1 /*- 2 * SPDX-License-Identifier: BSD-2-Clause-FreeBSD 3 * 4 * Copyright (c) 2004-2006 Pawel Jakub Dawidek <pjd@FreeBSD.org> 5 * All rights reserved. 6 * 7 * Redistribution and use in source and binary forms, with or without 8 * modification, are permitted provided that the following conditions 9 * are met: 10 * 1. Redistributions of source code must retain the above copyright 11 * notice, this list of conditions and the following disclaimer. 12 * 2. Redistributions in binary form must reproduce the above copyright 13 * notice, this list of conditions and the following disclaimer in the 14 * documentation and/or other materials provided with the distribution. 15 * 16 * THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND 17 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 18 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 19 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE 20 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 21 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 22 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 23 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 24 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 25 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 26 * SUCH DAMAGE. 27 */ 28 29 #include <sys/cdefs.h> 30 __FBSDID("$FreeBSD$"); 31 32 #include <sys/param.h> 33 #include <sys/systm.h> 34 #include <sys/kernel.h> 35 #include <sys/module.h> 36 #include <sys/limits.h> 37 #include <sys/lock.h> 38 #include <sys/mutex.h> 39 #include <sys/bio.h> 40 #include <sys/sbuf.h> 41 #include <sys/sysctl.h> 42 #include <sys/malloc.h> 43 #include <sys/eventhandler.h> 44 #include <vm/uma.h> 45 #include <geom/geom.h> 46 #include <geom/geom_dbg.h> 47 #include <sys/proc.h> 48 #include <sys/kthread.h> 49 #include <sys/sched.h> 50 #include <geom/raid3/g_raid3.h> 51 52 FEATURE(geom_raid3, "GEOM RAID-3 functionality"); 53 54 static MALLOC_DEFINE(M_RAID3, "raid3_data", "GEOM_RAID3 Data"); 55 56 SYSCTL_DECL(_kern_geom); 57 static SYSCTL_NODE(_kern_geom, OID_AUTO, raid3, CTLFLAG_RW | CTLFLAG_MPSAFE, 0, 58 "GEOM_RAID3 stuff"); 59 u_int g_raid3_debug = 0; 60 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, debug, CTLFLAG_RWTUN, &g_raid3_debug, 0, 61 "Debug level"); 62 static u_int g_raid3_timeout = 4; 63 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, timeout, CTLFLAG_RWTUN, &g_raid3_timeout, 64 0, "Time to wait on all raid3 components"); 65 static u_int g_raid3_idletime = 5; 66 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, idletime, CTLFLAG_RWTUN, 67 &g_raid3_idletime, 0, "Mark components as clean when idling"); 68 static u_int g_raid3_disconnect_on_failure = 1; 69 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, disconnect_on_failure, CTLFLAG_RWTUN, 70 &g_raid3_disconnect_on_failure, 0, "Disconnect component on I/O failure."); 71 static u_int g_raid3_syncreqs = 2; 72 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, sync_requests, CTLFLAG_RDTUN, 73 &g_raid3_syncreqs, 0, "Parallel synchronization I/O requests."); 74 static u_int g_raid3_use_malloc = 0; 75 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, use_malloc, CTLFLAG_RDTUN, 76 &g_raid3_use_malloc, 0, "Use malloc(9) instead of uma(9)."); 77 78 static u_int g_raid3_n64k = 50; 79 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, n64k, CTLFLAG_RDTUN, &g_raid3_n64k, 0, 80 "Maximum number of 64kB allocations"); 81 static u_int g_raid3_n16k = 200; 82 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, n16k, CTLFLAG_RDTUN, &g_raid3_n16k, 0, 83 "Maximum number of 16kB allocations"); 84 static u_int g_raid3_n4k = 1200; 85 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, n4k, CTLFLAG_RDTUN, &g_raid3_n4k, 0, 86 "Maximum number of 4kB allocations"); 87 88 static SYSCTL_NODE(_kern_geom_raid3, OID_AUTO, stat, 89 CTLFLAG_RW | CTLFLAG_MPSAFE, 0, 90 "GEOM_RAID3 statistics"); 91 static u_int g_raid3_parity_mismatch = 0; 92 SYSCTL_UINT(_kern_geom_raid3_stat, OID_AUTO, parity_mismatch, CTLFLAG_RD, 93 &g_raid3_parity_mismatch, 0, "Number of failures in VERIFY mode"); 94 95 #define MSLEEP(ident, mtx, priority, wmesg, timeout) do { \ 96 G_RAID3_DEBUG(4, "%s: Sleeping %p.", __func__, (ident)); \ 97 msleep((ident), (mtx), (priority), (wmesg), (timeout)); \ 98 G_RAID3_DEBUG(4, "%s: Woken up %p.", __func__, (ident)); \ 99 } while (0) 100 101 static eventhandler_tag g_raid3_post_sync = NULL; 102 static int g_raid3_shutdown = 0; 103 104 static int g_raid3_destroy_geom(struct gctl_req *req, struct g_class *mp, 105 struct g_geom *gp); 106 static g_taste_t g_raid3_taste; 107 static void g_raid3_init(struct g_class *mp); 108 static void g_raid3_fini(struct g_class *mp); 109 110 struct g_class g_raid3_class = { 111 .name = G_RAID3_CLASS_NAME, 112 .version = G_VERSION, 113 .ctlreq = g_raid3_config, 114 .taste = g_raid3_taste, 115 .destroy_geom = g_raid3_destroy_geom, 116 .init = g_raid3_init, 117 .fini = g_raid3_fini 118 }; 119 120 static void g_raid3_destroy_provider(struct g_raid3_softc *sc); 121 static int g_raid3_update_disk(struct g_raid3_disk *disk, u_int state); 122 static void g_raid3_update_device(struct g_raid3_softc *sc, boolean_t force); 123 static void g_raid3_dumpconf(struct sbuf *sb, const char *indent, 124 struct g_geom *gp, struct g_consumer *cp, struct g_provider *pp); 125 static void g_raid3_sync_stop(struct g_raid3_softc *sc, int type); 126 static int g_raid3_register_request(struct bio *pbp); 127 static void g_raid3_sync_release(struct g_raid3_softc *sc); 128 129 static const char * 130 g_raid3_disk_state2str(int state) 131 { 132 133 switch (state) { 134 case G_RAID3_DISK_STATE_NODISK: 135 return ("NODISK"); 136 case G_RAID3_DISK_STATE_NONE: 137 return ("NONE"); 138 case G_RAID3_DISK_STATE_NEW: 139 return ("NEW"); 140 case G_RAID3_DISK_STATE_ACTIVE: 141 return ("ACTIVE"); 142 case G_RAID3_DISK_STATE_STALE: 143 return ("STALE"); 144 case G_RAID3_DISK_STATE_SYNCHRONIZING: 145 return ("SYNCHRONIZING"); 146 case G_RAID3_DISK_STATE_DISCONNECTED: 147 return ("DISCONNECTED"); 148 default: 149 return ("INVALID"); 150 } 151 } 152 153 static const char * 154 g_raid3_device_state2str(int state) 155 { 156 157 switch (state) { 158 case G_RAID3_DEVICE_STATE_STARTING: 159 return ("STARTING"); 160 case G_RAID3_DEVICE_STATE_DEGRADED: 161 return ("DEGRADED"); 162 case G_RAID3_DEVICE_STATE_COMPLETE: 163 return ("COMPLETE"); 164 default: 165 return ("INVALID"); 166 } 167 } 168 169 const char * 170 g_raid3_get_diskname(struct g_raid3_disk *disk) 171 { 172 173 if (disk->d_consumer == NULL || disk->d_consumer->provider == NULL) 174 return ("[unknown]"); 175 return (disk->d_name); 176 } 177 178 static void * 179 g_raid3_alloc(struct g_raid3_softc *sc, size_t size, int flags) 180 { 181 void *ptr; 182 enum g_raid3_zones zone; 183 184 if (g_raid3_use_malloc || 185 (zone = g_raid3_zone(size)) == G_RAID3_NUM_ZONES) 186 ptr = malloc(size, M_RAID3, flags); 187 else { 188 ptr = uma_zalloc_arg(sc->sc_zones[zone].sz_zone, 189 &sc->sc_zones[zone], flags); 190 sc->sc_zones[zone].sz_requested++; 191 if (ptr == NULL) 192 sc->sc_zones[zone].sz_failed++; 193 } 194 return (ptr); 195 } 196 197 static void 198 g_raid3_free(struct g_raid3_softc *sc, void *ptr, size_t size) 199 { 200 enum g_raid3_zones zone; 201 202 if (g_raid3_use_malloc || 203 (zone = g_raid3_zone(size)) == G_RAID3_NUM_ZONES) 204 free(ptr, M_RAID3); 205 else { 206 uma_zfree_arg(sc->sc_zones[zone].sz_zone, 207 ptr, &sc->sc_zones[zone]); 208 } 209 } 210 211 static int 212 g_raid3_uma_ctor(void *mem, int size, void *arg, int flags) 213 { 214 struct g_raid3_zone *sz = arg; 215 216 if (sz->sz_max > 0 && sz->sz_inuse == sz->sz_max) 217 return (ENOMEM); 218 sz->sz_inuse++; 219 return (0); 220 } 221 222 static void 223 g_raid3_uma_dtor(void *mem, int size, void *arg) 224 { 225 struct g_raid3_zone *sz = arg; 226 227 sz->sz_inuse--; 228 } 229 230 #define g_raid3_xor(src, dst, size) \ 231 _g_raid3_xor((uint64_t *)(src), \ 232 (uint64_t *)(dst), (size_t)size) 233 static void 234 _g_raid3_xor(uint64_t *src, uint64_t *dst, size_t size) 235 { 236 237 KASSERT((size % 128) == 0, ("Invalid size: %zu.", size)); 238 for (; size > 0; size -= 128) { 239 *dst++ ^= (*src++); 240 *dst++ ^= (*src++); 241 *dst++ ^= (*src++); 242 *dst++ ^= (*src++); 243 *dst++ ^= (*src++); 244 *dst++ ^= (*src++); 245 *dst++ ^= (*src++); 246 *dst++ ^= (*src++); 247 *dst++ ^= (*src++); 248 *dst++ ^= (*src++); 249 *dst++ ^= (*src++); 250 *dst++ ^= (*src++); 251 *dst++ ^= (*src++); 252 *dst++ ^= (*src++); 253 *dst++ ^= (*src++); 254 *dst++ ^= (*src++); 255 } 256 } 257 258 static int 259 g_raid3_is_zero(struct bio *bp) 260 { 261 static const uint64_t zeros[] = { 262 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 263 }; 264 u_char *addr; 265 ssize_t size; 266 267 size = bp->bio_length; 268 addr = (u_char *)bp->bio_data; 269 for (; size > 0; size -= sizeof(zeros), addr += sizeof(zeros)) { 270 if (bcmp(addr, zeros, sizeof(zeros)) != 0) 271 return (0); 272 } 273 return (1); 274 } 275 276 /* 277 * --- Events handling functions --- 278 * Events in geom_raid3 are used to maintain disks and device status 279 * from one thread to simplify locking. 280 */ 281 static void 282 g_raid3_event_free(struct g_raid3_event *ep) 283 { 284 285 free(ep, M_RAID3); 286 } 287 288 int 289 g_raid3_event_send(void *arg, int state, int flags) 290 { 291 struct g_raid3_softc *sc; 292 struct g_raid3_disk *disk; 293 struct g_raid3_event *ep; 294 int error; 295 296 ep = malloc(sizeof(*ep), M_RAID3, M_WAITOK); 297 G_RAID3_DEBUG(4, "%s: Sending event %p.", __func__, ep); 298 if ((flags & G_RAID3_EVENT_DEVICE) != 0) { 299 disk = NULL; 300 sc = arg; 301 } else { 302 disk = arg; 303 sc = disk->d_softc; 304 } 305 ep->e_disk = disk; 306 ep->e_state = state; 307 ep->e_flags = flags; 308 ep->e_error = 0; 309 mtx_lock(&sc->sc_events_mtx); 310 TAILQ_INSERT_TAIL(&sc->sc_events, ep, e_next); 311 mtx_unlock(&sc->sc_events_mtx); 312 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, sc); 313 mtx_lock(&sc->sc_queue_mtx); 314 wakeup(sc); 315 wakeup(&sc->sc_queue); 316 mtx_unlock(&sc->sc_queue_mtx); 317 if ((flags & G_RAID3_EVENT_DONTWAIT) != 0) 318 return (0); 319 sx_assert(&sc->sc_lock, SX_XLOCKED); 320 G_RAID3_DEBUG(4, "%s: Sleeping %p.", __func__, ep); 321 sx_xunlock(&sc->sc_lock); 322 while ((ep->e_flags & G_RAID3_EVENT_DONE) == 0) { 323 mtx_lock(&sc->sc_events_mtx); 324 MSLEEP(ep, &sc->sc_events_mtx, PRIBIO | PDROP, "r3:event", 325 hz * 5); 326 } 327 error = ep->e_error; 328 g_raid3_event_free(ep); 329 sx_xlock(&sc->sc_lock); 330 return (error); 331 } 332 333 static struct g_raid3_event * 334 g_raid3_event_get(struct g_raid3_softc *sc) 335 { 336 struct g_raid3_event *ep; 337 338 mtx_lock(&sc->sc_events_mtx); 339 ep = TAILQ_FIRST(&sc->sc_events); 340 mtx_unlock(&sc->sc_events_mtx); 341 return (ep); 342 } 343 344 static void 345 g_raid3_event_remove(struct g_raid3_softc *sc, struct g_raid3_event *ep) 346 { 347 348 mtx_lock(&sc->sc_events_mtx); 349 TAILQ_REMOVE(&sc->sc_events, ep, e_next); 350 mtx_unlock(&sc->sc_events_mtx); 351 } 352 353 static void 354 g_raid3_event_cancel(struct g_raid3_disk *disk) 355 { 356 struct g_raid3_softc *sc; 357 struct g_raid3_event *ep, *tmpep; 358 359 sc = disk->d_softc; 360 sx_assert(&sc->sc_lock, SX_XLOCKED); 361 362 mtx_lock(&sc->sc_events_mtx); 363 TAILQ_FOREACH_SAFE(ep, &sc->sc_events, e_next, tmpep) { 364 if ((ep->e_flags & G_RAID3_EVENT_DEVICE) != 0) 365 continue; 366 if (ep->e_disk != disk) 367 continue; 368 TAILQ_REMOVE(&sc->sc_events, ep, e_next); 369 if ((ep->e_flags & G_RAID3_EVENT_DONTWAIT) != 0) 370 g_raid3_event_free(ep); 371 else { 372 ep->e_error = ECANCELED; 373 wakeup(ep); 374 } 375 } 376 mtx_unlock(&sc->sc_events_mtx); 377 } 378 379 /* 380 * Return the number of disks in the given state. 381 * If state is equal to -1, count all connected disks. 382 */ 383 u_int 384 g_raid3_ndisks(struct g_raid3_softc *sc, int state) 385 { 386 struct g_raid3_disk *disk; 387 u_int n, ndisks; 388 389 sx_assert(&sc->sc_lock, SX_LOCKED); 390 391 for (n = ndisks = 0; n < sc->sc_ndisks; n++) { 392 disk = &sc->sc_disks[n]; 393 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 394 continue; 395 if (state == -1 || disk->d_state == state) 396 ndisks++; 397 } 398 return (ndisks); 399 } 400 401 static u_int 402 g_raid3_nrequests(struct g_raid3_softc *sc, struct g_consumer *cp) 403 { 404 struct bio *bp; 405 u_int nreqs = 0; 406 407 mtx_lock(&sc->sc_queue_mtx); 408 TAILQ_FOREACH(bp, &sc->sc_queue.queue, bio_queue) { 409 if (bp->bio_from == cp) 410 nreqs++; 411 } 412 mtx_unlock(&sc->sc_queue_mtx); 413 return (nreqs); 414 } 415 416 static int 417 g_raid3_is_busy(struct g_raid3_softc *sc, struct g_consumer *cp) 418 { 419 420 if (cp->index > 0) { 421 G_RAID3_DEBUG(2, 422 "I/O requests for %s exist, can't destroy it now.", 423 cp->provider->name); 424 return (1); 425 } 426 if (g_raid3_nrequests(sc, cp) > 0) { 427 G_RAID3_DEBUG(2, 428 "I/O requests for %s in queue, can't destroy it now.", 429 cp->provider->name); 430 return (1); 431 } 432 return (0); 433 } 434 435 static void 436 g_raid3_destroy_consumer(void *arg, int flags __unused) 437 { 438 struct g_consumer *cp; 439 440 g_topology_assert(); 441 442 cp = arg; 443 G_RAID3_DEBUG(1, "Consumer %s destroyed.", cp->provider->name); 444 g_detach(cp); 445 g_destroy_consumer(cp); 446 } 447 448 static void 449 g_raid3_kill_consumer(struct g_raid3_softc *sc, struct g_consumer *cp) 450 { 451 struct g_provider *pp; 452 int retaste_wait; 453 454 g_topology_assert(); 455 456 cp->private = NULL; 457 if (g_raid3_is_busy(sc, cp)) 458 return; 459 G_RAID3_DEBUG(2, "Consumer %s destroyed.", cp->provider->name); 460 pp = cp->provider; 461 retaste_wait = 0; 462 if (cp->acw == 1) { 463 if ((pp->geom->flags & G_GEOM_WITHER) == 0) 464 retaste_wait = 1; 465 } 466 G_RAID3_DEBUG(2, "Access %s r%dw%de%d = %d", pp->name, -cp->acr, 467 -cp->acw, -cp->ace, 0); 468 if (cp->acr > 0 || cp->acw > 0 || cp->ace > 0) 469 g_access(cp, -cp->acr, -cp->acw, -cp->ace); 470 if (retaste_wait) { 471 /* 472 * After retaste event was send (inside g_access()), we can send 473 * event to detach and destroy consumer. 