every second to do timeout
422 schedule_delayed_work(®->hr_nego_timeout_work,
423 msecs_to_jiffies(1000));
428 printk(KERN_NOTICE "o2hb: all nodes hb write hung, maybe region %s (%s) is down.\n",
429 config_item_name(®->hr_item), reg->hr_dev_name);
430 /* approve negotiate timeout request. */
431 o2hb_arm_timeout(reg);
434 while ((i = find_next_bit(live_node_bitmap,
435 O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
436 if (i == master_node)
439 mlog(ML_HEARTBEAT, "send NEGO_APPROVE msg to node %d\n", i);
440 ret = o2hb_send_nego_msg(reg->hr_key,
441 O2HB_NEGO_APPROVE_MSG, i);
443 mlog(ML_ERROR, "send NEGO_APPROVE msg to node %d fail %d\n",
447 /* negotiate timeout with master node. */
448 printk(KERN_NOTICE "o2hb: node %d hb write hung for %ds on region %s (%s), negotiate timeout with node %d.\n",
449 o2nm_this_node(), O2HB_NEGO_TIMEOUT_MS/1000, config_item_name(®->hr_item),
450 reg->hr_dev_name, master_node);
451 ret = o2hb_send_nego_msg(reg->hr_key, O2HB_NEGO_TIMEOUT_MSG,
454 mlog(ML_ERROR, "send NEGO_TIMEOUT msg to node %d fail %d\n",
459 static int o2hb_nego_timeout_handler(struct o2net_msg *msg, u32 len, void *data,
462 struct o2hb_region *reg = data;
463 struct o2hb_nego_msg *nego_msg;
465 nego_msg = (struct o2hb_nego_msg *)msg->buf;
466 printk(KERN_NOTICE "o2hb: receive negotiate timeout message from node %d on region %s (%s).\n",
467 nego_msg->node_num, config_item_name(®->hr_item), reg->hr_dev_name);
468 if (nego_msg->node_num < O2NM_MAX_NODES)
469 set_bit(nego_msg->node_num, reg->hr_nego_node_bitmap);
471 mlog(ML_ERROR, "got nego timeout message from bad node.\n");
476 static int o2hb_nego_approve_handler(struct o2net_msg *msg, u32 len, void *data,
479 struct o2hb_region *reg = data;
481 printk(KERN_NOTICE "o2hb: negotiate timeout approved by master node on region %s (%s).\n",
482 config_item_name(®->hr_item), reg->hr_dev_name);
483 o2hb_arm_timeout(reg);
487 static inline void o2hb_bio_wait_init(struct o2hb_bio_wait_ctxt *wc)
489 atomic_set(&wc->wc_num_reqs, 1);
490 init_completion(&wc->wc_io_complete);
494 /* Used in error paths too */
495 static inline void o2hb_bio_wait_dec(struct o2hb_bio_wait_ctxt *wc,
498 /* sadly atomic_sub_and_test() isn't available on all platforms. The
499 * good news is that the fast path only completes one at a time */
501 if (atomic_dec_and_test(&wc->wc_num_reqs)) {
503 complete(&wc->wc_io_complete);
508 static void o2hb_wait_on_io(struct o2hb_region *reg,
509 struct o2hb_bio_wait_ctxt *wc)
511 o2hb_bio_wait_dec(wc, 1);
512 wait_for_completion(&wc->wc_io_complete);
515 static void o2hb_bio_end_io(struct bio *bio)
517 struct o2hb_bio_wait_ctxt *wc = bio->bi_private;
520 mlog(ML_ERROR, "IO Error %d\n", bio->bi_error);
521 wc->wc_error = bio->bi_error;
524 o2hb_bio_wait_dec(wc, 1);
528 /* Setup a Bio to cover I/O against num_slots slots starting at
530 static struct bio *o2hb_setup_one_bio(struct o2hb_region *reg,
531 struct o2hb_bio_wait_ctxt *wc,
532 unsigned int *current_slot,
533 unsigned int max_slots)
535 int len, current_page;
536 unsigned int vec_len, vec_start;
537 unsigned int bits = reg->hr_block_bits;
538 unsigned int spp = reg->hr_slots_per_page;
539 unsigned int cs = *current_slot;
543 /* Testing has shown this allocation to take long enough under
544 * GFP_KERNEL that the local node can get fenced. It would be
545 * nicest if we could pre-allocate these bios and avoid this
547 bio = bio_alloc(GFP_ATOMIC, 16);
549 mlog(ML_ERROR, "Could not alloc slots BIO!\n");
550 bio = ERR_PTR(-ENOMEM);
554 /* Must put everything in 512 byte sectors for the bio... */
555 bio->bi_iter.bi_sector = (reg->hr_start_block + cs) << (bits - 9);
556 bio->bi_bdev = reg->hr_bdev;
557 bio->bi_private = wc;
558 bio->bi_end_io = o2hb_bio_end_io;
560 vec_start = (cs << bits) % PAGE_SIZE;
561 while(cs < max_slots) {
562 current_page = cs / spp;
563 page = reg->hr_slot_data[current_page];
565 vec_len = min(PAGE_SIZE - vec_start,
566 (max_slots-cs) * (PAGE_SIZE/spp) );
568 mlog(ML_HB_BIO, "page %d, vec_len = %u, vec_start = %u\n",
569 current_page, vec_len, vec_start);
571 len = bio_add_page(bio, page, vec_len, vec_start);
572 if (len != vec_len) break;
574 cs += vec_len / (PAGE_SIZE/spp);
583 static int o2hb_read_slots(struct o2hb_region *reg,
584 unsigned int max_slots)
586 unsigned int current_slot=0;
588 struct o2hb_bio_wait_ctxt wc;
591 o2hb_bio_wait_init(&wc);
593 while(current_slot < max_slots) {
594 bio = o2hb_setup_one_bio(reg, &wc, ¤t_slot, max_slots);
596 status = PTR_ERR(bio);
601 atomic_inc(&wc.wc_num_reqs);
602 submit_bio(READ, bio);
608 o2hb_wait_on_io(reg, &wc);
609 if (wc.wc_error && !status)
610 status = wc.wc_error;
615 static int o2hb_issue_node_write(struct o2hb_region *reg,
616 struct o2hb_bio_wait_ctxt *write_wc)
622 o2hb_bio_wait_init(write_wc);
624 slot = o2nm_this_node();
626 bio = o2hb_setup_one_bio(reg, write_wc, &slot, slot+1);
628 status = PTR_ERR(bio);
633 atomic_inc(&write_wc->wc_num_reqs);
634 submit_bio(WRITE_SYNC, bio);
641 static u32 o2hb_compute_block_crc_le(struct o2hb_region *reg,
642 struct o2hb_disk_heartbeat_block *hb_block)
647 /* We want to compute the block crc with a 0 value in the
648 * hb_cksum field. Save it off here and replace after the
650 old_cksum = hb_block->hb_cksum;
651 hb_block->hb_cksum = 0;
653 ret = crc32_le(0, (unsigned char *) hb_block, reg->hr_block_bytes);
655 hb_block->hb_cksum = old_cksum;
660 static void o2hb_dump_slot(struct o2hb_disk_heartbeat_block *hb_block)
662 mlog(ML_ERROR, "Dump slot information: seq = 0x%llx, node = %u, "
663 "cksum = 0x%x, generation 0x%llx\n",
664 (long long)le64_to_cpu(hb_block->hb_seq),
665 hb_block->hb_node, le32_to_cpu(hb_block->hb_cksum),
666 (long long)le64_to_cpu(hb_block->hb_generation));
669 static int o2hb_verify_crc(struct o2hb_region *reg,
670 struct o2hb_disk_heartbeat_block *hb_block)
674 read = le32_to_cpu(hb_block->hb_cksum);
675 computed = o2hb_compute_block_crc_le(reg, hb_block);
677 return read == computed;
681 * Compare the slot data with what we wrote in the last iteration.
682 * If the match fails, print an appropriate error message. This is to
683 * detect errors like... another node hearting on the same slot,
684 * flaky device that is losing writes, etc.
685 * Returns 1 if check succeeds, 0 otherwise.
