fs/ocfs2/cluster/heartbeat.c

   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>
  36 #include <linux/debugfs.h>
  37 #include <linux/slab.h>
  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  */
  53 static 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  */
  59 static DEFINE_SPINLOCK(o2hb_live_lock);
  60 static struct list_head o2hb_live_slots[O2NM_MAX_NODES];
  61 static unsigned long o2hb_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
  62 static LIST_HEAD(o2hb_node_events);
  63 static DECLARE_WAIT_QUEUE_HEAD(o2hb_steady_queue);
  64
  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.
  68  *      - o2hb_live_region_bitmap tracks live regions (seen steady iterations).
  69  *      - o2hb_quorum_region_bitmap tracks live regions that have seen all nodes
  70  *              heartbeat on it.
  71  *      - o2hb_failed_region_bitmap tracks the regions that have seen io timeouts.
  72  */
  73 static unsigned long o2hb_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
  74 static unsigned long o2hb_live_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
  75 static unsigned long o2hb_quorum_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
  76 static unsigned long o2hb_failed_region_bitmap[BITS_TO_LONGS(O2NM_MAX_REGIONS)];
  77
  78 #define O2HB_DB_TYPE_LIVENODES          0
  79 #define O2HB_DB_TYPE_LIVEREGIONS        1
  80 #define O2HB_DB_TYPE_QUORUMREGIONS      2
  81 #define O2HB_DB_TYPE_FAILEDREGIONS      3
  82 #define O2HB_DB_TYPE_REGION_LIVENODES   4
  83 #define O2HB_DB_TYPE_REGION_NUMBER      5
  84 #define O2HB_DB_TYPE_REGION_ELAPSED_TIME        6
  85 #define O2HB_DB_TYPE_REGION_PINNED      7
  86 struct o2hb_debug_buf {
  87         int db_type;
  88         int db_size;
  89         int db_len;
  90         void *db_data;
  91 };
  92
  93 static struct o2hb_debug_buf *o2hb_db_livenodes;
  94 static struct o2hb_debug_buf *o2hb_db_liveregions;
  95 static struct o2hb_debug_buf *o2hb_db_quorumregions;
  96 static struct o2hb_debug_buf *o2hb_db_failedregions;
  97
  98 #define O2HB_DEBUG_DIR                  "o2hb"
  99 #define O2HB_DEBUG_LIVENODES            "livenodes"
 100 #define O2HB_DEBUG_LIVEREGIONS          "live_regions"
 101 #define O2HB_DEBUG_QUORUMREGIONS        "quorum_regions"
 102 #define O2HB_DEBUG_FAILEDREGIONS        "failed_regions"
 103 #define O2HB_DEBUG_REGION_NUMBER        "num"
 104 #define O2HB_DEBUG_REGION_ELAPSED_TIME  "elapsed_time_in_ms"
 105 #define O2HB_DEBUG_REGION_PINNED        "pinned"
 106
 107 static struct dentry *o2hb_debug_dir;
 108 static struct dentry *o2hb_debug_livenodes;
 109 static struct dentry *o2hb_debug_liveregions;
 110 static struct dentry *o2hb_debug_quorumregions;
 111 static struct dentry *o2hb_debug_failedregions;
 112
 113 static LIST_HEAD(o2hb_all_regions);
 114
 115 static struct o2hb_callback {
 116         struct list_head list;
 117 } o2hb_callbacks[O2HB_NUM_CB];
 118
 119 static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type);
 120
 121 #define O2HB_DEFAULT_BLOCK_BITS       9
 122
 123 enum o2hb_heartbeat_modes {
 124         O2HB_HEARTBEAT_LOCAL            = 0,
 125         O2HB_HEARTBEAT_GLOBAL,
 126         O2HB_HEARTBEAT_NUM_MODES,
 127 };
 128
 129 char *o2hb_heartbeat_mode_desc[O2HB_HEARTBEAT_NUM_MODES] = {
 130                 "local",        /* O2HB_HEARTBEAT_LOCAL */
 131                 "global",       /* O2HB_HEARTBEAT_GLOBAL */
 132 };
 133
 134 unsigned int o2hb_dead_threshold = O2HB_DEFAULT_DEAD_THRESHOLD;
 135 unsigned int o2hb_heartbeat_mode = O2HB_HEARTBEAT_LOCAL;
 136
 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  */
 143 unsigned 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  */
 161 static int o2hb_region_pin(const char *region_uuid);
 162 static void o2hb_region_unpin(const char *region_uuid);
 163
 164 /* Only sets a new threshold if there are no active regions.
 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. */
 169 static 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
 179 static 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
 195 struct 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
 202 struct 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 */
 214 struct o2hb_region {
 215         struct config_item      hr_item;
 216
 217         struct list_head        hr_all_item;
 218         unsigned                hr_unclean_stop:1,
 219                                 hr_aborted_start:1,
 220                                 hr_item_pinned:1,
 221                                 hr_item_dropped:1;
 222
 223         /* protected by the hr_callback_sem */
 224         struct task_struct      *hr_task;
 225
 226         unsigned int            hr_blocks;
 227         unsigned long long      hr_start_block;
 228
 229         unsigned int            hr_block_bits;
 230         unsigned int            hr_block_bytes;
 231
 232         unsigned int            hr_slots_per_page;
 233         unsigned int            hr_num_pages;
 234
 235         struct page             **hr_slot_data;
 236         struct block_device     *hr_bdev;
 237         struct o2hb_disk_slot   *hr_slots;
 238
 239         /* live node map of this region */
 240         unsigned long           hr_live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
 241         unsigned int            hr_region_num;
 242
 243         struct dentry           *hr_debug_dir;
 244         struct dentry           *hr_debug_livenodes;
 245         struct dentry           *hr_debug_regnum;
 246         struct dentry           *hr_debug_elapsed_time;
 247         struct dentry           *hr_debug_pinned;
 248         struct o2hb_debug_buf   *hr_db_livenodes;
 249         struct o2hb_debug_buf   *hr_db_regnum;
 250         struct o2hb_debug_buf   *hr_db_elapsed_time;
 251         struct o2hb_debug_buf   *hr_db_pinned;
 252
 253         /* let the person setting up hb wait for it to return until it
 254          * has reached a 'steady' state.  This will be fixed when we have
 255          * a more complete api that doesn't lead to this sort of fragility. */
 256         atomic_t                hr_steady_iterations;
 257
 258         /* terminate o2hb thread if it does not reach steady state
 259          * (hr_steady_iterations == 0) within hr_unsteady_iterations */
 260         atomic_t                hr_unsteady_iterations;
 261
 262         char                    hr_dev_name[BDEVNAME_SIZE];
 263
 264         unsigned int            hr_timeout_ms;
 265
 266         /* randomized as the region goes up and down so that a node
 267          * recognizes a node going up and down in one iteration */
 268         u64                     hr_generation;
 269
 270         struct delayed_work     hr_write_timeout_work;
 271         unsigned long           hr_last_timeout_start;
 272
 273         /* Used during o2hb_check_slot to hold a copy of the block
 274          * being checked because we temporarily have to zero out the
 275          * crc field. */
 276         struct o2hb_disk_heartbeat_block *hr_tmp_block;
 277 };
 278
 279 struct o2hb_bio_wait_ctxt {
 280         atomic_t          wc_num_reqs;
 281         struct completion wc_io_complete;
 282         int               wc_error;
 283 };
 284
 285 static int o2hb_pop_count(void *map, int count)
 286 {
 287         int i = -1, pop = 0;
 288
 289         while ((i = find_next_bit(map, count, i + 1)) < count)
 290                 pop++;
 291         return pop;
 292 }
 293
 294 static void o2hb_write_timeout(struct work_struct *work)
 295 {
 296         int failed, quorum;
 297         unsigned long flags;
 298         struct o2hb_region *reg =
 299                 container_of(work, struct o2hb_region,
 300                              hr_write_timeout_work.work);
 301
 302         mlog(ML_ERROR, "Heartbeat write timeout to device %s after %u "
 303              "milliseconds\n", reg->hr_dev_name,
 304              jiffies_to_msecs(jiffies - reg->hr_last_timeout_start));
 305
 306         if (o2hb_global_heartbeat_active()) {
 307                 spin_lock_irqsave(&o2hb_live_lock, flags);
 308                 if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
 309                         set_bit(reg->hr_region_num, o2hb_failed_region_bitmap);
 310                 failed = o2hb_pop_count(&o2hb_failed_region_bitmap,
 311                                         O2NM_MAX_REGIONS);
 312                 quorum = o2hb_pop_count(&o2hb_quorum_region_bitmap,
 313                                         O2NM_MAX_REGIONS);
 314                 spin_unlock_irqrestore(&o2hb_live_lock, flags);
 315
 316                 mlog(ML_HEARTBEAT, "Number of regions %d, failed regions %d\n",
 317                      quorum, failed);
 318
 319                 /*
 320                  * Fence if the number of failed regions >= half the number
 321                  * of  quorum regions
 322                  */
 323                 if ((failed << 1) < quorum)
 324                         return;
 325         }
 326
 327         o2quo_disk_timeout();
 328 }
 329
 330 static void o2hb_arm_write_timeout(struct o2hb_region *reg)
 331 {
 332         /* Arm writeout only after thread reaches steady state */
 333         if (atomic_read(&reg->hr_steady_iterations) != 0)
 334                 return;
 335
 336         mlog(ML_HEARTBEAT, "Queue write timeout for %u ms\n",
 337              O2HB_MAX_WRITE_TIMEOUT_MS);
 338
 339         if (o2hb_global_heartbeat_active()) {
 340                 spin_lock(&o2hb_live_lock);
 341                 clear_bit(reg->hr_region_num, o2hb_failed_region_bitmap);
 342                 spin_unlock(&o2hb_live_lock);
 343         }
 344         cancel_delayed_work(&reg->hr_write_timeout_work);
 345         reg->hr_last_timeout_start = jiffies;
 346         schedule_delayed_work(&reg->hr_write_timeout_work,
 347                               msecs_to_jiffies(O2HB_MAX_WRITE_TIMEOUT_MS));
 348 }
 349
 350 static void o2hb_disarm_write_timeout(struct o2hb_region *reg)
 351 {
 352         cancel_delayed_work_sync(&reg->hr_write_timeout_work);
 353 }
 354
 355 static inline void o2hb_bio_wait_init(struct o2hb_bio_wait_ctxt *wc)
 356 {
 357         atomic_set(&wc->wc_num_reqs, 1);
 358         init_completion(&wc->wc_io_complete);
 359         wc->wc_error = 0;
 360 }
 361
 362 /* Used in error paths too */
 363 static inline void o2hb_bio_wait_dec(struct o2hb_bio_wait_ctxt *wc,
 364                                      unsigned int num)
 365 {
 366         /* sadly atomic_sub_and_test() isn't available on all platforms.  The
 367          * good news is that the fast path only completes one at a time */
 368         while(num--) {
 369                 if (atomic_dec_and_test(&wc->wc_num_reqs)) {
 370                         BUG_ON(num > 0);
 371                         complete(&wc->wc_io_complete);
 372                 }
 373         }
 374 }
 375
 376 static void o2hb_wait_on_io(struct o2hb_region *reg,
 377                             struct o2hb_bio_wait_ctxt *wc)
 378 {
 379         o2hb_bio_wait_dec(wc, 1);
 380         wait_for_completion(&wc->wc_io_complete);
 381 }
 382
 383 static void o2hb_bio_end_io(struct bio *bio,
 384                            int error)
 385 {
 386         struct o2hb_bio_wait_ctxt *wc = bio->bi_private;
 387
 388         if (error) {
 389                 mlog(ML_ERROR, "IO Error %d\n", error);
 390                 wc->wc_error = error;
 391         }
 392
 393         o2hb_bio_wait_dec(wc, 1);
 394         bio_put(bio);
 395 }
 396
 397 /* Setup a Bio to cover I/O against num_slots slots starting at
 398  * start_slot. */
 399 static struct bio *o2hb_setup_one_bio(struct o2hb_region *reg,
 400                                       struct o2hb_bio_wait_ctxt *wc,
 401                                       unsigned int *current_slot,
 402                                       unsigned int max_slots)
 403 {
 404         int len, current_page;
 405         unsigned int vec_len, vec_start;
 406         unsigned int bits = reg->hr_block_bits;
 407         unsigned int spp = reg->hr_slots_per_page;
 408         unsigned int cs = *current_slot;
 409         struct bio *bio;
 410         struct page *page;
 411
 412         /* Testing has shown this allocation to take long enough under
 413          * GFP_KERNEL that the local node can get fenced. It would be
 414          * nicest if we could pre-allocate these bios and avoid this
 415          * all together. */
 416         bio = bio_alloc(GFP_ATOMIC, 16);
 417         if (!bio) {
 418                 mlog(ML_ERROR, "Could not alloc slots BIO!\n");
 419                 bio = ERR_PTR(-ENOMEM);
 420                 goto bail;
 421         }
 422
