rf_netbsdkintf.c revision 1.230.4.5
1 /* $NetBSD: rf_netbsdkintf.c,v 1.230.4.5 2007/11/14 19:04:35 joerg Exp $ */ 2 /*- 3 * Copyright (c) 1996, 1997, 1998 The NetBSD Foundation, Inc. 4 * All rights reserved. 5 * 6 * This code is derived from software contributed to The NetBSD Foundation 7 * by Greg Oster; Jason R. Thorpe. 8 * 9 * Redistribution and use in source and binary forms, with or without 10 * modification, are permitted provided that the following conditions 11 * are met: 12 * 1. Redistributions of source code must retain the above copyright 13 * notice, this list of conditions and the following disclaimer. 14 * 2. Redistributions in binary form must reproduce the above copyright 15 * notice, this list of conditions and the following disclaimer in the 16 * documentation and/or other materials provided with the distribution. 17 * 3. All advertising materials mentioning features or use of this software 18 * must display the following acknowledgement: 19 * This product includes software developed by the NetBSD 20 * Foundation, Inc. and its contributors. 21 * 4. Neither the name of The NetBSD Foundation nor the names of its 22 * contributors may be used to endorse or promote products derived 23 * from this software without specific prior written permission. 24 * 25 * THIS SOFTWARE IS PROVIDED BY THE NETBSD FOUNDATION, INC. AND CONTRIBUTORS 26 * ``AS IS'' AND ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED 27 * TO, THE IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR 28 * PURPOSE ARE DISCLAIMED. IN NO EVENT SHALL THE FOUNDATION OR CONTRIBUTORS 29 * BE LIABLE FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR 30 * CONSEQUENTIAL DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF 31 * SUBSTITUTE GOODS OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS 32 * INTERRUPTION) HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN 33 * CONTRACT, STRICT LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) 34 * ARISING IN ANY WAY OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE 35 * POSSIBILITY OF SUCH DAMAGE. 36 */ 37 38 /* 39 * Copyright (c) 1990, 1993 40 * The Regents of the University of California. All rights reserved. 41 * 42 * This code is derived from software contributed to Berkeley by 43 * the Systems Programming Group of the University of Utah Computer 44 * Science Department. 45 * 46 * Redistribution and use in source and binary forms, with or without 47 * modification, are permitted provided that the following conditions 48 * are met: 49 * 1. Redistributions of source code must retain the above copyright 50 * notice, this list of conditions and the following disclaimer. 51 * 2. Redistributions in binary form must reproduce the above copyright 52 * notice, this list of conditions and the following disclaimer in the 53 * documentation and/or other materials provided with the distribution. 54 * 3. Neither the name of the University nor the names of its contributors 55 * may be used to endorse or promote products derived from this software 56 * without specific prior written permission. 57 * 58 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 59 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 60 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 61 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 62 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 63 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 64 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 65 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 66 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 67 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 68 * SUCH DAMAGE. 69 * 70 * from: Utah $Hdr: cd.c 1.6 90/11/28$ 71 * 72 * @(#)cd.c 8.2 (Berkeley) 11/16/93 73 */ 74 75 /* 76 * Copyright (c) 1988 University of Utah. 77 * 78 * This code is derived from software contributed to Berkeley by 79 * the Systems Programming Group of the University of Utah Computer 80 * Science Department. 81 * 82 * Redistribution and use in source and binary forms, with or without 83 * modification, are permitted provided that the following conditions 84 * are met: 85 * 1. Redistributions of source code must retain the above copyright 86 * notice, this list of conditions and the following disclaimer. 87 * 2. Redistributions in binary form must reproduce the above copyright 88 * notice, this list of conditions and the following disclaimer in the 89 * documentation and/or other materials provided with the distribution. 90 * 3. All advertising materials mentioning features or use of this software 91 * must display the following acknowledgement: 92 * This product includes software developed by the University of 93 * California, Berkeley and its contributors. 94 * 4. Neither the name of the University nor the names of its contributors 95 * may be used to endorse or promote products derived from this software 96 * without specific prior written permission. 97 * 98 * THIS SOFTWARE IS PROVIDED BY THE REGENTS AND CONTRIBUTORS ``AS IS'' AND 99 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE 100 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE 101 * ARE DISCLAIMED. IN NO EVENT SHALL THE REGENTS OR CONTRIBUTORS BE LIABLE 102 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL 103 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS 104 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION) 105 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT 106 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY 107 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF 108 * SUCH DAMAGE. 109 * 110 * from: Utah $Hdr: cd.c 1.6 90/11/28$ 111 * 112 * @(#)cd.c 8.2 (Berkeley) 11/16/93 113 */ 114 115 /* 116 * Copyright (c) 1995 Carnegie-Mellon University. 117 * All rights reserved. 118 * 119 * Authors: Mark Holland, Jim Zelenka 120 * 121 * Permission to use, copy, modify and distribute this software and 122 * its documentation is hereby granted, provided that both the copyright 123 * notice and this permission notice appear in all copies of the 124 * software, derivative works or modified versions, and any portions 125 * thereof, and that both notices appear in supporting documentation. 126 * 127 * CARNEGIE MELLON ALLOWS FREE USE OF THIS SOFTWARE IN ITS "AS IS" 128 * CONDITION. CARNEGIE MELLON DISCLAIMS ANY LIABILITY OF ANY KIND 129 * FOR ANY DAMAGES WHATSOEVER RESULTING FROM THE USE OF THIS SOFTWARE. 130 * 131 * Carnegie Mellon requests users of this software to return to 132 * 133 * Software Distribution Coordinator or Software.Distribution (at) CS.CMU.EDU 134 * School of Computer Science 135 * Carnegie Mellon University 136 * Pittsburgh PA 15213-3890 137 * 138 * any improvements or extensions that they make and grant Carnegie the 139 * rights to redistribute these changes. 140 */ 141 142 /*********************************************************** 143 * 144 * rf_kintf.c -- the kernel interface routines for RAIDframe 145 * 146 ***********************************************************/ 147 148 #include <sys/cdefs.h> 149 __KERNEL_RCSID(0, "$NetBSD: rf_netbsdkintf.c,v 1.230.4.5 2007/11/14 19:04:35 joerg Exp $"); 150 151 #include <sys/param.h> 152 #include <sys/errno.h> 153 #include <sys/pool.h> 154 #include <sys/proc.h> 155 #include <sys/queue.h> 156 #include <sys/disk.h> 157 #include <sys/device.h> 158 #include <sys/stat.h> 159 #include <sys/ioctl.h> 160 #include <sys/fcntl.h> 161 #include <sys/systm.h> 162 #include <sys/vnode.h> 163 #include <sys/disklabel.h> 164 #include <sys/conf.h> 165 #include <sys/lock.h> 166 #include <sys/buf.h> 167 #include <sys/bufq.h> 168 #include <sys/user.h> 169 #include <sys/reboot.h> 170 #include <sys/kauth.h> 171 172 #include <prop/proplib.h> 173 174 #include <dev/raidframe/raidframevar.h> 175 #include <dev/raidframe/raidframeio.h> 176 #include "raid.h" 177 #include "opt_raid_autoconfig.h" 178 #include "rf_raid.h" 179 #include "rf_copyback.h" 180 #include "rf_dag.h" 181 #include "rf_dagflags.h" 182 #include "rf_desc.h" 183 #include "rf_diskqueue.h" 184 #include "rf_etimer.h" 185 #include "rf_general.h" 186 #include "rf_kintf.h" 187 #include "rf_options.h" 188 #include "rf_driver.h" 189 #include "rf_parityscan.h" 190 #include "rf_threadstuff.h" 191 192 #ifdef DEBUG 193 int rf_kdebug_level = 0; 194 #define db1_printf(a) if (rf_kdebug_level > 0) printf a 195 #else /* DEBUG */ 196 #define db1_printf(a) { } 197 #endif /* DEBUG */ 198 199 static RF_Raid_t **raidPtrs; /* global raid device descriptors */ 200 201 RF_DECLARE_STATIC_MUTEX(rf_sparet_wait_mutex) 202 203 static RF_SparetWait_t *rf_sparet_wait_queue; /* requests to install a 204 * spare table */ 205 static RF_SparetWait_t *rf_sparet_resp_queue; /* responses from 206 * installation process */ 207 208 MALLOC_DEFINE(M_RAIDFRAME, "RAIDframe", "RAIDframe structures"); 209 210 /* prototypes */ 211 static void KernelWakeupFunc(struct buf *); 212 static void InitBP(struct buf *, struct vnode *, unsigned, 213 dev_t, RF_SectorNum_t, RF_SectorCount_t, void *, void (*) (struct buf *), 214 void *, int, struct proc *); 215 static void raidinit(RF_Raid_t *); 216 217 void raidattach(int); 218 static int raid_match(struct device *, struct cfdata *, void *); 219 static void raid_attach(struct device *, struct device *, void *); 220 static int raid_detach(struct device *, int); 221 222 dev_type_open(raidopen); 223 dev_type_close(raidclose); 224 dev_type_read(raidread); 225 dev_type_write(raidwrite); 226 dev_type_ioctl(raidioctl); 227 dev_type_strategy(raidstrategy); 228 dev_type_dump(raiddump); 229 dev_type_size(raidsize); 230 231 const struct bdevsw raid_bdevsw = { 232 raidopen, raidclose, raidstrategy, raidioctl, 233 raiddump, raidsize, D_DISK 234 }; 235 236 const struct cdevsw raid_cdevsw = { 237 raidopen, raidclose, raidread, raidwrite, raidioctl, 238 nostop, notty, nopoll, nommap, nokqfilter, D_DISK 239 }; 240 241 static struct dkdriver rf_dkdriver = { raidstrategy, minphys }; 242 243 /* XXX Not sure if the following should be replacing the raidPtrs above, 244 or if it should be used in conjunction with that... 245 */ 246 247 struct raid_softc { 248 struct device *sc_dev; 249 int sc_flags; /* flags */ 250 int sc_cflags; /* configuration flags */ 251 uint64_t sc_size; /* size of the raid device */ 252 char sc_xname[20]; /* XXX external name */ 253 struct disk sc_dkdev; /* generic disk device info */ 254 struct bufq_state *buf_queue; /* used for the device queue */ 255 }; 256 /* sc_flags */ 257 #define RAIDF_INITED 0x01 /* unit has been initialized */ 258 #define RAIDF_WLABEL 0x02 /* label area is writable */ 259 #define RAIDF_LABELLING 0x04 /* unit is currently being labelled */ 260 #define RAIDF_WANTED 0x40 /* someone is waiting to obtain a lock */ 261 #define RAIDF_LOCKED 0x80 /* unit is locked */ 262 263 #define raidunit(x) DISKUNIT(x) 264 int numraid = 0; 265 266 extern struct cfdriver raid_cd; 267 CFATTACH_DECL(raid, sizeof(struct raid_softc), 268 raid_match, raid_attach, raid_detach, NULL); 269 270 /* 271 * Allow RAIDOUTSTANDING number of simultaneous IO's to this RAID device. 272 * Be aware that large numbers can allow the driver to consume a lot of 273 * kernel memory, especially on writes, and in degraded mode reads. 274 * 275 * For example: with a stripe width of 64 blocks (32k) and 5 disks, 276 * a single 64K write will typically require 64K for the old data, 277 * 64K for the old parity, and 64K for the new parity, for a total 278 * of 192K (if the parity buffer is not re-used immediately). 279 * Even it if is used immediately, that's still 128K, which when multiplied 280 * by say 10 requests, is 1280K, *on top* of the 640K of incoming data. 281 * 282 * Now in degraded mode, for example, a 64K read on the above setup may 283 * require data reconstruction, which will require *all* of the 4 remaining 284 * disks to participate -- 4 * 32K/disk == 128K again. 285 */ 286 287 #ifndef RAIDOUTSTANDING 288 #define RAIDOUTSTANDING 6 289 #endif 290 291 #define RAIDLABELDEV(dev) \ 292 (MAKEDISKDEV(major((dev)), raidunit((dev)), RAW_PART)) 293 294 /* declared here, and made public, for the benefit of KVM stuff.. */ 295 struct raid_softc *raid_softc; 296 297 static void raidgetdefaultlabel(RF_Raid_t *, struct raid_softc *, 298 struct disklabel *); 299 static void raidgetdisklabel(dev_t); 300 static void raidmakedisklabel(struct raid_softc *); 301 302 static int raidlock(struct raid_softc *); 303 static void raidunlock(struct raid_softc *); 304 305 static void rf_markalldirty(RF_Raid_t *); 306 static void rf_set_properties(struct raid_softc *, RF_Raid_t *); 307 308 void rf_ReconThread(struct rf_recon_req *); 309 void rf_RewriteParityThread(RF_Raid_t *raidPtr); 310 void rf_CopybackThread(RF_Raid_t *raidPtr); 311 void rf_ReconstructInPlaceThread(struct rf_recon_req *); 312 int rf_autoconfig(struct device *self); 313 void rf_buildroothack(RF_ConfigSet_t *); 314 315 RF_AutoConfig_t *rf_find_raid_components(void); 316 RF_ConfigSet_t *rf_create_auto_sets(RF_AutoConfig_t *); 317 static int rf_does_it_fit(RF_ConfigSet_t *,RF_AutoConfig_t *); 318 static int rf_reasonable_label(RF_ComponentLabel_t *); 319 void rf_create_configuration(RF_AutoConfig_t *,RF_Config_t *, RF_Raid_t *); 320 int rf_set_autoconfig(RF_Raid_t *, int); 321 int rf_set_rootpartition(RF_Raid_t *, int); 322 void rf_release_all_vps(RF_ConfigSet_t *); 323 void rf_cleanup_config_set(RF_ConfigSet_t *); 324 int rf_have_enough_components(RF_ConfigSet_t *); 325 int rf_auto_config_set(RF_ConfigSet_t *, int *); 326 327 static int raidautoconfig = 0; /* Debugging, mostly. Set to 0 to not 328 allow autoconfig to take place. 