687 static int o2hb_check_own_slot(struct o2hb_region *reg)
689 struct o2hb_disk_slot *slot;
690 struct o2hb_disk_heartbeat_block *hb_block;
693 slot = ®->hr_slots[o2nm_this_node()];
694 /* Don't check on our 1st timestamp */
695 if (!slot->ds_last_time)
698 hb_block = slot->ds_raw_block;
699 if (le64_to_cpu(hb_block->hb_seq) == slot->ds_last_time &&
700 le64_to_cpu(hb_block->hb_generation) == slot->ds_last_generation &&
701 hb_block->hb_node == slot->ds_node_num)
704 #define ERRSTR1 "Another node is heartbeating on device"
705 #define ERRSTR2 "Heartbeat generation mismatch on device"
706 #define ERRSTR3 "Heartbeat sequence mismatch on device"
708 if (hb_block->hb_node != slot->ds_node_num)
710 else if (le64_to_cpu(hb_block->hb_generation) !=
711 slot->ds_last_generation)
716 mlog(ML_ERROR, "%s (%s): expected(%u:0x%llx, 0x%llx), "
717 "ondisk(%u:0x%llx, 0x%llx)\n", errstr, reg->hr_dev_name,
718 slot->ds_node_num, (unsigned long long)slot->ds_last_generation,
719 (unsigned long long)slot->ds_last_time, hb_block->hb_node,
720 (unsigned long long)le64_to_cpu(hb_block->hb_generation),
721 (unsigned long long)le64_to_cpu(hb_block->hb_seq));
726 static inline void o2hb_prepare_block(struct o2hb_region *reg,
731 struct o2hb_disk_slot *slot;
732 struct o2hb_disk_heartbeat_block *hb_block;
734 node_num = o2nm_this_node();
735 slot = ®->hr_slots[node_num];
737 hb_block = (struct o2hb_disk_heartbeat_block *)slot->ds_raw_block;
738 memset(hb_block, 0, reg->hr_block_bytes);
739 /* TODO: time stuff */
740 cputime = CURRENT_TIME.tv_sec;
744 hb_block->hb_seq = cpu_to_le64(cputime);
745 hb_block->hb_node = node_num;
746 hb_block->hb_generation = cpu_to_le64(generation);
747 hb_block->hb_dead_ms = cpu_to_le32(o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS);
749 /* This step must always happen last! */
750 hb_block->hb_cksum = cpu_to_le32(o2hb_compute_block_crc_le(reg,
753 mlog(ML_HB_BIO, "our node generation = 0x%llx, cksum = 0x%x\n",
754 (long long)generation,
755 le32_to_cpu(hb_block->hb_cksum));
758 static void o2hb_fire_callbacks(struct o2hb_callback *hbcall,
759 struct o2nm_node *node,
762 struct o2hb_callback_func *f;
764 list_for_each_entry(f, &hbcall->list, hc_item) {
765 mlog(ML_HEARTBEAT, "calling funcs %p\n", f);
766 (f->hc_func)(node, idx, f->hc_data);
770 /* Will run the list in order until we process the passed event */
771 static void o2hb_run_event_list(struct o2hb_node_event *queued_event)
773 struct o2hb_callback *hbcall;
774 struct o2hb_node_event *event;
776 /* Holding callback sem assures we don't alter the callback
777 * lists when doing this, and serializes ourselves with other
778 * processes wanting callbacks. */
779 down_write(&o2hb_callback_sem);
781 spin_lock(&o2hb_live_lock);
782 while (!list_empty(&o2hb_node_events)
783 && !list_empty(&queued_event->hn_item)) {
784 event = list_entry(o2hb_node_events.next,
785 struct o2hb_node_event,
787 list_del_init(&event->hn_item);
788 spin_unlock(&o2hb_live_lock);
790 mlog(ML_HEARTBEAT, "Node %s event for %d\n",
791 event->hn_event_type == O2HB_NODE_UP_CB ? "UP" : "DOWN",
794 hbcall = hbcall_from_type(event->hn_event_type);
796 /* We should *never* have gotten on to the list with a
797 * bad type... This isn't something that we should try
798 * to recover from. */
799 BUG_ON(IS_ERR(hbcall));
801 o2hb_fire_callbacks(hbcall, event->hn_node, event->hn_node_num);
803 spin_lock(&o2hb_live_lock);
805 spin_unlock(&o2hb_live_lock);
807 up_write(&o2hb_callback_sem);
810 static void o2hb_queue_node_event(struct o2hb_node_event *event,
811 enum o2hb_callback_type type,
812 struct o2nm_node *node,
815 assert_spin_locked(&o2hb_live_lock);
817 BUG_ON((!node) && (type != O2HB_NODE_DOWN_CB));
819 event->hn_event_type = type;
820 event->hn_node = node;
821 event->hn_node_num = node_num;
823 mlog(ML_HEARTBEAT, "Queue node %s event for node %d\n",
824 type == O2HB_NODE_UP_CB ? "UP" : "DOWN", node_num);
826 list_add_tail(&event->hn_item, &o2hb_node_events);
829 static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot)
831 struct o2hb_node_event event =
832 { .hn_item = LIST_HEAD_INIT(event.hn_item), };
833 struct o2nm_node *node;
836 node = o2nm_get_node_by_num(slot->ds_node_num);
840 spin_lock(&o2hb_live_lock);
841 if (!list_empty(&slot->ds_live_item)) {
842 mlog(ML_HEARTBEAT, "Shutdown, node %d leaves region\n",
845 list_del_init(&slot->ds_live_item);
847 if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
848 clear_bit(slot->ds_node_num, o2hb_live_node_bitmap);
850 o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB, node,
855 spin_unlock(&o2hb_live_lock);
858 o2hb_run_event_list(&event);
863 static void o2hb_set_quorum_device(struct o2hb_region *reg)
865 if (!o2hb_global_heartbeat_active())
868 /* Prevent race with o2hb_heartbeat_group_drop_item() */
869 if (kthread_should_stop())
872 /* Tag region as quorum only after thread reaches steady state */
873 if (atomic_read(®->hr_steady_iterations) != 0)
876 spin_lock(&o2hb_live_lock);
878 if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
882 * A region can be added to the quorum only when it sees all
883 * live nodes heartbeat on it. In other words, the region has been
884 * added to all nodes.
886 if (memcmp(reg->hr_live_node_bitmap, o2hb_live_node_bitmap,
887 sizeof(o2hb_live_node_bitmap)))
890 printk(KERN_NOTICE "o2hb: Region %s (%s) is now a quorum device\n",
891 config_item_name(®->hr_item), reg->hr_dev_name);
893 set_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
896 * If global heartbeat active, unpin all regions if the
897 * region count > CUT_OFF
899 if (bitmap_weight(o2hb_quorum_region_bitmap,
900 O2NM_MAX_REGIONS) > O2HB_PIN_CUT_OFF)
901 o2hb_region_unpin(NULL);
903 spin_unlock(&o2hb_live_lock);
906 static int o2hb_check_slot(struct o2hb_region *reg,
907 struct o2hb_disk_slot *slot)
909 int changed = 0, gen_changed = 0;
910 struct o2hb_node_event event =
911 { .hn_item = LIST_HEAD_INIT(event.hn_item), };
912 struct o2nm_node *node;
913 struct o2hb_disk_heartbeat_block *hb_block = reg->hr_tmp_block;
915 unsigned int dead_ms = o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS;
916 unsigned int slot_dead_ms;
920 memcpy(hb_block, slot->ds_raw_block, reg->hr_block_bytes);
923 * If a node is no longer configured but is still in the livemap, we
924 * may need to clear that bit from the livemap.