 423         /* Must put everything in 512 byte sectors for the bio... */
 424         bio->bi_sector = (reg->hr_start_block + cs) << (bits - 9);
 425         bio->bi_bdev = reg->hr_bdev;
 426         bio->bi_private = wc;
 427         bio->bi_end_io = o2hb_bio_end_io;
 428
 429         vec_start = (cs << bits) % PAGE_CACHE_SIZE;
 430         while(cs < max_slots) {
 431                 current_page = cs / spp;
 432                 page = reg->hr_slot_data[current_page];
 433
 434                 vec_len = min(PAGE_CACHE_SIZE - vec_start,
 435                               (max_slots-cs) * (PAGE_CACHE_SIZE/spp) );
 436
 437                 mlog(ML_HB_BIO, "page %d, vec_len = %u, vec_start = %u\n",
 438                      current_page, vec_len, vec_start);
 439
 440                 len = bio_add_page(bio, page, vec_len, vec_start);
 441                 if (len != vec_len) break;
 442
 443                 cs += vec_len / (PAGE_CACHE_SIZE/spp);
 444                 vec_start = 0;
 445         }
 446
 447 bail:
 448         *current_slot = cs;
 449         return bio;
 450 }
 451
 452 static int o2hb_read_slots(struct o2hb_region *reg,
 453                            unsigned int max_slots)
 454 {
 455         unsigned int current_slot=0;
 456         int status;
 457         struct o2hb_bio_wait_ctxt wc;
 458         struct bio *bio;
 459
 460         o2hb_bio_wait_init(&wc);
 461
 462         while(current_slot < max_slots) {
 463                 bio = o2hb_setup_one_bio(reg, &wc, &current_slot, max_slots);
 464                 if (IS_ERR(bio)) {
 465                         status = PTR_ERR(bio);
 466                         mlog_errno(status);
 467                         goto bail_and_wait;
 468                 }
 469
 470                 atomic_inc(&wc.wc_num_reqs);
 471                 submit_bio(READ, bio);
 472         }
 473
 474         status = 0;
 475
 476 bail_and_wait:
 477         o2hb_wait_on_io(reg, &wc);
 478         if (wc.wc_error && !status)
 479                 status = wc.wc_error;
 480
 481         return status;
 482 }
 483
 484 static int o2hb_issue_node_write(struct o2hb_region *reg,
 485                                  struct o2hb_bio_wait_ctxt *write_wc)
 486 {
 487         int status;
 488         unsigned int slot;
 489         struct bio *bio;
 490
 491         o2hb_bio_wait_init(write_wc);
 492
 493         slot = o2nm_this_node();
 494
 495         bio = o2hb_setup_one_bio(reg, write_wc, &slot, slot+1);
 496         if (IS_ERR(bio)) {
 497                 status = PTR_ERR(bio);
 498                 mlog_errno(status);
 499                 goto bail;
 500         }
 501
 502         atomic_inc(&write_wc->wc_num_reqs);
 503         submit_bio(WRITE, bio);
 504
 505         status = 0;
 506 bail:
 507         return status;
 508 }
 509
 510 static u32 o2hb_compute_block_crc_le(struct o2hb_region *reg,
 511                                      struct o2hb_disk_heartbeat_block *hb_block)
 512 {
 513         __le32 old_cksum;
 514         u32 ret;
 515
 516         /* We want to compute the block crc with a 0 value in the
 517          * hb_cksum field. Save it off here and replace after the
 518          * crc. */
 519         old_cksum = hb_block->hb_cksum;
 520         hb_block->hb_cksum = 0;
 521
 522         ret = crc32_le(0, (unsigned char *) hb_block, reg->hr_block_bytes);
 523
 524         hb_block->hb_cksum = old_cksum;
 525
 526         return ret;
 527 }
 528
 529 static void o2hb_dump_slot(struct o2hb_disk_heartbeat_block *hb_block)
 530 {
 531         mlog(ML_ERROR, "Dump slot information: seq = 0x%llx, node = %u, "
 532              "cksum = 0x%x, generation 0x%llx\n",
 533              (long long)le64_to_cpu(hb_block->hb_seq),
 534              hb_block->hb_node, le32_to_cpu(hb_block->hb_cksum),
 535              (long long)le64_to_cpu(hb_block->hb_generation));
 536 }
 537
 538 static int o2hb_verify_crc(struct o2hb_region *reg,
 539                            struct o2hb_disk_heartbeat_block *hb_block)
 540 {
 541         u32 read, computed;
 542
 543         read = le32_to_cpu(hb_block->hb_cksum);
 544         computed = o2hb_compute_block_crc_le(reg, hb_block);
 545
 546         return read == computed;
 547 }
 548
 549 /*
 550  * Compare the slot data with what we wrote in the last iteration.
 551  * If the match fails, print an appropriate error message. This is to
 552  * detect errors like... another node hearting on the same slot,
 553  * flaky device that is losing writes, etc.
 554  * Returns 1 if check succeeds, 0 otherwise.
 555  */
 556 static int o2hb_check_own_slot(struct o2hb_region *reg)
 557 {
 558         struct o2hb_disk_slot *slot;
 559         struct o2hb_disk_heartbeat_block *hb_block;
 560         char *errstr;
 561
 562         slot = &reg->hr_slots[o2nm_this_node()];
 563         /* Don't check on our 1st timestamp */
 564         if (!slot->ds_last_time)
 565                 return 0;
 566
 567         hb_block = slot->ds_raw_block;
 568         if (le64_to_cpu(hb_block->hb_seq) == slot->ds_last_time &&
 569             le64_to_cpu(hb_block->hb_generation) == slot->ds_last_generation &&
 570             hb_block->hb_node == slot->ds_node_num)
 571                 return 1;
 572
 573 #define ERRSTR1         "Another node is heartbeating on device"
 574 #define ERRSTR2         "Heartbeat generation mismatch on device"
 575 #define ERRSTR3         "Heartbeat sequence mismatch on device"
 576
 577         if (hb_block->hb_node != slot->ds_node_num)
 578                 errstr = ERRSTR1;
 579         else if (le64_to_cpu(hb_block->hb_generation) !=
 580                  slot->ds_last_generation)
 581                 errstr = ERRSTR2;
 582         else
 583                 errstr = ERRSTR3;
 584
 585         mlog(ML_ERROR, "%s (%s): expected(%u:0x%llx, 0x%llx), "
 586              "ondisk(%u:0x%llx, 0x%llx)\n", errstr, reg->hr_dev_name,
 587              slot->ds_node_num, (unsigned long long)slot->ds_last_generation,
 588              (unsigned long long)slot->ds_last_time, hb_block->hb_node,
 589              (unsigned long long)le64_to_cpu(hb_block->hb_generation),
 590              (unsigned long long)le64_to_cpu(hb_block->hb_seq));
 591
 592         return 0;
 593 }
 594
 595 static inline void o2hb_prepare_block(struct o2hb_region *reg,
 596                                       u64 generation)
 597 {
 598         int node_num;
 599         u64 cputime;
 600         struct o2hb_disk_slot *slot;
 601         struct o2hb_disk_heartbeat_block *hb_block;
 602
 603         node_num = o2nm_this_node();
 604         slot = &reg->hr_slots[node_num];
 605
 606         hb_block = (struct o2hb_disk_heartbeat_block *)slot->ds_raw_block;
 607         memset(hb_block, 0, reg->hr_block_bytes);
 608         /* TODO: time stuff */
 609         cputime = CURRENT_TIME.tv_sec;
 610         if (!cputime)
 611                 cputime = 1;
 612
 613         hb_block->hb_seq = cpu_to_le64(cputime);
 614         hb_block->hb_node = node_num;
 615         hb_block->hb_generation = cpu_to_le64(generation);
 616         hb_block->hb_dead_ms = cpu_to_le32(o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS);
 617
 618         /* This step must always happen last! */
 619         hb_block->hb_cksum = cpu_to_le32(o2hb_compute_block_crc_le(reg,
 620                                                                    hb_block));
 621
 622         mlog(ML_HB_BIO, "our node generation = 0x%llx, cksum = 0x%x\n",
 623              (long long)generation,
 624              le32_to_cpu(hb_block->hb_cksum));
 625 }
 626
 627 static void o2hb_fire_callbacks(struct o2hb_callback *hbcall,
 628                                 struct o2nm_node *node,
 629                                 int idx)
 630 {
 631         struct list_head *iter;
 632         struct o2hb_callback_func *f;
 633
 634         list_for_each(iter, &hbcall->list) {
 635                 f = list_entry(iter, struct o2hb_callback_func, hc_item);
 636                 mlog(ML_HEARTBEAT, "calling funcs %p\n", f);
 637                 (f->hc_func)(node, idx, f->hc_data);
 638         }
 639 }
 640
 641 /* Will run the list in order until we process the passed event */
 642 static void o2hb_run_event_list(struct o2hb_node_event *queued_event)
 643 {
 644         int empty;
 645         struct o2hb_callback *hbcall;
 646         struct o2hb_node_event *event;
 647
 648         spin_lock(&o2hb_live_lock);
 649         empty = list_empty(&queued_event->hn_item);
 650         spin_unlock(&o2hb_live_lock);
 651         if (empty)
 652                 return;
 653
 654         /* Holding callback sem assures we don't alter the callback
 655          * lists when doing this, and serializes ourselves with other
 656          * processes wanting callbacks. */
 657         down_write(&o2hb_callback_sem);
 658
 659         spin_lock(&o2hb_live_lock);
 660         while (!list_empty(&o2hb_node_events)
 661                && !list_empty(&queued_event->hn_item)) {
 662                 event = list_entry(o2hb_node_events.next,
 663                                    struct o2hb_node_event,
 664                                    hn_item);
 665                 list_del_init(&event->hn_item);
 666                 spin_unlock(&o2hb_live_lock);
 667
 668                 mlog(ML_HEARTBEAT, "Node %s event for %d\n",
 669                      event->hn_event_type == O2HB_NODE_UP_CB ? "UP" : "DOWN",
 670                      event->hn_node_num);
 671
 672                 hbcall = hbcall_from_type(event->hn_event_type);
 673
 674                 /* We should *never* have gotten on to the list with a
 675                  * bad type... This isn't something that we should try
 676                  * to recover from. */
 677                 BUG_ON(IS_ERR(hbcall));
 678
 679                 o2hb_fire_callbacks(hbcall, event->hn_node, event->hn_node_num);
 680
 681                 spin_lock(&o2hb_live_lock);
 682         }
 683         spin_unlock(&o2hb_live_lock);
 684
 685         up_write(&o2hb_callback_sem);
 686 }
 687
 688 static void o2hb_queue_node_event(struct o2hb_node_event *event,
 689                                   enum o2hb_callback_type type,
 690                                   struct o2nm_node *node,
 691                                   int node_num)
 692 {
 693         assert_spin_locked(&o2hb_live_lock);
 694
 695         BUG_ON((!node) && (type != O2HB_NODE_DOWN_CB));
 696
 697         event->hn_event_type = type;
 698         event->hn_node = node;
 699         event->hn_node_num = node_num;
 700
 701         mlog(ML_HEARTBEAT, "Queue node %s event for node %d\n",
 702              type == O2HB_NODE_UP_CB ? "UP" : "DOWN", node_num);
 703
 704         list_add_tail(&event->hn_item, &o2hb_node_events);
 705 }
 706
 707 static void o2hb_shutdown_slot(struct o2hb_disk_slot *slot)
 708 {
 709         struct o2hb_node_event event =
 710                 { .hn_item = LIST_HEAD_INIT(event.hn_item), };
 711         struct o2nm_node *node;
 712
 713         node = o2nm_get_node_by_num(slot->ds_node_num);
 714         if (!node)
 715                 return;
 716
 717         spin_lock(&o2hb_live_lock);
 718         if (!list_empty(&slot->ds_live_item)) {
 719                 mlog(ML_HEARTBEAT, "Shutdown, node %d leaves region\n",
 720                      slot->ds_node_num);
 721
 722                 list_del_init(&slot->ds_live_item);
 723
 724                 if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
 725                         clear_bit(slot->ds_node_num, o2hb_live_node_bitmap);
 726
 727                         o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB, node,
 728                                               slot->ds_node_num);
 729                 }
 730         }
 731         spin_unlock(&o2hb_live_lock);
 732
 733         o2hb_run_event_list(&event);
 734
 735         o2nm_node_put(node);
 736 }
 737
 738 static void o2hb_set_quorum_device(struct o2hb_region *reg)
 739 {
 740         if (!o2hb_global_heartbeat_active())
 741                 return;
 742
 743         /* Prevent race with o2hb_heartbeat_group_drop_item() */
 744         if (kthread_should_stop())
 745                 return;
 746
 747         /* Tag region as quorum only after thread reaches steady state */
 748         if (atomic_read(&reg->hr_steady_iterations) != 0)
 749                 return;
 750
 751         spin_lock(&o2hb_live_lock);
 752
 753         if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
 754                 goto unlock;
 755
 756         /*
 757          * A region can be added to the quorum only when it sees all
 758          * live nodes heartbeat on it. In other words, the region has been
 759          * added to all nodes.