474 * A class, which has consumer to the given provider connected 475 * will not receive retaste event for the provider. 476 * This is the way how I ignore retaste events when I close 477 * consumers opened for write: I detach and destroy consumer 478 * after retaste event is sent. 479 */ 480 g_post_event(g_raid3_destroy_consumer, cp, M_WAITOK, NULL); 481 return; 482 } 483 G_RAID3_DEBUG(1, "Consumer %s destroyed.", pp->name); 484 g_detach(cp); 485 g_destroy_consumer(cp); 486 } 487 488 static int 489 g_raid3_connect_disk(struct g_raid3_disk *disk, struct g_provider *pp) 490 { 491 struct g_consumer *cp; 492 int error; 493 494 g_topology_assert_not(); 495 KASSERT(disk->d_consumer == NULL, 496 ("Disk already connected (device %s).", disk->d_softc->sc_name)); 497 498 g_topology_lock(); 499 cp = g_new_consumer(disk->d_softc->sc_geom); 500 error = g_attach(cp, pp); 501 if (error != 0) { 502 g_destroy_consumer(cp); 503 g_topology_unlock(); 504 return (error); 505 } 506 error = g_access(cp, 1, 1, 1); 507 g_topology_unlock(); 508 if (error != 0) { 509 g_detach(cp); 510 g_destroy_consumer(cp); 511 G_RAID3_DEBUG(0, "Cannot open consumer %s (error=%d).", 512 pp->name, error); 513 return (error); 514 } 515 disk->d_consumer = cp; 516 disk->d_consumer->private = disk; 517 disk->d_consumer->index = 0; 518 G_RAID3_DEBUG(2, "Disk %s connected.", g_raid3_get_diskname(disk)); 519 return (0); 520 } 521 522 static void 523 g_raid3_disconnect_consumer(struct g_raid3_softc *sc, struct g_consumer *cp) 524 { 525 526 g_topology_assert(); 527 528 if (cp == NULL) 529 return; 530 if (cp->provider != NULL) 531 g_raid3_kill_consumer(sc, cp); 532 else 533 g_destroy_consumer(cp); 534 } 535 536 /* 537 * Initialize disk. This means allocate memory, create consumer, attach it 538 * to the provider and open access (r1w1e1) to it. 539 */ 540 static struct g_raid3_disk * 541 g_raid3_init_disk(struct g_raid3_softc *sc, struct g_provider *pp, 542 struct g_raid3_metadata *md, int *errorp) 543 { 544 struct g_raid3_disk *disk; 545 int error; 546 547 disk = &sc->sc_disks[md->md_no]; 548 error = g_raid3_connect_disk(disk, pp); 549 if (error != 0) { 550 if (errorp != NULL) 551 *errorp = error; 552 return (NULL); 553 } 554 disk->d_state = G_RAID3_DISK_STATE_NONE; 555 disk->d_flags = md->md_dflags; 556 if (md->md_provider[0] != '\0') 557 disk->d_flags |= G_RAID3_DISK_FLAG_HARDCODED; 558 disk->d_sync.ds_consumer = NULL; 559 disk->d_sync.ds_offset = md->md_sync_offset; 560 disk->d_sync.ds_offset_done = md->md_sync_offset; 561 disk->d_genid = md->md_genid; 562 disk->d_sync.ds_syncid = md->md_syncid; 563 if (errorp != NULL) 564 *errorp = 0; 565 return (disk); 566 } 567 568 static void 569 g_raid3_destroy_disk(struct g_raid3_disk *disk) 570 { 571 struct g_raid3_softc *sc; 572 573 g_topology_assert_not(); 574 sc = disk->d_softc; 575 sx_assert(&sc->sc_lock, SX_XLOCKED); 576 577 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 578 return; 579 g_raid3_event_cancel(disk); 580 switch (disk->d_state) { 581 case G_RAID3_DISK_STATE_SYNCHRONIZING: 582 if (sc->sc_syncdisk != NULL) 583 g_raid3_sync_stop(sc, 1); 584 /* FALLTHROUGH */ 585 case G_RAID3_DISK_STATE_NEW: 586 case G_RAID3_DISK_STATE_STALE: 587 case G_RAID3_DISK_STATE_ACTIVE: 588 g_topology_lock(); 589 g_raid3_disconnect_consumer(sc, disk->d_consumer); 590 g_topology_unlock(); 591 disk->d_consumer = NULL; 592 break; 593 default: 594 KASSERT(0 == 1, ("Wrong disk state (%s, %s).", 595 g_raid3_get_diskname(disk), 596 g_raid3_disk_state2str(disk->d_state))); 597 } 598 disk->d_state = G_RAID3_DISK_STATE_NODISK; 599 } 600 601 static void 602 g_raid3_destroy_device(struct g_raid3_softc *sc) 603 { 604 struct g_raid3_event *ep; 605 struct g_raid3_disk *disk; 606 struct g_geom *gp; 607 struct g_consumer *cp; 608 u_int n; 609 610 g_topology_assert_not(); 611 sx_assert(&sc->sc_lock, SX_XLOCKED); 612 613 gp = sc->sc_geom; 614 if (sc->sc_provider != NULL) 615 g_raid3_destroy_provider(sc); 616 for (n = 0; n < sc->sc_ndisks; n++) { 617 disk = &sc->sc_disks[n]; 618 if (disk->d_state != G_RAID3_DISK_STATE_NODISK) { 619 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 620 g_raid3_update_metadata(disk); 621 g_raid3_destroy_disk(disk); 622 } 623 } 624 while ((ep = g_raid3_event_get(sc)) != NULL) { 625 g_raid3_event_remove(sc, ep); 626 if ((ep->e_flags & G_RAID3_EVENT_DONTWAIT) != 0) 627 g_raid3_event_free(ep); 628 else { 629 ep->e_error = ECANCELED; 630 ep->e_flags |= G_RAID3_EVENT_DONE; 631 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, ep); 632 mtx_lock(&sc->sc_events_mtx); 633 wakeup(ep); 634 mtx_unlock(&sc->sc_events_mtx); 635 } 636 } 637 callout_drain(&sc->sc_callout); 638 cp = LIST_FIRST(&sc->sc_sync.ds_geom->consumer); 639 g_topology_lock(); 640 if (cp != NULL) 641 g_raid3_disconnect_consumer(sc, cp); 642 g_wither_geom(sc->sc_sync.ds_geom, ENXIO); 643 G_RAID3_DEBUG(0, "Device %s destroyed.", gp->name); 644 g_wither_geom(gp, ENXIO); 645 g_topology_unlock(); 646 if (!g_raid3_use_malloc) { 647 uma_zdestroy(sc->sc_zones[G_RAID3_ZONE_64K].sz_zone); 648 uma_zdestroy(sc->sc_zones[G_RAID3_ZONE_16K].sz_zone); 649 uma_zdestroy(sc->sc_zones[G_RAID3_ZONE_4K].sz_zone); 650 } 651 mtx_destroy(&sc->sc_queue_mtx); 652 mtx_destroy(&sc->sc_events_mtx); 653 sx_xunlock(&sc->sc_lock); 654 sx_destroy(&sc->sc_lock); 655 } 656 657 static void 658 g_raid3_orphan(struct g_consumer *cp) 659 { 660 struct g_raid3_disk *disk; 661 662 g_topology_assert(); 663 664 disk = cp->private; 665 if (disk == NULL) 666 return; 667 disk->d_softc->sc_bump_id = G_RAID3_BUMP_SYNCID; 668 g_raid3_event_send(disk, G_RAID3_DISK_STATE_DISCONNECTED, 669 G_RAID3_EVENT_DONTWAIT); 670 } 671 672 static int 673 g_raid3_write_metadata(struct g_raid3_disk *disk, struct g_raid3_metadata *md) 674 { 675 struct g_raid3_softc *sc; 676 struct g_consumer *cp; 677 off_t offset, length; 678 u_char *sector; 679 int error = 0; 680 681 g_topology_assert_not(); 682 sc = disk->d_softc; 683 sx_assert(&sc->sc_lock, SX_LOCKED); 684 685 cp = disk->d_consumer; 686 KASSERT(cp != NULL, ("NULL consumer (%s).", sc->sc_name)); 687 KASSERT(cp->provider != NULL, ("NULL provider (%s).", sc->sc_name)); 688 KASSERT(cp->acr >= 1 && cp->acw >= 1 && cp->ace >= 1, 689 ("Consumer %s closed? (r%dw%de%d).", cp->provider->name, cp->acr, 690 cp->acw, cp->ace)); 691 length = cp->provider->sectorsize; 692 offset = cp->provider->mediasize - length; 693 sector = malloc((size_t)length, M_RAID3, M_WAITOK | M_ZERO); 694 if (md != NULL) 695 raid3_metadata_encode(md, sector); 696 error = g_write_data(cp, offset, sector, length); 697 free(sector, M_RAID3); 698 if (error != 0) { 699 if ((disk->d_flags & G_RAID3_DISK_FLAG_BROKEN) == 0) { 700 G_RAID3_DEBUG(0, "Cannot write metadata on %s " 701 "(device=%s, error=%d).", 702 g_raid3_get_diskname(disk), sc->sc_name, error); 703 disk->d_flags |= G_RAID3_DISK_FLAG_BROKEN; 704 } else { 705 G_RAID3_DEBUG(1, "Cannot write metadata on %s " 706 "(device=%s, error=%d).", 707 g_raid3_get_diskname(disk), sc->sc_name, error); 708 } 709 if (g_raid3_disconnect_on_failure && 710 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE) { 711 sc->sc_bump_id |= G_RAID3_BUMP_GENID; 712 g_raid3_event_send(disk, 713 G_RAID3_DISK_STATE_DISCONNECTED, 714 G_RAID3_EVENT_DONTWAIT); 715 } 716 } 717 return (error); 718 } 719 720 int 721 g_raid3_clear_metadata(struct g_raid3_disk *disk) 722 { 723 int error; 724 725 g_topology_assert_not(); 726 sx_assert(&disk->d_softc->sc_lock, SX_LOCKED); 727 728 error = g_raid3_write_metadata(disk, NULL); 729 if (error == 0) { 730 G_RAID3_DEBUG(2, "Metadata on %s cleared.", 731 g_raid3_get_diskname(disk)); 732 } else { 733 G_RAID3_DEBUG(0, 734 "Cannot clear metadata on disk %s (error=%d).", 735 g_raid3_get_diskname(disk), error); 736 } 737 return (error); 738 } 739 740 void 741 g_raid3_fill_metadata(struct g_raid3_disk *disk, struct g_raid3_metadata *md) 742 { 743 struct g_raid3_softc *sc; 744 struct g_provider *pp; 745 746 bzero(md, sizeof(*md)); 747 sc = disk->d_softc; 748 strlcpy(md->md_magic, G_RAID3_MAGIC, sizeof(md->md_magic)); 749 md->md_version = G_RAID3_VERSION; 750 strlcpy(md->md_name, sc->sc_name, sizeof(md->md_name)); 751 md->md_id = sc->sc_id; 752 md->md_all = sc->sc_ndisks; 753 md->md_genid = sc->sc_genid; 754 md->md_mediasize = sc->sc_mediasize; 755 md->md_sectorsize = sc->sc_sectorsize; 756 md->md_mflags = (sc->sc_flags & G_RAID3_DEVICE_FLAG_MASK); 757 md->md_no = disk->d_no; 758 md->md_syncid = disk->d_sync.ds_syncid; 759 md->md_dflags = (disk->d_flags & G_RAID3_DISK_FLAG_MASK); 760 if (disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 761 md->md_sync_offset = 762 disk->d_sync.ds_offset_done / (sc->sc_ndisks - 1); 763 } 764 if (disk->d_consumer != NULL && disk->d_consumer->provider != NULL) 765 pp = disk->d_consumer->provider; 766 else 767 pp = NULL; 768 if ((disk->d_flags & G_RAID3_DISK_FLAG_HARDCODED) != 0 && pp != NULL) 769 strlcpy(md->md_provider, pp->name, sizeof(md->md_provider)); 770 if (pp != NULL) 771 md->md_provsize = pp->mediasize; 772 } 773 774 void 775 g_raid3_update_metadata(struct g_raid3_disk *disk) 776 { 777 struct g_raid3_softc *sc; 778 struct g_raid3_metadata md; 779 int error; 780 781 g_topology_assert_not(); 782 sc = disk->d_softc; 783 sx_assert(&sc->sc_lock, SX_LOCKED); 784 785 g_raid3_fill_metadata(disk, &md); 786 error = g_raid3_write_metadata(disk, &md); 787 if (error == 0) { 788 G_RAID3_DEBUG(2, "Metadata on %s updated.", 789 g_raid3_get_diskname(disk)); 790 } else { 791 G_RAID3_DEBUG(0, 792 "Cannot update metadata on disk %s (error=%d).", 793 g_raid3_get_diskname(disk), error); 794 } 795 } 796 797 static void 798 g_raid3_bump_syncid(struct g_raid3_softc *sc) 799 { 800 struct g_raid3_disk *disk; 801 u_int n; 802 803 g_topology_assert_not(); 804 sx_assert(&sc->sc_lock, SX_XLOCKED); 805 KASSERT(g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) > 0, 806 ("%s called with no active disks (device=%s).", __func__, 807 sc->sc_name)); 808 809 sc->sc_syncid++; 810 G_RAID3_DEBUG(1, "Device %s: syncid bumped to %u.", sc->sc_name, 811 sc->sc_syncid); 812 for (n = 0; n < sc->sc_ndisks; n++) { 813 disk = &sc->sc_disks[n]; 814 if (disk->d_state == G_RAID3_DISK_STATE_ACTIVE || 815 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 816 disk->d_sync.ds_syncid = sc->sc_syncid; 817 g_raid3_update_metadata(disk); 818 } 819 } 820 } 821 822 static void 823 g_raid3_bump_genid(struct g_raid3_softc *sc) 824 { 825 struct g_raid3_disk *disk; 826 u_int n; 827 828 g_topology_assert_not(); 829 sx_assert(&sc->sc_lock, SX_XLOCKED); 830 KASSERT(g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) > 0, 831 ("%s called with no active disks (device=%s).", __func__, 832 sc->sc_name)); 833 834 sc->sc_genid++; 835 G_RAID3_DEBUG(1, "Device %s: genid bumped to %u.", sc->sc_name, 836 sc->sc_genid); 837 for (n = 0; n < sc->sc_ndisks; n++) { 838 disk = &sc->sc_disks[n]; 839 if (disk->d_state == G_RAID3_DISK_STATE_ACTIVE || 840 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 841 disk->d_genid = sc->sc_genid; 842 g_raid3_update_metadata(disk); 843 } 844 } 845 } 846 847 static int 848 g_raid3_idle(struct g_raid3_softc *sc, int acw) 849 { 850 struct g_raid3_disk *disk; 851 u_int i; 852 int timeout; 853 854 g_topology_assert_not(); 855 sx_assert(&sc->sc_lock, SX_XLOCKED); 856 857 if (sc->sc_provider == NULL) 858 return (0); 859 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_NOFAILSYNC) != 0) 860 return (0); 861 if (sc->sc_idle) 862 return (0); 863 if (sc->sc_writes > 0) 864 return (0); 865 if (acw > 0 || (acw == -1 && sc->sc_provider->acw > 0)) { 866 timeout = g_raid3_idletime - (time_uptime - sc->sc_last_write); 867 if (!g_raid3_shutdown && timeout > 0) 868 return (timeout); 869 } 870 sc->sc_idle = 1; 871 for (i = 0; i < sc->sc_ndisks; i++) { 872 disk = &sc->sc_disks[i]; 873 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) 874 continue; 875 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as clean.", 876 g_raid3_get_diskname(disk), sc->sc_name); 877 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 878 g_raid3_update_metadata(disk); 879 } 880 return (0); 881 } 882 883 static void 884 g_raid3_unidle(struct g_raid3_softc *sc) 885 { 886 struct g_raid3_disk *disk; 887 u_int i; 888 889 g_topology_assert_not(); 890 sx_assert(&sc->sc_lock, SX_XLOCKED); 891 892 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_NOFAILSYNC) != 0) 893 return; 894 sc->sc_idle = 0; 895 sc->sc_last_write = time_uptime; 896 for (i = 0; i < sc->sc_ndisks; i++) { 897 disk = &sc->sc_disks[i]; 898 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) 899 continue; 900 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as dirty.", 901 g_raid3_get_diskname(disk), sc->sc_name); 902 disk->d_flags |= G_RAID3_DISK_FLAG_DIRTY; 903 g_raid3_update_metadata(disk); 904 } 905 } 906 907 /* 908 * Treat bio_driver1 field in parent bio as list head and field bio_caller1 909 * in child bio as pointer to the next element on the list. 910 */ 911 #define