1 /****************************************************************************** 2 ******************************************************************************* 3 ** 4 ** Copyright (C) Sistina Software, Inc. 1997-2003 All rights reserved. 5 ** Copyright (C) 2005-2008 Red Hat, Inc. All rights reserved. 6 ** 7 ** This copyrighted material is made available to anyone wishing to use, 8 ** modify, copy, or redistribute it subject to the terms and conditions 9 ** of the GNU General Public License v.2. 10 ** 11 ******************************************************************************* 12 ******************************************************************************/ 13 14 #include "dlm_internal.h" 15 #include "lockspace.h" 16 #include "member.h" 17 #include "lowcomms.h" 18 #include "midcomms.h" 19 #include "rcom.h" 20 #include "recover.h" 21 #include "dir.h" 22 #include "config.h" 23 #include "memory.h" 24 #include "lock.h" 25 #include "util.h" 26 #include "member.h" 27 28 29 static int rcom_response(struct dlm_ls *ls) 30 { 31 return test_bit(LSFL_RCOM_READY, &ls->ls_flags); 32 } 33 34 static int create_rcom(struct dlm_ls *ls, int to_nodeid, int type, int len, 35 struct dlm_rcom **rc_ret, struct dlm_mhandle **mh_ret) 36 { 37 struct dlm_rcom *rc; 38 struct dlm_mhandle *mh; 39 char *mb; 40 int mb_len = sizeof(struct dlm_rcom) + len; 41 42 mh = dlm_lowcomms_get_buffer(to_nodeid, mb_len, GFP_NOFS, &mb); 43 if (!mh) { 44 log_print("create_rcom to %d type %d len %d ENOBUFS", 45 to_nodeid, type, len); 46 return -ENOBUFS; 47 } 48 memset(mb, 0, mb_len); 49 50 rc = (struct dlm_rcom *) mb; 51 52 rc->rc_header.h_version = (DLM_HEADER_MAJOR | DLM_HEADER_MINOR); 53 rc->rc_header.h_lockspace = ls->ls_global_id; 54 rc->rc_header.h_nodeid = dlm_our_nodeid(); 55 rc->rc_header.h_length = mb_len; 56 rc->rc_header.h_cmd = DLM_RCOM; 57 58 rc->rc_type = type; 59 60 spin_lock(&ls->ls_recover_lock); 61 rc->rc_seq = ls->ls_recover_seq; 62 spin_unlock(&ls->ls_recover_lock); 63 64 *mh_ret = mh; 65 *rc_ret = rc; 66 return 0; 67 } 68 69 static void send_rcom(struct dlm_ls *ls, struct dlm_mhandle *mh, 70 struct dlm_rcom *rc) 71 { 72 dlm_rcom_out(rc); 73 dlm_lowcomms_commit_buffer(mh); 74 } 75 76 static void set_rcom_status(struct dlm_ls *ls, struct rcom_status *rs, 77 uint32_t flags) 78 { 79 rs->rs_flags = cpu_to_le32(flags); 80 } 81 82 /* When replying to a status request, a node also sends back its 83 configuration values. The requesting node then checks that the remote 84 node is configured the same way as itself. */ 85 86 static void set_rcom_config(struct dlm_ls *ls, struct rcom_config *rf, 87 uint32_t num_slots) 88 { 89 rf->rf_lvblen = cpu_to_le32(ls->ls_lvblen); 90 rf->rf_lsflags = cpu_to_le32(ls->ls_exflags); 91 92 rf->rf_our_slot = cpu_to_le16(ls->ls_slot); 93 rf->rf_num_slots = cpu_to_le16(num_slots); 94 rf->rf_generation = cpu_to_le32(ls->ls_generation); 95 } 96 97 static int check_rcom_config(struct dlm_ls *ls, struct dlm_rcom *rc, int nodeid) 98 { 99 struct rcom_config *rf = (struct rcom_config *) rc->rc_buf; 100 101 if ((rc->rc_header.h_version & 0xFFFF0000) != DLM_HEADER_MAJOR) { 102 log_error(ls, "version mismatch: %x nodeid %d: %x", 103 DLM_HEADER_MAJOR | DLM_HEADER_MINOR, nodeid, 104 rc->rc_header.h_version); 105 