926 node = o2nm_get_node_by_num(slot->ds_node_num);
928 spin_lock(&o2hb_live_lock);
929 tmp = test_bit(slot->ds_node_num, o2hb_live_node_bitmap);
930 spin_unlock(&o2hb_live_lock);
935 if (!o2hb_verify_crc(reg, hb_block)) {
936 /* all paths from here will drop o2hb_live_lock for
938 spin_lock(&o2hb_live_lock);
940 /* Don't print an error on the console in this case -
941 * a freshly formatted heartbeat area will not have a
943 if (list_empty(&slot->ds_live_item))
946 /* The node is live but pushed out a bad crc. We
947 * consider it a transient miss but don't populate any
948 * other values as they may be junk. */
949 mlog(ML_ERROR, "Node %d has written a bad crc to %s\n",
950 slot->ds_node_num, reg->hr_dev_name);
951 o2hb_dump_slot(hb_block);
953 slot->ds_equal_samples++;
957 /* we don't care if these wrap.. the state transitions below
958 * clear at the right places */
959 cputime = le64_to_cpu(hb_block->hb_seq);
960 if (slot->ds_last_time != cputime)
961 slot->ds_changed_samples++;
963 slot->ds_equal_samples++;
964 slot->ds_last_time = cputime;
966 /* The node changed heartbeat generations. We assume this to
967 * mean it dropped off but came back before we timed out. We
968 * want to consider it down for the time being but don't want
969 * to lose any changed_samples state we might build up to
970 * considering it live again. */
971 if (slot->ds_last_generation != le64_to_cpu(hb_block->hb_generation)) {
973 slot->ds_equal_samples = 0;
974 mlog(ML_HEARTBEAT, "Node %d changed generation (0x%llx "
975 "to 0x%llx)\n", slot->ds_node_num,
976 (long long)slot->ds_last_generation,
977 (long long)le64_to_cpu(hb_block->hb_generation));
980 slot->ds_last_generation = le64_to_cpu(hb_block->hb_generation);
982 mlog(ML_HEARTBEAT, "Slot %d gen 0x%llx cksum 0x%x "
983 "seq %llu last %llu changed %u equal %u\n",
984 slot->ds_node_num, (long long)slot->ds_last_generation,
985 le32_to_cpu(hb_block->hb_cksum),
986 (unsigned long long)le64_to_cpu(hb_block->hb_seq),
987 (unsigned long long)slot->ds_last_time, slot->ds_changed_samples,
988 slot->ds_equal_samples);
990 spin_lock(&o2hb_live_lock);
993 /* dead nodes only come to life after some number of
994 * changes at any time during their dead time */
995 if (list_empty(&slot->ds_live_item) &&
996 slot->ds_changed_samples >= O2HB_LIVE_THRESHOLD) {
997 mlog(ML_HEARTBEAT, "Node %d (id 0x%llx) joined my region\n",
998 slot->ds_node_num, (long long)slot->ds_last_generation);
1000 set_bit(slot->ds_node_num, reg->hr_live_node_bitmap);
1002 /* first on the list generates a callback */
1003 if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
1004 mlog(ML_HEARTBEAT, "o2hb: Add node %d to live nodes "
1005 "bitmap\n", slot->ds_node_num);
1006 set_bit(slot->ds_node_num, o2hb_live_node_bitmap);
1008 o2hb_queue_node_event(&event, O2HB_NODE_UP_CB, node,
1015 list_add_tail(&slot->ds_live_item,
1016 &o2hb_live_slots[slot->ds_node_num]);
1018 slot->ds_equal_samples = 0;
1020 /* We want to be sure that all nodes agree on the
1021 * number of milliseconds before a node will be
1022 * considered dead. The self-fencing timeout is
1023 * computed from this value, and a discrepancy might
1024 * result in heartbeat calling a node dead when it
1025 * hasn't self-fenced yet. */
1026 slot_dead_ms = le32_to_cpu(hb_block->hb_dead_ms);
1027 if (slot_dead_ms && slot_dead_ms != dead_ms) {
1028 /* TODO: Perhaps we can fail the region here. */
1029 mlog(ML_ERROR, "Node %d on device %s has a dead count "
1030 "of %u ms, but our count is %u ms.\n"
1031 "Please double check your configuration values "
1032 "for 'O2CB_HEARTBEAT_THRESHOLD'\n",
1033 slot->ds_node_num, reg->hr_dev_name, slot_dead_ms,
1039 /* if the list is dead, we're done.. */
1040 if (list_empty(&slot->ds_live_item))
1043 /* live nodes only go dead after enough consequtive missed
1044 * samples.. reset the missed counter whenever we see
1046 if (slot->ds_equal_samples >= o2hb_dead_threshold || gen_changed) {
1047 mlog(ML_HEARTBEAT, "Node %d left my region\n",
1050 clear_bit(slot->ds_node_num, reg->hr_live_node_bitmap);
1052 /* last off the live_slot generates a callback */
1053 list_del_init(&slot->ds_live_item);
1054 if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
1055 mlog(ML_HEARTBEAT, "o2hb: Remove node %d from live "
1056 "nodes bitmap\n", slot->ds_node_num);
1057 clear_bit(slot->ds_node_num, o2hb_live_node_bitmap);
1059 /* node can be null */
1060 o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB,
1061 node, slot->ds_node_num);
1067 /* We don't clear this because the node is still
1068 * actually writing new blocks. */
1070 slot->ds_changed_samples = 0;
1073 if (slot->ds_changed_samples) {
1074 slot->ds_changed_samples = 0;
1075 slot->ds_equal_samples = 0;
1078 spin_unlock(&o2hb_live_lock);
1081 o2hb_run_event_list(&event);
1084 o2nm_node_put(node);
1088 static int o2hb_highest_node(unsigned long *nodes, int numbits)
1090 return find_last_bit(nodes, numbits);
1093 static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
1095 int i, ret, highest_node;
1096 int membership_change = 0, own_slot_ok = 0;
1097 unsigned long configured_nodes[BITS_TO_LONGS(O2NM_MAX_NODES)];
1098 unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
1099 struct o2hb_bio_wait_ctxt write_wc;
1101 ret = o2nm_configured_node_map(configured_nodes,
1102 sizeof(configured_nodes));
1109 * If a node is not configured but is in the livemap, we still need
1110 * to read the slot so as to be able to remove it from the livemap.
1112 o2hb_fill_node_map(live_node_bitmap, sizeof(live_node_bitmap));
1114 while ((i = find_next_bit(live_node_bitmap,
1115 O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
1116 set_bit(i, configured_nodes);
1119 highest_node = o2hb_highest_node(configured_nodes, O2NM_MAX_NODES);
1120 if (highest_node >= O2NM_MAX_NODES) {
1121 mlog(ML_NOTICE, "o2hb: No configured nodes found!\n");
1126 /* No sense in reading the slots of nodes that don't exist
1127 * yet. Of course, if the node definitions have holes in them
1128 * then we're reading an empty slot anyway... Consider this
1130 ret = o2hb_read_slots(reg, highest_node + 1);
1136 /* With an up to date view of the slots, we can check that no
1137 * other node has been improperly configured to heartbeat in
1139 own_slot_ok = o2hb_check_own_slot(reg);
1141 /* fill in the proper info for our next heartbeat */
1142 o2hb_prepare_block(reg, reg->hr_generation);
1144 ret = o2hb_issue_node_write(reg, &write_wc);
1151 while((i = find_next_bit(configured_nodes,
1152 O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
1153 membership_change |= o2hb_check_slot(reg, ®->hr_slots[i]);
1157 * We have to be sure we've advertised ourselves on disk
1158 * before we can go to steady state. This ensures that
1159 * people we find in our steady state have seen us.
1161 o2hb_wait_on_io(reg, &write_wc);
1162 if (write_wc.wc_error) {
1163 /* Do not re-arm the write timeout on I/O error - we
1164 * can't be sure that the new block ever made it to
1166 mlog(ML_ERROR, "Write error %d on device \"%s\"\n",
1167 write_wc.wc_error, reg->hr_dev_name);
1168 ret = write_wc.wc_error;
1172 /* Skip disarming the timeout if own slot has stale/bad data */
1174 o2hb_set_quorum_device(reg);
1175 o2hb_arm_timeout(reg);
1176 reg->hr_last_timeout_start = jiffies;
1180 /* let the person who launched us know when things are steady */
1181 if (atomic_read(®->hr_steady_iterations) != 0) {
1182 if (!ret && own_slot_ok && !membership_change) {
1183 if (atomic_dec_and_test(®->hr_steady_iterations))
1184 wake_up(&o2hb_steady_queue);
1188 if (atomic_read(®->hr_steady_iterations) != 0) {
1189 if (atomic_dec_and_test(®->hr_unsteady_iterations)) {
1190 printk(KERN_NOTICE "o2hb: Unable to stabilize "
1191 "heartbeart on region %s (%s)\n",
1192 config_item_name(®->hr_item),
1194 atomic_set(®->hr_steady_iterations, 0);
1195 reg->hr_aborted_start = 1;
1196 wake_up(&o2hb_steady_queue);
1205 * we ride the region ref that the region dir holds. before the region
1206 * dir is removed and drops it ref it will wait to tear down this
1209 static int o2hb_thread(void *data)
1212 struct o2hb_region *reg = data;
1213 struct o2hb_bio_wait_ctxt write_wc;
1214 ktime_t before_hb, after_hb;
1215 unsigned int elapsed_msec;
1217 mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread running\n");
1219 set_user_nice(current, MIN_NICE);
1222 ret = o2nm_depend_this_node();
1224 mlog(ML_ERROR, "Node has been deleted, ret = %d\n", ret);
1225 reg->hr_node_deleted = 1;
1226 wake_up(&o2hb_steady_queue);
1230 while (!kthread_should_stop() &&
1231 !reg->hr_unclean_stop && !reg->hr_aborted_start) {
1232 /* We track the time spent inside
1233 * o2hb_do_disk_heartbeat so that we avoid more than
1234 * hr_timeout_ms between disk writes. On busy systems
1235 * this should result in a heartbeat which is less
1236 * likely to time itself out. */
1237 before_hb = ktime_get_real();
1239 ret = o2hb_do_disk_heartbeat(reg);
1240 reg->hr_last_hb_status = ret;
1242 after_hb = ktime_get_real();
1244 elapsed_msec = (unsigned int)
1245 ktime_ms_delta(after_hb, before_hb);
1248 "start = %lld, end = %lld, msec = %u, ret = %d\n",
1249 before_hb.tv64, after_hb.tv64, elapsed_msec, ret);
1251 if (!kthread_should_stop() &&
1252 elapsed_msec < reg->hr_timeout_ms) {
1253 /* the kthread api has blocked signals for us so no
1254 * need to record the return value. */
1255 msleep_interruptible(reg->hr_timeout_ms - elapsed_msec);
1259 o2hb_disarm_timeout(reg);
1261 /* unclean stop is only used in very bad situation */
1262 for(i = 0; !reg->hr_unclean_stop && i < reg->hr_blocks; i++)
1263 o2hb_shutdown_slot(®->hr_slots[i]);
1265 /* Explicit down notification - avoid forcing the other nodes
1266 * to timeout on this region when we could just as easily
1267 * write a clear generation - thus indicating to them that
1268 * this node has left this region.