G_RAID3_HEAD_BIO(pbp) (pbp)->bio_driver1 912 913 #define G_RAID3_NEXT_BIO(cbp) (cbp)->bio_caller1 914 915 #define G_RAID3_FOREACH_BIO(pbp, bp) \ 916 for ((bp) = G_RAID3_HEAD_BIO(pbp); (bp) != NULL; \ 917 (bp) = G_RAID3_NEXT_BIO(bp)) 918 919 #define G_RAID3_FOREACH_SAFE_BIO(pbp, bp, tmpbp) \ 920 for ((bp) = G_RAID3_HEAD_BIO(pbp); \ 921 (bp) != NULL && ((tmpbp) = G_RAID3_NEXT_BIO(bp), 1); \ 922 (bp) = (tmpbp)) 923 924 static void 925 g_raid3_init_bio(struct bio *pbp) 926 { 927 928 G_RAID3_HEAD_BIO(pbp) = NULL; 929 } 930 931 static void 932 g_raid3_remove_bio(struct bio *cbp) 933 { 934 struct bio *pbp, *bp; 935 936 pbp = cbp->bio_parent; 937 if (G_RAID3_HEAD_BIO(pbp) == cbp) 938 G_RAID3_HEAD_BIO(pbp) = G_RAID3_NEXT_BIO(cbp); 939 else { 940 G_RAID3_FOREACH_BIO(pbp, bp) { 941 if (G_RAID3_NEXT_BIO(bp) == cbp) { 942 G_RAID3_NEXT_BIO(bp) = G_RAID3_NEXT_BIO(cbp); 943 break; 944 } 945 } 946 } 947 G_RAID3_NEXT_BIO(cbp) = NULL; 948 } 949 950 static void 951 g_raid3_replace_bio(struct bio *sbp, struct bio *dbp) 952 { 953 struct bio *pbp, *bp; 954 955 g_raid3_remove_bio(sbp); 956 pbp = dbp->bio_parent; 957 G_RAID3_NEXT_BIO(sbp) = G_RAID3_NEXT_BIO(dbp); 958 if (G_RAID3_HEAD_BIO(pbp) == dbp) 959 G_RAID3_HEAD_BIO(pbp) = sbp; 960 else { 961 G_RAID3_FOREACH_BIO(pbp, bp) { 962 if (G_RAID3_NEXT_BIO(bp) == dbp) { 963 G_RAID3_NEXT_BIO(bp) = sbp; 964 break; 965 } 966 } 967 } 968 G_RAID3_NEXT_BIO(dbp) = NULL; 969 } 970 971 static void 972 g_raid3_destroy_bio(struct g_raid3_softc *sc, struct bio *cbp) 973 { 974 struct bio *bp, *pbp; 975 size_t size; 976 977 pbp = cbp->bio_parent; 978 pbp->bio_children--; 979 KASSERT(cbp->bio_data != NULL, ("NULL bio_data")); 980 size = pbp->bio_length / (sc->sc_ndisks - 1); 981 g_raid3_free(sc, cbp->bio_data, size); 982 if (G_RAID3_HEAD_BIO(pbp) == cbp) { 983 G_RAID3_HEAD_BIO(pbp) = G_RAID3_NEXT_BIO(cbp); 984 G_RAID3_NEXT_BIO(cbp) = NULL; 985 g_destroy_bio(cbp); 986 } else { 987 G_RAID3_FOREACH_BIO(pbp, bp) { 988 if (G_RAID3_NEXT_BIO(bp) == cbp) 989 break; 990 } 991 if (bp != NULL) { 992 KASSERT(G_RAID3_NEXT_BIO(bp) != NULL, 993 ("NULL bp->bio_driver1")); 994 G_RAID3_NEXT_BIO(bp) = G_RAID3_NEXT_BIO(cbp); 995 G_RAID3_NEXT_BIO(cbp) = NULL; 996 } 997 g_destroy_bio(cbp); 998 } 999 } 1000 1001 static struct bio * 1002 g_raid3_clone_bio(struct g_raid3_softc *sc, struct bio *pbp) 1003 { 1004 struct bio *bp, *cbp; 1005 size_t size; 1006 int memflag; 1007 1008 cbp = g_clone_bio(pbp); 1009 if (cbp == NULL) 1010 return (NULL); 1011 size = pbp->bio_length / (sc->sc_ndisks - 1); 1012 if ((pbp->bio_cflags & G_RAID3_BIO_CFLAG_REGULAR) != 0) 1013 memflag = M_WAITOK; 1014 else 1015 memflag = M_NOWAIT; 1016 cbp->bio_data = g_raid3_alloc(sc, size, memflag); 1017 if (cbp->bio_data == NULL) { 1018 pbp->bio_children--; 1019 g_destroy_bio(cbp); 1020 return (NULL); 1021 } 1022 G_RAID3_NEXT_BIO(cbp) = NULL; 1023 if (G_RAID3_HEAD_BIO(pbp) == NULL) 1024 G_RAID3_HEAD_BIO(pbp) = cbp; 1025 else { 1026 G_RAID3_FOREACH_BIO(pbp, bp) { 1027 if (G_RAID3_NEXT_BIO(bp) == NULL) { 1028 G_RAID3_NEXT_BIO(bp) = cbp; 1029 break; 1030 } 1031 } 1032 } 1033 return (cbp); 1034 } 1035 1036 static void 1037 g_raid3_scatter(struct bio *pbp) 1038 { 1039 struct g_raid3_softc *sc; 1040 struct g_raid3_disk *disk; 1041 struct bio *bp, *cbp, *tmpbp; 1042 off_t atom, cadd, padd, left; 1043 int first; 1044 1045 sc = pbp->bio_to->geom->softc; 1046 bp = NULL; 1047 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_NOPARITY) == 0) { 1048 /* 1049 * Find bio for which we should calculate data. 1050 */ 1051 G_RAID3_FOREACH_BIO(pbp, cbp) { 1052 if ((cbp->bio_cflags & G_RAID3_BIO_CFLAG_PARITY) != 0) { 1053 bp = cbp; 1054 break; 1055 } 1056 } 1057 KASSERT(bp != NULL, ("NULL parity bio.")); 1058 } 1059 atom = sc->sc_sectorsize / (sc->sc_ndisks - 1); 1060 cadd = padd = 0; 1061 for (left = pbp->bio_length; left > 0; left -= sc->sc_sectorsize) { 1062 G_RAID3_FOREACH_BIO(pbp, cbp) { 1063 if (cbp == bp) 1064 continue; 1065 bcopy(pbp->bio_data + padd, cbp->bio_data + cadd, atom); 1066 padd += atom; 1067 } 1068 cadd += atom; 1069 } 1070 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_NOPARITY) == 0) { 1071 /* 1072 * Calculate parity. 1073 */ 1074 first = 1; 1075 G_RAID3_FOREACH_SAFE_BIO(pbp, cbp, tmpbp) { 1076 if (cbp == bp) 1077 continue; 1078 if (first) { 1079 bcopy(cbp->bio_data, bp->bio_data, 1080 bp->bio_length); 1081 first = 0; 1082 } else { 1083 g_raid3_xor(cbp->bio_data, bp->bio_data, 1084 bp->bio_length); 1085 } 1086 if ((cbp->bio_cflags & G_RAID3_BIO_CFLAG_NODISK) != 0) 1087 g_raid3_destroy_bio(sc, cbp); 1088 } 1089 } 1090 G_RAID3_FOREACH_SAFE_BIO(pbp, cbp, tmpbp) { 1091 struct g_consumer *cp; 1092 1093 disk = cbp->bio_caller2; 1094 cp = disk->d_consumer; 1095 cbp->bio_to = cp->provider; 1096 G_RAID3_LOGREQ(3, cbp, "Sending request."); 1097 KASSERT(cp->acr >= 1 && cp->acw >= 1 && cp->ace >= 1, 1098 ("Consumer %s not opened (r%dw%de%d).", cp->provider->name, 1099 cp->acr, cp->acw, cp->ace)); 1100 cp->index++; 1101 sc->sc_writes++; 1102 g_io_request(cbp, cp); 1103 } 1104 } 1105 1106 static void 1107 g_raid3_gather(struct bio *pbp) 1108 { 1109 struct g_raid3_softc *sc; 1110 struct g_raid3_disk *disk; 1111 struct bio *xbp, *fbp, *cbp; 1112 off_t atom, cadd, padd, left; 1113 1114 sc = pbp->bio_to->geom->softc; 1115 /* 1116 * Find bio for which we have to calculate data. 1117 * While going through this path, check if all requests 1118 * succeeded, if not, deny whole request. 1119 * If we're in COMPLETE mode, we allow one request to fail, 1120 * so if we find one, we're sending it to the parity consumer. 1121 * If there are more failed requests, we deny whole request. 1122 */ 1123 xbp = fbp = NULL; 1124 G_RAID3_FOREACH_BIO(pbp, cbp) { 1125 if ((cbp->bio_cflags & G_RAID3_BIO_CFLAG_PARITY) != 0) { 1126 KASSERT(xbp == NULL, ("More than one parity bio.")); 1127 xbp = cbp; 1128 } 1129 if (cbp->bio_error == 0) 1130 continue; 1131 /* 1132 * Found failed request. 1133 */ 1134 if (fbp == NULL) { 1135 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_DEGRADED) != 0) { 1136 /* 1137 * We are already in degraded mode, so we can't 1138 * accept any failures. 1139 */ 1140 if (pbp->bio_error == 0) 1141 pbp->bio_error = cbp->bio_error; 1142 } else { 1143 fbp = cbp; 1144 } 1145 } else { 1146 /* 1147 * Next failed request, that's too many. 1148 */ 1149 if (pbp->bio_error == 0) 1150 pbp->bio_error = fbp->bio_error; 1151 } 1152 disk = cbp->bio_caller2; 1153 if (disk == NULL) 1154 continue; 1155 if ((disk->d_flags & G_RAID3_DISK_FLAG_BROKEN) == 0) { 1156 disk->d_flags |= G_RAID3_DISK_FLAG_BROKEN; 1157 G_RAID3_LOGREQ(0, cbp, "Request failed (error=%d).", 1158 cbp->bio_error); 1159 } else { 1160 G_RAID3_LOGREQ(1, cbp, "Request failed (error=%d).", 1161 cbp->bio_error); 1162 } 1163 if (g_raid3_disconnect_on_failure && 1164 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE) { 1165 sc->sc_bump_id |= G_RAID3_BUMP_GENID; 1166 g_raid3_event_send(disk, 1167 G_RAID3_DISK_STATE_DISCONNECTED, 1168 G_RAID3_EVENT_DONTWAIT); 1169 } 1170 } 1171 if (pbp->bio_error != 0) 1172 goto finish; 1173 if (fbp != NULL && (pbp->bio_pflags & G_RAID3_BIO_PFLAG_VERIFY) != 0) { 1174 pbp->bio_pflags &= ~G_RAID3_BIO_PFLAG_VERIFY; 1175 if (xbp != fbp) 1176 g_raid3_replace_bio(xbp, fbp); 1177 g_raid3_destroy_bio(sc, fbp); 1178 } else if (fbp != NULL) { 1179 struct g_consumer *cp; 1180 1181 /* 1182 * One request failed, so send the same request to 1183 * the parity consumer. 1184 */ 1185 disk = pbp->bio_driver2; 1186 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) { 1187 pbp->bio_error = fbp->bio_error; 1188 goto finish; 1189 } 1190 pbp->bio_pflags |= G_RAID3_BIO_PFLAG_DEGRADED; 1191 pbp->bio_inbed--; 1192 fbp->bio_flags &= ~(BIO_DONE | BIO_ERROR); 1193 if (disk->d_no == sc->sc_ndisks - 1) 1194 fbp->bio_cflags |= G_RAID3_BIO_CFLAG_PARITY; 1195 fbp->bio_error = 0; 1196 fbp->bio_completed = 0; 1197 fbp->bio_children = 0; 1198 fbp->bio_inbed = 0; 1199 cp = disk->d_consumer; 1200 fbp->bio_caller2 = disk; 1201 fbp->bio_to = cp->provider; 1202 G_RAID3_LOGREQ(3, fbp, "Sending request (recover)."); 1203 KASSERT(cp->acr >= 1 && cp->acw >= 1 && cp->ace >= 1, 1204 ("Consumer %s not opened (r%dw%de%d).", cp->provider->name, 1205 cp->acr, cp->acw, cp->ace)); 1206 cp->index++; 1207 g_io_request(fbp, cp); 1208 return; 1209 } 1210 if (xbp != NULL) { 1211 /* 1212 * Calculate parity. 1213 */ 1214 G_RAID3_FOREACH_BIO(pbp, cbp) { 1215 if ((cbp->bio_cflags & G_RAID3_BIO_CFLAG_PARITY) != 0) 1216 continue; 1217 g_raid3_xor(cbp->bio_data, xbp->bio_data, 1218 xbp->bio_length); 1219 } 1220 xbp->bio_cflags &= ~G_RAID3_BIO_CFLAG_PARITY; 1221 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_VERIFY) != 0) { 1222 if (!g_raid3_is_zero(xbp)) { 1223 g_raid3_parity_mismatch++; 1224 pbp->bio_error = EIO; 1225 goto finish; 1226 } 1227 g_raid3_destroy_bio(sc, xbp); 1228 } 1229 } 1230 atom = sc->sc_sectorsize / (sc->sc_ndisks - 1); 1231 cadd = padd = 0; 1232 for (left = pbp->bio_length; left > 0; left -= sc->sc_sectorsize) { 1233 G_RAID3_FOREACH_BIO(pbp, cbp) { 1234 bcopy(cbp->bio_data + cadd, pbp->bio_data + padd, atom); 1235 pbp->bio_completed += atom; 1236 padd += atom; 1237 } 1238 cadd += atom; 1239 } 1240 finish: 1241 if (pbp->bio_error == 0) 1242 G_RAID3_LOGREQ(3, pbp, "Request finished."); 1243 else { 1244 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_VERIFY) != 0) 1245 G_RAID3_LOGREQ(1, pbp, "Verification error."); 1246 else 1247 G_RAID3_LOGREQ(0, pbp, "Request failed."); 1248 } 1249 pbp->bio_pflags &= ~G_RAID3_BIO_PFLAG_MASK; 1250 while ((cbp = G_RAID3_HEAD_BIO(pbp)) != NULL) 1251 g_raid3_destroy_bio(sc, cbp); 1252 g_io_deliver(pbp, pbp->bio_error); 1253 } 1254 1255 static void 1256 g_raid3_done(struct bio *bp) 1257 { 1258 struct g_raid3_softc *sc; 1259 1260 sc = bp->bio_from->geom->softc; 1261 bp->bio_cflags |= G_RAID3_BIO_CFLAG_REGULAR; 1262 G_RAID3_LOGREQ(3, bp, "Regular request done (error=%d).", bp->bio_error); 1263 mtx_lock(&sc->sc_queue_mtx); 1264 bioq_insert_head(&sc->sc_queue, bp); 1265 mtx_unlock(&sc->sc_queue_mtx); 1266 wakeup(sc); 1267 wakeup(&sc->sc_queue); 1268 } 1269 1270 static void 1271 g_raid3_regular_request(struct bio *cbp) 1272 { 1273 struct g_raid3_softc *sc; 1274 struct g_raid3_disk *disk; 1275 struct bio *pbp; 1276 1277 g_topology_assert_not(); 1278 1279 pbp = cbp->bio_parent; 1280 sc = pbp->bio_to->geom->softc; 1281 cbp->bio_from->index--; 1282 if (cbp->bio_cmd == BIO_WRITE) 1283 sc->sc_writes--; 1284 disk = cbp->bio_from->private; 1285 if (disk == NULL) { 1286 g_topology_lock(); 1287 g_raid3_kill_consumer(sc, cbp->bio_from); 1288 g_topology_unlock(); 1289 } 1290 1291 G_RAID3_LOGREQ(3, cbp, "Request finished."); 1292 pbp->bio_inbed++; 1293 KASSERT(pbp->bio_inbed <= pbp->bio_children, 1294 ("bio_inbed (%u) is bigger than bio_children (%u).", pbp->bio_inbed, 1295 pbp->bio_children)); 1296 if (pbp->bio_inbed != pbp->bio_children) 1297 return; 1298 switch (pbp->bio_cmd) { 1299 case BIO_READ: 1300 g_raid3_gather(pbp); 1301 break; 1302 case BIO_WRITE: 1303 case BIO_DELETE: 1304 { 1305 int error = 0; 1306 1307 pbp->bio_completed = pbp->bio_length; 1308 while ((cbp = G_RAID3_HEAD_BIO(pbp)) != NULL) { 1309 if (cbp->bio_error == 0) { 1310 g_raid3_destroy_bio(sc, cbp); 1311 continue; 1312 } 1313 1314 if (error == 0) 1315 error = cbp->bio_error; 1316 else if (pbp->bio_error == 0) { 1317 /* 1318 * Next failed request, that's too many. 1319 */ 1320 pbp->bio_error = error; 1321 } 1322 1323 disk = cbp->bio_caller2; 1324 if (disk == NULL) { 1325 g_raid3_destroy_bio(sc, cbp); 1326 continue; 1327 } 1328 1329 if ((disk->d_flags & G_RAID3_DISK_FLAG_BROKEN) == 0) { 1330 disk->d_flags |= G_RAID3_DISK_FLAG_BROKEN; 1331 G_RAID3_LOGREQ(0, cbp, 1332 "Request failed (error=%d).", 1333 cbp->bio_error); 1334 } else { 1335 G_RAID3_LOGREQ(1, cbp, 1336 "Request failed (error=%d).", 1337 cbp->bio_error); 1338 } 1339 if (g_raid3_disconnect_on_failure && 1340 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE) { 1341 sc->sc_bump_id |= G_RAID3_BUMP_GENID; 1342 g_raid3_event_send(disk, 1343 G_RAID3_DISK_STATE_DISCONNECTED, 1344 G_RAID3_EVENT_DONTWAIT); 1345 } 1346 g_raid3_destroy_bio(sc, cbp); 1347 } 1348 if (pbp->bio_error == 0) 1349 G_RAID3_LOGREQ(3, pbp, "Request finished."); 1350 else 1351 G_RAID3_LOGREQ(0, pbp, "Request failed."); 1352 pbp->bio_pflags &= ~G_RAID3_BIO_PFLAG_DEGRADED; 1353 pbp->bio_pflags &= ~G_RAID3_BIO_PFLAG_NOPARITY; 1354 bioq_remove(&sc->sc_inflight, pbp); 1355 /* Release delayed sync requests if possible. */ 1356 g_raid3_sync_release(sc); 1357 g_io_deliver(pbp, pbp->bio_error); 1358 break; 1359 } 1360 } 1361 } 1362 1363 static void 1364 g_raid3_sync_done(struct bio *bp) 1365 { 1366 struct g_raid3_softc *sc; 1367 1368 G_RAID3_LOGREQ(3, bp, "Synchronization request delivered."); 1369 sc = bp->bio_from->geom->softc; 1370 bp->bio_cflags |= G_RAID3_BIO_CFLAG_SYNC; 1371 mtx_lock(&sc->sc_queue_mtx); 1372 bioq_insert_head(&sc->sc_queue, bp); 1373 mtx_unlock(&sc->sc_queue_mtx); 1374 wakeup(sc); 1375 wakeup(&sc->sc_queue); 1376 } 1377 1378 static void 1379 g_raid3_flush(struct g_raid3_softc *sc, struct bio *bp) 1380 { 1381 struct bio_queue_head queue; 1382 struct g_raid3_disk *disk; 1383 struct g_consumer *cp; 1384 struct bio *cbp; 1385 u_int i; 1386 1387 bioq_init(&queue); 1388 for (i = 0; i < sc->sc_ndisks; i++) { 1389 disk = &sc->sc_disks[i]; 1390 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) 1391 continue; 1392 cbp = g_clone_bio(bp); 1393 if (cbp == NULL) { 1394 for (cbp = bioq_first(&queue); cbp != NULL; 1395 cbp = bioq_first(&queue)) { 1396 bioq_remove(&queue, cbp); 1397 g_destroy_bio(cbp); 1398 } 1399 if (bp->bio_error == 0) 1400 bp->bio_error = ENOMEM; 1401 g_io_deliver(bp, bp->bio_error); 1402 return; 1403 } 1404 bioq_insert_tail(&queue, cbp); 1405 cbp->bio_done = g_std_done; 1406 cbp->bio_caller1 = disk; 1407 cbp->bio_to = disk->d_consumer->provider; 1408 } 1409 for (cbp = bioq_first(&queue); cbp != NULL; cbp = bioq_first(&queue)) { 1410 bioq_remove(&queue, cbp); 1411 G_RAID3_LOGREQ(3, cbp, "Sending request."); 1412 disk = cbp->bio_caller1; 1413 cbp->bio_caller1 = NULL; 1414 cp = disk->d_consumer; 1415 KASSERT(cp->acr >= 1 && cp->acw >= 1 && cp->ace >= 1, 1416 ("Consumer %s not opened (r%dw%de%d).", cp->provider->name, 1417 cp->acr, cp->acw, cp->ace)); 1418 g_io_request(cbp, disk->d_consumer); 1419 } 1420 } 1421 1422 static void 1423 g_raid3_start(struct bio *bp) 1424 { 1425 struct g_raid3_softc *sc; 1426 1427 sc = bp->bio_to->geom->softc; 1428 /* 1429 * If sc == NULL or there are no valid disks, provider's error 1430 * should be set and g_raid3_start() should not be called at all. 1431 */ 1432 KASSERT(sc != NULL && (sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 1433 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE), 1434 ("Provider's error should be set (error=%d)(device=%s).", 1435 bp->bio_to->error, bp->bio_to->name)); 1436 G_RAID3_LOGREQ(3, bp, "Request received."); 1437 1438 switch (bp->bio_cmd) { 1439 case BIO_READ: 1440 case BIO_WRITE: 1441 case BIO_DELETE: 1442 break; 1443 case BIO_SPEEDUP: 1444 case BIO_FLUSH: 1445 g_raid3_flush(sc, bp); 1446 return; 1447 case BIO_GETATTR: 1448 default: 1449 g_io_deliver(bp, EOPNOTSUPP); 1450 return; 1451 } 1452 mtx_lock(&sc->sc_queue_mtx); 1453 bioq_insert_tail(&sc->sc_queue, bp); 1454 mtx_unlock(&sc->sc_queue_mtx); 1455 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, sc); 1456 wakeup(sc); 1457 } 1458 1459 /* 1460 * Return TRUE if the given request is colliding with a in-progress 1461 * synchronization request. 