1 /*- 2 * Copyright (c) 2004 Pawel Jakub Dawidek <pjd@FreeBSD.org> 3 * All rights reserved. 4 * 5 * Redistribution and use in source and binary forms, with or without 6 * modification, are permitted provided that the following conditions 7 * are met: 8 * 1. Redistributions of source code must retain the above copyright 9 * notice, this list of conditions and the following disclaimer. 10 * 2. Redistributions in binary form must reproduce the above copyright 11 * notice, this list of conditions and the following disclaimer in the 12 * documentation and/or other materials provided with the distribution. 13 * 14 * THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND 15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE 18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 24 * SUCH DAMAGE. 25 */ 26 27 #include <sys/cdefs.h> 28 __FBSDID("$FreeBSD$"); 29 30 #include <sys/param.h> 31 #include <sys/systm.h> 32 #include <sys/kernel.h> 33 #include <sys/module.h> 34 #include <sys/limits.h> 35 #include <sys/lock.h> 36 #include <sys/mutex.h> 37 #include <sys/bio.h> 38 #include <sys/sysctl.h> 39 #include <sys/malloc.h> 40 #include <sys/eventhandler.h> 41 #include <vm/uma.h> 42 #include <geom/geom.h> 43 #include <sys/proc.h> 44 #include <sys/kthread.h> 45 #include <sys/sched.h> 46 #include <geom/raid3/g_raid3.h> 47 48 49 static MALLOC_DEFINE(M_RAID3, "raid3 data", "GEOM_RAID3 Data"); 50 51 SYSCTL_DECL(_kern_geom); 52 SYSCTL_NODE(_kern_geom, OID_AUTO, raid3, CTLFLAG_RW, 0, "GEOM_RAID3 stuff"); 53 u_int g_raid3_debug = 0; 54 TUNABLE_INT("kern.geom.raid3.debug", &g_raid3_debug); 55 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, debug, CTLFLAG_RW, &g_raid3_debug, 0, 56 "Debug level"); 57 static u_int g_raid3_timeout = 4; 58 TUNABLE_INT("kern.geom.raid3.timeout", &g_raid3_timeout); 59 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, timeout, CTLFLAG_RW, &g_raid3_timeout, 60 0, "Time to wait on all raid3 components"); 61 static u_int g_raid3_idletime = 5; 62 TUNABLE_INT("kern.geom.raid3.idletime", &g_raid3_idletime); 63 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, idletime, CTLFLAG_RW, 64 &g_raid3_idletime, 0, "Mark components as clean when idling"); 65 static u_int g_raid3_reqs_per_sync = 5; 66 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, reqs_per_sync, CTLFLAG_RW, 67 &g_raid3_reqs_per_sync, 0, 68 "Number of regular I/O requests per synchronization request"); 69 static u_int g_raid3_syncs_per_sec = 1000; 70 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, syncs_per_sec, CTLFLAG_RW, 71 &g_raid3_syncs_per_sec, 0, 72 "Number of synchronizations requests per second"); 73 74 static u_int g_raid3_n64k = 50; 75 TUNABLE_INT("kern.geom.raid3.n64k", &g_raid3_n64k); 76 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, n64k, CTLFLAG_RD, &g_raid3_n64k, 0, 77 "Maximum number of 64kB allocations"); 78 static u_int g_raid3_n16k = 200; 79 TUNABLE_INT("kern.geom.raid3.n16k", &g_raid3_n16k); 80 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, n16k, CTLFLAG_RD, &g_raid3_n16k, 0, 81 "Maximum number of 16kB allocations"); 82 static u_int g_raid3_n4k = 1200; 83 TUNABLE_INT("kern.geom.raid3.n4k", &g_raid3_n4k); 84 SYSCTL_UINT(_kern_geom_raid3, OID_AUTO, n4k, CTLFLAG_RD, &g_raid3_n4k, 0, 85 "Maximum number of 4kB allocations"); 86 87 SYSCTL_NODE(_kern_geom_raid3, OID_AUTO, stat, CTLFLAG_RW, 0, 88 "GEOM_RAID3 statistics"); 89 static u_int g_raid3_parity_mismatch = 0; 90 SYSCTL_UINT(_kern_geom_raid3_stat, OID_AUTO, parity_mismatch, CTLFLAG_RD, 91 &g_raid3_parity_mismatch, 0, "Number of failures in VERIFY mode"); 92 static u_int g_raid3_64k_requested = 0; 93 SYSCTL_UINT(_kern_geom_raid3_stat, OID_AUTO, 64k_requested, CTLFLAG_RD, 94 &g_raid3_64k_requested, 0, "Number of requested 64kB allocations"); 95 static u_int g_raid3_64k_failed = 0; 96 SYSCTL_UINT(_kern_geom_raid3_stat, OID_AUTO, 64k_failed, CTLFLAG_RD, 97 &g_raid3_64k_failed, 0, "Number of failed 64kB allocations"); 98 static u_int g_raid3_16k_requested = 0; 99 SYSCTL_UINT(_kern_geom_raid3_stat, OID_AUTO, 16k_requested, CTLFLAG_RD, 100 &g_raid3_16k_requested, 0, "Number of requested 16kB allocations"); 101 static u_int g_raid3_16k_failed = 0; 102 SYSCTL_UINT(_kern_geom_raid3_stat, OID_AUTO, 16k_failed, CTLFLAG_RD, 103 &g_raid3_16k_failed, 0, "Number of failed 16kB allocations"); 104 static u_int g_raid3_4k_requested = 0; 105 SYSCTL_UINT(_kern_geom_raid3_stat, OID_AUTO, 4k_requested, CTLFLAG_RD, 106 &g_raid3_4k_requested, 0, "Number of requested 4kB allocations"); 107 static u_int g_raid3_4k_failed = 0; 108 SYSCTL_UINT(_kern_geom_raid3_stat, OID_AUTO, 4k_failed, CTLFLAG_RD, 109 &g_raid3_4k_failed, 0, "Number of failed 4kB allocations"); 110 111 #define MSLEEP(ident, mtx, priority, wmesg, timeout) do { \ 112 G_RAID3_DEBUG(4, "%s: Sleeping %p.", __func__, (ident)); \ 113 msleep((ident), (mtx), (priority), (wmesg), (timeout)); \ 114 G_RAID3_DEBUG(4, "%s: Woken up %p.", __func__, (ident)); \ 115 } while (0) 116 117 static eventhandler_tag g_raid3_ehtag = NULL; 118 119 static int g_raid3_destroy_geom(struct gctl_req *req, struct g_class *mp, 120 struct g_geom *gp); 121 static g_taste_t g_raid3_taste; 122 static void g_raid3_init(struct g_class *mp); 123 static void g_raid3_fini(struct g_class *mp); 124 125 struct g_class g_raid3_class = { 126 .name = G_RAID3_CLASS_NAME, 127 .version = G_VERSION, 128 .ctlreq = g_raid3_config, 129 .taste = g_raid3_taste, 130 .destroy_geom = g_raid3_destroy_geom, 131 .init = g_raid3_init, 132 .fini = g_raid3_fini 133 }; 134 135 136 static void g_raid3_destroy_provider(struct g_raid3_softc *sc); 137 static int g_raid3_update_disk(struct g_raid3_disk *disk, u_int state); 138 static void g_raid3_update_device(struct g_raid3_softc *sc, boolean_t force); 139 static void g_raid3_dumpconf(struct sbuf *sb, const char *indent, 140 struct g_geom *gp, struct g_consumer *cp, struct g_provider *pp); 141 static void g_raid3_sync_stop(struct g_raid3_softc *sc, int type); 142 143 144 static const char * 145 g_raid3_disk_state2str(int state) 146 { 147 148 switch (state) { 149 case G_RAID3_DISK_STATE_NODISK: 150 return ("NODISK"); 151 case G_RAID3_DISK_STATE_NONE: 152 return ("NONE"); 153 case G_RAID3_DISK_STATE_NEW: 154 return ("NEW"); 155 case G_RAID3_DISK_STATE_ACTIVE: 156 return ("ACTIVE"); 157 case G_RAID3_DISK_STATE_STALE: 158 return ("STALE"); 159 case G_RAID3_DISK_STATE_SYNCHRONIZING: 160 return ("SYNCHRONIZING"); 161 case G_RAID3_DISK_STATE_DISCONNECTED: 162 return ("DISCONNECTED"); 163 default: 164 return ("INVALID"); 165 } 166 } 167 168 static const char * 169 g_raid3_device_state2str(int state) 170 { 171 172 switch (state) { 173 case G_RAID3_DEVICE_STATE_STARTING: 174 return ("STARTING"); 175 case G_RAID3_DEVICE_STATE_DEGRADED: 176 return ("DEGRADED"); 177 case G_RAID3_DEVICE_STATE_COMPLETE: 178 return ("COMPLETE"); 179 default: 180 return ("INVALID"); 181 } 182 } 183 184 const char * 185 g_raid3_get_diskname(struct g_raid3_disk *disk) 186 { 187 188 if (disk->d_consumer == NULL || disk->d_consumer->provider == NULL) 189 return ("[unknown]"); 190 return (disk->d_name); 191 } 192 193 #define g_raid3_xor(src1, src2, dst, size) \ 194 _g_raid3_xor((uint64_t *)(src1), (uint64_t *)(src2), \ 195 (uint64_t *)(dst), (size_t)size) 196 static void 197 _g_raid3_xor(uint64_t *src1, uint64_t *src2, uint64_t *dst, size_t size) 198 { 199 200 KASSERT((size % 128) == 0, ("Invalid size: %zu.", size)); 201 for (; size > 0; size -= 128) { 202 *dst++ = (*src1++) ^ (*src2++); 203 *dst++ = (*src1++) ^ (*src2++); 204 *dst++ = (*src1++) ^ (*src2++); 205 *dst++ = (*src1++) ^ (*src2++); 206 *dst++ = (*src1++) ^ (*src2++); 207 *dst++ = (*src1++) ^ (*src2++); 208 *dst++ = (*src1++) ^ (*src2++); 209 *dst++ = (*src1++) ^ (*src2++); 210 *dst++ = (*src1++) ^ (*src2++); 211 *dst++ = (*src1++) ^ (*src2++); 212 *dst++ = (*src1++) ^ (*src2++); 213 *dst++ = (*src1++) ^ (*src2++); 214 *dst++ = (*src1++) ^ (*src2++); 215 *dst++ = (*src1++) ^ (*src2++); 216 *dst++ = (*src1++) ^ (*src2++); 217 *dst++ = (*src1++) ^ (*src2++); 218 } 219 } 220 221 static int 222 g_raid3_is_zero(struct bio *bp) 223 { 224 static const uint64_t zeros[] = { 225 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0, 0 226 }; 227 u_char *addr; 228 ssize_t size; 229 230 size = bp->bio_length; 231 addr = (u_char *)bp->bio_data; 232 for (; size > 0; size -= sizeof(zeros), addr += sizeof(zeros)) { 233 if (bcmp(addr, zeros, sizeof(zeros)) != 0) 234 return (0); 235 } 236 return (1); 237 } 238 239 /* 240 * --- Events handling functions --- 241 * Events in geom_raid3 are used to maintain disks and device status 242 * from one thread to simplify locking. 243 */ 244 static void 245 g_raid3_event_free(struct g_raid3_event *ep) 246 { 247 248 free(ep, M_RAID3); 249 } 250 251 int 252 g_raid3_event_send(void *arg, int state, int flags) 253 { 254 struct g_raid3_softc *sc; 255 struct g_raid3_disk *disk; 256 struct g_raid3_event *ep; 257 int error; 258 259 ep = malloc(sizeof(*ep), M_RAID3, M_WAITOK); 260 G_RAID3_DEBUG(4, "%s: Sending event %p.", __func__, ep); 261 if ((flags & G_RAID3_EVENT_DEVICE) != 0) { 262 disk = NULL; 263 sc = arg; 264 } else { 265 disk = arg; 266 sc = disk->d_softc; 267 } 268 ep->e_disk = disk; 269 ep->e_state = state; 270 ep->e_flags = flags; 271 ep->e_error = 0; 272 mtx_lock(&sc->sc_events_mtx); 273 TAILQ_INSERT_TAIL(&sc->sc_events, ep, e_next); 274 mtx_unlock(&sc->sc_events_mtx); 275 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, sc); 276 mtx_lock(&sc->sc_queue_mtx); 277 wakeup(sc); 278 wakeup(&sc->sc_queue); 279 mtx_unlock(&sc->sc_queue_mtx); 280 if ((flags & G_RAID3_EVENT_DONTWAIT) != 0) 281 return (0); 282 g_topology_assert(); 283 G_RAID3_DEBUG(4, "%s: Sleeping %p.", __func__, ep); 284 g_topology_unlock(); 285 while ((ep->e_flags & G_RAID3_EVENT_DONE) == 0) { 286 mtx_lock(&sc->sc_events_mtx); 287 MSLEEP(ep, &sc->sc_events_mtx, PRIBIO | PDROP, "r3:event", 288 hz * 5); 289 } 290 /* Don't even try to use 'sc' here, because it could be already dead. */ 291 g_topology_lock(); 292 error = ep->e_error; 293 g_raid3_event_free(ep); 294 return (error); 295 } 296 297 static struct g_raid3_event * 298 g_raid3_event_get(struct g_raid3_softc *sc) 299 { 300 struct g_raid3_event *ep; 301 302 mtx_lock(&sc->sc_events_mtx); 303 ep = TAILQ_FIRST(&sc->sc_events); 304 mtx_unlock(&sc->sc_events_mtx); 305 return (ep); 306 } 307 308 static void 309 g_raid3_event_remove(struct g_raid3_softc *sc, struct g_raid3_event *ep) 310 { 311 312 mtx_lock(&sc->sc_events_mtx); 313 TAILQ_REMOVE(&sc->sc_events, ep, e_next); 314 mtx_unlock(&sc->sc_events_mtx); 315 } 316 317 static void 318 g_raid3_event_cancel(struct g_raid3_disk *disk) 319 { 320 struct g_raid3_softc *sc; 321 struct g_raid3_event *ep, *tmpep; 322 323 g_topology_assert(); 324 325 sc = disk->d_softc; 326 mtx_lock(&sc->sc_events_mtx); 327 TAILQ_FOREACH_SAFE(ep, &sc->sc_events, e_next, tmpep) { 328 if ((ep->e_flags & G_RAID3_EVENT_DEVICE) != 0) 329 continue; 330 if (ep->e_disk != disk) 331 continue; 332 TAILQ_REMOVE(&sc->sc_events, ep, e_next); 333 if ((ep->e_flags & G_RAID3_EVENT_DONTWAIT) != 0) 334 g_raid3_event_free(ep); 335 else { 336 ep->e_error = ECANCELED; 337 wakeup(ep); 338 } 339 } 340 mtx_unlock(&sc->sc_events_mtx); 341 } 342 343 /* 344 * Return the number of disks in the given state. 345 * If state is equal to -1, count all connected disks. 