1 // SPDX-License-Identifier: GPL-2.0-only 2 /****************************************************************************** 3 ******************************************************************************* 4 ** 5 ** Copyright (C) 2005-2011 Red Hat, Inc. All rights reserved. 6 ** 7 ** 8 ******************************************************************************* 9 ******************************************************************************/ 10 11 #include "dlm_internal.h" 12 #include "lockspace.h" 13 #include "member.h" 14 #include "recoverd.h" 15 #include "recover.h" 16 #include "rcom.h" 17 #include "config.h" 18 #include "midcomms.h" 19 #include "lowcomms.h" 20 21 int dlm_slots_version(const struct dlm_header *h) 22 { 23 if ((le32_to_cpu(h->h_version) & 0x0000FFFF) < DLM_HEADER_SLOTS) 24 return 0; 25 return 1; 26 } 27 28 void dlm_slot_save(struct dlm_ls *ls, struct dlm_rcom *rc, 29 struct dlm_member *memb) 30 { 31 struct rcom_config *rf = (struct rcom_config *)rc->rc_buf; 32 33 if (!dlm_slots_version(&rc->rc_header)) 34 return; 35 36 memb->slot = le16_to_cpu(rf->rf_our_slot); 37 memb->generation = le32_to_cpu(rf->rf_generation); 38 } 39 40 void dlm_slots_copy_out(struct dlm_ls *ls, struct dlm_rcom *rc) 41 { 42 struct dlm_slot *slot; 43 struct rcom_slot *ro; 44 int i; 45 46 ro = (struct rcom_slot *)(rc->rc_buf + sizeof(struct rcom_config)); 47 48 /* ls_slots array is sparse, but not rcom_slots */ 49 50 for (i = 0; i < ls->ls_slots_size; i++) { 51 slot = &ls->ls_slots[i]; 52 if (!slot->nodeid) 53 continue; 54 ro->ro_nodeid = cpu_to_le32(slot->nodeid); 55 ro->ro_slot = cpu_to_le16(slot->slot); 56 ro++; 57 } 58 } 59 60 #define SLOT_DEBUG_LINE 128 61 62 static void log_slots(struct dlm_ls *ls, uint32_t gen, int num_slots, 63 struct rcom_slot *ro0, struct dlm_slot *array, 64 int array_size) 65 { 66 char line[SLOT_DEBUG_LINE]; 67 int len = SLOT_DEBUG_LINE - 1; 68 int pos = 0; 69 int ret, i; 70 71 memset(line, 0, sizeof(line)); 72 73 if (array) { 74 for (i = 0; i < array_size; i++) { 75 if (!array[i].nodeid) 76 continue; 77 78 ret = snprintf(line + pos, len - pos, " %d:%d", 79 array[i].slot, array[i].nodeid); 80 if (ret >= len - pos) 81 break; 82 pos += ret; 83 } 84 } else if (ro0) { 85 for (i = 0; i < num_slots; i++) { 86 ret = snprintf(line + pos, len - pos, " %d:%d", 87 ro0[i].ro_slot, ro0[i].ro_nodeid); 88 if (ret >= len - pos) 89 break; 90 pos += ret; 91 } 92 } 93 94 log_rinfo(ls, "generation %u slots %d%s", gen, num_slots, line); 95 } 96 97 int dlm_slots_copy_in(struct dlm_ls *ls) 98 { 99 struct dlm_member *memb; 100 struct dlm_rcom *rc = ls->ls_recover_buf; 101 struct rcom_config *rf = (struct rcom_config *)rc->rc_buf; 102 struct rcom_slot *ro0, *ro; 103 int our_nodeid = dlm_our_nodeid(); 104 int i, num_slots; 105 uint32_t gen; 106 107 if (!dlm_slots_version(&rc->rc_header)) 108 return -1; 109 110 gen = le32_to_cpu(rf->rf_generation); 111 if (gen <= ls->ls_generation) { 112 log_error(ls, "dlm_slots_copy_in gen %u old %u", 113 gen, ls->ls_generation); 114 } 115 ls->ls_generation = gen; 116 117 num_slots = le16_to_cpu(rf->rf_num_slots); 118 if (!num_slots) 119 return -1; 120 121 ro0 = (struct rcom_slot *)(rc->rc_buf + sizeof(struct rcom_config)); 122 123 log_slots(ls, gen, num_slots, ro0, NULL, 