projects / linux-2.6-block.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
Merge branch 'for-linus' of master.kernel.org:/home/rmk/linux-2.6-arm
[linux-2.6-block.git] / fs / ocfs2 / cluster / heartbeat.c
CommitLineData
a7f6a5fb
MF		 1/* -*- mode: c; c-basic-offset: 8; -*-
2 * vim: noexpandtab sw=8 ts=8 sts=0:
3 *
4 * Copyright (C) 2004, 2005 Oracle. All rights reserved.
5 *
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.
10 *
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.
15 *
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.
20 */
21
22#include <linux/kernel.h>
23#include <linux/sched.h>
24#include <linux/jiffies.h>
25#include <linux/module.h>
26#include <linux/fs.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>
87d3d3f3 36#include <linux/debugfs.h>
5a0e3ad6 37#include <linux/slab.h>
a7f6a5fb
MF		 38
39#include "heartbeat.h"
40#include "tcp.h"
41#include "nodemanager.h"
42#include "quorum.h"
43
44#include "masklog.h"
45
46
47/*
48 * The first heartbeat pass had one global thread that would serialize all hb
49 * callback calls. This global serializing sem should only be removed once
50 * we've made sure that all callees can deal with being called concurrently
51 * from multiple hb region threads.
52 */
53static DECLARE_RWSEM(o2hb_callback_sem);
54
55/*
56 * multiple hb threads are watching multiple regions. A node is live
57 * whenever any of the threads sees activity from the node in its region.
58 */
34af946a 59static DEFINE_SPINLOCK(o2hb_live_lock);
a7f6a5fb
MF		 60static struct list_head o2hb_live_slots[O2NM_MAX_NODES];
61static unsigned long o2hb_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
62static LIST_HEAD(o2hb_node_events);
63static DECLARE_WAIT_QUEUE_HEAD(o2hb_steady_queue);
64
536f0741
SM		 65/*
66 * In global heartbeat, we maintain a series of region bitmaps.
67 * - o2hb_region_bitmap allows us to limit the region number to max region.
e7d656ba 68 * - o2hb_live_region_bitmap tracks live regions (seen steady iterations).
43182d2a
SM		 69 * - o2hb_quorum_region_bitmap tracks live regions that have seen all nodes
70 * heartbeat on it.
b1c5ebfb 71 * - o2hb_failed_region_bitmap tracks the regions that have seen io timeouts.
536f0741
SM		 72 */
73static unsigned long o2hb_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
e7d656ba 74static unsigned long o2hb_live_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
43182d2a 75static unsigned long o2hb_quorum_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
b1c5ebfb 76static unsigned long o2hb_failed_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
536f0741 77
8ca8b0bb 78#define O2HB_DB_TYPE_LIVENODES 0
a6de0136
SM		 79#define O2HB_DB_TYPE_LIVEREGIONS 1
80#define O2HB_DB_TYPE_QUORUMREGIONS 2
81#define O2HB_DB_TYPE_FAILEDREGIONS 3
1f285305
SM		 82#define O2HB_DB_TYPE_REGION_LIVENODES 4
83#define O2HB_DB_TYPE_REGION_NUMBER 5
43695d09 84#define O2HB_DB_TYPE_REGION_ELAPSED_TIME 6
cb0586bd 85#define O2HB_DB_TYPE_REGION_PINNED 7
8ca8b0bb
SM		 86struct o2hb_debug_buf {
87 int db_type;
88 int db_size;
89 int db_len;
90 void *db_data;
91};
92
93static struct o2hb_debug_buf *o2hb_db_livenodes;
a6de0136
SM		 94static struct o2hb_debug_buf *o2hb_db_liveregions;
95static struct o2hb_debug_buf *o2hb_db_quorumregions;
96static struct o2hb_debug_buf *o2hb_db_failedregions;
8ca8b0bb 97
87d3d3f3
SM		 98#define O2HB_DEBUG_DIR "o2hb"
99#define O2HB_DEBUG_LIVENODES "livenodes"
a6de0136
SM		 100#define O2HB_DEBUG_LIVEREGIONS "live_regions"
101#define O2HB_DEBUG_QUORUMREGIONS "quorum_regions"
102#define O2HB_DEBUG_FAILEDREGIONS "failed_regions"
1f285305 103#define O2HB_DEBUG_REGION_NUMBER "num"
43695d09 104#define O2HB_DEBUG_REGION_ELAPSED_TIME "elapsed_time_in_ms"
cb0586bd 105#define O2HB_DEBUG_REGION_PINNED "pinned"
8ca8b0bb 106
87d3d3f3
SM		 107static struct dentry *o2hb_debug_dir;
108static struct dentry *o2hb_debug_livenodes;
a6de0136
SM		 109static struct dentry *o2hb_debug_liveregions;
110static struct dentry *o2hb_debug_quorumregions;
111static struct dentry *o2hb_debug_failedregions;
87d3d3f3 112
a7f6a5fb
MF		 113static LIST_HEAD(o2hb_all_regions);
114
115static struct o2hb_callback {
116 struct list_head list;
117} o2hb_callbacks[O2HB_NUM_CB];
118
119static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type);
120
121#define O2HB_DEFAULT_BLOCK_BITS 9
122
54b5187b
SM		 123enum o2hb_heartbeat_modes {
124 O2HB_HEARTBEAT_LOCAL = 0,
125 O2HB_HEARTBEAT_GLOBAL,
126 O2HB_HEARTBEAT_NUM_MODES,
127};
128
129char *o2hb_heartbeat_mode_desc[O2HB_HEARTBEAT_NUM_MODES] = {
130 "local", /* O2HB_HEARTBEAT_LOCAL */
131 "global", /* O2HB_HEARTBEAT_GLOBAL */
132};
133
a7f6a5fb 134unsigned int o2hb_dead_threshold = O2HB_DEFAULT_DEAD_THRESHOLD;
54b5187b 135unsigned int o2hb_heartbeat_mode = O2HB_HEARTBEAT_LOCAL;
a7f6a5fb 136
58a3158a
SM		 137/*
138 * o2hb_dependent_users tracks the number of registered callbacks that depend
139 * on heartbeat. o2net and o2dlm are two entities that register this callback.
140 * However only o2dlm depends on the heartbeat. It does not want the heartbeat
141 * to stop while a dlm domain is still active.
142 */
143unsigned int o2hb_dependent_users;
144
145/*
146 * In global heartbeat mode, all regions are pinned if there are one or more
147 * dependent users and the quorum region count is <= O2HB_PIN_CUT_OFF. All
148 * regions are unpinned if the region count exceeds the cut off or the number
149 * of dependent users falls to zero.
150 */
151#define O2HB_PIN_CUT_OFF 3
152
153/*
154 * In local heartbeat mode, we assume the dlm domain name to be the same as
155 * region uuid. This is true for domains created for the file system but not
156 * necessarily true for userdlm domains. This is a known limitation.
157 *
158 * In global heartbeat mode, we pin/unpin all o2hb regions. This solution
159 * works for both file system and userdlm domains.
160 */
161static int o2hb_region_pin(const char *region_uuid);
162static void o2hb_region_unpin(const char *region_uuid);
163
2bd63216 164/* Only sets a new threshold if there are no active regions.
a7f6a5fb
MF		 165 *
166 * No locking or otherwise interesting code is required for reading
167 * o2hb_dead_threshold as it can't change once regions are active and
168 * it's not interesting to anyone until then anyway. */
169static void o2hb_dead_threshold_set(unsigned int threshold)
170{
171 if (threshold > O2HB_MIN_DEAD_THRESHOLD) {
172 spin_lock(&o2hb_live_lock);
173 if (list_empty(&o2hb_all_regions))
174 o2hb_dead_threshold = threshold;
175 spin_unlock(&o2hb_live_lock);
176 }
177}
178
54b5187b
SM		 179static int o2hb_global_hearbeat_mode_set(unsigned int hb_mode)
180{
181 int ret = -1;
182
183 if (hb_mode < O2HB_HEARTBEAT_NUM_MODES) {
184 spin_lock(&o2hb_live_lock);
185 if (list_empty(&o2hb_all_regions)) {
186 o2hb_heartbeat_mode = hb_mode;
187 ret = 0;
188 }
189 spin_unlock(&o2hb_live_lock);
190 }
191
192 return ret;
193}
194
a7f6a5fb
MF		 195struct o2hb_node_event {
196 struct list_head hn_item;
197 enum o2hb_callback_type hn_event_type;
198 struct o2nm_node *hn_node;
199 int hn_node_num;
200};
201
202struct o2hb_disk_slot {
203 struct o2hb_disk_heartbeat_block *ds_raw_block;
204 u8 ds_node_num;
205 u64 ds_last_time;
206 u64 ds_last_generation;
207 u16 ds_equal_samples;
208 u16 ds_changed_samples;
209 struct list_head ds_live_item;
210};
211
212/* each thread owns a region.. when we're asked to tear down the region
213 * we ask the thread to stop, who cleans up the region */
214struct o2hb_region {
215 struct config_item hr_item;
216
217 struct list_head hr_all_item;
58a3158a
SM		 218 unsigned hr_unclean_stop:1,
219 hr_item_pinned:1,
220 hr_item_dropped:1;
a7f6a5fb
MF		 221
222 /* protected by the hr_callback_sem */
223 struct task_struct *hr_task;
224
225 unsigned int hr_blocks;
226 unsigned long long hr_start_block;
227
228 unsigned int hr_block_bits;
229 unsigned int hr_block_bytes;
230
231 unsigned int hr_slots_per_page;
232 unsigned int hr_num_pages;
233
234 struct page **hr_slot_data;
235 struct block_device *hr_bdev;
236 struct o2hb_disk_slot *hr_slots;
237
823a637a
SM		 238 /* live node map of this region */
239 unsigned long hr_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
536f0741 240 unsigned int hr_region_num;
823a637a 241
1f285305
SM		 242 struct dentry *hr_debug_dir;
243 struct dentry *hr_debug_livenodes;
244 struct dentry *hr_debug_regnum;
43695d09 245 struct dentry *hr_debug_elapsed_time;
cb0586bd 246 struct dentry *hr_debug_pinned;
1f285305
SM		 247 struct o2hb_debug_buf *hr_db_livenodes;
248 struct o2hb_debug_buf *hr_db_regnum;
43695d09 249 struct o2hb_debug_buf *hr_db_elapsed_time;
cb0586bd 250 struct o2hb_debug_buf *hr_db_pinned;
1f285305 251
a7f6a5fb
MF		 252 /* let the person setting up hb wait for it to return until it
253 * has reached a 'steady' state. This will be fixed when we have
254 * a more complete api that doesn't lead to this sort of fragility. */
255 atomic_t hr_steady_iterations;
256
257 char hr_dev_name[BDEVNAME_SIZE];
258
259 unsigned int hr_timeout_ms;
260
261 /* randomized as the region goes up and down so that a node
262 * recognizes a node going up and down in one iteration */
263 u64 hr_generation;
264
c4028958 265 struct delayed_work hr_write_timeout_work;
a7f6a5fb
MF		 266 unsigned long hr_last_timeout_start;
267
268 /* Used during o2hb_check_slot to hold a copy of the block
269 * being checked because we temporarily have to zero out the
270 * crc field. */
271 struct o2hb_disk_heartbeat_block *hr_tmp_block;
272};
273
274struct o2hb_bio_wait_ctxt {
275 atomic_t wc_num_reqs;
276 struct completion wc_io_complete;
a9e2ae39 277 int wc_error;
a7f6a5fb
MF		 278};
279
b1c5ebfb
SM		 280static int o2hb_pop_count(void *map, int count)
281{
282 int i = -1, pop = 0;
283
284 while ((i = find_next_bit(map, count, i + 1)) < count)
285 pop++;
286 return pop;
287}
288
c4028958 289static void o2hb_write_timeout(struct work_struct *work)
a7f6a5fb 290{
b1c5ebfb
SM		 291 int failed, quorum;
292 unsigned long flags;
c4028958
DH		 293 struct o2hb_region *reg =
294 container_of(work, struct o2hb_region,
295 hr_write_timeout_work.work);
a7f6a5fb
MF		 296
297 mlog(ML_ERROR, "Heartbeat write timeout to device %s after %u "
298 "milliseconds\n", reg->hr_dev_name,
2bd63216 299 jiffies_to_msecs(jiffies - reg->hr_last_timeout_start));
b1c5ebfb
SM		 300
301 if (o2hb_global_heartbeat_active()) {
302 spin_lock_irqsave(&o2hb_live_lock, flags);
303 if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
304 set_bit(reg->hr_region_num, o2hb_failed_region_bitmap);
305 failed = o2hb_pop_count(&o2hb_failed_region_bitmap,
306 O2NM_MAX_REGIONS);
307 quorum = o2hb_pop_count(&o2hb_quorum_region_bitmap,
308 O2NM_MAX_REGIONS);
309 spin_unlock_irqrestore(&o2hb_live_lock, flags);
310
311 mlog(ML_HEARTBEAT, "Number of regions %d, failed regions %d\n",
312 quorum, failed);
313
314 /*
315 * Fence if the number of failed regions >= half the number
316 * of quorum regions
317 */
318 if ((failed << 1) < quorum)
319 return;
320 }
321
a7f6a5fb
MF		 322 o2quo_disk_timeout();
323}
324
325static void o2hb_arm_write_timeout(struct o2hb_region *reg)
326{
b31d308d
TM		 327 mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n",
328 O2HB_MAX_WRITE_TIMEOUT_MS);
a7f6a5fb 329
b1c5ebfb
SM		 330 if (o2hb_global_heartbeat_active()) {
331 spin_lock(&o2hb_live_lock);
332 clear_bit(reg->hr_region_num, o2hb_failed_region_bitmap);
333 spin_unlock(&o2hb_live_lock);
334 }
a7f6a5fb
MF		 335 cancel_delayed_work(&reg->hr_write_timeout_work);
336 reg->hr_last_timeout_start = jiffies;
337 schedule_delayed_work(&reg->hr_write_timeout_work,
338 msecs_to_jiffies(O2HB_MAX_WRITE_TIMEOUT_MS));
339}
340
341static void o2hb_disarm_write_timeout(struct o2hb_region *reg)
342{
9b00a818 343 cancel_delayed_work_sync(&reg->hr_write_timeout_work);
a7f6a5fb
MF		 344}
345
b559292e 346static inline void o2hb_bio_wait_init(struct o2hb_bio_wait_ctxt *wc)
a7f6a5fb 347{
b559292e 348 atomic_set(&wc->wc_num_reqs, 1);
a7f6a5fb 349 init_completion(&wc->wc_io_complete);