1270 if (!reg->hr_unclean_stop && !reg->hr_aborted_start) {
1271 o2hb_prepare_block(reg, 0);
1272 ret = o2hb_issue_node_write(reg, &write_wc);
1274 o2hb_wait_on_io(reg, &write_wc);
1280 o2nm_undepend_this_node();
1282 mlog(ML_HEARTBEAT|ML_KTHREAD, "o2hb thread exiting\n");
1287 #ifdef CONFIG_DEBUG_FS
1288 static int o2hb_debug_open(struct inode *inode, struct file *file)
1290 struct o2hb_debug_buf *db = inode->i_private;
1291 struct o2hb_region *reg;
1292 unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)];
1298 /* max_nodes should be the largest bitmap we pass here */
1299 BUG_ON(sizeof(map) < db->db_size);
1301 buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
1305 switch (db->db_type) {
1306 case O2HB_DB_TYPE_LIVENODES:
1307 case O2HB_DB_TYPE_LIVEREGIONS:
1308 case O2HB_DB_TYPE_QUORUMREGIONS:
1309 case O2HB_DB_TYPE_FAILEDREGIONS:
1310 spin_lock(&o2hb_live_lock);
1311 memcpy(map, db->db_data, db->db_size);
1312 spin_unlock(&o2hb_live_lock);
1315 case O2HB_DB_TYPE_REGION_LIVENODES:
1316 spin_lock(&o2hb_live_lock);
1317 reg = (struct o2hb_region *)db->db_data;
1318 memcpy(map, reg->hr_live_node_bitmap, db->db_size);
1319 spin_unlock(&o2hb_live_lock);
1322 case O2HB_DB_TYPE_REGION_NUMBER:
1323 reg = (struct o2hb_region *)db->db_data;
1324 out += snprintf(buf + out, PAGE_SIZE - out, "%d\n",
1325 reg->hr_region_num);
1328 case O2HB_DB_TYPE_REGION_ELAPSED_TIME:
1329 reg = (struct o2hb_region *)db->db_data;
1330 lts = reg->hr_last_timeout_start;
1331 /* If 0, it has never been set before */
1333 lts = jiffies_to_msecs(jiffies - lts);
1334 out += snprintf(buf + out, PAGE_SIZE - out, "%lu\n", lts);
1337 case O2HB_DB_TYPE_REGION_PINNED:
1338 reg = (struct o2hb_region *)db->db_data;
1339 out += snprintf(buf + out, PAGE_SIZE - out, "%u\n",
1340 !!reg->hr_item_pinned);
1347 while ((i = find_next_bit(map, db->db_len, i + 1)) < db->db_len)
1348 out += snprintf(buf + out, PAGE_SIZE - out, "%d ", i);
1349 out += snprintf(buf + out, PAGE_SIZE - out, "\n");
1352 i_size_write(inode, out);
1354 file->private_data = buf;
1361 static int o2hb_debug_release(struct inode *inode, struct file *file)
1363 kfree(file->private_data);
1367 static ssize_t o2hb_debug_read(struct file *file, char __user *buf,
1368 size_t nbytes, loff_t *ppos)
1370 return simple_read_from_buffer(buf, nbytes, ppos, file->private_data,
1371 i_size_read(file->f_mapping->host));
1374 static int o2hb_debug_open(struct inode *inode, struct file *file)
1378 static int o2hb_debug_release(struct inode *inode, struct file *file)
1382 static ssize_t o2hb_debug_read(struct file *file, char __user *buf,
1383 size_t nbytes, loff_t *ppos)
1387 #endif /* CONFIG_DEBUG_FS */
1389 static const struct file_operations o2hb_debug_fops = {
1390 .open = o2hb_debug_open,
1391 .release = o2hb_debug_release,
1392 .read = o2hb_debug_read,
1393 .llseek = generic_file_llseek,
1396 void o2hb_exit(void)
1398 debugfs_remove(o2hb_debug_failedregions);
1399 debugfs_remove(o2hb_debug_quorumregions);
1400 debugfs_remove(o2hb_debug_liveregions);
1401 debugfs_remove(o2hb_debug_livenodes);
1402 debugfs_remove(o2hb_debug_dir);
1403 kfree(o2hb_db_livenodes);
1404 kfree(o2hb_db_liveregions);
1405 kfree(o2hb_db_quorumregions);
1406 kfree(o2hb_db_failedregions);
1409 static struct dentry *o2hb_debug_create(const char *name, struct dentry *dir,
1410 struct o2hb_debug_buf **db, int db_len,
1411 int type, int size, int len, void *data)
1413 *db = kmalloc(db_len, GFP_KERNEL);
1417 (*db)->db_type = type;
1418 (*db)->db_size = size;
1419 (*db)->db_len = len;
1420 (*db)->db_data = data;
1422 return debugfs_create_file(name, S_IFREG|S_IRUSR, dir, *db,
1426 static int o2hb_debug_init(void)
1430 o2hb_debug_dir = debugfs_create_dir(O2HB_DEBUG_DIR, NULL);
1431 if (!o2hb_debug_dir) {
1436 o2hb_debug_livenodes = o2hb_debug_create(O2HB_DEBUG_LIVENODES,
1439 sizeof(*o2hb_db_livenodes),
1440 O2HB_DB_TYPE_LIVENODES,
1441 sizeof(o2hb_live_node_bitmap),
1443 o2hb_live_node_bitmap);
1444 if (!o2hb_debug_livenodes) {
1449 o2hb_debug_liveregions = o2hb_debug_create(O2HB_DEBUG_LIVEREGIONS,
1451 &o2hb_db_liveregions,
1452 sizeof(*o2hb_db_liveregions),
1453 O2HB_DB_TYPE_LIVEREGIONS,
1454 sizeof(o2hb_live_region_bitmap),
1456 o2hb_live_region_bitmap);
1457 if (!o2hb_debug_liveregions) {
1462 o2hb_debug_quorumregions =
1463 o2hb_debug_create(O2HB_DEBUG_QUORUMREGIONS,
1465 &o2hb_db_quorumregions,
1466 sizeof(*o2hb_db_quorumregions),
1467 O2HB_DB_TYPE_QUORUMREGIONS,
1468 sizeof(o2hb_quorum_region_bitmap),
1470 o2hb_quorum_region_bitmap);
1471 if (!o2hb_debug_quorumregions) {
1476 o2hb_debug_failedregions =
1477 o2hb_debug_create(O2HB_DEBUG_FAILEDREGIONS,
1479 &o2hb_db_failedregions,
1480 sizeof(*o2hb_db_failedregions),
1481 O2HB_DB_TYPE_FAILEDREGIONS,
1482 sizeof(o2hb_failed_region_bitmap),
1484 o2hb_failed_region_bitmap);
1485 if (!o2hb_debug_failedregions) {
1502 for (i = 0; i < ARRAY_SIZE(o2hb_callbacks); i++)
1503 INIT_LIST_HEAD(&o2hb_callbacks[i].list);
1505 for (i = 0; i < ARRAY_SIZE(o2hb_live_slots); i++)
1506 INIT_LIST_HEAD(&o2hb_live_slots[i]);
1508 INIT_LIST_HEAD(&o2hb_node_events);
1510 memset(o2hb_live_node_bitmap, 0, sizeof(o2hb_live_node_bitmap));
1511 memset(o2hb_region_bitmap, 0, sizeof(o2hb_region_bitmap));
1512 memset(o2hb_live_region_bitmap, 0, sizeof(o2hb_live_region_bitmap));
1513 memset(o2hb_quorum_region_bitmap, 0, sizeof(o2hb_quorum_region_bitmap));
1514 memset(o2hb_failed_region_bitmap, 0, sizeof(o2hb_failed_region_bitmap));
1516 o2hb_dependent_users = 0;
1518 return o2hb_debug_init();
1521 /* if we're already in a callback then we're already serialized by the sem */
1522 static void o2hb_fill_node_map_from_callback(unsigned long *map,
1525 BUG_ON(bytes < (BITS_TO_LONGS(O2NM_MAX_NODES) * sizeof(unsigned long)));
1527 memcpy(map, &o2hb_live_node_bitmap, bytes);
1531 * get a map of all nodes that are heartbeating in any regions
1533 void o2hb_fill_node_map(unsigned long *map, unsigned bytes)
1535 /* callers want to serialize this map and callbacks so that they
1536 * can trust that they don't miss nodes coming to the party */
1537 down_read(&o2hb_callback_sem);
1538 spin_lock(&o2hb_live_lock);
1539 o2hb_fill_node_map_from_callback(map, bytes);
1540 spin_unlock(&o2hb_live_lock);
1541 up_read(&o2hb_callback_sem);
1543 EXPORT_SYMBOL_GPL(o2hb_fill_node_map);
1546 * heartbeat configfs bits. The heartbeat set is a default set under
1547 * the cluster set in nodemanager.c.