0); 124 125 list_for_each_entry(memb, &ls->ls_nodes, list) { 126 for (i = 0, ro = ro0; i < num_slots; i++, ro++) { 127 if (le32_to_cpu(ro->ro_nodeid) != memb->nodeid) 128 continue; 129 memb->slot = le16_to_cpu(ro->ro_slot); 130 memb->slot_prev = memb->slot; 131 break; 132 } 133 134 if (memb->nodeid == our_nodeid) { 135 if (ls->ls_slot && ls->ls_slot != memb->slot) { 136 log_error(ls, "dlm_slots_copy_in our slot " 137 "changed %d %d", ls->ls_slot, 138 memb->slot); 139 return -1; 140 } 141 142 if (!ls->ls_slot) 143 ls->ls_slot = memb->slot; 144 } 145 146 if (!memb->slot) { 147 log_error(ls, "dlm_slots_copy_in nodeid %d no slot", 148 memb->nodeid); 149 return -1; 150 } 151 } 152 153 return 0; 154 } 155 156 /* for any nodes that do not support slots, we will not have set memb->slot 157 in wait_status_all(), so memb->slot will remain -1, and we will not 158 assign slots or set ls_num_slots here */ 159 160 int dlm_slots_assign(struct dlm_ls *ls, int *num_slots, int *slots_size, 161 struct dlm_slot **slots_out, uint32_t *gen_out) 162 { 163 struct dlm_member *memb; 164 struct dlm_slot *array; 165 int our_nodeid = dlm_our_nodeid(); 166 int array_size, max_slots, i; 167 int need = 0; 168 int max = 0; 169 int num = 0; 170 uint32_t gen = 0; 171 172 /* our own memb struct will have slot -1 gen 0 */ 173 174 list_for_each_entry(memb, &ls->ls_nodes, list) { 175 if (memb->nodeid == our_nodeid) { 176 memb->slot = ls->ls_slot; 177 memb->generation = ls->ls_generation; 178 break; 179 } 180 } 181 182 list_for_each_entry(memb, &ls->ls_nodes, list) { 183 if (memb->generation > gen) 184 gen = memb->generation; 185 186 /* node doesn't support slots */ 187 188 if (memb->slot == -1) 189 return -1; 190 191 /* node needs a slot assigned */ 192 193 if (!memb->slot) 194 need++; 195 196 /* node has a slot assigned */ 197 198 num++; 199 200 if (!max || max < memb->slot) 201 max = memb->slot; 202 203 /* sanity check, once slot is assigned it shouldn't change */ 204 205 if (memb->slot_prev && memb->slot && memb->slot_prev != memb->slot) { 206 log_error(ls, "nodeid %d slot changed %d %d", 207 memb->nodeid, memb->slot_prev, memb->slot); 208 return -1; 209 } 210 memb->slot_prev = memb->slot; 211 } 212 213 array_size = max + need; 214 array = kcalloc(array_size, sizeof(*array), GFP_NOFS); 215 if (!array) 216 return -ENOMEM; 217 218 num = 0; 219 220 /* fill in slots (offsets) that are used */ 221 222 list_for_each_entry(memb, &ls->ls_nodes, list) { 223 if (!memb->slot) 224 continue; 225 226 if (memb->slot > array_size) { 227 log_error(ls, "invalid slot number %d", memb->slot); 228 kfree(array); 229 return -1; 230 } 231 232 array[memb->slot - 1].nodeid = memb->nodeid; 233 array[memb->slot - 1].slot = memb->slot; 234 num++; 235 } 236 237 /* assign new slots from unused offsets */ 238 239 list_for_each_entry(memb, &ls->ls_nodes, list) { 240 if (memb->slot) 241 continue; 242 243 for (i = 0; i < array_size; i++) { 244 if (array[i].nodeid) 245 continue; 246 247 memb->slot = i + 1; 248 memb->slot_prev = memb->slot; 249 array[i].nodeid = memb->nodeid; 250 array[i].slot = memb->slot; 251 num++; 252 253 if (!ls->ls_slot && memb->nodeid == our_nodeid) 254 ls->ls_slot = memb->slot; 255 break; 256 } 257 258 if (!memb->slot) { 259 log_error(ls, "no