return -EPROTO; 106 } 107 108 if (le32_to_cpu(rf->rf_lvblen) != ls->ls_lvblen || 109 le32_to_cpu(rf->rf_lsflags) != ls->ls_exflags) { 110 log_error(ls, "config mismatch: %d,%x nodeid %d: %d,%x", 111 ls->ls_lvblen, ls->ls_exflags, nodeid, 112 le32_to_cpu(rf->rf_lvblen), 113 le32_to_cpu(rf->rf_lsflags)); 114 return -EPROTO; 115 } 116 return 0; 117 } 118 119 static void allow_sync_reply(struct dlm_ls *ls, uint64_t *new_seq) 120 { 121 spin_lock(&ls->ls_rcom_spin); 122 *new_seq = ++ls->ls_rcom_seq; 123 set_bit(LSFL_RCOM_WAIT, &ls->ls_flags); 124 spin_unlock(&ls->ls_rcom_spin); 125 } 126 127 static void disallow_sync_reply(struct dlm_ls *ls) 128 { 129 spin_lock(&ls->ls_rcom_spin); 130 clear_bit(LSFL_RCOM_WAIT, &ls->ls_flags); 131 clear_bit(LSFL_RCOM_READY, &ls->ls_flags); 132 spin_unlock(&ls->ls_rcom_spin); 133 } 134 135 /* 136 * low nodeid gathers one slot value at a time from each node. 137 * it sets need_slots=0, and saves rf_our_slot returned from each 138 * rcom_config. 139 * 140 * other nodes gather all slot values at once from the low nodeid. 141 * they set need_slots=1, and ignore the rf_our_slot returned from each 142 * rcom_config. they use the rf_num_slots returned from the low 143 * node's rcom_config. 144 */ 145 146 int dlm_rcom_status(struct dlm_ls *ls, int nodeid, uint32_t status_flags) 147 { 148 struct dlm_rcom *rc; 149 struct dlm_mhandle *mh; 150 int error = 0; 151 152 ls->ls_recover_nodeid = nodeid; 153 154 if (nodeid == dlm_our_nodeid()) { 155 rc = ls->ls_recover_buf; 156 rc->rc_result = dlm_recover_status(ls); 157 goto out; 158 } 159 160 error = create_rcom(ls, nodeid, DLM_RCOM_STATUS, 161 sizeof(struct rcom_status), &rc, &mh); 162 if (error) 163 goto out; 164 165 set_rcom_status(ls, (struct rcom_status *)rc->rc_buf, status_flags); 166 167 allow_sync_reply(ls, &rc->rc_id); 168 memset(ls->ls_recover_buf, 0, dlm_config.ci_buffer_size); 169 170 send_rcom(ls, mh, rc); 171 172 error = dlm_wait_function(ls, &rcom_response); 173 disallow_sync_reply(ls); 174 if (error) 175 goto out; 176 177 rc = ls->ls_recover_buf; 178 179 if (rc->rc_result == -ESRCH) { 180 /* we pretend the remote lockspace exists with 0 status */ 181 log_debug(ls, "remote node %d not ready", nodeid); 182 rc->rc_result = 0; 183 error = 0; 184 } else { 185 error = check_rcom_config(ls, rc, nodeid); 186 } 187 188 /* the caller looks at rc_result for the remote recovery status */ 189 out: 190 return error; 191 } 192 193 static void receive_rcom_status(struct dlm_ls *ls, struct dlm_rcom *rc_in) 194 { 195 struct dlm_rcom *rc; 196 struct dlm_mhandle *mh; 197 struct rcom_status *rs; 198 uint32_t status; 199 int nodeid = rc_in->rc_header.h_nodeid; 200 int len = sizeof(struct rcom_config); 201 int num_slots = 0; 202 int error; 203 204 if (!dlm_slots_version(&rc_in->rc_header)) { 205 status = dlm_recover_status(ls); 206 goto do_create; 207 } 208 209 rs = (struct rcom_status *)rc_in->rc_buf; 210 211 if (!