329 Note that this is overridden by having 330 RAID_AUTOCONFIG as an option in the 331 kernel config file. */ 332 333 struct RF_Pools_s rf_pools; 334 335 void 336 raidattach(int num) 337 { 338 int raidID; 339 int i, rc; 340 341 #ifdef DEBUG 342 printf("raidattach: Asked for %d units\n", num); 343 #endif 344 345 if (num <= 0) { 346 #ifdef DIAGNOSTIC 347 panic("raidattach: count <= 0"); 348 #endif 349 return; 350 } 351 /* This is where all the initialization stuff gets done. */ 352 353 numraid = num; 354 355 /* Make some space for requested number of units... */ 356 357 RF_Malloc(raidPtrs, num * sizeof(RF_Raid_t *), (RF_Raid_t **)); 358 if (raidPtrs == NULL) { 359 panic("raidPtrs is NULL!!"); 360 } 361 362 rf_mutex_init(&rf_sparet_wait_mutex); 363 364 rf_sparet_wait_queue = rf_sparet_resp_queue = NULL; 365 366 for (i = 0; i < num; i++) 367 raidPtrs[i] = NULL; 368 rc = rf_BootRaidframe(); 369 if (rc == 0) 370 printf("Kernelized RAIDframe activated\n"); 371 else 372 panic("Serious error booting RAID!!"); 373 374 /* put together some datastructures like the CCD device does.. This 375 * lets us lock the device and what-not when it gets opened. */ 376 377 raid_softc = (struct raid_softc *) 378 malloc(num * sizeof(struct raid_softc), 379 M_RAIDFRAME, M_NOWAIT); 380 if (raid_softc == NULL) { 381 printf("WARNING: no memory for RAIDframe driver\n"); 382 return; 383 } 384 385 memset(raid_softc, 0, num * sizeof(struct raid_softc)); 386 387 for (raidID = 0; raidID < num; raidID++) { 388 bufq_alloc(&raid_softc[raidID].buf_queue, "fcfs", 0); 389 390 RF_Malloc(raidPtrs[raidID], sizeof(RF_Raid_t), 391 (RF_Raid_t *)); 392 if (raidPtrs[raidID] == NULL) { 393 printf("WARNING: raidPtrs[%d] is NULL\n", raidID); 394 numraid = raidID; 395 return; 396 } 397 } 398 399 if (config_cfattach_attach(raid_cd.cd_name, &raid_ca)) { 400 printf("config_cfattach_attach failed?\n"); 401 } 402 403 #ifdef RAID_AUTOCONFIG 404 raidautoconfig = 1; 405 #endif 406 407 /* 408 * Register a finalizer which will be used to auto-config RAID 409 * sets once all real hardware devices have been found. 410 */ 411 if (config_finalize_register(NULL, rf_autoconfig) != 0) 412 printf("WARNING: unable to register RAIDframe finalizer\n"); 413 } 414 415 int 416 rf_autoconfig(struct device *self) 417 { 418 RF_AutoConfig_t *ac_list; 419 RF_ConfigSet_t *config_sets; 420 421 if (raidautoconfig == 0) 422 return (0); 423 424 /* XXX This code can only be run once. */ 425 raidautoconfig = 0; 426 427 /* 1. locate all RAID components on the system */ 428 #ifdef DEBUG 429 printf("Searching for RAID components...\n"); 430 #endif 431 ac_list = rf_find_raid_components(); 432 433 /* 2. Sort them into their respective sets. */ 434 config_sets = rf_create_auto_sets(ac_list); 435 436 /* 437 * 3. Evaluate each set andconfigure the valid ones. 438 * This gets done in rf_buildroothack(). 439 */ 440 rf_buildroothack(config_sets); 441 442 return 1; 443 } 444 445 void 446 rf_buildroothack(RF_ConfigSet_t *config_sets) 447 { 448 RF_ConfigSet_t *cset; 449 RF_ConfigSet_t *next_cset; 450 int retcode; 451 int raidID; 452 int rootID; 453 int col; 454 int num_root; 455 char *devname; 456 457 rootID = 0; 458 num_root = 0; 459 cset = config_sets; 460 while(cset != NULL ) { 461 next_cset = cset->next; 462 if (rf_have_enough_components(cset) && 463 cset->ac->clabel->autoconfigure==1) { 464 retcode = rf_auto_config_set(cset,&raidID); 465 if (!retcode) { 466 #ifdef DEBUG 467 printf("raid%d: configured ok\n", raidID); 468 #endif 469 if (cset->rootable) { 470 rootID = raidID; 471 num_root++; 472 } 473 } else { 474 /* The autoconfig didn't work :( */ 475 #ifdef DEBUG 476 printf("Autoconfig failed with code %d for raid%d\n", retcode, raidID); 477 #endif 478 rf_release_all_vps(cset); 479 } 480 } else { 481 #ifdef DEBUG 482 printf("raid%d: not enough components\n", raidID); 483 #endif 484 /* we're not autoconfiguring this set... 485 release the associated resources */ 486 rf_release_all_vps(cset); 487 } 488 /* cleanup */ 489 rf_cleanup_config_set(cset); 490 cset = next_cset; 491 } 492 493 /* if the user has specified what the root device should be 494 then we don't touch booted_device or boothowto... */ 495 496 if (rootspec != NULL) 497 return; 498 499 /* we found something bootable... */ 500 501 if (num_root == 1) { 502 booted_device = raid_softc[rootID].sc_dev; 503 } else if (num_root > 1) { 504 505 /* 506 * Maybe the MD code can help. If it cannot, then 507 * setroot() will discover that we have no 508 * booted_device and will ask the user if nothing was 509 * hardwired in the kernel config file 510 */ 511 512 if (booted_device == NULL) 513 cpu_rootconf(); 514 if (booted_device == NULL) 515 return; 516 517 num_root = 0; 518 for (raidID = 0; raidID < numraid; raidID++) { 519 if (raidPtrs[raidID]->valid == 0) 520 continue; 521 522 if (raidPtrs[raidID]->root_partition == 0) 523 continue; 524 525 for (col = 0; col < raidPtrs[raidID]->numCol; col++) { 526 devname = raidPtrs[raidID]->Disks[col].devname; 527 devname += sizeof("/dev/") - 1; 528 if (strncmp(devname, booted_device->dv_xname, 529 strlen(booted_device->dv_xname)) != 0) 530 continue; 531 #ifdef DEBUG 532 printf("raid%d includes boot device %s\n", 533 raidID, devname); 534 #endif 535 num_root++; 536 rootID = raidID; 537 } 538 } 539 540 if (num_root == 1) { 541 booted_device = raid_softc[rootID].sc_dev; 542 } else { 543 /* we can't guess.. require the user to answer... */ 544 boothowto |= RB_ASKNAME; 545 } 546 } 547 } 548 549 550 int 551 raidsize(dev_t dev) 552 { 553 struct raid_softc *rs; 554 struct disklabel *lp; 555 int part, unit, omask, size; 556 557 unit = raidunit(dev); 558 if (unit >= numraid) 559 return (-1); 560 rs = &raid_softc[unit]; 561 562 if ((rs->sc_flags & RAIDF_INITED) == 0) 563 return (-1); 564 565 part = DISKPART(dev); 566 omask = rs->sc_dkdev.dk_openmask & (1 << part); 567 lp = rs->sc_dkdev.dk_label; 568 569 if (omask == 0 && raidopen(dev, 0, S_IFBLK, curlwp)) 570 return (-1); 571 572 if (lp->d_partitions[part].p_fstype != FS_SWAP) 573 size = -1; 574 else 575 size = lp->d_partitions[part].p_size * 576 (lp->d_secsize / DEV_BSIZE); 577 578 if (omask == 0 && raidclose(dev, 0, S_IFBLK, curlwp)) 579 return (-1); 580 581 return (size); 582 583 } 584 585 int 586 raiddump(dev_t dev, daddr_t blkno, void *va, size_t size) 587 { 588 int unit = raidunit(dev); 589 struct raid_softc *rs; 590 const struct bdevsw *bdev; 591 struct disklabel *lp; 592 RF_Raid_t *raidPtr; 593 daddr_t offset; 594 int part, c, sparecol, j, scol, dumpto; 595 int error = 0; 596 597 if (unit >= numraid) 598 return (ENXIO); 599 600 rs = &raid_softc[unit]; 601 raidPtr = raidPtrs[unit]; 602 603 if ((rs->sc_flags & RAIDF_INITED) == 0) 604 return ENXIO; 605 606 /* we only support dumping to RAID 1 sets */ 607 if (raidPtr->Layout.numDataCol != 1 || 608 raidPtr->Layout.numParityCol != 1) 609 return EINVAL; 610 611 612 if ((error = raidlock(rs)) != 0) 613 return error; 614 615 if (size % DEV_BSIZE != 0) { 616 error = EINVAL; 617 goto out; 618 } 619 620 if (blkno + size / DEV_BSIZE > rs->sc_size) { 621 printf("%s: blkno (%" PRIu64 ") + size / DEV_BSIZE (%zu) > " 622 "sc->sc_size (%" PRIu64 ")\n", __func__, blkno, 623 size / DEV_BSIZE, rs->sc_size); 624 error = EINVAL; 625 goto out; 626 } 627 628 part = DISKPART(dev); 629 lp = rs->sc_dkdev.dk_label; 630 offset = lp->d_partitions[part].p_offset + RF_PROTECTED_SECTORS; 631 632 /* figure out what device is alive.. */ 633 634 /* 635 Look for a component to dump to. The preference for the 636 component to dump to is as follows: 637 1) the master 638 2) a used_spare of the master 639 3) the slave 640 4) a used_spare of the slave 641 */ 642 643 dumpto = -1; 644 for (c = 0; c < raidPtr->numCol; c++) { 645 if (raidPtr->Disks[c].status == rf_ds_optimal) { 646 /* this might be the one */ 647 dumpto = c; 648 break; 649 } 650 } 651 652 /* 653 At this point we have possibly selected a live master or a 654 live slave. We now check to see if there is a spared 655 master (or a spared slave), if we didn't find a live master 656 or a live slave. 657 */ 658 659 for (c = 0; c < raidPtr->numSpare; c++) { 660 sparecol = raidPtr->numCol + c; 661 if (raidPtr->Disks[sparecol].status == rf_ds_used_spare) { 662 /* How about this one? */ 663 scol = -1; 664 for(j=0;j<raidPtr->numCol;j++) { 665 if (raidPtr->Disks[j].spareCol == sparecol) { 666 scol = j; 667 break; 668 } 669 } 670 if (scol == 0) { 671 /* 672 We must have found a spared master! 673 We'll take that over anything else 674 found so far. (We couldn't have 675 found a real master before, since 676 this is a used spare, and it's 677 saying that it's replacing the 678 master.) On reboot (with 679 autoconfiguration turned on) 680 sparecol will become the 1st 681 component (component0) of this set. 682 */ 683 dumpto = sparecol; 684 break; 685 } else if (scol != -1) { 686 /* 687 Must be a spared slave. We'll dump 688 to that if we havn't found anything 689 else so far. 690 */ 691 if (dumpto == -1) 692 dumpto = sparecol; 693 } 694 } 695 } 696 697 if (dumpto == -1) { 698 /* we couldn't find any live components to dump to!?!? 699 */ 700 error = EINVAL; 701 goto out; 702 } 703 704 bdev = bdevsw_lookup(raidPtr->Disks[dumpto].dev); 705 706 /* 707 Note that blkno is relative to this particular partition. 708 By adding the offset of this partition in the RAID 709 set, and also adding RF_PROTECTED_SECTORS, we get a 710 value that is relative to the partition used for the 711 underlying component. 712 */ 713 714 error = (*bdev->d_dump)(raidPtr->Disks[dumpto].dev, 715 blkno + offset, va, size); 716 717 out: 718 raidunlock(rs); 719 720 return error; 721 } 722 /* ARGSUSED */ 723 int 724 raidopen(dev_t dev, int flags, int fmt, 725 struct lwp *l) 726 { 727 int unit = raidunit(dev); 728 struct raid_softc *rs; 729 struct disklabel *lp; 730 int part, pmask; 731 int error = 0; 732 733 if (unit >= numraid) 734 return (ENXIO); 735 rs = &raid_softc[unit]; 736 737 if ((error = raidlock(rs)) != 0) 738 return (error); 739 lp = rs->sc_dkdev.dk_label; 740 741 part = DISKPART(dev); 742 743 /* 744 * If there are wedges, and this is not RAW_PART, then we 745 * need to fail. 746 */ 747 if (rs->sc_dkdev.dk_nwedges != 0 && part != RAW_PART) { 748 error = EBUSY; 749 goto bad; 750 } 751 pmask = (1 << part); 752 753 if ((rs->sc_flags & RAIDF_INITED) && 754 (rs->sc_dkdev.dk_openmask == 0)) 755 raidgetdisklabel(dev); 756 757 /* make sure that this partition exists */ 758 759 if (part != RAW_PART) { 760 if (((rs->sc_flags & RAIDF_INITED) == 0) || 761 ((part >= lp->d_npartitions) || 762 (lp->d_partitions[part].p_fstype == FS_UNUSED))) { 763 error = ENXIO; 764 goto bad; 765 } 766 } 767 /* Prevent this unit from being unconfigured while open. */ 768 switch (fmt) { 769 case S_IFCHR: 770 rs->sc_dkdev.dk_copenmask |= pmask; 771 break; 772 773 case S_IFBLK: 774 rs->sc_dkdev.dk_bopenmask |= pmask; 775 break; 776 } 777 778 if ((rs->sc_dkdev.dk_openmask == 0) && 779 ((rs->sc_flags & RAIDF_INITED) != 0)) { 780 /* First one... mark things as dirty... Note that we *MUST* 781 have done a configure before this. I DO NOT WANT TO BE 782 SCRIBBLING TO RANDOM COMPONENTS UNTIL IT'S BEEN DETERMINED 783 THAT THEY BELONG TOGETHER!!!!! */ 784 /* XXX should check to see if we're only open for reading 785 here... If so, we needn't do this, but then need some 786 other way of keeping track of what's happened.. */ 787 788 rf_markalldirty( raidPtrs[unit] ); 789 } 790 791 792 rs->sc_dkdev.dk_openmask = 793 rs->sc_dkdev.dk_copenmask | rs->sc_dkdev.dk_bopenmask; 794 795 bad: 796 raidunlock(rs); 797 798 return (error); 799 800 801 } 802 /* ARGSUSED */ 803 int 804 raidclose(dev_t dev, int flags, int fmt, struct lwp *l) 805 { 806 int unit = raidunit(dev); 807 struct cfdata *cf; 808 struct raid_softc *rs; 809 int error = 0; 810 int part; 811 812 if (unit >= numraid) 813 return (ENXIO); 814 rs = &raid_softc[unit]; 815 816 if ((error = raidlock(rs)) != 0) 817 return (error); 818 819 part = DISKPART(dev); 820 821 /* ...that much closer to allowing unconfiguration... */ 822 switch (fmt) { 823 case S_IFCHR: 824 rs->sc_dkdev.dk_copenmask &= ~(1 << part); 825 break; 826 827 case S_IFBLK: 828 rs->sc_dkdev.dk_bopenmask &= ~(1 << part); 829 break; 830 } 831 rs->sc_dkdev.dk_openmask = 832 rs->sc_dkdev.dk_copenmask | rs->sc_dkdev.dk_bopenmask; 833 834 if ((rs->sc_dkdev.dk_openmask == 0) && 835 ((rs->sc_flags & RAIDF_INITED) != 0)) { 836 /* Last one... device is not unconfigured yet. 837 Device shutdown has taken care of setting the 838 clean bits if RAIDF_INITED is not set 839 mark things as clean... */ 840 841 rf_update_component_labels(raidPtrs[unit], 842 RF_FINAL_COMPONENT_UPDATE); 843 if (doing_shutdown) { 844 /* last one, and we're going down, so 845 lights out for this RAID set too. */ 846 error = rf_Shutdown(raidPtrs[unit]); 847 848 /* It's no longer initialized... */ 849 rs->sc_flags &= ~RAIDF_INITED; 850 851 /* detach the device */ 852 853 cf = device_cfdata(rs->sc_dev); 854 error = config_detach(rs->sc_dev, DETACH_QUIET); 855 free(cf, M_RAIDFRAME); 856 857 /* Detach the disk. */ 858 disk_detach(&rs->sc_dkdev); 859 disk_destroy(&rs->sc_dkdev); 860 } 861 } 862 863 raidunlock(rs); 864 return (0); 865 866 } 867 868 void 869 raidstrategy(struct buf *bp) 870 { 871 int s; 872 873 unsigned int raidID = raidunit(bp->b_dev); 874 RF_Raid_t *raidPtr; 875 struct raid_softc *rs = &raid_softc[raidID]; 876 int wlabel; 877 878 if ((rs->sc_flags & RAIDF_INITED) ==0) { 879 bp->b_error = ENXIO; 880 goto done; 881 } 882 if (raidID >= numraid || !raidPtrs[raidID]) { 883 bp->b_error = ENODEV; 884 goto done; 885 } 886 raidPtr = raidPtrs[raidID]; 887 if (!raidPtr->valid) { 888 bp->b_error = ENODEV; 889 goto done; 890 } 891 if (bp->b_bcount == 0) { 892 db1_printf(("b_bcount is zero..