rf_netbsdkintf.c revision 1.226.2.5
1 /* $NetBSD: rf_netbsdkintf.c,v 1.226.2.5 2007/08/19 23:30:20 ad 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 1.226.2.5 2007/08/19 23:30:20 ad 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/namei.h> 163 #include <sys/vnode.h> 164 #include <sys/disklabel.h> 165 #include <sys/conf.h> 166 #include <sys/lock.h> 167 #include <sys/buf.h> 168 #include <sys/bufq.h> 169 #include <sys/user.h> 170 #include <sys/reboot.h> 171 #include <sys/kauth.h> 172 173 #include <dev/raidframe/raidframevar.h> 174 #include <dev/raidframe/raidframeio.h> 175 #include "raid.h" 176 #include "opt_raid_autoconfig.h" 177 #include "rf_raid.h" 178 #include "rf_copyback.h" 179 #include "rf_dag.h" 180 #include "rf_dagflags.h" 181 #include "rf_desc.h" 182 #include "rf_diskqueue.h" 183 #include "rf_etimer.h" 184 #include "rf_general.h" 185 #include "rf_kintf.h" 186 #include "rf_options.h" 187 #include "rf_driver.h" 188 #include "rf_parityscan.h" 189 #include "rf_threadstuff.h" 190 191 #ifdef DEBUG 192 int rf_kdebug_level = 0; 193 #define db1_printf(a) if (rf_kdebug_level > 0) printf a 194 #else /* DEBUG */ 195 #define db1_printf(a) { } 196 #endif /* DEBUG */ 197 198 static RF_Raid_t **raidPtrs; /* global raid device descriptors */ 199 200 RF_DECLARE_STATIC_MUTEX(rf_sparet_wait_mutex) 201 202 static RF_SparetWait_t *rf_sparet_wait_queue; /* requests to install a 203 * spare table */ 204 static RF_SparetWait_t *rf_sparet_resp_queue; /* responses from 205 * installation process */ 206 207 MALLOC_DEFINE(M_RAIDFRAME, "RAIDframe", "RAIDframe structures"); 208 209 /* prototypes */ 210 static void KernelWakeupFunc(struct buf *); 211 static void InitBP(struct buf *, struct vnode *, unsigned, 212 dev_t, RF_SectorNum_t, RF_SectorCount_t, void *, void (*) (struct buf *), 213 void *, int, struct proc *); 214 static void raidinit(RF_Raid_t *); 215 216 void raidattach(int); 217 static int raid_match(struct device *, struct cfdata *, void *); 218 static void raid_attach(struct device *, struct device *, void *); 219 static int raid_detach(struct device *, int); 220 221 dev_type_open(raidopen); 222 dev_type_close(raidclose); 223 dev_type_read(raidread); 224 dev_type_write(raidwrite); 225 dev_type_ioctl(raidioctl); 226 dev_type_strategy(raidstrategy); 227 dev_type_dump(raiddump); 228 dev_type_size(raidsize); 229 230 const struct bdevsw raid_bdevsw = { 231 raidopen, raidclose, raidstrategy, raidioctl, 232 raiddump, raidsize, D_DISK 233 }; 234 235 const struct cdevsw raid_cdevsw = { 236 raidopen, raidclose, raidread, raidwrite, raidioctl, 237 nostop, notty, nopoll, nommap, nokqfilter, D_DISK 238 }; 239 240 /* XXX Not sure if the following should be replacing the raidPtrs above, 241 or if it should be used in conjunction with that... 242 */ 243 244 struct raid_softc { 245 struct device *sc_dev; 246 int sc_flags; /* flags */ 247 int sc_cflags; /* configuration flags */ 248 uint64_t sc_size; /* size of the raid device */ 249 char sc_xname[20]; /* XXX external name */ 250 struct disk sc_dkdev; /* generic disk device