a9e2ae39 350 wc->wc_error = 0;
a7f6a5fb
MF		 351}
352
353/* Used in error paths too */
354static inline void o2hb_bio_wait_dec(struct o2hb_bio_wait_ctxt *wc,
355 unsigned int num)
356{
357 /* sadly atomic_sub_and_test() isn't available on all platforms. The
358 * good news is that the fast path only completes one at a time */
359 while(num--) {
360 if (atomic_dec_and_test(&wc->wc_num_reqs)) {
361 BUG_ON(num > 0);
362 complete(&wc->wc_io_complete);
363 }
364 }
365}
366
367static void o2hb_wait_on_io(struct o2hb_region *reg,
368 struct o2hb_bio_wait_ctxt *wc)
369{
b559292e 370 o2hb_bio_wait_dec(wc, 1);
a7f6a5fb
MF		 371 wait_for_completion(&wc->wc_io_complete);
372}
373
782e3b3b 374static void o2hb_bio_end_io(struct bio *bio,
a7f6a5fb
MF		 375 int error)
376{
377 struct o2hb_bio_wait_ctxt *wc = bio->bi_private;
378
a9e2ae39 379 if (error) {
a7f6a5fb 380 mlog(ML_ERROR, "IO Error %d\n", error);
a9e2ae39
MF		 381 wc->wc_error = error;
382 }
a7f6a5fb 383
a7f6a5fb 384 o2hb_bio_wait_dec(wc, 1);
b559292e 385 bio_put(bio);
a7f6a5fb
MF		 386}
387
388/* Setup a Bio to cover I/O against num_slots slots starting at
389 * start_slot. */
390static struct bio *o2hb_setup_one_bio(struct o2hb_region *reg,
391 struct o2hb_bio_wait_ctxt *wc,
b559292e
PR		 392 unsigned int *current_slot,
393 unsigned int max_slots)
a7f6a5fb 394{
b559292e 395 int len, current_page;
a7f6a5fb
MF		 396 unsigned int vec_len, vec_start;
397 unsigned int bits = reg->hr_block_bits;
398 unsigned int spp = reg->hr_slots_per_page;
b559292e 399 unsigned int cs = *current_slot;
a7f6a5fb
MF		 400 struct bio *bio;
401 struct page *page;
402
a7f6a5fb
MF		 403 /* Testing has shown this allocation to take long enough under
404 * GFP_KERNEL that the local node can get fenced. It would be
405 * nicest if we could pre-allocate these bios and avoid this
406 * all together. */
b559292e 407 bio = bio_alloc(GFP_ATOMIC, 16);
a7f6a5fb
MF		 408 if (!bio) {
409 mlog(ML_ERROR, "Could not alloc slots BIO!\n");
410 bio = ERR_PTR(-ENOMEM);
411 goto bail;
412 }
413
414 /* Must put everything in 512 byte sectors for the bio... */
b559292e 415 bio->bi_sector = (reg->hr_start_block + cs) << (bits - 9);
a7f6a5fb
MF		 416 bio->bi_bdev = reg->hr_bdev;
417 bio->bi_private = wc;
418 bio->bi_end_io = o2hb_bio_end_io;
419
b559292e
PR		 420 vec_start = (cs << bits) % PAGE_CACHE_SIZE;
421 while(cs < max_slots) {
422 current_page = cs / spp;
423 page = reg->hr_slot_data[current_page];
a7f6a5fb 424
bc7e97cb 425 vec_len = min(PAGE_CACHE_SIZE - vec_start,
b559292e 426 (max_slots-cs) * (PAGE_CACHE_SIZE/spp) );
a7f6a5fb
MF		 427
428 mlog(ML_HB_BIO, "page %d, vec_len = %u, vec_start = %u\n",
b559292e 429 current_page, vec_len, vec_start);
a7f6a5fb
MF		 430
431 len = bio_add_page(bio, page, vec_len, vec_start);
b559292e 432 if (len != vec_len) break;
a7f6a5fb 433
b559292e 434 cs += vec_len / (PAGE_CACHE_SIZE/spp);
a7f6a5fb
MF		 435 vec_start = 0;
436 }
437
438bail:
b559292e 439 *current_slot = cs;
a7f6a5fb
MF		 440 return bio;
441}
442
a7f6a5fb
MF		 443static int o2hb_read_slots(struct o2hb_region *reg,
444 unsigned int max_slots)
445{
b559292e
PR		 446 unsigned int current_slot=0;
447 int status;
a7f6a5fb 448 struct o2hb_bio_wait_ctxt wc;
a7f6a5fb
MF		 449 struct bio *bio;
450
b559292e 451 o2hb_bio_wait_init(&wc);
a7f6a5fb 452
b559292e
PR		 453 while(current_slot < max_slots) {
454 bio = o2hb_setup_one_bio(reg, &wc, &current_slot, max_slots);
a7f6a5fb 455 if (IS_ERR(bio)) {
a7f6a5fb
MF		 456 status = PTR_ERR(bio);
457 mlog_errno(status);
458 goto bail_and_wait;
459 }
a7f6a5fb 460
b559292e 461 atomic_inc(&wc.wc_num_reqs);
a7f6a5fb
MF		 462 submit_bio(READ, bio);
463 }
464
465 status = 0;
466
467bail_and_wait:
468 o2hb_wait_on_io(reg, &wc);
a9e2ae39
MF		 469 if (wc.wc_error && !status)
470 status = wc.wc_error;
a7f6a5fb 471
a7f6a5fb
MF		 472 return status;
473}
474
475static int o2hb_issue_node_write(struct o2hb_region *reg,
a7f6a5fb
MF		 476 struct o2hb_bio_wait_ctxt *write_wc)
477{
478 int status;
479 unsigned int slot;
480 struct bio *bio;
481
b559292e 482 o2hb_bio_wait_init(write_wc);
a7f6a5fb
MF		 483
484 slot = o2nm_this_node();
485
b559292e 486 bio = o2hb_setup_one_bio(reg, write_wc, &slot, slot+1);
a7f6a5fb
MF		 487 if (IS_ERR(bio)) {
488 status = PTR_ERR(bio);
489 mlog_errno(status);
490 goto bail;
491 }
492
b559292e 493 atomic_inc(&write_wc->wc_num_reqs);
a7f6a5fb
MF		 494 submit_bio(WRITE, bio);
495
a7f6a5fb
MF		 496 status = 0;
497bail:
498 return status;
499}
500
501static u32 o2hb_compute_block_crc_le(struct o2hb_region *reg,
502 struct o2hb_disk_heartbeat_block *hb_block)
503{
504 __le32 old_cksum;
505 u32 ret;
506
507 /* We want to compute the block crc with a 0 value in the
508 * hb_cksum field. Save it off here and replace after the
509 * crc. */
510 old_cksum = hb_block->hb_cksum;
511 hb_block->hb_cksum = 0;
512
513 ret = crc32_le(0, (unsigned char *) hb_block, reg->hr_block_bytes);
514
515 hb_block->hb_cksum = old_cksum;
516
517 return ret;
518}
519
520static void o2hb_dump_slot(struct o2hb_disk_heartbeat_block *hb_block)
521{
70bacbdb
MF		 522 mlog(ML_ERROR, "Dump slot information: seq = 0x%llx, node = %u, "
523 "cksum = 0x%x, generation 0x%llx\n",
524 (long long)le64_to_cpu(hb_block->hb_seq),
525 hb_block->hb_node, le32_to_cpu(hb_block->hb_cksum),
526 (long long)le64_to_cpu(hb_block->hb_generation));
a7f6a5fb
MF		 527}
528
529static int o2hb_verify_crc(struct o2hb_region *reg,
530 struct o2hb_disk_heartbeat_block *hb_block)
531{
532 u32 read, computed;
533
534 read = le32_to_cpu(hb_block->hb_cksum);
535 computed = o2hb_compute_block_crc_le(reg, hb_block);
536
537 return read == computed;
538}
539
540/* We want to make sure that nobody is heartbeating on top of us --
541 * this will help detect an invalid configuration. */
33c12a54 542static void o2hb_check_last_timestamp(struct o2hb_region *reg)
a7f6a5fb 543{
a7f6a5fb
MF		 544 struct o2hb_disk_slot *slot;
545 struct o2hb_disk_heartbeat_block *hb_block;
33c12a54 546 char *errstr;
a7f6a5fb 547
33c12a54 548 slot = &reg->hr_slots[o2nm_this_node()];
a7f6a5fb 549 /* Don't check on our 1st timestamp */
33c12a54
SM		 550 if (!slot->ds_last_time)
551 return;
a7f6a5fb 552
33c12a54
SM		 553 hb_block = slot->ds_raw_block;
554 if (le64_to_cpu(hb_block->hb_seq) == slot->ds_last_time &&
555 le64_to_cpu(hb_block->hb_generation) == slot->ds_last_generation &&
556 hb_block->hb_node == slot->ds_node_num)
557 return;
a7f6a5fb 558
33c12a54
SM		 559#define ERRSTR1 "Another node is heartbeating on device"
560#define ERRSTR2 "Heartbeat generation mismatch on device"
561#define ERRSTR3 "Heartbeat sequence mismatch on device"
562
563 if (hb_block->hb_node != slot->ds_node_num)
564 errstr = ERRSTR1;
565 else if (le64_to_cpu(hb_block->hb_generation) !=
566 slot->ds_last_generation)
567 errstr = ERRSTR2;
568 else
569 errstr = ERRSTR3;
570
571 mlog(ML_ERROR, "%s (%s): expected(%u:0x%llx, 0x%llx), "
572 "ondisk(%u:0x%llx, 0x%llx)\n", errstr, reg->hr_dev_name,
573 slot->ds_node_num, (unsigned long long)slot->ds_last_generation,
574 (unsigned long long)slot->ds_last_time, hb_block->hb_node,
575 (unsigned long long)le64_to_cpu(hb_block->hb_generation),
576 (unsigned long long)le64_to_cpu(hb_block->hb_seq));
a7f6a5fb
MF		 577}
578
579static inline void o2hb_prepare_block(struct o2hb_region *reg,
580 u64 generation)
581{
582 int node_num;
583 u64 cputime;
584 struct o2hb_disk_slot *slot;
585 struct o2hb_disk_heartbeat_block *hb_block;
586
587 node_num = o2nm_this_node();
588 slot = &reg->hr_slots[node_num];
589
590 hb_block = (struct o2hb_disk_heartbeat_block *)slot->ds_raw_block;
591 memset(hb_block, 0, reg->hr_block_bytes);
592 /* TODO: time stuff */
593 cputime = CURRENT_TIME.tv_sec;
594 if (!cputime)
595 cputime = 1;
596
597 hb_block->hb_seq = cpu_to_le64(cputime);
598 hb_block->hb_node = node_num;
599 hb_block->hb_generation = cpu_to_le64(generation);
0db638f4 600 hb_block->hb_dead_ms = cpu_to_le32(o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS);
a7f6a5fb
MF		 601
602 /* This step must always happen last! */
603 hb_block->hb_cksum = cpu_to_le32(o2hb_compute_block_crc_le(reg,
604 hb_block));
605
70bacbdb 606 mlog(ML_HB_BIO, "our node generation = 0x%llx, cksum = 0x%x\n",
5fdf1e67 607 (long long)generation,
70bacbdb 608 le32_to_cpu(hb_block->hb_cksum));
a7f6a5fb
MF		 609}
610
611static void o2hb_fire_callbacks(struct o2hb_callback *hbcall,
612 struct o2nm_node *node,
613 int idx)
614{
615 struct list_head *iter;
616 struct o2hb_callback_func *f;
617
618 list_for_each(iter, &hbcall->list) {
619 f = list_entry(iter, struct o2hb_callback_func, hc_item);
620 mlog(ML_HEARTBEAT, "calling funcs %p\n", f);
621 (f->hc_func)(node, idx, f->hc_data);
622 }
623}
624
625/* Will run the list in order until we process the passed event */
626static void o2hb_run_event_list(struct o2hb_node_event *queued_event)
627{
628 int empty;
629 struct o2hb_callback *hbcall;
630 struct o2hb_node_event *event;
631
632 spin_lock(&o2hb_live_lock);
633 empty = list_empty(&queued_event->hn_item);
634 spin_unlock(&o2hb_live_lock);
635 if (empty)
636 return;
637
638 /* Holding callback sem assures we don't alter the callback
639 * lists when doing this, and serializes ourselves with other
640 * processes wanting callbacks. */
641 down_write(&o2hb_callback_sem);
642
643 spin_lock(&o2hb_live_lock);
644 while (!list_empty(&o2hb_node_events)
645 && !list_empty(&queued_event->hn_item)) {
646 event = list_entry(o2hb_node_events.next,
647 struct o2hb_node_event,
648 hn_item);
649 list_del_init(&event->hn_item);
650 spin_unlock(&o2hb_live_lock);
651
652 mlog(ML_HEARTBEAT, "Node %s event for %d\n",
653 event->hn_event_type == O2HB_NODE_UP_CB ? "UP" : "DOWN",
654 event->hn_node_num);
655
656 hbcall = hbcall_from_type(event->hn_event_type);
657
658 /* We should *never* have gotten on to the list with a
659 * bad type... This isn't something that we should try
660 * to recover from. */
661 BUG_ON(IS_ERR(hbcall));
662
663 o2hb_fire_callbacks(hbcall, event->hn_node, event->hn_node_num);
664
665 spin_lock(&o2hb_live_lock);
666 }
667 spin_unlock(&o2hb_live_lock);
668
669 up_write(&o2hb_callback_sem);
670}
671
672static void o2hb_queue_node_event(struct o2hb_node_event *event,
673 enum o2hb_callback_type type,
674 struct o2nm_node *node,
675 int node_num)
676{
677 assert_spin_locked(&o2hb_live_lock);
678
0e105d37
SM		 679 BUG_ON((!node) && (type != O2HB_NODE_DOWN_CB));
680
a7f6a5fb
MF		 681 event->hn_event_type = type;
682 event->hn_node = node;
683 event->hn_node_num = node_num;
684
685 mlog(ML_HEARTBEAT, "Queue node %s event for node %d\n",
686 type == O2HB_NODE_UP_CB ? "UP" : "DOWN", node_num);
687
688 list_add_tail(&event->hn_item, &o2hb_node_events);
689}
690
691static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot)
692{
693 struct o2hb_node_event event =
694 { .hn_item = LIST_HEAD_INIT(event.hn_item), };
695 struct o2nm_node *node;
696
697 node = o2nm_get_node_by_num(slot->ds_node_num);
698 if (!node)
699 return;
700
701 spin_lock(&o2hb_live_lock);
702 if (!list_empty(&slot->ds_live_item)) {
703 mlog(ML_HEARTBEAT, "Shutdown, node %d leaves region\n",
704 slot->ds_node_num);
705
706 list_del_init(&slot->ds_live_item);
707
708 if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
709 clear_bit(slot->ds_node_num, o2hb_live_node_bitmap);
710
711 o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB, node,
712 slot->ds_node_num);
713 }
714 }
715 spin_unlock(&o2hb_live_lock);
716
717 o2hb_run_event_list(&event);
718
719 o2nm_node_put(node);
720}
721
43182d2a
SM		 722static void o2hb_set_quorum_device(struct o2hb_region *reg,
723 struct o2hb_disk_slot *slot)
724{
725 assert_spin_locked(&o2hb_live_lock);
726
727 if (!o2hb_global_heartbeat_active())
728 return;
729
730 if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
731 return;
732
733 /*
734 * A region can be added to the quorum only when it sees all
735 * live nodes heartbeat on it. In other words, the region has been
736 * added to all nodes.