\n")); 893 goto done; 894 } 895 896 /* 897 * Do bounds checking and adjust transfer. If there's an 898 * error, the bounds check will flag that for us. 899 */ 900 901 wlabel = rs->sc_flags & (RAIDF_WLABEL | RAIDF_LABELLING); 902 if (DISKPART(bp->b_dev) == RAW_PART) { 903 uint64_t size; /* device size in DEV_BSIZE unit */ 904 905 if (raidPtr->logBytesPerSector > DEV_BSHIFT) { 906 size = raidPtr->totalSectors << 907 (raidPtr->logBytesPerSector - DEV_BSHIFT); 908 } else { 909 size = raidPtr->totalSectors >> 910 (DEV_BSHIFT - raidPtr->logBytesPerSector); 911 } 912 if (bounds_check_with_mediasize(bp, DEV_BSIZE, size) <= 0) { 913 goto done; 914 } 915 } else { 916 if (bounds_check_with_label(&rs->sc_dkdev, bp, wlabel) <= 0) { 917 db1_printf(("Bounds check failed!!:%d %d\n", 918 (int) bp->b_blkno, (int) wlabel)); 919 goto done; 920 } 921 } 922 s = splbio(); 923 924 bp->b_resid = 0; 925 926 /* stuff it onto our queue */ 927 BUFQ_PUT(rs->buf_queue, bp); 928 929 /* scheduled the IO to happen at the next convenient time */ 930 wakeup(&(raidPtrs[raidID]->iodone)); 931 932 splx(s); 933 return; 934 935 done: 936 bp->b_resid = bp->b_bcount; 937 biodone(bp); 938 } 939 /* ARGSUSED */ 940 int 941 raidread(dev_t dev, struct uio *uio, int flags) 942 { 943 int unit = raidunit(dev); 944 struct raid_softc *rs; 945 946 if (unit >= numraid) 947 return (ENXIO); 948 rs = &raid_softc[unit]; 949 950 if ((rs->sc_flags & RAIDF_INITED) == 0) 951 return (ENXIO); 952 953 return (physio(raidstrategy, NULL, dev, B_READ, minphys, uio)); 954 955 } 956 /* ARGSUSED */ 957 int 958 raidwrite(dev_t dev, struct uio *uio, int flags) 959 { 960 int unit = raidunit(dev); 961 struct raid_softc *rs; 962 963 if (unit >= numraid) 964 return (ENXIO); 965 rs = &raid_softc[unit]; 966 967 if ((rs->sc_flags & RAIDF_INITED) == 0) 968 return (ENXIO); 969 970 return (physio(raidstrategy, NULL, dev, B_WRITE, minphys, uio)); 971 972 } 973 974 int 975 raidioctl(dev_t dev, u_long cmd, void *data, int flag, struct lwp *l) 976 { 977 int unit = raidunit(dev); 978 int error = 0; 979 int part, pmask; 980 struct cfdata *cf; 981 struct raid_softc *rs; 982 RF_Config_t *k_cfg, *u_cfg; 983 RF_Raid_t *raidPtr; 984 RF_RaidDisk_t *diskPtr; 985 RF_AccTotals_t *totals; 986 RF_DeviceConfig_t *d_cfg, **ucfgp; 987 u_char *specific_buf; 988 int retcode = 0; 989 int column; 990 int raidid; 991 struct rf_recon_req *rrcopy, *rr; 992 RF_ComponentLabel_t *clabel; 993 RF_ComponentLabel_t *ci_label; 994 RF_ComponentLabel_t **clabel_ptr; 995 RF_SingleComponent_t *sparePtr,*componentPtr; 996 RF_SingleComponent_t component; 997 RF_ProgressInfo_t progressInfo, **progressInfoPtr; 998 int i, j, d; 999 #ifdef __HAVE_OLD_DISKLABEL 1000 struct disklabel newlabel; 1001 #endif 1002 struct dkwedge_info *dkw; 1003 1004 if (unit >= numraid) 1005 return (ENXIO); 1006 rs = &raid_softc[unit]; 1007 raidPtr = raidPtrs[unit]; 1008 1009 db1_printf(("raidioctl: %d %d %d %d\n", (int) dev, 1010 (int) DISKPART(dev), (int) unit, (int) cmd)); 1011 1012 /* Must be open for writes for these commands... */ 1013 switch (cmd) { 1014 #ifdef DIOCGSECTORSIZE 1015 case DIOCGSECTORSIZE: 1016 *(u_int *)data = raidPtr->bytesPerSector; 1017 return 0; 1018 case DIOCGMEDIASIZE: 1019 *(off_t *)data = 1020 (off_t)raidPtr->totalSectors * raidPtr->bytesPerSector; 1021 return 0; 1022 #endif 1023 case DIOCSDINFO: 1024 case DIOCWDINFO: 1025 #ifdef __HAVE_OLD_DISKLABEL 1026 case ODIOCWDINFO: 1027 case ODIOCSDINFO: 1028 #endif 1029 case DIOCWLABEL: 1030 case DIOCAWEDGE: 1031 case DIOCDWEDGE: 1032 if ((flag & FWRITE) == 0) 1033 return (EBADF); 1034 } 1035 1036 /* Must be initialized for these... */ 1037 switch (cmd) { 1038 case DIOCGDINFO: 1039 case DIOCSDINFO: 1040 case DIOCWDINFO: 1041 #ifdef __HAVE_OLD_DISKLABEL 1042 case ODIOCGDINFO: 1043 case ODIOCWDINFO: 1044 case ODIOCSDINFO: 1045 case ODIOCGDEFLABEL: 1046 #endif 1047 case DIOCGPART: 1048 case DIOCWLABEL: 1049 case DIOCGDEFLABEL: 1050 case DIOCAWEDGE: 1051 case DIOCDWEDGE: 1052 case DIOCLWEDGES: 1053 case RAIDFRAME_SHUTDOWN: 1054 case RAIDFRAME_REWRITEPARITY: 1055 case RAIDFRAME_GET_INFO: 1056 case RAIDFRAME_RESET_ACCTOTALS: 1057 case RAIDFRAME_GET_ACCTOTALS: 1058 case RAIDFRAME_KEEP_ACCTOTALS: 1059 case RAIDFRAME_GET_SIZE: 1060 case RAIDFRAME_FAIL_DISK: 1061 case RAIDFRAME_COPYBACK: 1062 case RAIDFRAME_CHECK_RECON_STATUS: 1063 case RAIDFRAME_CHECK_RECON_STATUS_EXT: 1064 case RAIDFRAME_GET_COMPONENT_LABEL: 1065 case RAIDFRAME_SET_COMPONENT_LABEL: 1066 case RAIDFRAME_ADD_HOT_SPARE: 1067 case RAIDFRAME_REMOVE_HOT_SPARE: 1068 case RAIDFRAME_INIT_LABELS: 1069 case RAIDFRAME_REBUILD_IN_PLACE: 1070 case RAIDFRAME_CHECK_PARITY: 1071 case RAIDFRAME_CHECK_PARITYREWRITE_STATUS: 1072 case RAIDFRAME_CHECK_PARITYREWRITE_STATUS_EXT: 1073 case RAIDFRAME_CHECK_COPYBACK_STATUS: 1074 case RAIDFRAME_CHECK_COPYBACK_STATUS_EXT: 1075 case RAIDFRAME_SET_AUTOCONFIG: 1076 case RAIDFRAME_SET_ROOT: 1077 case RAIDFRAME_DELETE_COMPONENT: 1078 case RAIDFRAME_INCORPORATE_HOT_SPARE: 1079 if ((rs->sc_flags & RAIDF_INITED) == 0) 1080 return (ENXIO); 1081 } 1082 1083 switch (cmd) { 1084 1085 /* configure the system */ 1086 case RAIDFRAME_CONFIGURE: 1087 1088 if (raidPtr->valid) { 1089 /* There is a valid RAID set running on this unit! */ 1090 printf("raid%d: Device already configured!\n",unit); 1091 return(EINVAL); 1092 } 1093 1094 /* copy-in the configuration information */ 1095 /* data points to a pointer to the configuration structure */ 1096 1097 u_cfg = *((RF_Config_t **) data); 1098 RF_Malloc(k_cfg, sizeof(RF_Config_t), (RF_Config_t *)); 1099 if (k_cfg == NULL) { 1100 return (ENOMEM); 1101 } 1102 retcode = copyin(u_cfg, k_cfg, sizeof(RF_Config_t)); 1103 if (retcode) { 1104 RF_Free(k_cfg, sizeof(RF_Config_t)); 1105 db1_printf(("rf_ioctl: retcode=%d copyin.1\n", 1106 retcode)); 1107 return (retcode); 1108 } 1109 /* allocate a buffer for the layout-specific data, and copy it 1110 * in */ 1111 if (k_cfg->layoutSpecificSize) { 1112 if (k_cfg->layoutSpecificSize > 10000) { 1113 /* sanity check */ 1114 RF_Free(k_cfg, sizeof(RF_Config_t)); 1115 return (EINVAL); 1116 } 1117 RF_Malloc(specific_buf, k_cfg->layoutSpecificSize, 1118 (u_char *)); 1119 if (specific_buf == NULL) { 1120 RF_Free(k_cfg, sizeof(RF_Config_t)); 1121 return (ENOMEM); 1122 } 1123 retcode = copyin(k_cfg->layoutSpecific, specific_buf, 1124 k_cfg->layoutSpecificSize); 1125 if (retcode) { 1126 RF_Free(k_cfg, sizeof(RF_Config_t)); 1127 RF_Free(specific_buf, 1128 k_cfg->layoutSpecificSize); 1129 db1_printf(("rf_ioctl: retcode=%d copyin.2\n", 1130 retcode)); 1131 return (retcode); 1132 } 1133 } else 1134 specific_buf = NULL; 1135 k_cfg->layoutSpecific = specific_buf; 1136 1137 /* should do some kind of sanity check on the configuration. 1138 * Store the sum of all the bytes in the last byte? */ 1139 1140 /* configure the system */ 1141 1142 /* 1143 * Clear the entire RAID descriptor, just to make sure 1144 * there is no stale data left in the case of a 1145 * reconfiguration 1146 */ 1147 memset((char *) raidPtr, 0, sizeof(RF_Raid_t)); 1148 raidPtr->raidid = unit; 1149 1150 retcode = rf_Configure(raidPtr, k_cfg, NULL); 1151 1152 if (retcode == 0) { 1153 1154 /* allow this many simultaneous IO's to 1155 this RAID device */ 1156 raidPtr->openings = RAIDOUTSTANDING; 1157 1158 raidinit(raidPtr); 1159 rf_markalldirty(raidPtr); 1160 } 1161 /* free the buffers. No return code here. */ 1162 if (k_cfg->layoutSpecificSize) { 1163 RF_Free(specific_buf, k_cfg->layoutSpecificSize); 1164 } 1165 RF_Free(k_cfg, sizeof(RF_Config_t)); 1166 1167 return (retcode); 1168 1169 /* shutdown the system */ 1170 case RAIDFRAME_SHUTDOWN: 1171 1172 if ((error = raidlock(rs)) != 0) 1173 return (error); 1174 1175 /* 1176 * If somebody has a partition mounted, we shouldn't 1177 * shutdown. 1178 */ 1179 1180 part = DISKPART(dev); 1181 pmask = (1 << part); 1182 if ((rs->sc_dkdev.dk_openmask & ~pmask) || 1183 ((rs->sc_dkdev.dk_bopenmask & pmask) && 1184 (rs->sc_dkdev.dk_copenmask & pmask))) { 1185 raidunlock(rs); 1186 return (EBUSY); 1187 } 1188 1189 retcode = rf_Shutdown(raidPtr); 1190 1191 /* It's no longer initialized... */ 1192 rs->sc_flags &= ~RAIDF_INITED; 1193 1194 /* free the pseudo device attach bits */ 1195 1196 cf = device_cfdata(rs->sc_dev); 1197 /* XXX this causes us to not return any errors 1198 from the above call to rf_Shutdown() */ 1199 retcode = config_detach(rs->sc_dev, DETACH_QUIET); 1200 free(cf, M_RAIDFRAME); 1201 1202 /* Detach the disk. */ 1203 disk_detach(&rs->sc_dkdev); 1204 disk_destroy(&rs->sc_dkdev); 1205 1206 raidunlock(rs); 1207 1208 return (retcode); 1209 case RAIDFRAME_GET_COMPONENT_LABEL: 1210 clabel_ptr = (RF_ComponentLabel_t **) data; 1211 /* need to read the component label for the disk indicated 1212 by row,column in clabel */ 1213 1214 /* For practice, let's get it directly fromdisk, rather 1215 than from the in-core copy */ 1216 RF_Malloc( clabel, sizeof( RF_ComponentLabel_t ), 1217 (RF_ComponentLabel_t *)); 1218 if (clabel == NULL) 1219 return (ENOMEM); 1220 1221 retcode = copyin( *clabel_ptr, clabel, 1222 sizeof(RF_ComponentLabel_t)); 1223 1224 if (retcode) { 1225 RF_Free( clabel, sizeof(RF_ComponentLabel_t)); 1226 return(retcode); 1227 } 1228 1229 clabel->row = 0; /* Don't allow looking at anything else.*/ 1230 1231 column = clabel->column; 1232 1233 if ((column < 0) || (column >= raidPtr->numCol + 1234 raidPtr->numSpare)) { 1235 RF_Free( clabel, sizeof(RF_ComponentLabel_t)); 1236 return(EINVAL); 1237 } 1238 1239 retcode = raidread_component_label(raidPtr->Disks[column].dev, 1240 raidPtr->raid_cinfo[column].ci_vp, 1241 clabel ); 1242 1243 if (retcode == 0) { 1244 retcode = copyout(clabel, *clabel_ptr, 1245 sizeof(RF_ComponentLabel_t)); 1246 } 1247 RF_Free(clabel, sizeof(RF_ComponentLabel_t)); 1248 return (retcode); 1249 1250 case RAIDFRAME_SET_COMPONENT_LABEL: 1251 clabel = (RF_ComponentLabel_t *) data; 1252 1253 /* XXX check the label for valid stuff... */ 1254 /* Note that some things *should not* get modified -- 1255 the user should be re-initing the labels instead of 1256 trying to patch things. 1257 */ 1258 1259 raidid = raidPtr->raidid; 1260 #ifdef DEBUG 1261 printf("raid%d: Got component label:\n", raidid); 1262 printf("raid%d: Version: %d\n", raidid, clabel->version); 1263 printf("raid%d: Serial Number: %d\n", raidid, clabel->serial_number); 1264 printf("raid%d: Mod counter: %d\n", raidid, clabel->mod_counter); 1265 printf("raid%d: Column: %d\n", raidid, clabel->column); 1266 printf("raid%d: Num Columns: %d\n", raidid, clabel->num_columns); 1267 printf("raid%d: Clean: %d\n", raidid, clabel->clean); 1268 printf("raid%d: Status: %d\n", raidid, clabel->status); 1269 #endif 1270 clabel->row = 0; 1271 column = clabel->column; 1272 1273 if ((column < 0) || (column >= raidPtr->numCol)) { 1274 return(EINVAL); 1275 } 1276 1277 /* XXX this isn't allowed to do anything for now :-) */ 1278 1279 /* XXX and before it is, we need to fill in the rest 1280 of the fields!?!?!?! */ 1281 #if 0 1282 raidwrite_component_label( 1283 raidPtr->Disks[column].dev, 1284 raidPtr->raid_cinfo[column].ci_vp, 1285 clabel ); 1286 #endif 1287 return (0); 1288 1289 case RAIDFRAME_INIT_LABELS: 1290 clabel = (RF_ComponentLabel_t *) data; 1291 /* 1292 we only want the serial number from 1293 the above. We get all the rest of the information 1294 from the config that was used to create this RAID 1295 set. 1296 */ 1297 1298 raidPtr->serial_number = clabel->serial_number; 1299 1300 RF_Malloc(ci_label, sizeof(RF_ComponentLabel_t), 1301 (RF_ComponentLabel_t *)); 1302 if (ci_label == NULL) 1303 return (ENOMEM); 1304 1305 raid_init_component_label(raidPtr, ci_label); 1306 ci_label->serial_number = clabel->serial_number; 1307 ci_label->row = 0; /* we dont' pretend to support more */ 1308 1309 for(column=0;column<raidPtr->numCol;column++) { 1310 diskPtr = &raidPtr->Disks[column]; 1311 if (!RF_DEAD_DISK(diskPtr->status)) { 1312 ci_label->partitionSize = diskPtr->partitionSize; 1313 ci_label->column = column; 1314 raidwrite_component_label( 1315 raidPtr->Disks[column].dev, 1316 raidPtr->raid_cinfo[column].ci_vp, 1317 ci_label ); 1318 } 1319 } 1320 RF_Free(ci_label, sizeof(RF_ComponentLabel_t)); 1321 1322 return (retcode); 1323 case RAIDFRAME_SET_AUTOCONFIG: 1324 d = rf_set_autoconfig(raidPtr, *(int *) data); 1325 printf("raid%d: New autoconfig value is: %d\n", 1326 raidPtr->raidid, d); 1327 *(int *) data = d; 1328 return (retcode); 1329 1330 case RAIDFRAME_SET_ROOT: 1331 d = rf_set_rootpartition(raidPtr, *(int *) data); 1332 