 760          */
 761         if (memcmp(reg->hr_live_node_bitmap, o2hb_live_node_bitmap,
 762                    sizeof(o2hb_live_node_bitmap)))
 763                 goto unlock;
 764
 765         printk(KERN_NOTICE "o2hb: Region %s (%s) is now a quorum device\n",
 766                config_item_name(&reg->hr_item), reg->hr_dev_name);
 767
 768         set_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
 769
 770         /*
 771          * If global heartbeat active, unpin all regions if the
 772          * region count > CUT_OFF
 773          */
 774         if (o2hb_pop_count(&o2hb_quorum_region_bitmap,
 775                            O2NM_MAX_REGIONS) > O2HB_PIN_CUT_OFF)
 776                 o2hb_region_unpin(NULL);
 777 unlock:
 778         spin_unlock(&o2hb_live_lock);
 779 }
 780
 781 static int o2hb_check_slot(struct o2hb_region *reg,
 782                            struct o2hb_disk_slot *slot)
 783 {
 784         int changed = 0, gen_changed = 0;
 785         struct o2hb_node_event event =
 786                 { .hn_item = LIST_HEAD_INIT(event.hn_item), };
 787         struct o2nm_node *node;
 788         struct o2hb_disk_heartbeat_block *hb_block = reg->hr_tmp_block;
 789         u64 cputime;
 790         unsigned int dead_ms = o2hb_dead_threshold * O2HB_REGION_TIMEOUT_MS;
 791         unsigned int slot_dead_ms;
 792         int tmp;
 793
 794         memcpy(hb_block, slot->ds_raw_block, reg->hr_block_bytes);
 795
 796         /*
 797          * If a node is no longer configured but is still in the livemap, we
 798          * may need to clear that bit from the livemap.
 799          */
 800         node = o2nm_get_node_by_num(slot->ds_node_num);
 801         if (!node) {
 802                 spin_lock(&o2hb_live_lock);
 803                 tmp = test_bit(slot->ds_node_num, o2hb_live_node_bitmap);
 804                 spin_unlock(&o2hb_live_lock);
 805                 if (!tmp)
 806                         return 0;
 807         }
 808
 809         if (!o2hb_verify_crc(reg, hb_block)) {
 810                 /* all paths from here will drop o2hb_live_lock for
 811                  * us. */
 812                 spin_lock(&o2hb_live_lock);
 813
 814                 /* Don't print an error on the console in this case -
 815                  * a freshly formatted heartbeat area will not have a
 816                  * crc set on it. */
 817                 if (list_empty(&slot->ds_live_item))
 818                         goto out;
 819
 820                 /* The node is live but pushed out a bad crc. We
 821                  * consider it a transient miss but don't populate any
 822                  * other values as they may be junk. */
 823                 mlog(ML_ERROR, "Node %d has written a bad crc to %s\n",
 824                      slot->ds_node_num, reg->hr_dev_name);
 825                 o2hb_dump_slot(hb_block);
 826
 827                 slot->ds_equal_samples++;
 828                 goto fire_callbacks;
 829         }
 830
 831         /* we don't care if these wrap.. the state transitions below
 832          * clear at the right places */
 833         cputime = le64_to_cpu(hb_block->hb_seq);
 834         if (slot->ds_last_time != cputime)
 835                 slot->ds_changed_samples++;
 836         else
 837                 slot->ds_equal_samples++;
 838         slot->ds_last_time = cputime;
 839
 840         /* The node changed heartbeat generations. We assume this to
 841          * mean it dropped off but came back before we timed out. We
 842          * want to consider it down for the time being but don't want
 843          * to lose any changed_samples state we might build up to
 844          * considering it live again. */
 845         if (slot->ds_last_generation != le64_to_cpu(hb_block->hb_generation)) {
 846                 gen_changed = 1;
 847                 slot->ds_equal_samples = 0;
 848                 mlog(ML_HEARTBEAT, "Node %d changed generation (0x%llx "
 849                      "to 0x%llx)\n", slot->ds_node_num,
 850                      (long long)slot->ds_last_generation,
 851                      (long long)le64_to_cpu(hb_block->hb_generation));
 852         }
 853
 854         slot->ds_last_generation = le64_to_cpu(hb_block->hb_generation);
 855
 856         mlog(ML_HEARTBEAT, "Slot %d gen 0x%llx cksum 0x%x "
 857              "seq %llu last %llu changed %u equal %u\n",
 858              slot->ds_node_num, (long long)slot->ds_last_generation,
 859              le32_to_cpu(hb_block->hb_cksum),
 860              (unsigned long long)le64_to_cpu(hb_block->hb_seq),
 861              (unsigned long long)slot->ds_last_time, slot->ds_changed_samples,
 862              slot->ds_equal_samples);
 863
 864         spin_lock(&o2hb_live_lock);
 865
 866 fire_callbacks:
 867         /* dead nodes only come to life after some number of
 868          * changes at any time during their dead time */
 869         if (list_empty(&slot->ds_live_item) &&
 870             slot->ds_changed_samples >= O2HB_LIVE_THRESHOLD) {
 871                 mlog(ML_HEARTBEAT, "Node %d (id 0x%llx) joined my region\n",
 872                      slot->ds_node_num, (long long)slot->ds_last_generation);
 873
 874                 set_bit(slot->ds_node_num, reg->hr_live_node_bitmap);
 875
 876                 /* first on the list generates a callback */
 877                 if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
 878                         mlog(ML_HEARTBEAT, "o2hb: Add node %d to live nodes "
 879                              "bitmap\n", slot->ds_node_num);
 880                         set_bit(slot->ds_node_num, o2hb_live_node_bitmap);
 881
 882                         o2hb_queue_node_event(&event, O2HB_NODE_UP_CB, node,
 883                                               slot->ds_node_num);
 884
 885                         changed = 1;
 886                 }
 887
 888                 list_add_tail(&slot->ds_live_item,
 889                               &o2hb_live_slots[slot->ds_node_num]);
 890
 891                 slot->ds_equal_samples = 0;
 892
 893                 /* We want to be sure that all nodes agree on the
 894                  * number of milliseconds before a node will be
 895                  * considered dead. The self-fencing timeout is
 896                  * computed from this value, and a discrepancy might
 897                  * result in heartbeat calling a node dead when it
 898                  * hasn't self-fenced yet. */
 899                 slot_dead_ms = le32_to_cpu(hb_block->hb_dead_ms);
 900                 if (slot_dead_ms && slot_dead_ms != dead_ms) {
 901                         /* TODO: Perhaps we can fail the region here. */
 902                         mlog(ML_ERROR, "Node %d on device %s has a dead count "
 903                              "of %u ms, but our count is %u ms.\n"
 904                              "Please double check your configuration values "
 905                              "for 'O2CB_HEARTBEAT_THRESHOLD'\n",
 906                              slot->ds_node_num, reg->hr_dev_name, slot_dead_ms,
 907                              dead_ms);
 908                 }
 909                 goto out;
 910         }
 911
 912         /* if the list is dead, we're done.. */
 913         if (list_empty(&slot->ds_live_item))
 914                 goto out;
 915
 916         /* live nodes only go dead after enough consequtive missed
 917          * samples..  reset the missed counter whenever we see
 918          * activity */
 919         if (slot->ds_equal_samples >= o2hb_dead_threshold || gen_changed) {
 920                 mlog(ML_HEARTBEAT, "Node %d left my region\n",
 921                      slot->ds_node_num);
 922
 923                 clear_bit(slot->ds_node_num, reg->hr_live_node_bitmap);
 924
 925                 /* last off the live_slot generates a callback */
 926                 list_del_init(&slot->ds_live_item);
 927                 if (list_empty(&o2hb_live_slots[slot->ds_node_num])) {
 928                         mlog(ML_HEARTBEAT, "o2hb: Remove node %d from live "
 929                              "nodes bitmap\n", slot->ds_node_num);
 930                         clear_bit(slot->ds_node_num, o2hb_live_node_bitmap);
 931
 932                         /* node can be null */
 933                         o2hb_queue_node_event(&event, O2HB_NODE_DOWN_CB,
 934                                               node, slot->ds_node_num);
 935
 936                         changed = 1;
 937                 }
 938
 939                 /* We don't clear this because the node is still
 940                  * actually writing new blocks. */
 941                 if (!gen_changed)
 942                         slot->ds_changed_samples = 0;
 943                 goto out;
 944         }
 945         if (slot->ds_changed_samples) {
 946                 slot->ds_changed_samples = 0;
 947                 slot->ds_equal_samples = 0;
 948         }
 949 out:
 950         spin_unlock(&o2hb_live_lock);
 951
 952         o2hb_run_event_list(&event);
 953
 954         if (node)
 955                 o2nm_node_put(node);
 956         return changed;
 957 }
 958
 959 /* This could be faster if we just implmented a find_last_bit, but I
 960  * don't think the circumstances warrant it. */
 961 static int o2hb_highest_node(unsigned long *nodes,
 962                              int numbits)
 963 {
 964         int highest, node;
 965
 966         highest = numbits;
 967         node = -1;
 968         while ((node = find_next_bit(nodes, numbits, node + 1)) != -1) {
 969                 if (node >= numbits)
 970                         break;
 971
 972                 highest = node;
 973         }
 974
 975         return highest;
 976 }
 977
 978 static int o2hb_do_disk_heartbeat(struct o2hb_region *reg)
 979 {
 980         int i, ret, highest_node;
 981         int membership_change = 0, own_slot_ok = 0;
 982         unsigned long configured_nodes[BITS_TO_LONGS(O2NM_MAX_NODES)];
 983         unsigned long live_node_bitmap[BITS_TO_LONGS(O2NM_MAX_NODES)];
 984         struct o2hb_bio_wait_ctxt write_wc;
 985
 986         ret = o2nm_configured_node_map(configured_nodes,
 987                                        sizeof(configured_nodes));
 988         if (ret) {
 989                 mlog_errno(ret);
 990                 goto bail;
 991         }
 992
 993         /*
 994          * If a node is not configured but is in the livemap, we still need
 995          * to read the slot so as to be able to remove it from the livemap.
 996          */
 997         o2hb_fill_node_map(live_node_bitmap, sizeof(live_node_bitmap));
 998         i = -1;
 999         while ((i = find_next_bit(live_node_bitmap,
1000                                   O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
1001                 set_bit(i, configured_nodes);
1002         }
1003
1004         highest_node = o2hb_highest_node(configured_nodes, O2NM_MAX_NODES);
1005         if (highest_node >= O2NM_MAX_NODES) {
1006                 mlog(ML_NOTICE, "o2hb: No configured nodes found!\n");
1007                 ret = -EINVAL;
1008                 goto bail;
1009         }
1010
1011         /* No sense in reading the slots of nodes that don't exist
1012          * yet. Of course, if the node definitions have holes in them
1013          * then we're reading an empty slot anyway... Consider this
1014          * best-effort. */
1015         ret = o2hb_read_slots(reg, highest_node + 1);
1016         if (ret < 0) {
1017                 mlog_errno(ret);
1018                 goto bail;
1019         }
1020
1021         /* With an up to date view of the slots, we can check that no
1022          * other node has been improperly configured to heartbeat in
1023          * our slot. */
1024         own_slot_ok = o2hb_check_own_slot(reg);
1025
1026         /* fill in the proper info for our next heartbeat */
1027         o2hb_prepare_block(reg, reg->hr_generation);
1028
1029         ret = o2hb_issue_node_write(reg, &write_wc);
1030         if (ret < 0) {
1031                 mlog_errno(ret);
1032                 goto bail;
1033         }
1034
1035         i = -1;
1036         while((i = find_next_bit(configured_nodes,
1037                                  O2NM_MAX_NODES, i + 1)) < O2NM_MAX_NODES) {
1038                 membership_change |= o2hb_check_slot(reg, &reg->hr_slots[i]);
1039         }
1040
1041         /*
1042          * We have to be sure we've advertised ourselves on disk
1043          * before we can go to steady state.  This ensures that
1044          * people we find in our steady state have seen us.