info */ 251 struct bufq_state *buf_queue; /* used for the device queue */ 252 }; 253 /* sc_flags */ 254 #define RAIDF_INITED 0x01 /* unit has been initialized */ 255 #define RAIDF_WLABEL 0x02 /* label area is writable */ 256 #define RAIDF_LABELLING 0x04 /* unit is currently being labelled */ 257 #define RAIDF_WANTED 0x40 /* someone is waiting to obtain a lock */ 258 #define RAIDF_LOCKED 0x80 /* unit is locked */ 259 260 #define raidunit(x) DISKUNIT(x) 261 int numraid = 0; 262 263 extern struct cfdriver raid_cd; 264 CFATTACH_DECL(raid, sizeof(struct raid_softc), 265 raid_match, raid_attach, raid_detach, NULL); 266 267 /* 268 * Allow RAIDOUTSTANDING number of simultaneous IO's to this RAID device. 269 * Be aware that large numbers can allow the driver to consume a lot of 270 * kernel memory, especially on writes, and in degraded mode reads. 271 * 272 * For example: with a stripe width of 64 blocks (32k) and 5 disks, 273 * a single 64K write will typically require 64K for the old data, 274 * 64K for the old parity, and 64K for the new parity, for a total 275 * of 192K (if the parity buffer is not re-used immediately). 276 * Even it if is used immediately, that's still 128K, which when multiplied 277 * by say 10 requests, is 1280K, *on top* of the 640K of incoming data. 278 * 279 * Now in degraded mode, for example, a 64K read on the above setup may 280 * require data reconstruction, which will require *all* of the 4 remaining 281 * disks to participate -- 4 * 32K/disk == 128K again. 282 */ 283 284 #ifndef RAIDOUTSTANDING 285 #define RAIDOUTSTANDING 6 286 #endif 287 288 #define RAIDLABELDEV(dev) \ 289 (MAKEDISKDEV(major((dev)), raidunit((dev)), RAW_PART)) 290 291 /* declared here, and made public, for the benefit of KVM stuff.. */ 292 struct raid_softc *raid_softc; 293 294 static void raidgetdefaultlabel(RF_Raid_t *, struct raid_softc *, 295 struct disklabel *); 296 static void raidgetdisklabel(dev_t); 297 static void raidmakedisklabel(struct raid_softc *); 298 299 static int raidlock(struct raid_softc *); 300 static void raidunlock(struct raid_softc *); 301 302 static void rf_markalldirty(RF_Raid_t *); 303 304 void rf_ReconThread(struct rf_recon_req *); 305 void rf_RewriteParityThread(RF_Raid_t *raidPtr); 306 void rf_CopybackThread(RF_Raid_t *raidPtr); 307 void rf_ReconstructInPlaceThread(struct rf_recon_req *); 308 int rf_autoconfig(struct device *self); 309 void rf_buildroothack(RF_ConfigSet_t *); 310 311 RF_AutoConfig_t *rf_find_raid_components(void); 312 RF_ConfigSet_t *rf_create_auto_sets(RF_AutoConfig_t *); 313 static int rf_does_it_fit(RF_ConfigSet_t *,RF_AutoConfig_t *); 314 static int rf_reasonable_label(RF_ComponentLabel_t *); 315 void rf_create_configuration(RF_AutoConfig_t *,RF_Config_t *, RF_Raid_t *); 316 int rf_set_autoconfig(RF_Raid_t *, int); 317 int rf_set_rootpartition(RF_Raid_t *, int); 318 void rf_release_all_vps(RF_ConfigSet_t *); 319 void rf_cleanup_config_set(RF_ConfigSet_t *); 320 int rf_have_enough_components(RF_ConfigSet_t *); 321 int rf_auto_config_set(RF_ConfigSet_t *, int *); 322 323 static int raidautoconfig = 0; /* Debugging, mostly. Set to 0 to not 324 allow autoconfig to take place. 