printf("raid%d: New rootpartition value is: %d\n", 1333 raidPtr->raidid, d); 1334 *(int *) data = d; 1335 return (retcode); 1336 1337 /* initialize all parity */ 1338 case RAIDFRAME_REWRITEPARITY: 1339 1340 if (raidPtr->Layout.map->faultsTolerated == 0) { 1341 /* Parity for RAID 0 is trivially correct */ 1342 raidPtr->parity_good = RF_RAID_CLEAN; 1343 return(0); 1344 } 1345 1346 if (raidPtr->parity_rewrite_in_progress == 1) { 1347 /* Re-write is already in progress! */ 1348 return(EINVAL); 1349 } 1350 1351 retcode = RF_CREATE_THREAD(raidPtr->parity_rewrite_thread, 1352 rf_RewriteParityThread, 1353 raidPtr,"raid_parity"); 1354 return (retcode); 1355 1356 1357 case RAIDFRAME_ADD_HOT_SPARE: 1358 sparePtr = (RF_SingleComponent_t *) data; 1359 memcpy( &component, sparePtr, sizeof(RF_SingleComponent_t)); 1360 retcode = rf_add_hot_spare(raidPtr, &component); 1361 return(retcode); 1362 1363 case RAIDFRAME_REMOVE_HOT_SPARE: 1364 return(retcode); 1365 1366 case RAIDFRAME_DELETE_COMPONENT: 1367 componentPtr = (RF_SingleComponent_t *)data; 1368 memcpy( &component, componentPtr, 1369 sizeof(RF_SingleComponent_t)); 1370 retcode = rf_delete_component(raidPtr, &component); 1371 return(retcode); 1372 1373 case RAIDFRAME_INCORPORATE_HOT_SPARE: 1374 componentPtr = (RF_SingleComponent_t *)data; 1375 memcpy( &component, componentPtr, 1376 sizeof(RF_SingleComponent_t)); 1377 retcode = rf_incorporate_hot_spare(raidPtr, &component); 1378 return(retcode); 1379 1380 case RAIDFRAME_REBUILD_IN_PLACE: 1381 1382 if (raidPtr->Layout.map->faultsTolerated == 0) { 1383 /* Can't do this on a RAID 0!! */ 1384 return(EINVAL); 1385 } 1386 1387 if (raidPtr->recon_in_progress == 1) { 1388 /* a reconstruct is already in progress! */ 1389 return(EINVAL); 1390 } 1391 1392 componentPtr = (RF_SingleComponent_t *) data; 1393 memcpy( &component, componentPtr, 1394 sizeof(RF_SingleComponent_t)); 1395 component.row = 0; /* we don't support any more */ 1396 column = component.column; 1397 1398 if ((column < 0) || (column >= raidPtr->numCol)) { 1399 return(EINVAL); 1400 } 1401 1402 RF_LOCK_MUTEX(raidPtr->mutex); 1403 if ((raidPtr->Disks[column].status == rf_ds_optimal) && 1404 (raidPtr->numFailures > 0)) { 1405 /* XXX 0 above shouldn't be constant!!! */ 1406 /* some component other than this has failed. 1407 Let's not make things worse than they already 1408 are... */ 1409 printf("raid%d: Unable to reconstruct to disk at:\n", 1410 raidPtr->raidid); 1411 printf("raid%d: Col: %d Too many failures.\n", 1412 raidPtr->raidid, column); 1413 RF_UNLOCK_MUTEX(raidPtr->mutex); 1414 return (EINVAL); 1415 } 1416 if (raidPtr->Disks[column].status == 1417 rf_ds_reconstructing) { 1418 printf("raid%d: Unable to reconstruct to disk at:\n", 1419 raidPtr->raidid); 1420 printf("raid%d: Col: %d Reconstruction already occuring!\n", raidPtr->raidid, column); 1421 1422 RF_UNLOCK_MUTEX(raidPtr->mutex); 1423 return (EINVAL); 1424 } 1425 if (raidPtr->Disks[column].status == rf_ds_spared) { 1426 RF_UNLOCK_MUTEX(raidPtr->mutex); 1427 return (EINVAL); 1428 } 1429 RF_UNLOCK_MUTEX(raidPtr->mutex); 1430 1431 RF_Malloc(rrcopy, sizeof(*rrcopy), (struct rf_recon_req *)); 1432 if (rrcopy == NULL) 1433 return(ENOMEM); 1434 1435 rrcopy->raidPtr = (void *) raidPtr; 1436 rrcopy->col = column; 1437 1438 retcode = RF_CREATE_THREAD(raidPtr->recon_thread, 1439 rf_ReconstructInPlaceThread, 1440 rrcopy,"raid_reconip"); 1441 return(retcode); 1442 1443 case RAIDFRAME_GET_INFO: 1444 if (!raidPtr->valid) 1445 return (ENODEV); 1446 ucfgp = (RF_DeviceConfig_t **) data; 1447 RF_Malloc(d_cfg, sizeof(RF_DeviceConfig_t), 1448 (RF_DeviceConfig_t *)); 1449 if (d_cfg == NULL) 1450 return (ENOMEM); 1451 d_cfg->rows = 1; /* there is only 1 row now */ 1452 d_cfg->cols = raidPtr->numCol; 1453 d_cfg->ndevs = raidPtr->numCol; 1454 if (d_cfg->ndevs >= RF_MAX_DISKS) { 1455 RF_Free(d_cfg, sizeof(RF_DeviceConfig_t)); 1456 return (ENOMEM); 1457 } 1458 d_cfg->nspares = raidPtr->numSpare; 1459 if (d_cfg->nspares >= RF_MAX_DISKS) { 1460 RF_Free(d_cfg, sizeof(RF_DeviceConfig_t)); 1461 return (ENOMEM); 1462 } 1463 d_cfg->maxqdepth = raidPtr->maxQueueDepth; 1464 d = 0; 1465 for (j = 0; j < d_cfg->cols; j++) { 1466 d_cfg->devs[d] = raidPtr->Disks[j]; 1467 d++; 1468 } 1469 for (j = d_cfg->cols, i = 0; i < d_cfg->nspares; i++, j++) { 1470 d_cfg->spares[i] = raidPtr->Disks[j]; 1471 } 1472 retcode = copyout(d_cfg, *ucfgp, sizeof(RF_DeviceConfig_t)); 1473 RF_Free(d_cfg, sizeof(RF_DeviceConfig_t)); 1474 1475 return (retcode); 1476 1477 case RAIDFRAME_CHECK_PARITY: 1478 *(int *) data = raidPtr->parity_good; 1479 return (0); 1480 1481 case RAIDFRAME_RESET_ACCTOTALS: 1482 memset(&raidPtr->acc_totals, 0, sizeof(raidPtr->acc_totals)); 1483 return (0); 1484 1485 case RAIDFRAME_GET_ACCTOTALS: 1486 totals = (RF_AccTotals_t *) data; 1487 *totals = raidPtr->acc_totals; 1488 return (0); 1489 1490 case RAIDFRAME_KEEP_ACCTOTALS: 1491 raidPtr->keep_acc_totals = *(int *)data; 1492 return (0); 1493 1494 case RAIDFRAME_GET_SIZE: 1495 *(int *) data = raidPtr->totalSectors; 1496 return (0); 1497 1498 /* fail a disk & optionally start reconstruction */ 1499 case RAIDFRAME_FAIL_DISK: 1500 1501 if (raidPtr->Layout.map->faultsTolerated == 0) { 1502 /* Can't do this on a RAID 0!! */ 1503 return(EINVAL); 1504 } 1505 1506 rr = (struct rf_recon_req *) data; 1507 rr->row = 0; 1508 if (rr->col < 0 || rr->col >= raidPtr->numCol) 1509 return (EINVAL); 1510 1511 1512 RF_LOCK_MUTEX(raidPtr->mutex); 1513 if (raidPtr->status == rf_rs_reconstructing) { 1514 /* you can't fail a disk while we're reconstructing! */ 1515 /* XXX wrong for RAID6 */ 1516 RF_UNLOCK_MUTEX(raidPtr->mutex); 1517 return (EINVAL); 1518 } 1519 if ((raidPtr->Disks[rr->col].status == 1520 rf_ds_optimal) && (raidPtr->numFailures > 0)) { 1521 /* some other component has failed. Let's not make 1522 things worse. XXX wrong for RAID6 */ 1523 RF_UNLOCK_MUTEX(raidPtr->mutex); 1524 return (EINVAL); 1525 } 1526 if (raidPtr->Disks[rr->col].status == rf_ds_spared) { 1527 /* Can't fail a spared disk! */ 1528 RF_UNLOCK_MUTEX(raidPtr->mutex); 1529 return (EINVAL); 1530 } 1531 RF_UNLOCK_MUTEX(raidPtr->mutex); 1532 1533 /* make a copy of the recon request so that we don't rely on 1534 * the user's buffer */ 1535 RF_Malloc(rrcopy, sizeof(*rrcopy), (struct rf_recon_req *)); 1536 if (rrcopy == NULL) 1537 return(ENOMEM); 1538 memcpy(rrcopy, rr, sizeof(*rr)); 1539 rrcopy->raidPtr = (void *) raidPtr; 1540 1541 retcode = RF_CREATE_THREAD(raidPtr->recon_thread, 1542 rf_ReconThread, 1543 rrcopy,"raid_recon"); 1544 return (0); 1545 1546 /* invoke a copyback operation after recon on whatever disk 1547 * needs it, if any */ 1548 case RAIDFRAME_COPYBACK: 1549 1550 if (raidPtr->Layout.map->faultsTolerated == 0) { 1551 /* This makes no sense on a RAID 0!! */ 1552 return(EINVAL); 1553 } 1554 1555 if (raidPtr->copyback_in_progress == 1) { 1556 /* Copyback is already in progress! */ 1557 return(EINVAL); 1558 } 1559 1560 retcode = RF_CREATE_THREAD(raidPtr->copyback_thread, 1561 rf_CopybackThread, 1562 raidPtr,"raid_copyback"); 1563 return (retcode); 1564 1565 /* return the percentage completion of reconstruction */ 1566 case RAIDFRAME_CHECK_RECON_STATUS: 1567 if (raidPtr->Layout.map->faultsTolerated == 0) { 1568 /* This makes no sense on a RAID 0, so tell the 1569 user it's done. */ 1570 *(int *) data = 100; 1571 return(0); 1572 } 1573 if (raidPtr->status != rf_rs_reconstructing) 1574 *(int *) data = 100; 1575 else { 1576 if (raidPtr->reconControl->numRUsTotal > 0) { 1577 *(int *) data = (raidPtr->reconControl->numRUsComplete * 100 / raidPtr->reconControl->numRUsTotal); 1578 } else { 1579 *(int *) data = 0; 1580 } 1581 } 1582 return (0); 1583 case RAIDFRAME_CHECK_RECON_STATUS_EXT: 1584 progressInfoPtr = (RF_ProgressInfo_t **) data; 1585 if (raidPtr->status != rf_rs_reconstructing) { 1586 progressInfo.remaining = 0; 1587 progressInfo.completed = 100; 1588 progressInfo.total = 100; 1589 } else { 1590 progressInfo.total = 1591 raidPtr->reconControl->numRUsTotal; 1592 progressInfo.completed = 1593 raidPtr->reconControl->numRUsComplete; 1594 progressInfo.remaining = progressInfo.total - 1595 progressInfo.completed; 1596 } 1597 retcode = copyout(&progressInfo, *progressInfoPtr, 1598 sizeof(RF_ProgressInfo_t)); 1599 return (retcode); 1600 1601 case RAIDFRAME_CHECK_PARITYREWRITE_STATUS: 1602 if (raidPtr->Layout.map->faultsTolerated == 0) { 1603 /* This makes no sense on a RAID 0, so tell the 1604 user it's done. */ 1605 *(int *) data = 100; 1606 return(0); 1607 } 1608 if (raidPtr->parity_rewrite_in_progress == 1) { 1609 *(int *) data = 100 * 1610 raidPtr->parity_rewrite_stripes_done / 1611 raidPtr->Layout.numStripe; 1612 } else { 1613 *(int *) data = 100; 1614 } 1615 return (0); 1616 1617 case RAIDFRAME_CHECK_PARITYREWRITE_STATUS_EXT: 1618 progressInfoPtr = (RF_ProgressInfo_t **) data; 1619 if (raidPtr->parity_rewrite_in_progress == 1) { 1620 progressInfo.total = raidPtr->Layout.numStripe; 1621 progressInfo.completed = 1622 raidPtr->parity_rewrite_stripes_done; 1623 progressInfo.remaining = progressInfo.total - 1624 progressInfo.completed; 1625 } else { 1626 progressInfo.remaining = 0; 1627 progressInfo.completed = 100; 1628 progressInfo.total = 100; 1629 } 1630 retcode = copyout(&progressInfo, *progressInfoPtr, 1631 sizeof(RF_ProgressInfo_t)); 1632 return (retcode); 1633 1634 case RAIDFRAME_CHECK_COPYBACK_STATUS: 1635 if (raidPtr->Layout.map->faultsTolerated == 0) { 1636 /* This makes no sense on a RAID 0 */ 1637 *(int *) data = 100; 1638 return(0); 1639 } 1640 if (raidPtr->copyback_in_progress == 1) { 1641 *(int *) data = 100 * raidPtr->copyback_stripes_done / 1642 raidPtr->Layout.numStripe; 1643 } else { 1644 *(int *) data = 100; 1645 } 1646 return (0); 1647 1648 case RAIDFRAME_CHECK_COPYBACK_STATUS_EXT: 1649 progressInfoPtr = (RF_ProgressInfo_t **) data; 1650 if (raidPtr->copyback_in_progress == 1) { 1651 progressInfo.total = raidPtr->Layout.numStripe; 1652 progressInfo.completed = 1653 raidPtr->copyback_stripes_done; 1654 progressInfo.remaining = progressInfo.total - 1655 progressInfo.completed; 1656 } else { 1657 progressInfo.remaining = 0; 1658 progressInfo.completed = 100; 1659 progressInfo.total = 100; 1660 } 1661 retcode = copyout(&progressInfo, *progressInfoPtr, 1662 sizeof(RF_ProgressInfo_t)); 1663 return (retcode); 1664 1665 /* the sparetable daemon calls this to wait for the kernel to 1666 * need a spare table. this ioctl does not return until a 1667 * spare table is needed. XXX -- calling mpsleep here in the 1668 * ioctl code is almost certainly wrong and evil. -- XXX XXX 1669 * -- I should either compute the spare table in the kernel, 1670 * or have a different -- XXX XXX -- interface (a different 1671 * character device) for delivering the table -- XXX */ 1672 #if 0 1673 case RAIDFRAME_SPARET_WAIT: 1674 RF_LOCK_MUTEX(rf_sparet_wait_mutex); 1675 while (!rf_sparet_wait_queue) 1676 mpsleep(&rf_sparet_wait_queue, (PZERO + 1) | PCATCH, "sparet wait", 0, (void *) simple_lock_addr(rf_sparet_wait_mutex), MS_LOCK_SIMPLE); 1677 waitreq = rf_sparet_wait_queue; 1678 rf_sparet_wait_queue = rf_sparet_wait_queue->next; 1679 RF_UNLOCK_MUTEX(rf_sparet_wait_mutex); 1680 1681 /* structure assignment */ 1682 *((RF_SparetWait_t *) data) = *waitreq; 1683 1684 RF_Free(waitreq, sizeof(*waitreq)); 1685 return (0); 1686 1687 /* wakes up a process waiting on SPARET_WAIT and puts an error 1688 * code in it that will cause the dameon to exit */ 1689 case RAIDFRAME_ABORT_SPARET_WAIT: 1690 RF_Malloc(waitreq, sizeof(*waitreq), (RF_SparetWait_t *)); 1691 waitreq->fcol = -1; 1692 RF_LOCK_MUTEX(rf_sparet_wait_mutex); 1693 waitreq->next = rf_sparet_wait_queue; 1694 rf_sparet_wait_queue = waitreq; 1695 RF_UNLOCK_MUTEX(rf_sparet_wait_mutex); 1696 wakeup(&rf_sparet_wait_queue); 1697 return (0); 1698 1699 /* used by the spare table daemon to deliver a spare table 1700 * into the kernel */ 1701 case RAIDFRAME_SEND_SPARET: 1702 1703 /* install the spare table */ 1704 retcode = rf_SetSpareTable(raidPtr, *(void **) data); 1705 1706 /* respond to the requestor. the return status of the spare 1707 * table installation is passed in the "fcol" field */ 1708 RF_Malloc(waitreq, sizeof(*waitreq), (RF_SparetWait_t *)); 1709 waitreq->fcol = retcode; 1710 RF_LOCK_MUTEX(rf_sparet_wait_mutex); 1711 waitreq->next = rf_sparet_resp_queue; 1712 rf_sparet_resp_queue = waitreq; 1713 wakeup(&rf_sparet_resp_queue); 1714 RF_UNLOCK_MUTEX(rf_sparet_wait_mutex); 1715 1716 return (retcode); 1717 #endif 1718 1719 default: 1720 break; /* fall through to the os-specific code below */ 1721 1722 } 1723 1724 if (!raidPtr->valid) 1725 return (EINVAL); 1726 1727 /* 1728 * Add support for "regular" device ioctls here. 1729 */ 1730 1731 switch (cmd) { 1732 case DIOCGDINFO: 1733 *(struct disklabel *) data = *(rs->sc_dkdev.dk_label); 1734 break; 1735 #ifdef __HAVE_OLD_DISKLABEL 1736 case ODIOCGDINFO: 1737 newlabel = *(rs->sc_dkdev.dk_label); 1738 if (newlabel.d_npartitions > OLDMAXPARTITIONS) 1739 return ENOTTY; 1740 memcpy(data, &newlabel, sizeof (struct olddisklabel)); 1741 break; 1742 #endif 1743 1744 case DIOCGPART: 1745 ((struct partinfo *) data)->disklab = rs->sc_dkdev.dk_label; 1746 ((struct partinfo *) data)->part = 1747 &rs->sc_dkdev.dk_label->d_partitions[DISKPART(dev)]; 1748 break; 1749 1750 case DIOCWDINFO: 1751 case