1550 static struct o2hb_region *to_o2hb_region(struct config_item *item)
1552 return item ? container_of(item, struct o2hb_region, hr_item) : NULL;
1555 /* drop_item only drops its ref after killing the thread, nothing should
1556 * be using the region anymore. this has to clean up any state that
1557 * attributes might have built up. */
1558 static void o2hb_region_release(struct config_item *item)
1562 struct o2hb_region *reg = to_o2hb_region(item);
1564 mlog(ML_HEARTBEAT, "hb region release (%s)\n", reg->hr_dev_name);
1566 kfree(reg->hr_tmp_block);
1568 if (reg->hr_slot_data) {
1569 for (i = 0; i < reg->hr_num_pages; i++) {
1570 page = reg->hr_slot_data[i];
1574 kfree(reg->hr_slot_data);
1578 blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE);
1580 kfree(reg->hr_slots);
1582 debugfs_remove(reg->hr_debug_livenodes);
1583 debugfs_remove(reg->hr_debug_regnum);
1584 debugfs_remove(reg->hr_debug_elapsed_time);
1585 debugfs_remove(reg->hr_debug_pinned);
1586 debugfs_remove(reg->hr_debug_dir);
1587 kfree(reg->hr_db_livenodes);
1588 kfree(reg->hr_db_regnum);
1589 kfree(reg->hr_db_elapsed_time);
1590 kfree(reg->hr_db_pinned);
1592 spin_lock(&o2hb_live_lock);
1593 list_del(®->hr_all_item);
1594 spin_unlock(&o2hb_live_lock);
1596 o2net_unregister_handler_list(®->hr_handler_list);
1600 static int o2hb_read_block_input(struct o2hb_region *reg,
1602 unsigned long *ret_bytes,
1603 unsigned int *ret_bits)
1605 unsigned long bytes;
1606 char *p = (char *)page;
1608 bytes = simple_strtoul(p, &p, 0);
1609 if (!p || (*p && (*p != '\n')))
1612 /* Heartbeat and fs min / max block sizes are the same. */
1613 if (bytes > 4096 || bytes < 512)
1615 if (hweight16(bytes) != 1)
1621 *ret_bits = ffs(bytes) - 1;
1626 static ssize_t o2hb_region_block_bytes_show(struct config_item *item,
1629 return sprintf(page, "%u\n", to_o2hb_region(item)->hr_block_bytes);
1632 static ssize_t o2hb_region_block_bytes_store(struct config_item *item,
1636 struct o2hb_region *reg = to_o2hb_region(item);
1638 unsigned long block_bytes;
1639 unsigned int block_bits;
1644 status = o2hb_read_block_input(reg, page, &block_bytes,
1649 reg->hr_block_bytes = (unsigned int)block_bytes;
1650 reg->hr_block_bits = block_bits;
1655 static ssize_t o2hb_region_start_block_show(struct config_item *item,
1658 return sprintf(page, "%llu\n", to_o2hb_region(item)->hr_start_block);
1661 static ssize_t o2hb_region_start_block_store(struct config_item *item,
1665 struct o2hb_region *reg = to_o2hb_region(item);
1666 unsigned long long tmp;
1667 char *p = (char *)page;
1672 tmp = simple_strtoull(p, &p, 0);
1673 if (!p || (*p && (*p != '\n')))
1676 reg->hr_start_block = tmp;
1681 static ssize_t o2hb_region_blocks_show(struct config_item *item, char *page)
1683 return sprintf(page, "%d\n", to_o2hb_region(item)->hr_blocks);
1686 static ssize_t o2hb_region_blocks_store(struct config_item *item,
1690 struct o2hb_region *reg = to_o2hb_region(item);
1692 char *p = (char *)page;
1697 tmp = simple_strtoul(p, &p, 0);
1698 if (!p || (*p && (*p != '\n')))
1701 if (tmp > O2NM_MAX_NODES || tmp == 0)
1704 reg->hr_blocks = (unsigned int)tmp;
1709 static ssize_t o2hb_region_dev_show(struct config_item *item, char *page)
1711 unsigned int ret = 0;
1713 if (to_o2hb_region(item)->hr_bdev)
1714 ret = sprintf(page, "%s\n", to_o2hb_region(item)->hr_dev_name);
1719 static void o2hb_init_region_params(struct o2hb_region *reg)
1721 reg->hr_slots_per_page = PAGE_SIZE >> reg->hr_block_bits;
1722 reg->hr_timeout_ms = O2HB_REGION_TIMEOUT_MS;
1724 mlog(ML_HEARTBEAT, "hr_start_block = %llu, hr_blocks = %u\n",
1725 reg->hr_start_block, reg->hr_blocks);
1726 mlog(ML_HEARTBEAT, "hr_block_bytes = %u, hr_block_bits = %u\n",
1727 reg->hr_block_bytes, reg->hr_block_bits);
1728 mlog(ML_HEARTBEAT, "hr_timeout_ms = %u\n", reg->hr_timeout_ms);
1729 mlog(ML_HEARTBEAT, "dead threshold = %u\n", o2hb_dead_threshold);
1732 static int o2hb_map_slot_data(struct o2hb_region *reg)
1735 unsigned int last_slot;
1736 unsigned int spp = reg->hr_slots_per_page;
1739 struct o2hb_disk_slot *slot;
1741 reg->hr_tmp_block = kmalloc(reg->hr_block_bytes, GFP_KERNEL);
1742 if (reg->hr_tmp_block == NULL)
1745 reg->hr_slots = kcalloc(reg->hr_blocks,
1746 sizeof(struct o2hb_disk_slot), GFP_KERNEL);
1747 if (reg->hr_slots == NULL)
1750 for(i = 0; i < reg->hr_blocks; i++) {
1751 slot = ®->hr_slots[i];
1752 slot->ds_node_num = i;
1753 INIT_LIST_HEAD(&slot->ds_live_item);
1754 slot->ds_raw_block = NULL;
1757 reg->hr_num_pages = (reg->hr_blocks + spp - 1) / spp;
1758 mlog(ML_HEARTBEAT, "Going to require %u pages to cover %u blocks "
1759 "at %u blocks per page\n",
1760 reg->hr_num_pages, reg->hr_blocks, spp);
1762 reg->hr_slot_data = kcalloc(reg->hr_num_pages, sizeof(struct page *),
1764 if (!reg->hr_slot_data)
1767 for(i = 0; i < reg->hr_num_pages; i++) {
1768 page = alloc_page(GFP_KERNEL);
1772 reg->hr_slot_data[i] = page;
1774 last_slot = i * spp;
1775 raw = page_address(page);
1777 (j < spp) && ((j + last_slot) < reg->hr_blocks);
1779 BUG_ON((j + last_slot) >= reg->hr_blocks);
1781 slot = ®->hr_slots[j + last_slot];
1782 slot->ds_raw_block =
1783 (struct o2hb_disk_heartbeat_block *) raw;
1785 raw += reg->hr_block_bytes;
1792 /* Read in all the slots available and populate the tracking
1793 * structures so that we can start with a baseline idea of what's
1795 static int o2hb_populate_slot_data(struct o2hb_region *reg)
1798 struct o2hb_disk_slot *slot;
1799 struct o2hb_disk_heartbeat_block *hb_block;
1801 ret = o2hb_read_slots(reg, reg->hr_blocks);
1805 /* We only want to get an idea of the values initially in each
1806 * slot, so we do no verification - o2hb_check_slot will
1807 * actually determine if each configured slot is valid and
1808 * whether any values have changed. */
1809 for(i = 0; i < reg->hr_blocks; i++) {
1810 slot = ®->hr_slots[i];
1811 hb_block = (struct o2hb_disk_heartbeat_block *) slot->ds_raw_block;
1813 /* Only fill the values that o2hb_check_slot uses to
1814 * determine changing slots */
1815 slot->ds_last_time = le64_to_cpu(hb_block->hb_seq);
1816 slot->ds_last_generation = le64_to_cpu(hb_block->hb_generation);
1823 /* this is acting as commit; we set up all of hr_bdev and hr_task or nothing */
1824 static ssize_t o2hb_region_dev_store(struct config_item *item,
1828 struct o2hb_region *reg = to_o2hb_region(item);
1829 struct task_struct *hb_task;