346 */ 347 u_int 348 g_raid3_ndisks(struct g_raid3_softc *sc, int state) 349 { 350 struct g_raid3_disk *disk; 351 u_int n, ndisks; 352 353 for (n = ndisks = 0; n < sc->sc_ndisks; n++) { 354 disk = &sc->sc_disks[n]; 355 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 356 continue; 357 if (state == -1 || disk->d_state == state) 358 ndisks++; 359 } 360 return (ndisks); 361 } 362 363 static u_int 364 g_raid3_nrequests(struct g_raid3_softc *sc, struct g_consumer *cp) 365 { 366 struct bio *bp; 367 u_int nreqs = 0; 368 369 mtx_lock(&sc->sc_queue_mtx); 370 TAILQ_FOREACH(bp, &sc->sc_queue.queue, bio_queue) { 371 if (bp->bio_from == cp) 372 nreqs++; 373 } 374 mtx_unlock(&sc->sc_queue_mtx); 375 return (nreqs); 376 } 377 378 static int 379 g_raid3_is_busy(struct g_raid3_softc *sc, struct g_consumer *cp) 380 { 381 382 if (cp->index > 0) { 383 G_RAID3_DEBUG(2, 384 "I/O requests for %s exist, can't destroy it now.", 385 cp->provider->name); 386 return (1); 387 } 388 if (g_raid3_nrequests(sc, cp) > 0) { 389 G_RAID3_DEBUG(2, 390 "I/O requests for %s in queue, can't destroy it now.", 391 cp->provider->name); 392 return (1); 393 } 394 return (0); 395 } 396 397 static void 398 g_raid3_destroy_consumer(void *arg, int flags __unused) 399 { 400 struct g_consumer *cp; 401 402 cp = arg; 403 G_RAID3_DEBUG(1, "Consumer %s destroyed.", cp->provider->name); 404 g_detach(cp); 405 g_destroy_consumer(cp); 406 } 407 408 static void 409 g_raid3_kill_consumer(struct g_raid3_softc *sc, struct g_consumer *cp) 410 { 411 struct g_provider *pp; 412 int retaste_wait; 413 414 g_topology_assert(); 415 416 cp->private = NULL; 417 if (g_raid3_is_busy(sc, cp)) 418 return; 419 G_RAID3_DEBUG(2, "Consumer %s destroyed.", cp->provider->name); 420 pp = cp->provider; 421 retaste_wait = 0; 422 if (cp->acw == 1) { 423 if ((pp->geom->flags & G_GEOM_WITHER) == 0) 424 retaste_wait = 1; 425 } 426 G_RAID3_DEBUG(2, "Access %s r%dw%de%d = %d", pp->name, -cp->acr, 427 -cp->acw, -cp->ace, 0); 428 if (cp->acr > 0 || cp->acw > 0 || cp->ace > 0) 429 g_access(cp, -cp->acr, -cp->acw, -cp->ace); 430 if (retaste_wait) { 431 /* 432 * After retaste event was send (inside g_access()), we can send 433 * event to detach and destroy consumer. 434 * A class, which has consumer to the given provider connected 435 * will not receive retaste event for the provider. 436 * This is the way how I ignore retaste events when I close 437 * consumers opened for write: I detach and destroy consumer 438 * after retaste event is sent. 439 */ 440 g_post_event(g_raid3_destroy_consumer, cp, M_WAITOK, NULL); 441 return; 442 } 443 G_RAID3_DEBUG(1, "Consumer %s destroyed.", pp->name); 444 g_detach(cp); 445 g_destroy_consumer(cp); 446 } 447 448 static int 449 g_raid3_connect_disk(struct g_raid3_disk *disk, struct g_provider *pp) 450 { 451 int error; 452 453 g_topology_assert(); 454 KASSERT(disk->d_consumer == NULL, 455 ("Disk already connected (device %s).", disk->d_softc->sc_name)); 456 457 disk->d_consumer = g_new_consumer(disk->d_softc->sc_geom); 458 disk->d_consumer->private = disk; 459 disk->d_consumer->index = 0; 460 error = g_attach(disk->d_consumer, pp); 461 if (error != 0) 462 return (error); 463 error = g_access(disk->d_consumer, 1, 1, 1); 464 if (error != 0) { 465 G_RAID3_DEBUG(0, "Cannot open consumer %s (error=%d).", 466 pp->name, error); 467 return (error); 468 } 469 G_RAID3_DEBUG(2, "Disk %s connected.", g_raid3_get_diskname(disk)); 470 return (0); 471 } 472 473 static void 474 g_raid3_disconnect_consumer(struct g_raid3_softc *sc, struct g_consumer *cp) 475 { 476 477 g_topology_assert(); 478 479 if (cp == NULL) 480 return; 481 if (cp->provider != NULL) 482 g_raid3_kill_consumer(sc, cp); 483 else 484 g_destroy_consumer(cp); 485 } 486 487 /* 488 * Initialize disk. This means allocate memory, create consumer, attach it 489 * to the provider and open access (r1w1e1) to it. 490 */ 491 static struct g_raid3_disk * 492 g_raid3_init_disk(struct g_raid3_softc *sc, struct g_provider *pp, 493 struct g_raid3_metadata *md, int *errorp) 494 { 495 struct g_raid3_disk *disk; 496 int error; 497 498 disk = &sc->sc_disks[md->md_no]; 499 error = g_raid3_connect_disk(disk, pp); 500 if (error != 0) 501 goto fail; 502 disk->d_state = G_RAID3_DISK_STATE_NONE; 503 disk->d_flags = md->md_dflags; 504 if (md->md_provider[0] != '\0') 505 disk->d_flags |= G_RAID3_DISK_FLAG_HARDCODED; 506 disk->d_sync.ds_consumer = NULL; 507 disk->d_sync.ds_offset = md->md_sync_offset; 508 disk->d_sync.ds_offset_done = md->md_sync_offset; 509 disk->d_sync.ds_resync = -1; 510 disk->d_genid = md->md_genid; 511 disk->d_sync.ds_syncid = md->md_syncid; 512 if (errorp != NULL) 513 *errorp = 0; 514 return (disk); 515 fail: 516 if (errorp != NULL) 517 *errorp = error; 518 if (disk != NULL) 519 g_raid3_disconnect_consumer(sc, disk->d_consumer); 520 return (NULL); 521 } 522 523 static void 524 g_raid3_destroy_disk(struct g_raid3_disk *disk) 525 { 526 struct g_raid3_softc *sc; 527 528 g_topology_assert(); 529 530 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 531 return; 532 g_raid3_event_cancel(disk); 533 sc = disk->d_softc; 534 switch (disk->d_state) { 535 case G_RAID3_DISK_STATE_SYNCHRONIZING: 536 if (sc->sc_syncdisk != NULL) 537 g_raid3_sync_stop(sc, 1); 538 /* FALLTHROUGH */ 539 case G_RAID3_DISK_STATE_NEW: 540 case G_RAID3_DISK_STATE_STALE: 541 case G_RAID3_DISK_STATE_ACTIVE: 542 g_raid3_disconnect_consumer(sc, disk->d_consumer); 543 disk->d_consumer = NULL; 544 break; 545 default: 546 KASSERT(0 == 1, ("Wrong disk state (%s, %s).", 547 g_raid3_get_diskname(disk), 548 g_raid3_disk_state2str(disk->d_state))); 549 } 550 disk->d_state = G_RAID3_DISK_STATE_NODISK; 551 } 552 553 static void 554 g_raid3_destroy_device(struct g_raid3_softc *sc) 555 { 556 struct g_raid3_event *ep; 557 struct g_raid3_disk *disk; 558 struct g_geom *gp; 559 struct g_consumer *cp; 560 u_int n; 561 562 g_topology_assert(); 563 564 gp = sc->sc_geom; 565 if (sc->sc_provider != NULL) 566 g_raid3_destroy_provider(sc); 567 for (n = 0; n < sc->sc_ndisks; n++) { 568 disk = &sc->sc_disks[n]; 569 if (disk->d_state != G_RAID3_DISK_STATE_NODISK) { 570 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 571 g_raid3_update_metadata(disk); 572 g_raid3_destroy_disk(disk); 573 } 574 } 575 while ((ep = g_raid3_event_get(sc)) != NULL) { 576 g_raid3_event_remove(sc, ep); 577 if ((ep->e_flags & G_RAID3_EVENT_DONTWAIT) != 0) 578 g_raid3_event_free(ep); 579 else { 580 ep->e_error = ECANCELED; 581 ep->e_flags |= G_RAID3_EVENT_DONE; 582 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, ep); 583 mtx_lock(&sc->sc_events_mtx); 584 wakeup(ep); 585 mtx_unlock(&sc->sc_events_mtx); 586 } 587 } 588 callout_drain(&sc->sc_callout); 589 gp->softc = NULL; 590 cp = LIST_FIRST(&sc->sc_sync.ds_geom->consumer); 591 if (cp != NULL) 592 g_raid3_disconnect_consumer(sc, cp); 593 sc->sc_sync.ds_geom->softc = NULL; 594 g_wither_geom(sc->sc_sync.ds_geom, ENXIO); 595 uma_zdestroy(sc->sc_zone_64k); 596 uma_zdestroy(sc->sc_zone_16k); 597 uma_zdestroy(sc->sc_zone_4k); 598 mtx_destroy(&sc->sc_queue_mtx); 599 mtx_destroy(&sc->sc_events_mtx); 600 G_RAID3_DEBUG(0, "Device %s destroyed.", gp->name); 601 g_wither_geom(gp, ENXIO); 602 } 603 604 static void 605 g_raid3_orphan(struct g_consumer *cp) 606 { 607 struct g_raid3_disk *disk; 608 609 g_topology_assert(); 610 611 disk = cp->private; 612 if (disk == NULL) 613 return; 614 disk->d_softc->sc_bump_id = G_RAID3_BUMP_SYNCID; 615 g_raid3_event_send(disk, G_RAID3_DISK_STATE_DISCONNECTED, 616 G_RAID3_EVENT_DONTWAIT); 617 } 618 619 static int 620 g_raid3_write_metadata(struct g_raid3_disk *disk, struct g_raid3_metadata *md) 621 { 622 struct g_raid3_softc *sc; 623 struct g_consumer *cp; 624 off_t offset, length; 625 u_char *sector; 626 int error = 0; 627 628 g_topology_assert(); 629 630 sc = disk->d_softc; 631 cp = disk->d_consumer; 632 KASSERT(cp != NULL, ("NULL consumer (%s).", sc->sc_name)); 633 KASSERT(cp->provider != NULL, ("NULL provider (%s).", sc->sc_name)); 634 KASSERT(cp->acr == 1 && cp->acw == 1 && cp->ace == 1, 635 ("Consumer %s closed? (r%dw%de%d).", cp->provider->name, cp->acr, 636 cp->acw, cp->ace)); 637 length = cp->provider->sectorsize; 638 offset = cp->provider->mediasize - length; 639 sector = malloc((size_t)length, M_RAID3, M_WAITOK | M_ZERO); 640 if (md != NULL) 641 raid3_metadata_encode(md, sector); 642 g_topology_unlock(); 643 error = g_write_data(cp, offset, sector, length); 644 g_topology_lock(); 645 free(sector, M_RAID3); 646 if (error != 0) { 647 disk->d_softc->sc_bump_id = G_RAID3_BUMP_GENID; 648 g_raid3_event_send(disk, G_RAID3_DISK_STATE_DISCONNECTED, 649 G_RAID3_EVENT_DONTWAIT); 650 } 651 return (error); 652 } 653 654 int 655 g_raid3_clear_metadata(struct g_raid3_disk *disk) 656 { 657 int error; 658 659 g_topology_assert(); 660 error = g_raid3_write_metadata(disk, NULL); 661 if (error == 0) { 662 G_RAID3_DEBUG(2, "Metadata on %s cleared.", 663 g_raid3_get_diskname(disk)); 664 } else { 665 G_RAID3_DEBUG(0, 666 "Cannot clear metadata on disk %s (error=%d).", 667 g_raid3_get_diskname(disk), error); 668 } 669 return (error); 670 } 671 672 void 673 g_raid3_fill_metadata(struct g_raid3_disk *disk, struct g_raid3_metadata *md) 674 { 675 struct g_raid3_softc *sc; 676 677 sc = disk->d_softc; 678 strlcpy(md->md_magic, G_RAID3_MAGIC, sizeof(md->md_magic)); 679 md->md_version = G_RAID3_VERSION; 680 strlcpy(md->md_name, sc->sc_name, sizeof(md->md_name)); 681 md->md_id = sc->sc_id; 682 md->md_all = sc->sc_ndisks; 683 md->md_genid = sc->sc_genid; 684 md->md_mediasize = sc->sc_mediasize; 685 md->md_sectorsize = sc->sc_sectorsize; 686 md->md_mflags = (sc->sc_flags & G_RAID3_DEVICE_FLAG_MASK); 687 md->md_no = disk->d_no; 688 md->md_syncid = disk->d_sync.ds_syncid; 689 md->md_dflags = (disk->d_flags & G_RAID3_DISK_FLAG_MASK); 690 if (disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) 691 md->md_sync_offset = disk->d_sync.ds_offset_done; 692 else 693 md->md_sync_offset = 0; 694 if ((disk->d_flags & G_RAID3_DISK_FLAG_HARDCODED) != 0 && 695 disk->d_consumer != NULL && disk->d_consumer->provider != NULL) { 696 strlcpy(md->md_provider, disk->d_consumer->provider->name, 697 sizeof(md->md_provider)); 698 } else { 699 bzero(md->md_provider, sizeof(md->md_provider)); 700 } 701 } 702 703 void 704 g_raid3_update_metadata(struct g_raid3_disk *disk) 705 { 706 struct g_raid3_metadata md; 707 int error; 708 709 g_topology_assert(); 710 g_raid3_fill_metadata(disk, &md); 711 error = g_raid3_write_metadata(disk, &md); 712 if (error == 0) { 713 G_RAID3_DEBUG(2, "Metadata on %s updated.", 714 g_raid3_get_diskname(disk)); 715 } else { 716 G_RAID3_DEBUG(0, 717 "Cannot update metadata on disk %s (error=%d).", 718 g_raid3_get_diskname(disk), error); 719 } 720 } 721 722 static void 723 g_raid3_bump_syncid(struct g_raid3_softc *sc) 724 { 725 struct g_raid3_disk *disk; 726 u_int n; 727 728 g_topology_assert(); 729 KASSERT(g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) > 0, 730 ("%s called with no active disks (device=%s).", __func__, 731 sc->sc_name)); 732 733 sc->sc_syncid++; 734 G_RAID3_DEBUG(1, "Device %s: syncid bumped to %u.", sc->sc_name, 735 sc->sc_syncid); 736 for (n = 0; n < sc->sc_ndisks; n++) { 737 disk = &sc->sc_disks[n]; 738 if (disk->d_state == G_RAID3_DISK_STATE_ACTIVE || 739 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 740 disk->d_sync.ds_syncid = sc->sc_syncid; 741 g_raid3_update_metadata(disk); 742 } 743 } 744 } 745 746 static void 747 g_raid3_bump_genid(struct g_raid3_softc *sc) 748 { 749 struct g_raid3_disk *disk; 750 u_int n; 751 752 g_topology_assert(); 753 KASSERT(g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) > 0, 754 ("%s called with no active disks (device=%s).", __func__, 755 sc->sc_name)); 756 757 sc->sc_genid++; 758 G_RAID3_DEBUG(1, "Device %s: genid bumped to %u.", sc->sc_name, 759 sc->sc_genid); 760 for (n = 0; n < sc->sc_ndisks; n++) { 761 disk = &sc->sc_disks[n]; 762 if (disk->d_state == G_RAID3_DISK_STATE_ACTIVE || 763 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 764 disk->d_genid = sc->sc_genid; 765 g_raid3_update_metadata(disk); 766 } 767 } 768 } 769 770 static void 771 g_raid3_idle(struct g_raid3_softc *sc) 772 { 773 struct g_raid3_disk *disk; 774 u_int i; 775 776 if (sc->sc_provider == NULL || sc->sc_provider->acw == 0) 777 return; 778 sc->sc_idle = 1; 779 g_topology_lock(); 780 for (i = 0; i < sc->sc_ndisks; i++) { 781 disk = &sc->sc_disks[i]; 782 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) 783 continue; 784 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as clean.", 785 g_raid3_get_diskname(disk), sc->sc_name); 786 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 787 g_raid3_update_metadata(disk); 788 } 789 g_topology_unlock(); 790 } 791 792 static void 793 g_raid3_unidle(struct g_raid3_softc *sc) 794 { 795 struct g_raid3_disk *disk; 796 u_int i; 797 798 sc->sc_idle = 0; 799 g_topology_lock(); 800 for (i = 0; i < sc->sc_ndisks; i++) { 801 disk = &sc->sc_disks[i]; 802 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) 803 continue; 804 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as dirty.", 805 g_raid3_get_diskname(disk), sc->sc_name); 806 disk->d_flags |= G_RAID3_DISK_FLAG_DIRTY; 807 g_raid3_update_metadata(disk); 808 } 809 g_topology_unlock(); 810 } 811 812 /* 813 * Return 1 if we should check if RAID3 device is idling. 