1462 */ 1463 static int 1464 g_raid3_sync_collision(struct g_raid3_softc *sc, struct bio *bp) 1465 { 1466 struct g_raid3_disk *disk; 1467 struct bio *sbp; 1468 off_t rstart, rend, sstart, send; 1469 int i; 1470 1471 disk = sc->sc_syncdisk; 1472 if (disk == NULL) 1473 return (0); 1474 rstart = bp->bio_offset; 1475 rend = bp->bio_offset + bp->bio_length; 1476 for (i = 0; i < g_raid3_syncreqs; i++) { 1477 sbp = disk->d_sync.ds_bios[i]; 1478 if (sbp == NULL) 1479 continue; 1480 sstart = sbp->bio_offset; 1481 send = sbp->bio_length; 1482 if (sbp->bio_cmd == BIO_WRITE) { 1483 sstart *= sc->sc_ndisks - 1; 1484 send *= sc->sc_ndisks - 1; 1485 } 1486 send += sstart; 1487 if (rend > sstart && rstart < send) 1488 return (1); 1489 } 1490 return (0); 1491 } 1492 1493 /* 1494 * Return TRUE if the given sync request is colliding with a in-progress regular 1495 * request. 1496 */ 1497 static int 1498 g_raid3_regular_collision(struct g_raid3_softc *sc, struct bio *sbp) 1499 { 1500 off_t rstart, rend, sstart, send; 1501 struct bio *bp; 1502 1503 if (sc->sc_syncdisk == NULL) 1504 return (0); 1505 sstart = sbp->bio_offset; 1506 send = sstart + sbp->bio_length; 1507 TAILQ_FOREACH(bp, &sc->sc_inflight.queue, bio_queue) { 1508 rstart = bp->bio_offset; 1509 rend = bp->bio_offset + bp->bio_length; 1510 if (rend > sstart && rstart < send) 1511 return (1); 1512 } 1513 return (0); 1514 } 1515 1516 /* 1517 * Puts request onto delayed queue. 1518 */ 1519 static void 1520 g_raid3_regular_delay(struct g_raid3_softc *sc, struct bio *bp) 1521 { 1522 1523 G_RAID3_LOGREQ(2, bp, "Delaying request."); 1524 bioq_insert_head(&sc->sc_regular_delayed, bp); 1525 } 1526 1527 /* 1528 * Puts synchronization request onto delayed queue. 1529 */ 1530 static void 1531 g_raid3_sync_delay(struct g_raid3_softc *sc, struct bio *bp) 1532 { 1533 1534 G_RAID3_LOGREQ(2, bp, "Delaying synchronization request."); 1535 bioq_insert_tail(&sc->sc_sync_delayed, bp); 1536 } 1537 1538 /* 1539 * Releases delayed regular requests which don't collide anymore with sync 1540 * requests. 1541 */ 1542 static void 1543 g_raid3_regular_release(struct g_raid3_softc *sc) 1544 { 1545 struct bio *bp, *bp2; 1546 1547 TAILQ_FOREACH_SAFE(bp, &sc->sc_regular_delayed.queue, bio_queue, bp2) { 1548 if (g_raid3_sync_collision(sc, bp)) 1549 continue; 1550 bioq_remove(&sc->sc_regular_delayed, bp); 1551 G_RAID3_LOGREQ(2, bp, "Releasing delayed request (%p).", bp); 1552 mtx_lock(&sc->sc_queue_mtx); 1553 bioq_insert_head(&sc->sc_queue, bp); 1554 #if 0 1555 /* 1556 * wakeup() is not needed, because this function is called from 1557 * the worker thread. 1558 */ 1559 wakeup(&sc->sc_queue); 1560 #endif 1561 mtx_unlock(&sc->sc_queue_mtx); 1562 } 1563 } 1564 1565 /* 1566 * Releases delayed sync requests which don't collide anymore with regular 1567 * requests. 1568 */ 1569 static void 1570 g_raid3_sync_release(struct g_raid3_softc *sc) 1571 { 1572 struct bio *bp, *bp2; 1573 1574 TAILQ_FOREACH_SAFE(bp, &sc->sc_sync_delayed.queue, bio_queue, bp2) { 1575 if (g_raid3_regular_collision(sc, bp)) 1576 continue; 1577 bioq_remove(&sc->sc_sync_delayed, bp); 1578 G_RAID3_LOGREQ(2, bp, 1579 "Releasing delayed synchronization request."); 1580 g_io_request(bp, bp->bio_from); 1581 } 1582 } 1583 1584 /* 1585 * Handle synchronization requests. 1586 * Every synchronization request is two-steps process: first, READ request is 1587 * send to active provider and then WRITE request (with read data) to the provider 1588 * being synchronized. When WRITE is finished, new synchronization request is 1589 * send. 1590 */ 1591 static void 1592 g_raid3_sync_request(struct bio *bp) 1593 { 1594 struct g_raid3_softc *sc; 1595 struct g_raid3_disk *disk; 1596 1597 bp->bio_from->index--; 1598 sc = bp->bio_from->geom->softc; 1599 disk = bp->bio_from->private; 1600 if (disk == NULL) { 1601 sx_xunlock(&sc->sc_lock); /* Avoid recursion on sc_lock. */ 1602 g_topology_lock(); 1603 g_raid3_kill_consumer(sc, bp->bio_from); 1604 g_topology_unlock(); 1605 free(bp->bio_data, M_RAID3); 1606 g_destroy_bio(bp); 1607 sx_xlock(&sc->sc_lock); 1608 return; 1609 } 1610 1611 /* 1612 * Synchronization request. 1613 */ 1614 switch (bp->bio_cmd) { 1615 case BIO_READ: 1616 { 1617 struct g_consumer *cp; 1618 u_char *dst, *src; 1619 off_t left; 1620 u_int atom; 1621 1622 if (bp->bio_error != 0) { 1623 G_RAID3_LOGREQ(0, bp, 1624 "Synchronization request failed (error=%d).", 1625 bp->bio_error); 1626 g_destroy_bio(bp); 1627 return; 1628 } 1629 G_RAID3_LOGREQ(3, bp, "Synchronization request finished."); 1630 atom = sc->sc_sectorsize / (sc->sc_ndisks - 1); 1631 dst = src = bp->bio_data; 1632 if (disk->d_no == sc->sc_ndisks - 1) { 1633 u_int n; 1634 1635 /* Parity component. */ 1636 for (left = bp->bio_length; left > 0; 1637 left -= sc->sc_sectorsize) { 1638 bcopy(src, dst, atom); 1639 src += atom; 1640 for (n = 1; n < sc->sc_ndisks - 1; n++) { 1641 g_raid3_xor(src, dst, atom); 1642 src += atom; 1643 } 1644 dst += atom; 1645 } 1646 } else { 1647 /* Regular component. */ 1648 src += atom * disk->d_no; 1649 for (left = bp->bio_length; left > 0; 1650 left -= sc->sc_sectorsize) { 1651 bcopy(src, dst, atom); 1652 src += sc->sc_sectorsize; 1653 dst += atom; 1654 } 1655 } 1656 bp->bio_driver1 = bp->bio_driver2 = NULL; 1657 bp->bio_pflags = 0; 1658 bp->bio_offset /= sc->sc_ndisks - 1; 1659 bp->bio_length /= sc->sc_ndisks - 1; 1660 bp->bio_cmd = BIO_WRITE; 1661 bp->bio_cflags = 0; 1662 bp->bio_children = bp->bio_inbed = 0; 1663 cp = disk->d_consumer; 1664 KASSERT(cp->acr >= 1 && cp->acw >= 1 && cp->ace >= 1, 1665 ("Consumer %s not opened (r%dw%de%d).", cp->provider->name, 1666 cp->acr, cp->acw, cp->ace)); 1667 cp->index++; 1668 g_io_request(bp, cp); 1669 return; 1670 } 1671 case BIO_WRITE: 1672 { 1673 struct g_raid3_disk_sync *sync; 1674 off_t boffset, moffset; 1675 void *data; 1676 int i; 1677 1678 if (bp->bio_error != 0) { 1679 G_RAID3_LOGREQ(0, bp, 1680 "Synchronization request failed (error=%d).", 1681 bp->bio_error); 1682 g_destroy_bio(bp); 1683 sc->sc_bump_id |= G_RAID3_BUMP_GENID; 1684 g_raid3_event_send(disk, 1685 G_RAID3_DISK_STATE_DISCONNECTED, 1686 G_RAID3_EVENT_DONTWAIT); 1687 return; 1688 } 1689 G_RAID3_LOGREQ(3, bp, "Synchronization request finished."); 1690 sync = &disk->d_sync; 1691 if (sync->ds_offset == sc->sc_mediasize / (sc->sc_ndisks - 1) || 1692 sync->ds_consumer == NULL || 1693 (sc->sc_flags & G_RAID3_DEVICE_FLAG_DESTROY) != 0) { 1694 /* Don't send more synchronization requests. */ 1695 sync->ds_inflight--; 1696 if (sync->ds_bios != NULL) { 1697 i = (int)(uintptr_t)bp->bio_caller1; 1698 sync->ds_bios[i] = NULL; 1699 } 1700 free(bp->bio_data, M_RAID3); 1701 g_destroy_bio(bp); 1702 if (sync->ds_inflight > 0) 1703 return; 1704 if (sync->ds_consumer == NULL || 1705 (sc->sc_flags & G_RAID3_DEVICE_FLAG_DESTROY) != 0) { 1706 return; 1707 } 1708 /* 1709 * Disk up-to-date, activate it. 1710 */ 1711 g_raid3_event_send(disk, G_RAID3_DISK_STATE_ACTIVE, 1712 G_RAID3_EVENT_DONTWAIT); 1713 return; 1714 } 1715 1716 /* Send next synchronization request. */ 1717 data = bp->bio_data; 1718 g_reset_bio(bp); 1719 bp->bio_cmd = BIO_READ; 1720 bp->bio_offset = sync->ds_offset * (sc->sc_ndisks - 1); 1721 bp->bio_length = MIN(maxphys, sc->sc_mediasize - bp->bio_offset); 1722 sync->ds_offset += bp->bio_length / (sc->sc_ndisks - 1); 1723 bp->bio_done = g_raid3_sync_done; 1724 bp->bio_data = data; 1725 bp->bio_from = sync->ds_consumer; 1726 bp->bio_to = sc->sc_provider; 1727 G_RAID3_LOGREQ(3, bp, "Sending synchronization request."); 1728 sync->ds_consumer->index++; 1729 /* 1730 * Delay the request if it is colliding with a regular request. 1731 */ 1732 if (g_raid3_regular_collision(sc, bp)) 1733 g_raid3_sync_delay(sc, bp); 1734 else 1735 g_io_request(bp, sync->ds_consumer); 1736 1737 /* Release delayed requests if possible. */ 1738 g_raid3_regular_release(sc); 1739 1740 /* Find the smallest offset. */ 1741 moffset = sc->sc_mediasize; 1742 for (i = 0; i < g_raid3_syncreqs; i++) { 1743 bp = sync->ds_bios[i]; 1744 boffset = bp->bio_offset; 1745 if (bp->bio_cmd == BIO_WRITE) 1746 boffset *= sc->sc_ndisks - 1; 1747 if (boffset < moffset) 1748 moffset = boffset; 1749 } 1750 if (sync->ds_offset_done + maxphys * 100 < moffset) { 1751 /* Update offset_done on every 100 blocks. */ 1752 sync->ds_offset_done = moffset; 1753 g_raid3_update_metadata(disk); 1754 } 1755 return; 1756 } 1757 default: 1758 KASSERT(1 == 0, ("Invalid command here: %u (device=%s)", 1759 bp->bio_cmd, sc->sc_name)); 1760 break; 1761 } 1762 } 1763 1764 static int 1765 g_raid3_register_request(struct bio *pbp) 1766 { 1767 struct g_raid3_softc *sc; 1768 struct g_raid3_disk *disk; 1769 struct g_consumer *cp; 1770 struct bio *cbp, *tmpbp; 1771 off_t offset, length; 1772 u_int n, ndisks; 1773 int round_robin, verify; 1774 1775 ndisks = 0; 1776 sc = pbp->bio_to->geom->softc; 1777 if ((pbp->bio_cflags & G_RAID3_BIO_CFLAG_REGSYNC) != 0 && 1778 sc->sc_syncdisk == NULL) { 1779 g_io_deliver(pbp, EIO); 1780 return (0); 1781 } 1782 g_raid3_init_bio(pbp); 1783 length = pbp->bio_length / (sc->sc_ndisks - 1); 1784 offset = pbp->bio_offset / (sc->sc_ndisks - 1); 1785 round_robin = verify = 0; 1786 switch (pbp->bio_cmd) { 1787 case BIO_READ: 1788 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_VERIFY) != 0 && 1789 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE) { 1790 pbp->bio_pflags |= G_RAID3_BIO_PFLAG_VERIFY; 1791 verify = 1; 1792 ndisks = sc->sc_ndisks; 1793 } else { 1794 verify = 0; 1795 ndisks = sc->sc_ndisks - 1; 1796 } 1797 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_ROUND_ROBIN) != 0 && 1798 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE) { 1799 round_robin = 1; 1800 } else { 1801 round_robin = 0; 1802 } 1803 KASSERT(!round_robin || !verify, 1804 ("ROUND-ROBIN and VERIFY are mutually exclusive.")); 1805 pbp->bio_driver2 = &sc->sc_disks[sc->sc_ndisks - 1]; 1806 break; 1807 case BIO_WRITE: 1808 case BIO_DELETE: 1809 /* 1810 * Delay the request if it is colliding with a synchronization 1811 * request. 1812 */ 1813 if (g_raid3_sync_collision(sc, pbp)) { 1814 g_raid3_regular_delay(sc, pbp); 1815 return (0); 1816 } 1817 1818 if (sc->sc_idle) 1819 g_raid3_unidle(sc); 1820 else 1821 sc->sc_last_write = time_uptime; 1822 1823 ndisks = sc->sc_ndisks; 1824 break; 1825 } 1826 for (n = 0; n < ndisks; n++) { 1827 disk = &sc->sc_disks[n]; 1828 cbp = g_raid3_clone_bio(sc, pbp); 1829 if (cbp == NULL) { 1830 while ((cbp = G_RAID3_HEAD_BIO(pbp)) != NULL) 1831 g_raid3_destroy_bio(sc, cbp); 1832 /* 1833 * To prevent deadlock, we must run back up 1834 * with the ENOMEM for failed requests of any 1835 * of our consumers. Our own sync requests 1836 * can stick around, as they are finite. 1837 */ 1838 if ((pbp->bio_cflags & 1839 G_RAID3_BIO_CFLAG_REGULAR) != 0) { 1840 g_io_deliver(pbp, ENOMEM); 1841 return (0); 1842 } 1843 return (ENOMEM); 1844 } 1845 cbp->bio_offset = offset; 1846 cbp->bio_length = length; 1847 cbp->bio_done = g_raid3_done; 1848 switch (pbp->bio_cmd) { 1849 case BIO_READ: 1850 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) { 1851 /* 1852 * Replace invalid component with the parity 1853 * component. 1854 */ 1855 disk = &sc->sc_disks[sc->sc_ndisks - 1]; 1856 cbp->bio_cflags |= G_RAID3_BIO_CFLAG_PARITY; 1857 pbp->bio_pflags |= G_RAID3_BIO_PFLAG_DEGRADED; 1858 } else if (round_robin && 1859 disk->d_no == sc->sc_round_robin) { 1860 /* 1861 * In round-robin mode skip one data component 1862 * and use parity component when reading. 