1045          */
1046         o2hb_wait_on_io(reg, &write_wc);
1047         if (write_wc.wc_error) {
1048                 /* Do not re-arm the write timeout on I/O error - we
1049                  * can't be sure that the new block ever made it to
1050                  * disk */
1051                 mlog(ML_ERROR, "Write error %d on device \"%s\"\n",
1052                      write_wc.wc_error, reg->hr_dev_name);
1053                 ret = write_wc.wc_error;
1054                 goto bail;
1055         }
1056
1057         /* Skip disarming the timeout if own slot has stale/bad data */
1058         if (own_slot_ok) {
1059                 o2hb_set_quorum_device(reg);
1060                 o2hb_arm_write_timeout(reg);
1061         }
1062
1063 bail:
1064         /* let the person who launched us know when things are steady */
1065         if (atomic_read(&reg->hr_steady_iterations) != 0) {
1066                 if (!ret && own_slot_ok && !membership_change) {
1067                         if (atomic_dec_and_test(&reg->hr_steady_iterations))
1068                                 wake_up(&o2hb_steady_queue);
1069                 }
1070         }
1071
1072         if (atomic_read(&reg->hr_steady_iterations) != 0) {
1073                 if (atomic_dec_and_test(&reg->hr_unsteady_iterations)) {
1074                         printk(KERN_NOTICE "o2hb: Unable to stabilize "
1075                                "heartbeart on region %s (%s)\n",
1076                                config_item_name(&reg->hr_item),
1077                                reg->hr_dev_name);
1078                         atomic_set(&reg->hr_steady_iterations, 0);
1079                         reg->hr_aborted_start = 1;
1080                         wake_up(&o2hb_steady_queue);
1081                         ret = -EIO;
1082                 }
1083         }
1084
1085         return ret;
1086 }
1087
1088 /* Subtract b from a, storing the result in a. a *must* have a larger
1089  * value than b. */
1090 static void o2hb_tv_subtract(struct timeval *a,
1091                              struct timeval *b)
1092 {
1093         /* just return 0 when a is after b */
1094         if (a->tv_sec < b->tv_sec ||
1095             (a->tv_sec == b->tv_sec && a->tv_usec < b->tv_usec)) {
1096                 a->tv_sec = 0;
1097                 a->tv_usec = 0;
1098                 return;
1099         }
1100
1101         a->tv_sec -= b->tv_sec;
1102         a->tv_usec -= b->tv_usec;
1103         while ( a->tv_usec < 0 ) {
1104                 a->tv_sec--;
1105                 a->tv_usec += 1000000;
1106         }
1107 }
1108
1109 static unsigned int o2hb_elapsed_msecs(struct timeval *start,
1110                                        struct timeval *end)
1111 {
1112         struct timeval res = *end;
1113
1114         o2hb_tv_subtract(&res, start);
1115
1116         return res.tv_sec * 1000 + res.tv_usec / 1000;
1117 }
1118
1119 /*
1120  * we ride the region ref that the region dir holds.  before the region
1121  * dir is removed and drops it ref it will wait to tear down this
1122  * thread.
1123  */
1124 static int o2hb_thread(void *data)
1125 {
1126         int i, ret;
1127         struct o2hb_region *reg = data;
1128         struct o2hb_bio_wait_ctxt write_wc;
1129         struct timeval before_hb, after_hb;
1130         unsigned int elapsed_msec;
1131
1132         mlog(ML_HEARTBEAT|ML_KTHREAD, "hb thread running\n");
1133
1134         set_user_nice(current, -20);
1135
1136         /* Pin node */
1137         o2nm_depend_this_node();
1138
1139         while (!kthread_should_stop() &&
1140                !reg->hr_unclean_stop && !reg->hr_aborted_start) {
1141                 /* We track the time spent inside
1142                  * o2hb_do_disk_heartbeat so that we avoid more than
1143                  * hr_timeout_ms between disk writes. On busy systems
1144                  * this should result in a heartbeat which is less
1145                  * likely to time itself out. */
1146                 do_gettimeofday(&before_hb);
1147
1148                 ret = o2hb_do_disk_heartbeat(reg);
1149
1150                 do_gettimeofday(&after_hb);
1151                 elapsed_msec = o2hb_elapsed_msecs(&before_hb, &after_hb);
1152
1153                 mlog(ML_HEARTBEAT,
1154                      "start = %lu.%lu, end = %lu.%lu, msec = %u\n",
1155                      before_hb.tv_sec, (unsigned long) before_hb.tv_usec,
1156                      after_hb.tv_sec, (unsigned long) after_hb.tv_usec,
1157                      elapsed_msec);
1158
1159                 if (!kthread_should_stop() &&
1160                     elapsed_msec < reg->hr_timeout_ms) {
1161                         /* the kthread api has blocked signals for us so no
1162                          * need to record the return value. */
1163                         msleep_interruptible(reg->hr_timeout_ms - elapsed_msec);
1164                 }
1165         }
1166
1167         o2hb_disarm_write_timeout(reg);
1168
1169         /* unclean stop is only used in very bad situation */
1170         for(i = 0; !reg->hr_unclean_stop && i < reg->hr_blocks; i++)
1171                 o2hb_shutdown_slot(&reg->hr_slots[i]);
1172
1173         /* Explicit down notification - avoid forcing the other nodes
1174          * to timeout on this region when we could just as easily
1175          * write a clear generation - thus indicating to them that
1176          * this node has left this region.
1177          */
1178         if (!reg->hr_unclean_stop && !reg->hr_aborted_start) {
1179                 o2hb_prepare_block(reg, 0);
1180                 ret = o2hb_issue_node_write(reg, &write_wc);
1181                 if (ret == 0)
1182                         o2hb_wait_on_io(reg, &write_wc);
1183                 else
1184                         mlog_errno(ret);
1185         }
1186
1187         /* Unpin node */
1188         o2nm_undepend_this_node();
1189
1190         mlog(ML_HEARTBEAT|ML_KTHREAD, "o2hb thread exiting\n");
1191
1192         return 0;
1193 }
1194
1195 #ifdef CONFIG_DEBUG_FS
1196 static int o2hb_debug_open(struct inode *inode, struct file *file)
1197 {
1198         struct o2hb_debug_buf *db = inode->i_private;
1199         struct o2hb_region *reg;
1200         unsigned long map[BITS_TO_LONGS(O2NM_MAX_NODES)];
1201         unsigned long lts;
1202         char *buf = NULL;
1203         int i = -1;
1204         int out = 0;
1205
1206         /* max_nodes should be the largest bitmap we pass here */
1207         BUG_ON(sizeof(map) < db->db_size);
1208
1209         buf = kmalloc(PAGE_SIZE, GFP_KERNEL);
1210         if (!buf)
1211                 goto bail;
1212
1213         switch (db->db_type) {
1214         case O2HB_DB_TYPE_LIVENODES:
1215         case O2HB_DB_TYPE_LIVEREGIONS:
1216         case O2HB_DB_TYPE_QUORUMREGIONS:
1217         case O2HB_DB_TYPE_FAILEDREGIONS:
1218                 spin_lock(&o2hb_live_lock);
1219                 memcpy(map, db->db_data, db->db_size);
1220                 spin_unlock(&o2hb_live_lock);
1221                 break;
1222
1223         case O2HB_DB_TYPE_REGION_LIVENODES:
1224                 spin_lock(&o2hb_live_lock);
1225                 reg = (struct o2hb_region *)db->db_data;
1226                 memcpy(map, reg->hr_live_node_bitmap, db->db_size);
1227                 spin_unlock(&o2hb_live_lock);
1228                 break;
1229
1230         case O2HB_DB_TYPE_REGION_NUMBER:
1231                 reg = (struct o2hb_region *)db->db_data;
1232                 out += snprintf(buf + out, PAGE_SIZE - out, "%d\n",
1233                                 reg->hr_region_num);
1234                 goto done;
1235
1236         case O2HB_DB_TYPE_REGION_ELAPSED_TIME:
1237                 reg = (struct o2hb_region *)db->db_data;
1238                 lts = reg->hr_last_timeout_start;
1239                 /* If 0, it has never been set before */
1240                 if (lts)
1241                         lts = jiffies_to_msecs(jiffies - lts);
1242                 out += snprintf(buf + out, PAGE_SIZE - out, "%lu\n", lts);
1243                 goto done;
1244
1245         case O2HB_DB_TYPE_REGION_PINNED:
1246                 reg = (struct o2hb_region *)db->db_data;
1247                 out += snprintf(buf + out, PAGE_SIZE - out, "%u\n",
1248                                 !!reg->hr_item_pinned);
1249                 goto done;
1250
1251         default:
1252                 goto done;
1253         }
1254
1255         while ((i = find_next_bit(map, db->db_len, i + 1)) < db->db_len)
1256                 out += snprintf(buf + out, PAGE_SIZE - out, "%d ", i);
1257         out += snprintf(buf + out, PAGE_SIZE - out, "\n");
1258
1259 done:
1260         i_size_write(inode, out);
1261
1262         file->private_data = buf;
1263
1264         return 0;
1265 bail:
1266         return -ENOMEM;
1267 }
1268
1269 static int o2hb_debug_release(struct inode *inode, struct file *file)
1270 {
1271         kfree(file->private_data);
1272         return 0;
1273 }
1274
1275 static ssize_t o2hb_debug_read(struct file *file, char __user *buf,
1276                                  size_t nbytes, loff_t *ppos)
1277 {
1278         return simple_read_from_buffer(buf, nbytes, ppos, file->private_data,
1279                                        i_size_read(file->f_mapping->host));
1280 }
1281 #else
1282 static int o2hb_debug_open(struct inode *inode, struct file *file)
1283 {
1284         return 0;
1285 }
1286 static int o2hb_debug_release(struct inode *inode, struct file *file)
1287 {
1288         return 0;
1289 }
1290 static ssize_t o2hb_debug_read(struct file *file, char __user *buf,
1291                                size_t nbytes, loff_t *ppos)
1292 {
1293         return 0;
1294 }
1295 #endif  /* CONFIG_DEBUG_FS */
1296
1297 static const struct file_operations o2hb_debug_fops = {
1298         .open =         o2hb_debug_open,
1299         .release =      o2hb_debug_release,
1300         .read =         o2hb_debug_read,
1301         .llseek =       generic_file_llseek,
1302 };
1303
1304 void o2hb_exit(void)
1305 {
1306         kfree(o2hb_db_livenodes);
1307         kfree(o2hb_db_liveregions);
1308         kfree(o2hb_db_quorumregions);
1309         kfree(o2hb_db_failedregions);
1310         debugfs_remove(o2hb_debug_failedregions);
1311         debugfs_remove(o2hb_debug_quorumregions);
1312         debugfs_remove(o2hb_debug_liveregions);
1313         debugfs_remove(o2hb_debug_livenodes);
1314         debugfs_remove(o2hb_debug_dir);
1315 }
1316
1317 static struct dentry *o2hb_debug_create(const char *name, struct dentry *dir,
1318                                         struct o2hb_debug_buf **db, int db_len,
1319                                         int type, int size, int len, void *data)
1320 {
1321         *db = kmalloc(db_len, GFP_KERNEL);
1322         if (!*db)
1323                 return NULL;
1324
1325         (*db)->db_type = type;
1326         (*db)->db_size = size;
1327         (*db)->db_len = len;
1328         (*db)->db_data = data;
1329
1330         return debugfs_create_file(name, S_IFREG|S_IRUSR, dir, *db,
1331                                    &o2hb_debug_fops);
1332 }
1333
1334 static int o2hb_debug_init(void)
1335 {
1336         int ret = -ENOMEM;
1337
1338         o2hb_debug_dir = debugfs_create_dir(O2HB_DEBUG_DIR, NULL);
1339         if (!o2hb_debug_dir) {
1340                 mlog_errno(ret);
1341                 goto bail;
1342         }
1343
1344         o2hb_debug_livenodes = o2hb_debug_create(O2HB_DEBUG_LIVENODES,
1345                                                  o2hb_debug_dir,
1346                                                  &o2hb_db_livenodes,
1347                                                  sizeof(*o2hb_db_livenodes),
1348                                                  O2HB_DB_TYPE_LIVENODES,
1349                                                  sizeof(o2hb_live_node_bitmap),
1350                                                  O2NM_MAX_NODES,
1351                                                  o2hb_live_node_bitmap);
1352         if (!o2hb_debug_livenodes) {
1353                 mlog_errno(ret);
1354                 goto bail;
1355         }
1356
1357         o2hb_debug_liveregions = o2hb_debug_create(O2HB_DEBUG_LIVEREGIONS,
1358                                                    o2hb_debug_dir,
1359                                                    &o2hb_db_liveregions,