814 */ 815 static int 816 g_raid3_check_idle(struct g_raid3_softc *sc) 817 { 818 struct g_raid3_disk *disk; 819 u_int i; 820 821 if (sc->sc_idle) 822 return (0); 823 if (sc->sc_provider != NULL && sc->sc_provider->acw == 0) 824 return (0); 825 /* 826 * Check if there are no in-flight requests. 827 */ 828 for (i = 0; i < sc->sc_ndisks; i++) { 829 disk = &sc->sc_disks[i]; 830 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) 831 continue; 832 if (disk->d_consumer->index > 0) 833 return (0); 834 } 835 return (1); 836 } 837 838 /* 839 * Treat bio_driver1 field in parent bio as list head and field bio_caller1 840 * in child bio as pointer to the next element on the list. 841 */ 842 #define G_RAID3_HEAD_BIO(pbp) (pbp)->bio_driver1 843 844 #define G_RAID3_NEXT_BIO(cbp) (cbp)->bio_caller1 845 846 #define G_RAID3_FOREACH_BIO(pbp, bp) \ 847 for ((bp) = G_RAID3_HEAD_BIO(pbp); (bp) != NULL; \ 848 (bp) = G_RAID3_NEXT_BIO(bp)) 849 850 #define G_RAID3_FOREACH_SAFE_BIO(pbp, bp, tmpbp) \ 851 for ((bp) = G_RAID3_HEAD_BIO(pbp); \ 852 (bp) != NULL && ((tmpbp) = G_RAID3_NEXT_BIO(bp), 1); \ 853 (bp) = (tmpbp)) 854 855 static void 856 g_raid3_init_bio(struct bio *pbp) 857 { 858 859 G_RAID3_HEAD_BIO(pbp) = NULL; 860 } 861 862 static void 863 g_raid3_remove_bio(struct bio *cbp) 864 { 865 struct bio *pbp, *bp; 866 867 pbp = cbp->bio_parent; 868 if (G_RAID3_HEAD_BIO(pbp) == cbp) 869 G_RAID3_HEAD_BIO(pbp) = G_RAID3_NEXT_BIO(cbp); 870 else { 871 G_RAID3_FOREACH_BIO(pbp, bp) { 872 if (G_RAID3_NEXT_BIO(bp) == cbp) { 873 G_RAID3_NEXT_BIO(bp) = G_RAID3_NEXT_BIO(cbp); 874 break; 875 } 876 } 877 } 878 G_RAID3_NEXT_BIO(cbp) = NULL; 879 } 880 881 static void 882 g_raid3_replace_bio(struct bio *sbp, struct bio *dbp) 883 { 884 struct bio *pbp, *bp; 885 886 g_raid3_remove_bio(sbp); 887 pbp = dbp->bio_parent; 888 G_RAID3_NEXT_BIO(sbp) = G_RAID3_NEXT_BIO(dbp); 889 if (G_RAID3_HEAD_BIO(pbp) == dbp) 890 G_RAID3_HEAD_BIO(pbp) = sbp; 891 else { 892 G_RAID3_FOREACH_BIO(pbp, bp) { 893 if (G_RAID3_NEXT_BIO(bp) == dbp) { 894 G_RAID3_NEXT_BIO(bp) = sbp; 895 break; 896 } 897 } 898 } 899 G_RAID3_NEXT_BIO(dbp) = NULL; 900 } 901 902 static void 903 g_raid3_destroy_bio(struct g_raid3_softc *sc, struct bio *cbp) 904 { 905 struct bio *bp, *pbp; 906 size_t size; 907 908 pbp = cbp->bio_parent; 909 pbp->bio_children--; 910 KASSERT(cbp->bio_data != NULL, ("NULL bio_data")); 911 size = pbp->bio_length / (sc->sc_ndisks - 1); 912 if (size > 16384) 913 uma_zfree(sc->sc_zone_64k, cbp->bio_data); 914 else if (size > 4096) 915 uma_zfree(sc->sc_zone_16k, cbp->bio_data); 916 else 917 uma_zfree(sc->sc_zone_4k, cbp->bio_data); 918 if (G_RAID3_HEAD_BIO(pbp) == cbp) { 919 G_RAID3_HEAD_BIO(pbp) = G_RAID3_NEXT_BIO(cbp); 920 G_RAID3_NEXT_BIO(cbp) = NULL; 921 g_destroy_bio(cbp); 922 } else { 923 G_RAID3_FOREACH_BIO(pbp, bp) { 924 if (G_RAID3_NEXT_BIO(bp) == cbp) 925 break; 926 } 927 if (bp != NULL) { 928 KASSERT(G_RAID3_NEXT_BIO(bp) != NULL, 929 ("NULL bp->bio_driver1")); 930 G_RAID3_NEXT_BIO(bp) = G_RAID3_NEXT_BIO(cbp); 931 G_RAID3_NEXT_BIO(cbp) = NULL; 932 } 933 g_destroy_bio(cbp); 934 } 935 } 936 937 static struct bio * 938 g_raid3_clone_bio(struct g_raid3_softc *sc, struct bio *pbp) 939 { 940 struct bio *bp, *cbp; 941 size_t size; 942 943 cbp = g_clone_bio(pbp); 944 if (cbp == NULL) 945 return (NULL); 946 size = pbp->bio_length / (sc->sc_ndisks - 1); 947 if (size > 16384) { 948 cbp->bio_data = uma_zalloc(sc->sc_zone_64k, M_NOWAIT); 949 g_raid3_64k_requested++; 950 } else if (size > 4096) { 951 cbp->bio_data = uma_zalloc(sc->sc_zone_16k, M_NOWAIT); 952 g_raid3_16k_requested++; 953 } else { 954 cbp->bio_data = uma_zalloc(sc->sc_zone_4k, M_NOWAIT); 955 g_raid3_4k_requested++; 956 } 957 if (cbp->bio_data == NULL) { 958 if (size > 16384) 959 g_raid3_64k_failed++; 960 if (size > 4096) 961 g_raid3_16k_failed++; 962 else 963 g_raid3_4k_failed++; 964 pbp->bio_children--; 965 g_destroy_bio(cbp); 966 return (NULL); 967 } 968 G_RAID3_NEXT_BIO(cbp) = NULL; 969 if (G_RAID3_HEAD_BIO(pbp) == NULL) 970 G_RAID3_HEAD_BIO(pbp) = cbp; 971 else { 972 G_RAID3_FOREACH_BIO(pbp, bp) { 973 if (G_RAID3_NEXT_BIO(bp) == NULL) { 974 G_RAID3_NEXT_BIO(bp) = cbp; 975 break; 976 } 977 } 978 } 979 return (cbp); 980 } 981 982 static void 983 g_raid3_scatter(struct bio *pbp) 984 { 985 struct g_raid3_softc *sc; 986 struct g_raid3_disk *disk; 987 struct bio *bp, *cbp; 988 off_t atom, cadd, padd, left; 989 990 sc = pbp->bio_to->geom->softc; 991 bp = NULL; 992 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_NOPARITY) == 0) { 993 /* 994 * Find bio for which we should calculate data. 995 */ 996 G_RAID3_FOREACH_BIO(pbp, cbp) { 997 if ((cbp->bio_cflags & G_RAID3_BIO_CFLAG_PARITY) != 0) { 998 bp = cbp; 999 break; 1000 } 1001 } 1002 KASSERT(bp != NULL, ("NULL parity bio.")); 1003 } 1004 atom = sc->sc_sectorsize / (sc->sc_ndisks - 1); 1005 cadd = padd = 0; 1006 for (left = pbp->bio_length; left > 0; left -= sc->sc_sectorsize) { 1007 G_RAID3_FOREACH_BIO(pbp, cbp) { 1008 if (cbp == bp) 1009 continue; 1010 bcopy(pbp->bio_data + padd, cbp->bio_data + cadd, atom); 1011 padd += atom; 1012 } 1013 cadd += atom; 1014 } 1015 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_NOPARITY) == 0) { 1016 struct bio *tmpbp; 1017 1018 /* 1019 * Calculate parity. 1020 */ 1021 bzero(bp->bio_data, bp->bio_length); 1022 G_RAID3_FOREACH_SAFE_BIO(pbp, cbp, tmpbp) { 1023 if (cbp == bp) 1024 continue; 1025 g_raid3_xor(cbp->bio_data, bp->bio_data, bp->bio_data, 1026 bp->bio_length); 1027 if ((cbp->bio_cflags & G_RAID3_BIO_CFLAG_NODISK) != 0) 1028 g_raid3_destroy_bio(sc, cbp); 1029 } 1030 } 1031 G_RAID3_FOREACH_BIO(pbp, cbp) { 1032 struct g_consumer *cp; 1033 1034 disk = cbp->bio_caller2; 1035 cp = disk->d_consumer; 1036 cbp->bio_to = cp->provider; 1037 G_RAID3_LOGREQ(3, cbp, "Sending request."); 1038 KASSERT(cp->acr == 1 && cp->acw == 1 && cp->ace == 1, 1039 ("Consumer %s not opened (r%dw%de%d).", cp->provider->name, 1040 cp->acr, cp->acw, cp->ace)); 1041 cp->index++; 1042 g_io_request(cbp, cp); 1043 } 1044 } 1045 1046 static void 1047 g_raid3_gather(struct bio *pbp) 1048 { 1049 struct g_raid3_softc *sc; 1050 struct g_raid3_disk *disk; 1051 struct bio *xbp, *fbp, *cbp; 1052 off_t atom, cadd, padd, left; 1053 1054 sc = pbp->bio_to->geom->softc; 1055 /* 1056 * Find bio for which we have to calculate data. 1057 * While going through this path, check if all requests 1058 * succeeded, if not, deny whole request. 1059 * If we're in COMPLETE mode, we allow one request to fail, 1060 * so if we find one, we're sending it to the parity consumer. 1061 * If there are more failed requests, we deny whole request. 1062 */ 1063 xbp = fbp = NULL; 1064 G_RAID3_FOREACH_BIO(pbp, cbp) { 1065 if ((cbp->bio_cflags & G_RAID3_BIO_CFLAG_PARITY) != 0) { 1066 KASSERT(xbp == NULL, ("More than one parity bio.")); 1067 xbp = cbp; 1068 } 1069 if (cbp->bio_error == 0) 1070 continue; 1071 /* 1072 * Found failed request. 1073 */ 1074 G_RAID3_LOGREQ(0, cbp, "Request failed."); 1075 disk = cbp->bio_caller2; 1076 if (disk != NULL) { 1077 /* 1078 * Actually this is pointless to bump genid, 1079 * because whole device is fucked up. 1080 */ 1081 sc->sc_bump_id |= G_RAID3_BUMP_GENID; 1082 g_raid3_event_send(disk, 1083 G_RAID3_DISK_STATE_DISCONNECTED, 1084 G_RAID3_EVENT_DONTWAIT); 1085 } 1086 if (fbp == NULL) { 1087 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_DEGRADED) != 0) { 1088 /* 1089 * We are already in degraded mode, so we can't 1090 * accept any failures. 1091 */ 1092 if (pbp->bio_error == 0) 1093 pbp->bio_error = fbp->bio_error; 1094 } else { 1095 fbp = cbp; 1096 } 1097 } else { 1098 /* 1099 * Next failed request, that's too many. 1100 */ 1101 if (pbp->bio_error == 0) 1102 pbp->bio_error = fbp->bio_error; 1103 } 1104 } 1105 if (pbp->bio_error != 0) 1106 goto finish; 1107 if (fbp != NULL && (pbp->bio_pflags & G_RAID3_BIO_PFLAG_VERIFY) != 0) { 1108 pbp->bio_pflags &= ~G_RAID3_BIO_PFLAG_VERIFY; 1109 if (xbp != fbp) 1110 g_raid3_replace_bio(xbp, fbp); 1111 g_raid3_destroy_bio(sc, fbp); 1112 } else if (fbp != NULL) { 1113 struct g_consumer *cp; 1114 1115 /* 1116 * One request failed, so send the same request to 1117 * the parity consumer. 1118 */ 1119 disk = pbp->bio_driver2; 1120 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) { 1121 pbp->bio_error = fbp->bio_error; 1122 goto finish; 1123 } 1124 pbp->bio_pflags |= G_RAID3_BIO_PFLAG_DEGRADED; 1125 pbp->bio_inbed--; 1126 fbp->bio_flags &= ~(BIO_DONE | BIO_ERROR); 1127 if (disk->d_no == sc->sc_ndisks - 1) 1128 fbp->bio_cflags |= G_RAID3_BIO_CFLAG_PARITY; 1129 fbp->bio_error = 0; 1130 fbp->bio_completed = 0; 1131 fbp->bio_children = 0; 1132 fbp->bio_inbed = 0; 1133 cp = disk->d_consumer; 1134 fbp->bio_caller2 = disk; 1135 fbp->bio_to = cp->provider; 1136 G_RAID3_LOGREQ(3, fbp, "Sending request (recover)."); 1137 KASSERT(cp->acr == 1 && cp->acw == 1 && cp->ace == 1, 1138 ("Consumer %s not opened (r%dw%de%d).", cp->provider->name, 1139 cp->acr, cp->acw, cp->ace)); 1140 cp->index++; 1141 g_io_request(fbp, cp); 1142 return; 1143 } 1144 if (xbp != NULL) { 1145 /* 1146 * Calculate parity. 