free slot found"); 260 kfree(array); 261 return -1; 262 } 263 } 264 265 gen++; 266 267 log_slots(ls, gen, num, NULL, array, array_size); 268 269 max_slots = (DLM_MAX_APP_BUFSIZE - sizeof(struct dlm_rcom) - 270 sizeof(struct rcom_config)) / sizeof(struct rcom_slot); 271 272 if (num > max_slots) { 273 log_error(ls, "num_slots %d exceeds max_slots %d", 274 num, max_slots); 275 kfree(array); 276 return -1; 277 } 278 279 *gen_out = gen; 280 *slots_out = array; 281 *slots_size = array_size; 282 *num_slots = num; 283 return 0; 284 } 285 286 static void add_ordered_member(struct dlm_ls *ls, struct dlm_member *new) 287 { 288 struct dlm_member *memb = NULL; 289 struct list_head *tmp; 290 struct list_head *newlist = &new->list; 291 struct list_head *head = &ls->ls_nodes; 292 293 list_for_each(tmp, head) { 294 memb = list_entry(tmp, struct dlm_member, list); 295 if (new->nodeid < memb->nodeid) 296 break; 297 } 298 299 if (!memb) 300 list_add_tail(newlist, head); 301 else { 302 list_add_tail(newlist, tmp); 303 } 304 } 305 306 static int add_remote_member(int nodeid) 307 { 308 int error; 309 310 if (nodeid == dlm_our_nodeid()) 311 return 0; 312 313 error = dlm_lowcomms_connect_node(nodeid); 314 if (error < 0) 315 return error; 316 317 dlm_midcomms_add_member(nodeid); 318 return 0; 319 } 320 321 static int dlm_add_member(struct dlm_ls *ls, struct dlm_config_node *node) 322 { 323 struct dlm_member *memb; 324 int error; 325 326 memb = kzalloc(sizeof(*memb), GFP_NOFS); 327 if (!memb) 328 return -ENOMEM; 329 330 memb->nodeid = node->nodeid; 331 memb->weight = node->weight; 332 memb->comm_seq = node->comm_seq; 333 334 error = add_remote_member(node->nodeid); 335 if (error < 0) { 336 kfree(memb); 337 return error; 338 } 339 340 add_ordered_member(ls, memb); 341 ls->ls_num_nodes++; 342 return 0; 343 } 344 345 static struct dlm_member *find_memb(struct list_head *head, int nodeid) 346 { 347 struct dlm_member *memb; 348 349 list_for_each_entry(memb, head, list) { 350 if (memb->nodeid == nodeid) 351 return memb; 352 } 353 return NULL; 354 } 355 356 int dlm_is_member(struct dlm_ls *ls, int nodeid) 357 { 358 if (find_memb(&ls->ls_nodes, nodeid)) 359 return 1; 360 return 0; 361 } 362 363 int dlm_is_removed(struct dlm_ls *ls, int nodeid) 364 { 365 WARN_ON_ONCE(!nodeid || nodeid == -1); 366 367 if (find_memb(&ls->ls_nodes_gone, nodeid)) 368 return 1; 369 return 0; 370 } 371 372 static void clear_memb_list(struct list_head *head, 373 void (*after_del)(int nodeid)) 374 { 375 struct dlm_member *memb; 376 377 while (!list_empty(head)) { 378 memb = list_entry(head->next, struct dlm_member, list); 379 list_del(&memb->list); 380 if (after_del) 381 after_del(memb->nodeid); 382 kfree(memb); 383 } 384 } 385 386 static void remove_remote_member(int nodeid) 387 { 388 if (nodeid == dlm_our_nodeid()) 389 return; 390 391 dlm_midcomms_remove_member(nodeid); 392 } 393 394 void dlm_clear_members(struct dlm_ls *ls) 395 { 396 clear_memb_list(&ls->ls_nodes, remove_remote_member); 397 ls->ls_num_nodes = 0; 398 } 399 400 void dlm_clear_members_gone(struct dlm_ls *ls) 401 { 402 clear_memb_list(&ls->ls_nodes_gone, NULL); 403 } 404 405 static void make_member_array(struct dlm_ls *ls) 406 { 407 struct dlm_member *memb; 408 int i, w, x = 0, total = 0, all_zero = 0, *array; 409 410 kfree(ls->ls_node_array); 411 ls->ls_node_array = NULL; 412 413 list_for_each_entry(memb, &ls->ls_nodes, list) { 414 if (memb->weight) 415 total += memb->weight; 416 } 417 418 /* all nodes revert to weight of 1 if all have weight 0 */ 419 420 if (!total) { 421 total = ls->ls_num_nodes; 422 all_zero = 1; 423 } 424 425 ls->ls_total_weight = total; 426 array = kmalloc_array(total, sizeof(*array), GFP_NOFS); 427 if (!array) 428 return; 429 430 list_for_each_entry(memb, &ls->ls_nodes, list) { 431 if (!all_zero && !memb->weight) 432 continue; 433 434 if (all_zero) 435 w = 1; 436 else 437 w = memb->weight; 438 439 DLM_ASSERT(x < total, printk("total %d x %d\n", total, x);); 440 441 for (i = 0; i < w; i++) 442 array[x++] = memb->nodeid; 443 } 444 445 ls->ls_node_array = array; 446 } 447 448 /* send a status request to all members just to establish comms connections */ 449 450 static int ping_members(struct dlm_ls *ls, uint64_t seq) 451 { 452 struct dlm_member *memb; 453 int error = 0; 454 455 list_for_each_entry(memb, &ls->ls_nodes, list) { 456 if (dlm_recovery_stopped(ls)) { 457 error = -EINTR; 458 break; 459 } 460 error = dlm_rcom_status(ls, memb->nodeid, 0, seq); 461 if (error) 462 break; 463 } 464 if (error) 465 log_rinfo(ls, "ping_members aborted %d last nodeid %d", 466 error, ls->ls_recover_nodeid); 467 return error; 468 } 469 470 static void dlm_lsop_recover_prep(struct dlm_ls *ls) 471 { 472 if (!ls->ls_ops || !ls->ls_ops->recover_prep) 473 return; 474 ls->ls_ops->recover_prep(ls->ls_ops_arg); 475 } 476 477 static void dlm_lsop_recover_slot(struct dlm_ls *ls, struct dlm_member *memb, 478 unsigned int release_recover) 479 { 480 struct dlm_slot slot; 481 uint32_t seq; 482 int error; 483 484 if (!ls->ls_ops || !ls->ls_ops->recover_slot) 485 return; 486 487 /* if there is no comms connection with this node 488 or the present comms connection is newer 489 than the one when this member was added, then 490 we consider the node to have failed (versus 491 being removed due to dlm_release_lockspace) */ 492 493 error = dlm_comm_seq(memb->nodeid, &seq, false); 494 495 if (!release_recover && !error && seq == memb->comm_seq) 496 return; 497 498 slot.nodeid = memb->nodeid; 499 slot.slot = memb->slot; 500 501 ls->ls_ops->recover_slot(ls->ls_ops_arg, &slot); 502 } 503 504 void dlm_lsop_recover_done(struct dlm_ls *ls) 505 { 506 struct dlm_member *memb; 507 struct dlm_slot *slots; 508 int i, num; 509 510 if (!ls->ls_ops || !ls->ls_ops->recover_done) 511 return; 512 513 num = ls->ls_num_nodes; 514 slots = kcalloc(num, sizeof(*slots), GFP_KERNEL); 515 if (!slots) 516 return; 517 518 i = 0; 519 list_for_each_entry(memb, &ls->ls_nodes, list) { 520 if (i == num) { 521 log_error(ls, "dlm_lsop_recover_done bad num %d", num); 522 goto out; 523 } 524 slots[i].nodeid = memb->nodeid; 525 slots[i].slot = memb->slot; 526 i++; 527 } 528 529 ls->ls_ops->recover_done(ls->ls_ops_arg, slots, num, 530 ls->ls_slot, ls->ls_generation); 531 out: 532 kfree(slots); 533 } 534 535 static struct dlm_config_node *find_config_node(struct dlm_recover *rv, 536 int nodeid) 537 { 538 int i; 539 540 for (i = 0; i < rv->nodes_count; i++) { 541 if (rv->nodes[i].nodeid == nodeid) 542 return &rv->nodes[i]; 543 } 544 return NULL; 