1863 */ 1864 pbp->bio_driver2 = disk; 1865 disk = &sc->sc_disks[sc->sc_ndisks - 1]; 1866 cbp->bio_cflags |= G_RAID3_BIO_CFLAG_PARITY; 1867 sc->sc_round_robin++; 1868 round_robin = 0; 1869 } else if (verify && disk->d_no == sc->sc_ndisks - 1) { 1870 cbp->bio_cflags |= G_RAID3_BIO_CFLAG_PARITY; 1871 } 1872 break; 1873 case BIO_WRITE: 1874 case BIO_DELETE: 1875 if (disk->d_state == G_RAID3_DISK_STATE_ACTIVE || 1876 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 1877 if (n == ndisks - 1) { 1878 /* 1879 * Active parity component, mark it as such. 1880 */ 1881 cbp->bio_cflags |= 1882 G_RAID3_BIO_CFLAG_PARITY; 1883 } 1884 } else { 1885 pbp->bio_pflags |= G_RAID3_BIO_PFLAG_DEGRADED; 1886 if (n == ndisks - 1) { 1887 /* 1888 * Parity component is not connected, 1889 * so destroy its request. 1890 */ 1891 pbp->bio_pflags |= 1892 G_RAID3_BIO_PFLAG_NOPARITY; 1893 g_raid3_destroy_bio(sc, cbp); 1894 cbp = NULL; 1895 } else { 1896 cbp->bio_cflags |= 1897 G_RAID3_BIO_CFLAG_NODISK; 1898 disk = NULL; 1899 } 1900 } 1901 break; 1902 } 1903 if (cbp != NULL) 1904 cbp->bio_caller2 = disk; 1905 } 1906 switch (pbp->bio_cmd) { 1907 case BIO_READ: 1908 if (round_robin) { 1909 /* 1910 * If we are in round-robin mode and 'round_robin' is 1911 * still 1, it means, that we skipped parity component 1912 * for this read and must reset sc_round_robin field. 1913 */ 1914 sc->sc_round_robin = 0; 1915 } 1916 G_RAID3_FOREACH_SAFE_BIO(pbp, cbp, tmpbp) { 1917 disk = cbp->bio_caller2; 1918 cp = disk->d_consumer; 1919 cbp->bio_to = cp->provider; 1920 G_RAID3_LOGREQ(3, cbp, "Sending request."); 1921 KASSERT(cp->acr >= 1 && cp->acw >= 1 && cp->ace >= 1, 1922 ("Consumer %s not opened (r%dw%de%d).", 1923 cp->provider->name, cp->acr, cp->acw, cp->ace)); 1924 cp->index++; 1925 g_io_request(cbp, cp); 1926 } 1927 break; 1928 case BIO_WRITE: 1929 case BIO_DELETE: 1930 /* 1931 * Put request onto inflight queue, so we can check if new 1932 * synchronization requests don't collide with it. 1933 */ 1934 bioq_insert_tail(&sc->sc_inflight, pbp); 1935 1936 /* 1937 * Bump syncid on first write. 1938 */ 1939 if ((sc->sc_bump_id & G_RAID3_BUMP_SYNCID) != 0) { 1940 sc->sc_bump_id &= ~G_RAID3_BUMP_SYNCID; 1941 g_raid3_bump_syncid(sc); 1942 } 1943 g_raid3_scatter(pbp); 1944 break; 1945 } 1946 return (0); 1947 } 1948 1949 static int 1950 g_raid3_can_destroy(struct g_raid3_softc *sc) 1951 { 1952 struct g_geom *gp; 1953 struct g_consumer *cp; 1954 1955 g_topology_assert(); 1956 gp = sc->sc_geom; 1957 if (gp->softc == NULL) 1958 return (1); 1959 LIST_FOREACH(cp, &gp->consumer, consumer) { 1960 if (g_raid3_is_busy(sc, cp)) 1961 return (0); 1962 } 1963 gp = sc->sc_sync.ds_geom; 1964 LIST_FOREACH(cp, &gp->consumer, consumer) { 1965 if (g_raid3_is_busy(sc, cp)) 1966 return (0); 1967 } 1968 G_RAID3_DEBUG(2, "No I/O requests for %s, it can be destroyed.", 1969 sc->sc_name); 1970 return (1); 1971 } 1972 1973 static int 1974 g_raid3_try_destroy(struct g_raid3_softc *sc) 1975 { 1976 1977 g_topology_assert_not(); 1978 sx_assert(&sc->sc_lock, SX_XLOCKED); 1979 1980 if (sc->sc_rootmount != NULL) { 1981 G_RAID3_DEBUG(1, "root_mount_rel[%u] %p", __LINE__, 1982 sc->sc_rootmount); 1983 root_mount_rel(sc->sc_rootmount); 1984 sc->sc_rootmount = NULL; 1985 } 1986 1987 g_topology_lock(); 1988 if (!g_raid3_can_destroy(sc)) { 1989 g_topology_unlock(); 1990 return (0); 1991 } 1992 sc->sc_geom->softc = NULL; 1993 sc->sc_sync.ds_geom->softc = NULL; 1994 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_WAIT) != 0) { 1995 g_topology_unlock(); 1996 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, 1997 &sc->sc_worker); 1998 /* Unlock sc_lock here, as it can be destroyed after wakeup. */ 1999 sx_xunlock(&sc->sc_lock); 2000 wakeup(&sc->sc_worker); 2001 sc->sc_worker = NULL; 2002 } else { 2003 g_topology_unlock(); 2004 g_raid3_destroy_device(sc); 2005 free(sc->sc_disks, M_RAID3); 2006 free(sc, M_RAID3); 2007 } 2008 return (1); 2009 } 2010 2011 /* 2012 * Worker thread. 2013 */ 2014 static void 2015 g_raid3_worker(void *arg) 2016 { 2017 struct g_raid3_softc *sc; 2018 struct g_raid3_event *ep; 2019 struct bio *bp; 2020 int timeout; 2021 2022 sc = arg; 2023 thread_lock(curthread); 2024 sched_prio(curthread, PRIBIO); 2025 thread_unlock(curthread); 2026 2027 sx_xlock(&sc->sc_lock); 2028 for (;;) { 2029 G_RAID3_DEBUG(5, "%s: Let's see...", __func__); 2030 /* 2031 * First take a look at events. 2032 * This is important to handle events before any I/O requests. 2033 */ 2034 ep = g_raid3_event_get(sc); 2035 if (ep != NULL) { 2036 g_raid3_event_remove(sc, ep); 2037 if ((ep->e_flags & G_RAID3_EVENT_DEVICE) != 0) { 2038 /* Update only device status. */ 2039 G_RAID3_DEBUG(3, 2040 "Running event for device %s.", 2041 sc->sc_name); 2042 ep->e_error = 0; 2043 g_raid3_update_device(sc, 1); 2044 } else { 2045 /* Update disk status. */ 2046 G_RAID3_DEBUG(3, "Running event for disk %s.", 2047 g_raid3_get_diskname(ep->e_disk)); 2048 ep->e_error = g_raid3_update_disk(ep->e_disk, 2049 ep->e_state); 2050 if (ep->e_error == 0) 2051 g_raid3_update_device(sc, 0); 2052 } 2053 if ((ep->e_flags & G_RAID3_EVENT_DONTWAIT) != 0) { 2054 KASSERT(ep->e_error == 0, 2055 ("Error cannot be handled.")); 2056 g_raid3_event_free(ep); 2057 } else { 2058 ep->e_flags |= G_RAID3_EVENT_DONE; 2059 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, 2060 ep); 2061 mtx_lock(&sc->sc_events_mtx); 2062 wakeup(ep); 2063 mtx_unlock(&sc->sc_events_mtx); 2064 } 2065 if ((sc->sc_flags & 2066 G_RAID3_DEVICE_FLAG_DESTROY) != 0) { 2067 if (g_raid3_try_destroy(sc)) { 2068 curthread->td_pflags &= ~TDP_GEOM; 2069 G_RAID3_DEBUG(1, "Thread exiting."); 2070 kproc_exit(0); 2071 } 2072 } 2073 G_RAID3_DEBUG(5, "%s: I'm here 1.", __func__); 2074 continue; 2075 } 2076 /* 2077 * Check if we can mark array as CLEAN and if we can't take 2078 * how much seconds should we wait. 2079 */ 2080 timeout = g_raid3_idle(sc, -1); 2081 /* 2082 * Now I/O requests. 2083 */ 2084 /* Get first request from the queue. */ 2085 mtx_lock(&sc->sc_queue_mtx); 2086 bp = bioq_first(&sc->sc_queue); 2087 if (bp == NULL) { 2088 if ((sc->sc_flags & 2089 G_RAID3_DEVICE_FLAG_DESTROY) != 0) { 2090 mtx_unlock(&sc->sc_queue_mtx); 2091 if (g_raid3_try_destroy(sc)) { 2092 curthread->td_pflags &= ~TDP_GEOM; 2093 G_RAID3_DEBUG(1, "Thread exiting."); 2094 kproc_exit(0); 2095 } 2096 mtx_lock(&sc->sc_queue_mtx); 2097 } 2098 sx_xunlock(&sc->sc_lock); 2099 /* 2100 * XXX: We can miss an event here, because an event 2101 * can be added without sx-device-lock and without 2102 * mtx-queue-lock. Maybe I should just stop using 2103 * dedicated mutex for events synchronization and 2104 * stick with the queue lock? 2105 * The event will hang here until next I/O request 2106 * or next event is received. 2107 */ 2108 MSLEEP(sc, &sc->sc_queue_mtx, PRIBIO | PDROP, "r3:w1", 2109 timeout * hz); 2110 sx_xlock(&sc->sc_lock); 2111 G_RAID3_DEBUG(5, "%s: I'm here 4.", __func__); 2112 continue; 2113 } 2114 process: 2115 bioq_remove(&sc->sc_queue, bp); 2116 mtx_unlock(&sc->sc_queue_mtx); 2117 2118 if (bp->bio_from->geom == sc->sc_sync.ds_geom && 2119 (bp->bio_cflags & G_RAID3_BIO_CFLAG_SYNC) != 0) { 2120 g_raid3_sync_request(bp); /* READ */ 2121 } else if (bp->bio_to != sc->sc_provider) { 2122 if ((bp->bio_cflags & G_RAID3_BIO_CFLAG_REGULAR) != 0) 2123 g_raid3_regular_request(bp); 2124 else if ((bp->bio_cflags & G_RAID3_BIO_CFLAG_SYNC) != 0) 2125 g_raid3_sync_request(bp); /* WRITE */ 2126 else { 2127 KASSERT(0, 2128 ("Invalid request cflags=0x%hx to=%s.", 2129 bp->bio_cflags, bp->bio_to->name)); 2130 } 2131 } else if (g_raid3_register_request(bp) != 0) { 2132 mtx_lock(&sc->sc_queue_mtx); 2133 bioq_insert_head(&sc->sc_queue, bp); 2134 /* 2135 * We are short in memory, let see if there are finished 2136 * request we can free. 2137 */ 2138 TAILQ_FOREACH(bp, &sc->sc_queue.queue, bio_queue) { 2139 if (bp->bio_cflags & G_RAID3_BIO_CFLAG_REGULAR) 2140 goto process; 2141 } 2142 /* 2143 * No finished regular request, so at least keep 2144 * synchronization running. 2145 */ 2146 TAILQ_FOREACH(bp, &sc->sc_queue.queue, bio_queue) { 2147 if (bp->bio_cflags & G_RAID3_BIO_CFLAG_SYNC) 2148 goto process; 2149 } 2150 sx_xunlock(&sc->sc_lock); 2151 MSLEEP(&sc->sc_queue, &sc->sc_queue_mtx, PRIBIO | PDROP, 2152 "r3:lowmem", hz / 10); 2153 sx_xlock(&sc->sc_lock); 2154 } 2155 G_RAID3_DEBUG(5, "%s: I'm here 9.", __func__); 2156 } 2157 } 2158 2159 static void 2160 g_raid3_update_idle(struct g_raid3_softc *sc, struct g_raid3_disk *disk) 2161 { 2162 2163 sx_assert(&sc->sc_lock, SX_LOCKED); 2164 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_NOFAILSYNC) != 0) 2165 return; 2166 if (!sc->sc_idle && (disk->d_flags & G_RAID3_DISK_FLAG_DIRTY) == 0) { 2167 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as dirty.", 2168 g_raid3_get_diskname(disk), sc->sc_name); 2169 disk->d_flags |= G_RAID3_DISK_FLAG_DIRTY; 2170 } else if (sc->sc_idle && 2171 (disk->d_flags & G_RAID3_DISK_FLAG_DIRTY) != 0) { 2172 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as clean.", 2173 g_raid3_get_diskname(disk), sc->sc_name); 2174 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 2175 } 2176 } 2177 2178 static void 2179 g_raid3_sync_start(struct g_raid3_softc *sc) 2180 { 2181 struct g_raid3_disk *disk; 2182 struct g_consumer *cp; 2183 struct bio *bp; 2184 int error; 2185 u_int n; 2186 2187 g_topology_assert_not(); 2188 sx_assert(&sc->sc_lock, SX_XLOCKED); 2189 2190 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED, 2191 ("Device not in DEGRADED state (%s, %u).", sc->sc_name, 2192 sc->sc_state)); 2193 KASSERT(sc->sc_syncdisk == NULL, ("Syncdisk is not NULL (%s, %u).", 2194 sc->sc_name, sc->sc_state)); 2195 disk = NULL; 2196 for (n = 0; n < sc->sc_ndisks; n++) { 2197 if (sc->sc_disks[n].d_state != G_RAID3_DISK_STATE_SYNCHRONIZING) 2198 continue; 2199 disk = &sc->sc_disks[n]; 2200 break; 2201 } 2202 if (disk == NULL) 2203 return; 2204 2205 sx_xunlock(&sc->sc_lock); 2206 g_topology_lock(); 2207 cp = g_new_consumer(sc->sc_sync.ds_geom); 2208 error = g_attach(cp, sc->sc_provider); 2209 KASSERT(error == 0, 2210 ("Cannot attach to %s (error=%d).", sc->sc_name, error)); 2211 error = g_access(cp, 1, 0, 0); 2212 KASSERT(error == 0, ("Cannot open %s (error=%d).", sc->sc_name, error)); 2213 g_topology_unlock(); 2214 sx_xlock(&sc->sc_lock); 2215 2216 G_RAID3_DEBUG(0, "Device %s: rebuilding provider %s.", sc->sc_name, 2217 g_raid3_get_diskname(disk)); 2218 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_NOFAILSYNC) == 0) 2219 disk->d_flags |= G_RAID3_DISK_FLAG_DIRTY; 2220 KASSERT(disk->d_sync.ds_consumer == NULL, 2221 ("Sync consumer already exists (device=%s, disk=%s).", 2222 sc->sc_name, g_raid3_get_diskname(disk))); 2223 2224 disk->d_sync.ds_consumer = cp; 2225 disk->d_sync.ds_consumer->private = disk; 2226 disk->d_sync.ds_consumer->index = 0; 2227 sc->sc_syncdisk = disk; 2228 2229 /* 2230 * Allocate memory for synchronization bios and initialize them. 2231 */ 2232 disk->d_sync.ds_bios = malloc(sizeof(struct bio *) * g_raid3_syncreqs, 2233 M_RAID3, M_WAITOK); 2234 for (n = 0; n < g_raid3_syncreqs; n++) { 2235 bp = g_alloc_bio(); 2236 disk->d_sync.ds_bios[n] = bp; 2237 bp->bio_parent = NULL; 2238 bp->bio_cmd = BIO_READ; 2239 bp->bio_data = malloc(maxphys, M_RAID3, M_WAITOK); 2240 bp->bio_cflags = 0; 2241 bp->bio_offset = disk->d_sync.ds_offset * (sc->sc_ndisks - 1); 2242 bp->bio_length = MIN(maxphys, sc->sc_mediasize - bp->bio_offset); 2243 disk->d_sync.ds_offset += bp->bio_length / (sc->sc_ndisks - 1); 2244 bp->bio_done = g_raid3_sync_done; 2245 bp->bio_from = disk->d_sync.ds_consumer; 2246 bp->bio_to = sc->sc_provider; 2247 bp->bio_caller1 = (void *)(uintptr_t)n; 2248 } 2249 2250 /* Set the number of in-flight synchronization requests. */ 2251 disk->d_sync.ds_inflight = g_raid3_syncreqs; 2252 2253 /* 2254 * Fire off first synchronization requests. 2255 */ 2256 for (n = 0; n < g_raid3_syncreqs; n++) { 2257 bp = disk->d_sync.ds_bios[n]; 2258 G_RAID3_LOGREQ(3, bp, "Sending synchronization request."); 2259 disk->d_sync.ds_consumer->index++; 2260 /* 2261 * Delay the request if it is colliding with a regular request. 2262 */ 2263 if (g_raid3_regular_collision(sc, bp)) 2264 g_raid3_sync_delay(sc, bp); 2265 else 2266 g_io_request(bp, disk->d_sync.ds_consumer); 2267 } 2268 } 2269 2270 /* 2271 * Stop synchronization process. 