1147 */ 1148 G_RAID3_FOREACH_BIO(pbp, cbp) { 1149 if ((cbp->bio_cflags & G_RAID3_BIO_CFLAG_PARITY) != 0) 1150 continue; 1151 g_raid3_xor(cbp->bio_data, xbp->bio_data, xbp->bio_data, 1152 xbp->bio_length); 1153 } 1154 xbp->bio_cflags &= ~G_RAID3_BIO_CFLAG_PARITY; 1155 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_VERIFY) != 0) { 1156 if (!g_raid3_is_zero(xbp)) { 1157 g_raid3_parity_mismatch++; 1158 pbp->bio_error = EIO; 1159 goto finish; 1160 } 1161 g_raid3_destroy_bio(sc, xbp); 1162 } 1163 } 1164 atom = sc->sc_sectorsize / (sc->sc_ndisks - 1); 1165 cadd = padd = 0; 1166 for (left = pbp->bio_length; left > 0; left -= sc->sc_sectorsize) { 1167 G_RAID3_FOREACH_BIO(pbp, cbp) { 1168 bcopy(cbp->bio_data + cadd, pbp->bio_data + padd, atom); 1169 pbp->bio_completed += atom; 1170 padd += atom; 1171 } 1172 cadd += atom; 1173 } 1174 finish: 1175 if (pbp->bio_error == 0) 1176 G_RAID3_LOGREQ(3, pbp, "Request finished."); 1177 else { 1178 if ((pbp->bio_pflags & G_RAID3_BIO_PFLAG_VERIFY) != 0) 1179 G_RAID3_LOGREQ(1, pbp, "Verification error."); 1180 else 1181 G_RAID3_LOGREQ(0, pbp, "Request failed."); 1182 } 1183 pbp->bio_pflags &= ~G_RAID3_BIO_PFLAG_MASK; 1184 g_io_deliver(pbp, pbp->bio_error); 1185 while ((cbp = G_RAID3_HEAD_BIO(pbp)) != NULL) 1186 g_raid3_destroy_bio(sc, cbp); 1187 } 1188 1189 static void 1190 g_raid3_done(struct bio *bp) 1191 { 1192 struct g_raid3_softc *sc; 1193 1194 sc = bp->bio_from->geom->softc; 1195 bp->bio_cflags |= G_RAID3_BIO_CFLAG_REGULAR; 1196 G_RAID3_LOGREQ(3, bp, "Regular request done (error=%d).", bp->bio_error); 1197 mtx_lock(&sc->sc_queue_mtx); 1198 bioq_insert_head(&sc->sc_queue, bp); 1199 wakeup(sc); 1200 wakeup(&sc->sc_queue); 1201 mtx_unlock(&sc->sc_queue_mtx); 1202 } 1203 1204 static void 1205 g_raid3_regular_request(struct bio *cbp) 1206 { 1207 struct g_raid3_softc *sc; 1208 struct g_raid3_disk *disk; 1209 struct bio *pbp; 1210 1211 g_topology_assert_not(); 1212 1213 cbp->bio_from->index--; 1214 pbp = cbp->bio_parent; 1215 sc = pbp->bio_to->geom->softc; 1216 disk = cbp->bio_from->private; 1217 if (disk == NULL) { 1218 g_topology_lock(); 1219 g_raid3_kill_consumer(sc, cbp->bio_from); 1220 g_topology_unlock(); 1221 } 1222 1223 G_RAID3_LOGREQ(3, cbp, "Request finished."); 1224 pbp->bio_inbed++; 1225 KASSERT(pbp->bio_inbed <= pbp->bio_children, 1226 ("bio_inbed (%u) is bigger than bio_children (%u).", pbp->bio_inbed, 1227 pbp->bio_children)); 1228 if (pbp->bio_inbed != pbp->bio_children) 1229 return; 1230 switch (pbp->bio_cmd) { 1231 case BIO_READ: 1232 g_raid3_gather(pbp); 1233 break; 1234 case BIO_WRITE: 1235 case BIO_DELETE: 1236 { 1237 int error = 0; 1238 1239 pbp->bio_completed = pbp->bio_length; 1240 while ((cbp = G_RAID3_HEAD_BIO(pbp)) != NULL) { 1241 if (cbp->bio_error != 0) { 1242 disk = cbp->bio_caller2; 1243 if (disk != NULL) { 1244 sc->sc_bump_id |= G_RAID3_BUMP_GENID; 1245 g_raid3_event_send(disk, 1246 G_RAID3_DISK_STATE_DISCONNECTED, 1247 G_RAID3_EVENT_DONTWAIT); 1248 } 1249 if (error == 0) 1250 error = cbp->bio_error; 1251 else if (pbp->bio_error == 0) { 1252 /* 1253 * Next failed request, that's too many. 1254 */ 1255 pbp->bio_error = error; 1256 } 1257 } 1258 g_raid3_destroy_bio(sc, cbp); 1259 } 1260 if (pbp->bio_error == 0) 1261 G_RAID3_LOGREQ(3, pbp, "Request finished."); 1262 else 1263 G_RAID3_LOGREQ(0, pbp, "Request failed."); 1264 pbp->bio_pflags &= ~G_RAID3_BIO_PFLAG_DEGRADED; 1265 pbp->bio_pflags &= ~G_RAID3_BIO_PFLAG_NOPARITY; 1266 g_io_deliver(pbp, pbp->bio_error); 1267 break; 1268 } 1269 } 1270 } 1271 1272 static void 1273 g_raid3_sync_done(struct bio *bp) 1274 { 1275 struct g_raid3_softc *sc; 1276 1277 G_RAID3_LOGREQ(3, bp, "Synchronization request delivered."); 1278 sc = bp->bio_from->geom->softc; 1279 bp->bio_cflags |= G_RAID3_BIO_CFLAG_SYNC; 1280 mtx_lock(&sc->sc_queue_mtx); 1281 bioq_insert_head(&sc->sc_queue, bp); 1282 wakeup(sc); 1283 wakeup(&sc->sc_queue); 1284 mtx_unlock(&sc->sc_queue_mtx); 1285 } 1286 1287 static void 1288 g_raid3_start(struct bio *bp) 1289 { 1290 struct g_raid3_softc *sc; 1291 1292 sc = bp->bio_to->geom->softc; 1293 /* 1294 * If sc == NULL or there are no valid disks, provider's error 1295 * should be set and g_raid3_start() should not be called at all. 1296 */ 1297 KASSERT(sc != NULL && (sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 1298 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE), 1299 ("Provider's error should be set (error=%d)(device=%s).", 1300 bp->bio_to->error, bp->bio_to->name)); 1301 G_RAID3_LOGREQ(3, bp, "Request received."); 1302 1303 switch (bp->bio_cmd) { 1304 case BIO_READ: 1305 case BIO_WRITE: 1306 case BIO_DELETE: 1307 break; 1308 case BIO_GETATTR: 1309 default: 1310 g_io_deliver(bp, EOPNOTSUPP); 1311 return; 1312 } 1313 mtx_lock(&sc->sc_queue_mtx); 1314 bioq_insert_tail(&sc->sc_queue, bp); 1315 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, sc); 1316 wakeup(sc); 1317 mtx_unlock(&sc->sc_queue_mtx); 1318 } 1319 1320 /* 1321 * Send one synchronization request. 1322 */ 1323 static void 1324 g_raid3_sync_one(struct g_raid3_softc *sc) 1325 { 1326 struct g_raid3_disk *disk; 1327 struct bio *bp; 1328 1329 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED, 1330 ("Wrong device state (%s, %s).", sc->sc_name, 1331 g_raid3_device_state2str(sc->sc_state))); 1332 disk = sc->sc_syncdisk; 1333 KASSERT(disk != NULL, ("No sync disk (%s).", sc->sc_name)); 1334 KASSERT(disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING, 1335 ("Disk %s is not marked for synchronization.", 1336 g_raid3_get_diskname(disk))); 1337 1338 bp = g_new_bio(); 1339 if (bp == NULL) 1340 return; 1341 bp->bio_parent = NULL; 1342 bp->bio_cmd = BIO_READ; 1343 bp->bio_offset = disk->d_sync.ds_offset * (sc->sc_ndisks - 1); 1344 bp->bio_length = MIN(MAXPHYS, sc->sc_mediasize - bp->bio_offset); 1345 bp->bio_cflags = 0; 1346 bp->bio_done = g_raid3_sync_done; 1347 bp->bio_data = disk->d_sync.ds_data; 1348 if (bp->bio_data == NULL) { 1349 g_destroy_bio(bp); 1350 return; 1351 } 1352 bp->bio_cflags = G_RAID3_BIO_CFLAG_REGSYNC; 1353 disk->d_sync.ds_offset += bp->bio_length / (sc->sc_ndisks - 1); 1354 bp->bio_to = sc->sc_provider; 1355 G_RAID3_LOGREQ(3, bp, "Sending synchronization request."); 1356 disk->d_sync.ds_consumer->index++; 1357 g_io_request(bp, disk->d_sync.ds_consumer); 1358 } 1359 1360 static void 1361 g_raid3_sync_request(struct bio *bp) 1362 { 1363 struct g_raid3_softc *sc; 1364 struct g_raid3_disk *disk; 1365 1366 bp->bio_from->index--; 1367 sc = bp->bio_from->geom->softc; 1368 disk = bp->bio_from->private; 1369 if (disk == NULL) { 1370 g_topology_lock(); 1371 g_raid3_kill_consumer(sc, bp->bio_from); 1372 g_topology_unlock(); 1373 g_destroy_bio(bp); 1374 return; 1375 } 1376 1377 /* 1378 * Synchronization request. 1379 */ 1380 switch (bp->bio_cmd) { 1381 case BIO_READ: 1382 { 1383 struct g_consumer *cp; 1384 u_char *dst, *src; 1385 off_t left; 1386 u_int atom; 1387 1388 if (bp->bio_error != 0) { 1389 G_RAID3_LOGREQ(0, bp, 1390 "Synchronization request failed (error=%d).", 1391 bp->bio_error); 1392 g_destroy_bio(bp); 1393 return; 1394 } 1395 G_RAID3_LOGREQ(3, bp, "Synchronization request finished."); 1396 atom = sc->sc_sectorsize / (sc->sc_ndisks - 1); 1397 dst = src = bp->bio_data; 1398 if (disk->d_no == sc->sc_ndisks - 1) { 1399 u_int n; 1400 1401 /* Parity component. */ 1402 for (left = bp->bio_length; left > 0; 1403 left -= sc->sc_sectorsize) { 1404 bcopy(src, dst, atom); 1405 src += atom; 1406 for (n = 1; n < sc->sc_ndisks - 1; n++) { 1407 g_raid3_xor(src, dst, dst, atom); 1408 src += atom; 1409 } 1410 dst += atom; 1411 } 1412 } else { 1413 /* Regular component. */ 1414 src += atom * disk->d_no; 1415 for (left = bp->bio_length; left > 0; 1416 left -= sc->sc_sectorsize) { 1417 bcopy(src, dst, atom); 1418 src += sc->sc_sectorsize; 1419 dst += atom; 1420 } 1421 } 1422 bp->bio_offset /= sc->sc_ndisks - 1; 1423 bp->bio_length /= sc->sc_ndisks - 1; 1424 bp->bio_cmd = BIO_WRITE; 1425 bp->bio_cflags = 0; 1426 bp->bio_children = bp->bio_inbed = 0; 1427 cp = disk->d_consumer; 1428 KASSERT(cp->acr == 1 && cp->acw == 1 && cp->ace == 1, 1429 ("Consumer %s not opened (r%dw%de%d).", cp->provider->name, 1430 cp->acr, cp->acw, cp->ace)); 1431 cp->index++; 1432 g_io_request(bp, cp); 1433 return; 1434 } 1435 case BIO_WRITE: 1436 { 1437 struct g_raid3_disk_sync *sync; 1438 1439 if (bp->bio_error != 0) { 1440 G_RAID3_LOGREQ(0, bp, 1441 "Synchronization request failed (error=%d).", 1442 bp->bio_error); 1443 g_destroy_bio(bp); 1444 sc->sc_bump_id |= G_RAID3_BUMP_GENID; 1445 g_raid3_event_send(disk, 1446 G_RAID3_DISK_STATE_DISCONNECTED, 1447 G_RAID3_EVENT_DONTWAIT); 1448 return; 1449 } 1450 G_RAID3_LOGREQ(3, bp, "Synchronization request finished."); 1451 sync = &disk->d_sync; 1452 sync->ds_offset_done = bp->bio_offset + bp->bio_length; 1453 g_destroy_bio(bp); 1454 if (sync->ds_resync != -1) 1455 return; 1456 if (sync->ds_offset_done == 1457 sc->sc_mediasize / (sc->sc_ndisks - 1)) { 1458 /* 1459 * Disk up-to-date, activate it. 1460 */ 1461 g_raid3_event_send(disk, G_RAID3_DISK_STATE_ACTIVE, 1462 G_RAID3_EVENT_DONTWAIT); 1463 return; 1464 } else if (sync->ds_offset_done % (MAXPHYS * 100) == 0) { 1465 /* 1466 * Update offset_done on every 100 blocks. 1467 * XXX: This should be configurable. 1468 */ 1469 g_topology_lock(); 1470 g_raid3_update_metadata(disk); 1471 g_topology_unlock(); 1472 } 1473 return; 1474 } 1475 default: 1476 KASSERT(1 == 0, ("Invalid command here: %u (device=%s)", 1477 bp->bio_cmd, sc->sc_name)); 1478 break; 1479 } 1480 } 1481 1482 static int 1483 g_raid3_register_request(struct bio *pbp) 1484 { 1485 struct g_raid3_softc *sc; 1486 struct g_raid3_disk *disk; 1487 struct g_consumer *cp; 1488 struct bio *cbp; 1489 off_t offset, length; 1490 u_int n, ndisks; 1491 int round_robin, verify; 1492 1493 ndisks = 0; 1494 sc = pbp->bio_to->geom->softc; 1495 if ((pbp->bio_cflags & G_RAID3_BIO_CFLAG_REGSYNC) != 0 && 1496 sc->sc_syncdisk == NULL) { 1497 g_io_deliver(pbp, EIO); 1498 return (0); 1499 } 1500 g_raid3_init_bio(pbp); 1501 length = pbp->bio_length / (sc->sc_ndisks - 1); 1502 offset = pbp->bio_offset / (sc->sc_ndisks - 1); 1503 round_robin = verify = 0; 1504 switch (pbp->bio_cmd) { 1505 case BIO_READ: 1506 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_VERIFY) != 0 && 1507 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE) { 1508 pbp->bio_pflags |= G_RAID3_BIO_PFLAG_VERIFY; 1509 verify = 1; 1510 ndisks = sc->sc_ndisks; 1511 } else { 1512 verify = 0; 1513 ndisks = sc->sc_ndisks - 1; 1514 } 1515 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_ROUND_ROBIN) != 0 && 1516 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE) { 1517 round_robin = 1; 1518 } else { 1519 round_robin = 0; 1520 } 1521 KASSERT(!round_robin || !verify, 1522 ("ROUND-ROBIN and VERIFY are mutually exclusive.")); 1523 pbp->bio_driver2 = &sc->sc_disks[sc->sc_ndisks - 1]; 1524 break; 1525 case BIO_WRITE: 1526 case BIO_DELETE: 1527 { 1528 struct g_raid3_disk_sync *sync; 1529 1530 if (sc->sc_idle) 1531 g_raid3_unidle(sc); 1532 1533 ndisks = sc->sc_ndisks; 1534 1535 if (sc->sc_syncdisk == NULL) 1536 break; 1537 sync = &sc->sc_syncdisk->d_sync; 1538 if (offset >= sync->ds_offset) 1539 break; 1540 if (offset + length <= sync->ds_offset_done) 1541 break; 1542 if (offset >= sync->ds_resync && sync->ds_resync != -1) 1543 break; 1544 sync->ds_resync = offset - (offset % MAXPHYS); 1545 break; 1546 } 1547 } 1548 for (n = 0; n < ndisks; n++) { 1549 disk = &sc->sc_disks[n]; 1550 cbp = g_raid3_clone_bio(sc, pbp); 1551 if (cbp == NULL) { 1552 while ((cbp = G_RAID3_HEAD_BIO(pbp)) != NULL) 1553 g_raid3_destroy_bio(sc, cbp); 1554 return (ENOMEM); 1555 } 1556 cbp->bio_offset = offset; 1557 cbp->bio_length = length; 1558 cbp->bio_done = g_raid3_done; 1559 switch (pbp->bio_cmd) { 1560 case BIO_READ: 1561 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) { 1562 /* 1563 * Replace invalid component with the parity 1564 * component. 1565 */ 1566 disk = &sc->sc_disks[sc->sc_ndisks - 1]; 1567 cbp->bio_cflags |= G_RAID3_BIO_CFLAG_PARITY; 1568 pbp->bio_pflags |= G_RAID3_BIO_PFLAG_DEGRADED; 1569 } else if (round_robin && 1570 disk->d_no == sc->sc_round_robin) { 1571 /* 1572 * In round-robin mode skip one data component 1573 * and use parity component when reading. 