737 */
738 if (memcmp(reg->hr_live_node_bitmap, o2hb_live_node_bitmap,
739 sizeof(o2hb_live_node_bitmap)))
740 return;
741
742 if (slot->ds_changed_samples < O2HB_LIVE_THRESHOLD)
743 return;
744
745 printk(KERN_NOTICE "o2hb: Region %s is now a quorum device\n",
746 config_item_name(&reg->hr_item));
747
748 set_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
58a3158a
SM		 749
750 /*
751 * If global heartbeat active, unpin all regions if the
752 * region count > CUT_OFF
753 */
754 if (o2hb_pop_count(&o2hb_quorum_region_bitmap,
755 O2NM_MAX_REGIONS) > O2HB_PIN_CUT_OFF)
756 o2hb_region_unpin(NULL);
43182d2a
SM		 757}
758
a7f6a5fb
MF		 759static int o2hb_check_slot(struct o2hb_region *reg,
760 struct o2hb_disk_slot *slot)
761{
762 int changed = 0, gen_changed = 0;
763 struct o2hb_node_event event =
764 { .hn_item = LIST_HEAD_INIT(event.hn_item), };
765 struct o2nm_node *node;
766 struct o2hb_disk_heartbeat_block *hb_block = reg->hr_tmp_block;
767 u64 cputime;
0db638f4
MF		 768 unsigned int dead_ms = o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS;
769 unsigned int slot_dead_ms;
0e105d37 770 int tmp;
a7f6a5fb
MF		 771
772 memcpy(hb_block, slot->ds_raw_block, reg->hr_block_bytes);
773
0e105d37
SM		 774 /*
775 * If a node is no longer configured but is still in the livemap, we
776 * may need to clear that bit from the livemap.
777 */
a7f6a5fb 778 node = o2nm_get_node_by_num(slot->ds_node_num);
0e105d37
SM		 779 if (!node) {
780 spin_lock(&o2hb_live_lock);
781 tmp = test_bit(slot->ds_node_num, o2hb_live_node_bitmap);
782 spin_unlock(&o2hb_live_lock);
783 if (!tmp)
784 return 0;
785 }
a7f6a5fb
MF		 786
787 if (!o2hb_verify_crc(reg, hb_block)) {
788 /* all paths from here will drop o2hb_live_lock for
789 * us. */
790 spin_lock(&o2hb_live_lock);
791
792 /* Don't print an error on the console in this case -
793 * a freshly formatted heartbeat area will not have a
794 * crc set on it. */
795 if (list_empty(&slot->ds_live_item))
796 goto out;
797
798 /* The node is live but pushed out a bad crc. We
799 * consider it a transient miss but don't populate any
800 * other values as they may be junk. */
801 mlog(ML_ERROR, "Node %d has written a bad crc to %s\n",
802 slot->ds_node_num, reg->hr_dev_name);
803 o2hb_dump_slot(hb_block);
804
805 slot->ds_equal_samples++;
806 goto fire_callbacks;
807 }
808
809 /* we don't care if these wrap.. the state transitions below
810 * clear at the right places */
811 cputime = le64_to_cpu(hb_block->hb_seq);
812 if (slot->ds_last_time != cputime)
813 slot->ds_changed_samples++;
814 else
815 slot->ds_equal_samples++;
816 slot->ds_last_time = cputime;
817
818 /* The node changed heartbeat generations. We assume this to
819 * mean it dropped off but came back before we timed out. We
820 * want to consider it down for the time being but don't want
821 * to lose any changed_samples state we might build up to
822 * considering it live again. */
823 if (slot->ds_last_generation != le64_to_cpu(hb_block->hb_generation)) {
824 gen_changed = 1;
825 slot->ds_equal_samples = 0;
70bacbdb
MF		 826 mlog(ML_HEARTBEAT, "Node %d changed generation (0x%llx "
827 "to 0x%llx)\n", slot->ds_node_num,
828 (long long)slot->ds_last_generation,
829 (long long)le64_to_cpu(hb_block->hb_generation));
a7f6a5fb
MF		 830 }
831
832 slot->ds_last_generation = le64_to_cpu(hb_block->hb_generation);
833
70bacbdb
MF		 834 mlog(ML_HEARTBEAT, "Slot %d gen 0x%llx cksum 0x%x "
835 "seq %llu last %llu changed %u equal %u\n",
836 slot->ds_node_num, (long long)slot->ds_last_generation,
837 le32_to_cpu(hb_block->hb_cksum),
2bd63216 838 (unsigned long long)le64_to_cpu(hb_block->hb_seq),
70bacbdb 839 (unsigned long long)slot->ds_last_time, slot->ds_changed_samples,
a7f6a5fb
MF		 840 slot->ds_equal_samples);
841
842 spin_lock(&o2hb_live_lock);
843
844fire_callbacks:
845 /* dead nodes only come to life after some number of
846 * changes at any time during their dead time */
847 if (list_empty(&slot->ds_live_item) &&
848 slot->ds_changed_samples >= O2HB_LIVE_THRESHOLD) {
70bacbdb
MF		 849 mlog(ML_HEARTBEAT, "Node %d (id 0x%llx) joined my region\n",
850 slot->ds_node_num, (long long)slot->ds_last_generation);
a7f6a5fb 851
823a637a
SM		 852 set_bit(slot->ds_node_num, reg->hr_live_node_bitmap);
853
a7f6a5fb
MF		 854 /* first on the list generates a callback */
855 if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
d6aa1c7c
SM		 856 mlog(ML_HEARTBEAT, "o2hb: Add node %d to live nodes "
857 "bitmap\n", slot->ds_node_num);
a7f6a5fb
MF		 858 set_bit(slot->ds_node_num, o2hb_live_node_bitmap);
859
860 o2hb_queue_node_event(&event, O2HB_NODE_UP_CB, node,
861 slot->ds_node_num);
862
863 changed = 1;
864 }
865
866 list_add_tail(&slot->ds_live_item,
867 &o2hb_live_slots[slot->ds_node_num]);
868
869 slot->ds_equal_samples = 0;
0db638f4
MF		 870
871 /* We want to be sure that all nodes agree on the
872 * number of milliseconds before a node will be
873 * considered dead. The self-fencing timeout is
874 * computed from this value, and a discrepancy might
875 * result in heartbeat calling a node dead when it
876 * hasn't self-fenced yet. */
877 slot_dead_ms = le32_to_cpu(hb_block->hb_dead_ms);
878 if (slot_dead_ms && slot_dead_ms != dead_ms) {
879 /* TODO: Perhaps we can fail the region here. */
880 mlog(ML_ERROR, "Node %d on device %s has a dead count "
881 "of %u ms, but our count is %u ms.\n"
882 "Please double check your configuration values "
883 "for 'O2CB_HEARTBEAT_THRESHOLD'\n",
884 slot->ds_node_num, reg->hr_dev_name, slot_dead_ms,
885 dead_ms);
886 }
a7f6a5fb
MF		 887 goto out;
888 }
889
890 /* if the list is dead, we're done.. */
891 if (list_empty(&slot->ds_live_item))
892 goto out;
893
894 /* live nodes only go dead after enough consequtive missed
895 * samples.. reset the missed counter whenever we see
896 * activity */
897 if (slot->ds_equal_samples >= o2hb_dead_threshold || gen_changed) {
898 mlog(ML_HEARTBEAT, "Node %d left my region\n",
899 slot->ds_node_num);
900
823a637a
SM		 901 clear_bit(slot->ds_node_num, reg->hr_live_node_bitmap);
902
a7f6a5fb
MF		 903 /* last off the live_slot generates a callback */
904 list_del_init(&slot->ds_live_item);
905 if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
d6aa1c7c
SM		 906 mlog(ML_HEARTBEAT, "o2hb: Remove node %d from live "
907 "nodes bitmap\n", slot->ds_node_num);
a7f6a5fb
MF		 908 clear_bit(slot->ds_node_num, o2hb_live_node_bitmap);
909
0e105d37
SM		 910 /* node can be null */
911 o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB,
912 node, slot->ds_node_num);
a7f6a5fb
MF		 913
914 changed = 1;
915 }
916
917 /* We don't clear this because the node is still
918 * actually writing new blocks. */
919 if (!gen_changed)
920 slot->ds_changed_samples = 0;
921 goto out;
922 }
923 if (slot->ds_changed_samples) {
924 slot->ds_changed_samples = 0;
925 slot->ds_equal_samples = 0;
926 }
927out:
43182d2a
SM		 928 o2hb_set_quorum_device(reg, slot);
929
a7f6a5fb
MF		 930 spin_unlock(&o2hb_live_lock);
931
932 o2hb_run_event_list(&event);
933
0e105d37
SM		 934 if (node)
935 o2nm_node_put(node);
a7f6a5fb
MF		 936 return changed;
937}
938
939/* This could be faster if we just implmented a find_last_bit, but I
940 * don't think the circumstances warrant it. */
941static int o2hb_highest_node(unsigned long *nodes,
942 int numbits)
943{
944 int highest, node;
945
946 highest = numbits;
947 node = -1;
948 while ((node = find_next_bit(nodes, numbits, node + 1)) != -1) {
949 if (node >= numbits)
950 break;
951
952 highest = node;
953 }
954
955 return highest;
956}
957
a9e2ae39 958static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
a7f6a5fb
MF		 959{
960 int i, ret, highest_node, change = 0;
961 unsigned long configured_nodes[BITS_TO_LONGS(O2NM_MAX_NODES)];
0e105d37 962 unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
a7f6a5fb
MF		 963 struct o2hb_bio_wait_ctxt write_wc;
964
a9e2ae39
MF		 965 ret = o2nm_configured_node_map(configured_nodes,
966 sizeof(configured_nodes));
967 if (ret) {
968 mlog_errno(ret);
969 return ret;
970 }
a7f6a5fb 971
0e105d37
SM		 972 /*
973 * If a node is not configured but is in the livemap, we still need
974 * to read the slot so as to be able to remove it from the livemap.