1832 char *p = (char *)page;
1834 struct inode *inode;
1835 ssize_t ret = -EINVAL;
1841 /* We can't heartbeat without having had our node number
1842 * configured yet. */
1843 if (o2nm_this_node() == O2NM_MAX_NODES)
1846 fd = simple_strtol(p, &p, 0);
1847 if (!p || (*p && (*p != '\n')))
1850 if (fd < 0 || fd >= INT_MAX)
1857 if (reg->hr_blocks == 0 || reg->hr_start_block == 0 ||
1858 reg->hr_block_bytes == 0)
1861 inode = igrab(f.file->f_mapping->host);
1865 if (!S_ISBLK(inode->i_mode))
1868 reg->hr_bdev = I_BDEV(f.file->f_mapping->host);
1869 ret = blkdev_get(reg->hr_bdev, FMODE_WRITE | FMODE_READ, NULL);
1871 reg->hr_bdev = NULL;
1876 bdevname(reg->hr_bdev, reg->hr_dev_name);
1878 sectsize = bdev_logical_block_size(reg->hr_bdev);
1879 if (sectsize != reg->hr_block_bytes) {
1881 "blocksize %u incorrect for device, expected %d",
1882 reg->hr_block_bytes, sectsize);
1887 o2hb_init_region_params(reg);
1889 /* Generation of zero is invalid */
1891 get_random_bytes(®->hr_generation,
1892 sizeof(reg->hr_generation));
1893 } while (reg->hr_generation == 0);
1895 ret = o2hb_map_slot_data(reg);
1901 ret = o2hb_populate_slot_data(reg);
1907 INIT_DELAYED_WORK(®->hr_write_timeout_work, o2hb_write_timeout);
1908 INIT_DELAYED_WORK(®->hr_nego_timeout_work, o2hb_nego_timeout);
1911 * A node is considered live after it has beat LIVE_THRESHOLD
1912 * times. We're not steady until we've given them a chance
1913 * _after_ our first read.
1914 * The default threshold is bare minimum so as to limit the delay
1915 * during mounts. For global heartbeat, the threshold doubled for the
1918 live_threshold = O2HB_LIVE_THRESHOLD;
1919 if (o2hb_global_heartbeat_active()) {
1920 spin_lock(&o2hb_live_lock);
1921 if (bitmap_weight(o2hb_region_bitmap, O2NM_MAX_REGIONS) == 1)
1922 live_threshold <<= 1;
1923 spin_unlock(&o2hb_live_lock);
1926 atomic_set(®->hr_steady_iterations, live_threshold);
1927 /* unsteady_iterations is triple the steady_iterations */
1928 atomic_set(®->hr_unsteady_iterations, (live_threshold * 3));
1930 hb_task = kthread_run(o2hb_thread, reg, "o2hb-%s",
1931 reg->hr_item.ci_name);
1932 if (IS_ERR(hb_task)) {
1933 ret = PTR_ERR(hb_task);
1938 spin_lock(&o2hb_live_lock);
1939 reg->hr_task = hb_task;
1940 spin_unlock(&o2hb_live_lock);
1942 ret = wait_event_interruptible(o2hb_steady_queue,
1943 atomic_read(®->hr_steady_iterations) == 0 ||
1944 reg->hr_node_deleted);
1946 atomic_set(®->hr_steady_iterations, 0);
1947 reg->hr_aborted_start = 1;
1950 if (reg->hr_aborted_start) {
1955 if (reg->hr_node_deleted) {
1960 /* Ok, we were woken. Make sure it wasn't by drop_item() */
1961 spin_lock(&o2hb_live_lock);
1962 hb_task = reg->hr_task;
1963 if (o2hb_global_heartbeat_active())
1964 set_bit(reg->hr_region_num, o2hb_live_region_bitmap);
1965 spin_unlock(&o2hb_live_lock);
1972 if (hb_task && o2hb_global_heartbeat_active())
1973 printk(KERN_NOTICE "o2hb: Heartbeat started on region %s (%s)\n",
1974 config_item_name(®->hr_item), reg->hr_dev_name);
1983 blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE);
1984 reg->hr_bdev = NULL;
1990 static ssize_t o2hb_region_pid_show(struct config_item *item, char *page)
1992 struct o2hb_region *reg = to_o2hb_region(item);
1995 spin_lock(&o2hb_live_lock);
1997 pid = task_pid_nr(reg->hr_task);
1998 spin_unlock(&o2hb_live_lock);
2003 return sprintf(page, "%u\n", pid);
2006 CONFIGFS_ATTR(o2hb_region_, block_bytes);
2007 CONFIGFS_ATTR(o2hb_region_, start_block);
2008 CONFIGFS_ATTR(o2hb_region_, blocks);
2009 CONFIGFS_ATTR(o2hb_region_, dev);
2010 CONFIGFS_ATTR_RO(o2hb_region_, pid);
2012 static struct configfs_attribute *o2hb_region_attrs[] = {
2013 &o2hb_region_attr_block_bytes,
2014 &o2hb_region_attr_start_block,
2015 &o2hb_region_attr_blocks,
2016 &o2hb_region_attr_dev,
2017 &o2hb_region_attr_pid,
2021 static struct configfs_item_operations o2hb_region_item_ops = {
2022 .release = o2hb_region_release,
2025 static struct config_item_type o2hb_region_type = {
2026 .ct_item_ops = &o2hb_region_item_ops,
2027 .ct_attrs = o2hb_region_attrs,
2028 .ct_owner = THIS_MODULE,
2033 struct o2hb_heartbeat_group {
2034 struct config_group hs_group;
2038 static struct o2hb_heartbeat_group *to_o2hb_heartbeat_group(struct config_group *group)
2041 container_of(group, struct o2hb_heartbeat_group, hs_group)
2045 static int o2hb_debug_region_init(struct o2hb_region *reg, struct dentry *dir)
2050 debugfs_create_dir(config_item_name(®->hr_item), dir);
2051 if (!reg->hr_debug_dir) {
2056 reg->hr_debug_livenodes =
2057 o2hb_debug_create(O2HB_DEBUG_LIVENODES,
2059 &(reg->hr_db_livenodes),
2060 sizeof(*(reg->hr_db_livenodes)),
2061 O2HB_DB_TYPE_REGION_LIVENODES,
2062 sizeof(reg->hr_live_node_bitmap),
2063 O2NM_MAX_NODES, reg);
2064 if (!reg->hr_debug_livenodes) {
2069 reg->hr_debug_regnum =
2070 o2hb_debug_create(O2HB_DEBUG_REGION_NUMBER,
2072 &(reg->hr_db_regnum),
2073 sizeof(*(reg->hr_db_regnum)),
2074 O2HB_DB_TYPE_REGION_NUMBER,
2075 0, O2NM_MAX_NODES, reg);
2076 if (!reg->hr_debug_regnum) {
2081 reg->hr_debug_elapsed_time =
2082 o2hb_debug_create(O2HB_DEBUG_REGION_ELAPSED_TIME,
2084 &(reg->hr_db_elapsed_time),
2085 sizeof(*(reg->hr_db_elapsed_time)),
2086 O2HB_DB_TYPE_REGION_ELAPSED_TIME,
2088 if (!reg->hr_debug_elapsed_time) {
2093 reg->hr_debug_pinned =
2094 o2hb_debug_create(O2HB_DEBUG_REGION_PINNED,
2096 &(reg->hr_db_pinned),
2097 sizeof(*(reg->hr_db_pinned)),
2098 O2HB_DB_TYPE_REGION_PINNED,
2100 if (!reg->hr_debug_pinned) {
2110 static struct config_item *o2hb_heartbeat_group_make_item(struct config_group *group,
2113 struct o2hb_region *reg = NULL;
2116 reg = kzalloc(sizeof(struct o2hb_region), GFP_KERNEL);
2118 return ERR_PTR(-ENOMEM);
2120 if (strlen(name) > O2HB_MAX_REGION_NAME_LEN) {
2121 ret = -ENAMETOOLONG;
2125 spin_lock(&o2hb_live_lock);
2126 reg->hr_region_num = 0;
2127 if (o2hb_global_heartbeat_active()) {
2128 reg->hr_region_num = find_first_zero_bit(o2hb_region_bitmap,
2130 if (reg->hr_region_num >= O2NM_MAX_REGIONS) {
2131 spin_unlock(&o2hb_live_lock);
2135 set_bit(reg->hr_region_num, o2hb_region_bitmap);
2137 list_add_tail(®->hr_all_item, &o2hb_all_regions);
2138 spin_unlock(&o2hb_live_lock);
2140 config_item_init_type_name(®->hr_item, name, &o2hb_region_type);
2142 /* this is the same way to generate msg key as dlm, for local heartbeat,
2143 * name is also the same, so make initial crc value different to avoid
2144 * message key conflict.