1574 */ 1575 pbp->bio_driver2 = disk; 1576 disk = &sc->sc_disks[sc->sc_ndisks - 1]; 1577 cbp->bio_cflags |= G_RAID3_BIO_CFLAG_PARITY; 1578 sc->sc_round_robin++; 1579 round_robin = 0; 1580 } else if (verify && disk->d_no == sc->sc_ndisks - 1) { 1581 cbp->bio_cflags |= G_RAID3_BIO_CFLAG_PARITY; 1582 } 1583 break; 1584 case BIO_WRITE: 1585 case BIO_DELETE: 1586 if (disk->d_state == G_RAID3_DISK_STATE_ACTIVE || 1587 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 1588 if (n == ndisks - 1) { 1589 /* 1590 * Active parity component, mark it as such. 1591 */ 1592 cbp->bio_cflags |= 1593 G_RAID3_BIO_CFLAG_PARITY; 1594 } 1595 } else { 1596 pbp->bio_pflags |= G_RAID3_BIO_PFLAG_DEGRADED; 1597 if (n == ndisks - 1) { 1598 /* 1599 * Parity component is not connected, 1600 * so destroy its request. 1601 */ 1602 pbp->bio_pflags |= 1603 G_RAID3_BIO_PFLAG_NOPARITY; 1604 g_raid3_destroy_bio(sc, cbp); 1605 cbp = NULL; 1606 } else { 1607 cbp->bio_cflags |= 1608 G_RAID3_BIO_CFLAG_NODISK; 1609 disk = NULL; 1610 } 1611 } 1612 break; 1613 } 1614 if (cbp != NULL) 1615 cbp->bio_caller2 = disk; 1616 } 1617 switch (pbp->bio_cmd) { 1618 case BIO_READ: 1619 if (round_robin) { 1620 /* 1621 * If we are in round-robin mode and 'round_robin' is 1622 * still 1, it means, that we skipped parity component 1623 * for this read and must reset sc_round_robin field. 1624 */ 1625 sc->sc_round_robin = 0; 1626 } 1627 G_RAID3_FOREACH_BIO(pbp, cbp) { 1628 disk = cbp->bio_caller2; 1629 cp = disk->d_consumer; 1630 cbp->bio_to = cp->provider; 1631 G_RAID3_LOGREQ(3, cbp, "Sending request."); 1632 KASSERT(cp->acr == 1 && cp->acw == 1 && cp->ace == 1, 1633 ("Consumer %s not opened (r%dw%de%d).", 1634 cp->provider->name, cp->acr, cp->acw, cp->ace)); 1635 cp->index++; 1636 g_io_request(cbp, cp); 1637 } 1638 break; 1639 case BIO_WRITE: 1640 case BIO_DELETE: 1641 /* 1642 * Bump syncid on first write. 1643 */ 1644 if ((sc->sc_bump_id & G_RAID3_BUMP_SYNCID) != 0) { 1645 sc->sc_bump_id &= ~G_RAID3_BUMP_SYNCID; 1646 g_topology_lock(); 1647 g_raid3_bump_syncid(sc); 1648 g_topology_unlock(); 1649 } 1650 g_raid3_scatter(pbp); 1651 break; 1652 } 1653 return (0); 1654 } 1655 1656 static int 1657 g_raid3_can_destroy(struct g_raid3_softc *sc) 1658 { 1659 struct g_geom *gp; 1660 struct g_consumer *cp; 1661 1662 g_topology_assert(); 1663 gp = sc->sc_geom; 1664 LIST_FOREACH(cp, &gp->consumer, consumer) { 1665 if (g_raid3_is_busy(sc, cp)) 1666 return (0); 1667 } 1668 gp = sc->sc_sync.ds_geom; 1669 LIST_FOREACH(cp, &gp->consumer, consumer) { 1670 if (g_raid3_is_busy(sc, cp)) 1671 return (0); 1672 } 1673 G_RAID3_DEBUG(2, "No I/O requests for %s, it can be destroyed.", 1674 sc->sc_name); 1675 return (1); 1676 } 1677 1678 static int 1679 g_raid3_try_destroy(struct g_raid3_softc *sc) 1680 { 1681 1682 g_topology_lock(); 1683 if (!g_raid3_can_destroy(sc)) { 1684 g_topology_unlock(); 1685 return (0); 1686 } 1687 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_WAIT) != 0) { 1688 g_topology_unlock(); 1689 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, 1690 &sc->sc_worker); 1691 wakeup(&sc->sc_worker); 1692 sc->sc_worker = NULL; 1693 } else { 1694 g_raid3_destroy_device(sc); 1695 g_topology_unlock(); 1696 free(sc->sc_disks, M_RAID3); 1697 free(sc, M_RAID3); 1698 } 1699 return (1); 1700 } 1701 1702 /* 1703 * Worker thread. 1704 */ 1705 static void 1706 g_raid3_worker(void *arg) 1707 { 1708 struct g_raid3_softc *sc; 1709 struct g_raid3_disk *disk; 1710 struct g_raid3_disk_sync *sync; 1711 struct g_raid3_event *ep; 1712 struct bio *bp; 1713 u_int nreqs; 1714 1715 sc = arg; 1716 mtx_lock_spin(&sched_lock); 1717 sched_prio(curthread, PRIBIO); 1718 mtx_unlock_spin(&sched_lock); 1719 1720 nreqs = 0; 1721 for (;;) { 1722 G_RAID3_DEBUG(5, "%s: Let's see...", __func__); 1723 /* 1724 * First take a look at events. 1725 * This is important to handle events before any I/O requests. 1726 */ 1727 ep = g_raid3_event_get(sc); 1728 if (ep != NULL && g_topology_try_lock()) { 1729 g_raid3_event_remove(sc, ep); 1730 if ((ep->e_flags & G_RAID3_EVENT_DEVICE) != 0) { 1731 /* Update only device status. */ 1732 G_RAID3_DEBUG(3, 1733 "Running event for device %s.", 1734 sc->sc_name); 1735 ep->e_error = 0; 1736 g_raid3_update_device(sc, 1); 1737 } else { 1738 /* Update disk status. */ 1739 G_RAID3_DEBUG(3, "Running event for disk %s.", 1740 g_raid3_get_diskname(ep->e_disk)); 1741 ep->e_error = g_raid3_update_disk(ep->e_disk, 1742 ep->e_state); 1743 if (ep->e_error == 0) 1744 g_raid3_update_device(sc, 0); 1745 } 1746 g_topology_unlock(); 1747 if ((ep->e_flags & G_RAID3_EVENT_DONTWAIT) != 0) { 1748 KASSERT(ep->e_error == 0, 1749 ("Error cannot be handled.")); 1750 g_raid3_event_free(ep); 1751 } else { 1752 ep->e_flags |= G_RAID3_EVENT_DONE; 1753 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, 1754 ep); 1755 mtx_lock(&sc->sc_events_mtx); 1756 wakeup(ep); 1757 mtx_unlock(&sc->sc_events_mtx); 1758 } 1759 if ((sc->sc_flags & 1760 G_RAID3_DEVICE_FLAG_DESTROY) != 0) { 1761 if (g_raid3_try_destroy(sc)) 1762 kthread_exit(0); 1763 } 1764 G_RAID3_DEBUG(5, "%s: I'm here 1.", __func__); 1765 continue; 1766 } 1767 /* 1768 * Now I/O requests. 1769 */ 1770 /* Get first request from the queue. */ 1771 mtx_lock(&sc->sc_queue_mtx); 1772 bp = bioq_first(&sc->sc_queue); 1773 if (bp == NULL) { 1774 if (ep != NULL) { 1775 /* 1776 * No I/O requests and topology lock was 1777 * already held? Try again. 1778 */ 1779 mtx_unlock(&sc->sc_queue_mtx); 1780 tsleep(ep, PRIBIO, "r3:top1", hz / 5); 1781 continue; 1782 } 1783 if ((sc->sc_flags & 1784 G_RAID3_DEVICE_FLAG_DESTROY) != 0) { 1785 mtx_unlock(&sc->sc_queue_mtx); 1786 if (g_raid3_try_destroy(sc)) 1787 kthread_exit(0); 1788 mtx_lock(&sc->sc_queue_mtx); 1789 } 1790 } 1791 if (sc->sc_syncdisk != NULL && 1792 (bp == NULL || nreqs > g_raid3_reqs_per_sync)) { 1793 mtx_unlock(&sc->sc_queue_mtx); 1794 /* 1795 * It is time for synchronization... 1796 */ 1797 nreqs = 0; 1798 disk = sc->sc_syncdisk; 1799 sync = &disk->d_sync; 1800 if (sync->ds_offset < 1801 sc->sc_mediasize / (sc->sc_ndisks - 1) && 1802 sync->ds_offset == sync->ds_offset_done) { 1803 if (sync->ds_resync != -1) { 1804 sync->ds_offset = sync->ds_resync; 1805 sync->ds_offset_done = sync->ds_resync; 1806 sync->ds_resync = -1; 1807 } 1808 g_raid3_sync_one(sc); 1809 } 1810 G_RAID3_DEBUG(5, "%s: I'm here 2.", __func__); 1811 goto sleep; 1812 } 1813 if (bp == NULL) { 1814 if (g_raid3_check_idle(sc)) { 1815 u_int idletime; 1816 1817 idletime = g_raid3_idletime; 1818 if (idletime == 0) 1819 idletime = 1; 1820 idletime *= hz; 1821 if (msleep(sc, &sc->sc_queue_mtx, PRIBIO | PDROP, 1822 "r3:w1", idletime) == EWOULDBLOCK) { 1823 G_RAID3_DEBUG(5, "%s: I'm here 3.", 1824 __func__); 1825 /* 1826 * No I/O requests in 'idletime' 1827 * seconds, so mark components as clean. 1828 */ 1829 g_raid3_idle(sc); 1830 } 1831 G_RAID3_DEBUG(5, "%s: I'm here 4.", __func__); 1832 } else { 1833 MSLEEP(sc, &sc->sc_queue_mtx, PRIBIO | PDROP, 1834 "r3:w2", 0); 1835 G_RAID3_DEBUG(5, "%s: I'm here 5.", __func__); 1836 } 1837 continue; 1838 } 1839 nreqs++; 1840 bioq_remove(&sc->sc_queue, bp); 1841 mtx_unlock(&sc->sc_queue_mtx); 1842 1843 if ((bp->bio_cflags & G_RAID3_BIO_CFLAG_REGULAR) != 0) { 1844 g_raid3_regular_request(bp); 1845 } else if ((bp->bio_cflags & G_RAID3_BIO_CFLAG_SYNC) != 0) { 1846 u_int timeout, sps; 1847 1848 g_raid3_sync_request(bp); 1849 sleep: 1850 sps = atomic_load_acq_int(&g_raid3_syncs_per_sec); 1851 if (sps == 0) { 1852 G_RAID3_DEBUG(5, "%s: I'm here 6.", __func__); 1853 continue; 1854 } 1855 if (ep != NULL) { 1856 /* 1857 * We have some pending events, don't sleep now. 1858 */ 1859 G_RAID3_DEBUG(5, "%s: I'm here 7.", __func__); 1860 tsleep(ep, PRIBIO, "r3:top2", hz / 5); 1861 continue; 1862 } 1863 mtx_lock(&sc->sc_queue_mtx); 1864 if (bioq_first(&sc->sc_queue) != NULL) { 1865 mtx_unlock(&sc->sc_queue_mtx); 1866 G_RAID3_DEBUG(5, "%s: I'm here 8.", __func__); 1867 continue; 1868 } 1869 timeout = hz / sps; 1870 if (timeout == 0) 1871 timeout = 1; 1872 MSLEEP(sc, &sc->sc_queue_mtx, PRIBIO | PDROP, "r3:w2", 1873 timeout); 1874 } else { 1875 if (g_raid3_register_request(bp) != 0) { 1876 mtx_lock(&sc->sc_queue_mtx); 1877 bioq_insert_tail(&sc->sc_queue, bp); 1878 MSLEEP(&sc->sc_queue, &sc->sc_queue_mtx, 1879 PRIBIO | PDROP, "r3:lowmem", hz / 10); 1880 } 1881 } 1882 G_RAID3_DEBUG(5, "%s: I'm here 9.", __func__); 1883 } 1884 } 1885 1886 /* 1887 * Open disk's consumer if needed. 1888 */ 1889 static void 1890 g_raid3_update_access(struct g_raid3_disk *disk) 1891 { 1892 struct g_provider *pp; 1893 1894 g_topology_assert(); 1895 1896 pp = disk->d_softc->sc_provider; 1897 if (pp == NULL) 1898 return; 1899 if (pp->acw > 0) { 1900 if ((disk->d_flags & G_RAID3_DISK_FLAG_DIRTY) == 0) { 1901 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as dirty.", 1902 g_raid3_get_diskname(disk), disk->d_softc->sc_name); 1903 disk->d_flags |= G_RAID3_DISK_FLAG_DIRTY; 1904 } 1905 } else if (pp->acw == 0) { 1906 if ((disk->d_flags & G_RAID3_DISK_FLAG_DIRTY) != 0) { 1907 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as clean.", 1908 g_raid3_get_diskname(disk), disk->d_softc->sc_name); 1909 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 1910 } 1911 } 1912 } 1913 1914 static void 1915 g_raid3_sync_start(struct g_raid3_softc *sc) 1916 { 1917 struct g_raid3_disk *disk; 1918 int error; 1919 u_int n; 1920 1921 g_topology_assert(); 1922 1923 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED, 1924 ("Device not in DEGRADED state (%s, %u).", sc->sc_name, 1925 sc->sc_state)); 1926 KASSERT(sc->sc_syncdisk == NULL, ("Syncdisk is not NULL (%s, %u).", 1927 sc->sc_name, sc->sc_state)); 1928 disk = NULL; 1929 for (n = 0; n < sc->sc_ndisks; n++) { 1930 if (sc->sc_disks[n].d_state != G_RAID3_DISK_STATE_SYNCHRONIZING) 1931 continue; 1932 disk = &sc->sc_disks[n]; 1933 break; 1934 } 1935 if (disk == NULL) 1936 return; 1937 1938 G_RAID3_DEBUG(0, "Device %s: rebuilding provider %s.", sc->sc_name, 1939 g_raid3_get_diskname(disk)); 1940 disk->d_flags |= G_RAID3_DISK_FLAG_DIRTY; 1941 KASSERT(disk->d_sync.ds_consumer == NULL, 1942 ("Sync consumer already exists (device=%s, disk=%s).", 1943 sc->sc_name, g_raid3_get_diskname(disk))); 1944 disk->d_sync.ds_consumer = g_new_consumer(sc->sc_sync.ds_geom); 1945 disk->d_sync.ds_consumer->private = disk; 1946 disk->d_sync.ds_consumer->index = 0; 1947 error = g_attach(disk->d_sync.ds_consumer, disk->d_softc->sc_provider); 1948 KASSERT(error == 0, ("Cannot attach to %s (error=%d).", 1949 disk->d_softc->sc_name, error)); 1950 error = g_access(disk->d_sync.ds_consumer, 1, 0, 0); 1951 KASSERT(error == 0, ("Cannot open %s (error=%d).", 1952 disk->d_softc->sc_name, error)); 1953 disk->d_sync.ds_data = malloc(MAXPHYS, M_RAID3, M_WAITOK); 1954 sc->sc_syncdisk = disk; 1955 } 1956 1957 /* 1958 * Stop synchronization process. 