975 */
976 o2hb_fill_node_map(live_node_bitmap, sizeof(live_node_bitmap));
977 i = -1;
978 while ((i = find_next_bit(live_node_bitmap,
979 O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
980 set_bit(i, configured_nodes);
981 }
982
a7f6a5fb
MF		 983 highest_node = o2hb_highest_node(configured_nodes, O2NM_MAX_NODES);
984 if (highest_node >= O2NM_MAX_NODES) {
985 mlog(ML_NOTICE, "ocfs2_heartbeat: no configured nodes found!\n");
a9e2ae39 986 return -EINVAL;
a7f6a5fb
MF		 987 }
988
989 /* No sense in reading the slots of nodes that don't exist
990 * yet. Of course, if the node definitions have holes in them
991 * then we're reading an empty slot anyway... Consider this
992 * best-effort. */
993 ret = o2hb_read_slots(reg, highest_node + 1);
994 if (ret < 0) {
995 mlog_errno(ret);
a9e2ae39 996 return ret;
a7f6a5fb
MF		 997 }
998
999 /* With an up to date view of the slots, we can check that no
1000 * other node has been improperly configured to heartbeat in
1001 * our slot. */
33c12a54 1002 o2hb_check_last_timestamp(reg);
a7f6a5fb
MF		 1003
1004 /* fill in the proper info for our next heartbeat */
1005 o2hb_prepare_block(reg, reg->hr_generation);
1006
1007 /* And fire off the write. Note that we don't wait on this I/O
1008 * until later. */
b559292e 1009 ret = o2hb_issue_node_write(reg, &write_wc);
a7f6a5fb
MF		 1010 if (ret < 0) {
1011 mlog_errno(ret);
a9e2ae39 1012 return ret;
a7f6a5fb
MF		 1013 }
1014
1015 i = -1;
33c12a54
SM		 1016 while((i = find_next_bit(configured_nodes,
1017 O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
a7f6a5fb
MF		 1018 change |= o2hb_check_slot(reg, &reg->hr_slots[i]);
1019 }
1020
1021 /*
1022 * We have to be sure we've advertised ourselves on disk
1023 * before we can go to steady state. This ensures that
1024 * people we find in our steady state have seen us.
1025 */
1026 o2hb_wait_on_io(reg, &write_wc);
a9e2ae39
MF		 1027 if (write_wc.wc_error) {
1028 /* Do not re-arm the write timeout on I/O error - we
1029 * can't be sure that the new block ever made it to
1030 * disk */
1031 mlog(ML_ERROR, "Write error %d on device \"%s\"\n",
1032 write_wc.wc_error, reg->hr_dev_name);
1033 return write_wc.wc_error;
1034 }
1035
a7f6a5fb
MF		 1036 o2hb_arm_write_timeout(reg);
1037
1038 /* let the person who launched us know when things are steady */
1039 if (!change && (atomic_read(&reg->hr_steady_iterations) != 0)) {
1040 if (atomic_dec_and_test(&reg->hr_steady_iterations))
1041 wake_up(&o2hb_steady_queue);
1042 }
a9e2ae39
MF		 1043
1044 return 0;
a7f6a5fb
MF		 1045}
1046
1047/* Subtract b from a, storing the result in a. a *must* have a larger
1048 * value than b. */
1049static void o2hb_tv_subtract(struct timeval *a,
1050 struct timeval *b)
1051{
1052 /* just return 0 when a is after b */
1053 if (a->tv_sec < b->tv_sec ||
1054 (a->tv_sec == b->tv_sec && a->tv_usec < b->tv_usec)) {
1055 a->tv_sec = 0;
1056 a->tv_usec = 0;
1057 return;
1058 }
1059
1060 a->tv_sec -= b->tv_sec;
1061 a->tv_usec -= b->tv_usec;
1062 while ( a->tv_usec < 0 ) {
1063 a->tv_sec--;
1064 a->tv_usec += 1000000;
1065 }
1066}
1067
1068static unsigned int o2hb_elapsed_msecs(struct timeval *start,
1069 struct timeval *end)
1070{
1071 struct timeval res = *end;
1072
1073 o2hb_tv_subtract(&res, start);
1074
1075 return res.tv_sec * 1000 + res.tv_usec / 1000;
1076}
1077
1078/*
1079 * we ride the region ref that the region dir holds. before the region
1080 * dir is removed and drops it ref it will wait to tear down this
1081 * thread.
1082 */
1083static int o2hb_thread(void *data)
1084{
1085 int i, ret;
1086 struct o2hb_region *reg = data;
a7f6a5fb
MF		 1087 struct o2hb_bio_wait_ctxt write_wc;
1088 struct timeval before_hb, after_hb;
1089 unsigned int elapsed_msec;
1090
1091 mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread running\n");
1092
1093 set_user_nice(current, -20);
1094
cfc069d3
SM		 1095 /* Pin node */
1096 o2nm_depend_this_node();
1097
a7f6a5fb
MF		 1098 while (!kthread_should_stop() && !reg->hr_unclean_stop) {
1099 /* We track the time spent inside
025dfdaf 1100 * o2hb_do_disk_heartbeat so that we avoid more than
a7f6a5fb
MF		 1101 * hr_timeout_ms between disk writes. On busy systems
1102 * this should result in a heartbeat which is less
1103 * likely to time itself out. */
1104 do_gettimeofday(&before_hb);
1105
a9e2ae39
MF		 1106 i = 0;
1107 do {
1108 ret = o2hb_do_disk_heartbeat(reg);
1109 } while (ret && ++i < 2);
a7f6a5fb
MF		 1110
1111 do_gettimeofday(&after_hb);
1112 elapsed_msec = o2hb_elapsed_msecs(&before_hb, &after_hb);
1113
b31d308d
TM		 1114 mlog(ML_HEARTBEAT,
1115 "start = %lu.%lu, end = %lu.%lu, msec = %u\n",
215c7f9f
MF		 1116 before_hb.tv_sec, (unsigned long) before_hb.tv_usec,
1117 after_hb.tv_sec, (unsigned long) after_hb.tv_usec,
1118 elapsed_msec);
a7f6a5fb
MF		 1119
1120 if (elapsed_msec < reg->hr_timeout_ms) {
1121 /* the kthread api has blocked signals for us so no
1122 * need to record the return value. */
1123 msleep_interruptible(reg->hr_timeout_ms - elapsed_msec);
1124 }
1125 }
1126
1127 o2hb_disarm_write_timeout(reg);
1128
1129 /* unclean stop is only used in very bad situation */
1130 for(i = 0; !reg->hr_unclean_stop && i < reg->hr_blocks; i++)
1131 o2hb_shutdown_slot(&reg->hr_slots[i]);
1132
1133 /* Explicit down notification - avoid forcing the other nodes
1134 * to timeout on this region when we could just as easily
1135 * write a clear generation - thus indicating to them that
1136 * this node has left this region.
1137 *
1138 * XXX: Should we skip this on unclean_stop? */
1139 o2hb_prepare_block(reg, 0);
b559292e 1140 ret = o2hb_issue_node_write(reg, &write_wc);
a7f6a5fb
MF		 1141 if (ret == 0) {
1142 o2hb_wait_on_io(reg, &write_wc);
a7f6a5fb
MF		 1143 } else {
1144 mlog_errno(ret);
1145 }
1146
cfc069d3
SM		 1147 /* Unpin node */
1148 o2nm_undepend_this_node();
a7f6a5fb
MF		 1149
1150 mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread exiting\n");
1151
1152 return 0;
1153}
1154
87d3d3f3
SM		 1155#ifdef CONFIG_DEBUG_FS
1156static int o2hb_debug_open(struct inode *inode, struct file *file)
1157{
8ca8b0bb 1158 struct o2hb_debug_buf *db = inode->i_private;
1f285305 1159 struct o2hb_region *reg;
87d3d3f3
SM		 1160 unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)];
1161 char *buf = NULL;
1162 int i = -1;
1163 int out = 0;
1164
8ca8b0bb
SM		 1165 /* max_nodes should be the largest bitmap we pass here */
1166 BUG_ON(sizeof(map) < db->db_size);
1167
87d3d3f3
SM		 1168 buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
1169 if (!buf)
1170 goto bail;
1171
8ca8b0bb
SM		 1172 switch (db->db_type) {
1173 case O2HB_DB_TYPE_LIVENODES:
a6de0136
SM		 1174 case O2HB_DB_TYPE_LIVEREGIONS:
1175 case O2HB_DB_TYPE_QUORUMREGIONS:
1176 case O2HB_DB_TYPE_FAILEDREGIONS:
8ca8b0bb
SM		 1177 spin_lock(&o2hb_live_lock);
1178 memcpy(map, db->db_data, db->db_size);
1179 spin_unlock(&o2hb_live_lock);
1180 break;
87d3d3f3 1181
1f285305
SM		 1182 case O2HB_DB_TYPE_REGION_LIVENODES:
1183 spin_lock(&o2hb_live_lock);
1184 reg = (struct o2hb_region *)db->db_data;
1185 memcpy(map, reg->hr_live_node_bitmap, db->db_size);
1186 spin_unlock(&o2hb_live_lock);
1187 break;
1188
1189 case O2HB_DB_TYPE_REGION_NUMBER:
1190 reg = (struct o2hb_region *)db->db_data;
1191 out += snprintf(buf + out, PAGE_SIZE - out, "%d\n",
1192 reg->hr_region_num);
1193 goto done;
1194
43695d09
SM		 1195 case O2HB_DB_TYPE_REGION_ELAPSED_TIME:
1196 reg = (struct o2hb_region *)db->db_data;
1197 out += snprintf(buf + out, PAGE_SIZE - out, "%u\n",
1198 jiffies_to_msecs(jiffies -
1199 reg->hr_last_timeout_start));
1200 goto done;
1201
cb0586bd
SM		 1202 case O2HB_DB_TYPE_REGION_PINNED:
1203 reg = (struct o2hb_region *)db->db_data;
1204 out += snprintf(buf + out, PAGE_SIZE - out, "%u\n",
1205 !!reg->hr_item_pinned);
1206 goto done;
1207
8ca8b0bb
SM		 1208 default:
1209 goto done;
1210 }
1211
1212 while ((i = find_next_bit(map, db->db_len, i + 1)) < db->db_len)
87d3d3f3
SM		 1213 out += snprintf(buf + out, PAGE_SIZE - out, "%d ", i);
1214 out += snprintf(buf + out, PAGE_SIZE - out, "\n");
1215
8ca8b0bb 1216done:
87d3d3f3
SM		 1217 i_size_write(inode, out);
1218
1219 file->private_data = buf;
1220
1221 return 0;
1222bail:
1223 return -ENOMEM;
1224}
1225
1226static int o2hb_debug_release(struct inode *inode, struct file *file)
1227{
1228 kfree(file->private_data);
1229 return 0;
1230}
1231
1232static ssize_t o2hb_debug_read(struct file *file, char __user *buf,
1233 size_t nbytes, loff_t *ppos)
1234{
1235 return simple_read_from_buffer(buf, nbytes, ppos, file->private_data,
1236 i_size_read(file->f_mapping->host));
1237}
1238#else
1239static int o2hb_debug_open(struct inode *inode, struct file *file)
1240{
1241 return 0;
1242}
1243static int o2hb_debug_release(struct inode *inode, struct file *file)
1244{
1245 return 0;
1246}
1247static ssize_t o2hb_debug_read(struct file *file, char __user *buf,
1248 size_t nbytes, loff_t *ppos)
1249{
1250 return 0;
1251}
1252#endif /* CONFIG_DEBUG_FS */
1253
828c0950 1254static const struct file_operations o2hb_debug_fops = {
87d3d3f3
SM		 1255 .open = o2hb_debug_open,
1256 .release = o2hb_debug_release,
1257 .read = o2hb_debug_read,
1258 .llseek = generic_file_llseek,
1259};
1260
1261void o2hb_exit(void)
1262{
8ca8b0bb 1263 kfree(o2hb_db_livenodes);
a6de0136
SM		 1264 kfree(o2hb_db_liveregions);
1265 kfree(o2hb_db_quorumregions);
1266 kfree(o2hb_db_failedregions);
1267 debugfs_remove(o2hb_debug_failedregions);
1268 debugfs_remove(o2hb_debug_quorumregions);
1269 debugfs_remove(o2hb_debug_liveregions);
8ca8b0bb
SM		 1270 debugfs_remove(o2hb_debug_livenodes);
1271 debugfs_remove(o2hb_debug_dir);
1272}
1273
1274static struct dentry *o2hb_debug_create(const char *name, struct dentry *dir,
1275 struct o2hb_debug_buf **db, int db_len,
1276 int type, int size, int len, void *data)
1277{
1278 *db = kmalloc(db_len, GFP_KERNEL);
1279 if (!*db)
1280 return NULL;
1281
1282 (*db)->db_type = type;