(rs->rs_flags & DLM_RSF_NEED_SLOTS)) { 212 status = dlm_recover_status(ls); 213 goto do_create; 214 } 215 216 spin_lock(&ls->ls_recover_lock); 217 status = ls->ls_recover_status; 218 num_slots = ls->ls_num_slots; 219 spin_unlock(&ls->ls_recover_lock); 220 len += num_slots * sizeof(struct rcom_slot); 221 222 do_create: 223 error = create_rcom(ls, nodeid, DLM_RCOM_STATUS_REPLY, 224 len, &rc, &mh); 225 if (error) 226 return; 227 228 rc->rc_id = rc_in->rc_id; 229 rc->rc_seq_reply = rc_in->rc_seq; 230 rc->rc_result = status; 231 232 set_rcom_config(ls, (struct rcom_config *)rc->rc_buf, num_slots); 233 234 if (!num_slots) 235 goto do_send; 236 237 spin_lock(&ls->ls_recover_lock); 238 if (ls->ls_num_slots != num_slots) { 239 spin_unlock(&ls->ls_recover_lock); 240 log_debug(ls, "receive_rcom_status num_slots %d to %d", 241 num_slots, ls->ls_num_slots); 242 rc->rc_result = 0; 243 set_rcom_config(ls, (struct rcom_config *)rc->rc_buf, 0); 244 goto do_send; 245 } 246 247 dlm_slots_copy_out(ls, rc); 248 spin_unlock(&ls->ls_recover_lock); 249 250 do_send: 251 send_rcom(ls, mh, rc); 252 } 253 254 static void receive_sync_reply(struct dlm_ls *ls, struct dlm_rcom *rc_in) 255 { 256 spin_lock(&ls->ls_rcom_spin); 257 if (!test_bit(LSFL_RCOM_WAIT, &ls->ls_flags) || 258 rc_in->rc_id != ls->ls_rcom_seq) { 259 log_debug(ls, "reject reply %d from %d seq %llx expect %llx", 260 rc_in->rc_type, rc_in->rc_header.h_nodeid, 261 (unsigned long long)rc_in->rc_id, 262 (unsigned long long)ls->ls_rcom_seq); 263 goto out; 264 } 265 memcpy(ls->ls_recover_buf, rc_in, rc_in->rc_header.h_length); 266 set_bit(LSFL_RCOM_READY, &ls->ls_flags); 267 clear_bit(LSFL_RCOM_WAIT, &ls->ls_flags); 268 wake_up(&ls->ls_wait_general); 269 out: 270 spin_unlock(&ls->ls_rcom_spin); 271 } 272 273 int dlm_rcom_names(struct dlm_ls *ls, int nodeid, char *last_name, int last_len) 274 { 275 struct dlm_rcom *rc; 276 struct dlm_mhandle *mh; 277 int error = 0; 278 int max_size = dlm_config.ci_buffer_size - sizeof(struct dlm_rcom); 279 280 ls->ls_recover_nodeid = nodeid; 281 282 if (nodeid == dlm_our_nodeid()) { 283 ls->ls_recover_buf->rc_header.h_length = 284 dlm_config.ci_buffer_size; 285 dlm_copy_master_names(ls, last_name, last_len, 286 ls->ls_recover_buf->rc_buf, 287 max_size, nodeid); 288 goto out; 289 } 290 291 error = create_rcom(ls, nodeid, DLM_RCOM_NAMES, last_len, &rc, &mh); 292 if (error) 293 goto out; 294 memcpy(rc->rc_buf, last_name, last_len); 295 296 allow_sync_reply(ls, &rc->rc_id); 297 memset(ls->ls_recover_buf, 0, dlm_config.ci_buffer_size); 298 299 send_rcom(ls, mh, rc); 300 301 error = dlm_wait_function(ls, &rcom_response); 302 disallow_sync_reply(ls); 303 out: 304 return error; 305 } 306 307 static void receive_rcom_names(struct dlm_ls *ls, struct dlm_rcom *rc_in) 308 { 309 struct dlm_rcom *rc; 310 struct dlm_mhandle *mh; 311 int error, inlen, outlen, nodeid; 312 313 nodeid = rc_in->rc_header.h_nodeid; 314 inlen = rc_in->rc_header.h_length - sizeof(struct dlm_rcom); 315 outlen = dlm_config.ci_buffer_size - sizeof(struct dlm_rcom); 316 317 error = create_rcom(ls, nodeid, DLM_RCOM_NAMES_REPLY, outlen, &rc, &mh); 318 if (error) 319 return; 320 rc->rc_id = rc_in->rc_id; 321 rc->rc_seq_reply = rc_in->rc_seq; 322 323 dlm_copy_master_names(ls, rc_in->rc_buf, inlen, rc->rc_buf, outlen, 324 nodeid); 325 send_rcom(ls, mh, rc); 326 } 327 328 int dlm_send_rcom_lookup(struct dlm_rsb *r, int dir_nodeid) 329 { 330 struct dlm_rcom *rc; 331 struct dlm_mhandle *mh; 332 struct dlm_ls *ls = r->res_ls; 333 int error; 334 335 error = create_rcom(ls, dir_nodeid, DLM_RCOM_LOOKUP, r->res_length, 336 &rc, &mh); 337 