1959 * type: 0 - synchronization finished 1960 * 1 - synchronization stopped 1961 */ 1962 static void 1963 g_raid3_sync_stop(struct g_raid3_softc *sc, int type) 1964 { 1965 struct g_raid3_disk *disk; 1966 1967 g_topology_assert(); 1968 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED, 1969 ("Device not in DEGRADED state (%s, %u).", sc->sc_name, 1970 sc->sc_state)); 1971 disk = sc->sc_syncdisk; 1972 sc->sc_syncdisk = NULL; 1973 KASSERT(disk != NULL, ("No disk was synchronized (%s).", sc->sc_name)); 1974 KASSERT(disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING, 1975 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 1976 g_raid3_disk_state2str(disk->d_state))); 1977 if (disk->d_sync.ds_consumer == NULL) 1978 return; 1979 1980 if (type == 0) { 1981 G_RAID3_DEBUG(0, "Device %s: rebuilding provider %s finished.", 1982 disk->d_softc->sc_name, g_raid3_get_diskname(disk)); 1983 } else /* if (type == 1) */ { 1984 G_RAID3_DEBUG(0, "Device %s: rebuilding provider %s stopped.", 1985 disk->d_softc->sc_name, g_raid3_get_diskname(disk)); 1986 } 1987 g_raid3_kill_consumer(disk->d_softc, disk->d_sync.ds_consumer); 1988 free(disk->d_sync.ds_data, M_RAID3); 1989 disk->d_sync.ds_consumer = NULL; 1990 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 1991 } 1992 1993 static void 1994 g_raid3_launch_provider(struct g_raid3_softc *sc) 1995 { 1996 struct g_provider *pp; 1997 1998 g_topology_assert(); 1999 2000 pp = g_new_providerf(sc->sc_geom, "raid3/%s", sc->sc_name); 2001 pp->mediasize = sc->sc_mediasize; 2002 pp->sectorsize = sc->sc_sectorsize; 2003 sc->sc_provider = pp; 2004 g_error_provider(pp, 0); 2005 G_RAID3_DEBUG(0, "Device %s: provider %s launched.", sc->sc_name, 2006 pp->name); 2007 if (sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED) 2008 g_raid3_sync_start(sc); 2009 } 2010 2011 static void 2012 g_raid3_destroy_provider(struct g_raid3_softc *sc) 2013 { 2014 struct bio *bp; 2015 2016 g_topology_assert(); 2017 KASSERT(sc->sc_provider != NULL, ("NULL provider (device=%s).", 2018 sc->sc_name)); 2019 2020 g_error_provider(sc->sc_provider, ENXIO); 2021 mtx_lock(&sc->sc_queue_mtx); 2022 while ((bp = bioq_first(&sc->sc_queue)) != NULL) { 2023 bioq_remove(&sc->sc_queue, bp); 2024 g_io_deliver(bp, ENXIO); 2025 } 2026 mtx_unlock(&sc->sc_queue_mtx); 2027 G_RAID3_DEBUG(0, "Device %s: provider %s destroyed.", sc->sc_name, 2028 sc->sc_provider->name); 2029 sc->sc_provider->flags |= G_PF_WITHER; 2030 g_orphan_provider(sc->sc_provider, ENXIO); 2031 sc->sc_provider = NULL; 2032 if (sc->sc_syncdisk != NULL) 2033 g_raid3_sync_stop(sc, 1); 2034 } 2035 2036 static void 2037 g_raid3_go(void *arg) 2038 { 2039 struct g_raid3_softc *sc; 2040 2041 sc = arg; 2042 G_RAID3_DEBUG(0, "Force device %s start due to timeout.", sc->sc_name); 2043 g_raid3_event_send(sc, 0, 2044 G_RAID3_EVENT_DONTWAIT | G_RAID3_EVENT_DEVICE); 2045 } 2046 2047 static u_int 2048 g_raid3_determine_state(struct g_raid3_disk *disk) 2049 { 2050 struct g_raid3_softc *sc; 2051 u_int state; 2052 2053 sc = disk->d_softc; 2054 if (sc->sc_syncid == disk->d_sync.ds_syncid) { 2055 if ((disk->d_flags & 2056 G_RAID3_DISK_FLAG_SYNCHRONIZING) == 0) { 2057 /* Disk does not need synchronization. */ 2058 state = G_RAID3_DISK_STATE_ACTIVE; 2059 } else { 2060 if ((sc->sc_flags & 2061 G_RAID3_DEVICE_FLAG_NOAUTOSYNC) == 0 || 2062 (disk->d_flags & 2063 G_RAID3_DISK_FLAG_FORCE_SYNC) != 0) { 2064 /* 2065 * We can start synchronization from 2066 * the stored offset. 2067 */ 2068 state = G_RAID3_DISK_STATE_SYNCHRONIZING; 2069 } else { 2070 state = G_RAID3_DISK_STATE_STALE; 2071 } 2072 } 2073 } else if (disk->d_sync.ds_syncid < sc->sc_syncid) { 2074 /* 2075 * Reset all synchronization data for this disk, 2076 * because if it even was synchronized, it was 2077 * synchronized to disks with different syncid. 2078 */ 2079 disk->d_flags |= G_RAID3_DISK_FLAG_SYNCHRONIZING; 2080 disk->d_sync.ds_offset = 0; 2081 disk->d_sync.ds_offset_done = 0; 2082 disk->d_sync.ds_syncid = sc->sc_syncid; 2083 if ((sc->sc_flags & G_RAID3_DEVICE_FLAG_NOAUTOSYNC) == 0 || 2084 (disk->d_flags & G_RAID3_DISK_FLAG_FORCE_SYNC) != 0) { 2085 state = G_RAID3_DISK_STATE_SYNCHRONIZING; 2086 } else { 2087 state = G_RAID3_DISK_STATE_STALE; 2088 } 2089 } else /* if (sc->sc_syncid < disk->d_sync.ds_syncid) */ { 2090 /* 2091 * Not good, NOT GOOD! 2092 * It means that device was started on stale disks 2093 * and more fresh disk just arrive. 2094 * If there were writes, device is fucked up, sorry. 2095 * I think the best choice here is don't touch 2096 * this disk and inform the user laudly. 2097 */ 2098 G_RAID3_DEBUG(0, "Device %s was started before the freshest " 2099 "disk (%s) arrives!! It will not be connected to the " 2100 "running device.", sc->sc_name, 2101 g_raid3_get_diskname(disk)); 2102 g_raid3_destroy_disk(disk); 2103 state = G_RAID3_DISK_STATE_NONE; 2104 /* Return immediately, because disk was destroyed. */ 2105 return (state); 2106 } 2107 G_RAID3_DEBUG(3, "State for %s disk: %s.", 2108 g_raid3_get_diskname(disk), g_raid3_disk_state2str(state)); 2109 return (state); 2110 } 2111 2112 /* 2113 * Update device state. 2114 */ 2115 static void 2116 g_raid3_update_device(struct g_raid3_softc *sc, boolean_t force) 2117 { 2118 struct g_raid3_disk *disk; 2119 u_int state; 2120 2121 g_topology_assert(); 2122 2123 switch (sc->sc_state) { 2124 case G_RAID3_DEVICE_STATE_STARTING: 2125 { 2126 u_int n, ndirty, ndisks, genid, syncid; 2127 2128 KASSERT(sc->sc_provider == NULL, 2129 ("Non-NULL provider in STARTING state (%s).", sc->sc_name)); 2130 /* 2131 * Are we ready? We are, if all disks are connected or 2132 * one disk is missing and 'force' is true. 2133 */ 2134 if (g_raid3_ndisks(sc, -1) + force == sc->sc_ndisks) { 2135 if (!force) 2136 callout_drain(&sc->sc_callout); 2137 } else { 2138 if (force) { 2139 /* 2140 * Timeout expired, so destroy device. 2141 */ 2142 sc->sc_flags |= G_RAID3_DEVICE_FLAG_DESTROY; 2143 } 2144 return; 2145 } 2146 2147 /* 2148 * Find the biggest genid. 2149 */ 2150 genid = 0; 2151 for (n = 0; n < sc->sc_ndisks; n++) { 2152 disk = &sc->sc_disks[n]; 2153 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 2154 continue; 2155 if (disk->d_genid > genid) 2156 genid = disk->d_genid; 2157 } 2158 sc->sc_genid = genid; 2159 /* 2160 * Remove all disks without the biggest genid. 2161 */ 2162 for (n = 0; n < sc->sc_ndisks; n++) { 2163 disk = &sc->sc_disks[n]; 2164 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 2165 continue; 2166 if (disk->d_genid < genid) { 2167 G_RAID3_DEBUG(0, 2168 "Component %s (device %s) broken, skipping.", 2169 g_raid3_get_diskname(disk), sc->sc_name); 2170 g_raid3_destroy_disk(disk); 2171 } 2172 } 2173 2174 /* 2175 * There must be at least 'sc->sc_ndisks - 1' components 2176 * with the same syncid and without SYNCHRONIZING flag. 2177 */ 2178 2179 /* 2180 * Find the biggest syncid, number of valid components and 2181 * number of dirty components. 2182 */ 2183 ndirty = ndisks = syncid = 0; 2184 for (n = 0; n < sc->sc_ndisks; n++) { 2185 disk = &sc->sc_disks[n]; 2186 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 2187 continue; 2188 if ((disk->d_flags & G_RAID3_DISK_FLAG_DIRTY) != 0) 2189 ndirty++; 2190 if (disk->d_sync.ds_syncid > syncid) { 2191 syncid = disk->d_sync.ds_syncid; 2192 ndisks = 0; 2193 } else if (disk->d_sync.ds_syncid < syncid) { 2194 continue; 2195 } 2196 if ((disk->d_flags & 2197 G_RAID3_DISK_FLAG_SYNCHRONIZING) != 0) { 2198 continue; 2199 } 2200 ndisks++; 2201 } 2202 /* 2203 * Do we have enough valid components? 2204 */ 2205 if (ndisks + 1 < sc->sc_ndisks) { 2206 G_RAID3_DEBUG(0, 2207 "Device %s is broken, too few valid components.", 2208 sc->sc_name); 2209 sc->sc_flags |= G_RAID3_DEVICE_FLAG_DESTROY; 2210 return; 2211 } 2212 /* 2213 * If there is one DIRTY component and all disks are present, 2214 * mark it for synchronization. If there is more than one DIRTY 2215 * component, mark parity component for synchronization. 2216 */ 2217 if (ndisks == sc->sc_ndisks && ndirty == 1) { 2218 for (n = 0; n < sc->sc_ndisks; n++) { 2219 disk = &sc->sc_disks[n]; 2220 if ((disk->d_flags & 2221 G_RAID3_DISK_FLAG_DIRTY) == 0) { 2222 continue; 2223 } 2224 disk->d_flags |= 2225 G_RAID3_DISK_FLAG_SYNCHRONIZING; 2226 } 2227 } else if (ndisks == sc->sc_ndisks && ndirty > 1) { 2228 disk = &sc->sc_disks[sc->sc_ndisks - 1]; 2229 disk->d_flags |= G_RAID3_DISK_FLAG_SYNCHRONIZING; 2230 } 2231 2232 sc->sc_syncid = syncid; 2233 if (force) { 2234 /* Remember to bump syncid on first write. */ 2235 sc->sc_bump_id |= G_RAID3_BUMP_SYNCID; 2236 } 2237 if (ndisks == sc->sc_ndisks) 2238 state = G_RAID3_DEVICE_STATE_COMPLETE; 2239 else /* if (ndisks == sc->sc_ndisks - 1) */ 2240 state = G_RAID3_DEVICE_STATE_DEGRADED; 2241 G_RAID3_DEBUG(1, "Device %s state changed from %s to %s.", 2242 sc->sc_name, g_raid3_device_state2str(sc->sc_state), 2243 g_raid3_device_state2str(state)); 2244 sc->sc_state = state; 2245 for (n = 0; n < sc->sc_ndisks; n++) { 2246 disk = &sc->sc_disks[n]; 2247 if (disk->d_state == G_RAID3_DISK_STATE_NODISK) 2248 continue; 2249 state = g_raid3_determine_state(disk); 2250 g_raid3_event_send(disk, state, G_RAID3_EVENT_DONTWAIT); 2251 if (state == G_RAID3_DISK_STATE_STALE) 2252 sc->sc_bump_id |= G_RAID3_BUMP_SYNCID; 2253 } 2254 break; 2255 } 2256 case G_RAID3_DEVICE_STATE_DEGRADED: 2257 /* 2258 * Genid need to be bumped immediately, so do it here. 2259 */ 2260 if ((sc->sc_bump_id & G_RAID3_BUMP_GENID) != 0) { 2261 sc->sc_bump_id &= ~G_RAID3_BUMP_GENID; 2262 g_raid3_bump_genid(sc); 2263 } 2264 2265 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_NEW) > 0) 2266 return; 2267 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) < 2268 sc->sc_ndisks - 1) { 2269 if (sc->sc_provider != NULL) 2270 g_raid3_destroy_provider(sc); 2271 sc->sc_flags |= G_RAID3_DEVICE_FLAG_DESTROY; 2272 return; 2273 } 2274 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) == 2275 sc->sc_ndisks) { 2276 state = G_RAID3_DEVICE_STATE_COMPLETE; 2277 G_RAID3_DEBUG(1, 2278 "Device %s state changed from %s to %s.", 2279 sc->sc_name, g_raid3_device_state2str(sc->sc_state), 2280 g_raid3_device_state2str(state)); 2281 sc->sc_state = state; 2282 } 2283 if (sc->sc_provider == NULL) 2284 g_raid3_launch_provider(sc); 2285 break; 2286 case G_RAID3_DEVICE_STATE_COMPLETE: 2287 /* 2288 * Genid need to be bumped immediately, so do it here. 2289 */ 2290 if ((sc->sc_bump_id & G_RAID3_BUMP_GENID) != 0) { 2291 sc->sc_bump_id &= ~G_RAID3_BUMP_GENID; 2292 g_raid3_bump_genid(sc); 2293 } 2294 2295 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_NEW) > 0) 2296 return; 2297 KASSERT(g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) >= 2298 sc->sc_ndisks - 1, 2299 ("Too few ACTIVE components in COMPLETE state (device %s).", 2300 sc->sc_name)); 2301 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) == 2302 sc->sc_ndisks - 1) { 2303 state = G_RAID3_DEVICE_STATE_DEGRADED; 2304 G_RAID3_DEBUG(1, 2305 "Device %s state changed from %s to %s.", 2306 sc->sc_name, g_raid3_device_state2str(sc->sc_state), 2307 g_raid3_device_state2str(state)); 2308 sc->sc_state = state; 2309 } 2310 if (sc->sc_provider == NULL) 2311 g_raid3_launch_provider(sc); 2312 break; 2313 default: 2314 KASSERT(1 == 0, ("Wrong device state (%s, %s).", sc->sc_name, 2315 g_raid3_device_state2str(sc->sc_state))); 2316 break; 2317 } 2318 } 2319 2320 /* 2321 * Update disk state and device state if needed. 2322 */ 2323 #define DISK_STATE_CHANGED() G_RAID3_DEBUG(1, \ 2324 "Disk %s state changed from %s to %s (device %s).", \ 2325 g_raid3_get_diskname(disk), \ 2326 g_raid3_disk_state2str(disk->d_state), \ 2327 g_raid3_disk_state2str(state), sc->sc_name) 2328 static int 2329 g_raid3_update_disk(struct g_raid3_disk *disk, u_int state) 2330 { 2331 struct g_raid3_softc *sc; 2332 2333 g_topology_assert(); 2334 2335 sc = disk->d_softc; 2336 again: 2337 G_RAID3_DEBUG(3, "Changing disk %s state from %s to %s.", 2338 g_raid3_get_diskname(disk), g_raid3_disk_state2str(disk->d_state), 2339 g_raid3_disk_state2str(state)); 2340 switch (state) { 2341 case G_RAID3_DISK_STATE_NEW: 2342 /* 2343 * Possible scenarios: 2344 * 1. New disk arrive. 