1283 (*db)->db_size = size;
1284 (*db)->db_len = len;
1285 (*db)->db_data = data;
1286
1287 return debugfs_create_file(name, S_IFREG|S_IRUSR, dir, *db,
1288 &o2hb_debug_fops);
1289}
1290
1291static int o2hb_debug_init(void)
1292{
1293 int ret = -ENOMEM;
1294
1295 o2hb_debug_dir = debugfs_create_dir(O2HB_DEBUG_DIR, NULL);
1296 if (!o2hb_debug_dir) {
1297 mlog_errno(ret);
1298 goto bail;
1299 }
1300
1301 o2hb_debug_livenodes = o2hb_debug_create(O2HB_DEBUG_LIVENODES,
1302 o2hb_debug_dir,
1303 &o2hb_db_livenodes,
1304 sizeof(*o2hb_db_livenodes),
1305 O2HB_DB_TYPE_LIVENODES,
1306 sizeof(o2hb_live_node_bitmap),
1307 O2NM_MAX_NODES,
1308 o2hb_live_node_bitmap);
1309 if (!o2hb_debug_livenodes) {
1310 mlog_errno(ret);
1311 goto bail;
1312 }
a6de0136
SM		 1313
1314 o2hb_debug_liveregions = o2hb_debug_create(O2HB_DEBUG_LIVEREGIONS,
1315 o2hb_debug_dir,
1316 &o2hb_db_liveregions,
1317 sizeof(*o2hb_db_liveregions),
1318 O2HB_DB_TYPE_LIVEREGIONS,
1319 sizeof(o2hb_live_region_bitmap),
1320 O2NM_MAX_REGIONS,
1321 o2hb_live_region_bitmap);
1322 if (!o2hb_debug_liveregions) {
1323 mlog_errno(ret);
1324 goto bail;
1325 }
1326
1327 o2hb_debug_quorumregions =
1328 o2hb_debug_create(O2HB_DEBUG_QUORUMREGIONS,
1329 o2hb_debug_dir,
1330 &o2hb_db_quorumregions,
1331 sizeof(*o2hb_db_quorumregions),
1332 O2HB_DB_TYPE_QUORUMREGIONS,
1333 sizeof(o2hb_quorum_region_bitmap),
1334 O2NM_MAX_REGIONS,
1335 o2hb_quorum_region_bitmap);
1336 if (!o2hb_debug_quorumregions) {
1337 mlog_errno(ret);
1338 goto bail;
1339 }
1340
1341 o2hb_debug_failedregions =
1342 o2hb_debug_create(O2HB_DEBUG_FAILEDREGIONS,
1343 o2hb_debug_dir,
1344 &o2hb_db_failedregions,
1345 sizeof(*o2hb_db_failedregions),
1346 O2HB_DB_TYPE_FAILEDREGIONS,
1347 sizeof(o2hb_failed_region_bitmap),
1348 O2NM_MAX_REGIONS,
1349 o2hb_failed_region_bitmap);
1350 if (!o2hb_debug_failedregions) {
1351 mlog_errno(ret);
1352 goto bail;
1353 }
1354
8ca8b0bb
SM		 1355 ret = 0;
1356bail:
1357 if (ret)
1358 o2hb_exit();
1359
1360 return ret;
87d3d3f3
SM		 1361}
1362
1363int o2hb_init(void)
a7f6a5fb
MF		 1364{
1365 int i;
1366
1367 for (i = 0; i < ARRAY_SIZE(o2hb_callbacks); i++)
1368 INIT_LIST_HEAD(&o2hb_callbacks[i].list);
1369
1370 for (i = 0; i < ARRAY_SIZE(o2hb_live_slots); i++)
1371 INIT_LIST_HEAD(&o2hb_live_slots[i]);
1372
1373 INIT_LIST_HEAD(&o2hb_node_events);
1374
1375 memset(o2hb_live_node_bitmap, 0, sizeof(o2hb_live_node_bitmap));
536f0741 1376 memset(o2hb_region_bitmap, 0, sizeof(o2hb_region_bitmap));
e7d656ba 1377 memset(o2hb_live_region_bitmap, 0, sizeof(o2hb_live_region_bitmap));
43182d2a 1378 memset(o2hb_quorum_region_bitmap, 0, sizeof(o2hb_quorum_region_bitmap));
b1c5ebfb 1379 memset(o2hb_failed_region_bitmap, 0, sizeof(o2hb_failed_region_bitmap));
87d3d3f3 1380
58a3158a
SM		 1381 o2hb_dependent_users = 0;
1382
8ca8b0bb 1383 return o2hb_debug_init();
a7f6a5fb
MF		 1384}
1385
1386/* if we're already in a callback then we're already serialized by the sem */
1387static void o2hb_fill_node_map_from_callback(unsigned long *map,
1388 unsigned bytes)
1389{
1390 BUG_ON(bytes < (BITS_TO_LONGS(O2NM_MAX_NODES) * sizeof(unsigned long)));
1391
1392 memcpy(map, &o2hb_live_node_bitmap, bytes);
1393}
1394
1395/*
1396 * get a map of all nodes that are heartbeating in any regions
1397 */
1398void o2hb_fill_node_map(unsigned long *map, unsigned bytes)
1399{
1400 /* callers want to serialize this map and callbacks so that they
1401 * can trust that they don't miss nodes coming to the party */
1402 down_read(&o2hb_callback_sem);
1403 spin_lock(&o2hb_live_lock);
1404 o2hb_fill_node_map_from_callback(map, bytes);
1405 spin_unlock(&o2hb_live_lock);
1406 up_read(&o2hb_callback_sem);
1407}
1408EXPORT_SYMBOL_GPL(o2hb_fill_node_map);
1409
1410/*
1411 * heartbeat configfs bits. The heartbeat set is a default set under
1412 * the cluster set in nodemanager.c.
1413 */
1414
1415static struct o2hb_region *to_o2hb_region(struct config_item *item)
1416{
1417 return item ? container_of(item, struct o2hb_region, hr_item) : NULL;
1418}
1419
1420/* drop_item only drops its ref after killing the thread, nothing should
1421 * be using the region anymore. this has to clean up any state that
1422 * attributes might have built up. */
1423static void o2hb_region_release(struct config_item *item)
1424{
1425 int i;
1426 struct page *page;
1427 struct o2hb_region *reg = to_o2hb_region(item);
1428
1429 if (reg->hr_tmp_block)
1430 kfree(reg->hr_tmp_block);
1431
1432 if (reg->hr_slot_data) {
1433 for (i = 0; i < reg->hr_num_pages; i++) {
1434 page = reg->hr_slot_data[i];
1435 if (page)
1436 __free_page(page);
1437 }
1438 kfree(reg->hr_slot_data);
1439 }
1440
1441 if (reg->hr_bdev)
9a1c3542 1442 blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE);
a7f6a5fb
MF		 1443
1444 if (reg->hr_slots)
1445 kfree(reg->hr_slots);
1446
1f285305
SM		 1447 kfree(reg->hr_db_regnum);
1448 kfree(reg->hr_db_livenodes);
1449 debugfs_remove(reg->hr_debug_livenodes);
1450 debugfs_remove(reg->hr_debug_regnum);
d4396eaf 1451 debugfs_remove(reg->hr_debug_elapsed_time);
cb0586bd 1452 debugfs_remove(reg->hr_debug_pinned);
1f285305
SM		 1453 debugfs_remove(reg->hr_debug_dir);
1454
a7f6a5fb
MF		 1455 spin_lock(&o2hb_live_lock);
1456 list_del(&reg->hr_all_item);
1457 spin_unlock(&o2hb_live_lock);
1458
1459 kfree(reg);
1460}
1461
1462static int o2hb_read_block_input(struct o2hb_region *reg,
1463 const char *page,
1464 size_t count,
1465 unsigned long *ret_bytes,
1466 unsigned int *ret_bits)
1467{
1468 unsigned long bytes;
1469 char *p = (char *)page;
1470
1471 bytes = simple_strtoul(p, &p, 0);
1472 if (!p || (*p && (*p != '\n')))
1473 return -EINVAL;
1474
1475 /* Heartbeat and fs min / max block sizes are the same. */
1476 if (bytes > 4096 || bytes < 512)
1477 return -ERANGE;
1478 if (hweight16(bytes) != 1)
1479 return -EINVAL;
1480
1481 if (ret_bytes)
1482 *ret_bytes = bytes;
1483 if (ret_bits)
1484 *ret_bits = ffs(bytes) - 1;
1485
1486 return 0;
1487}
1488
1489static ssize_t o2hb_region_block_bytes_read(struct o2hb_region *reg,
1490 char *page)
1491{
1492 return sprintf(page, "%u\n", reg->hr_block_bytes);
1493}
1494
1495static ssize_t o2hb_region_block_bytes_write(struct o2hb_region *reg,
1496 const char *page,
1497 size_t count)
1498{
1499 int status;
1500 unsigned long block_bytes;
1501 unsigned int block_bits;
1502
1503 if (reg->hr_bdev)
1504 return -EINVAL;
1505
1506 status = o2hb_read_block_input(reg, page, count,
1507 &block_bytes, &block_bits);
1508 if (status)
1509 return status;
1510
1511 reg->hr_block_bytes = (unsigned int)block_bytes;
1512 reg->hr_block_bits = block_bits;
1513
1514 return count;
1515}
1516
1517static ssize_t o2hb_region_start_block_read(struct o2hb_region *reg,
1518 char *page)
1519{
1520 return sprintf(page, "%llu\n", reg->hr_start_block);
1521}
1522
1523static ssize_t o2hb_region_start_block_write(struct o2hb_region *reg,
1524 const char *page,
1525 size_t count)
1526{
1527 unsigned long long tmp;
1528 char *p = (char *)page;
1529
1530 if (reg->hr_bdev)
1531 return -EINVAL;
1532
1533 tmp = simple_strtoull(p, &p, 0);
1534 if (!p || (*p && (*p != '\n')))
1535 return -EINVAL;
1536
1537 reg->hr_start_block = tmp;
1538
1539 return count;
1540}
1541
1542static ssize_t o2hb_region_blocks_read(struct o2hb_region *reg,
1543 char *page)
1544{
1545 return sprintf(page, "%d\n", reg->hr_blocks);
1546}
1547
1548static ssize_t o2hb_region_blocks_write(struct o2hb_region *reg,
1549 const char *page,
1550 size_t count)
1551{
1552 unsigned long tmp;
1553 char *p = (char *)page;
1554
1555 if (reg->hr_bdev)
1556 return -EINVAL;
1557
1558 tmp = simple_strtoul(p, &p, 0);
1559 if (!p || (*p && (*p != '\n')))
1560 return -EINVAL;
1561
1562 if (tmp > O2NM_MAX_NODES || tmp == 0)
1563 return -ERANGE;
1564
1565 reg->hr_blocks = (unsigned int)tmp;
1566
1567 return count;
1568}
1569
1570static ssize_t o2hb_region_dev_read(struct o2hb_region *reg,
1571 char *page)
1572{
1573 unsigned int ret = 0;
1574
1575 if (reg->hr_bdev)
1576 ret = sprintf(page, "%s\n", reg->hr_dev_name);
1577
1578 return ret;
1579}
1580
1581static void o2hb_init_region_params(struct o2hb_region *reg)
1582{
1583 reg->hr_slots_per_page = PAGE_CACHE_SIZE >> reg->hr_block_bits;
1584 reg->hr_timeout_ms = O2HB_REGION_TIMEOUT_MS;
1585
1586 mlog(ML_HEARTBEAT, "hr_start_block = %llu, hr_blocks = %u\n",
1587 reg->hr_start_block, reg->hr_blocks);
1588 mlog(ML_HEARTBEAT, "hr_block_bytes = %u, hr_block_bits = %u\n",
1589 reg->hr_block_bytes, reg->hr_block_bits);
1590 mlog(ML_HEARTBEAT, "hr_timeout_ms = %u\n", reg->hr_timeout_ms);
1591 mlog(ML_HEARTBEAT, "dead threshold = %u\n", o2hb_dead_threshold);
1592}
1593
1594static int o2hb_map_slot_data(struct o2hb_region *reg)
1595{
1596 int i, j;
1597 unsigned int last_slot;
1598 unsigned int spp = reg->hr_slots_per_page;
1599 struct page *page;
1600 char *raw;
1601 struct o2hb_disk_slot *slot;
1602
1603 reg->hr_tmp_block = kmalloc(reg->hr_block_bytes, GFP_KERNEL);
1604 if (reg->hr_tmp_block == NULL) {
1605 mlog_errno(-ENOMEM);
1606 return -ENOMEM;
1607 }
1608
1609 reg->hr_slots = kcalloc(reg->hr_blocks,
1610 sizeof(struct o2hb_disk_slot), GFP_KERNEL);
1611 if (reg->hr_slots == NULL) {
1612 mlog_errno(-ENOMEM);
1613 return -ENOMEM;
1614 }
1615
1616 for(i = 0; i < reg->hr_blocks; i++) {
1617 slot = &reg->hr_slots[i];
1618 slot->ds_node_num = i;
1619 INIT_LIST_HEAD(&slot->ds_live_item);
1620 slot->ds_raw_block = NULL;
1621 }
1622
1623 reg->hr_num_pages = (reg->hr_blocks + spp - 1) / spp;
1624 mlog(ML_HEARTBEAT, "Going to require %u pages to cover %u blocks "
1625 "at %u blocks per page\n",
1626 reg->hr_num_pages, reg->hr_blocks, spp);
1627
1628 reg->hr_slot_data = kcalloc(reg->hr_num_pages, sizeof(struct page *),
1629 GFP_KERNEL);
1630 if (!reg->hr_slot_data) {
1631 mlog_errno(-ENOMEM);
1632 return -ENOMEM;
1633 }
1634
1635 for(i = 0; i < reg->hr_num_pages; i++) {
1636 page = alloc_page(GFP_KERNEL);
1637 if (!page) {
1638 mlog_errno(-ENOMEM);
1639 return -ENOMEM;