1360                                                    sizeof(*o2hb_db_liveregions),
1361                                                    O2HB_DB_TYPE_LIVEREGIONS,
1362                                                    sizeof(o2hb_live_region_bitmap),
1363                                                    O2NM_MAX_REGIONS,
1364                                                    o2hb_live_region_bitmap);
1365         if (!o2hb_debug_liveregions) {
1366                 mlog_errno(ret);
1367                 goto bail;
1368         }
1369
1370         o2hb_debug_quorumregions =
1371                         o2hb_debug_create(O2HB_DEBUG_QUORUMREGIONS,
1372                                           o2hb_debug_dir,
1373                                           &o2hb_db_quorumregions,
1374                                           sizeof(*o2hb_db_quorumregions),
1375                                           O2HB_DB_TYPE_QUORUMREGIONS,
1376                                           sizeof(o2hb_quorum_region_bitmap),
1377                                           O2NM_MAX_REGIONS,
1378                                           o2hb_quorum_region_bitmap);
1379         if (!o2hb_debug_quorumregions) {
1380                 mlog_errno(ret);
1381                 goto bail;
1382         }
1383
1384         o2hb_debug_failedregions =
1385                         o2hb_debug_create(O2HB_DEBUG_FAILEDREGIONS,
1386                                           o2hb_debug_dir,
1387                                           &o2hb_db_failedregions,
1388                                           sizeof(*o2hb_db_failedregions),
1389                                           O2HB_DB_TYPE_FAILEDREGIONS,
1390                                           sizeof(o2hb_failed_region_bitmap),
1391                                           O2NM_MAX_REGIONS,
1392                                           o2hb_failed_region_bitmap);
1393         if (!o2hb_debug_failedregions) {
1394                 mlog_errno(ret);
1395                 goto bail;
1396         }
1397
1398         ret = 0;
1399 bail:
1400         if (ret)
1401                 o2hb_exit();
1402
1403         return ret;
1404 }
1405
1406 int o2hb_init(void)
1407 {
1408         int i;
1409
1410         for (i = 0; i < ARRAY_SIZE(o2hb_callbacks); i++)
1411                 INIT_LIST_HEAD(&o2hb_callbacks[i].list);
1412
1413         for (i = 0; i < ARRAY_SIZE(o2hb_live_slots); i++)
1414                 INIT_LIST_HEAD(&o2hb_live_slots[i]);
1415
1416         INIT_LIST_HEAD(&o2hb_node_events);
1417
1418         memset(o2hb_live_node_bitmap, 0, sizeof(o2hb_live_node_bitmap));
1419         memset(o2hb_region_bitmap, 0, sizeof(o2hb_region_bitmap));
1420         memset(o2hb_live_region_bitmap, 0, sizeof(o2hb_live_region_bitmap));
1421         memset(o2hb_quorum_region_bitmap, 0, sizeof(o2hb_quorum_region_bitmap));
1422         memset(o2hb_failed_region_bitmap, 0, sizeof(o2hb_failed_region_bitmap));
1423
1424         o2hb_dependent_users = 0;
1425
1426         return o2hb_debug_init();
1427 }
1428
1429 /* if we're already in a callback then we're already serialized by the sem */
1430 static void o2hb_fill_node_map_from_callback(unsigned long *map,
1431                                              unsigned bytes)
1432 {
1433         BUG_ON(bytes < (BITS_TO_LONGS(O2NM_MAX_NODES) * sizeof(unsigned long)));
1434
1435         memcpy(map, &o2hb_live_node_bitmap, bytes);
1436 }
1437
1438 /*
1439  * get a map of all nodes that are heartbeating in any regions
1440  */
1441 void o2hb_fill_node_map(unsigned long *map, unsigned bytes)
1442 {
1443         /* callers want to serialize this map and callbacks so that they
1444          * can trust that they don't miss nodes coming to the party */
1445         down_read(&o2hb_callback_sem);
1446         spin_lock(&o2hb_live_lock);
1447         o2hb_fill_node_map_from_callback(map, bytes);
1448         spin_unlock(&o2hb_live_lock);
1449         up_read(&o2hb_callback_sem);
1450 }
1451 EXPORT_SYMBOL_GPL(o2hb_fill_node_map);
1452
1453 /*
1454  * heartbeat configfs bits.  The heartbeat set is a default set under
1455  * the cluster set in nodemanager.c.
1456  */
1457
1458 static struct o2hb_region *to_o2hb_region(struct config_item *item)
1459 {
1460         return item ? container_of(item, struct o2hb_region, hr_item) : NULL;
1461 }
1462
1463 /* drop_item only drops its ref after killing the thread, nothing should
1464  * be using the region anymore.  this has to clean up any state that
1465  * attributes might have built up. */
1466 static void o2hb_region_release(struct config_item *item)
1467 {
1468         int i;
1469         struct page *page;
1470         struct o2hb_region *reg = to_o2hb_region(item);
1471
1472         mlog(ML_HEARTBEAT, "hb region release (%s)\n", reg->hr_dev_name);
1473
1474         if (reg->hr_tmp_block)
1475                 kfree(reg->hr_tmp_block);
1476
1477         if (reg->hr_slot_data) {
1478                 for (i = 0; i < reg->hr_num_pages; i++) {
1479                         page = reg->hr_slot_data[i];
1480                         if (page)
1481                                 __free_page(page);
1482                 }
1483                 kfree(reg->hr_slot_data);
1484         }
1485
1486         if (reg->hr_bdev)
1487                 blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE);
1488
1489         if (reg->hr_slots)
1490                 kfree(reg->hr_slots);
1491
1492         kfree(reg->hr_db_regnum);
1493         kfree(reg->hr_db_livenodes);
1494         debugfs_remove(reg->hr_debug_livenodes);
1495         debugfs_remove(reg->hr_debug_regnum);
1496         debugfs_remove(reg->hr_debug_elapsed_time);
1497         debugfs_remove(reg->hr_debug_pinned);
1498         debugfs_remove(reg->hr_debug_dir);
1499
1500         spin_lock(&o2hb_live_lock);
1501         list_del(&reg->hr_all_item);
1502         spin_unlock(&o2hb_live_lock);
1503
1504         kfree(reg);
1505 }
1506
1507 static int o2hb_read_block_input(struct o2hb_region *reg,
1508                                  const char *page,
1509                                  size_t count,
1510                                  unsigned long *ret_bytes,
1511                                  unsigned int *ret_bits)
1512 {
1513         unsigned long bytes;
1514         char *p = (char *)page;
1515
1516         bytes = simple_strtoul(p, &p, 0);
1517         if (!p || (*p && (*p != '\n')))
1518                 return -EINVAL;
1519
1520         /* Heartbeat and fs min / max block sizes are the same. */
1521         if (bytes > 4096 || bytes < 512)
1522                 return -ERANGE;
1523         if (hweight16(bytes) != 1)
1524                 return -EINVAL;
1525
1526         if (ret_bytes)
1527                 *ret_bytes = bytes;
1528         if (ret_bits)
1529                 *ret_bits = ffs(bytes) - 1;
1530
1531         return 0;
1532 }
1533
1534 static ssize_t o2hb_region_block_bytes_read(struct o2hb_region *reg,
1535                                             char *page)
1536 {
1537         return sprintf(page, "%u\n", reg->hr_block_bytes);
1538 }
1539
1540 static ssize_t o2hb_region_block_bytes_write(struct o2hb_region *reg,
1541                                              const char *page,
1542                                              size_t count)
1543 {
1544         int status;
1545         unsigned long block_bytes;
1546         unsigned int block_bits;
1547
1548         if (reg->hr_bdev)
1549                 return -EINVAL;
1550
1551         status = o2hb_read_block_input(reg, page, count,
1552                                        &block_bytes, &block_bits);
1553         if (status)
1554                 return status;
1555
1556         reg->hr_block_bytes = (unsigned int)block_bytes;
1557         reg->hr_block_bits = block_bits;
1558
1559         return count;
1560 }
1561
1562 static ssize_t o2hb_region_start_block_read(struct o2hb_region *reg,
1563                                             char *page)
1564 {
1565         return sprintf(page, "%llu\n", reg->hr_start_block);
1566 }
1567
1568 static ssize_t o2hb_region_start_block_write(struct o2hb_region *reg,
1569                                              const char *page,
1570                                              size_t count)
1571 {
1572         unsigned long long tmp;
1573         char *p = (char *)page;
1574
1575         if (reg->hr_bdev)
1576                 return -EINVAL;
1577
1578         tmp = simple_strtoull(p, &p, 0);
1579         if (!p || (*p && (*p != '\n')))
1580                 return -EINVAL;
1581
1582         reg->hr_start_block = tmp;
1583
1584         return count;
1585 }
1586
1587 static ssize_t o2hb_region_blocks_read(struct o2hb_region *reg,
1588                                        char *page)
1589 {
1590         return sprintf(page, "%d\n", reg->hr_blocks);
1591 }
1592
1593 static ssize_t o2hb_region_blocks_write(struct o2hb_region *reg,
1594                                         const char *page,
1595                                         size_t count)
1596 {
1597         unsigned long tmp;
1598         char *p = (char *)page;
1599
1600         if (reg->hr_bdev)
1601                 return -EINVAL;
1602
1603         tmp = simple_strtoul(p, &p, 0);
1604         if (!p || (*p && (*p != '\n')))
1605                 return -EINVAL;
1606
1607         if (tmp > O2NM_MAX_NODES || tmp == 0)
1608                 return -ERANGE;
1609
1610         reg->hr_blocks = (unsigned int)tmp;
1611
1612         return count;
1613 }
1614
1615 static ssize_t o2hb_region_dev_read(struct o2hb_region *reg,
1616                                     char *page)
1617 {
1618         unsigned int ret = 0;
1619
1620         if (reg->hr_bdev)
1621                 ret = sprintf(page, "%s\n", reg->hr_dev_name);
1622
1623         return ret;
1624 }
1625
1626 static void o2hb_init_region_params(struct o2hb_region *reg)
1627 {
1628         reg->hr_slots_per_page = PAGE_CACHE_SIZE >> reg->hr_block_bits;
1629         reg->hr_timeout_ms = O2HB_REGION_TIMEOUT_MS;
1630
1631         mlog(ML_HEARTBEAT, "hr_start_block = %llu, hr_blocks = %u\n",
1632              reg->hr_start_block, reg->hr_blocks);
1633         mlog(ML_HEARTBEAT, "hr_block_bytes = %u, hr_block_bits = %u\n",
1634              reg->hr_block_bytes, reg->hr_block_bits);
1635         mlog(ML_HEARTBEAT, "hr_timeout_ms = %u\n", reg->hr_timeout_ms);
1636         mlog(ML_HEARTBEAT, "dead threshold = %u\n", o2hb_dead_threshold);
1637 }
1638
1639 static int o2hb_map_slot_data(struct o2hb_region *reg)
1640 {
1641         int i, j;
1642         unsigned int last_slot;
1643         unsigned int spp = reg->hr_slots_per_page;
1644         struct page *page;
1645         char *raw;
1646         struct o2hb_disk_slot *slot;
1647
1648         reg->hr_tmp_block = kmalloc(reg->hr_block_bytes, GFP_KERNEL);
1649         if (reg->hr_tmp_block == NULL) {
1650                 mlog_errno(-ENOMEM);
1651                 return -ENOMEM;
1652         }
1653
1654         reg->hr_slots = kcalloc(reg->hr_blocks,
1655                                 sizeof(struct o2hb_disk_slot), GFP_KERNEL);
1656         if (reg->hr_slots == NULL) {
1657                 mlog_errno(-ENOMEM);
1658                 return -ENOMEM;
1659         }
1660
1661         for(i = 0; i < reg->hr_blocks; i++) {
1662                 slot = &reg->hr_slots[i];
1663                 slot->ds_node_num = i;
1664                 INIT_LIST_HEAD(&slot->ds_live_item);
1665                 slot->ds_raw_block = NULL;
1666         }
1667
1668         reg->hr_num_pages = (reg->hr_blocks + spp - 1) / spp;
1669         mlog(ML_HEARTBEAT, "Going to require %u pages to cover %u blocks "
1670                            "at %u blocks per page\n",
1671              reg->hr_num_pages, reg->hr_blocks, spp);
1672
1673         reg->hr_slot_data = kcalloc(reg->hr_num_pages, sizeof(struct page *),
1674                                     GFP_KERNEL);
1675         if (!reg->hr_slot_data) {
1676                 mlog_errno(-ENOMEM);
1677                 return -ENOMEM;
1678         }
1679
1680         for(i = 0; i < reg->hr_num_pages; i++) {
1681                 page = alloc_page(GFP_KERNEL);
1682                 if (!page) {
1683                         mlog_errno(-ENOMEM);
1684                         return -ENOMEM;
1685                 }
1686