if (error) 338 goto out; 339 memcpy(rc->rc_buf, r->res_name, r->res_length); 340 rc->rc_id = (unsigned long) r; 341 342 send_rcom(ls, mh, rc); 343 out: 344 return error; 345 } 346 347 static void receive_rcom_lookup(struct dlm_ls *ls, struct dlm_rcom *rc_in) 348 { 349 struct dlm_rcom *rc; 350 struct dlm_mhandle *mh; 351 int error, ret_nodeid, nodeid = rc_in->rc_header.h_nodeid; 352 int len = rc_in->rc_header.h_length - sizeof(struct dlm_rcom); 353 354 error = create_rcom(ls, nodeid, DLM_RCOM_LOOKUP_REPLY, 0, &rc, &mh); 355 if (error) 356 return; 357 358 error = dlm_dir_lookup(ls, nodeid, rc_in->rc_buf, len, &ret_nodeid); 359 if (error) 360 ret_nodeid = error; 361 rc->rc_result = ret_nodeid; 362 rc->rc_id = rc_in->rc_id; 363 rc->rc_seq_reply = rc_in->rc_seq; 364 365 send_rcom(ls, mh, rc); 366 } 367 368 static void receive_rcom_lookup_reply(struct dlm_ls *ls, struct dlm_rcom *rc_in) 369 { 370 dlm_recover_master_reply(ls, rc_in); 371 } 372 373 static void pack_rcom_lock(struct dlm_rsb *r, struct dlm_lkb *lkb, 374 struct rcom_lock *rl) 375 { 376 memset(rl, 0, sizeof(*rl)); 377 378 rl->rl_ownpid = cpu_to_le32(lkb->lkb_ownpid); 379 rl->rl_lkid = cpu_to_le32(lkb->lkb_id); 380 rl->rl_exflags = cpu_to_le32(lkb->lkb_exflags); 381 rl->rl_flags = cpu_to_le32(lkb->lkb_flags); 382 rl->rl_lvbseq = cpu_to_le32(lkb->lkb_lvbseq); 383 rl->rl_rqmode = lkb->lkb_rqmode; 384 rl->rl_grmode = lkb->lkb_grmode; 385 rl->rl_status = lkb->lkb_status; 386 rl->rl_wait_type = cpu_to_le16(lkb->lkb_wait_type); 387 388 if (lkb->lkb_bastfn) 389 rl->rl_asts |= DLM_CB_BAST; 390 if (lkb->lkb_astfn) 391 rl->rl_asts |= DLM_CB_CAST; 392 393 rl->rl_namelen = cpu_to_le16(r->res_length); 394 memcpy(rl->rl_name, r->res_name, r->res_length); 395 396 /* FIXME: might we have an lvb without DLM_LKF_VALBLK set ? 397 If so, receive_rcom_lock_args() won't take this copy. */ 398 399 if (lkb->lkb_lvbptr) 400 memcpy(rl->rl_lvb, lkb->lkb_lvbptr, r->res_ls->ls_lvblen); 401 } 402 403 int dlm_send_rcom_lock(struct dlm_rsb *r, struct dlm_lkb *lkb) 404 { 405 struct dlm_ls *ls = r->res_ls; 406 struct dlm_rcom *rc; 407 struct dlm_mhandle *mh; 408 struct rcom_lock *rl; 409 int error, len = sizeof(struct rcom_lock); 410 411 if (lkb->lkb_lvbptr) 412 len += ls->ls_lvblen; 413 414 error = create_rcom(ls, r->res_nodeid, DLM_RCOM_LOCK, len, &rc, &mh); 415 if (error) 416 goto out; 417 418 rl = (struct rcom_lock *) rc->rc_buf; 419 pack_rcom_lock(r, lkb, rl); 420 rc->rc_id = (unsigned long) r; 421 422 send_rcom(ls, mh, rc); 423 out: 424 return error; 425 } 426 427 /* needs at least dlm_rcom + rcom_lock */ 428 static void receive_rcom_lock(struct dlm_ls *ls, struct dlm_rcom *rc_in) 429 { 430 struct dlm_rcom *rc; 431 struct dlm_mhandle *mh; 432 int error, nodeid = rc_in->rc_header.h_nodeid; 433 434 dlm_recover_master_copy(ls, rc_in); 435 436 error = create_rcom(ls, nodeid, DLM_RCOM_LOCK_REPLY, 437 sizeof(struct rcom_lock), &rc, &mh); 438 if (error) 439 return; 440 441 /* We send back the same rcom_lock struct we received, but 442 dlm_recover_master_copy() has filled in rl_remid and rl_result */ 443 444 memcpy(rc->rc_buf, rc_in->rc_buf, sizeof(struct rcom_lock)); 445 rc->rc_id = rc_in->rc_id; 446 rc->rc_seq_reply = rc_in->rc_seq; 447 448 send_rcom(ls, mh, rc); 449 } 450 451 /* If