1640 }
1641
1642 reg->hr_slot_data[i] = page;
1643
1644 last_slot = i * spp;
1645 raw = page_address(page);
1646 for (j = 0;
1647 (j < spp) && ((j + last_slot) < reg->hr_blocks);
1648 j++) {
1649 BUG_ON((j + last_slot) >= reg->hr_blocks);
1650
1651 slot = &reg->hr_slots[j + last_slot];
1652 slot->ds_raw_block =
1653 (struct o2hb_disk_heartbeat_block *) raw;
1654
1655 raw += reg->hr_block_bytes;
1656 }
1657 }
1658
1659 return 0;
1660}
1661
1662/* Read in all the slots available and populate the tracking
1663 * structures so that we can start with a baseline idea of what's
1664 * there. */
1665static int o2hb_populate_slot_data(struct o2hb_region *reg)
1666{
1667 int ret, i;
1668 struct o2hb_disk_slot *slot;
1669 struct o2hb_disk_heartbeat_block *hb_block;
1670
a7f6a5fb
MF		 1671 ret = o2hb_read_slots(reg, reg->hr_blocks);
1672 if (ret) {
1673 mlog_errno(ret);
1674 goto out;
1675 }
1676
1677 /* We only want to get an idea of the values initially in each
1678 * slot, so we do no verification - o2hb_check_slot will
1679 * actually determine if each configured slot is valid and
1680 * whether any values have changed. */
1681 for(i = 0; i < reg->hr_blocks; i++) {
1682 slot = &reg->hr_slots[i];
1683 hb_block = (struct o2hb_disk_heartbeat_block *) slot->ds_raw_block;
1684
1685 /* Only fill the values that o2hb_check_slot uses to
1686 * determine changing slots */
1687 slot->ds_last_time = le64_to_cpu(hb_block->hb_seq);
1688 slot->ds_last_generation = le64_to_cpu(hb_block->hb_generation);
1689 }
1690
1691out:
a7f6a5fb
MF		 1692 return ret;
1693}
1694
1695/* this is acting as commit; we set up all of hr_bdev and hr_task or nothing */
1696static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
1697 const char *page,
1698 size_t count)
1699{
e6c352db 1700 struct task_struct *hb_task;
a7f6a5fb
MF		 1701 long fd;
1702 int sectsize;
1703 char *p = (char *)page;
1704 struct file *filp = NULL;
1705 struct inode *inode = NULL;
1706 ssize_t ret = -EINVAL;
76d9fc29 1707 int live_threshold;
a7f6a5fb
MF		 1708
1709 if (reg->hr_bdev)
1710 goto out;
1711
1712 /* We can't heartbeat without having had our node number
1713 * configured yet. */
1714 if (o2nm_this_node() == O2NM_MAX_NODES)
1715 goto out;
1716
1717 fd = simple_strtol(p, &p, 0);
1718 if (!p || (*p && (*p != '\n')))
1719 goto out;
1720
1721 if (fd < 0 || fd >= INT_MAX)
1722 goto out;
1723
1724 filp = fget(fd);
1725 if (filp == NULL)
1726 goto out;
1727
1728 if (reg->hr_blocks == 0 || reg->hr_start_block == 0 ||
1729 reg->hr_block_bytes == 0)
1730 goto out;
1731
1732 inode = igrab(filp->f_mapping->host);
1733 if (inode == NULL)
1734 goto out;
1735
1736 if (!S_ISBLK(inode->i_mode))
1737 goto out;
1738
1739 reg->hr_bdev = I_BDEV(filp->f_mapping->host);
e525fd89 1740 ret = blkdev_get(reg->hr_bdev, FMODE_WRITE | FMODE_READ, NULL);
a7f6a5fb
MF		 1741 if (ret) {
1742 reg->hr_bdev = NULL;
1743 goto out;
1744 }
1745 inode = NULL;
1746
1747 bdevname(reg->hr_bdev, reg->hr_dev_name);
1748
e1defc4f 1749 sectsize = bdev_logical_block_size(reg->hr_bdev);
a7f6a5fb
MF		 1750 if (sectsize != reg->hr_block_bytes) {
1751 mlog(ML_ERROR,
1752 "blocksize %u incorrect for device, expected %d",
1753 reg->hr_block_bytes, sectsize);
1754 ret = -EINVAL;
1755 goto out;
1756 }
1757
1758 o2hb_init_region_params(reg);
1759
1760 /* Generation of zero is invalid */
1761 do {
1762 get_random_bytes(&reg->hr_generation,
1763 sizeof(reg->hr_generation));
1764 } while (reg->hr_generation == 0);
1765
1766 ret = o2hb_map_slot_data(reg);
1767 if (ret) {
1768 mlog_errno(ret);
1769 goto out;
1770 }
1771
1772 ret = o2hb_populate_slot_data(reg);
1773 if (ret) {
1774 mlog_errno(ret);
1775 goto out;
1776 }
1777
c4028958 1778 INIT_DELAYED_WORK(&reg->hr_write_timeout_work, o2hb_write_timeout);
a7f6a5fb
MF		 1779
1780 /*
1781 * A node is considered live after it has beat LIVE_THRESHOLD
1782 * times. We're not steady until we've given them a chance
1783 * _after_ our first read.
76d9fc29
SM		 1784 * The default threshold is bare minimum so as to limit the delay
1785 * during mounts. For global heartbeat, the threshold doubled for the
1786 * first region.
a7f6a5fb 1787 */
76d9fc29
SM		 1788 live_threshold = O2HB_LIVE_THRESHOLD;
1789 if (o2hb_global_heartbeat_active()) {
1790 spin_lock(&o2hb_live_lock);
1791 if (o2hb_pop_count(&o2hb_region_bitmap, O2NM_MAX_REGIONS) == 1)
1792 live_threshold <<= 1;
1793 spin_unlock(&o2hb_live_lock);
1794 }
1795 atomic_set(&reg->hr_steady_iterations, live_threshold + 1);
a7f6a5fb 1796
e6c352db
JB		 1797 hb_task = kthread_run(o2hb_thread, reg, "o2hb-%s",
1798 reg->hr_item.ci_name);
1799 if (IS_ERR(hb_task)) {
1800 ret = PTR_ERR(hb_task);
a7f6a5fb 1801 mlog_errno(ret);
a7f6a5fb
MF		 1802 goto out;
1803 }
1804
e6c352db
JB		 1805 spin_lock(&o2hb_live_lock);
1806 reg->hr_task = hb_task;
1807 spin_unlock(&o2hb_live_lock);
1808
a7f6a5fb
MF		 1809 ret = wait_event_interruptible(o2hb_steady_queue,
1810 atomic_read(&reg->hr_steady_iterations) == 0);
1811 if (ret) {
e6df3a66 1812 /* We got interrupted (hello ptrace!). Clean up */
e6c352db
JB		 1813 spin_lock(&o2hb_live_lock);
1814 hb_task = reg->hr_task;
a7f6a5fb 1815 reg->hr_task = NULL;
e6c352db
JB		 1816 spin_unlock(&o2hb_live_lock);
1817
1818 if (hb_task)
1819 kthread_stop(hb_task);
a7f6a5fb
MF		 1820 goto out;
1821 }
1822
e6df3a66
JB		 1823 /* Ok, we were woken. Make sure it wasn't by drop_item() */
1824 spin_lock(&o2hb_live_lock);
1825 hb_task = reg->hr_task;
e7d656ba
SM		 1826 if (o2hb_global_heartbeat_active())
1827 set_bit(reg->hr_region_num, o2hb_live_region_bitmap);
e6df3a66
JB		 1828 spin_unlock(&o2hb_live_lock);
1829
1830 if (hb_task)
1831 ret = count;
1832 else
1833 ret = -EIO;
1834
18c50cb0
SM		 1835 if (hb_task && o2hb_global_heartbeat_active())
1836 printk(KERN_NOTICE "o2hb: Heartbeat started on region %s\n",
1837 config_item_name(&reg->hr_item));
1838
a7f6a5fb
MF		 1839out:
1840 if (filp)
1841 fput(filp);
1842 if (inode)
1843 iput(inode);
1844 if (ret < 0) {
1845 if (reg->hr_bdev) {
9a1c3542 1846 blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE);
a7f6a5fb
MF		 1847 reg->hr_bdev = NULL;
1848 }
1849 }
1850 return ret;
1851}
1852
92efc152
ZW		 1853static ssize_t o2hb_region_pid_read(struct o2hb_region *reg,
1854 char *page)
1855{
e6c352db
JB		 1856 pid_t pid = 0;
1857
1858 spin_lock(&o2hb_live_lock);
1859 if (reg->hr_task)
ba25f9dc 1860 pid = task_pid_nr(reg->hr_task);
e6c352db
JB		 1861 spin_unlock(&o2hb_live_lock);
1862
1863 if (!pid)
92efc152
ZW		 1864 return 0;
1865
e6c352db 1866 return sprintf(page, "%u\n", pid);
92efc152
ZW		 1867}
1868
a7f6a5fb
MF		 1869struct o2hb_region_attribute {
1870 struct configfs_attribute attr;
1871 ssize_t (*show)(struct o2hb_region *, char *);
1872 ssize_t (*store)(struct o2hb_region *, const char *, size_t);
1873};
1874
1875static struct o2hb_region_attribute o2hb_region_attr_block_bytes = {
1876 .attr = { .ca_owner = THIS_MODULE,
1877 .ca_name = "block_bytes",
1878 .ca_mode = S_IRUGO | S_IWUSR },
1879 .show = o2hb_region_block_bytes_read,
1880 .store = o2hb_region_block_bytes_write,
1881};
1882
1883static struct o2hb_region_attribute o2hb_region_attr_start_block = {
1884 .attr = { .ca_owner = THIS_MODULE,
1885 .ca_name = "start_block",
1886 .ca_mode = S_IRUGO | S_IWUSR },
1887 .show = o2hb_region_start_block_read,
1888 .store = o2hb_region_start_block_write,
1889};
1890
1891static struct o2hb_region_attribute o2hb_region_attr_blocks = {
1892 .attr = { .ca_owner = THIS_MODULE,
1893 .ca_name = "blocks",
1894 .ca_mode = S_IRUGO | S_IWUSR },
1895 .show = o2hb_region_blocks_read,
1896 .store = o2hb_region_blocks_write,
1897};
1898
1899static struct o2hb_region_attribute o2hb_region_attr_dev = {
1900 .attr = { .ca_owner = THIS_MODULE,
1901 .ca_name = "dev",
1902 .ca_mode = S_IRUGO | S_IWUSR },
1903 .show = o2hb_region_dev_read,
1904 .store = o2hb_region_dev_write,
1905};
1906
92efc152
ZW		 1907static struct o2hb_region_attribute o2hb_region_attr_pid = {
1908 .attr = { .ca_owner = THIS_MODULE,
1909 .ca_name = "pid",
1910 .ca_mode = S_IRUGO | S_IRUSR },
1911 .show = o2hb_region_pid_read,
1912};
1913
a7f6a5fb
MF		 1914static struct configfs_attribute *o2hb_region_attrs[] = {
1915 &o2hb_region_attr_block_bytes.attr,
1916 &o2hb_region_attr_start_block.attr,
1917 &o2hb_region_attr_blocks.attr,
1918 &o2hb_region_attr_dev.attr,
92efc152 1919 &o2hb_region_attr_pid.attr,
a7f6a5fb
MF		 1920 NULL,
1921};
1922
1923static ssize_t o2hb_region_show(struct config_item *item,
1924 struct configfs_attribute *attr,
1925 char *page)
1926{
1927 struct o2hb_region *reg = to_o2hb_region(item);
1928 struct o2hb_region_attribute *o2hb_region_attr =
1929 container_of(attr, struct o2hb_region_attribute, attr);
1930 ssize_t ret = 0;
1931
1932 if (o2hb_region_attr->show)
1933 ret = o2hb_region_attr->show(reg, page);
1934 return ret;
1935}
1936
1937static ssize_t o2hb_region_store(struct config_item *item,
1938 struct configfs_attribute *attr,
1939 const char *page, size_t count)
1940{
1941 struct o2hb_region *reg = to_o2hb_region(item);
1942 struct o2hb_region_attribute *o2hb_region_attr =
1943 container_of(attr, struct o2hb_region_attribute, attr);
1944 ssize_t ret = -EINVAL;
1945
1946 if (o2hb_region_attr->store)
1947 ret = o2hb_region_attr->store(reg, page, count);
1948 return ret;
1949}
1950
1951static struct configfs_item_operations o2hb_region_item_ops = {
1952 .release = o2hb_region_release,
1953 .show_attribute = o2hb_region_show,
1954 .store_attribute = o2hb_region_store,
1955};
1956
1957static struct config_item_type o2hb_region_type = {
1958 .ct_item_ops = &o2hb_region_item_ops,
1959 .ct_attrs = o2hb_region_attrs,
1960 .ct_owner = THIS_MODULE,
1961};
1962
1963/* heartbeat set */
1964
1965struct o2hb_heartbeat_group {
1966 struct config_group hs_group;
1967 /* some stuff? */
1968};
1969
1970static struct o2hb_heartbeat_group *to_o2hb_heartbeat_group(struct config_group *group)
1971{
1972 return group ?