1687                 reg->hr_slot_data[i] = page;
1688
1689                 last_slot = i * spp;
1690                 raw = page_address(page);
1691                 for (j = 0;
1692                      (j < spp) && ((j + last_slot) < reg->hr_blocks);
1693                      j++) {
1694                         BUG_ON((j + last_slot) >= reg->hr_blocks);
1695
1696                         slot = &reg->hr_slots[j + last_slot];
1697                         slot->ds_raw_block =
1698                                 (struct o2hb_disk_heartbeat_block *) raw;
1699
1700                         raw += reg->hr_block_bytes;
1701                 }
1702         }
1703
1704         return 0;
1705 }
1706
1707 /* Read in all the slots available and populate the tracking
1708  * structures so that we can start with a baseline idea of what's
1709  * there. */
1710 static int o2hb_populate_slot_data(struct o2hb_region *reg)
1711 {
1712         int ret, i;
1713         struct o2hb_disk_slot *slot;
1714         struct o2hb_disk_heartbeat_block *hb_block;
1715
1716         ret = o2hb_read_slots(reg, reg->hr_blocks);
1717         if (ret) {
1718                 mlog_errno(ret);
1719                 goto out;
1720         }
1721
1722         /* We only want to get an idea of the values initially in each
1723          * slot, so we do no verification - o2hb_check_slot will
1724          * actually determine if each configured slot is valid and
1725          * whether any values have changed. */
1726         for(i = 0; i < reg->hr_blocks; i++) {
1727                 slot = &reg->hr_slots[i];
1728                 hb_block = (struct o2hb_disk_heartbeat_block *) slot->ds_raw_block;
1729
1730                 /* Only fill the values that o2hb_check_slot uses to
1731                  * determine changing slots */
1732                 slot->ds_last_time = le64_to_cpu(hb_block->hb_seq);
1733                 slot->ds_last_generation = le64_to_cpu(hb_block->hb_generation);
1734         }
1735
1736 out:
1737         return ret;
1738 }
1739
1740 /* this is acting as commit; we set up all of hr_bdev and hr_task or nothing */
1741 static ssize_t o2hb_region_dev_write(struct o2hb_region *reg,
1742                                      const char *page,
1743                                      size_t count)
1744 {
1745         struct task_struct *hb_task;
1746         long fd;
1747         int sectsize;
1748         char *p = (char *)page;
1749         struct file *filp = NULL;
1750         struct inode *inode = NULL;
1751         ssize_t ret = -EINVAL;
1752         int live_threshold;
1753
1754         if (reg->hr_bdev)
1755                 goto out;
1756
1757         /* We can't heartbeat without having had our node number
1758          * configured yet. */
1759         if (o2nm_this_node() == O2NM_MAX_NODES)
1760                 goto out;
1761
1762         fd = simple_strtol(p, &p, 0);
1763         if (!p || (*p && (*p != '\n')))
1764                 goto out;
1765
1766         if (fd < 0 || fd >= INT_MAX)
1767                 goto out;
1768
1769         filp = fget(fd);
1770         if (filp == NULL)
1771                 goto out;
1772
1773         if (reg->hr_blocks == 0 || reg->hr_start_block == 0 ||
1774             reg->hr_block_bytes == 0)
1775                 goto out;
1776
1777         inode = igrab(filp->f_mapping->host);
1778         if (inode == NULL)
1779                 goto out;
1780
1781         if (!S_ISBLK(inode->i_mode))
1782                 goto out;
1783
1784         reg->hr_bdev = I_BDEV(filp->f_mapping->host);
1785         ret = blkdev_get(reg->hr_bdev, FMODE_WRITE | FMODE_READ, NULL);
1786         if (ret) {
1787                 reg->hr_bdev = NULL;
1788                 goto out;
1789         }
1790         inode = NULL;
1791
1792         bdevname(reg->hr_bdev, reg->hr_dev_name);
1793
1794         sectsize = bdev_logical_block_size(reg->hr_bdev);
1795         if (sectsize != reg->hr_block_bytes) {
1796                 mlog(ML_ERROR,
1797                      "blocksize %u incorrect for device, expected %d",
1798                      reg->hr_block_bytes, sectsize);
1799                 ret = -EINVAL;
1800                 goto out;
1801         }
1802
1803         o2hb_init_region_params(reg);
1804
1805         /* Generation of zero is invalid */
1806         do {
1807                 get_random_bytes(&reg->hr_generation,
1808                                  sizeof(reg->hr_generation));
1809         } while (reg->hr_generation == 0);
1810
1811         ret = o2hb_map_slot_data(reg);
1812         if (ret) {
1813                 mlog_errno(ret);
1814                 goto out;
1815         }
1816
1817         ret = o2hb_populate_slot_data(reg);
1818         if (ret) {
1819                 mlog_errno(ret);
1820                 goto out;
1821         }
1822
1823         INIT_DELAYED_WORK(&reg->hr_write_timeout_work, o2hb_write_timeout);
1824
1825         /*
1826          * A node is considered live after it has beat LIVE_THRESHOLD
1827          * times.  We're not steady until we've given them a chance
1828          * _after_ our first read.
1829          * The default threshold is bare minimum so as to limit the delay
1830          * during mounts. For global heartbeat, the threshold doubled for the
1831          * first region.
1832          */
1833         live_threshold = O2HB_LIVE_THRESHOLD;
1834         if (o2hb_global_heartbeat_active()) {
1835                 spin_lock(&o2hb_live_lock);
1836                 if (o2hb_pop_count(&o2hb_region_bitmap, O2NM_MAX_REGIONS) == 1)
1837                         live_threshold <<= 1;
1838                 spin_unlock(&o2hb_live_lock);
1839         }
1840         ++live_threshold;
1841         atomic_set(&reg->hr_steady_iterations, live_threshold);
1842         /* unsteady_iterations is double the steady_iterations */
1843         atomic_set(&reg->hr_unsteady_iterations, (live_threshold << 1));
1844
1845         hb_task = kthread_run(o2hb_thread, reg, "o2hb-%s",
1846                               reg->hr_item.ci_name);
1847         if (IS_ERR(hb_task)) {
1848                 ret = PTR_ERR(hb_task);
1849                 mlog_errno(ret);
1850                 goto out;
1851         }
1852
1853         spin_lock(&o2hb_live_lock);
1854         reg->hr_task = hb_task;
1855         spin_unlock(&o2hb_live_lock);
1856
1857         ret = wait_event_interruptible(o2hb_steady_queue,
1858                                 atomic_read(&reg->hr_steady_iterations) == 0);
1859         if (ret) {
1860                 atomic_set(&reg->hr_steady_iterations, 0);
1861                 reg->hr_aborted_start = 1;
1862         }
1863
1864         if (reg->hr_aborted_start) {
1865                 ret = -EIO;
1866                 goto out;
1867         }
1868
1869         /* Ok, we were woken.  Make sure it wasn't by drop_item() */
1870         spin_lock(&o2hb_live_lock);
1871         hb_task = reg->hr_task;
1872         if (o2hb_global_heartbeat_active())
1873                 set_bit(reg->hr_region_num, o2hb_live_region_bitmap);
1874         spin_unlock(&o2hb_live_lock);
1875
1876         if (hb_task)
1877                 ret = count;
1878         else
1879                 ret = -EIO;
1880
1881         if (hb_task && o2hb_global_heartbeat_active())
1882                 printk(KERN_NOTICE "o2hb: Heartbeat started on region %s (%s)\n",
1883                        config_item_name(&reg->hr_item), reg->hr_dev_name);
1884
1885 out:
1886         if (filp)
1887                 fput(filp);
1888         if (inode)
1889                 iput(inode);
1890         if (ret < 0) {
1891                 if (reg->hr_bdev) {
1892                         blkdev_put(reg->hr_bdev, FMODE_READ|FMODE_WRITE);
1893                         reg->hr_bdev = NULL;
1894                 }
1895         }
1896         return ret;
1897 }
1898
1899 static ssize_t o2hb_region_pid_read(struct o2hb_region *reg,
1900                                       char *page)
1901 {
1902         pid_t pid = 0;
1903
1904         spin_lock(&o2hb_live_lock);
1905         if (reg->hr_task)
1906                 pid = task_pid_nr(reg->hr_task);
1907         spin_unlock(&o2hb_live_lock);
1908
1909         if (!pid)
1910                 return 0;
1911
1912         return sprintf(page, "%u\n", pid);
1913 }
1914
1915 struct o2hb_region_attribute {
1916         struct configfs_attribute attr;
1917         ssize_t (*show)(struct o2hb_region *, char *);
1918         ssize_t (*store)(struct o2hb_region *, const char *, size_t);
1919 };
1920
1921 static struct o2hb_region_attribute o2hb_region_attr_block_bytes = {
1922         .attr   = { .ca_owner = THIS_MODULE,
1923                     .ca_name = "block_bytes",
1924                     .ca_mode = S_IRUGO | S_IWUSR },
1925         .show   = o2hb_region_block_bytes_read,
1926         .store  = o2hb_region_block_bytes_write,
1927 };
1928
1929 static struct o2hb_region_attribute o2hb_region_attr_start_block = {
1930         .attr   = { .ca_owner = THIS_MODULE,
1931                     .ca_name = "start_block",
1932                     .ca_mode = S_IRUGO | S_IWUSR },
1933         .show   = o2hb_region_start_block_read,
1934         .store  = o2hb_region_start_block_write,
1935 };
1936
1937 static struct o2hb_region_attribute o2hb_region_attr_blocks = {
1938         .attr   = { .ca_owner = THIS_MODULE,
1939                     .ca_name = "blocks",
1940                     .ca_mode = S_IRUGO | S_IWUSR },
1941         .show   = o2hb_region_blocks_read,
1942         .store  = o2hb_region_blocks_write,
1943 };
1944
1945 static struct o2hb_region_attribute o2hb_region_attr_dev = {
1946         .attr   = { .ca_owner = THIS_MODULE,
1947                     .ca_name = "dev",
1948                     .ca_mode = S_IRUGO | S_IWUSR },
1949         .show   = o2hb_region_dev_read,
1950         .store  = o2hb_region_dev_write,
1951 };
1952
1953 static struct o2hb_region_attribute o2hb_region_attr_pid = {
1954        .attr   = { .ca_owner = THIS_MODULE,
1955                    .ca_name = "pid",
1956                    .ca_mode = S_IRUGO | S_IRUSR },
1957        .show   = o2hb_region_pid_read,
1958 };
1959
1960 static struct configfs_attribute *o2hb_region_attrs[] = {
1961         &o2hb_region_attr_block_bytes.attr,
1962         &o2hb_region_attr_start_block.attr,
1963         &o2hb_region_attr_blocks.attr,
1964         &o2hb_region_attr_dev.attr,
1965         &o2hb_region_attr_pid.attr,
1966         NULL,
1967 };
1968
1969 static ssize_t o2hb_region_show(struct config_item *item,
1970                                 struct configfs_attribute *attr,
1971                                 char *page)
1972 {
1973         struct o2hb_region *reg = to_o2hb_region(item);
1974         struct o2hb_region_attribute *o2hb_region_attr =
1975                 container_of(attr, struct o2hb_region_attribute, attr);
1976         ssize_t ret = 0;
1977
1978         if (o2hb_region_attr->show)
1979                 ret = o2hb_region_attr->show(reg, page);
1980         return ret;
1981 }
1982
1983 static ssize_t o2hb_region_store(struct config_item *item,
1984                                  struct configfs_attribute *attr,
1985                                  const char *page, size_t count)
1986 {
1987         struct o2hb_region *reg = to_o2hb_region(item);
1988         struct o2hb_region_attribute *o2hb_region_attr =
1989                 container_of(attr, struct o2hb_region_attribute, attr);
1990         ssize_t ret = -EINVAL;
1991
1992         if (o2hb_region_attr->store)
1993                 ret = o2hb_region_attr->store(reg, page, count);
1994         return ret;
1995 }
1996
1997 static struct configfs_item_operations o2hb_region_item_ops = {
1998         .release                = o2hb_region_release,
1999         .show_attribute         = o2hb_region_show,
2000         .store_attribute        = o2hb_region_store,
2001 };
2002
2003 static struct config_item_type o2hb_region_type = {
2004         .ct_item_ops    = &o2hb_region_item_ops,
2005         .ct_attrs       = o2hb_region_attrs,
2006         .ct_owner       = THIS_MODULE,
2007 };
2008
2009 /* heartbeat set */
2010
2011 struct o2hb_heartbeat_group {
2012         struct config_group hs_group;
2013         /* some stuff? */
2014 };
2015
2016 static struct o2hb_heartbeat_group *to_o2hb_heartbeat_group(struct config_group *group)
2017 {
2018         return group ?