the lockspace doesn't exist then still send a status message 452 back; it's possible that it just doesn't have its global_id yet. */ 453 454 int dlm_send_ls_not_ready(int nodeid, struct dlm_rcom *rc_in) 455 { 456 struct dlm_rcom *rc; 457 struct rcom_config *rf; 458 struct dlm_mhandle *mh; 459 char *mb; 460 int mb_len = sizeof(struct dlm_rcom) + sizeof(struct rcom_config); 461 462 mh = dlm_lowcomms_get_buffer(nodeid, mb_len, GFP_NOFS, &mb); 463 if (!mh) 464 return -ENOBUFS; 465 memset(mb, 0, mb_len); 466 467 rc = (struct dlm_rcom *) mb; 468 469 rc->rc_header.h_version = (DLM_HEADER_MAJOR | DLM_HEADER_MINOR); 470 rc->rc_header.h_lockspace = rc_in->rc_header.h_lockspace; 471 rc->rc_header.h_nodeid = dlm_our_nodeid(); 472 rc->rc_header.h_length = mb_len; 473 rc->rc_header.h_cmd = DLM_RCOM; 474 475 rc->rc_type = DLM_RCOM_STATUS_REPLY; 476 rc->rc_id = rc_in->rc_id; 477 rc->rc_seq_reply = rc_in->rc_seq; 478 rc->rc_result = -ESRCH; 479 480 rf = (struct rcom_config *) rc->rc_buf; 481 rf->rf_lvblen = cpu_to_le32(~0U); 482 483 dlm_rcom_out(rc); 484 dlm_lowcomms_commit_buffer(mh); 485 486 return 0; 487 } 488 489 static int is_old_reply(struct dlm_ls *ls, struct dlm_rcom *rc) 490 { 491 uint64_t seq; 492 int rv = 0; 493 494 switch (rc->rc_type) { 495 case DLM_RCOM_STATUS_REPLY: 496 case DLM_RCOM_NAMES_REPLY: 497 case DLM_RCOM_LOOKUP_REPLY: 498 case DLM_RCOM_LOCK_REPLY: 499 spin_lock(&ls->ls_recover_lock); 500 seq = ls->ls_recover_seq; 501 spin_unlock(&ls->ls_recover_lock); 502 if (rc->rc_seq_reply != seq) { 503 log_debug(ls, "ignoring old reply %x from %d " 504 "seq_reply %llx expect %llx", 505 rc->rc_type, rc->rc_header.h_nodeid, 506 (unsigned long long)rc->rc_seq_reply, 507 (unsigned long long)seq); 508 rv = 1; 509 } 510 } 511 return rv; 512 } 513 514 /* Called by dlm_recv; corresponds to dlm_receive_message() but special 515 recovery-only comms are sent through here. */ 516 517 void dlm_receive_rcom(struct dlm_ls *ls, struct dlm_rcom *rc, int nodeid) 518 { 519 int lock_size = sizeof(struct dlm_rcom) + sizeof(struct rcom_lock); 520 521 if (dlm_recovery_stopped(ls) && (rc->rc_type != DLM_RCOM_STATUS)) { 522 log_debug(ls, "ignoring recovery message %x from %d", 523 rc->rc_type, nodeid); 524 goto out; 525 } 526 527 if (is_old_reply(ls, rc)) 528 goto out; 529 530 switch (rc->rc_type) { 531 case DLM_RCOM_STATUS: 532 receive_rcom_status(ls, rc); 533 break; 534 535 case DLM_RCOM_NAMES: 536 receive_rcom_names(ls, rc); 537 break; 538 539 case DLM_RCOM_LOOKUP: 540 receive_rcom_lookup(ls, rc); 541 break; 542 543 case DLM_RCOM_LOCK: 544 if (rc->rc_header.h_length < lock_size) 545 goto Eshort; 546 receive_rcom_lock(ls, rc); 547 break; 548 549 case DLM_RCOM_STATUS_REPLY: 550 receive_sync_reply(ls, rc); 551 break; 552 553 case DLM_RCOM_NAMES_REPLY: 554 receive_sync_reply(ls, rc); 555 break; 556 557 case DLM_RCOM_LOOKUP_REPLY: 558 receive_rcom_lookup_reply(ls, rc); 559 break; 560 561 case DLM_RCOM_LOCK_REPLY: 562 if (rc->rc_header.h_length < lock_size) 563 goto Eshort; 564 dlm_recover_process_copy(ls, rc); 565 break; 566 567 default: 568 log_error(ls, "receive_rcom bad type %d", rc->rc_type); 569 } 570 out: 571 return; 572 Eshort: 573 log_error(ls, "recovery message %x from %d is too short", 574 rc->rc_type, nodeid); 575 } 576 577