1973 container_of(group, struct o2hb_heartbeat_group, hs_group)
1974 : NULL;
1975}
1976
1f285305
SM		 1977static int o2hb_debug_region_init(struct o2hb_region *reg, struct dentry *dir)
1978{
1979 int ret = -ENOMEM;
1980
1981 reg->hr_debug_dir =
1982 debugfs_create_dir(config_item_name(&reg->hr_item), dir);
1983 if (!reg->hr_debug_dir) {
1984 mlog_errno(ret);
1985 goto bail;
1986 }
1987
1988 reg->hr_debug_livenodes =
1989 o2hb_debug_create(O2HB_DEBUG_LIVENODES,
1990 reg->hr_debug_dir,
1991 &(reg->hr_db_livenodes),
1992 sizeof(*(reg->hr_db_livenodes)),
1993 O2HB_DB_TYPE_REGION_LIVENODES,
1994 sizeof(reg->hr_live_node_bitmap),
1995 O2NM_MAX_NODES, reg);
1996 if (!reg->hr_debug_livenodes) {
1997 mlog_errno(ret);
1998 goto bail;
1999 }
2000
2001 reg->hr_debug_regnum =
2002 o2hb_debug_create(O2HB_DEBUG_REGION_NUMBER,
2003 reg->hr_debug_dir,
2004 &(reg->hr_db_regnum),
2005 sizeof(*(reg->hr_db_regnum)),
2006 O2HB_DB_TYPE_REGION_NUMBER,
2007 0, O2NM_MAX_NODES, reg);
2008 if (!reg->hr_debug_regnum) {
2009 mlog_errno(ret);
2010 goto bail;
2011 }
2012
43695d09
SM		 2013 reg->hr_debug_elapsed_time =
2014 o2hb_debug_create(O2HB_DEBUG_REGION_ELAPSED_TIME,
2015 reg->hr_debug_dir,
2016 &(reg->hr_db_elapsed_time),
2017 sizeof(*(reg->hr_db_elapsed_time)),
2018 O2HB_DB_TYPE_REGION_ELAPSED_TIME,
2019 0, 0, reg);
2020 if (!reg->hr_debug_elapsed_time) {
2021 mlog_errno(ret);
2022 goto bail;
2023 }
2024
cb0586bd
SM		 2025 reg->hr_debug_pinned =
2026 o2hb_debug_create(O2HB_DEBUG_REGION_PINNED,
2027 reg->hr_debug_dir,
2028 &(reg->hr_db_pinned),
2029 sizeof(*(reg->hr_db_pinned)),
2030 O2HB_DB_TYPE_REGION_PINNED,
2031 0, 0, reg);
2032 if (!reg->hr_debug_pinned) {
2033 mlog_errno(ret);
2034 goto bail;
2035 }
2036
1f285305
SM		 2037 ret = 0;
2038bail:
2039 return ret;
2040}
2041
f89ab861
JB		 2042static struct config_item *o2hb_heartbeat_group_make_item(struct config_group *group,
2043 const char *name)
a7f6a5fb
MF		 2044{
2045 struct o2hb_region *reg = NULL;
1f285305 2046 int ret;
a7f6a5fb 2047
cd861280 2048 reg = kzalloc(sizeof(struct o2hb_region), GFP_KERNEL);
f89ab861 2049 if (reg == NULL)
a6795e9e 2050 return ERR_PTR(-ENOMEM);
a7f6a5fb 2051
1cf257f5
JS		 2052 if (strlen(name) > O2HB_MAX_REGION_NAME_LEN) {
2053 ret = -ENAMETOOLONG;
2054 goto free;
2055 }
b3c85c4c 2056
a7f6a5fb 2057 spin_lock(&o2hb_live_lock);
536f0741
SM		 2058 reg->hr_region_num = 0;
2059 if (o2hb_global_heartbeat_active()) {
2060 reg->hr_region_num = find_first_zero_bit(o2hb_region_bitmap,
2061 O2NM_MAX_REGIONS);
2062 if (reg->hr_region_num >= O2NM_MAX_REGIONS) {
2063 spin_unlock(&o2hb_live_lock);
1cf257f5
JS		 2064 ret = -EFBIG;
2065 goto free;
536f0741
SM		 2066 }
2067 set_bit(reg->hr_region_num, o2hb_region_bitmap);
2068 }
a7f6a5fb
MF		 2069 list_add_tail(&reg->hr_all_item, &o2hb_all_regions);
2070 spin_unlock(&o2hb_live_lock);
a7f6a5fb 2071
536f0741
SM		 2072 config_item_init_type_name(&reg->hr_item, name, &o2hb_region_type);
2073
1f285305
SM		 2074 ret = o2hb_debug_region_init(reg, o2hb_debug_dir);
2075 if (ret) {
2076 config_item_put(&reg->hr_item);
1cf257f5 2077 goto free;
1f285305
SM		 2078 }
2079
a6795e9e 2080 return &reg->hr_item;
1cf257f5
JS		 2081free:
2082 kfree(reg);
2083 return ERR_PTR(ret);
a7f6a5fb
MF		 2084}
2085
2086static void o2hb_heartbeat_group_drop_item(struct config_group *group,
2087 struct config_item *item)
2088{
e6c352db 2089 struct task_struct *hb_task;
a7f6a5fb 2090 struct o2hb_region *reg = to_o2hb_region(item);
58a3158a 2091 int quorum_region = 0;
a7f6a5fb
MF		 2092
2093 /* stop the thread when the user removes the region dir */
e6c352db 2094 spin_lock(&o2hb_live_lock);
e7d656ba 2095 if (o2hb_global_heartbeat_active()) {
536f0741 2096 clear_bit(reg->hr_region_num, o2hb_region_bitmap);
e7d656ba 2097 clear_bit(reg->hr_region_num, o2hb_live_region_bitmap);
58a3158a
SM		 2098 if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
2099 quorum_region = 1;
ffee223a 2100 clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
e7d656ba 2101 }
e6c352db
JB		 2102 hb_task = reg->hr_task;
2103 reg->hr_task = NULL;
58a3158a 2104 reg->hr_item_dropped = 1;
e6c352db
JB		 2105 spin_unlock(&o2hb_live_lock);
2106
2107 if (hb_task)
2108 kthread_stop(hb_task);
a7f6a5fb 2109
e6df3a66
JB		 2110 /*
2111 * If we're racing a dev_write(), we need to wake them. They will
2112 * check reg->hr_task
2113 */
2114 if (atomic_read(&reg->hr_steady_iterations) != 0) {
2115 atomic_set(&reg->hr_steady_iterations, 0);
2116 wake_up(&o2hb_steady_queue);
2117 }
2118
18c50cb0
SM		 2119 if (o2hb_global_heartbeat_active())
2120 printk(KERN_NOTICE "o2hb: Heartbeat stopped on region %s\n",
2121 config_item_name(&reg->hr_item));
58a3158a 2122
a7f6a5fb 2123 config_item_put(item);
58a3158a
SM		 2124
2125 if (!o2hb_global_heartbeat_active() || !quorum_region)
2126 return;
2127
2128 /*
2129 * If global heartbeat active and there are dependent users,
2130 * pin all regions if quorum region count <= CUT_OFF
2131 */
2132 spin_lock(&o2hb_live_lock);
2133
2134 if (!o2hb_dependent_users)
2135 goto unlock;
2136
2137 if (o2hb_pop_count(&o2hb_quorum_region_bitmap,
2138 O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
2139 o2hb_region_pin(NULL);
2140
2141unlock:
2142 spin_unlock(&o2hb_live_lock);
a7f6a5fb
MF		 2143}
2144
2145struct o2hb_heartbeat_group_attribute {
2146 struct configfs_attribute attr;
2147 ssize_t (*show)(struct o2hb_heartbeat_group *, char *);
2148 ssize_t (*store)(struct o2hb_heartbeat_group *, const char *, size_t);
2149};
2150
2151static ssize_t o2hb_heartbeat_group_show(struct config_item *item,
2152 struct configfs_attribute *attr,
2153 char *page)
2154{
2155 struct o2hb_heartbeat_group *reg = to_o2hb_heartbeat_group(to_config_group(item));
2156 struct o2hb_heartbeat_group_attribute *o2hb_heartbeat_group_attr =
2157 container_of(attr, struct o2hb_heartbeat_group_attribute, attr);
2158 ssize_t ret = 0;
2159
2160 if (o2hb_heartbeat_group_attr->show)
2161 ret = o2hb_heartbeat_group_attr->show(reg, page);
2162 return ret;
2163}
2164
2165static ssize_t o2hb_heartbeat_group_store(struct config_item *item,
2166 struct configfs_attribute *attr,
2167 const char *page, size_t count)
2168{
2169 struct o2hb_heartbeat_group *reg = to_o2hb_heartbeat_group(to_config_group(item));
2170 struct o2hb_heartbeat_group_attribute *o2hb_heartbeat_group_attr =
2171 container_of(attr, struct o2hb_heartbeat_group_attribute, attr);
2172 ssize_t ret = -EINVAL;
2173
2174 if (o2hb_heartbeat_group_attr->store)
2175 ret = o2hb_heartbeat_group_attr->store(reg, page, count);
2176 return ret;
2177}
2178
2179static ssize_t o2hb_heartbeat_group_threshold_show(struct o2hb_heartbeat_group *group,
2180 char *page)
2181{
2182 return sprintf(page, "%u\n", o2hb_dead_threshold);
2183}
2184
2185static ssize_t o2hb_heartbeat_group_threshold_store(struct o2hb_heartbeat_group *group,
2186 const char *page,
2187 size_t count)
2188{
2189 unsigned long tmp;
2190 char *p = (char *)page;
2191
2192 tmp = simple_strtoul(p, &p, 10);
2193 if (!p || (*p && (*p != '\n')))
2194 return -EINVAL;
2195
2196 /* this will validate ranges for us. */
2197 o2hb_dead_threshold_set((unsigned int) tmp);
2198
2199 return count;
2200}
2201
54b5187b
SM		 2202static
2203ssize_t o2hb_heartbeat_group_mode_show(struct o2hb_heartbeat_group *group,
2204 char *page)
2205{
2206 return sprintf(page, "%s\n",
2207 o2hb_heartbeat_mode_desc[o2hb_heartbeat_mode]);
2208}
2209
2210static
2211ssize_t o2hb_heartbeat_group_mode_store(struct o2hb_heartbeat_group *group,
2212 const char *page, size_t count)
2213{
2214 unsigned int i;
2215 int ret;
2216 size_t len;
2217
2218 len = (page[count - 1] == '\n') ? count - 1 : count;
2219 if (!len)
2220 return -EINVAL;
2221
2222 for (i = 0; i < O2HB_HEARTBEAT_NUM_MODES; ++i) {
2223 if (strnicmp(page, o2hb_heartbeat_mode_desc[i], len))
2224 continue;
2225
2226 ret = o2hb_global_hearbeat_mode_set(i);
2227 if (!ret)
18c50cb0 2228 printk(KERN_NOTICE "o2hb: Heartbeat mode set to %s\n",
54b5187b
SM		 2229 o2hb_heartbeat_mode_desc[i]);
2230 return count;
2231 }
2232
2233 return -EINVAL;
2234
2235}
2236
a7f6a5fb
MF		 2237static struct o2hb_heartbeat_group_attribute o2hb_heartbeat_group_attr_threshold = {
2238 .attr = { .ca_owner = THIS_MODULE,
2239 .ca_name = "dead_threshold",
2240 .ca_mode = S_IRUGO | S_IWUSR },
2241 .show = o2hb_heartbeat_group_threshold_show,
2242 .store = o2hb_heartbeat_group_threshold_store,
2243};
2244
54b5187b
SM		 2245static struct o2hb_heartbeat_group_attribute o2hb_heartbeat_group_attr_mode = {
2246 .attr = { .ca_owner = THIS_MODULE,
2247 .ca_name = "mode",
2248 .ca_mode = S_IRUGO | S_IWUSR },
2249 .show = o2hb_heartbeat_group_mode_show,
2250 .store = o2hb_heartbeat_group_mode_store,
2251};
2252
a7f6a5fb
MF		 2253static struct configfs_attribute *o2hb_heartbeat_group_attrs[] = {
2254 &o2hb_heartbeat_group_attr_threshold.attr,
54b5187b 2255 &o2hb_heartbeat_group_attr_mode.attr,
a7f6a5fb
MF		 2256 NULL,
2257};
2258
2259static struct configfs_item_operations o2hb_hearbeat_group_item_ops = {
2260 .show_attribute = o2hb_heartbeat_group_show,
2261 .store_attribute = o2hb_heartbeat_group_store,
2262};
2263
2264static struct configfs_group_operations o2hb_heartbeat_group_group_ops = {
2265 .make_item = o2hb_heartbeat_group_make_item,
2266 .drop_item = o2hb_heartbeat_group_drop_item,
2267};
2268
2269static struct config_item_type o2hb_heartbeat_group_type = {
2270 .ct_group_ops = &o2hb_heartbeat_group_group_ops,
2271 .ct_item_ops = &o2hb_hearbeat_group_item_ops,
2272 .ct_attrs = o2hb_heartbeat_group_attrs,
2273 .ct_owner = THIS_MODULE,
2274};
2275
2276/* this is just here to avoid touching group in heartbeat.h which the
2277 * entire damn world #includes */
2278struct config_group *o2hb_alloc_hb_set(void)
2279{
2280 struct o2hb_heartbeat_group *hs = NULL;
2281 struct config_group *ret = NULL;
2282
cd861280 2283 hs = kzalloc(sizeof(struct o2hb_heartbeat_group), GFP_KERNEL);
a7f6a5fb
MF		 2284 if (hs == NULL)
2285 goto out;
2286
2287 config_group_init_type_name(&hs->hs_group, "heartbeat",
2288 &o2hb_heartbeat_group_type);
2289
2290 ret = &hs->hs_group;
2291out:
2292 if (ret == NULL)
2293 kfree(hs);
2294 return ret;
2295}
2296
2297void o2hb_free_hb_set(struct config_group *group)
2298{
2299 struct o2hb_heartbeat_group *hs = to_o2hb_heartbeat_group(group);
2300 kfree(hs);
2301}
2302
25985edc 2303/* hb callback registration and issuing */
a7f6a5fb
MF		 2304
2305static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type)
2306{
2307 if (type == O2HB_NUM_CB)
2308 return ERR_PTR(-EINVAL);
2309
2310 return &o2hb_callbacks[type];
2311}
2312
2313void o2hb_setup_callback(struct o2hb_callback_func *hc,
2314 enum o2hb_callback_type type,
2315 o2hb_cb_func *func,
2316 void *data,
2317 int priority)
2318{
2319 INIT_LIST_HEAD(&hc->hc_item);
2320 hc->hc_func = func;
2321 hc->hc_data = data;
2322 hc->hc_priority = priority;
2323 hc->hc_type = type;
2324 hc->hc_magic = O2HB_CB_MAGIC;
2325}
2326EXPORT_SYMBOL_GPL(o2hb_setup_callback);
2327
58a3158a
SM		 2328/*
2329 * In local heartbeat mode, region_uuid passed matches the dlm domain name.