2345 */ 2346 /* Previous state should be NONE. */ 2347 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NONE, 2348 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 2349 g_raid3_disk_state2str(disk->d_state))); 2350 DISK_STATE_CHANGED(); 2351 2352 disk->d_state = state; 2353 G_RAID3_DEBUG(0, "Device %s: provider %s detected.", 2354 sc->sc_name, g_raid3_get_diskname(disk)); 2355 if (sc->sc_state == G_RAID3_DEVICE_STATE_STARTING) 2356 break; 2357 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2358 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE, 2359 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2360 g_raid3_device_state2str(sc->sc_state), 2361 g_raid3_get_diskname(disk), 2362 g_raid3_disk_state2str(disk->d_state))); 2363 state = g_raid3_determine_state(disk); 2364 if (state != G_RAID3_DISK_STATE_NONE) 2365 goto again; 2366 break; 2367 case G_RAID3_DISK_STATE_ACTIVE: 2368 /* 2369 * Possible scenarios: 2370 * 1. New disk does not need synchronization. 2371 * 2. Synchronization process finished successfully. 2372 */ 2373 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2374 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE, 2375 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2376 g_raid3_device_state2str(sc->sc_state), 2377 g_raid3_get_diskname(disk), 2378 g_raid3_disk_state2str(disk->d_state))); 2379 /* Previous state should be NEW or SYNCHRONIZING. */ 2380 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NEW || 2381 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING, 2382 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 2383 g_raid3_disk_state2str(disk->d_state))); 2384 DISK_STATE_CHANGED(); 2385 2386 if (disk->d_state == G_RAID3_DISK_STATE_NEW) 2387 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 2388 else if (disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 2389 disk->d_flags &= ~G_RAID3_DISK_FLAG_SYNCHRONIZING; 2390 disk->d_flags &= ~G_RAID3_DISK_FLAG_FORCE_SYNC; 2391 g_raid3_sync_stop(sc, 0); 2392 } 2393 disk->d_state = state; 2394 disk->d_sync.ds_offset = 0; 2395 disk->d_sync.ds_offset_done = 0; 2396 g_raid3_update_access(disk); 2397 g_raid3_update_metadata(disk); 2398 G_RAID3_DEBUG(0, "Device %s: provider %s activated.", 2399 sc->sc_name, g_raid3_get_diskname(disk)); 2400 break; 2401 case G_RAID3_DISK_STATE_STALE: 2402 /* 2403 * Possible scenarios: 2404 * 1. Stale disk was connected. 2405 */ 2406 /* Previous state should be NEW. */ 2407 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NEW, 2408 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 2409 g_raid3_disk_state2str(disk->d_state))); 2410 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2411 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE, 2412 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2413 g_raid3_device_state2str(sc->sc_state), 2414 g_raid3_get_diskname(disk), 2415 g_raid3_disk_state2str(disk->d_state))); 2416 /* 2417 * STALE state is only possible if device is marked 2418 * NOAUTOSYNC. 2419 */ 2420 KASSERT((sc->sc_flags & G_RAID3_DEVICE_FLAG_NOAUTOSYNC) != 0, 2421 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2422 g_raid3_device_state2str(sc->sc_state), 2423 g_raid3_get_diskname(disk), 2424 g_raid3_disk_state2str(disk->d_state))); 2425 DISK_STATE_CHANGED(); 2426 2427 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 2428 disk->d_state = state; 2429 g_raid3_update_metadata(disk); 2430 G_RAID3_DEBUG(0, "Device %s: provider %s is stale.", 2431 sc->sc_name, g_raid3_get_diskname(disk)); 2432 break; 2433 case G_RAID3_DISK_STATE_SYNCHRONIZING: 2434 /* 2435 * Possible scenarios: 2436 * 1. Disk which needs synchronization was connected. 2437 */ 2438 /* Previous state should be NEW. */ 2439 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NEW, 2440 ("Wrong disk state (%s, %s).", g_raid3_get_diskname(disk), 2441 g_raid3_disk_state2str(disk->d_state))); 2442 KASSERT(sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2443 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE, 2444 ("Wrong device state (%s, %s, %s, %s).", sc->sc_name, 2445 g_raid3_device_state2str(sc->sc_state), 2446 g_raid3_get_diskname(disk), 2447 g_raid3_disk_state2str(disk->d_state))); 2448 DISK_STATE_CHANGED(); 2449 2450 if (disk->d_state == G_RAID3_DISK_STATE_NEW) 2451 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 2452 disk->d_state = state; 2453 if (sc->sc_provider != NULL) { 2454 g_raid3_sync_start(sc); 2455 g_raid3_update_metadata(disk); 2456 } 2457 break; 2458 case G_RAID3_DISK_STATE_DISCONNECTED: 2459 /* 2460 * Possible scenarios: 2461 * 1. Device wasn't running yet, but disk disappear. 2462 * 2. Disk was active and disapppear. 2463 * 3. Disk disappear during synchronization process. 2464 */ 2465 if (sc->sc_state == G_RAID3_DEVICE_STATE_DEGRADED || 2466 sc->sc_state == G_RAID3_DEVICE_STATE_COMPLETE) { 2467 /* 2468 * Previous state should be ACTIVE, STALE or 2469 * SYNCHRONIZING. 2470 */ 2471 KASSERT(disk->d_state == G_RAID3_DISK_STATE_ACTIVE || 2472 disk->d_state == G_RAID3_DISK_STATE_STALE || 2473 disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING, 2474 ("Wrong disk state (%s, %s).", 2475 g_raid3_get_diskname(disk), 2476 g_raid3_disk_state2str(disk->d_state))); 2477 } else if (sc->sc_state == G_RAID3_DEVICE_STATE_STARTING) { 2478 /* Previous state should be NEW. */ 2479 KASSERT(disk->d_state == G_RAID3_DISK_STATE_NEW, 2480 ("Wrong disk state (%s, %s).", 2481 g_raid3_get_diskname(disk), 2482 g_raid3_disk_state2str(disk->d_state))); 2483 /* 2484 * Reset bumping syncid if disk disappeared in STARTING 2485 * state. 2486 */ 2487 if ((sc->sc_bump_id & G_RAID3_BUMP_SYNCID) != 0) 2488 sc->sc_bump_id &= ~G_RAID3_BUMP_SYNCID; 2489 #ifdef INVARIANTS 2490 } else { 2491 KASSERT(1 == 0, ("Wrong device state (%s, %s, %s, %s).", 2492 sc->sc_name, 2493 g_raid3_device_state2str(sc->sc_state), 2494 g_raid3_get_diskname(disk), 2495 g_raid3_disk_state2str(disk->d_state))); 2496 #endif 2497 } 2498 DISK_STATE_CHANGED(); 2499 G_RAID3_DEBUG(0, "Device %s: provider %s disconnected.", 2500 sc->sc_name, g_raid3_get_diskname(disk)); 2501 2502 g_raid3_destroy_disk(disk); 2503 break; 2504 default: 2505 KASSERT(1 == 0, ("Unknown state (%u).", state)); 2506 break; 2507 } 2508 return (0); 2509 } 2510 #undef DISK_STATE_CHANGED 2511 2512 int 2513 g_raid3_read_metadata(struct g_consumer *cp, struct g_raid3_metadata *md) 2514 { 2515 struct g_provider *pp; 2516 u_char *buf; 2517 int error; 2518 2519 g_topology_assert(); 2520 2521 error = g_access(cp, 1, 0, 0); 2522 if (error != 0) 2523 return (error); 2524 pp = cp->provider; 2525 g_topology_unlock(); 2526 /* Metadata are stored on last sector. */ 2527 buf = g_read_data(cp, pp->mediasize - pp->sectorsize, pp->sectorsize, 2528 &error); 2529 g_topology_lock(); 2530 g_access(cp, -1, 0, 0); 2531 if (error != 0) { 2532 G_RAID3_DEBUG(1, "Cannot read metadata from %s (error=%d).", 2533 cp->provider->name, error); 2534 if (buf != NULL) 2535 g_free(buf); 2536 return (error); 2537 } 2538 2539 /* Decode metadata. */ 2540 error = raid3_metadata_decode(buf, md); 2541 g_free(buf); 2542 if (strcmp(md->md_magic, G_RAID3_MAGIC) != 0) 2543 return (EINVAL); 2544 if (md->md_version > G_RAID3_VERSION) { 2545 G_RAID3_DEBUG(0, 2546 "Kernel module is too old to handle metadata from %s.", 2547 cp->provider->name); 2548 return (EINVAL); 2549 } 2550 if (error != 0) { 2551 G_RAID3_DEBUG(1, "MD5 metadata hash mismatch for provider %s.", 2552 cp->provider->name); 2553 return (error); 2554 } 2555 2556 return (0); 2557 } 2558 2559 static int 2560 g_raid3_check_metadata(struct g_raid3_softc *sc, struct g_provider *pp, 2561 struct g_raid3_metadata *md) 2562 { 2563 2564 if (md->md_no >= sc->sc_ndisks) { 2565 G_RAID3_DEBUG(1, "Invalid disk %s number (no=%u), skipping.", 2566 pp->name, md->md_no); 2567 return (EINVAL); 2568 } 2569 if (sc->sc_disks[md->md_no].d_state != G_RAID3_DISK_STATE_NODISK) { 2570 G_RAID3_DEBUG(1, "Disk %s (no=%u) already exists, skipping.", 2571 pp->name, md->md_no); 2572 return (EEXIST); 2573 } 2574 if (md->md_all != sc->sc_ndisks) { 2575 G_RAID3_DEBUG(1, 2576 "Invalid '%s' field on disk %s (device %s), skipping.", 2577 "md_all", pp->name, sc->sc_name); 2578 return (EINVAL); 2579 } 2580 if (md->md_mediasize != sc->sc_mediasize) { 2581 G_RAID3_DEBUG(1, 2582 "Invalid '%s' field on disk %s (device %s), skipping.", 2583 "md_mediasize", pp->name, sc->sc_name); 2584 return (EINVAL); 2585 } 2586 if ((md->md_mediasize % (sc->sc_ndisks - 1)) != 0) { 2587 G_RAID3_DEBUG(1, 2588 "Invalid '%s' field on disk %s (device %s), skipping.", 2589 "md_mediasize", pp->name, sc->sc_name); 2590 return (EINVAL); 2591 } 2592 if ((sc->sc_mediasize / (sc->sc_ndisks - 1)) > pp->mediasize) { 2593 G_RAID3_DEBUG(1, 2594 "Invalid size of disk %s (device %s), skipping.", pp->name, 2595 sc->sc_name); 2596 return (EINVAL); 2597 } 2598 if ((md->md_sectorsize / pp->sectorsize) < sc->sc_ndisks - 1) { 2599 G_RAID3_DEBUG(1, 2600 "Invalid '%s' field on disk %s (device %s), skipping.", 2601 "md_sectorsize", pp->name, sc->sc_name); 2602 return (EINVAL); 2603 } 2604 if (md->md_sectorsize != sc->sc_sectorsize) { 2605 G_RAID3_DEBUG(1, 2606 "Invalid '%s' field on disk %s (device %s), skipping.", 2607 "md_sectorsize", pp->name, sc->sc_name); 2608 return (EINVAL); 2609 } 2610 if ((sc->sc_sectorsize % pp->sectorsize) != 0) { 2611 G_RAID3_DEBUG(1, 2612 "Invalid sector size of disk %s (device %s), skipping.", 2613 pp->name, sc->sc_name); 2614 return (EINVAL); 2615 } 2616 if ((md->md_mflags & ~G_RAID3_DEVICE_FLAG_MASK) != 0) { 2617 G_RAID3_DEBUG(1, 2618 "Invalid device flags on disk %s (device %s), skipping.", 2619 pp->name, sc->sc_name); 2620 return (EINVAL); 2621 } 2622 if ((md->md_mflags & G_RAID3_DEVICE_FLAG_VERIFY) != 0 && 2623 (md->md_mflags & G_RAID3_DEVICE_FLAG_ROUND_ROBIN) != 0) { 2624 /* 2625 * VERIFY and ROUND-ROBIN options are mutally exclusive. 2626 */ 2627 G_RAID3_DEBUG(1, "Both VERIFY and ROUND-ROBIN flags exist on " 2628 "disk %s (device %s), skipping.", pp->name, sc->sc_name); 2629 return (EINVAL); 2630 } 2631 if ((md->md_dflags & ~G_RAID3_DISK_FLAG_MASK) != 0) { 2632 G_RAID3_DEBUG(1, 2633 "Invalid disk flags on disk %s (device %s), skipping.", 2634 pp->name, sc->sc_name); 2635 return (EINVAL); 2636 } 2637 return (0); 2638 } 2639 2640 int 2641 g_raid3_add_disk(struct g_raid3_softc *sc, struct g_provider *pp, 2642 struct g_raid3_metadata *md) 2643 { 2644 struct g_raid3_disk *disk; 2645 int error; 2646 2647 g_topology_assert(); 2648 G_RAID3_DEBUG(2, "Adding disk %s.", pp->name); 2649 2650 error = g_raid3_check_metadata(sc, pp, md); 2651 if (error != 0) 2652 return (error); 2653 if (sc->sc_state != G_RAID3_DEVICE_STATE_STARTING && 2654 md->md_genid < sc->sc_genid) { 2655 G_RAID3_DEBUG(0, "Component %s (device %s) broken, skipping.", 2656 pp->name, sc->sc_name); 2657 return (EINVAL); 2658 } 2659 disk = g_raid3_init_disk(sc, pp, md, &error); 2660 if (disk == NULL) 2661 return (error); 2662 error = g_raid3_event_send(disk, G_RAID3_DISK_STATE_NEW, 2663 G_RAID3_EVENT_WAIT); 2664 if (error != 0) 2665 return (error); 2666 if (md->md_version < G_RAID3_VERSION) { 2667 G_RAID3_DEBUG(0, "Upgrading metadata on %s (v%d->v%d).", 2668 pp->name, md->md_version, G_RAID3_VERSION); 2669 g_raid3_update_metadata(disk); 2670 } 2671 return (0); 2672 } 2673 2674 static int 2675 g_raid3_access(struct g_provider *pp, int acr, int acw, int ace) 2676 { 2677 struct g_raid3_softc *sc; 2678 struct g_raid3_disk *disk; 2679 int dcr, dcw, dce; 2680 u_int n; 2681 2682 g_topology_assert(); 2683 G_RAID3_DEBUG(2, "Access request for %s: r%dw%de%d.", pp->name, acr, 2684 acw, ace); 2685 2686 dcr = pp->acr + acr; 2687 dcw = pp->acw + acw; 2688 dce = pp->ace + ace; 2689 2690 sc = pp->geom->softc; 2691 if (sc == NULL || 2692 g_raid3_ndisks(sc, G_RAID3_DISK_STATE_ACTIVE) < sc->sc_ndisks - 1 || 2693 (sc->sc_flags & G_RAID3_DEVICE_FLAG_DESTROY) != 0) { 2694 if (acr <= 0 && acw <= 0 && ace <= 0) 2695 return (0); 2696 else 2697 return (ENXIO); 2698 } 2699 for (n = 0; n < sc->sc_ndisks; n++) { 2700 disk = &sc->sc_disks[n]; 2701 if (disk->d_state != G_RAID3_DISK_STATE_ACTIVE) 2702 continue; 2703 /* 2704 * Mark disk as dirty on open and unmark on close. 