2146 reg->hr_key = crc32_le(reg->hr_region_num + O2NM_MAX_REGIONS,
2147 name, strlen(name));
2148 INIT_LIST_HEAD(®->hr_handler_list);
2149 ret = o2net_register_handler(O2HB_NEGO_TIMEOUT_MSG, reg->hr_key,
2150 sizeof(struct o2hb_nego_msg),
2151 o2hb_nego_timeout_handler,
2152 reg, NULL, ®->hr_handler_list);
2156 ret = o2net_register_handler(O2HB_NEGO_APPROVE_MSG, reg->hr_key,
2157 sizeof(struct o2hb_nego_msg),
2158 o2hb_nego_approve_handler,
2159 reg, NULL, ®->hr_handler_list);
2161 goto unregister_handler;
2163 ret = o2hb_debug_region_init(reg, o2hb_debug_dir);
2165 config_item_put(®->hr_item);
2166 goto unregister_handler;
2169 return ®->hr_item;
2172 o2net_unregister_handler_list(®->hr_handler_list);
2175 return ERR_PTR(ret);
2178 static void o2hb_heartbeat_group_drop_item(struct config_group *group,
2179 struct config_item *item)
2181 struct task_struct *hb_task;
2182 struct o2hb_region *reg = to_o2hb_region(item);
2183 int quorum_region = 0;
2185 /* stop the thread when the user removes the region dir */
2186 spin_lock(&o2hb_live_lock);
2187 hb_task = reg->hr_task;
2188 reg->hr_task = NULL;
2189 reg->hr_item_dropped = 1;
2190 spin_unlock(&o2hb_live_lock);
2193 kthread_stop(hb_task);
2195 if (o2hb_global_heartbeat_active()) {
2196 spin_lock(&o2hb_live_lock);
2197 clear_bit(reg->hr_region_num, o2hb_region_bitmap);
2198 clear_bit(reg->hr_region_num, o2hb_live_region_bitmap);
2199 if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
2201 clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
2202 spin_unlock(&o2hb_live_lock);
2203 printk(KERN_NOTICE "o2hb: Heartbeat %s on region %s (%s)\n",
2204 ((atomic_read(®->hr_steady_iterations) == 0) ?
2205 "stopped" : "start aborted"), config_item_name(item),
2210 * If we're racing a dev_write(), we need to wake them. They will
2211 * check reg->hr_task
2213 if (atomic_read(®->hr_steady_iterations) != 0) {
2214 reg->hr_aborted_start = 1;
2215 atomic_set(®->hr_steady_iterations, 0);
2216 wake_up(&o2hb_steady_queue);
2219 config_item_put(item);
2221 if (!o2hb_global_heartbeat_active() || !quorum_region)
2225 * If global heartbeat active and there are dependent users,
2226 * pin all regions if quorum region count <= CUT_OFF
2228 spin_lock(&o2hb_live_lock);
2230 if (!o2hb_dependent_users)
2233 if (bitmap_weight(o2hb_quorum_region_bitmap,
2234 O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
2235 o2hb_region_pin(NULL);
2238 spin_unlock(&o2hb_live_lock);
2241 static ssize_t o2hb_heartbeat_group_threshold_show(struct config_item *item,
2244 return sprintf(page, "%u\n", o2hb_dead_threshold);
2247 static ssize_t o2hb_heartbeat_group_threshold_store(struct config_item *item,
2248 const char *page, size_t count)
2251 char *p = (char *)page;
2253 tmp = simple_strtoul(p, &p, 10);
2254 if (!p || (*p && (*p != '\n')))
2257 /* this will validate ranges for us. */
2258 o2hb_dead_threshold_set((unsigned int) tmp);
2263 static ssize_t o2hb_heartbeat_group_mode_show(struct config_item *item,
2266 return sprintf(page, "%s\n",
2267 o2hb_heartbeat_mode_desc[o2hb_heartbeat_mode]);
2270 static ssize_t o2hb_heartbeat_group_mode_store(struct config_item *item,
2271 const char *page, size_t count)
2277 len = (page[count - 1] == '\n') ? count - 1 : count;
2281 for (i = 0; i < O2HB_HEARTBEAT_NUM_MODES; ++i) {
2282 if (strncasecmp(page, o2hb_heartbeat_mode_desc[i], len))
2285 ret = o2hb_global_heartbeat_mode_set(i);
2287 printk(KERN_NOTICE "o2hb: Heartbeat mode set to %s\n",
2288 o2hb_heartbeat_mode_desc[i]);
2296 CONFIGFS_ATTR(o2hb_heartbeat_group_, threshold);
2297 CONFIGFS_ATTR(o2hb_heartbeat_group_, mode);
2299 static struct configfs_attribute *o2hb_heartbeat_group_attrs[] = {
2300 &o2hb_heartbeat_group_attr_threshold,
2301 &o2hb_heartbeat_group_attr_mode,
2305 static struct configfs_group_operations o2hb_heartbeat_group_group_ops = {
2306 .make_item = o2hb_heartbeat_group_make_item,
2307 .drop_item = o2hb_heartbeat_group_drop_item,
2310 static struct config_item_type o2hb_heartbeat_group_type = {
2311 .ct_group_ops = &o2hb_heartbeat_group_group_ops,
2312 .ct_attrs = o2hb_heartbeat_group_attrs,
2313 .ct_owner = THIS_MODULE,
2316 /* this is just here to avoid touching group in heartbeat.h which the
2317 * entire damn world #includes */
2318 struct config_group *o2hb_alloc_hb_set(void)
2320 struct o2hb_heartbeat_group *hs = NULL;
2321 struct config_group *ret = NULL;
2323 hs = kzalloc(sizeof(struct o2hb_heartbeat_group), GFP_KERNEL);
2327 config_group_init_type_name(&hs->hs_group, "heartbeat",
2328 &o2hb_heartbeat_group_type);
2330 ret = &hs->hs_group;
2337 void o2hb_free_hb_set(struct config_group *group)
2339 struct o2hb_heartbeat_group *hs = to_o2hb_heartbeat_group(group);
2343 /* hb callback registration and issuing */
2345 static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type)
2347 if (type == O2HB_NUM_CB)
2348 return ERR_PTR(-EINVAL);
2350 return &o2hb_callbacks[type];
2353 void o2hb_setup_callback(struct o2hb_callback_func *hc,
2354 enum o2hb_callback_type type,
2359 INIT_LIST_HEAD(&hc->hc_item);
2362 hc->hc_priority = priority;
2364 hc->hc_magic = O2HB_CB_MAGIC;
2366 EXPORT_SYMBOL_GPL(o2hb_setup_callback);
2369 * In local heartbeat mode, region_uuid passed matches the dlm domain name.