2330 * In global heartbeat mode, region_uuid passed is NULL.
2331 *
2332 * In local, we only pin the matching region. In global we pin all the active
2333 * regions.
2334 */
2335static int o2hb_region_pin(const char *region_uuid)
14829422 2336{
58a3158a
SM		 2337 int ret = 0, found = 0;
2338 struct o2hb_region *reg;
2339 char *uuid;
14829422
JB		 2340
2341 assert_spin_locked(&o2hb_live_lock);
2342
58a3158a
SM		 2343 list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2344 uuid = config_item_name(&reg->hr_item);
2345
2346 /* local heartbeat */
2347 if (region_uuid) {
2348 if (strcmp(region_uuid, uuid))
2349 continue;
2350 found = 1;
14829422 2351 }
58a3158a
SM		 2352
2353 if (reg->hr_item_pinned || reg->hr_item_dropped)
2354 goto skip_pin;
2355
2356 /* Ignore ENOENT only for local hb (userdlm domain) */
2357 ret = o2nm_depend_item(&reg->hr_item);
2358 if (!ret) {
2359 mlog(ML_CLUSTER, "Pin region %s\n", uuid);
2360 reg->hr_item_pinned = 1;
2361 } else {
2362 if (ret == -ENOENT && found)
2363 ret = 0;
2364 else {
2365 mlog(ML_ERROR, "Pin region %s fails with %d\n",
2366 uuid, ret);
2367 break;
2368 }
14829422 2369 }
58a3158a
SM		 2370skip_pin:
2371 if (found)
2372 break;
14829422
JB		 2373 }
2374
58a3158a 2375 return ret;
14829422
JB		 2376}
2377
58a3158a
SM		 2378/*
2379 * In local heartbeat mode, region_uuid passed matches the dlm domain name.
2380 * In global heartbeat mode, region_uuid passed is NULL.
2381 *
2382 * In local, we only unpin the matching region. In global we unpin all the
2383 * active regions.
2384 */
2385static void o2hb_region_unpin(const char *region_uuid)
14829422 2386{
14829422 2387 struct o2hb_region *reg;
58a3158a
SM		 2388 char *uuid;
2389 int found = 0;
14829422 2390
58a3158a 2391 assert_spin_locked(&o2hb_live_lock);
14829422 2392
58a3158a
SM		 2393 list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2394 uuid = config_item_name(&reg->hr_item);
2395 if (region_uuid) {
2396 if (strcmp(region_uuid, uuid))
2397 continue;
2398 found = 1;
2399 }
14829422 2400
58a3158a
SM		 2401 if (reg->hr_item_pinned) {
2402 mlog(ML_CLUSTER, "Unpin region %s\n", uuid);
2403 o2nm_undepend_item(&reg->hr_item);
2404 reg->hr_item_pinned = 0;
2405 }
2406 if (found)
2407 break;
2408 }
2409}
16c6a4f2 2410
58a3158a
SM		 2411static int o2hb_region_inc_user(const char *region_uuid)
2412{
2413 int ret = 0;
14829422 2414
58a3158a 2415 spin_lock(&o2hb_live_lock);
16c6a4f2 2416
58a3158a
SM		 2417 /* local heartbeat */
2418 if (!o2hb_global_heartbeat_active()) {
2419 ret = o2hb_region_pin(region_uuid);
2420 goto unlock;
2421 }
2422
2423 /*
2424 * if global heartbeat active and this is the first dependent user,
2425 * pin all regions if quorum region count <= CUT_OFF
2426 */
2427 o2hb_dependent_users++;
2428 if (o2hb_dependent_users > 1)
2429 goto unlock;
2430
2431 if (o2hb_pop_count(&o2hb_quorum_region_bitmap,
2432 O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
2433 ret = o2hb_region_pin(NULL);
2434
2435unlock:
2436 spin_unlock(&o2hb_live_lock);
14829422
JB		 2437 return ret;
2438}
2439
58a3158a 2440void o2hb_region_dec_user(const char *region_uuid)
14829422 2441{
14829422
JB		 2442 spin_lock(&o2hb_live_lock);
2443
58a3158a
SM		 2444 /* local heartbeat */
2445 if (!o2hb_global_heartbeat_active()) {
2446 o2hb_region_unpin(region_uuid);
2447 goto unlock;
2448 }
14829422 2449
58a3158a
SM		 2450 /*
2451 * if global heartbeat active and there are no dependent users,
2452 * unpin all quorum regions
2453 */
2454 o2hb_dependent_users--;
2455 if (!o2hb_dependent_users)
2456 o2hb_region_unpin(NULL);
14829422 2457
58a3158a
SM		 2458unlock:
2459 spin_unlock(&o2hb_live_lock);
14829422
JB		 2460}
2461
2462int o2hb_register_callback(const char *region_uuid,
2463 struct o2hb_callback_func *hc)
a7f6a5fb
MF		 2464{
2465 struct o2hb_callback_func *tmp;
2466 struct list_head *iter;
2467 struct o2hb_callback *hbcall;
2468 int ret;
2469
2470 BUG_ON(hc->hc_magic != O2HB_CB_MAGIC);
2471 BUG_ON(!list_empty(&hc->hc_item));
2472
2473 hbcall = hbcall_from_type(hc->hc_type);
2474 if (IS_ERR(hbcall)) {
2475 ret = PTR_ERR(hbcall);
2476 goto out;
2477 }
2478
14829422 2479 if (region_uuid) {
58a3158a
SM		 2480 ret = o2hb_region_inc_user(region_uuid);
2481 if (ret) {
2482 mlog_errno(ret);
14829422 2483 goto out;
58a3158a 2484 }
14829422
JB		 2485 }
2486
a7f6a5fb
MF		 2487 down_write(&o2hb_callback_sem);
2488
2489 list_for_each(iter, &hbcall->list) {
2490 tmp = list_entry(iter, struct o2hb_callback_func, hc_item);
2491 if (hc->hc_priority < tmp->hc_priority) {
2492 list_add_tail(&hc->hc_item, iter);
2493 break;
2494 }
2495 }
2496 if (list_empty(&hc->hc_item))
2497 list_add_tail(&hc->hc_item, &hbcall->list);
2498
2499 up_write(&o2hb_callback_sem);
2500 ret = 0;
2501out:
58a3158a 2502 mlog(ML_CLUSTER, "returning %d on behalf of %p for funcs %p\n",
a7f6a5fb
MF		 2503 ret, __builtin_return_address(0), hc);
2504 return ret;
2505}
2506EXPORT_SYMBOL_GPL(o2hb_register_callback);
2507
14829422
JB		 2508void o2hb_unregister_callback(const char *region_uuid,
2509 struct o2hb_callback_func *hc)
a7f6a5fb
MF		 2510{
2511 BUG_ON(hc->hc_magic != O2HB_CB_MAGIC);
2512
58a3158a 2513 mlog(ML_CLUSTER, "on behalf of %p for funcs %p\n",
a7f6a5fb
MF		 2514 __builtin_return_address(0), hc);
2515
14829422 2516 /* XXX Can this happen _with_ a region reference? */
a7f6a5fb 2517 if (list_empty(&hc->hc_item))
c24f72cc 2518 return;
a7f6a5fb 2519
14829422 2520 if (region_uuid)
58a3158a 2521 o2hb_region_dec_user(region_uuid);
14829422 2522
a7f6a5fb
MF		 2523 down_write(&o2hb_callback_sem);
2524
2525 list_del_init(&hc->hc_item);
2526
2527 up_write(&o2hb_callback_sem);
a7f6a5fb
MF		 2528}
2529EXPORT_SYMBOL_GPL(o2hb_unregister_callback);
2530
2531int o2hb_check_node_heartbeating(u8 node_num)
2532{
2533 unsigned long testing_map[BITS_TO_LONGS(O2NM_MAX_NODES)];
2534
2535 o2hb_fill_node_map(testing_map, sizeof(testing_map));
2536 if (!test_bit(node_num, testing_map)) {
2537 mlog(ML_HEARTBEAT,
2538 "node (%u) does not have heartbeating enabled.\n",
2539 node_num);
2540 return 0;
2541 }
2542
2543 return 1;
2544}
2545EXPORT_SYMBOL_GPL(o2hb_check_node_heartbeating);
2546
2547int o2hb_check_node_heartbeating_from_callback(u8 node_num)
2548{
2549 unsigned long testing_map[BITS_TO_LONGS(O2NM_MAX_NODES)];
2550
2551 o2hb_fill_node_map_from_callback(testing_map, sizeof(testing_map));
2552 if (!test_bit(node_num, testing_map)) {
2553 mlog(ML_HEARTBEAT,
2554 "node (%u) does not have heartbeating enabled.\n",
2555 node_num);
2556 return 0;
2557 }
2558
2559 return 1;
2560}
2561EXPORT_SYMBOL_GPL(o2hb_check_node_heartbeating_from_callback);
2562
2563/* Makes sure our local node is configured with a node number, and is
2564 * heartbeating. */
2565int o2hb_check_local_node_heartbeating(void)
2566{
2567 u8 node_num;
2568
2569 /* if this node was set then we have networking */
2570 node_num = o2nm_this_node();
2571 if (node_num == O2NM_MAX_NODES) {
2572 mlog(ML_HEARTBEAT, "this node has not been configured.\n");
2573 return 0;
2574 }
2575
2576 return o2hb_check_node_heartbeating(node_num);
2577}
2578EXPORT_SYMBOL_GPL(o2hb_check_local_node_heartbeating);
2579
2580/*
2581 * this is just a hack until we get the plumbing which flips file systems
2582 * read only and drops the hb ref instead of killing the node dead.
2583 */
2584void o2hb_stop_all_regions(void)
2585{
2586 struct o2hb_region *reg;
2587
2588 mlog(ML_ERROR, "stopping heartbeat on all active regions.\n");
2589
2590 spin_lock(&o2hb_live_lock);
2591
2592 list_for_each_entry(reg, &o2hb_all_regions, hr_all_item)
2593 reg->hr_unclean_stop = 1;
2594
2595 spin_unlock(&o2hb_live_lock);
2596}
2597EXPORT_SYMBOL_GPL(o2hb_stop_all_regions);
b3c85c4c
SM		 2598
2599int o2hb_get_all_regions(char *region_uuids, u8 max_regions)
2600{
2601 struct o2hb_region *reg;
2602 int numregs = 0;
2603 char *p;
2604
2605 spin_lock(&o2hb_live_lock);
2606
2607 p = region_uuids;
2608 list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2609 mlog(0, "Region: %s\n", config_item_name(&reg->hr_item));
2610 if (numregs < max_regions) {
2611 memcpy(p, config_item_name(&reg->hr_item),
2612 O2HB_MAX_REGION_NAME_LEN);
2613 p += O2HB_MAX_REGION_NAME_LEN;
2614 }
2615 numregs++;
2616 }
2617
2618 spin_unlock(&o2hb_live_lock);
2619
2620 return numregs;
2621}
2622EXPORT_SYMBOL_GPL(o2hb_get_all_regions);
2623
2624int o2hb_global_heartbeat_active(void)
2625{
4d94aa1b 2626 return (o2hb_heartbeat_mode == O2HB_HEARTBEAT_GLOBAL);
b3c85c4c
SM		 2627}
2628EXPORT_SYMBOL(o2hb_global_heartbeat_active);
Linux 6.x block layer and io_uring tree(s)