DIOCSDINFO: 1752 #ifdef __HAVE_OLD_DISKLABEL 1753 case ODIOCWDINFO: 1754 case ODIOCSDINFO: 1755 #endif 1756 { 1757 struct disklabel *lp; 1758 #ifdef __HAVE_OLD_DISKLABEL 1759 if (cmd == ODIOCSDINFO || cmd == ODIOCWDINFO) { 1760 memset(&newlabel, 0, sizeof newlabel); 1761 memcpy(&newlabel, data, sizeof (struct olddisklabel)); 1762 lp = &newlabel; 1763 } else 1764 #endif 1765 lp = (struct disklabel *)data; 1766 1767 if ((error = raidlock(rs)) != 0) 1768 return (error); 1769 1770 rs->sc_flags |= RAIDF_LABELLING; 1771 1772 error = setdisklabel(rs->sc_dkdev.dk_label, 1773 lp, 0, rs->sc_dkdev.dk_cpulabel); 1774 if (error == 0) { 1775 if (cmd == DIOCWDINFO 1776 #ifdef __HAVE_OLD_DISKLABEL 1777 || cmd == ODIOCWDINFO 1778 #endif 1779 ) 1780 error = writedisklabel(RAIDLABELDEV(dev), 1781 raidstrategy, rs->sc_dkdev.dk_label, 1782 rs->sc_dkdev.dk_cpulabel); 1783 } 1784 rs->sc_flags &= ~RAIDF_LABELLING; 1785 1786 raidunlock(rs); 1787 1788 if (error) 1789 return (error); 1790 break; 1791 } 1792 1793 case DIOCWLABEL: 1794 if (*(int *) data != 0) 1795 rs->sc_flags |= RAIDF_WLABEL; 1796 else 1797 rs->sc_flags &= ~RAIDF_WLABEL; 1798 break; 1799 1800 case DIOCGDEFLABEL: 1801 raidgetdefaultlabel(raidPtr, rs, (struct disklabel *) data); 1802 break; 1803 1804 #ifdef __HAVE_OLD_DISKLABEL 1805 case ODIOCGDEFLABEL: 1806 raidgetdefaultlabel(raidPtr, rs, &newlabel); 1807 if (newlabel.d_npartitions > OLDMAXPARTITIONS) 1808 return ENOTTY; 1809 memcpy(data, &newlabel, sizeof (struct olddisklabel)); 1810 break; 1811 #endif 1812 1813 case DIOCAWEDGE: 1814 case DIOCDWEDGE: 1815 dkw = (void *)data; 1816 1817 /* If the ioctl happens here, the parent is us. */ 1818 (void)strcpy(dkw->dkw_parent, rs->sc_xname); 1819 return cmd == DIOCAWEDGE ? dkwedge_add(dkw) : dkwedge_del(dkw); 1820 1821 case DIOCLWEDGES: 1822 return dkwedge_list(&rs->sc_dkdev, 1823 (struct dkwedge_list *)data, l); 1824 1825 default: 1826 retcode = ENOTTY; 1827 } 1828 return (retcode); 1829 1830 } 1831 1832 1833 /* raidinit -- complete the rest of the initialization for the 1834 RAIDframe device. */ 1835 1836 1837 static void 1838 raidinit(RF_Raid_t *raidPtr) 1839 { 1840 struct cfdata *cf; 1841 struct raid_softc *rs; 1842 int unit; 1843 1844 unit = raidPtr->raidid; 1845 1846 rs = &raid_softc[unit]; 1847 1848 /* XXX should check return code first... */ 1849 rs->sc_flags |= RAIDF_INITED; 1850 1851 /* XXX doesn't check bounds. */ 1852 snprintf(rs->sc_xname, sizeof(rs->sc_xname), "raid%d", unit); 1853 1854 /* attach the pseudo device */ 1855 cf = malloc(sizeof(*cf), M_RAIDFRAME, M_WAITOK); 1856 cf->cf_name = raid_cd.cd_name; 1857 cf->cf_atname = raid_cd.cd_name; 1858 cf->cf_unit = unit; 1859 cf->cf_fstate = FSTATE_STAR; 1860 1861 rs->sc_dev = config_attach_pseudo(cf); 1862 1863 if (rs->sc_dev==NULL) { 1864 printf("raid%d: config_attach_pseudo failed\n", 1865 raidPtr->raidid); 1866 } 1867 1868 /* disk_attach actually creates space for the CPU disklabel, among 1869 * other things, so it's critical to call this *BEFORE* we try putzing 1870 * with disklabels. */ 1871 1872 disk_init(&rs->sc_dkdev, rs->sc_xname, &rf_dkdriver); 1873 disk_attach(&rs->sc_dkdev); 1874 1875 /* XXX There may be a weird interaction here between this, and 1876 * protectedSectors, as used in RAIDframe. */ 1877 1878 rs->sc_size = raidPtr->totalSectors; 1879 1880 dkwedge_discover(&rs->sc_dkdev); 1881 1882 rf_set_properties(rs, raidPtr); 1883 1884 } 1885 #if (RF_INCLUDE_PARITY_DECLUSTERING_DS > 0) 1886 /* wake up the daemon & tell it to get us a spare table 1887 * XXX 1888 * the entries in the queues should be tagged with the raidPtr 1889 * so that in the extremely rare case that two recons happen at once, 1890 * we know for which device were requesting a spare table 1891 * XXX 1892 * 1893 * XXX This code is not currently used. GO 1894 */ 1895 int 1896 rf_GetSpareTableFromDaemon(RF_SparetWait_t *req) 1897 { 1898 int retcode; 1899 1900 RF_LOCK_MUTEX(rf_sparet_wait_mutex); 1901 req->next = rf_sparet_wait_queue; 1902 rf_sparet_wait_queue = req; 1903 wakeup(&rf_sparet_wait_queue); 1904 1905 /* mpsleep unlocks the mutex */ 1906 while (!rf_sparet_resp_queue) { 1907 tsleep(&rf_sparet_resp_queue, PRIBIO, 1908 "raidframe getsparetable", 0); 1909 } 1910 req = rf_sparet_resp_queue; 1911 rf_sparet_resp_queue = req->next; 1912 RF_UNLOCK_MUTEX(rf_sparet_wait_mutex); 1913 1914 retcode = req->fcol; 1915 RF_Free(req, sizeof(*req)); /* this is not the same req as we 1916 * alloc'd */ 1917 return (retcode); 1918 } 1919 #endif 1920 1921 /* a wrapper around rf_DoAccess that extracts appropriate info from the 1922 * bp & passes it down. 1923 * any calls originating in the kernel must use non-blocking I/O 1924 * do some extra sanity checking to return "appropriate" error values for 1925 * certain conditions (to make some standard utilities work) 1926 * 1927 * Formerly known as: rf_DoAccessKernel 1928 */ 1929 void 1930 raidstart(RF_Raid_t *raidPtr) 1931 { 1932 RF_SectorCount_t num_blocks, pb, sum; 1933 RF_RaidAddr_t raid_addr; 1934 struct partition *pp; 1935 daddr_t blocknum; 1936 int unit; 1937 struct raid_softc *rs; 1938 int do_async; 1939 struct buf *bp; 1940 int rc; 1941 1942 unit = raidPtr->raidid; 1943 rs = &raid_softc[unit]; 1944 1945 /* quick check to see if anything has died recently */ 1946 RF_LOCK_MUTEX(raidPtr->mutex); 1947 if (raidPtr->numNewFailures > 0) { 1948 RF_UNLOCK_MUTEX(raidPtr->mutex); 1949 rf_update_component_labels(raidPtr, 1950 RF_NORMAL_COMPONENT_UPDATE); 1951 RF_LOCK_MUTEX(raidPtr->mutex); 1952 raidPtr->numNewFailures--; 1953 } 1954 1955 /* Check to see if we're at the limit... */ 1956 while (raidPtr->openings > 0) { 1957 RF_UNLOCK_MUTEX(raidPtr->mutex); 1958 1959 /* get the next item, if any, from the queue */ 1960 if ((bp = BUFQ_GET(rs->buf_queue)) == NULL) { 1961 /* nothing more to do */ 1962 return; 1963 } 1964 1965 /* Ok, for the bp we have here, bp->b_blkno is relative to the 1966 * partition.. Need to make it absolute to the underlying 1967 * device.. */ 1968 1969 blocknum = bp->b_blkno; 1970 if (DISKPART(bp->b_dev) != RAW_PART) { 1971 pp = &rs->sc_dkdev.dk_label->d_partitions[DISKPART(bp->b_dev)]; 1972 blocknum += pp->p_offset; 1973 } 1974 1975 db1_printf(("Blocks: %d, %d\n", (int) bp->b_blkno, 1976 (int) blocknum)); 1977 1978 db1_printf(("bp->b_bcount = %d\n", (int) bp->b_bcount)); 1979 db1_printf(("bp->b_resid = %d\n", (int) bp->b_resid)); 1980 1981 /* *THIS* is where we adjust what block we're going to... 1982 * but DO NOT TOUCH bp->b_blkno!!! */ 1983 raid_addr = blocknum; 1984 1985 num_blocks = bp->b_bcount >> raidPtr->logBytesPerSector; 1986 pb = (bp->b_bcount & raidPtr->sectorMask) ? 1 : 0; 1987 sum = raid_addr + num_blocks + pb; 1988 if (1 || rf_debugKernelAccess) { 1989 db1_printf(("raid_addr=%d sum=%d num_blocks=%d(+%d) (%d)\n", 1990 (int) raid_addr, (int) sum, (int) num_blocks, 1991 (int) pb, (int) bp->b_resid)); 1992 } 1993 if ((sum > raidPtr->totalSectors) || (sum < raid_addr) 1994 || (sum < num_blocks) || (sum < pb)) { 1995 bp->b_error = ENOSPC; 1996 bp->b_resid = bp->b_bcount; 1997 biodone(bp); 1998 RF_LOCK_MUTEX(raidPtr->mutex); 1999 continue; 2000 } 2001 /* 2002 * XXX rf_DoAccess() should do this, not just DoAccessKernel() 2003 */ 2004 2005 if (bp->b_bcount & raidPtr->sectorMask) { 2006 bp->b_error = EINVAL; 2007 bp->b_resid = bp->b_bcount; 2008 biodone(bp); 2009 RF_LOCK_MUTEX(raidPtr->mutex); 2010 continue; 2011 2012 } 2013 db1_printf(("Calling DoAccess..\n")); 2014 2015 2016 RF_LOCK_MUTEX(raidPtr->mutex); 2017 raidPtr->openings--; 2018 RF_UNLOCK_MUTEX(raidPtr->mutex); 2019 2020 /* 2021 * Everything is async. 2022 */ 2023 do_async = 1; 2024 2025 disk_busy(&rs->sc_dkdev); 2026 2027 /* XXX we're still at splbio() here... do we *really* 2028 need to be? */ 2029 2030 /* don't ever condition on bp->b_flags & B_WRITE. 2031 * always condition on B_READ instead */ 2032 2033 rc = rf_DoAccess(raidPtr, (bp->b_flags & B_READ) ? 2034 RF_IO_TYPE_READ : RF_IO_TYPE_WRITE, 2035 do_async, raid_addr, num_blocks, 2036 bp->b_data, bp, RF_DAG_NONBLOCKING_IO); 2037 2038 if (rc) { 2039 bp->b_error = rc; 2040 bp->b_resid = bp->b_bcount; 2041 biodone(bp); 2042 /* continue loop */ 2043 } 2044 2045 RF_LOCK_MUTEX(raidPtr->mutex); 2046 } 2047 RF_UNLOCK_MUTEX(raidPtr->mutex); 2048 } 2049 2050 2051 2052 2053 /* invoke an I/O from kernel mode. Disk queue should be locked upon entry */ 2054 2055 int 2056 rf_DispatchKernelIO(RF_DiskQueue_t *queue, RF_DiskQueueData_t *req) 2057 { 2058 int op = (req->type == RF_IO_TYPE_READ) ? B_READ : B_WRITE; 2059 struct buf *bp; 2060 2061 req->queue = queue; 2062 2063 #if DIAGNOSTIC 2064 if (queue->raidPtr->raidid >= numraid) { 2065 printf("Invalid unit number: %d %d\n", queue->raidPtr->raidid, 2066 numraid); 2067 panic("Invalid Unit number in rf_DispatchKernelIO"); 2068 } 2069 #endif 2070 2071 bp = req->bp; 2072 2073 switch (req->type) { 2074 case RF_IO_TYPE_NOP: /* used primarily to unlock a locked queue */ 2075 /* XXX need to do something extra here.. */ 2076 /* I'm leaving this in, as I've never actually seen it used, 2077 * and I'd like folks to report it... GO */ 2078 printf(("WAKEUP CALLED\n")); 2079 queue->numOutstanding++; 2080 2081 bp->b_flags = 0; 2082 bp->b_private = req; 2083 2084 KernelWakeupFunc(bp); 2085 break; 2086 2087 case RF_IO_TYPE_READ: 2088 case RF_IO_TYPE_WRITE: 2089 #if RF_ACC_TRACE > 0 2090 if (req->tracerec) { 2091 RF_ETIMER_START(req->tracerec->timer); 2092 } 2093 #endif 2094 InitBP(bp, queue->rf_cinfo->ci_vp, 2095 op, queue->rf_cinfo->ci_dev, 2096 req->sectorOffset, req->numSector, 2097 req->buf, KernelWakeupFunc, (void *) req, 2098 queue->raidPtr->logBytesPerSector, req->b_proc); 2099 2100 if (rf_debugKernelAccess) { 2101 db1_printf(("dispatch: bp->b_blkno = %ld\n", 2102 (long) bp->b_blkno)); 2103 } 2104 queue->numOutstanding++; 2105 queue->last_deq_sector = req->sectorOffset; 2106 /* acc wouldn't have been let in if there were any pending 2107 * reqs at any other priority */ 2108 queue->curPriority = req->priority; 2109 2110 db1_printf(("Going for %c to unit %d col %d\n", 2111 req->type, queue->raidPtr->raidid, 2112 queue->col)); 2113 db1_printf(("sector %d count %d (%d bytes) %d\n", 2114 (int) req->sectorOffset, (int) req->numSector, 2115 (int) (req->numSector << 2116 queue->raidPtr->logBytesPerSector), 2117 (int) queue->raidPtr->logBytesPerSector)); 2118 VOP_STRATEGY(bp->b_vp, bp); 2119 2120 break; 2121 2122 default: 2123 panic("bad req->type in rf_DispatchKernelIO"); 2124 } 2125 db1_printf(("Exiting from DispatchKernelIO\n")); 2126 2127 return (0); 2128 } 2129 /* this is the callback function associated with a I/O invoked from 2130 kernel code. 2131 */ 2132 static void 2133 KernelWakeupFunc(struct buf *bp) 2134 { 2135 RF_DiskQueueData_t *req = NULL; 2136 RF_DiskQueue_t *queue; 2137 int s; 2138 2139 s = splbio(); 2140 db1_printf(("recovering the request queue:\n")); 2141 req = bp->b_private; 2142 2143 queue = (RF_DiskQueue_t *) req->queue; 2144 2145 #if RF_ACC_TRACE > 0 2146 if (req->tracerec) { 2147 RF_ETIMER_STOP(req->tracerec->timer); 2148 RF_ETIMER_EVAL(req->tracerec->timer); 2149 RF_LOCK_MUTEX(rf_tracing_mutex); 2150 req->tracerec->diskwait_us += RF_ETIMER_VAL_US(req->tracerec->timer); 2151 req->tracerec->phys_io_us += RF_ETIMER_VAL_US(req->tracerec->timer); 2152 req->tracerec->num_phys_ios++; 2153 RF_UNLOCK_MUTEX(rf_tracing_mutex); 2154 } 2155 #endif 2156 2157 /* XXX Ok, let's get aggressive... If b_error is set, let's go 2158 * ballistic, and mark the component as hosed... */ 2159 2160 if (bp->b_error != 0) { 2161 /* Mark the disk as dead */ 2162 /* but only mark it once... */ 2163 /* and only if it wouldn't leave this RAID set 2164 completely broken */ 2165 if (((queue->raidPtr->Disks[queue->col].status == 2166 rf_ds_optimal) || 2167 (queue->raidPtr->Disks[queue->col].status == 2168 rf_ds_used_spare)) && 2169 (queue->raidPtr->numFailures < 2170 queue->raidPtr->Layout.map->faultsTolerated)) { 2171 printf("raid%d: IO Error. Marking %s as failed.\n", 2172 queue->raidPtr->raidid, 2173 queue->raidPtr->Disks[queue->col].devname); 2174 queue->raidPtr->Disks[queue->col].status = 2175 rf_ds_failed; 2176 queue->raidPtr->status = rf_rs_degraded; 2177 queue->raidPtr->numFailures++; 2178 queue->raidPtr->numNewFailures++; 2179 } else { /* Disk is already dead... */ 2180 /* printf("Disk already marked as dead!\n"); */ 2181 } 2182 2183 } 2184 2185 /* Fill in the error value */ 2186 2187 req->error = bp->b_error; 2188 2189 simple_lock(&queue->raidPtr->iodone_lock); 2190 2191 /* Drop this one on the "finished" queue... */ 2192 TAILQ_INSERT_TAIL(&(queue->raidPtr->iodone), req, iodone_entries); 2193 2194 /* Let the raidio thread know there is work to be done. */ 2195 wakeup(&(queue->raidPtr->iodone)); 2196 2197 simple_unlock(&queue->raidPtr->iodone_lock); 2198 2199 splx(s); 2200 } 2201 2202 2203 2204 /* 2205 * initialize a buf structure for doing an I/O in the kernel. 