545 } 546 547 int dlm_recover_members(struct dlm_ls *ls, struct dlm_recover *rv, int *neg_out) 548 { 549 struct dlm_member *memb, *safe; 550 struct dlm_config_node *node; 551 int i, error, neg = 0, low = -1; 552 unsigned int release_recover; 553 554 /* previously removed members that we've not finished removing need to 555 * count as a negative change so the "neg" recovery steps will happen 556 * 557 * This functionality must report all member changes to lsops or 558 * midcomms layer and must never return before. 559 */ 560 561 list_for_each_entry(memb, &ls->ls_nodes_gone, list) { 562 log_rinfo(ls, "prev removed member %d", memb->nodeid); 563 neg++; 564 } 565 566 /* move departed members from ls_nodes to ls_nodes_gone */ 567 568 list_for_each_entry_safe(memb, safe, &ls->ls_nodes, list) { 569 node = find_config_node(rv, memb->nodeid); 570 if (!node) { 571 log_error(ls, "remove member %d invalid", 572 memb->nodeid); 573 return -EFAULT; 574 } 575 576 if (!node->new && !node->gone) 577 continue; 578 579 release_recover = 0; 580 581 if (node->gone) { 582 release_recover = node->release_recover; 583 log_rinfo(ls, "remove member %d%s", memb->nodeid, 584 release_recover ? " (release_recover)" : ""); 585 } else { 586 /* removed and re-added */ 587 log_rinfo(ls, "remove member %d comm_seq %u %u", 588 memb->nodeid, memb->comm_seq, node->comm_seq); 589 } 590 591 neg++; 592 list_move(&memb->list, &ls->ls_nodes_gone); 593 remove_remote_member(memb->nodeid); 594 ls->ls_num_nodes--; 595 dlm_lsop_recover_slot(ls, memb, release_recover); 596 } 597 598 /* add new members to ls_nodes */ 599 600 for (i = 0; i < rv->nodes_count; i++) { 601 node = &rv->nodes[i]; 602 if (node->gone) 603 continue; 604 605 if (dlm_is_member(ls, node->nodeid)) 606 continue; 607 error = dlm_add_member(ls, node); 608 if (error) 609 return error; 610 611 log_rinfo(ls, "add member %d", node->nodeid); 612 } 613 614 list_for_each_entry(memb, &ls->ls_nodes, list) { 615 if (low == -1 || memb->nodeid < low) 616 low = memb->nodeid; 617 } 618 ls->ls_low_nodeid = low; 619 620 make_member_array(ls); 621 *neg_out = neg; 622 623 error = ping_members(ls, rv->seq); 624 log_rinfo(ls, "dlm_recover_members %d nodes", ls->ls_num_nodes); 625 return error; 626 } 627 628 /* Userspace guarantees that dlm_ls_stop() has completed on all nodes before 629 dlm_ls_start() is called on any of them to start the new recovery. */ 630 631 int dlm_ls_stop(struct dlm_ls *ls) 632 { 633 int new; 634 635 /* 636 * Prevent dlm_recv from being in the middle of something when we do 637 * the stop. This includes ensuring dlm_recv isn't processing a 638 * recovery message (rcom), while dlm_recoverd is aborting and 639 * resetting things from an in-progress recovery. i.e. we want 640 * dlm_recoverd to abort its recovery without worrying about dlm_recv 641 * processing an rcom at the same time. Stopping dlm_recv also makes 642 * it easy for dlm_receive_message() to check locking stopped and add a 643 * message to the requestqueue without races. 644 */ 645 646 write_lock_bh(&ls->ls_recv_active); 647 648 /* 649 * Abort any recovery that's in progress (see RECOVER_STOP, 650 * dlm_recovery_stopped()) and tell any other threads running in the 651 * dlm to quit any processing (see RUNNING, dlm_locking_stopped()). 