2272 * type: 0 - synchronization finished 2273 * 1 - synchronization stopped 2274 */ 2275 static void 2276 g_raid3_sync_stop(struct g_raid3_softc *sc, int type) 2277 { 2278 struct g_raid3_disk *disk; 2279 struct g_consumer *cp; 2280 2281 g_topology_assert_not(); 2282 sx_assert(&sc->sc_lock, SX_LOCKED); 2283 2284 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED, 2285 ("Device not in DEGRADED state (%s, %u).", sc->sc_name, 2286 sc->sc_state)); 2287 disk = sc->sc_syncdisk; 2288 sc->sc_syncdisk = NULL; 2289 KASSERT(disk != NULL, ("No disk was synchronized (%s).", sc->sc_name)); 2290 KASSERT(disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING, 2291 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 2292 g_raid3_disk_state2str(disk->d_state))); 2293 if (disk->d_sync.ds_consumer == NULL) 2294 return; 2295 2296 if (type == 0) { 2297 G_RAID3_DEBUG(0, "Device %s: rebuilding provider %s finished.", 2298 sc->sc_name, g_raid3_get_diskname(disk)); 2299 } else /* if (type == 1) */ { 2300 G_RAID3_DEBUG(0, "Device %s: rebuilding provider %s stopped.", 2301 sc->sc_name, g_raid3_get_diskname(disk)); 2302 } 2303 free(disk->d_sync.ds_bios, M_RAID3); 2304 disk->d_sync.ds_bios = NULL; 2305 cp = disk->d_sync.ds_consumer; 2306 disk->d_sync.ds_consumer = NULL; 2307 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 2308 sx_xunlock(&sc->sc_lock); /* Avoid recursion on sc_lock. */ 2309 g_topology_lock(); 2310 g_raid3_kill_consumer(sc, cp); 2311 g_topology_unlock(); 2312 sx_xlock(&sc->sc_lock); 2313 } 2314 2315 static void 2316 g_raid3_launch_provider(struct g_raid3_softc *sc) 2317 { 2318 struct g_provider *pp; 2319 struct g_raid3_disk *disk; 2320 int n; 2321 2322 sx_assert(&sc->sc_lock, SX_LOCKED); 2323 2324 g_topology_lock(); 2325 pp = g_new_providerf(sc->sc_geom, "raid3/%s", sc->sc_name); 2326 pp->mediasize = sc->sc_mediasize; 2327 pp->sectorsize = sc->sc_sectorsize; 2328 pp->stripesize = 0; 2329 pp->stripeoffset = 0; 2330 for (n = 0; n < sc->sc_ndisks; n++) { 2331 disk = &sc->sc_disks[n]; 2332 if (disk->d_consumer && disk->d_consumer->provider && 2333 disk->d_consumer->provider->stripesize > pp->stripesize) { 2334 pp->stripesize = disk->d_consumer->provider->stripesize; 2335 pp->stripeoffset = disk->d_consumer->provider->stripeoffset; 2336 } 2337 } 2338 pp->stripesize *= sc->sc_ndisks - 1; 2339 pp->stripeoffset *= sc->sc_ndisks - 1; 2340 sc->sc_provider = pp; 2341 g_error_provider(pp, 0); 2342 g_topology_unlock(); 2343 G_RAID3_DEBUG(0, "Device %s launched (%u/%u).", pp->name, 2344 g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE), sc->sc_ndisks); 2345 2346 if (sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED) 2347 g_raid3_sync_start(sc); 2348 } 2349 2350 static void 2351 g_raid3_destroy_provider(struct g_raid3_softc *sc) 2352 { 2353 struct bio *bp; 2354 2355 g_topology_assert_not(); 2356 KASSERT(sc->sc_provider != NULL, ("NULL provider (device=%s).", 2357 sc->sc_name)); 2358 2359 g_topology_lock(); 2360 g_error_provider(sc->sc_provider, ENXIO); 2361 mtx_lock(&sc->sc_queue_mtx); 2362 while ((bp = bioq_first(&sc->sc_queue)) != NULL) { 2363 bioq_remove(&sc->sc_queue, bp); 2364 g_io_deliver(bp, ENXIO); 2365 } 2366 mtx_unlock(&sc->sc_queue_mtx); 2367 G_RAID3_DEBUG(0, "Device %s: provider %s destroyed.", sc->sc_name, 2368 sc->sc_provider->name); 2369 g_wither_provider(sc->sc_provider, ENXIO); 2370 g_topology_unlock(); 2371 sc->sc_provider = NULL; 2372 if (sc->sc_syncdisk != NULL) 2373 g_raid3_sync_stop(sc, 1); 2374 } 2375 2376 static void 2377 g_raid3_go(void *arg) 2378 { 2379 struct g_raid3_softc *sc; 2380 2381 sc = arg; 2382 G_RAID3_DEBUG(0, "Force device %s start due to timeout.", sc->sc_name); 2383 g_raid3_event_send(sc, 0, 2384 G_RAID3_EVENT_DONTWAIT | G_RAID3_EVENT_DEVICE); 2385 } 2386 2387 static u_int 2388 g_raid3_determine_state(struct g_raid3_disk *disk) 2389 { 2390 struct g_raid3_softc *sc; 2391 u_int state; 2392 2393 sc = disk->d_softc; 2394 if (sc->sc_syncid == disk->d_sync.ds_syncid) { 2395 if ((disk->d_flags & 2396 G_RAID3_DISK_FLAG_SYNCHRONIZING) == 0) { 2397 /* Disk does not need synchronization. */ 2398 state = G_RAID3_DISK_STATE_ACTIVE; 2399 } else { 2400 if ((sc->sc_flags & 2401 G_RAID3_DEVICE_FLAG_NOAUTOSYNC) == 0 || 2402 (disk->d_flags & 2403 G_RAID3_DISK_FLAG_FORCE_SYNC) != 0) { 2404 /* 2405 * We can start synchronization from 2406 * the stored offset. 2407 */ 2408 state = G_RAID3_DISK_STATE_SYNCHRONIZING; 2409 } else { 2410 state = G_RAID3_DISK_STATE_STALE; 2411 } 2412 } 2413 } else if (disk->d_sync.ds_syncid < sc->sc_syncid) { 2414 /* 2415 * Reset all synchronization data for this disk, 2416 * because if it even was synchronized, it was 2417 * synchronized to disks with different syncid. 2418 */ 2419 disk->d_flags |= G_RAID3_DISK_FLAG_SYNCHRONIZING; 2420 disk->d_sync.ds_offset = 0; 2421 disk->d_sync.ds_offset_done = 0; 2422 disk->d_sync.ds_syncid = sc->sc_syncid; 2423 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_NOAUTOSYNC) == 0 || 2424 (disk->d_flags & G_RAID3_DISK_FLAG_FORCE_SYNC) != 0) { 2425 state = G_RAID3_DISK_STATE_SYNCHRONIZING; 2426 } else { 2427 state = G_RAID3_DISK_STATE_STALE; 2428 } 2429 } else /* if (sc->sc_syncid < disk->d_sync.ds_syncid) */ { 2430 /* 2431 * Not good, NOT GOOD! 2432 * It means that device was started on stale disks 2433 * and more fresh disk just arrive. 2434 * If there were writes, device is broken, sorry. 2435 * I think the best choice here is don't touch 2436 * this disk and inform the user loudly. 2437 */ 2438 G_RAID3_DEBUG(0, "Device %s was started before the freshest " 2439 "disk (%s) arrives!! It will not be connected to the " 2440 "running device.", sc->sc_name, 2441 g_raid3_get_diskname(disk)); 2442 g_raid3_destroy_disk(disk); 2443 state = G_RAID3_DISK_STATE_NONE; 2444 /* Return immediately, because disk was destroyed. */ 2445 return (state); 2446 } 2447 G_RAID3_DEBUG(3, "State for %s disk: %s.", 2448 g_raid3_get_diskname(disk), g_raid3_disk_state2str(state)); 2449 return (state); 2450 } 2451 2452 /* 2453 * Update device state. 2454 */ 2455 static void 2456 g_raid3_update_device(struct g_raid3_softc *sc, boolean_t force) 2457 { 2458 struct g_raid3_disk *disk; 2459 u_int state; 2460 2461 sx_assert(&sc->sc_lock, SX_XLOCKED); 2462 2463 switch (sc->sc_state) { 2464 case G_RAID3_DEVICE_STATE_STARTING: 2465 { 2466 u_int n, ndirty, ndisks, genid, syncid; 2467 2468 KASSERT(sc->sc_provider == NULL, 2469 ("Non-NULL provider in STARTING state (%s).", sc->sc_name)); 2470 /* 2471 * Are we ready? We are, if all disks are connected or 2472 * one disk is missing and 'force' is true. 2473 */ 2474 if (g_raid3_ndisks(sc, -1) + force == sc->sc_ndisks) { 2475 if (!force) 2476 callout_drain(&sc->sc_callout); 2477 } else { 2478 if (force) { 2479 /* 2480 * Timeout expired, so destroy device. 2481 */ 2482 sc->sc_flags |= G_RAID3_DEVICE_FLAG_DESTROY; 2483 G_RAID3_DEBUG(1, "root_mount_rel[%u] %p", 2484 __LINE__, sc->sc_rootmount); 2485 root_mount_rel(sc->sc_rootmount); 2486 sc->sc_rootmount = NULL; 2487 } 2488 return; 2489 } 2490 2491 /* 2492 * Find the biggest genid. 2493 */ 2494 genid = 0; 2495 for (n = 0; n < sc->sc_ndisks; n++) { 2496 disk = &sc->sc_disks[n]; 2497 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 2498 continue; 2499 if (disk->d_genid > genid) 2500 genid = disk->d_genid; 2501 } 2502 sc->sc_genid = genid; 2503 /* 2504 * Remove all disks without the biggest genid. 2505 */ 2506 for (n = 0; n < sc->sc_ndisks; n++) { 2507 disk = &sc->sc_disks[n]; 2508 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 2509 continue; 2510 if (disk->d_genid < genid) { 2511 G_RAID3_DEBUG(0, 2512 "Component %s (device %s) broken, skipping.", 2513 g_raid3_get_diskname(disk), sc->sc_name); 2514 g_raid3_destroy_disk(disk); 2515 } 2516 } 2517 2518 /* 2519 * There must be at least 'sc->sc_ndisks - 1' components 2520 * with the same syncid and without SYNCHRONIZING flag. 2521 */ 2522 2523 /* 2524 * Find the biggest syncid, number of valid components and 2525 * number of dirty components. 2526 */ 2527 ndirty = ndisks = syncid = 0; 2528 for (n = 0; n < sc->sc_ndisks; n++) { 2529 disk = &sc->sc_disks[n]; 2530 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 2531 continue; 2532 if ((disk->d_flags & G_RAID3_DISK_FLAG_DIRTY) != 0) 2533 ndirty++; 2534 if (disk->d_sync.ds_syncid > syncid) { 2535 syncid = disk->d_sync.ds_syncid; 2536 ndisks = 0; 2537 } else if (disk->d_sync.ds_syncid < syncid) { 2538 continue; 2539 } 2540 if ((disk->d_flags & 2541 G_RAID3_DISK_FLAG_SYNCHRONIZING) != 0) { 2542 continue; 2543 } 2544 ndisks++; 2545 } 2546 /* 2547 * Do we have enough valid components? 2548 */ 2549 if (ndisks + 1 < sc->sc_ndisks) { 2550 G_RAID3_DEBUG(0, 2551 "Device %s is broken, too few valid components.", 2552 sc->sc_name); 2553 sc->sc_flags |= G_RAID3_DEVICE_FLAG_DESTROY; 2554 return; 2555 } 2556 /* 2557 * If there is one DIRTY component and all disks are present, 2558 * mark it for synchronization. If there is more than one DIRTY 2559 * component, mark parity component for synchronization. 2560 */ 2561 if (ndisks == sc->sc_ndisks && ndirty == 1) { 2562 for (n = 0; n < sc->sc_ndisks; n++) { 2563 disk = &sc->sc_disks[n]; 2564 if ((disk->d_flags & 2565 G_RAID3_DISK_FLAG_DIRTY) == 0) { 2566 continue; 2567 } 2568 disk->d_flags |= 2569 G_RAID3_DISK_FLAG_SYNCHRONIZING; 2570 } 2571 } else if (ndisks == sc->sc_ndisks && ndirty > 1) { 2572 disk = &sc->sc_disks[sc->sc_ndisks - 1]; 2573 disk->d_flags |= G_RAID3_DISK_FLAG_SYNCHRONIZING; 2574 } 2575 2576 sc->sc_syncid = syncid; 2577 if (force) { 2578 /* Remember to bump syncid on first write. */ 2579 sc->sc_bump_id |= G_RAID3_BUMP_SYNCID; 2580 } 2581 if (ndisks == sc->sc_ndisks) 2582 state = G_RAID3_DEVICE_STATE_COMPLETE; 2583 else /* if (ndisks == sc->sc_ndisks - 1) */ 2584 state = G_RAID3_DEVICE_STATE_DEGRADED; 2585 G_RAID3_DEBUG(1, "Device %s state changed from %s to %s.", 2586 sc->sc_name, g_raid3_device_state2str(sc->sc_state), 2587 g_raid3_device_state2str(state)); 2588 sc->sc_state = state; 2589 for (n = 0; n < sc->sc_ndisks; n++) { 2590 disk = &sc->sc_disks[n]; 2591 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 2592 continue; 2593 state = g_raid3_determine_state(disk); 2594 g_raid3_event_send(disk, state, G_RAID3_EVENT_DONTWAIT); 2595 if (state == G_RAID3_DISK_STATE_STALE) 2596 sc->sc_bump_id |= G_RAID3_BUMP_SYNCID; 2597 } 2598 break; 2599 } 2600 case G_RAID3_DEVICE_STATE_DEGRADED: 2601 /* 2602 * Genid need to be bumped immediately, so do it here. 2603 */ 2604 if ((sc->sc_bump_id & G_RAID3_BUMP_GENID) != 0) { 2605 sc->sc_bump_id &= ~G_RAID3_BUMP_GENID; 2606 g_raid3_bump_genid(sc); 2607 } 2608 2609 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_NEW) > 0) 2610 return; 2611 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) < 2612 sc->sc_ndisks - 1) { 2613 if (sc->sc_provider != NULL) 2614 g_raid3_destroy_provider(sc); 2615 sc->sc_flags |= G_RAID3_DEVICE_FLAG_DESTROY; 2616 return; 2617 } 2618 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) == 2619 sc->sc_ndisks) { 2620 state = G_RAID3_DEVICE_STATE_COMPLETE; 2621 G_RAID3_DEBUG(1, 2622 "Device %s state changed from %s to %s.", 2623 sc->sc_name, g_raid3_device_state2str(sc->sc_state), 2624 g_raid3_device_state2str(state)); 2625 sc->sc_state = state; 2626 } 2627 if (sc->sc_provider == NULL) 2628 g_raid3_launch_provider(sc); 2629 if (sc->sc_rootmount != NULL) { 2630 G_RAID3_DEBUG(1, "root_mount_rel[%u] %p", __LINE__, 2631 sc->sc_rootmount); 2632 root_mount_rel(sc->sc_rootmount); 2633 sc->sc_rootmount = NULL; 2634 } 2635 break; 2636 case G_RAID3_DEVICE_STATE_COMPLETE: 2637 /* 2638 * Genid need to be bumped immediately, so do it here. 2639 */ 2640 if ((sc->sc_bump_id & G_RAID3_BUMP_GENID) != 0) { 2641 sc->sc_bump_id &= ~G_RAID3_BUMP_GENID; 2642 g_raid3_bump_genid(sc); 2643 } 2644 2645 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_NEW) > 0) 2646 return; 2647 KASSERT(g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) >= 2648 sc->sc_ndisks - 1, 2649 ("Too few ACTIVE components in COMPLETE state (device %s).", 2650 sc->sc_name)); 2651 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) == 2652 sc->sc_ndisks - 1) { 2653 state = G_RAID3_DEVICE_STATE_DEGRADED; 2654 G_RAID3_DEBUG(1, 2655 "Device %s state changed from %s to %s.", 2656 sc->sc_name, g_raid3_device_state2str(sc->sc_state), 2657 g_raid3_device_state2str(state)); 2658 sc->sc_state = state; 2659 } 2660 if (sc->sc_provider == NULL) 2661 g_raid3_launch_provider(sc); 2662 if (sc->sc_rootmount != NULL) { 2663 G_RAID3_DEBUG(1, "root_mount_rel[%u] %p", __LINE__, 2664 sc->sc_rootmount); 2665 root_mount_rel(sc->sc_rootmount); 2666 sc->sc_rootmount = NULL; 2667 } 2668 break; 2669 default: 2670 KASSERT(1 == 0, ("Wrong device state (%s, %s).", sc->sc_name, 2671 g_raid3_device_state2str(sc->sc_state))); 2672 break; 2673 } 2674 } 2675 2676 /* 2677 * Update disk state and device state if needed. 2678 */ 2679 #define DISK_STATE_CHANGED() G_RAID3_DEBUG(1, \ 2680 "Disk %s state changed from %s to %s (device %s).", \ 2681 g_raid3_get_diskname(disk), \ 2682 g_raid3_disk_state2str(disk->d_state), \ 2683 g_raid3_disk_state2str(state), sc->sc_name) 2684 static int 2685 g_raid3_update_disk(struct g_raid3_disk *disk, u_int state) 2686 { 2687 struct g_raid3_softc *sc; 2688 2689 sc = disk->d_softc; 2690 sx_assert(&sc->sc_lock, SX_XLOCKED); 2691 2692 again: 2693 G_RAID3_DEBUG(3, "Changing disk %s state from %s to %s.", 2694 g_raid3_get_diskname(disk), g_raid3_disk_state2str(disk->d_state), 2695 g_raid3_disk_state2str(state)); 2696 switch (state) { 2697 case G_RAID3_DISK_STATE_NEW: 2698 /* 2699 * Possible scenarios: 2700 * 1. New disk arrive. 2701 */ 2702 /* Previous state should be NONE. */ 2703 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NONE, 2704 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 2705 g_raid3_disk_state2str(disk->d_state))); 2706 DISK_STATE_CHANGED(); 2707 2708 disk->d_state = state; 2709 G_RAID3_DEBUG(1, "Device %s: provider %s detected.", 2710 sc->sc_name, g_raid3_get_diskname(disk)); 2711 if (sc->sc_state == G_RAID3_DEVICE_STATE_STARTING) 2712 break; 2713 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2714 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE, 2715 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2716 g_raid3_device_state2str(sc->sc_state), 2717 g_raid3_get_diskname(disk), 2718 g_raid3_disk_state2str(disk->d_state))); 2719 state = g_raid3_determine_state(disk); 2720 if (state != G_RAID3_DISK_STATE_NONE) 2721 goto again; 2722 break; 2723 case G_RAID3_DISK_STATE_ACTIVE: 2724 /* 2725 * Possible scenarios: 2726 * 1. New disk does not need synchronization. 2727 * 2. Synchronization process finished successfully. 