2206 */ 2207 static void 2208 InitBP(struct buf *bp, struct vnode *b_vp, unsigned rw_flag, dev_t dev, 2209 RF_SectorNum_t startSect, RF_SectorCount_t numSect, void *bf, 2210 void (*cbFunc) (struct buf *), void *cbArg, int logBytesPerSector, 2211 struct proc *b_proc) 2212 { 2213 /* bp->b_flags = B_PHYS | rw_flag; */ 2214 bp->b_flags = B_CALL | rw_flag; /* XXX need B_PHYS here too??? */ 2215 bp->b_bcount = numSect << logBytesPerSector; 2216 bp->b_bufsize = bp->b_bcount; 2217 bp->b_error = 0; 2218 bp->b_dev = dev; 2219 bp->b_data = bf; 2220 bp->b_blkno = startSect; 2221 bp->b_resid = bp->b_bcount; /* XXX is this right!??!?!! */ 2222 if (bp->b_bcount == 0) { 2223 panic("bp->b_bcount is zero in InitBP!!"); 2224 } 2225 bp->b_proc = b_proc; 2226 bp->b_iodone = cbFunc; 2227 bp->b_private = cbArg; 2228 bp->b_vp = b_vp; 2229 if ((bp->b_flags & B_READ) == 0) { 2230 bp->b_vp->v_numoutput++; 2231 } 2232 2233 } 2234 2235 static void 2236 raidgetdefaultlabel(RF_Raid_t *raidPtr, struct raid_softc *rs, 2237 struct disklabel *lp) 2238 { 2239 memset(lp, 0, sizeof(*lp)); 2240 2241 /* fabricate a label... */ 2242 lp->d_secperunit = raidPtr->totalSectors; 2243 lp->d_secsize = raidPtr->bytesPerSector; 2244 lp->d_nsectors = raidPtr->Layout.dataSectorsPerStripe; 2245 lp->d_ntracks = 4 * raidPtr->numCol; 2246 lp->d_ncylinders = raidPtr->totalSectors / 2247 (lp->d_nsectors * lp->d_ntracks); 2248 lp->d_secpercyl = lp->d_ntracks * lp->d_nsectors; 2249 2250 strncpy(lp->d_typename, "raid", sizeof(lp->d_typename)); 2251 lp->d_type = DTYPE_RAID; 2252 strncpy(lp->d_packname, "fictitious", sizeof(lp->d_packname)); 2253 lp->d_rpm = 3600; 2254 lp->d_interleave = 1; 2255 lp->d_flags = 0; 2256 2257 lp->d_partitions[RAW_PART].p_offset = 0; 2258 lp->d_partitions[RAW_PART].p_size = raidPtr->totalSectors; 2259 lp->d_partitions[RAW_PART].p_fstype = FS_UNUSED; 2260 lp->d_npartitions = RAW_PART + 1; 2261 2262 lp->d_magic = DISKMAGIC; 2263 lp->d_magic2 = DISKMAGIC; 2264 lp->d_checksum = dkcksum(rs->sc_dkdev.dk_label); 2265 2266 } 2267 /* 2268 * Read the disklabel from the raid device. If one is not present, fake one 2269 * up. 2270 */ 2271 static void 2272 raidgetdisklabel(dev_t dev) 2273 { 2274 int unit = raidunit(dev); 2275 struct raid_softc *rs = &raid_softc[unit]; 2276 const char *errstring; 2277 struct disklabel *lp = rs->sc_dkdev.dk_label; 2278 struct cpu_disklabel *clp = rs->sc_dkdev.dk_cpulabel; 2279 RF_Raid_t *raidPtr; 2280 2281 db1_printf(("Getting the disklabel...\n")); 2282 2283 memset(clp, 0, sizeof(*clp)); 2284 2285 raidPtr = raidPtrs[unit]; 2286 2287 raidgetdefaultlabel(raidPtr, rs, lp); 2288 2289 /* 2290 * Call the generic disklabel extraction routine. 2291 */ 2292 errstring = readdisklabel(RAIDLABELDEV(dev), raidstrategy, 2293 rs->sc_dkdev.dk_label, rs->sc_dkdev.dk_cpulabel); 2294 if (errstring) 2295 raidmakedisklabel(rs); 2296 else { 2297 int i; 2298 struct partition *pp; 2299 2300 /* 2301 * Sanity check whether the found disklabel is valid. 2302 * 2303 * This is necessary since total size of the raid device 2304 * may vary when an interleave is changed even though exactly 2305 * same components are used, and old disklabel may used 2306 * if that is found. 2307 */ 2308 if (lp->d_secperunit != rs->sc_size) 2309 printf("raid%d: WARNING: %s: " 2310 "total sector size in disklabel (%d) != " 2311 "the size of raid (%ld)\n", unit, rs->sc_xname, 2312 lp->d_secperunit, (long) rs->sc_size); 2313 for (i = 0; i < lp->d_npartitions; i++) { 2314 pp = &lp->d_partitions[i]; 2315 if (pp->p_offset + pp->p_size > rs->sc_size) 2316 printf("raid%d: WARNING: %s: end of partition `%c' " 2317 "exceeds the size of raid (%ld)\n", 2318 unit, rs->sc_xname, 'a' + i, (long) rs->sc_size); 2319 } 2320 } 2321 2322 } 2323 /* 2324 * Take care of things one might want to take care of in the event 2325 * that a disklabel isn't present. 2326 */ 2327 static void 2328 raidmakedisklabel(struct raid_softc *rs) 2329 { 2330 struct disklabel *lp = rs->sc_dkdev.dk_label; 2331 db1_printf(("Making a label..\n")); 2332 2333 /* 2334 * For historical reasons, if there's no disklabel present 2335 * the raw partition must be marked FS_BSDFFS. 2336 */ 2337 2338 lp->d_partitions[RAW_PART].p_fstype = FS_BSDFFS; 2339 2340 strncpy(lp->d_packname, "default label", sizeof(lp->d_packname)); 2341 2342 lp->d_checksum = dkcksum(lp); 2343 } 2344 /* 2345 * Wait interruptibly for an exclusive lock. 2346 * 2347 * XXX 2348 * Several drivers do this; it should be abstracted and made MP-safe. 2349 * (Hmm... where have we seen this warning before :-> GO ) 2350 */ 2351 static int 2352 raidlock(struct raid_softc *rs) 2353 { 2354 int error; 2355 2356 while ((rs->sc_flags & RAIDF_LOCKED) != 0) { 2357 rs->sc_flags |= RAIDF_WANTED; 2358 if ((error = 2359 tsleep(rs, PRIBIO | PCATCH, "raidlck", 0)) != 0) 2360 return (error); 2361 } 2362 rs->sc_flags |= RAIDF_LOCKED; 2363 return (0); 2364 } 2365 /* 2366 * Unlock and wake up any waiters. 2367 */ 2368 static void 2369 raidunlock(struct raid_softc *rs) 2370 { 2371 2372 rs->sc_flags &= ~RAIDF_LOCKED; 2373 if ((rs->sc_flags & RAIDF_WANTED) != 0) { 2374 rs->sc_flags &= ~RAIDF_WANTED; 2375 wakeup(rs); 2376 } 2377 } 2378 2379 2380 #define RF_COMPONENT_INFO_OFFSET 16384 /* bytes */ 2381 #define RF_COMPONENT_INFO_SIZE 1024 /* bytes */ 2382 2383 int 2384 raidmarkclean(dev_t dev, struct vnode *b_vp, int mod_counter) 2385 { 2386 RF_ComponentLabel_t clabel; 2387 raidread_component_label(dev, b_vp, &clabel); 2388 clabel.mod_counter = mod_counter; 2389 clabel.clean = RF_RAID_CLEAN; 2390 raidwrite_component_label(dev, b_vp, &clabel); 2391 return(0); 2392 } 2393 2394 2395 int 2396 raidmarkdirty(dev_t dev, struct vnode *b_vp, int mod_counter) 2397 { 2398 RF_ComponentLabel_t clabel; 2399 raidread_component_label(dev, b_vp, &clabel); 2400 clabel.mod_counter = mod_counter; 2401 clabel.clean = RF_RAID_DIRTY; 2402 raidwrite_component_label(dev, b_vp, &clabel); 2403 return(0); 2404 } 2405 2406 /* ARGSUSED */ 2407 int 2408 raidread_component_label(dev_t dev, struct vnode *b_vp, 2409 RF_ComponentLabel_t *clabel) 2410 { 2411 struct buf *bp; 2412 const struct bdevsw *bdev; 2413 int error; 2414 2415 /* XXX should probably ensure that we don't try to do this if 2416 someone has changed rf_protected_sectors. */ 2417 2418 if (b_vp == NULL) { 2419 /* For whatever reason, this component is not valid. 2420 Don't try to read a component label from it. */ 2421 return(EINVAL); 2422 } 2423 2424 /* get a block of the appropriate size... */ 2425 bp = geteblk((int)RF_COMPONENT_INFO_SIZE); 2426 bp->b_dev = dev; 2427 2428 /* get our ducks in a row for the read */ 2429 bp->b_blkno = RF_COMPONENT_INFO_OFFSET / DEV_BSIZE; 2430 bp->b_bcount = RF_COMPONENT_INFO_SIZE; 2431 bp->b_flags |= B_READ; 2432 bp->b_resid = RF_COMPONENT_INFO_SIZE / DEV_BSIZE; 2433 2434 bdev = bdevsw_lookup(bp->b_dev); 2435 if (bdev == NULL) 2436 return (ENXIO); 2437 (*bdev->d_strategy)(bp); 2438 2439 error = biowait(bp); 2440 2441 if (!error) { 2442 memcpy(clabel, bp->b_data, 2443 sizeof(RF_ComponentLabel_t)); 2444 } 2445 2446 brelse(bp, 0); 2447 return(error); 2448 } 2449 /* ARGSUSED */ 2450 int 2451 raidwrite_component_label(dev_t dev, struct vnode *b_vp, 2452 RF_ComponentLabel_t *clabel) 2453 { 2454 struct buf *bp; 2455 const struct bdevsw *bdev; 2456 int error; 2457 2458 /* get a block of the appropriate size... */ 2459 bp = geteblk((int)RF_COMPONENT_INFO_SIZE); 2460 bp->b_dev = dev; 2461 2462 /* get our ducks in a row for the write */ 2463 bp->b_blkno = RF_COMPONENT_INFO_OFFSET / DEV_BSIZE; 2464 bp->b_bcount = RF_COMPONENT_INFO_SIZE; 2465 bp->b_flags |= B_WRITE; 2466 bp->b_resid = RF_COMPONENT_INFO_SIZE / DEV_BSIZE; 2467 2468 memset(bp->b_data, 0, RF_COMPONENT_INFO_SIZE ); 2469 2470 memcpy(bp->b_data, clabel, sizeof(RF_ComponentLabel_t)); 2471 2472 bdev = bdevsw_lookup(bp->b_dev); 2473 if (bdev == NULL) 2474 return (ENXIO); 2475 (*bdev->d_strategy)(bp); 2476 error = biowait(bp); 2477 brelse(bp, 0); 2478 if (error) { 2479 #if 1 2480 printf("Failed to write RAID component info!\n"); 2481 #endif 2482 } 2483 2484 return(error); 2485 } 2486 2487 void 2488 rf_markalldirty(RF_Raid_t *raidPtr) 2489 { 2490 RF_ComponentLabel_t clabel; 2491 int sparecol; 2492 int c; 2493 int j; 2494 int scol = -1; 2495 2496 raidPtr->mod_counter++; 2497 for (c = 0; c < raidPtr->numCol; c++) { 2498 /* we don't want to touch (at all) a disk that has 2499 failed */ 2500 if (!RF_DEAD_DISK(raidPtr->Disks[c].status)) { 2501 raidread_component_label( 2502 raidPtr->Disks[c].dev, 2503 raidPtr->raid_cinfo[c].ci_vp, 2504 &clabel); 2505 if (clabel.status == rf_ds_spared) { 2506 /* XXX do something special... 2507 but whatever you do, don't 2508 try to access it!! */ 2509 } else { 2510 raidmarkdirty( 2511 raidPtr->Disks[c].dev, 2512 raidPtr->raid_cinfo[c].ci_vp, 2513 raidPtr->mod_counter); 2514 } 2515 } 2516 } 2517 2518 for( c = 0; c < raidPtr->numSpare ; c++) { 2519 sparecol = raidPtr->numCol + c; 2520 if (raidPtr->Disks[sparecol].status == rf_ds_used_spare) { 2521 /* 2522 2523 we claim this disk is "optimal" if it's 2524 rf_ds_used_spare, as that means it should be 2525 directly substitutable for the disk it replaced. 2526 We note that too... 2527 2528 */ 2529 2530 for(j=0;j<raidPtr->numCol;j++) { 2531 if (raidPtr->Disks[j].spareCol == sparecol) { 2532 scol = j; 2533 break; 2534 } 2535 } 2536 2537 raidread_component_label( 2538 raidPtr->Disks[sparecol].dev, 2539 raidPtr->raid_cinfo[sparecol].ci_vp, 2540 &clabel); 2541 /* make sure status is noted */ 2542 2543 raid_init_component_label(raidPtr, &clabel); 2544 2545 clabel.row = 0; 2546 clabel.column = scol; 2547 /* Note: we *don't* change status from rf_ds_used_spare 2548 to rf_ds_optimal */ 2549 /* clabel.status = rf_ds_optimal; */ 2550 2551 raidmarkdirty(raidPtr->Disks[sparecol].dev, 2552 raidPtr->raid_cinfo[sparecol].ci_vp, 2553 raidPtr->mod_counter); 2554 } 2555 } 2556 } 2557 2558 2559 void 2560 rf_update_component_labels(RF_Raid_t *raidPtr, int final) 2561 { 2562 RF_ComponentLabel_t clabel; 2563 int sparecol; 2564 int c; 2565 int j; 2566 int scol; 2567 2568 scol = -1; 2569 2570 /* XXX should do extra checks to make sure things really are clean, 2571 rather than blindly setting the clean bit... */ 2572 2573 raidPtr->mod_counter++; 2574 2575 for (c = 0; c < raidPtr->numCol; c++) { 2576 if (raidPtr->Disks[c].status == rf_ds_optimal) { 2577 raidread_component_label( 2578 raidPtr->Disks[c].dev, 2579 raidPtr->raid_cinfo[c].ci_vp, 2580 &clabel); 2581 /* make sure status is noted */ 2582 clabel.status = rf_ds_optimal; 2583 2584 /* bump the counter */ 2585 clabel.mod_counter = raidPtr->mod_counter; 2586 2587 /* note what unit we are configured as */ 2588 clabel.last_unit = raidPtr->raidid; 2589 2590 raidwrite_component_label( 2591 raidPtr->Disks[c].dev, 2592 raidPtr->raid_cinfo[c].ci_vp, 2593 &clabel); 2594 if (final == RF_FINAL_COMPONENT_UPDATE) { 2595 if (raidPtr->parity_good == RF_RAID_CLEAN) { 2596 raidmarkclean( 2597 raidPtr->Disks[c].dev, 2598 raidPtr->raid_cinfo[c].ci_vp, 2599 raidPtr->mod_counter); 2600 } 2601 } 2602 } 2603 /* else we don't touch it.. */ 2604 } 2605 2606 for( c = 0; c < raidPtr->numSpare ; c++) { 2607 sparecol = raidPtr->numCol + c; 2608 /* Need to ensure that the reconstruct actually completed! */ 2609 if (raidPtr->Disks[sparecol].status == rf_ds_used_spare) { 2610 /* 2611 2612 we claim this disk is "optimal" if it's 2613 rf_ds_used_spare, as that means it should be 2614 directly substitutable for the disk it replaced. 2615 We note that too... 2616 2617 */ 2618 2619 for(j=0;j<raidPtr->numCol;j++) { 2620 if (raidPtr->Disks[j].spareCol == sparecol) { 2621 scol = j; 2622 break; 2623 } 2624 } 2625 2626 /* XXX shouldn't *really* need this... */ 2627 raidread_component_label( 2628 raidPtr->Disks[sparecol].dev, 2629 raidPtr->raid_cinfo[sparecol].ci_vp, 2630 &clabel); 2631 /* make sure status is noted */ 2632 2633 raid_init_component_label(raidPtr, &clabel); 2634 2635 clabel.mod_counter = raidPtr->mod_counter; 2636 clabel.column = scol; 2637 clabel.status = rf_ds_optimal; 2638 clabel.last_unit = raidPtr->raidid; 2639 2640 raidwrite_component_label( 2641 raidPtr->Disks[sparecol].dev, 2642 raidPtr->raid_cinfo[sparecol].ci_vp, 2643 &clabel); 2644 if (final == RF_FINAL_COMPONENT_UPDATE) { 2645 if (raidPtr->parity_good == RF_RAID_CLEAN) { 2646 raidmarkclean( raidPtr->Disks[sparecol].dev, 2647 raidPtr->raid_cinfo[sparecol].ci_vp, 2648 raidPtr->mod_counter); 2649 } 2650 } 2651 } 2652 } 2653 } 2654 2655 void 2656 rf_close_component(RF_Raid_t *raidPtr, struct vnode *vp, int auto_configured) 2657 { 2658 2659 if (vp != NULL) { 2660 if (auto_configured == 1) { 2661 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 2662 VOP_CLOSE(vp, FREAD | FWRITE, NOCRED, 0); 2663 vput(vp); 2664 2665 } else { 2666 (void) vn_close(vp, FREAD | FWRITE, curlwp->l_cred, curlwp); 2667 } 2668 } 2669 } 2670 2671 2672 void 2673 rf_UnconfigureVnodes(RF_Raid_t *raidPtr) 2674 { 2675 int r,c; 2676 struct vnode *vp; 2677 int acd; 2678 2679 2680 /* We take this opportunity to close the vnodes like we should.. */ 2681 2682 for (c = 0; c < raidPtr->numCol; c++) { 2683 vp = raidPtr->raid_cinfo[c].ci_vp; 2684 acd = raidPtr->Disks[c].auto_configured; 2685 rf_close_component(raidPtr, vp, acd); 2686 raidPtr->raid_cinfo[c].ci_vp = NULL; 2687 raidPtr->Disks[c].auto_configured = 0; 2688 } 2689 2690 for (r = 0; r < raidPtr->numSpare; r++) { 2691 vp = raidPtr->raid_cinfo[raidPtr->numCol + r].ci_vp; 2692 acd = raidPtr->Disks[raidPtr->numCol + r].auto_configured; 2693 rf_close_component(raidPtr, vp, acd); 2694 raidPtr->raid_cinfo[raidPtr->numCol + r].ci_vp = NULL; 2695 raidPtr->Disks[raidPtr->numCol + r].auto_configured = 0; 2696 } 2697 } 2698 2699 2700 void 2701 rf_ReconThread(struct rf_recon_req *req) 2702 { 2703 int s; 2704 RF_Raid_t *raidPtr; 2705 2706 s = splbio(); 2707 raidPtr = (RF_Raid_t *) req->raidPtr; 2708 raidPtr->recon_in_progress = 1; 2709 2710 rf_FailDisk((RF_Raid_t *) req->raidPtr, req->col, 2711 ((req->flags & RF_FDFLAGS_RECON) ? 