652 */ 653 654 spin_lock_bh(&ls->ls_recover_lock); 655 set_bit(LSFL_RECOVER_STOP, &ls->ls_flags); 656 new = test_and_clear_bit(LSFL_RUNNING, &ls->ls_flags); 657 if (new) 658 timer_delete_sync(&ls->ls_scan_timer); 659 ls->ls_recover_seq++; 660 661 /* activate requestqueue and stop processing */ 662 write_lock_bh(&ls->ls_requestqueue_lock); 663 set_bit(LSFL_RECV_MSG_BLOCKED, &ls->ls_flags); 664 write_unlock_bh(&ls->ls_requestqueue_lock); 665 spin_unlock_bh(&ls->ls_recover_lock); 666 667 /* 668 * Let dlm_recv run again, now any normal messages will be saved on the 669 * requestqueue for later. 670 */ 671 672 write_unlock_bh(&ls->ls_recv_active); 673 674 /* 675 * This in_recovery lock does two things: 676 * 1) Keeps this function from returning until all threads are out 677 * of locking routines and locking is truly stopped. 678 * 2) Keeps any new requests from being processed until it's unlocked 679 * when recovery is complete. 680 */ 681 682 if (new) { 683 set_bit(LSFL_RECOVER_DOWN, &ls->ls_flags); 684 wake_up_process(ls->ls_recoverd_task); 685 wait_event(ls->ls_recover_lock_wait, 686 test_bit(LSFL_RECOVER_LOCK, &ls->ls_flags)); 687 } 688 689 /* 690 * The recoverd suspend/resume makes sure that dlm_recoverd (if 691 * running) has noticed RECOVER_STOP above and quit processing the 692 * previous recovery. 693 */ 694 695 dlm_recoverd_suspend(ls); 696 697 spin_lock_bh(&ls->ls_recover_lock); 698 kfree(ls->ls_slots); 699 ls->ls_slots = NULL; 700 ls->ls_num_slots = 0; 701 ls->ls_slots_size = 0; 702 ls->ls_recover_status = 0; 703 spin_unlock_bh(&ls->ls_recover_lock); 704 705 dlm_recoverd_resume(ls); 706 707 if (!ls->ls_recover_begin) 708 ls->ls_recover_begin = jiffies; 709 710 /* call recover_prep ops only once and not multiple times 711 * for each possible dlm_ls_stop() when recovery is already 712 * stopped. 713 * 714 * If we successful was able to clear LSFL_RUNNING bit and 715 * it was set we know it is the first dlm_ls_stop() call. 716 */ 717 if (new) 718 dlm_lsop_recover_prep(ls); 719 720 return 0; 721 } 722 723 int dlm_ls_start(struct dlm_ls *ls) 724 { 725 struct dlm_recover *rv, *rv_old; 726 struct dlm_config_node *nodes = NULL; 727 int error, count; 728 729 rv = kzalloc(sizeof(*rv), GFP_NOFS); 730 if (!rv) 731 return -ENOMEM; 732 733 error = dlm_config_nodes(ls->ls_name, &nodes, &count); 734 if (error < 0) 735 goto fail_rv; 736 737 spin_lock_bh(&ls->ls_recover_lock); 738 739 /* the lockspace needs to be stopped before it can be started */ 740 741 if (!dlm_locking_stopped(ls)) { 742 spin_unlock_bh(&ls->ls_recover_lock); 743 log_error(ls, "start ignored: lockspace running"); 744 error = -EINVAL; 745 goto fail; 746 } 747 748 rv->nodes = nodes; 749 rv->nodes_count = count; 750 rv->seq = ++ls->ls_recover_seq; 751 rv_old = ls->ls_recover_args; 752 ls->ls_recover_args = rv; 753 spin_unlock_bh(&ls->ls_recover_lock); 754 755 if (rv_old) { 756 log_error(ls, "unused recovery %llx %d", 757 (unsigned long long)rv_old->seq, rv_old->nodes_count); 758 kfree(rv_old->nodes); 759 kfree(rv_old); 760 } 761 762 set_bit(LSFL_RECOVER_WORK, &ls->ls_flags); 763 wake_up_process(ls->ls_recoverd_task); 764 return 0; 765 766 fail: 767 kfree(nodes); 768 fail_rv: 769 kfree(rv); 770 return error; 771 } 772 773