2728 */ 2729 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2730 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE, 2731 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2732 g_raid3_device_state2str(sc->sc_state), 2733 g_raid3_get_diskname(disk), 2734 g_raid3_disk_state2str(disk->d_state))); 2735 /* Previous state should be NEW or SYNCHRONIZING. */ 2736 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NEW || 2737 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING, 2738 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 2739 g_raid3_disk_state2str(disk->d_state))); 2740 DISK_STATE_CHANGED(); 2741 2742 if (disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 2743 disk->d_flags &= ~G_RAID3_DISK_FLAG_SYNCHRONIZING; 2744 disk->d_flags &= ~G_RAID3_DISK_FLAG_FORCE_SYNC; 2745 g_raid3_sync_stop(sc, 0); 2746 } 2747 disk->d_state = state; 2748 disk->d_sync.ds_offset = 0; 2749 disk->d_sync.ds_offset_done = 0; 2750 g_raid3_update_idle(sc, disk); 2751 g_raid3_update_metadata(disk); 2752 G_RAID3_DEBUG(1, "Device %s: provider %s activated.", 2753 sc->sc_name, g_raid3_get_diskname(disk)); 2754 break; 2755 case G_RAID3_DISK_STATE_STALE: 2756 /* 2757 * Possible scenarios: 2758 * 1. Stale disk was connected. 2759 */ 2760 /* Previous state should be NEW. */ 2761 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NEW, 2762 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 2763 g_raid3_disk_state2str(disk->d_state))); 2764 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2765 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE, 2766 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2767 g_raid3_device_state2str(sc->sc_state), 2768 g_raid3_get_diskname(disk), 2769 g_raid3_disk_state2str(disk->d_state))); 2770 /* 2771 * STALE state is only possible if device is marked 2772 * NOAUTOSYNC. 2773 */ 2774 KASSERT((sc->sc_flags & G_RAID3_DEVICE_FLAG_NOAUTOSYNC) != 0, 2775 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2776 g_raid3_device_state2str(sc->sc_state), 2777 g_raid3_get_diskname(disk), 2778 g_raid3_disk_state2str(disk->d_state))); 2779 DISK_STATE_CHANGED(); 2780 2781 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 2782 disk->d_state = state; 2783 g_raid3_update_metadata(disk); 2784 G_RAID3_DEBUG(0, "Device %s: provider %s is stale.", 2785 sc->sc_name, g_raid3_get_diskname(disk)); 2786 break; 2787 case G_RAID3_DISK_STATE_SYNCHRONIZING: 2788 /* 2789 * Possible scenarios: 2790 * 1. Disk which needs synchronization was connected. 2791 */ 2792 /* Previous state should be NEW. */ 2793 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NEW, 2794 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 2795 g_raid3_disk_state2str(disk->d_state))); 2796 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2797 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE, 2798 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2799 g_raid3_device_state2str(sc->sc_state), 2800 g_raid3_get_diskname(disk), 2801 g_raid3_disk_state2str(disk->d_state))); 2802 DISK_STATE_CHANGED(); 2803 2804 if (disk->d_state == G_RAID3_DISK_STATE_NEW) 2805 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 2806 disk->d_state = state; 2807 if (sc->sc_provider != NULL) { 2808 g_raid3_sync_start(sc); 2809 g_raid3_update_metadata(disk); 2810 } 2811 break; 2812 case G_RAID3_DISK_STATE_DISCONNECTED: 2813 /* 2814 * Possible scenarios: 2815 * 1. Device wasn't running yet, but disk disappear. 2816 * 2. Disk was active and disapppear. 2817 * 3. Disk disappear during synchronization process. 2818 */ 2819 if (sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2820 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE) { 2821 /* 2822 * Previous state should be ACTIVE, STALE or 2823 * SYNCHRONIZING. 2824 */ 2825 KASSERT(disk->d_state == G_RAID3_DISK_STATE_ACTIVE || 2826 disk->d_state == G_RAID3_DISK_STATE_STALE || 2827 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING, 2828 ("Wrong disk state (%s, %s).", 2829 g_raid3_get_diskname(disk), 2830 g_raid3_disk_state2str(disk->d_state))); 2831 } else if (sc->sc_state == G_RAID3_DEVICE_STATE_STARTING) { 2832 /* Previous state should be NEW. */ 2833 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NEW, 2834 ("Wrong disk state (%s, %s).", 2835 g_raid3_get_diskname(disk), 2836 g_raid3_disk_state2str(disk->d_state))); 2837 /* 2838 * Reset bumping syncid if disk disappeared in STARTING 2839 * state. 2840 */ 2841 if ((sc->sc_bump_id & G_RAID3_BUMP_SYNCID) != 0) 2842 sc->sc_bump_id &= ~G_RAID3_BUMP_SYNCID; 2843 #ifdef INVARIANTS 2844 } else { 2845 KASSERT(1 == 0, ("Wrong device state (%s, %s, %s, %s).", 2846 sc->sc_name, 2847 g_raid3_device_state2str(sc->sc_state), 2848 g_raid3_get_diskname(disk), 2849 g_raid3_disk_state2str(disk->d_state))); 2850 #endif 2851 } 2852 DISK_STATE_CHANGED(); 2853 G_RAID3_DEBUG(0, "Device %s: provider %s disconnected.", 2854 sc->sc_name, g_raid3_get_diskname(disk)); 2855 2856 g_raid3_destroy_disk(disk); 2857 break; 2858 default: 2859 KASSERT(1 == 0, ("Unknown state (%u).", state)); 2860 break; 2861 } 2862 return (0); 2863 } 2864 #undef DISK_STATE_CHANGED 2865 2866 int 2867 g_raid3_read_metadata(struct g_consumer *cp, struct g_raid3_metadata *md) 2868 { 2869 struct g_provider *pp; 2870 u_char *buf; 2871 int error; 2872 2873 g_topology_assert(); 2874 2875 error = g_access(cp, 1, 0, 0); 2876 if (error != 0) 2877 return (error); 2878 pp = cp->provider; 2879 g_topology_unlock(); 2880 /* Metadata are stored on last sector. */ 2881 buf = g_read_data(cp, pp->mediasize - pp->sectorsize, pp->sectorsize, 2882 &error); 2883 g_topology_lock(); 2884 g_access(cp, -1, 0, 0); 2885 if (buf == NULL) { 2886 G_RAID3_DEBUG(1, "Cannot read metadata from %s (error=%d).", 2887 cp->provider->name, error); 2888 return (error); 2889 } 2890 2891 /* Decode metadata. */ 2892 error = raid3_metadata_decode(buf, md); 2893 g_free(buf); 2894 if (strcmp(md->md_magic, G_RAID3_MAGIC) != 0) 2895 return (EINVAL); 2896 if (md->md_version > G_RAID3_VERSION) { 2897 G_RAID3_DEBUG(0, 2898 "Kernel module is too old to handle metadata from %s.", 2899 cp->provider->name); 2900 return (EINVAL); 2901 } 2902 if (error != 0) { 2903 G_RAID3_DEBUG(1, "MD5 metadata hash mismatch for provider %s.", 2904 cp->provider->name); 2905 return (error); 2906 } 2907 if (md->md_sectorsize > maxphys) { 2908 G_RAID3_DEBUG(0, "The blocksize is too big."); 2909 return (EINVAL); 2910 } 2911 2912 return (0); 2913 } 2914 2915 static int 2916 g_raid3_check_metadata(struct g_raid3_softc *sc, struct g_provider *pp, 2917 struct g_raid3_metadata *md) 2918 { 2919 2920 if (md->md_no >= sc->sc_ndisks) { 2921 G_RAID3_DEBUG(1, "Invalid disk %s number (no=%u), skipping.", 2922 pp->name, md->md_no); 2923 return (EINVAL); 2924 } 2925 if (sc->sc_disks[md->md_no].d_state != G_RAID3_DISK_STATE_NODISK) { 2926 G_RAID3_DEBUG(1, "Disk %s (no=%u) already exists, skipping.", 2927 pp->name, md->md_no); 2928 return (EEXIST); 2929 } 2930 if (md->md_all != sc->sc_ndisks) { 2931 G_RAID3_DEBUG(1, 2932 "Invalid '%s' field on disk %s (device %s), skipping.", 2933 "md_all", pp->name, sc->sc_name); 2934 return (EINVAL); 2935 } 2936 if ((md->md_mediasize % md->md_sectorsize) != 0) { 2937 G_RAID3_DEBUG(1, "Invalid metadata (mediasize %% sectorsize != " 2938 "0) on disk %s (device %s), skipping.", pp->name, 2939 sc->sc_name); 2940 return (EINVAL); 2941 } 2942 if (md->md_mediasize != sc->sc_mediasize) { 2943 G_RAID3_DEBUG(1, 2944 "Invalid '%s' field on disk %s (device %s), skipping.", 2945 "md_mediasize", pp->name, sc->sc_name); 2946 return (EINVAL); 2947 } 2948 if ((md->md_mediasize % (sc->sc_ndisks - 1)) != 0) { 2949 G_RAID3_DEBUG(1, 2950 "Invalid '%s' field on disk %s (device %s), skipping.", 2951 "md_mediasize", pp->name, sc->sc_name); 2952 return (EINVAL); 2953 } 2954 if ((sc->sc_mediasize / (sc->sc_ndisks - 1)) > pp->mediasize) { 2955 G_RAID3_DEBUG(1, 2956 "Invalid size of disk %s (device %s), skipping.", pp->name, 2957 sc->sc_name); 2958 return (EINVAL); 2959 } 2960 if ((md->md_sectorsize / pp->sectorsize) < sc->sc_ndisks - 1) { 2961 G_RAID3_DEBUG(1, 2962 "Invalid '%s' field on disk %s (device %s), skipping.", 2963 "md_sectorsize", pp->name, sc->sc_name); 2964 return (EINVAL); 2965 } 2966 if (md->md_sectorsize != sc->sc_sectorsize) { 2967 G_RAID3_DEBUG(1, 2968 "Invalid '%s' field on disk %s (device %s), skipping.", 2969 "md_sectorsize", pp->name, sc->sc_name); 2970 return (EINVAL); 2971 } 2972 if ((sc->sc_sectorsize % pp->sectorsize) != 0) { 2973 G_RAID3_DEBUG(1, 2974 "Invalid sector size of disk %s (device %s), skipping.", 2975 pp->name, sc->sc_name); 2976 return (EINVAL); 2977 } 2978 if ((md->md_mflags & ~G_RAID3_DEVICE_FLAG_MASK) != 0) { 2979 G_RAID3_DEBUG(1, 2980 "Invalid device flags on disk %s (device %s), skipping.", 2981 pp->name, sc->sc_name); 2982 return (EINVAL); 2983 } 2984 if ((md->md_mflags & G_RAID3_DEVICE_FLAG_VERIFY) != 0 && 2985 (md->md_mflags & G_RAID3_DEVICE_FLAG_ROUND_ROBIN) != 0) { 2986 /* 2987 * VERIFY and ROUND-ROBIN options are mutally exclusive. 2988 */ 2989 G_RAID3_DEBUG(1, "Both VERIFY and ROUND-ROBIN flags exist on " 2990 "disk %s (device %s), skipping.", pp->name, sc->sc_name); 2991 return (EINVAL); 2992 } 2993 if ((md->md_dflags & ~G_RAID3_DISK_FLAG_MASK) != 0) { 2994 G_RAID3_DEBUG(1, 2995 "Invalid disk flags on disk %s (device %s), skipping.", 2996 pp->name, sc->sc_name); 2997 return (EINVAL); 2998 } 2999 return (0); 3000 } 3001 3002 int 3003 g_raid3_add_disk(struct g_raid3_softc *sc, struct g_provider *pp, 3004 struct g_raid3_metadata *md) 3005 { 3006 struct g_raid3_disk *disk; 3007 int error; 3008 3009 g_topology_assert_not(); 3010 G_RAID3_DEBUG(2, "Adding disk %s.", pp->name); 3011 3012 error = g_raid3_check_metadata(sc, pp, md); 3013 if (error != 0) 3014 return (error); 3015 if (sc->sc_state != G_RAID3_DEVICE_STATE_STARTING && 3016 md->md_genid < sc->sc_genid) { 3017 G_RAID3_DEBUG(0, "Component %s (device %s) broken, skipping.", 3018 pp->name, sc->sc_name); 3019 return (EINVAL); 3020 } 3021 disk = g_raid3_init_disk(sc, pp, md, &error); 3022 if (disk == NULL) 3023 return (error); 3024 error = g_raid3_event_send(disk, G_RAID3_DISK_STATE_NEW, 3025 G_RAID3_EVENT_WAIT); 3026 if (error != 0) 3027 return (error); 3028 if (md->md_version < G_RAID3_VERSION) { 3029 G_RAID3_DEBUG(0, "Upgrading metadata on %s (v%d->v%d).", 3030 pp->name, md->md_version, G_RAID3_VERSION); 3031 g_raid3_update_metadata(disk); 3032 } 3033 return (0); 3034 } 3035 3036 static void 3037 g_raid3_destroy_delayed(void *arg, int flag) 3038 { 3039 struct g_raid3_softc *sc; 3040 int error; 3041 3042 if (flag == EV_CANCEL) { 3043 G_RAID3_DEBUG(1, "Destroying canceled."); 3044 return; 3045 } 3046 sc = arg; 3047 g_topology_unlock(); 3048 sx_xlock(&sc->sc_lock); 3049 KASSERT((sc->sc_flags & G_RAID3_DEVICE_FLAG_DESTROY) == 0, 3050 ("DESTROY flag set on %s.", sc->sc_name)); 3051 KASSERT((sc->sc_flags & G_RAID3_DEVICE_FLAG_DESTROYING) != 0, 3052 ("DESTROYING flag not set on %s.", sc->sc_name)); 3053 G_RAID3_DEBUG(0, "Destroying %s (delayed).", sc->sc_name); 3054 error = g_raid3_destroy(sc, G_RAID3_DESTROY_SOFT); 3055 if (error != 0) { 3056 G_RAID3_DEBUG(0, "Cannot destroy %s.", sc->sc_name); 3057 sx_xunlock(&sc->sc_lock); 3058 } 3059 g_topology_lock(); 3060 } 3061 3062 static int 3063 g_raid3_access(struct g_provider *pp, int acr, int acw, int ace) 3064 { 3065 struct g_raid3_softc *sc; 3066 int dcr, dcw, dce, error = 0; 3067 3068 g_topology_assert(); 3069 G_RAID3_DEBUG(2, "Access request for %s: r%dw%de%d.", pp->name, acr, 3070 acw, ace); 3071 3072 sc = pp->geom->softc; 3073 if (sc == NULL && acr <= 0 && acw <= 0 && ace <= 0) 3074 return (0); 3075 KASSERT(sc != NULL, ("NULL softc (provider=%s).", pp->name)); 3076 3077 dcr = pp->acr + acr; 3078 dcw = pp->acw + acw; 3079 dce = pp->ace + ace; 3080 3081 g_topology_unlock(); 3082 sx_xlock(&sc->sc_lock); 3083 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_DESTROY) != 0 || 3084 g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) < sc->sc_ndisks - 1) { 3085 if (acr > 0 || acw > 0 || ace > 0) 3086 error = ENXIO; 3087 goto end; 3088 } 3089 if (dcw == 0) 3090 g_raid3_idle(sc, dcw); 3091 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_DESTROYING) != 0) { 3092 if (acr > 0 || acw > 0 || ace > 0) { 3093 error = ENXIO; 3094 goto end; 3095 } 3096 if (dcr == 0 && dcw == 0 && dce == 0) { 3097 g_post_event(g_raid3_destroy_delayed, sc, M_WAITOK, 3098 sc, NULL); 3099 } 3100 } 3101 end: 3102 sx_xunlock(&sc->sc_lock); 3103 g_topology_lock(); 3104 return (error); 3105 } 3106 3107 static struct g_geom * 3108 g_raid3_create(struct g_class *mp, const struct g_raid3_metadata *md) 3109 { 3110 struct g_raid3_softc *sc; 3111 struct g_geom *gp; 3112 int error, timeout; 3113 u_int n; 3114 3115 g_topology_assert(); 3116 G_RAID3_DEBUG(1, "Creating device %s (id=%u).", md->md_name, md->md_id); 3117 3118 /* One disk is minimum. */ 3119 if (md->md_all < 1) 3120 return (NULL); 3121 /* 3122 * Action geom. 3123 */ 3124 gp = g_new_geomf(mp, "%s", md->md_name); 3125 sc = malloc(sizeof(*sc), M_RAID3, M_WAITOK | M_ZERO); 3126 sc->sc_disks = malloc(sizeof(struct g_raid3_disk) * md->md_all, M_RAID3, 3127 M_WAITOK | M_ZERO); 3128 gp->start = g_raid3_start; 3129 gp->orphan = g_raid3_orphan; 3130 gp->access = g_raid3_access; 3131 gp->dumpconf = g_raid3_dumpconf; 3132 3133 sc->sc_id = md->md_id; 3134 sc->sc_mediasize = md->md_mediasize; 3135 sc->sc_sectorsize = md->md_sectorsize; 3136 sc->sc_ndisks = md->md_all; 3137 sc->sc_round_robin = 0; 3138 sc->sc_flags = md->md_mflags; 3139 sc->sc_bump_id = 0; 3140 sc->sc_idle = 1; 3141 sc->sc_last_write = time_uptime; 3142 sc->sc_writes = 0; 3143 for (n = 0; n < sc->sc_ndisks; n++) { 3144 sc->sc_disks[n].d_softc = sc; 3145 sc->sc_disks[n].d_no = n; 3146 sc->sc_disks[n].d_state = G_RAID3_DISK_STATE_NODISK; 3147 } 3148 sx_init(&sc->sc_lock, "graid3:lock"); 3149 bioq_init(&sc->sc_queue); 3150 mtx_init(&sc->sc_queue_mtx, "graid3:queue", NULL, MTX_DEF); 3151 bioq_init(&sc->sc_regular_delayed); 3152 bioq_init(&sc->sc_inflight); 3153 bioq_init(&sc->sc_sync_delayed); 3154 TAILQ_INIT(&sc->sc_events); 3155 mtx_init(&sc->sc_events_mtx, "graid3:events", NULL, MTX_DEF); 3156 callout_init(&sc->sc_callout, 1); 3157 sc->sc_state = G_RAID3_DEVICE_STATE_STARTING; 3158 gp->softc = sc; 3159 sc->sc_geom = gp; 3160 sc->sc_provider = NULL; 3161 /* 3162 * Synchronization geom. 3163 */ 3164 gp = g_new_geomf(mp, "%s.sync", md->md_name); 3165 gp->softc = sc; 3166 gp->orphan = g_raid3_orphan; 3167 sc->sc_sync.ds_geom = gp; 3168 3169 if (!g_raid3_use_malloc) { 3170 sc->sc_zones[G_RAID3_ZONE_64K].sz_zone = uma_zcreate("gr3:64k", 3171 65536, g_raid3_uma_ctor, g_raid3_uma_dtor, NULL, NULL, 3172 UMA_ALIGN_PTR, 0); 3173 sc->sc_zones[G_RAID3_ZONE_64K].sz_inuse = 0; 3174 sc->sc_zones[G_RAID3_ZONE_64K].sz_max = g_raid3_n64k; 3175 sc->sc_zones[G_RAID3_ZONE_64K].sz_requested = 3176 sc->sc_zones[G_RAID3_ZONE_64K].sz_failed = 0; 3177 sc->sc_zones[G_RAID3_ZONE_16K].sz_zone = uma_zcreate("gr3:16k", 3178 16384, g_raid3_uma_ctor, g_raid3_uma_dtor, NULL, NULL, 3179 UMA_ALIGN_PTR, 0); 3180 sc->sc_zones[G_RAID3_ZONE_16K].sz_inuse = 0; 3181 sc->sc_zones[G_RAID3_ZONE_16K].sz_max = g_raid3_n16k; 3182 sc->sc_zones[G_RAID3_ZONE_16K].sz_requested = 3183 sc->sc_zones[G_RAID3_ZONE_16K].sz_failed = 0; 3184 sc->sc_zones[G_RAID3_ZONE_4K].sz_zone = uma_zcreate("gr3:4k", 3185 4096, g_raid3_uma_ctor, g_raid3_uma_dtor, NULL, NULL, 3186 UMA_ALIGN_PTR, 0); 3187 sc->sc_zones[G_RAID3_ZONE_4K].sz_inuse = 0; 3188 sc->sc_zones[G_RAID3_ZONE_4K].sz_max = g_raid3_n4k; 3189 sc->sc_zones[G_RAID3_ZONE_4K].sz_requested = 3190 sc->sc_zones[G_RAID3_ZONE_4K].sz_failed = 0; 3191 } 3192 3193 error = kproc_create(g_raid3_worker, sc, &sc->sc_worker, 0, 0, 3194 "g_raid3 %s", md->md_name); 3195 if (error != 0) { 3196 G_RAID3_DEBUG(1, "Cannot create kernel thread for %s.", 3197 sc->sc_name); 3198 if (!g_raid3_use_malloc) { 3199 uma_zdestroy(sc->sc_zones[G_RAID3_ZONE_64K].sz_zone); 3200 uma_zdestroy(sc->sc_zones[G_RAID3_ZONE_16K].sz_zone); 3201 uma_zdestroy(sc->sc_zones[G_RAID3_ZONE_4K].sz_zone); 3202 } 3203 g_destroy_geom(sc->sc_sync.ds_geom); 3204 mtx_destroy(&sc->sc_events_mtx); 3205 mtx_destroy(&sc->sc_queue_mtx); 3206 sx_destroy(&sc->sc_lock); 3207 g_destroy_geom(sc->sc_geom); 3208 free(sc->sc_disks, M_RAID3); 3209 free(sc, M_RAID3); 3210 return (NULL); 3211 } 3212 3213 G_RAID3_DEBUG(1, "Device %s created (%u components, id=%u).", 3214 sc->sc_name, sc->sc_ndisks, sc->sc_id); 3215 3216 sc->sc_rootmount = root_mount_hold("GRAID3"); 3217 G_RAID3_DEBUG(1, "root_mount_hold %p", sc->sc_rootmount); 3218 3219 /* 3220 * Run timeout. 3221 */ 3222 timeout = atomic_load_acq_int(&g_raid3_timeout); 3223 callout_reset(&sc->sc_callout, timeout * hz, g_raid3_go, sc); 3224 return (sc->sc_geom); 3225 } 3226 3227 int 3228 g_raid3_destroy(struct g_raid3_softc *sc, int how) 3229 { 3230 struct g_provider *pp; 3231 3232 g_topology_assert_not(); 3233 if (sc == NULL) 3234 return (ENXIO); 3235 sx_assert(&sc->sc_lock, SX_XLOCKED); 3236 3237 pp = sc->sc_provider; 3238 if (pp != NULL && (pp->acr != 0 || pp->acw != 0 || pp->ace != 0)) { 3239 switch (how) { 3240 case G_RAID3_DESTROY_SOFT: 3241 G_RAID3_DEBUG(1, 3242 "Device %s is still open (r%dw%de%d).", pp->name, 3243 pp->acr, pp->acw, pp->ace); 3244 return (EBUSY); 3245 case G_RAID3_DESTROY_DELAYED: 3246 G_RAID3_DEBUG(1, 3247 "Device %s will be destroyed on last close.", 3248 pp->name); 3249 if (sc->sc_syncdisk != NULL) 3250 g_raid3_sync_stop(sc, 1); 3251 sc->sc_flags |= G_RAID3_DEVICE_FLAG_DESTROYING; 3252 return (EBUSY); 3253 case G_RAID3_DESTROY_HARD: 3254 G_RAID3_DEBUG(1, "Device %s is still open, so it " 3255 "can't be definitely removed.", pp->name); 3256 break; 3257 } 3258 } 3259 3260 g_topology_lock(); 3261 if (sc->sc_geom->softc == NULL) { 3262 g_topology_unlock(); 3263 return (0); 3264 } 3265 sc->sc_geom->softc = NULL; 3266 sc->sc_sync.ds_geom->softc = NULL; 3267 g_topology_unlock(); 3268 3269 sc->sc_flags |= G_RAID3_DEVICE_FLAG_DESTROY; 3270 sc->sc_flags |= G_RAID3_DEVICE_FLAG_WAIT; 3271 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, sc); 3272 sx_xunlock(&sc->sc_lock); 3273 mtx_lock(&sc->sc_queue_mtx); 3274 wakeup(sc); 3275 wakeup(&sc->sc_queue); 3276 mtx_unlock(&sc->sc_queue_mtx); 3277 G_RAID3_DEBUG(4, "%s: Sleeping %p.", __func__, &sc->sc_worker); 3278 while (sc->sc_worker != NULL) 3279 tsleep(&sc->sc_worker, PRIBIO, "r3:destroy", hz / 5); 3280 G_RAID3_DEBUG(4, "%s: Woken up %p.", __func__, &sc->sc_worker); 3281 sx_xlock(&sc->sc_lock); 3282 g_raid3_destroy_device(sc); 3283 free(sc->sc_disks, M_RAID3); 3284 free(sc, M_RAID3); 3285 return (0); 3286 } 3287 3288 static void 3289 g_raid3_taste_orphan(struct g_consumer *cp) 3290 { 3291 3292 KASSERT(1 == 0, ("%s called while tasting %s.", __func__, 3293 cp->provider->name)); 3294 } 3295 3296 static struct g_geom * 3297 g_raid3_taste(struct g_class *mp, struct g_provider *pp, int flags __unused) 3298 { 3299 struct g_raid3_metadata md; 3300 struct g_raid3_softc *sc; 3301 struct g_consumer *cp; 3302 struct g_geom *gp; 3303 int error; 3304 3305 g_topology_assert(); 3306 g_trace(G_T_TOPOLOGY, "%s(%s, %s)", __func__, mp->name, pp->name); 3307 G_RAID3_DEBUG(2, "Tasting %s.", pp->name); 3308 3309 gp = g_new_geomf(mp, "raid3:taste"); 3310 /* This orphan function should be never called. */ 3311 gp->orphan = g_raid3_taste_orphan; 3312 cp = g_new_consumer(gp); 3313 error = g_attach(cp, pp); 3314 if (error == 0) { 3315 error = g_raid3_read_metadata(cp, &md); 3316 g_detach(cp); 3317 } 3318 g_destroy_consumer(cp); 3319 g_destroy_geom(gp); 3320 if (error != 0) 3321 return (NULL); 3322 gp = NULL; 3323 3324 if (md.md_provider[0] != '\0' && 3325 !g_compare_names(md.md_provider, pp->name)) 3326 return (NULL); 3327 if (md.md_provsize != 0 && md.md_provsize != pp->mediasize) 3328 return (NULL); 3329 if (g_raid3_debug >= 2) 3330 raid3_metadata_dump(&md); 3331 3332 /* 3333 * Let's check if device already exists. 3334 */ 3335 sc = NULL; 3336 LIST_FOREACH(gp, &mp->geom, geom) { 3337 sc = gp->softc; 3338 if (sc == NULL) 3339 continue; 3340 if (sc->sc_sync.ds_geom == gp) 3341 continue; 3342 if (strcmp(md.md_name, sc->sc_name) != 0) 3343 continue; 3344 if (md.md_id != sc->sc_id) { 3345 G_RAID3_DEBUG(0, "Device %s already configured.", 3346 sc->sc_name); 3347 return (NULL); 3348 } 3349 break; 3350 } 3351 if (gp == NULL) { 3352 gp = g_raid3_create(mp, &md); 3353 if (gp == NULL) { 3354 G_RAID3_DEBUG(0, "Cannot create device %s.", 3355 md.md_name); 3356 return (NULL); 3357 } 3358 sc = gp->softc; 3359 } 3360 G_RAID3_DEBUG(1, "Adding disk %s to %s.", pp->name, gp->name); 3361 g_topology_unlock(); 3362 sx_xlock(&sc->sc_lock); 3363 error = g_raid3_add_disk(sc, pp, &md); 3364 if (error != 0) { 3365 G_RAID3_DEBUG(0, "Cannot add disk %s to %s (error=%d).", 3366 pp->name, gp->name, error); 3367 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_NODISK) == 3368 sc->sc_ndisks) { 3369 g_cancel_event(sc); 3370 g_raid3_destroy(sc, G_RAID3_DESTROY_HARD); 3371 g_topology_lock(); 3372 return (NULL); 3373 } 3374 gp = NULL; 3375 } 3376 sx_xunlock(&sc->sc_lock); 3377 g_topology_lock(); 3378 return (gp); 3379 } 3380 3381 static int 3382 g_raid3_destroy_geom(struct gctl_req *req __unused, struct g_class *mp __unused, 3383 struct g_geom *gp) 3384 { 3385 struct g_raid3_softc *sc; 3386 int error; 3387 3388 g_topology_unlock(); 3389 sc = gp->softc; 3390 sx_xlock(&sc->sc_lock); 3391 g_cancel_event(sc); 3392 error = g_raid3_destroy(gp->softc, G_RAID3_DESTROY_SOFT); 3393 if (error != 0) 3394 sx_xunlock(&sc->sc_lock); 3395 g_topology_lock(); 3396 return (error); 3397 } 3398 3399 static void 3400 g_raid3_dumpconf(struct sbuf *sb, const char *indent, struct g_geom *gp, 3401 struct g_consumer *cp, struct g_provider *pp) 3402 { 3403 struct g_raid3_softc *sc; 3404 3405 g_topology_assert(); 3406 3407 sc = gp->softc; 3408 if (sc == NULL) 3409 return; 3410 /* Skip synchronization geom. */ 3411 if (gp == sc->sc_sync.ds_geom) 3412 return; 3413 if (pp != NULL) { 3414 /* Nothing here. */ 3415 } else if (cp != NULL) { 3416 struct g_raid3_disk *disk; 3417 3418 disk = cp->private; 3419 if (disk == NULL) 3420 return; 3421 g_topology_unlock(); 3422 sx_xlock(&sc->sc_lock); 3423 sbuf_printf(sb, "%s<Type>", indent); 3424 if (disk->d_no == sc->sc_ndisks - 1) 3425 sbuf_cat(sb, "PARITY"); 3426 else 3427 sbuf_cat(sb, "DATA"); 3428 sbuf_cat(sb, "</Type>\n"); 3429 sbuf_printf(sb, "%s<Number>%u</Number>\n", indent, 3430 (u_int)disk->d_no); 3431 if (disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 3432 sbuf_printf(sb, "%s<Synchronized>", indent); 3433 if (disk->d_sync.ds_offset == 0) 3434 sbuf_cat(sb, "0%"); 3435 else { 3436 sbuf_printf(sb, "%u%%", 3437 (u_int)((disk->d_sync.ds_offset * 100) / 3438 (sc->sc_mediasize / (sc->sc_ndisks - 1)))); 3439 } 3440 sbuf_cat(sb, "</Synchronized>\n"); 3441 if (disk->d_sync.ds_offset > 0) { 3442 sbuf_printf(sb, "%s<BytesSynced>%jd" 3443 "</BytesSynced>\n", indent, 3444 (intmax_t)disk->d_sync.ds_offset); 3445 } 3446 } 3447 sbuf_printf(sb, "%s<SyncID>%u</SyncID>\n", indent, 3448 disk->d_sync.ds_syncid); 3449 sbuf_printf(sb, "%s<GenID>%u</GenID>\n", indent, disk->d_genid); 3450 sbuf_printf(sb, "%s<Flags>", indent); 3451 if (disk->d_flags == 0) 3452 sbuf_cat(sb, "NONE"); 3453 else { 3454 int first = 1; 3455 3456 #define ADD_FLAG(flag, name) do { \ 3457 if ((disk->d_flags & (flag)) != 0) { \ 3458 if (!first) \ 3459 sbuf_cat(sb, ", "); \ 3460 else \ 3461 first = 0; \ 3462 sbuf_cat(sb, name); \ 3463 } \ 3464 } while (0) 3465 ADD_FLAG(G_RAID3_DISK_FLAG_DIRTY, "DIRTY"); 3466 ADD_FLAG(G_RAID3_DISK_FLAG_HARDCODED, "HARDCODED"); 3467 ADD_FLAG(G_RAID3_DISK_FLAG_SYNCHRONIZING, 3468 "SYNCHRONIZING"); 3469 ADD_FLAG(G_RAID3_DISK_FLAG_FORCE_SYNC, "FORCE_SYNC"); 3470 ADD_FLAG(G_RAID3_DISK_FLAG_BROKEN, "BROKEN"); 3471 #undef ADD_FLAG 3472 } 3473 sbuf_cat(sb, "</Flags>\n"); 3474 sbuf_printf(sb, "%s<State>%s</State>\n", indent, 3475 g_raid3_disk_state2str(disk->d_state)); 3476 sx_xunlock(&sc->sc_lock); 3477 g_topology_lock(); 3478 } else { 3479 g_topology_unlock(); 3480 sx_xlock(&sc->sc_lock); 3481 if (!g_raid3_use_malloc) { 3482 sbuf_printf(sb, 3483 "%s<Zone4kRequested>%u</Zone4kRequested>\n", indent, 3484 sc->sc_zones[G_RAID3_ZONE_4K].sz_requested); 3485 sbuf_printf(sb, 3486 "%s<Zone4kFailed>%u</Zone4kFailed>\n", indent, 3487 sc->sc_zones[G_RAID3_ZONE_4K].sz_failed); 3488 sbuf_printf(sb, 3489 "%s<Zone16kRequested>%u</Zone16kRequested>\n", indent, 3490 sc->sc_zones[G_RAID3_ZONE_16K].sz_requested); 3491 sbuf_printf(sb, 3492 "%s<Zone16kFailed>%u</Zone16kFailed>\n", indent, 3493 sc->sc_zones[G_RAID3_ZONE_16K].sz_failed); 3494 sbuf_printf(sb, 3495 "%s<Zone64kRequested>%u</Zone64kRequested>\n", indent, 3496 sc->sc_zones[G_RAID3_ZONE_64K].sz_requested); 3497 sbuf_printf(sb, 3498 "%s<Zone64kFailed>%u</Zone64kFailed>\n", indent, 3499 sc->sc_zones[G_RAID3_ZONE_64K].sz_failed); 3500 } 3501 sbuf_printf(sb, "%s<ID>%u</ID>\n", indent, (u_int)sc->sc_id); 3502 sbuf_printf(sb, "%s<SyncID>%u</SyncID>\n", indent, sc->sc_syncid); 3503 sbuf_printf(sb, "%s<GenID>%u</GenID>\n", indent, sc->sc_genid); 3504 sbuf_printf(sb, "%s<Flags>", indent); 3505 if (sc->sc_flags == 0) 3506 sbuf_cat(sb, "NONE"); 3507 else { 3508 int first = 1; 3509 3510 #define ADD_FLAG(flag, name) do { \ 3511 if ((sc->sc_flags & (flag)) != 0) { \ 3512 if (!first) \ 3513 sbuf_cat(sb, ", "); \ 3514 else \ 3515 first = 0; \ 3516 sbuf_cat(sb, name); \ 3517 } \ 3518 } while (0) 3519 ADD_FLAG(G_RAID3_DEVICE_FLAG_NOFAILSYNC, "NOFAILSYNC"); 3520 ADD_FLAG(G_RAID3_DEVICE_FLAG_NOAUTOSYNC, "NOAUTOSYNC"); 3521 ADD_FLAG(G_RAID3_DEVICE_FLAG_ROUND_ROBIN, 3522 "ROUND-ROBIN"); 3523 ADD_FLAG(G_RAID3_DEVICE_FLAG_VERIFY, "VERIFY"); 3524 #undef ADD_FLAG 3525 } 3526 sbuf_cat(sb, "</Flags>\n"); 3527 sbuf_printf(sb, "%s<Components>%u</Components>\n", indent, 3528 sc->sc_ndisks); 3529 sbuf_printf(sb, "%s<State>%s</State>\n", indent, 3530 g_raid3_device_state2str(sc->sc_state)); 3531 sx_xunlock(&sc->sc_lock); 3532 g_topology_lock(); 3533 } 3534 } 3535 3536 static void 3537 g_raid3_shutdown_post_sync(void *arg, int howto) 3538 { 3539 struct g_class *mp; 3540 struct g_geom *gp, *gp2; 3541 struct g_raid3_softc *sc; 3542 int error; 3543 3544 mp = arg; 3545 g_topology_lock(); 3546 g_raid3_shutdown = 1; 3547 LIST_FOREACH_SAFE(gp, &mp->geom, geom, gp2) { 3548 if ((sc = gp->softc) == NULL) 3549 continue; 3550 /* Skip synchronization geom. */ 3551 if (gp == sc->sc_sync.ds_geom) 3552 continue; 3553 g_topology_unlock(); 3554 sx_xlock(&sc->sc_lock); 3555 g_raid3_idle(sc, -1); 3556 g_cancel_event(sc); 3557 error = g_raid3_destroy(sc, G_RAID3_DESTROY_DELAYED); 3558 if (error != 0) 3559 sx_xunlock(&sc->sc_lock); 3560 g_topology_lock(); 3561 } 3562 g_topology_unlock(); 3563 } 3564 3565 static void 3566 g_raid3_init(struct g_class *mp) 3567 { 3568 3569 g_raid3_post_sync = EVENTHANDLER_REGISTER(shutdown_post_sync, 3570 g_raid3_shutdown_post_sync, mp, SHUTDOWN_PRI_FIRST); 3571 if (g_raid3_post_sync == NULL) 3572 G_RAID3_DEBUG(0, "Warning! Cannot register shutdown event."); 3573 } 3574 3575 static void 3576 g_raid3_fini(struct g_class *mp) 3577 { 3578 3579 if (g_raid3_post_sync != NULL) 3580 EVENTHANDLER_DEREGISTER(shutdown_post_sync, g_raid3_post_sync); 3581 } 3582 3583 DECLARE_GEOM_CLASS(g_raid3_class, g_raid3); 3584 MODULE_VERSION(geom_raid3, 0); 3585