1 : 0)); 2712 2713 RF_Free(req, sizeof(*req)); 2714 2715 raidPtr->recon_in_progress = 0; 2716 splx(s); 2717 2718 /* That's all... */ 2719 kthread_exit(0); /* does not return */ 2720 } 2721 2722 void 2723 rf_RewriteParityThread(RF_Raid_t *raidPtr) 2724 { 2725 int retcode; 2726 int s; 2727 2728 raidPtr->parity_rewrite_stripes_done = 0; 2729 raidPtr->parity_rewrite_in_progress = 1; 2730 s = splbio(); 2731 retcode = rf_RewriteParity(raidPtr); 2732 splx(s); 2733 if (retcode) { 2734 printf("raid%d: Error re-writing parity!\n",raidPtr->raidid); 2735 } else { 2736 /* set the clean bit! If we shutdown correctly, 2737 the clean bit on each component label will get 2738 set */ 2739 raidPtr->parity_good = RF_RAID_CLEAN; 2740 } 2741 raidPtr->parity_rewrite_in_progress = 0; 2742 2743 /* Anyone waiting for us to stop? If so, inform them... */ 2744 if (raidPtr->waitShutdown) { 2745 wakeup(&raidPtr->parity_rewrite_in_progress); 2746 } 2747 2748 /* That's all... */ 2749 kthread_exit(0); /* does not return */ 2750 } 2751 2752 2753 void 2754 rf_CopybackThread(RF_Raid_t *raidPtr) 2755 { 2756 int s; 2757 2758 raidPtr->copyback_in_progress = 1; 2759 s = splbio(); 2760 rf_CopybackReconstructedData(raidPtr); 2761 splx(s); 2762 raidPtr->copyback_in_progress = 0; 2763 2764 /* That's all... */ 2765 kthread_exit(0); /* does not return */ 2766 } 2767 2768 2769 void 2770 rf_ReconstructInPlaceThread(struct rf_recon_req *req) 2771 { 2772 int s; 2773 RF_Raid_t *raidPtr; 2774 2775 s = splbio(); 2776 raidPtr = req->raidPtr; 2777 raidPtr->recon_in_progress = 1; 2778 rf_ReconstructInPlace(raidPtr, req->col); 2779 RF_Free(req, sizeof(*req)); 2780 raidPtr->recon_in_progress = 0; 2781 splx(s); 2782 2783 /* That's all... */ 2784 kthread_exit(0); /* does not return */ 2785 } 2786 2787 static RF_AutoConfig_t * 2788 rf_get_component(RF_AutoConfig_t *ac_list, dev_t dev, struct vnode *vp, 2789 const char *cname, RF_SectorCount_t size) 2790 { 2791 int good_one = 0; 2792 RF_ComponentLabel_t *clabel; 2793 RF_AutoConfig_t *ac; 2794 2795 clabel = malloc(sizeof(RF_ComponentLabel_t), M_RAIDFRAME, M_NOWAIT); 2796 if (clabel == NULL) { 2797 oomem: 2798 while(ac_list) { 2799 ac = ac_list; 2800 if (ac->clabel) 2801 free(ac->clabel, M_RAIDFRAME); 2802 ac_list = ac_list->next; 2803 free(ac, M_RAIDFRAME); 2804 } 2805 printf("RAID auto config: out of memory!\n"); 2806 return NULL; /* XXX probably should panic? */ 2807 } 2808 2809 if (!raidread_component_label(dev, vp, clabel)) { 2810 /* Got the label. Does it look reasonable? */ 2811 if (rf_reasonable_label(clabel) && 2812 (clabel->partitionSize <= size)) { 2813 #ifdef DEBUG 2814 printf("Component on: %s: %llu\n", 2815 cname, (unsigned long long)size); 2816 rf_print_component_label(clabel); 2817 #endif 2818 /* if it's reasonable, add it, else ignore it. */ 2819 ac = malloc(sizeof(RF_AutoConfig_t), M_RAIDFRAME, 2820 M_NOWAIT); 2821 if (ac == NULL) { 2822 free(clabel, M_RAIDFRAME); 2823 goto oomem; 2824 } 2825 strlcpy(ac->devname, cname, sizeof(ac->devname)); 2826 ac->dev = dev; 2827 ac->vp = vp; 2828 ac->clabel = clabel; 2829 ac->next = ac_list; 2830 ac_list = ac; 2831 good_one = 1; 2832 } 2833 } 2834 if (!good_one) { 2835 /* cleanup */ 2836 free(clabel, M_RAIDFRAME); 2837 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 2838 VOP_CLOSE(vp, FREAD | FWRITE, NOCRED, 0); 2839 vput(vp); 2840 } 2841 return ac_list; 2842 } 2843 2844 RF_AutoConfig_t * 2845 rf_find_raid_components() 2846 { 2847 struct vnode *vp; 2848 struct disklabel label; 2849 struct device *dv; 2850 dev_t dev; 2851 int bmajor, bminor, wedge; 2852 int error; 2853 int i; 2854 RF_AutoConfig_t *ac_list; 2855 2856 2857 /* initialize the AutoConfig list */ 2858 ac_list = NULL; 2859 2860 /* we begin by trolling through *all* the devices on the system */ 2861 2862 for (dv = alldevs.tqh_first; dv != NULL; 2863 dv = dv->dv_list.tqe_next) { 2864 2865 /* we are only interested in disks... */ 2866 if (device_class(dv) != DV_DISK) 2867 continue; 2868 2869 /* we don't care about floppies... */ 2870 if (device_is_a(dv, "fd")) { 2871 continue; 2872 } 2873 2874 /* we don't care about CD's... */ 2875 if (device_is_a(dv, "cd")) { 2876 continue; 2877 } 2878 2879 /* hdfd is the Atari/Hades floppy driver */ 2880 if (device_is_a(dv, "hdfd")) { 2881 continue; 2882 } 2883 2884 /* fdisa is the Atari/Milan floppy driver */ 2885 if (device_is_a(dv, "fdisa")) { 2886 continue; 2887 } 2888 2889 /* need to find the device_name_to_block_device_major stuff */ 2890 bmajor = devsw_name2blk(dv->dv_xname, NULL, 0); 2891 2892 /* get a vnode for the raw partition of this disk */ 2893 2894 wedge = device_is_a(dv, "dk"); 2895 bminor = minor(device_unit(dv)); 2896 dev = wedge ? makedev(bmajor, bminor) : 2897 MAKEDISKDEV(bmajor, bminor, RAW_PART); 2898 if (bdevvp(dev, &vp)) 2899 panic("RAID can't alloc vnode"); 2900 2901 error = VOP_OPEN(vp, FREAD, NOCRED, 0); 2902 2903 if (error) { 2904 /* "Who cares." Continue looking 2905 for something that exists*/ 2906 vput(vp); 2907 continue; 2908 } 2909 2910 if (wedge) { 2911 struct dkwedge_info dkw; 2912 error = VOP_IOCTL(vp, DIOCGWEDGEINFO, &dkw, FREAD, 2913 NOCRED, 0); 2914 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 2915 VOP_CLOSE(vp, FREAD | FWRITE, NOCRED, 0); 2916 vput(vp); 2917 if (error) { 2918 printf("RAIDframe: can't get wedge info for " 2919 "dev %s (%d)\n", dv->dv_xname, error); 2920 continue; 2921 } 2922 2923 if (strcmp(dkw.dkw_ptype, DKW_PTYPE_RAIDFRAME) != 0) 2924 continue; 2925 2926 ac_list = rf_get_component(ac_list, dev, vp, 2927 dv->dv_xname, dkw.dkw_size); 2928 continue; 2929 } 2930 2931 /* Ok, the disk exists. Go get the disklabel. */ 2932 error = VOP_IOCTL(vp, DIOCGDINFO, &label, FREAD, NOCRED, 0); 2933 if (error) { 2934 /* 2935 * XXX can't happen - open() would 2936 * have errored out (or faked up one) 2937 */ 2938 if (error != ENOTTY) 2939 printf("RAIDframe: can't get label for dev " 2940 "%s (%d)\n", dv->dv_xname, error); 2941 } 2942 2943 /* don't need this any more. We'll allocate it again 2944 a little later if we really do... */ 2945 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 2946 VOP_CLOSE(vp, FREAD | FWRITE, NOCRED, 0); 2947 vput(vp); 2948 2949 if (error) 2950 continue; 2951 2952 for (i = 0; i < label.d_npartitions; i++) { 2953 char cname[sizeof(ac_list->devname)]; 2954 2955 /* We only support partitions marked as RAID */ 2956 if (label.d_partitions[i].p_fstype != FS_RAID) 2957 continue; 2958 2959 dev = MAKEDISKDEV(bmajor, device_unit(dv), i); 2960 if (bdevvp(dev, &vp)) 2961 panic("RAID can't alloc vnode"); 2962 2963 error = VOP_OPEN(vp, FREAD, NOCRED, 0); 2964 if (error) { 2965 /* Whatever... */ 2966 vput(vp); 2967 continue; 2968 } 2969 snprintf(cname, sizeof(cname), "%s%c", 2970 dv->dv_xname, 'a' + i); 2971 ac_list = rf_get_component(ac_list, dev, vp, cname, 2972 label.d_partitions[i].p_size); 2973 } 2974 } 2975 return ac_list; 2976 } 2977 2978 2979 static int 2980 rf_reasonable_label(RF_ComponentLabel_t *clabel) 2981 { 2982 2983 if (((clabel->version==RF_COMPONENT_LABEL_VERSION_1) || 2984 (clabel->version==RF_COMPONENT_LABEL_VERSION)) && 2985 ((clabel->clean == RF_RAID_CLEAN) || 2986 (clabel->clean == RF_RAID_DIRTY)) && 2987 clabel->row >=0 && 2988 clabel->column >= 0 && 2989 clabel->num_rows > 0 && 2990 clabel->num_columns > 0 && 2991 clabel->row < clabel->num_rows && 2992 clabel->column < clabel->num_columns && 2993 clabel->blockSize > 0 && 2994 clabel->numBlocks > 0) { 2995 /* label looks reasonable enough... */ 2996 return(1); 2997 } 2998 return(0); 2999 } 3000 3001 3002 #ifdef DEBUG 3003 void 3004 rf_print_component_label(RF_ComponentLabel_t *clabel) 3005 { 3006 printf(" Row: %d Column: %d Num Rows: %d Num Columns: %d\n", 3007 clabel->row, clabel->column, 3008 clabel->num_rows, clabel->num_columns); 3009 printf(" Version: %d Serial Number: %d Mod Counter: %d\n", 3010 clabel->version, clabel->serial_number, 3011 clabel->mod_counter); 3012 printf(" Clean: %s Status: %d\n", 3013 clabel->clean ? "Yes" : "No", clabel->status ); 3014 printf(" sectPerSU: %d SUsPerPU: %d SUsPerRU: %d\n", 3015 clabel->sectPerSU, clabel->SUsPerPU, clabel->SUsPerRU); 3016 printf(" RAID Level: %c blocksize: %d numBlocks: %d\n", 3017 (char) clabel->parityConfig, clabel->blockSize, 3018 clabel->numBlocks); 3019 printf(" Autoconfig: %s\n", clabel->autoconfigure ? "Yes" : "No" ); 3020 printf(" Contains root partition: %s\n", 3021 clabel->root_partition ? "Yes" : "No" ); 3022 printf(" Last configured as: raid%d\n", clabel->last_unit ); 3023 #if 0 3024 printf(" Config order: %d\n", clabel->config_order); 3025 #endif 3026 3027 } 3028 #endif 3029 3030 RF_ConfigSet_t * 3031 rf_create_auto_sets(RF_AutoConfig_t *ac_list) 3032 { 3033 RF_AutoConfig_t *ac; 3034 RF_ConfigSet_t *config_sets; 3035 RF_ConfigSet_t *cset; 3036 RF_AutoConfig_t *ac_next; 3037 3038 3039 config_sets = NULL; 3040 3041 /* Go through the AutoConfig list, and figure out which components 3042 belong to what sets. */ 3043 ac = ac_list; 3044 while(ac!=NULL) { 3045 /* we're going to putz with ac->next, so save it here 3046 for use at the end of the loop */ 3047 ac_next = ac->next; 3048 3049 if (config_sets == NULL) { 3050 /* will need at least this one... */ 3051 config_sets = (RF_ConfigSet_t *) 3052 malloc(sizeof(RF_ConfigSet_t), 3053 M_RAIDFRAME, M_NOWAIT); 3054 if (config_sets == NULL) { 3055 panic("rf_create_auto_sets: No memory!"); 3056 } 3057 /* this one is easy :) */ 3058 config_sets->ac = ac; 3059 config_sets->next = NULL; 3060 config_sets->rootable = 0; 3061 ac->next = NULL; 3062 } else { 3063 /* which set does this component fit into? */ 3064 cset = config_sets; 3065 while(cset!=NULL) { 3066 if (rf_does_it_fit(cset, ac)) { 3067 /* looks like it matches... */ 3068 ac->next = cset->ac; 3069 cset->ac = ac; 3070 break; 3071 } 3072 cset = cset->next; 3073 } 3074 if (cset==NULL) { 3075 /* didn't find a match above... new set..*/ 3076 cset = (RF_ConfigSet_t *) 3077 malloc(sizeof(RF_ConfigSet_t), 3078 M_RAIDFRAME, M_NOWAIT); 3079 if (cset == NULL) { 3080 panic("rf_create_auto_sets: No memory!"); 3081 } 3082 cset->ac = ac; 3083 ac->next = NULL; 3084 cset->next = config_sets; 3085 cset->rootable = 0; 3086 config_sets = cset; 3087 } 3088 } 3089 ac = ac_next; 3090 } 3091 3092 3093 return(config_sets); 3094 } 3095 3096 static int 3097 rf_does_it_fit(RF_ConfigSet_t *cset, RF_AutoConfig_t *ac) 3098 { 3099 RF_ComponentLabel_t *clabel1, *clabel2; 3100 3101 /* If this one matches the *first* one in the set, that's good 3102 enough, since the other members of the set would have been 3103 through here too... */ 3104 /* note that we are not checking partitionSize here.. 3105 3106 Note that we are also not checking the mod_counters here. 3107 If everything else matches execpt the mod_counter, that's 3108 good enough for this test. We will deal with the mod_counters 3109 a little later in the autoconfiguration process. 3110 3111 (clabel1->mod_counter == clabel2->mod_counter) && 3112 3113 The reason we don't check for this is that failed disks 3114 will have lower modification counts. If those disks are 3115 not added to the set they used to belong to, then they will 3116 form their own set, which may result in 2 different sets, 3117 for example, competing to be configured at raid0, and 3118 perhaps competing to be the root filesystem set. If the 3119 wrong ones get configured, or both attempt to become /, 3120 weird behaviour and or serious lossage will occur. Thus we 3121 need to bring them into the fold here, and kick them out at 3122 a later point. 3123 3124 */ 3125 3126 clabel1 = cset->ac->clabel; 3127 clabel2 = ac->clabel; 3128 if ((clabel1->version == clabel2->version) && 3129 (clabel1->serial_number == clabel2->serial_number) && 3130 (clabel1->num_rows == clabel2->num_rows) && 3131 (clabel1->num_columns == clabel2->num_columns) && 3132 (clabel1->sectPerSU == clabel2->sectPerSU) && 3133 (clabel1->SUsPerPU == clabel2->SUsPerPU) && 3134 (clabel1->SUsPerRU == clabel2->SUsPerRU) && 3135 (clabel1->parityConfig == clabel2->parityConfig) && 3136 (clabel1->maxOutstanding == clabel2->maxOutstanding) && 3137 (clabel1->blockSize == clabel2->blockSize) && 3138 (clabel1->numBlocks == clabel2->numBlocks) && 3139 (clabel1->autoconfigure == clabel2->autoconfigure) && 3140 (clabel1->root_partition == clabel2->root_partition) && 3141 (clabel1->last_unit == clabel2->last_unit) && 3142 (clabel1->config_order == clabel2->config_order)) { 3143 /* if it get's here, it almost *has* to be a match */ 3144 } else { 3145 /* it's not consistent with somebody in the set.. 3146 punt */ 3147 return(0); 3148 } 3149 /* all was fine.. it must fit... */ 3150 return(1); 3151 } 3152 3153 int 3154 rf_have_enough_components(RF_ConfigSet_t *cset) 3155 { 3156 RF_AutoConfig_t *ac; 3157 RF_AutoConfig_t *auto_config; 3158 RF_ComponentLabel_t *clabel; 3159 int c; 3160 int num_cols; 3161 int num_missing; 3162 int mod_counter; 3163 int mod_counter_found; 3164 int even_pair_failed; 3165 char parity_type; 3166 3167 3168 /* check to see that we have enough 'live' components 3169 of this set. If so, we can configure it if necessary */ 3170 3171 num_cols = cset->ac->clabel->num_columns; 3172 parity_type = cset->ac->clabel->parityConfig; 3173 3174 /* XXX Check for duplicate components!?!?!? */ 3175 3176 /* Determine what the mod_counter is supposed to be for this set. */ 3177 3178 mod_counter_found = 0; 3179 mod_counter = 0; 3180 ac = cset->ac; 3181 while(ac!