2019                 container_of(group, struct o2hb_heartbeat_group, hs_group)
2020                 : NULL;
2021 }
2022
2023 static int o2hb_debug_region_init(struct o2hb_region *reg, struct dentry *dir)
2024 {
2025         int ret = -ENOMEM;
2026
2027         reg->hr_debug_dir =
2028                 debugfs_create_dir(config_item_name(&reg->hr_item), dir);
2029         if (!reg->hr_debug_dir) {
2030                 mlog_errno(ret);
2031                 goto bail;
2032         }
2033
2034         reg->hr_debug_livenodes =
2035                         o2hb_debug_create(O2HB_DEBUG_LIVENODES,
2036                                           reg->hr_debug_dir,
2037                                           &(reg->hr_db_livenodes),
2038                                           sizeof(*(reg->hr_db_livenodes)),
2039                                           O2HB_DB_TYPE_REGION_LIVENODES,
2040                                           sizeof(reg->hr_live_node_bitmap),
2041                                           O2NM_MAX_NODES, reg);
2042         if (!reg->hr_debug_livenodes) {
2043                 mlog_errno(ret);
2044                 goto bail;
2045         }
2046
2047         reg->hr_debug_regnum =
2048                         o2hb_debug_create(O2HB_DEBUG_REGION_NUMBER,
2049                                           reg->hr_debug_dir,
2050                                           &(reg->hr_db_regnum),
2051                                           sizeof(*(reg->hr_db_regnum)),
2052                                           O2HB_DB_TYPE_REGION_NUMBER,
2053                                           0, O2NM_MAX_NODES, reg);
2054         if (!reg->hr_debug_regnum) {
2055                 mlog_errno(ret);
2056                 goto bail;
2057         }
2058
2059         reg->hr_debug_elapsed_time =
2060                         o2hb_debug_create(O2HB_DEBUG_REGION_ELAPSED_TIME,
2061                                           reg->hr_debug_dir,
2062                                           &(reg->hr_db_elapsed_time),
2063                                           sizeof(*(reg->hr_db_elapsed_time)),
2064                                           O2HB_DB_TYPE_REGION_ELAPSED_TIME,
2065                                           0, 0, reg);
2066         if (!reg->hr_debug_elapsed_time) {
2067                 mlog_errno(ret);
2068                 goto bail;
2069         }
2070
2071         reg->hr_debug_pinned =
2072                         o2hb_debug_create(O2HB_DEBUG_REGION_PINNED,
2073                                           reg->hr_debug_dir,
2074                                           &(reg->hr_db_pinned),
2075                                           sizeof(*(reg->hr_db_pinned)),
2076                                           O2HB_DB_TYPE_REGION_PINNED,
2077                                           0, 0, reg);
2078         if (!reg->hr_debug_pinned) {
2079                 mlog_errno(ret);
2080                 goto bail;
2081         }
2082
2083         ret = 0;
2084 bail:
2085         return ret;
2086 }
2087
2088 static struct config_item *o2hb_heartbeat_group_make_item(struct config_group *group,
2089                                                           const char *name)
2090 {
2091         struct o2hb_region *reg = NULL;
2092         int ret;
2093
2094         reg = kzalloc(sizeof(struct o2hb_region), GFP_KERNEL);
2095         if (reg == NULL)
2096                 return ERR_PTR(-ENOMEM);
2097
2098         if (strlen(name) > O2HB_MAX_REGION_NAME_LEN) {
2099                 ret = -ENAMETOOLONG;
2100                 goto free;
2101         }
2102
2103         spin_lock(&o2hb_live_lock);
2104         reg->hr_region_num = 0;
2105         if (o2hb_global_heartbeat_active()) {
2106                 reg->hr_region_num = find_first_zero_bit(o2hb_region_bitmap,
2107                                                          O2NM_MAX_REGIONS);
2108                 if (reg->hr_region_num >= O2NM_MAX_REGIONS) {
2109                         spin_unlock(&o2hb_live_lock);
2110                         ret = -EFBIG;
2111                         goto free;
2112                 }
2113                 set_bit(reg->hr_region_num, o2hb_region_bitmap);
2114         }
2115         list_add_tail(&reg->hr_all_item, &o2hb_all_regions);
2116         spin_unlock(&o2hb_live_lock);
2117
2118         config_item_init_type_name(&reg->hr_item, name, &o2hb_region_type);
2119
2120         ret = o2hb_debug_region_init(reg, o2hb_debug_dir);
2121         if (ret) {
2122                 config_item_put(&reg->hr_item);
2123                 goto free;
2124         }
2125
2126         return &reg->hr_item;
2127 free:
2128         kfree(reg);
2129         return ERR_PTR(ret);
2130 }
2131
2132 static void o2hb_heartbeat_group_drop_item(struct config_group *group,
2133                                            struct config_item *item)
2134 {
2135         struct task_struct *hb_task;
2136         struct o2hb_region *reg = to_o2hb_region(item);
2137         int quorum_region = 0;
2138
2139         /* stop the thread when the user removes the region dir */
2140         spin_lock(&o2hb_live_lock);
2141         hb_task = reg->hr_task;
2142         reg->hr_task = NULL;
2143         reg->hr_item_dropped = 1;
2144         spin_unlock(&o2hb_live_lock);
2145
2146         if (hb_task)
2147                 kthread_stop(hb_task);
2148
2149         if (o2hb_global_heartbeat_active()) {
2150                 spin_lock(&o2hb_live_lock);
2151                 clear_bit(reg->hr_region_num, o2hb_region_bitmap);
2152                 clear_bit(reg->hr_region_num, o2hb_live_region_bitmap);
2153                 if (test_bit(reg->hr_region_num, o2hb_quorum_region_bitmap))
2154                         quorum_region = 1;
2155                 clear_bit(reg->hr_region_num, o2hb_quorum_region_bitmap);
2156                 spin_unlock(&o2hb_live_lock);
2157                 printk(KERN_NOTICE "o2hb: Heartbeat %s on region %s (%s)\n",
2158                        ((atomic_read(&reg->hr_steady_iterations) == 0) ?
2159                         "stopped" : "start aborted"), config_item_name(item),
2160                        reg->hr_dev_name);
2161         }
2162
2163         /*
2164          * If we're racing a dev_write(), we need to wake them.  They will
2165          * check reg->hr_task
2166          */
2167         if (atomic_read(&reg->hr_steady_iterations) != 0) {
2168                 reg->hr_aborted_start = 1;
2169                 atomic_set(&reg->hr_steady_iterations, 0);
2170                 wake_up(&o2hb_steady_queue);
2171         }
2172
2173         config_item_put(item);
2174
2175         if (!o2hb_global_heartbeat_active() || !quorum_region)
2176                 return;
2177
2178         /*
2179          * If global heartbeat active and there are dependent users,
2180          * pin all regions if quorum region count <= CUT_OFF
2181          */
2182         spin_lock(&o2hb_live_lock);
2183
2184         if (!o2hb_dependent_users)
2185                 goto unlock;
2186
2187         if (o2hb_pop_count(&o2hb_quorum_region_bitmap,
2188                            O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
2189                 o2hb_region_pin(NULL);
2190
2191 unlock:
2192         spin_unlock(&o2hb_live_lock);
2193 }
2194
2195 struct o2hb_heartbeat_group_attribute {
2196         struct configfs_attribute attr;
2197         ssize_t (*show)(struct o2hb_heartbeat_group *, char *);
2198         ssize_t (*store)(struct o2hb_heartbeat_group *, const char *, size_t);
2199 };
2200
2201 static ssize_t o2hb_heartbeat_group_show(struct config_item *item,
2202                                          struct configfs_attribute *attr,
2203                                          char *page)
2204 {
2205         struct o2hb_heartbeat_group *reg = to_o2hb_heartbeat_group(to_config_group(item));
2206         struct o2hb_heartbeat_group_attribute *o2hb_heartbeat_group_attr =
2207                 container_of(attr, struct o2hb_heartbeat_group_attribute, attr);
2208         ssize_t ret = 0;
2209
2210         if (o2hb_heartbeat_group_attr->show)
2211                 ret = o2hb_heartbeat_group_attr->show(reg, page);
2212         return ret;
2213 }
2214
2215 static ssize_t o2hb_heartbeat_group_store(struct config_item *item,
2216                                           struct configfs_attribute *attr,
2217                                           const char *page, size_t count)
2218 {
2219         struct o2hb_heartbeat_group *reg = to_o2hb_heartbeat_group(to_config_group(item));
2220         struct o2hb_heartbeat_group_attribute *o2hb_heartbeat_group_attr =
2221                 container_of(attr, struct o2hb_heartbeat_group_attribute, attr);
2222         ssize_t ret = -EINVAL;
2223
2224         if (o2hb_heartbeat_group_attr->store)
2225                 ret = o2hb_heartbeat_group_attr->store(reg, page, count);
2226         return ret;
2227 }
2228
2229 static ssize_t o2hb_heartbeat_group_threshold_show(struct o2hb_heartbeat_group *group,
2230                                                      char *page)
2231 {
2232         return sprintf(page, "%u\n", o2hb_dead_threshold);
2233 }
2234
2235 static ssize_t o2hb_heartbeat_group_threshold_store(struct o2hb_heartbeat_group *group,
2236                                                     const char *page,
2237                                                     size_t count)
2238 {
2239         unsigned long tmp;
2240         char *p = (char *)page;
2241
2242         tmp = simple_strtoul(p, &p, 10);
2243         if (!p || (*p && (*p != '\n')))
2244                 return -EINVAL;
2245
2246         /* this will validate ranges for us. */
2247         o2hb_dead_threshold_set((unsigned int) tmp);
2248
2249         return count;
2250 }
2251
2252 static
2253 ssize_t o2hb_heartbeat_group_mode_show(struct o2hb_heartbeat_group *group,
2254                                        char *page)
2255 {
2256         return sprintf(page, "%s\n",
2257                        o2hb_heartbeat_mode_desc[o2hb_heartbeat_mode]);
2258 }
2259
2260 static
2261 ssize_t o2hb_heartbeat_group_mode_store(struct o2hb_heartbeat_group *group,
2262                                         const char *page, size_t count)
2263 {
2264         unsigned int i;
2265         int ret;
2266         size_t len;
2267
2268         len = (page[count - 1] == '\n') ? count - 1 : count;
2269         if (!len)
2270                 return -EINVAL;
2271
2272         for (i = 0; i < O2HB_HEARTBEAT_NUM_MODES; ++i) {
2273                 if (strnicmp(page, o2hb_heartbeat_mode_desc[i], len))
2274                         continue;
2275
2276                 ret = o2hb_global_hearbeat_mode_set(i);
2277                 if (!ret)
2278                         printk(KERN_NOTICE "o2hb: Heartbeat mode set to %s\n",
2279                                o2hb_heartbeat_mode_desc[i]);
2280                 return count;
2281         }
2282
2283         return -EINVAL;
2284
2285 }
2286
2287 static struct o2hb_heartbeat_group_attribute o2hb_heartbeat_group_attr_threshold = {
2288         .attr   = { .ca_owner = THIS_MODULE,
2289                     .ca_name = "dead_threshold",
2290                     .ca_mode = S_IRUGO | S_IWUSR },
2291         .show   = o2hb_heartbeat_group_threshold_show,
2292         .store  = o2hb_heartbeat_group_threshold_store,
2293 };
2294
2295 static struct o2hb_heartbeat_group_attribute o2hb_heartbeat_group_attr_mode = {
2296         .attr   = { .ca_owner = THIS_MODULE,
2297                 .ca_name = "mode",
2298                 .ca_mode = S_IRUGO | S_IWUSR },
2299         .show   = o2hb_heartbeat_group_mode_show,
2300         .store  = o2hb_heartbeat_group_mode_store,
2301 };
2302
2303 static struct configfs_attribute *o2hb_heartbeat_group_attrs[] = {
2304         &o2hb_heartbeat_group_attr_threshold.attr,
2305         &o2hb_heartbeat_group_attr_mode.attr,
2306         NULL,
2307 };
2308
2309 static struct configfs_item_operations o2hb_hearbeat_group_item_ops = {
2310         .show_attribute         = o2hb_heartbeat_group_show,
2311         .store_attribute        = o2hb_heartbeat_group_store,
2312 };
2313
2314 static struct configfs_group_operations o2hb_heartbeat_group_group_ops = {
2315         .make_item      = o2hb_heartbeat_group_make_item,
2316         .drop_item      = o2hb_heartbeat_group_drop_item,
2317 };
2318
2319 static struct config_item_type o2hb_heartbeat_group_type = {
2320         .ct_group_ops   = &o2hb_heartbeat_group_group_ops,
2321         .ct_item_ops    = &o2hb_hearbeat_group_item_ops,
2322         .ct_attrs       = o2hb_heartbeat_group_attrs,
2323         .ct_owner       = THIS_MODULE,
2324 };
2325
2326 /* this is just here to avoid touching group in heartbeat.h which the
2327  * entire damn world #includes */
2328 struct config_group *o2hb_alloc_hb_set(void)
2329 {
2330         struct o2hb_heartbeat_group *hs = NULL;
2331         struct config_group *ret = NULL;
2332
2333         hs = kzalloc(sizeof(struct o2hb_heartbeat_group), GFP_KERNEL);
2334         if (hs == NULL)
2335                 goto out;
2336
2337         config_group_init_type_name(&hs->hs_group, "heartbeat",
2338                                     &o2hb_heartbeat_group_type);
2339
2340         ret = &hs->hs_group;
2341 out:
2342         if (ret == NULL)
2343                 kfree(hs);
2344         return ret;
2345 }
2346
2347 void o2hb_free_hb_set(struct config_group *group)
2348 {
2349         struct o2hb_heartbeat_group *hs = to_o2hb_heartbeat_group(group);
2350         kfree(hs);
2351 }
2352
2353 /* hb callback registration and issuing */
2354
2355 static struct o2hb_callback *hbcall_from_type(enum o2hb_callback_type type)
2356 {
2357         if (type == O2HB_NUM_CB)
2358                 return ERR_PTR(-EINVAL);
2359
2360         return &o2hb_callbacks[type];
2361 }
2362
2363 void o2hb_setup_callback(struct o2hb_callback_func *hc,
2364                          enum o2hb_callback_type type,
2365                          o2hb_cb_func *func,
2366                          void *data,
2367                          int priority)
2368 {
2369         INIT_LIST_HEAD(&hc->hc_item);
2370         hc->hc_func = func;
2371         hc->hc_data = data;
2372         hc->hc_priority = priority;
2373         hc->hc_type = type;
2374         hc->hc_magic = O2HB_CB_MAGIC;
2375 }
2376 EXPORT_SYMBOL_GPL(o2hb_setup_callback);
2377
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 pin the matching region. In global we pin all the active
2383  * regions.