325 Note that this is overridden by having 326 RAID_AUTOCONFIG as an option in the 327 kernel config file. */ 328 329 struct RF_Pools_s rf_pools; 330 331 void 332 raidattach(int num) 333 { 334 int raidID; 335 int i, rc; 336 337 #ifdef DEBUG 338 printf("raidattach: Asked for %d units\n", num); 339 #endif 340 341 if (num <= 0) { 342 #ifdef DIAGNOSTIC 343 panic("raidattach: count <= 0"); 344 #endif 345 return; 346 } 347 /* This is where all the initialization stuff gets done. */ 348 349 numraid = num; 350 351 /* Make some space for requested number of units... */ 352 353 RF_Malloc(raidPtrs, num * sizeof(RF_Raid_t *), (RF_Raid_t **)); 354 if (raidPtrs == NULL) { 355 panic("raidPtrs is NULL!!"); 356 } 357 358 rf_mutex_init(&rf_sparet_wait_mutex); 359 360 rf_sparet_wait_queue = rf_sparet_resp_queue = NULL; 361 362 for (i = 0; i < num; i++) 363 raidPtrs[i] = NULL; 364 rc = rf_BootRaidframe(); 365 if (rc == 0) 366 printf("Kernelized RAIDframe activated\n"); 367 else 368 panic("Serious error booting RAID!!"); 369 370 /* put together some datastructures like the CCD device does.. This 371 * lets us lock the device and what-not when it gets opened. */ 372 373 raid_softc = (struct raid_softc *) 374 malloc(num * sizeof(struct raid_softc), 375 M_RAIDFRAME, M_NOWAIT); 376 if (raid_softc == NULL) { 377 printf("WARNING: no memory for RAIDframe driver\n"); 378 return; 379 } 380 381 memset(raid_softc, 0, num * sizeof(struct raid_softc)); 382 383 for (raidID = 0; raidID < num; raidID++) { 384 bufq_alloc(&raid_softc[raidID].buf_queue, "fcfs", 0); 385 386 RF_Malloc(raidPtrs[raidID], sizeof(RF_Raid_t), 387 (RF_Raid_t *)); 388 if (raidPtrs[raidID] == NULL) { 389 printf("WARNING: raidPtrs[%d] is NULL\n", raidID); 390 numraid = raidID; 391 return; 392 } 393 } 394 395 if (config_cfattach_attach(raid_cd.cd_name, &raid_ca)) { 396 printf("config_cfattach_attach failed?\n"); 397 } 398 399 #ifdef RAID_AUTOCONFIG 400 raidautoconfig = 1; 401 #endif 402 403 /* 404 * Register a finalizer which will be used to auto-config RAID 405 * sets once all real hardware devices have been found. 406 */ 407 if (config_finalize_register(NULL, rf_autoconfig) != 0) 408 printf("WARNING: unable to register RAIDframe finalizer\n"); 409 } 410 411 int 412 rf_autoconfig(struct device *self) 413 { 414 RF_AutoConfig_t *ac_list; 415 RF_ConfigSet_t *config_sets; 416 int i; 417 418 if (raidautoconfig == 0) 419 return (0); 420 421 /* XXX This code can only be run once. */ 422 raidautoconfig = 0; 423 424 /* 1. locate all RAID components on the system */ 425 #ifdef DEBUG 426 printf("Searching for RAID components...\n"); 427 #endif 428 ac_list = rf_find_raid_components(); 429 430 /* 2. Sort them into their respective sets. */ 431 config_sets = rf_create_auto_sets(ac_list); 432 433 /* 434 * 3. Evaluate each set andconfigure the valid ones. 435 * This gets done in rf_buildroothack(). 