=NULL) { 3182 if (mod_counter_found==0) { 3183 mod_counter = ac->clabel->mod_counter; 3184 mod_counter_found = 1; 3185 } else { 3186 if (ac->clabel->mod_counter > mod_counter) { 3187 mod_counter = ac->clabel->mod_counter; 3188 } 3189 } 3190 ac = ac->next; 3191 } 3192 3193 num_missing = 0; 3194 auto_config = cset->ac; 3195 3196 even_pair_failed = 0; 3197 for(c=0; c<num_cols; c++) { 3198 ac = auto_config; 3199 while(ac!=NULL) { 3200 if ((ac->clabel->column == c) && 3201 (ac->clabel->mod_counter == mod_counter)) { 3202 /* it's this one... */ 3203 #ifdef DEBUG 3204 printf("Found: %s at %d\n", 3205 ac->devname,c); 3206 #endif 3207 break; 3208 } 3209 ac=ac->next; 3210 } 3211 if (ac==NULL) { 3212 /* Didn't find one here! */ 3213 /* special case for RAID 1, especially 3214 where there are more than 2 3215 components (where RAIDframe treats 3216 things a little differently :( ) */ 3217 if (parity_type == '1') { 3218 if (c%2 == 0) { /* even component */ 3219 even_pair_failed = 1; 3220 } else { /* odd component. If 3221 we're failed, and 3222 so is the even 3223 component, it's 3224 "Good Night, Charlie" */ 3225 if (even_pair_failed == 1) { 3226 return(0); 3227 } 3228 } 3229 } else { 3230 /* normal accounting */ 3231 num_missing++; 3232 } 3233 } 3234 if ((parity_type == '1') && (c%2 == 1)) { 3235 /* Just did an even component, and we didn't 3236 bail.. reset the even_pair_failed flag, 3237 and go on to the next component.... */ 3238 even_pair_failed = 0; 3239 } 3240 } 3241 3242 clabel = cset->ac->clabel; 3243 3244 if (((clabel->parityConfig == '0') && (num_missing > 0)) || 3245 ((clabel->parityConfig == '4') && (num_missing > 1)) || 3246 ((clabel->parityConfig == '5') && (num_missing > 1))) { 3247 /* XXX this needs to be made *much* more general */ 3248 /* Too many failures */ 3249 return(0); 3250 } 3251 /* otherwise, all is well, and we've got enough to take a kick 3252 at autoconfiguring this set */ 3253 return(1); 3254 } 3255 3256 void 3257 rf_create_configuration(RF_AutoConfig_t *ac, RF_Config_t *config, 3258 RF_Raid_t *raidPtr) 3259 { 3260 RF_ComponentLabel_t *clabel; 3261 int i; 3262 3263 clabel = ac->clabel; 3264 3265 /* 1. Fill in the common stuff */ 3266 config->numRow = clabel->num_rows = 1; 3267 config->numCol = clabel->num_columns; 3268 config->numSpare = 0; /* XXX should this be set here? */ 3269 config->sectPerSU = clabel->sectPerSU; 3270 config->SUsPerPU = clabel->SUsPerPU; 3271 config->SUsPerRU = clabel->SUsPerRU; 3272 config->parityConfig = clabel->parityConfig; 3273 /* XXX... */ 3274 strcpy(config->diskQueueType,"fifo"); 3275 config->maxOutstandingDiskReqs = clabel->maxOutstanding; 3276 config->layoutSpecificSize = 0; /* XXX ?? */ 3277 3278 while(ac!=NULL) { 3279 /* row/col values will be in range due to the checks 3280 in reasonable_label() */ 3281 strcpy(config->devnames[0][ac->clabel->column], 3282 ac->devname); 3283 ac = ac->next; 3284 } 3285 3286 for(i=0;i<RF_MAXDBGV;i++) { 3287 config->debugVars[i][0] = 0; 3288 } 3289 } 3290 3291 int 3292 rf_set_autoconfig(RF_Raid_t *raidPtr, int new_value) 3293 { 3294 RF_ComponentLabel_t clabel; 3295 struct vnode *vp; 3296 dev_t dev; 3297 int column; 3298 int sparecol; 3299 3300 raidPtr->autoconfigure = new_value; 3301 3302 for(column=0; column<raidPtr->numCol; column++) { 3303 if (raidPtr->Disks[column].status == rf_ds_optimal) { 3304 dev = raidPtr->Disks[column].dev; 3305 vp = raidPtr->raid_cinfo[column].ci_vp; 3306 raidread_component_label(dev, vp, &clabel); 3307 clabel.autoconfigure = new_value; 3308 raidwrite_component_label(dev, vp, &clabel); 3309 } 3310 } 3311 for(column = 0; column < raidPtr->numSpare ; column++) { 3312 sparecol = raidPtr->numCol + column; 3313 if (raidPtr->Disks[sparecol].status == rf_ds_used_spare) { 3314 dev = raidPtr->Disks[sparecol].dev; 3315 vp = raidPtr->raid_cinfo[sparecol].ci_vp; 3316 raidread_component_label(dev, vp, &clabel); 3317 clabel.autoconfigure = new_value; 3318 raidwrite_component_label(dev, vp, &clabel); 3319 } 3320 } 3321 return(new_value); 3322 } 3323 3324 int 3325 rf_set_rootpartition(RF_Raid_t *raidPtr, int new_value) 3326 { 3327 RF_ComponentLabel_t clabel; 3328 struct vnode *vp; 3329 dev_t dev; 3330 int column; 3331 int sparecol; 3332 3333 raidPtr->root_partition = new_value; 3334 for(column=0; column<raidPtr->numCol; column++) { 3335 if (raidPtr->Disks[column].status == rf_ds_optimal) { 3336 dev = raidPtr->Disks[column].dev; 3337 vp = raidPtr->raid_cinfo[column].ci_vp; 3338 raidread_component_label(dev, vp, &clabel); 3339 clabel.root_partition = new_value; 3340 raidwrite_component_label(dev, vp, &clabel); 3341 } 3342 } 3343 for(column = 0; column < raidPtr->numSpare ; column++) { 3344 sparecol = raidPtr->numCol + column; 3345 if (raidPtr->Disks[sparecol].status == rf_ds_used_spare) { 3346 dev = raidPtr->Disks[sparecol].dev; 3347 vp = raidPtr->raid_cinfo[sparecol].ci_vp; 3348 raidread_component_label(dev, vp, &clabel); 3349 clabel.root_partition = new_value; 3350 raidwrite_component_label(dev, vp, &clabel); 3351 } 3352 } 3353 return(new_value); 3354 } 3355 3356 void 3357 rf_release_all_vps(RF_ConfigSet_t *cset) 3358 { 3359 RF_AutoConfig_t *ac; 3360 3361 ac = cset->ac; 3362 while(ac!=NULL) { 3363 /* Close the vp, and give it back */ 3364 if (ac->vp) { 3365 vn_lock(ac->vp, LK_EXCLUSIVE | LK_RETRY); 3366 VOP_CLOSE(ac->vp, FREAD, NOCRED, 0); 3367 vput(ac->vp); 3368 ac->vp = NULL; 3369 } 3370 ac = ac->next; 3371 } 3372 } 3373 3374 3375 void 3376 rf_cleanup_config_set(RF_ConfigSet_t *cset) 3377 { 3378 RF_AutoConfig_t *ac; 3379 RF_AutoConfig_t *next_ac; 3380 3381 ac = cset->ac; 3382 while(ac!=NULL) { 3383 next_ac = ac->next; 3384 /* nuke the label */ 3385 free(ac->clabel, M_RAIDFRAME); 3386 /* cleanup the config structure */ 3387 free(ac, M_RAIDFRAME); 3388 /* "next.." */ 3389 ac = next_ac; 3390 } 3391 /* and, finally, nuke the config set */ 3392 free(cset, M_RAIDFRAME); 3393 } 3394 3395 3396 void 3397 raid_init_component_label(RF_Raid_t *raidPtr, RF_ComponentLabel_t *clabel) 3398 { 3399 /* current version number */ 3400 clabel->version = RF_COMPONENT_LABEL_VERSION; 3401 clabel->serial_number = raidPtr->serial_number; 3402 clabel->mod_counter = raidPtr->mod_counter; 3403 clabel->num_rows = 1; 3404 clabel->num_columns = raidPtr->numCol; 3405 clabel->clean = RF_RAID_DIRTY; /* not clean */ 3406 clabel->status = rf_ds_optimal; /* "It's good!" */ 3407 3408 clabel->sectPerSU = raidPtr->Layout.sectorsPerStripeUnit; 3409 clabel->SUsPerPU = raidPtr->Layout.SUsPerPU; 3410 clabel->SUsPerRU = raidPtr->Layout.SUsPerRU; 3411 3412 clabel->blockSize = raidPtr->bytesPerSector; 3413 clabel->numBlocks = raidPtr->sectorsPerDisk; 3414 3415 /* XXX not portable */ 3416 clabel->parityConfig = raidPtr->Layout.map->parityConfig; 3417 clabel->maxOutstanding = raidPtr->maxOutstanding; 3418 clabel->autoconfigure = raidPtr->autoconfigure; 3419 clabel->root_partition = raidPtr->root_partition; 3420 clabel->last_unit = raidPtr->raidid; 3421 clabel->config_order = raidPtr->config_order; 3422 } 3423 3424 int 3425 rf_auto_config_set(RF_ConfigSet_t *cset, int *unit) 3426 { 3427 RF_Raid_t *raidPtr; 3428 RF_Config_t *config; 3429 int raidID; 3430 int retcode; 3431 3432 #ifdef DEBUG 3433 printf("RAID autoconfigure\n"); 3434 #endif 3435 3436 retcode = 0; 3437 *unit = -1; 3438 3439 /* 1. Create a config structure */ 3440 3441 config = (RF_Config_t *)malloc(sizeof(RF_Config_t), 3442 M_RAIDFRAME, 3443 M_NOWAIT); 3444 if (config==NULL) { 3445 printf("Out of mem!?!?\n"); 3446 /* XXX do something more intelligent here. */ 3447 return(1); 3448 } 3449 3450 memset(config, 0, sizeof(RF_Config_t)); 3451 3452 /* 3453 2. Figure out what RAID ID this one is supposed to live at 3454 See if we can get the same RAID dev that it was configured 3455 on last time.. 3456 */ 3457 3458 raidID = cset->ac->clabel->last_unit; 3459 if ((raidID < 0) || (raidID >= numraid)) { 3460 /* let's not wander off into lala land. */ 3461 raidID = numraid - 1; 3462 } 3463 if (raidPtrs[raidID]->valid != 0) { 3464 3465 /* 3466 Nope... Go looking for an alternative... 3467 Start high so we don't immediately use raid0 if that's 3468 not taken. 3469 */ 3470 3471 for(raidID = numraid - 1; raidID >= 0; raidID--) { 3472 if (raidPtrs[raidID]->valid == 0) { 3473 /* can use this one! */ 3474 break; 3475 } 3476 } 3477 } 3478 3479 if (raidID < 0) { 3480 /* punt... */ 3481 printf("Unable to auto configure this set!\n"); 3482 printf("(Out of RAID devs!)\n"); 3483 free(config, M_RAIDFRAME); 3484 return(1); 3485 } 3486 3487 #ifdef DEBUG 3488 printf("Configuring raid%d:\n",raidID); 3489 #endif 3490 3491 raidPtr = raidPtrs[raidID]; 3492 3493 /* XXX all this stuff should be done SOMEWHERE ELSE! */ 3494 raidPtr->raidid = raidID; 3495 raidPtr->openings = RAIDOUTSTANDING; 3496 3497 /* 3. Build the configuration structure */ 3498 rf_create_configuration(cset->ac, config, raidPtr); 3499 3500 /* 4. Do the configuration */ 3501 retcode = rf_Configure(raidPtr, config, cset->ac); 3502 3503 if (retcode == 0) { 3504 3505 raidinit(raidPtrs[raidID]); 3506 3507 rf_markalldirty(raidPtrs[raidID]); 3508 raidPtrs[raidID]->autoconfigure = 1; /* XXX do this here? */ 3509 if (cset->ac->clabel->root_partition==1) { 3510 /* everything configured just fine. Make a note 3511 that this set is eligible to be root. */ 3512 cset->rootable = 1; 3513 /* XXX do this here? */ 3514 raidPtrs[raidID]->root_partition = 1; 3515 } 3516 } 3517 3518 /* 5. Cleanup */ 3519 free(config, M_RAIDFRAME); 3520 3521 *unit = raidID; 3522 return(retcode); 3523 } 3524 3525 void 3526 rf_disk_unbusy(RF_RaidAccessDesc_t *desc) 3527 { 3528 struct buf *bp; 3529 3530 bp = (struct buf *)desc->bp; 3531 disk_unbusy(&raid_softc[desc->raidPtr->raidid].sc_dkdev, 3532 (bp->b_bcount - bp->b_resid), (bp->b_flags & B_READ)); 3533 } 3534 3535 void 3536 rf_pool_init(struct pool *p, size_t size, const char *w_chan, 3537 size_t xmin, size_t xmax) 3538 { 3539 pool_init(p, size, 0, 0, 0, w_chan, NULL, IPL_BIO); 3540 pool_sethiwat(p, xmax); 3541 pool_prime(p, xmin); 3542 pool_setlowat(p, xmin); 3543 } 3544 3545 /* 3546 * rf_buf_queue_check(int raidid) -- looks into the buf_queue to see 3547 * if there is IO pending and if that IO could possibly be done for a 3548 * given RAID set. Returns 0 if IO is waiting and can be done, 1 3549 * otherwise. 3550 * 3551 */ 3552 3553 int 3554 rf_buf_queue_check(int raidid) 3555 { 3556 if ((BUFQ_PEEK(raid_softc[raidid].buf_queue) != NULL) && 3557 raidPtrs[raidid]->openings > 0) { 3558 /* there is work to do */ 3559 return 0; 3560 } 3561 /* default is nothing to do */ 3562 return 1; 3563 } 3564 3565 int 3566 rf_getdisksize(struct vnode *vp, struct lwp *l, RF_RaidDisk_t *diskPtr) 3567 { 3568 struct partinfo dpart; 3569 struct dkwedge_info dkw; 3570 int error; 3571 3572 error = VOP_IOCTL(vp, DIOCGPART, &dpart, FREAD, l->l_cred, l); 3573 if (error == 0) { 3574 diskPtr->blockSize = dpart.disklab->d_secsize; 3575 diskPtr->numBlocks = dpart.part->p_size - rf_protectedSectors; 3576 diskPtr->partitionSize = dpart.part->p_size; 3577 return 0; 3578 } 3579 3580 error = VOP_IOCTL(vp, DIOCGWEDGEINFO, &dkw, FREAD, l->l_cred, l); 3581 if (error == 0) { 3582 diskPtr->blockSize = 512; /* XXX */ 3583 diskPtr->numBlocks = dkw.dkw_size - rf_protectedSectors; 3584 diskPtr->partitionSize = dkw.dkw_size; 3585 return 0; 3586 } 3587 return error; 3588 } 3589 3590 static int 3591 raid_match(struct device *self, struct cfdata *cfdata, 3592 void *aux) 3593 { 3594 return 1; 3595 } 3596 3597 static void 3598 raid_attach(struct device *parent, struct device *self, 3599 void *aux) 3600 { 3601 3602 } 3603 3604 3605 static int 3606 raid_detach(struct device *self, int flags) 3607 { 3608 struct raid_softc *rs = (struct raid_softc *)self; 3609 3610 if (rs->sc_flags & RAIDF_INITED) 3611 return EBUSY; 3612 3613 return 0; 3614 } 3615 3616 static void 3617 rf_set_properties(struct raid_softc *rs, RF_Raid_t *raidPtr) 3618 { 3619 prop_dictionary_t disk_info, odisk_info, geom; 3620 disk_info = prop_dictionary_create(); 3621 geom = prop_dictionary_create(); 3622 prop_dictionary_set_uint64(geom, "sectors-per-unit", 3623 raidPtr->totalSectors); 3624 prop_dictionary_set_uint32(geom, "sector-size", 3625 raidPtr->bytesPerSector); 3626 3627 prop_dictionary_set_uint16(geom, "sectors-per-track", 3628 raidPtr->Layout.dataSectorsPerStripe); 3629 prop_dictionary_set_uint16(geom, "tracks-per-cylinder", 3630 4 * raidPtr->numCol); 3631 3632 prop_dictionary_set_uint64(geom, "cylinders-per-unit", 3633 raidPtr->totalSectors / (raidPtr->Layout.dataSectorsPerStripe * 3634 (4 * raidPtr->numCol))); 3635 3636 prop_dictionary_set(disk_info, "geometry", geom); 3637 prop_object_release(geom); 3638 prop_dictionary_set(device_properties(rs->sc_dev), 3639 "disk-info", disk_info); 3640 odisk_info = rs->sc_dkdev.dk_info; 3641 rs->sc_dkdev.dk_info = disk_info; 3642 if (odisk_info) 3643 prop_object_release(odisk_info); 3644 } 3645
Indexes created Tue Sep 30 11:09:46 GMT 2025