2705 */ 2706 if (pp->acw == 0 && dcw > 0) { 2707 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as dirty.", 2708 g_raid3_get_diskname(disk), sc->sc_name); 2709 disk->d_flags |= G_RAID3_DISK_FLAG_DIRTY; 2710 g_raid3_update_metadata(disk); 2711 } else if (pp->acw > 0 && dcw == 0) { 2712 G_RAID3_DEBUG(1, "Disk %s (device %s) marked as clean.", 2713 g_raid3_get_diskname(disk), sc->sc_name); 2714 disk->d_flags &= ~G_RAID3_DISK_FLAG_DIRTY; 2715 g_raid3_update_metadata(disk); 2716 } 2717 } 2718 return (0); 2719 } 2720 2721 static struct g_geom * 2722 g_raid3_create(struct g_class *mp, const struct g_raid3_metadata *md) 2723 { 2724 struct g_raid3_softc *sc; 2725 struct g_geom *gp; 2726 int error, timeout; 2727 u_int n; 2728 2729 g_topology_assert(); 2730 G_RAID3_DEBUG(1, "Creating device %s (id=%u).", md->md_name, md->md_id); 2731 2732 /* One disk is minimum. */ 2733 if (md->md_all < 1) 2734 return (NULL); 2735 /* 2736 * Action geom. 2737 */ 2738 gp = g_new_geomf(mp, "%s", md->md_name); 2739 sc = malloc(sizeof(*sc), M_RAID3, M_WAITOK | M_ZERO); 2740 sc->sc_disks = malloc(sizeof(struct g_raid3_disk) * md->md_all, M_RAID3, 2741 M_WAITOK | M_ZERO); 2742 gp->start = g_raid3_start; 2743 gp->orphan = g_raid3_orphan; 2744 gp->access = g_raid3_access; 2745 gp->dumpconf = g_raid3_dumpconf; 2746 2747 sc->sc_id = md->md_id; 2748 sc->sc_mediasize = md->md_mediasize; 2749 sc->sc_sectorsize = md->md_sectorsize; 2750 sc->sc_ndisks = md->md_all; 2751 sc->sc_round_robin = 0; 2752 sc->sc_flags = md->md_mflags; 2753 sc->sc_bump_id = 0; 2754 sc->sc_idle = 0; 2755 for (n = 0; n < sc->sc_ndisks; n++) { 2756 sc->sc_disks[n].d_softc = sc; 2757 sc->sc_disks[n].d_no = n; 2758 sc->sc_disks[n].d_state = G_RAID3_DISK_STATE_NODISK; 2759 } 2760 bioq_init(&sc->sc_queue); 2761 mtx_init(&sc->sc_queue_mtx, "graid3:queue", NULL, MTX_DEF); 2762 TAILQ_INIT(&sc->sc_events); 2763 mtx_init(&sc->sc_events_mtx, "graid3:events", NULL, MTX_DEF); 2764 callout_init(&sc->sc_callout, CALLOUT_MPSAFE); 2765 sc->sc_state = G_RAID3_DEVICE_STATE_STARTING; 2766 gp->softc = sc; 2767 sc->sc_geom = gp; 2768 sc->sc_provider = NULL; 2769 /* 2770 * Synchronization geom. 2771 */ 2772 gp = g_new_geomf(mp, "%s.sync", md->md_name); 2773 gp->softc = sc; 2774 gp->orphan = g_raid3_orphan; 2775 sc->sc_sync.ds_geom = gp; 2776 sc->sc_zone_64k = uma_zcreate("gr3:64k", 65536, NULL, NULL, NULL, NULL, 2777 UMA_ALIGN_PTR, 0); 2778 uma_zone_set_max(sc->sc_zone_64k, g_raid3_n64k); 2779 sc->sc_zone_16k = uma_zcreate("gr3:16k", 16384, NULL, NULL, NULL, NULL, 2780 UMA_ALIGN_PTR, 0); 2781 uma_zone_set_max(sc->sc_zone_64k, g_raid3_n16k); 2782 sc->sc_zone_4k = uma_zcreate("gr3:4k", 4096, NULL, NULL, NULL, NULL, 2783 UMA_ALIGN_PTR, 0); 2784 uma_zone_set_max(sc->sc_zone_4k, g_raid3_n4k); 2785 error = kthread_create(g_raid3_worker, sc, &sc->sc_worker, 0, 0, 2786 "g_raid3 %s", md->md_name); 2787 if (error != 0) { 2788 G_RAID3_DEBUG(1, "Cannot create kernel thread for %s.", 2789 sc->sc_name); 2790 uma_zdestroy(sc->sc_zone_64k); 2791 uma_zdestroy(sc->sc_zone_16k); 2792 uma_zdestroy(sc->sc_zone_4k); 2793 g_destroy_geom(sc->sc_sync.ds_geom); 2794 mtx_destroy(&sc->sc_events_mtx); 2795 mtx_destroy(&sc->sc_queue_mtx); 2796 g_destroy_geom(sc->sc_geom); 2797 free(sc->sc_disks, M_RAID3); 2798 free(sc, M_RAID3); 2799 return (NULL); 2800 } 2801 2802 G_RAID3_DEBUG(0, "Device %s created (id=%u).", sc->sc_name, sc->sc_id); 2803 2804 /* 2805 * Run timeout. 2806 */ 2807 timeout = atomic_load_acq_int(&g_raid3_timeout); 2808 callout_reset(&sc->sc_callout, timeout * hz, g_raid3_go, sc); 2809 return (sc->sc_geom); 2810 } 2811 2812 int 2813 g_raid3_destroy(struct g_raid3_softc *sc, boolean_t force) 2814 { 2815 struct g_provider *pp; 2816 2817 g_topology_assert(); 2818 2819 if (sc == NULL) 2820 return (ENXIO); 2821 pp = sc->sc_provider; 2822 if (pp != NULL && (pp->acr != 0 || pp->acw != 0 || pp->ace != 0)) { 2823 if (force) { 2824 G_RAID3_DEBUG(1, "Device %s is still open, so it " 2825 "can't be definitely removed.", pp->name); 2826 } else { 2827 G_RAID3_DEBUG(1, 2828 "Device %s is still open (r%dw%de%d).", pp->name, 2829 pp->acr, pp->acw, pp->ace); 2830 return (EBUSY); 2831 } 2832 } 2833 2834 sc->sc_flags |= G_RAID3_DEVICE_FLAG_DESTROY; 2835 sc->sc_flags |= G_RAID3_DEVICE_FLAG_WAIT; 2836 g_topology_unlock(); 2837 G_RAID3_DEBUG(4, "%s: Waking up %p.", __func__, sc); 2838 mtx_lock(&sc->sc_queue_mtx); 2839 wakeup(sc); 2840 wakeup(&sc->sc_queue); 2841 mtx_unlock(&sc->sc_queue_mtx); 2842 G_RAID3_DEBUG(4, "%s: Sleeping %p.", __func__, &sc->sc_worker); 2843 while (sc->sc_worker != NULL) 2844 tsleep(&sc->sc_worker, PRIBIO, "r3:destroy", hz / 5); 2845 G_RAID3_DEBUG(4, "%s: Woken up %p.", __func__, &sc->sc_worker); 2846 g_topology_lock(); 2847 g_raid3_destroy_device(sc); 2848 free(sc->sc_disks, M_RAID3); 2849 free(sc, M_RAID3); 2850 return (0); 2851 } 2852 2853 static void 2854 g_raid3_taste_orphan(struct g_consumer *cp) 2855 { 2856 2857 KASSERT(1 == 0, ("%s called while tasting %s.", __func__, 2858 cp->provider->name)); 2859 } 2860 2861 static struct g_geom * 2862 g_raid3_taste(struct g_class *mp, struct g_provider *pp, int flags __unused) 2863 { 2864 struct g_raid3_metadata md; 2865 struct g_raid3_softc *sc; 2866 struct g_consumer *cp; 2867 struct g_geom *gp; 2868 int error; 2869 2870 g_topology_assert(); 2871 g_trace(G_T_TOPOLOGY, "%s(%s, %s)", __func__, mp->name, pp->name); 2872 G_RAID3_DEBUG(2, "Tasting %s.", pp->name); 2873 2874 gp = g_new_geomf(mp, "raid3:taste"); 2875 /* This orphan function should be never called. */ 2876 gp->orphan = g_raid3_taste_orphan; 2877 cp = g_new_consumer(gp); 2878 g_attach(cp, pp); 2879 error = g_raid3_read_metadata(cp, &md); 2880 g_detach(cp); 2881 g_destroy_consumer(cp); 2882 g_destroy_geom(gp); 2883 if (error != 0) 2884 return (NULL); 2885 gp = NULL; 2886 2887 if (md.md_provider[0] != '\0' && strcmp(md.md_provider, pp->name) != 0) 2888 return (NULL); 2889 if (g_raid3_debug >= 2) 2890 raid3_metadata_dump(&md); 2891 2892 /* 2893 * Let's check if device already exists. 2894 */ 2895 sc = NULL; 2896 LIST_FOREACH(gp, &mp->geom, geom) { 2897 sc = gp->softc; 2898 if (sc == NULL) 2899 continue; 2900 if (sc->sc_sync.ds_geom == gp) 2901 continue; 2902 if (strcmp(md.md_name, sc->sc_name) != 0) 2903 continue; 2904 if (md.md_id != sc->sc_id) { 2905 G_RAID3_DEBUG(0, "Device %s already configured.", 2906 sc->sc_name); 2907 return (NULL); 2908 } 2909 break; 2910 } 2911 if (gp == NULL) { 2912 gp = g_raid3_create(mp, &md); 2913 if (gp == NULL) { 2914 G_RAID3_DEBUG(0, "Cannot create device %s.", 2915 md.md_name); 2916 return (NULL); 2917 } 2918 sc = gp->softc; 2919 } 2920 G_RAID3_DEBUG(1, "Adding disk %s to %s.", pp->name, gp->name); 2921 error = g_raid3_add_disk(sc, pp, &md); 2922 if (error != 0) { 2923 G_RAID3_DEBUG(0, "Cannot add disk %s to %s (error=%d).", 2924 pp->name, gp->name, error); 2925 if (g_raid3_ndisks(sc, G_RAID3_DISK_STATE_NODISK) == 2926 sc->sc_ndisks) { 2927 g_raid3_destroy(sc, 1); 2928 } 2929 return (NULL); 2930 } 2931 return (gp); 2932 } 2933 2934 static int 2935 g_raid3_destroy_geom(struct gctl_req *req __unused, struct g_class *mp __unused, 2936 struct g_geom *gp) 2937 { 2938 2939 return (g_raid3_destroy(gp->softc, 0)); 2940 } 2941 2942 static void 2943 g_raid3_dumpconf(struct sbuf *sb, const char *indent, struct g_geom *gp, 2944 struct g_consumer *cp, struct g_provider *pp) 2945 { 2946 struct g_raid3_softc *sc; 2947 2948 g_topology_assert(); 2949 2950 sc = gp->softc; 2951 if (sc == NULL) 2952 return; 2953 /* Skip synchronization geom. */ 2954 if (gp == sc->sc_sync.ds_geom) 2955 return; 2956 if (pp != NULL) { 2957 /* Nothing here. */ 2958 } else if (cp != NULL) { 2959 struct g_raid3_disk *disk; 2960 2961 disk = cp->private; 2962 if (disk == NULL) 2963 return; 2964 sbuf_printf(sb, "%s<Type>", indent); 2965 if (disk->d_no == sc->sc_ndisks - 1) 2966 sbuf_printf(sb, "PARITY"); 2967 else 2968 sbuf_printf(sb, "DATA"); 2969 sbuf_printf(sb, "</Type>\n"); 2970 sbuf_printf(sb, "%s<Number>%u</Number>\n", indent, 2971 (u_int)disk->d_no); 2972 if (disk->d_state == G_RAID3_DISK_STATE_SYNCHRONIZING) { 2973 sbuf_printf(sb, "%s<Synchronized>", indent); 2974 if (disk->d_sync.ds_offset_done == 0) 2975 sbuf_printf(sb, "0%%"); 2976 else { 2977 sbuf_printf(sb, "%u%%", 2978 (u_int)((disk->d_sync.ds_offset_done * 100) / 2979 (sc->sc_mediasize / (sc->sc_ndisks - 1)))); 2980 } 2981 sbuf_printf(sb, "</Synchronized>\n"); 2982 } 2983 sbuf_printf(sb, "%s<SyncID>%u</SyncID>\n", indent, 2984 disk->d_sync.ds_syncid); 2985 sbuf_printf(sb, "%s<GenID>%u</GenID>\n", indent, disk->d_genid); 2986 sbuf_printf(sb, "%s<Flags>", indent); 2987 if (disk->d_flags == 0) 2988 sbuf_printf(sb, "NONE"); 2989 else { 2990 int first = 1; 2991 2992 #define ADD_FLAG(flag, name) do { \ 2993 if ((disk->d_flags & (flag)) != 0) { \ 2994 if (!first) \ 2995 sbuf_printf(sb, ", "); \ 2996 else \ 2997 first = 0; \ 2998 sbuf_printf(sb, name); \ 2999 } \ 3000 } while (0) 3001 ADD_FLAG(G_RAID3_DISK_FLAG_DIRTY, "DIRTY"); 3002 ADD_FLAG(G_RAID3_DISK_FLAG_HARDCODED, "HARDCODED"); 3003 ADD_FLAG(G_RAID3_DISK_FLAG_SYNCHRONIZING, 3004 "SYNCHRONIZING"); 3005 ADD_FLAG(G_RAID3_DISK_FLAG_FORCE_SYNC, "FORCE_SYNC"); 3006 #undef ADD_FLAG 3007 } 3008 sbuf_printf(sb, "</Flags>\n"); 3009 sbuf_printf(sb, "%s<State>%s</State>\n", indent, 3010 g_raid3_disk_state2str(disk->d_state)); 3011 } else { 3012 sbuf_printf(sb, "%s<ID>%u</ID>\n", indent, (u_int)sc->sc_id); 3013 sbuf_printf(sb, "%s<SyncID>%u</SyncID>\n", indent, sc->sc_syncid); 3014 sbuf_printf(sb, "%s<GenID>%u</GenID>\n", indent, sc->sc_genid); 3015 sbuf_printf(sb, "%s<Flags>", indent); 3016 if (sc->sc_flags == 0) 3017 sbuf_printf(sb, "NONE"); 3018 else { 3019 int first = 1; 3020 3021 #define ADD_FLAG(flag, name) do { \ 3022 if ((sc->sc_flags & (flag)) != 0) { \ 3023 if (!first) \ 3024 sbuf_printf(sb, ", "); \ 3025 else \ 3026 first = 0; \ 3027 sbuf_printf(sb, name); \ 3028 } \ 3029 } while (0) 3030 ADD_FLAG(G_RAID3_DEVICE_FLAG_NOAUTOSYNC, "NOAUTOSYNC"); 3031 ADD_FLAG(G_RAID3_DEVICE_FLAG_ROUND_ROBIN, 3032 "ROUND-ROBIN"); 3033 ADD_FLAG(G_RAID3_DEVICE_FLAG_VERIFY, "VERIFY"); 3034 #undef ADD_FLAG 3035 } 3036 sbuf_printf(sb, "</Flags>\n"); 3037 sbuf_printf(sb, "%s<Components>%u</Components>\n", indent, 3038 sc->sc_ndisks); 3039 sbuf_printf(sb, "%s<State>%s</State>\n", indent, 3040 g_raid3_device_state2str(sc->sc_state)); 3041 } 3042 } 3043 3044 static void 3045 g_raid3_shutdown(void *arg, int howto) 3046 { 3047 struct g_class *mp; 3048 struct g_geom *gp, *gp2; 3049 3050 mp = arg; 3051 DROP_GIANT(); 3052 g_topology_lock(); 3053 LIST_FOREACH_SAFE(gp, &mp->geom, geom, gp2) { 3054 if (gp->softc == NULL) 3055 continue; 3056 g_raid3_destroy(gp->softc, 1); 3057 } 3058 g_topology_unlock(); 3059 PICKUP_GIANT(); 3060 #if 0 3061 tsleep(&gp, PRIBIO, "r3:shutdown", hz * 20); 3062 #endif 3063 } 3064 3065 static void 3066 g_raid3_init(struct g_class *mp) 3067 { 3068 3069 g_raid3_ehtag = EVENTHANDLER_REGISTER(shutdown_post_sync, 3070 g_raid3_shutdown, mp, SHUTDOWN_PRI_FIRST); 3071 if (g_raid3_ehtag == NULL) 3072 G_RAID3_DEBUG(0, "Warning! Cannot register shutdown event."); 3073 } 3074 3075 static void 3076 g_raid3_fini(struct g_class *mp) 3077 { 3078 3079 if (g_raid3_ehtag == NULL) 3080 return; 3081 EVENTHANDLER_DEREGISTER(shutdown_post_sync, g_raid3_ehtag); 3082 } 3083 3084 DECLARE_GEOM_CLASS(g_raid3_class, g_raid3); 3085