2384  */
2385 static int o2hb_region_pin(const char *region_uuid)
2386 {
2387         int ret = 0, found = 0;
2388         struct o2hb_region *reg;
2389         char *uuid;
2390
2391         assert_spin_locked(&o2hb_live_lock);
2392
2393         list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2394                 uuid = config_item_name(&reg->hr_item);
2395
2396                 /* local heartbeat */
2397                 if (region_uuid) {
2398                         if (strcmp(region_uuid, uuid))
2399                                 continue;
2400                         found = 1;
2401                 }
2402
2403                 if (reg->hr_item_pinned || reg->hr_item_dropped)
2404                         goto skip_pin;
2405
2406                 /* Ignore ENOENT only for local hb (userdlm domain) */
2407                 ret = o2nm_depend_item(&reg->hr_item);
2408                 if (!ret) {
2409                         mlog(ML_CLUSTER, "Pin region %s\n", uuid);
2410                         reg->hr_item_pinned = 1;
2411                 } else {
2412                         if (ret == -ENOENT && found)
2413                                 ret = 0;
2414                         else {
2415                                 mlog(ML_ERROR, "Pin region %s fails with %d\n",
2416                                      uuid, ret);
2417                                 break;
2418                         }
2419                 }
2420 skip_pin:
2421                 if (found)
2422                         break;
2423         }
2424
2425         return ret;
2426 }
2427
2428 /*
2429  * In local heartbeat mode, region_uuid passed matches the dlm domain name.
2430  * In global heartbeat mode, region_uuid passed is NULL.
2431  *
2432  * In local, we only unpin the matching region. In global we unpin all the
2433  * active regions.
2434  */
2435 static void o2hb_region_unpin(const char *region_uuid)
2436 {
2437         struct o2hb_region *reg;
2438         char *uuid;
2439         int found = 0;
2440
2441         assert_spin_locked(&o2hb_live_lock);
2442
2443         list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2444                 uuid = config_item_name(&reg->hr_item);
2445                 if (region_uuid) {
2446                         if (strcmp(region_uuid, uuid))
2447                                 continue;
2448                         found = 1;
2449                 }
2450
2451                 if (reg->hr_item_pinned) {
2452                         mlog(ML_CLUSTER, "Unpin region %s\n", uuid);
2453                         o2nm_undepend_item(&reg->hr_item);
2454                         reg->hr_item_pinned = 0;
2455                 }
2456                 if (found)
2457                         break;
2458         }
2459 }
2460
2461 static int o2hb_region_inc_user(const char *region_uuid)
2462 {
2463         int ret = 0;
2464
2465         spin_lock(&o2hb_live_lock);
2466
2467         /* local heartbeat */
2468         if (!o2hb_global_heartbeat_active()) {
2469             ret = o2hb_region_pin(region_uuid);
2470             goto unlock;
2471         }
2472
2473         /*
2474          * if global heartbeat active and this is the first dependent user,
2475          * pin all regions if quorum region count <= CUT_OFF
2476          */
2477         o2hb_dependent_users++;
2478         if (o2hb_dependent_users > 1)
2479                 goto unlock;
2480
2481         if (o2hb_pop_count(&o2hb_quorum_region_bitmap,
2482                            O2NM_MAX_REGIONS) <= O2HB_PIN_CUT_OFF)
2483                 ret = o2hb_region_pin(NULL);
2484
2485 unlock:
2486         spin_unlock(&o2hb_live_lock);
2487         return ret;
2488 }
2489
2490 void o2hb_region_dec_user(const char *region_uuid)
2491 {
2492         spin_lock(&o2hb_live_lock);
2493
2494         /* local heartbeat */
2495         if (!o2hb_global_heartbeat_active()) {
2496             o2hb_region_unpin(region_uuid);
2497             goto unlock;
2498         }
2499
2500         /*
2501          * if global heartbeat active and there are no dependent users,
2502          * unpin all quorum regions
2503          */
2504         o2hb_dependent_users--;
2505         if (!o2hb_dependent_users)
2506                 o2hb_region_unpin(NULL);
2507
2508 unlock:
2509         spin_unlock(&o2hb_live_lock);
2510 }
2511
2512 int o2hb_register_callback(const char *region_uuid,
2513                            struct o2hb_callback_func *hc)
2514 {
2515         struct o2hb_callback_func *tmp;
2516         struct list_head *iter;
2517         struct o2hb_callback *hbcall;
2518         int ret;
2519
2520         BUG_ON(hc->hc_magic != O2HB_CB_MAGIC);
2521         BUG_ON(!list_empty(&hc->hc_item));
2522
2523         hbcall = hbcall_from_type(hc->hc_type);
2524         if (IS_ERR(hbcall)) {
2525                 ret = PTR_ERR(hbcall);
2526                 goto out;
2527         }
2528
2529         if (region_uuid) {
2530                 ret = o2hb_region_inc_user(region_uuid);
2531                 if (ret) {
2532                         mlog_errno(ret);
2533                         goto out;
2534                 }
2535         }
2536
2537         down_write(&o2hb_callback_sem);
2538
2539         list_for_each(iter, &hbcall->list) {
2540                 tmp = list_entry(iter, struct o2hb_callback_func, hc_item);
2541                 if (hc->hc_priority < tmp->hc_priority) {
2542                         list_add_tail(&hc->hc_item, iter);
2543                         break;
2544                 }
2545         }
2546         if (list_empty(&hc->hc_item))
2547                 list_add_tail(&hc->hc_item, &hbcall->list);
2548
2549         up_write(&o2hb_callback_sem);
2550         ret = 0;
2551 out:
2552         mlog(ML_CLUSTER, "returning %d on behalf of %p for funcs %p\n",
2553              ret, __builtin_return_address(0), hc);
2554         return ret;
2555 }
2556 EXPORT_SYMBOL_GPL(o2hb_register_callback);
2557
2558 void o2hb_unregister_callback(const char *region_uuid,
2559                               struct o2hb_callback_func *hc)
2560 {
2561         BUG_ON(hc->hc_magic != O2HB_CB_MAGIC);
2562
2563         mlog(ML_CLUSTER, "on behalf of %p for funcs %p\n",
2564              __builtin_return_address(0), hc);
2565
2566         /* XXX Can this happen _with_ a region reference? */
2567         if (list_empty(&hc->hc_item))
2568                 return;
2569
2570         if (region_uuid)
2571                 o2hb_region_dec_user(region_uuid);
2572
2573         down_write(&o2hb_callback_sem);
2574
2575         list_del_init(&hc->hc_item);
2576
2577         up_write(&o2hb_callback_sem);
2578 }
2579 EXPORT_SYMBOL_GPL(o2hb_unregister_callback);
2580
2581 int o2hb_check_node_heartbeating(u8 node_num)
2582 {
2583         unsigned long testing_map[BITS_TO_LONGS(O2NM_MAX_NODES)];
2584
2585         o2hb_fill_node_map(testing_map, sizeof(testing_map));
2586         if (!test_bit(node_num, testing_map)) {
2587                 mlog(ML_HEARTBEAT,
2588                      "node (%u) does not have heartbeating enabled.\n",
2589                      node_num);
2590                 return 0;
2591         }
2592
2593         return 1;
2594 }
2595 EXPORT_SYMBOL_GPL(o2hb_check_node_heartbeating);
2596
2597 int o2hb_check_node_heartbeating_from_callback(u8 node_num)
2598 {
2599         unsigned long testing_map[BITS_TO_LONGS(O2NM_MAX_NODES)];
2600
2601         o2hb_fill_node_map_from_callback(testing_map, sizeof(testing_map));
2602         if (!test_bit(node_num, testing_map)) {
2603                 mlog(ML_HEARTBEAT,
2604                      "node (%u) does not have heartbeating enabled.\n",
2605                      node_num);
2606                 return 0;
2607         }
2608
2609         return 1;
2610 }
2611 EXPORT_SYMBOL_GPL(o2hb_check_node_heartbeating_from_callback);
2612
2613 /* Makes sure our local node is configured with a node number, and is
2614  * heartbeating. */
2615 int o2hb_check_local_node_heartbeating(void)
2616 {
2617         u8 node_num;
2618
2619         /* if this node was set then we have networking */
2620         node_num = o2nm_this_node();
2621         if (node_num == O2NM_MAX_NODES) {
2622                 mlog(ML_HEARTBEAT, "this node has not been configured.\n");
2623                 return 0;
2624         }
2625
2626         return o2hb_check_node_heartbeating(node_num);
2627 }
2628 EXPORT_SYMBOL_GPL(o2hb_check_local_node_heartbeating);
2629
2630 /*
2631  * this is just a hack until we get the plumbing which flips file systems
2632  * read only and drops the hb ref instead of killing the node dead.
2633  */
2634 void o2hb_stop_all_regions(void)
2635 {
2636         struct o2hb_region *reg;
2637
2638         mlog(ML_ERROR, "stopping heartbeat on all active regions.\n");
2639
2640         spin_lock(&o2hb_live_lock);
2641
2642         list_for_each_entry(reg, &o2hb_all_regions, hr_all_item)
2643                 reg->hr_unclean_stop = 1;
2644
2645         spin_unlock(&o2hb_live_lock);
2646 }
2647 EXPORT_SYMBOL_GPL(o2hb_stop_all_regions);
2648
2649 int o2hb_get_all_regions(char *region_uuids, u8 max_regions)
2650 {
2651         struct o2hb_region *reg;
2652         int numregs = 0;
2653         char *p;
2654
2655         spin_lock(&o2hb_live_lock);
2656
2657         p = region_uuids;
2658         list_for_each_entry(reg, &o2hb_all_regions, hr_all_item) {
2659                 mlog(0, "Region: %s\n", config_item_name(&reg->hr_item));
2660                 if (numregs < max_regions) {
2661                         memcpy(p, config_item_name(&reg->hr_item),
2662                                O2HB_MAX_REGION_NAME_LEN);
2663                         p += O2HB_MAX_REGION_NAME_LEN;
2664                 }
2665                 numregs++;
2666         }
2667
2668         spin_unlock(&o2hb_live_lock);
2669
2670         return numregs;
2671 }
2672 EXPORT_SYMBOL_GPL(o2hb_get_all_regions);
2673
2674 int o2hb_global_heartbeat_active(void)
2675 {
2676         return (o2hb_heartbeat_mode == O2HB_HEARTBEAT_GLOBAL);
2677 }
2678 EXPORT_SYMBOL(o2hb_global_heartbeat_active);