2370 * In global heartbeat mode, region_uuid passed is NULL.
2372 * In local, we only pin the matching region. In global we pin all the active
2375 static int o2hb_region_pin(const char *region_uuid)
2377 int ret = 0, found = 0;
2378 struct o2hb_region *reg;
2381 assert_spin_locked(&o2hb_live_lock);
2383 list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2384 if (reg->hr_item_dropped)
2387 uuid = config_item_name(®->hr_item);
2389 /* local heartbeat */
2391 if (strcmp(region_uuid, uuid))
2396 if (reg->hr_item_pinned || reg->hr_item_dropped)
2399 /* Ignore ENOENT only for local hb (userdlm domain) */
2400 ret = o2nm_depend_item(®->hr_item);
2402 mlog(ML_CLUSTER, "Pin region %s\n", uuid);
2403 reg->hr_item_pinned = 1;
2405 if (ret == -ENOENT && found)
2408 mlog(ML_ERROR, "Pin region %s fails with %d\n",
2422 * In local heartbeat mode, region_uuid passed matches the dlm domain name.
2423 * In global heartbeat mode, region_uuid passed is NULL.
2425 * In local, we only unpin the matching region. In global we unpin all the
2428 static void o2hb_region_unpin(const char *region_uuid)
2430 struct o2hb_region *reg;
2434 assert_spin_locked(&o2hb_live_lock);
2436 list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2437 if (reg->hr_item_dropped)
2440 uuid = config_item_name(®->hr_item);
2442 if (strcmp(region_uuid, uuid))
2447 if (reg->hr_item_pinned) {
2448 mlog(ML_CLUSTER, "Unpin region %s\n", uuid);
2449 o2nm_undepend_item(®->hr_item);
2450 reg->hr_item_pinned = 0;
2457 static int o2hb_region_inc_user(const char *region_uuid)
2461 spin_lock(&o2hb_live_lock);
2463 /* local heartbeat */
2464 if (!o2hb_global_heartbeat_active()) {
2465 ret = o2hb_region_pin(region_uuid);
2470 * if global heartbeat active and this is the first dependent user,
2471 * pin all regions if quorum region count <= CUT_OFF
2473 o2hb_dependent_users++;
2474 if (o2hb_dependent_users > 1)
2477 if (bitmap_weight(o2hb_quorum_region_bitmap,
2478 O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
2479 ret = o2hb_region_pin(NULL);
2482 spin_unlock(&o2hb_live_lock);
2486 void o2hb_region_dec_user(const char *region_uuid)
2488 spin_lock(&o2hb_live_lock);
2490 /* local heartbeat */
2491 if (!o2hb_global_heartbeat_active()) {
2492 o2hb_region_unpin(region_uuid);
2497 * if global heartbeat active and there are no dependent users,
2498 * unpin all quorum regions
2500 o2hb_dependent_users--;
2501 if (!o2hb_dependent_users)
2502 o2hb_region_unpin(NULL);
2505 spin_unlock(&o2hb_live_lock);
2508 int o2hb_register_callback(const char *region_uuid,
2509 struct o2hb_callback_func *hc)
2511 struct o2hb_callback_func *f;
2512 struct o2hb_callback *hbcall;
2515 BUG_ON(hc->hc_magic != O2HB_CB_MAGIC);
2516 BUG_ON(!list_empty(&hc->hc_item));
2518 hbcall = hbcall_from_type(hc->hc_type);
2519 if (IS_ERR(hbcall)) {
2520 ret = PTR_ERR(hbcall);
2525 ret = o2hb_region_inc_user(region_uuid);
2532 down_write(&o2hb_callback_sem);
2534 list_for_each_entry(f, &hbcall->list, hc_item) {
2535 if (hc->hc_priority < f->hc_priority) {
2536 list_add_tail(&hc->hc_item, &f->hc_item);
2540 if (list_empty(&hc->hc_item))
2541 list_add_tail(&hc->hc_item, &hbcall->list);
2543 up_write(&o2hb_callback_sem);
2546 mlog(ML_CLUSTER, "returning %d on behalf of %p for funcs %p\n",
2547 ret, __builtin_return_address(0), hc);
2550 EXPORT_SYMBOL_GPL(o2hb_register_callback);
2552 void o2hb_unregister_callback(const char *region_uuid,
2553 struct o2hb_callback_func *hc)
2555 BUG_ON(hc->hc_magic != O2HB_CB_MAGIC);
2557 mlog(ML_CLUSTER, "on behalf of %p for funcs %p\n",
2558 __builtin_return_address(0), hc);
2560 /* XXX Can this happen _with_ a region reference? */
2561 if (list_empty(&hc->hc_item))
2565 o2hb_region_dec_user(region_uuid);
2567 down_write(&o2hb_callback_sem);
2569 list_del_init(&hc->hc_item);
2571 up_write(&o2hb_callback_sem);
2573 EXPORT_SYMBOL_GPL(o2hb_unregister_callback);
2575 int o2hb_check_node_heartbeating(u8 node_num)
2577 unsigned long testing_map[BITS_TO_LONGS(O2NM_MAX_NODES)];
2579 o2hb_fill_node_map(testing_map, sizeof(testing_map));
2580 if (!test_bit(node_num, testing_map)) {
2582 "node (%u) does not have heartbeating enabled.\n",
2589 EXPORT_SYMBOL_GPL(o2hb_check_node_heartbeating);
2591 int o2hb_check_node_heartbeating_no_sem(u8 node_num)
2593 unsigned long testing_map[BITS_TO_LONGS(O2NM_MAX_NODES)];
2595 spin_lock(&o2hb_live_lock);
2596 o2hb_fill_node_map_from_callback(testing_map, sizeof(testing_map));
2597 spin_unlock(&o2hb_live_lock);
2598 if (!test_bit(node_num, testing_map)) {
2600 "node (%u) does not have heartbeating enabled.\n",
2607 EXPORT_SYMBOL_GPL(o2hb_check_node_heartbeating_no_sem);
2609 int o2hb_check_node_heartbeating_from_callback(u8 node_num)
2611 unsigned long testing_map[BITS_TO_LONGS(O2NM_MAX_NODES)];
2613 o2hb_fill_node_map_from_callback(testing_map, sizeof(testing_map));
2614 if (!test_bit(node_num, testing_map)) {
2616 "node (%u) does not have heartbeating enabled.\n",
2623 EXPORT_SYMBOL_GPL(o2hb_check_node_heartbeating_from_callback);
2625 /* Makes sure our local node is configured with a node number, and is
2627 int o2hb_check_local_node_heartbeating(void)
2631 /* if this node was set then we have networking */
2632 node_num = o2nm_this_node();
2633 if (node_num == O2NM_MAX_NODES) {
2634 mlog(ML_HEARTBEAT, "this node has not been configured.\n");
2638 return o2hb_check_node_heartbeating(node_num);
2640 EXPORT_SYMBOL_GPL(o2hb_check_local_node_heartbeating);
2643 * this is just a hack until we get the plumbing which flips file systems
2644 * read only and drops the hb ref instead of killing the node dead.
2646 void o2hb_stop_all_regions(void)
2648 struct o2hb_region *reg;
2650 mlog(ML_ERROR, "stopping heartbeat on all active regions.\n");
2652 spin_lock(&o2hb_live_lock);
2654 list_for_each_entry(reg, &o2hb_all_regions, hr_all_item)
2655 reg->hr_unclean_stop = 1;
2657 spin_unlock(&o2hb_live_lock);
2659 EXPORT_SYMBOL_GPL(o2hb_stop_all_regions);
2661 int o2hb_get_all_regions(char *region_uuids, u8 max_regions)
2663 struct o2hb_region *reg;
2667 spin_lock(&o2hb_live_lock);
2670 list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2671 if (reg->hr_item_dropped)
2674 mlog(0, "Region: %s\n", config_item_name(®->hr_item));
2675 if (numregs < max_regions) {
2676 memcpy(p, config_item_name(®->hr_item),
2677 O2HB_MAX_REGION_NAME_LEN);
2678 p += O2HB_MAX_REGION_NAME_LEN;
2683 spin_unlock(&o2hb_live_lock);
2687 EXPORT_SYMBOL_GPL(o2hb_get_all_regions);
2689 int o2hb_global_heartbeat_active(void)
2691 return (o2hb_heartbeat_mode == O2HB_HEARTBEAT_GLOBAL);
2693 EXPORT_SYMBOL(o2hb_global_heartbeat_active);