436 */ 437 rf_buildroothack(config_sets); 438 439 for (i = 0; i < numraid; i++) 440 if (raidPtrs[i] != NULL && raidPtrs[i]->valid) 441 dkwedge_discover(&raid_softc[i].sc_dkdev); 442 443 return 1; 444 } 445 446 void 447 rf_buildroothack(RF_ConfigSet_t *config_sets) 448 { 449 RF_ConfigSet_t *cset; 450 RF_ConfigSet_t *next_cset; 451 int retcode; 452 int raidID; 453 int rootID; 454 int col; 455 int num_root; 456 char *devname; 457 458 rootID = 0; 459 num_root = 0; 460 cset = config_sets; 461 while(cset != NULL ) { 462 next_cset = cset->next; 463 if (rf_have_enough_components(cset) && 464 cset->ac->clabel->autoconfigure==1) { 465 retcode = rf_auto_config_set(cset,&raidID); 466 if (!retcode) { 467 #ifdef DEBUG 468 printf("raid%d: configured ok\n", raidID); 469 #endif 470 if (cset->rootable) { 471 rootID = raidID; 472 num_root++; 473 } 474 } else { 475 /* The autoconfig didn't work :( */ 476 #ifdef DEBUG 477 printf("Autoconfig failed with code %d for raid%d\n", retcode, raidID); 478 #endif 479 rf_release_all_vps(cset); 480 } 481 } else { 482 #ifdef DEBUG 483 printf("raid%d: not enough components\n", raidID); 484 #endif 485 /* we're not autoconfiguring this set... 486 release the associated resources */ 487 rf_release_all_vps(cset); 488 } 489 /* cleanup */ 490 rf_cleanup_config_set(cset); 491 cset = next_cset; 492 } 493 494 /* if the user has specified what the root device should be 495 then we don't touch booted_device or boothowto... */ 496 497 if (rootspec != NULL) 498 return; 499 500 /* we found something bootable... */ 501 502 if (num_root == 1) { 503 booted_device = raid_softc[rootID].sc_dev; 504 } else if (num_root > 1) { 505 506 /* 507 * Maybe the MD code can help. If it cannot, then 508 * setroot() will discover that we have no 509 * booted_device and will ask the user if nothing was 510 * hardwired in the kernel config file 511 */ 512 513 if (booted_device == NULL) 514 cpu_rootconf(); 515 if (booted_device == NULL) 516 return; 517 518 num_root = 0; 519 for (raidID = 0; raidID < numraid; raidID++) { 520 if (raidPtrs[raidID]->valid == 0) 521 continue; 522 523 if (raidPtrs[raidID]->root_partition == 0) 524 continue; 525 526 for (col = 0; col < raidPtrs[raidID]->numCol; col++) { 527 devname = raidPtrs[raidID]->Disks[col].devname; 528 devname += sizeof("/dev/") - 1; 529 if (strncmp(devname, booted_device->dv_xname, 530 strlen(booted_device->dv_xname)) != 0) 531 continue; 532 #ifdef DEBUG 533 printf("raid%d includes boot device %s\n", 534 raidID, devname); 535 #endif 536 num_root++; 537 rootID = raidID; 538 } 539 } 540 541 if (num_root == 1) { 542 booted_device = raid_softc[rootID].sc_dev; 543 } else { 544 /* we can't guess.. require the user to answer... */ 545 boothowto |= RB_ASKNAME; 546 } 547 } 548 } 549 550 551 int 552 raidsize(dev_t dev) 553 { 554 struct raid_softc *rs; 555 struct disklabel *lp; 556 int part, unit, omask, size; 557 558 unit = raidunit(dev); 559 if (unit >= numraid) 560 return (-1); 561 rs = &raid_softc[unit]; 562 563 if ((rs->sc_flags & RAIDF_INITED) == 0) 564 return (-1); 565 566 part = DISKPART(dev); 567 omask = rs->sc_dkdev.dk_openmask & (1 << part); 568 lp = rs->sc_dkdev.dk_label; 569 570 if (omask == 0 && raidopen(dev, 0, S_IFBLK, curlwp)) 571 return (-1); 572 573 if (lp->d_partitions[part].p_fstype != FS_SWAP) 574 size = -1; 575 else 576 size = lp->d_partitions[part].p_size * 577 (lp->d_secsize / DEV_BSIZE); 578 579 if (omask == 0 && raidclose(dev, 0, S_IFBLK, curlwp)) 580 return (-1); 581 582 return (size); 583 584 } 585 586 int 587 raiddump(dev_t dev, daddr_t blkno, void *va, 588 size_t size) 589 { 590 /* Not implemented. */ 591 return ENXIO; 592 } 593 /* ARGSUSED */ 594 int 595 raidopen(dev_t dev, int flags, int fmt, 596 struct lwp *l) 597 { 598 int unit = raidunit(dev); 599 struct raid_softc *rs; 600 struct disklabel *lp; 601 int part, pmask; 602 int error = 0; 603 604 if (unit >= numraid) 605 return (ENXIO); 606 rs = &raid_softc[unit]; 607 608 if ((error = raidlock(rs)) != 0) 609 return (error); 610 lp = rs->sc_dkdev.dk_label; 611 612 part = DISKPART(dev); 613 614 /* 615 * If there are wedges, and this is not RAW_PART, then we 616 * need to fail. 617 */ 618 if (rs->sc_dkdev.dk_nwedges != 0 && part != RAW_PART) { 619 error = EBUSY; 620 goto bad; 621 } 622 pmask = (1 << part); 623 624 if ((rs->sc_flags & RAIDF_INITED) && 625 (rs->sc_dkdev.dk_openmask == 0)) 626 raidgetdisklabel(dev); 627 628 /* make sure that this partition exists */ 629 630 if (part != RAW_PART) { 631 if (((rs->sc_flags & RAIDF_INITED) == 0) || 632 ((part >= lp->d_npartitions) || 633 (lp->d_partitions[part].p_fstype == FS_UNUSED))) { 634 error = ENXIO; 635 goto bad; 636 } 637 } 638 /* Prevent this unit from being unconfigured while open. */ 639 switch (fmt) { 640 case S_IFCHR: 641 rs->sc_dkdev.dk_copenmask |= pmask; 642 break; 643 644 case S_IFBLK: 645 rs->sc_dkdev.dk_bopenmask |= pmask; 646 break; 647 } 648 649 if ((rs->sc_dkdev.dk_openmask == 0) && 650 ((rs->sc_flags & RAIDF_INITED) != 0)) { 651 /* First one... mark things as dirty... Note that we *MUST* 652 have done a configure before this. I DO NOT WANT TO BE 653 SCRIBBLING TO RANDOM COMPONENTS UNTIL IT'S BEEN DETERMINED 654 THAT THEY BELONG TOGETHER!!!!! */ 655 /* XXX should check to see if we're only open for reading 656 here... If so, we needn't do this, but then need some 657 other way of keeping track of what's happened.. */ 658 659 rf_markalldirty( raidPtrs[unit] ); 660 } 661 662 663 rs->sc_dkdev.dk_openmask = 664 rs->sc_dkdev.dk_copenmask | rs->sc_dkdev.dk_bopenmask; 665 666 bad: 667 raidunlock(rs); 668 669 return (error); 670 671 672 } 673 /* ARGSUSED */ 674 int 675 raidclose(dev_t dev, int flags, int fmt, struct lwp *l) 676 { 677 int unit = raidunit(dev); 678 struct cfdata *cf; 679 struct raid_softc *rs; 680 int error = 0; 681 int part; 682 683 if (unit >= numraid) 684 return (ENXIO); 685 rs = &raid_softc[unit]; 686 687 if ((error = raidlock(rs)) != 0) 688 return (error); 689 690 part = DISKPART(dev); 691 692 /* ...that much closer to allowing unconfiguration... */ 693 switch (fmt) { 694 case S_IFCHR: 695 rs->sc_dkdev.dk_copenmask &= ~(1 << part); 696 break; 697 698 case S_IFBLK: 699 rs->sc_dkdev.dk_bopenmask &= ~(1 << part); 700 break; 701 } 702 rs->sc_dkdev.dk_openmask = 703 rs->sc_dkdev.dk_copenmask | rs->sc_dkdev.dk_bopenmask; 704 705 if ((rs->sc_dkdev.dk_openmask == 0) && 706 ((rs->sc_flags & RAIDF_INITED) != 0)) { 707 /* Last one... device is not unconfigured yet. 708 Device shutdown has taken care of setting the 709 clean bits if RAIDF_INITED is not set 710 mark things as clean... */ 711 712 rf_update_component_labels(raidPtrs[unit], 713 RF_FINAL_COMPONENT_UPDATE); 714 if (doing_shutdown) { 715 /* last one, and we're going down, so 716 lights out for this RAID set too. */ 717 error = rf_Shutdown(raidPtrs[unit]); 718 719 /* It's no longer initialized... */ 720 rs->sc_flags &= ~RAIDF_INITED; 721 722 /* detach the device */ 723 724 cf = device_cfdata(rs->sc_dev); 725 error = config_detach(rs->sc_dev, DETACH_QUIET); 726 free(cf, M_RAIDFRAME); 727 728 /* Detach the disk. */ 729 pseudo_disk_detach(&rs->sc_dkdev); 730 } 731 } 732 733 raidunlock(rs); 734 return (0); 735 736 } 737 738 void 739 raidstrategy(struct buf *bp) 740 { 741 int s; 742 743 unsigned int raidID = raidunit(bp->b_dev); 744 RF_Raid_t *raidPtr; 745 struct raid_softc *rs = &raid_softc[raidID]; 746 int wlabel; 747 748 if ((rs->sc_flags & RAIDF_INITED) ==0) { 749 bp->b_error = ENXIO; 750 goto done; 751 } 752 if (raidID >= numraid || !raidPtrs[raidID]) { 753 bp->b_error = ENODEV; 754 goto done; 755 } 756 raidPtr = raidPtrs[raidID]; 757 if (!raidPtr->valid) { 758 bp->b_error = ENODEV; 759 goto done; 760 } 761 if (bp->b_bcount == 0) { 762 db1_printf(("b_bcount is zero..\n")); 763 goto done; 764 } 765 766 /* 767 * Do bounds checking and adjust transfer. If there's an 768 * error, the bounds check will flag that for us. 769 */ 770 771 wlabel = rs->sc_flags & (RAIDF_WLABEL | RAIDF_LABELLING); 772 if (DISKPART(bp->b_dev) == RAW_PART) { 773 uint64_t size; /* device size in DEV_BSIZE unit */ 774 775 if (raidPtr->logBytesPerSector > DEV_BSHIFT) { 776 size = raidPtr->totalSectors << 777 (raidPtr->logBytesPerSector - DEV_BSHIFT); 778 } else { 779 size = raidPtr->totalSectors >> 780 (DEV_BSHIFT - raidPtr->logBytesPerSector); 781 } 782 if (bounds_check_with_mediasize(bp, DEV_BSIZE, size) <= 0) { 783 goto done; 784 } 785 } else { 786 if (bounds_check_with_label(&rs->sc_dkdev, bp, wlabel) <= 0) { 787 db1_printf(("Bounds check failed!!:%d %d\n", 788 (int) bp->b_blkno, (int) wlabel)); 789 goto done; 790 } 791 } 792 s = splbio(); 793 794 bp->b_resid = 0; 795 796 /* stuff it onto our queue */ 797 BUFQ_PUT(rs->buf_queue, bp); 798 799 /* scheduled the IO to happen at the next convenient time */ 800 wakeup(&(raidPtrs[raidID]->iodone)); 801 802 splx(s); 803 return; 804 805 done: 806 bp->b_resid = bp->b_bcount; 807 biodone(bp); 808 } 809 /* ARGSUSED */ 810 int 811 raidread(dev_t dev, struct uio *uio, int flags) 812 { 813 int unit = raidunit(dev); 814 struct raid_softc *rs; 815 816 if (unit >= numraid) 817 return (ENXIO); 818 rs = &raid_softc[unit]; 819 820 if ((rs->sc_flags & RAIDF_INITED) == 0) 821 return (ENXIO); 822 823 return (physio(raidstrategy, NULL, dev, B_READ, minphys, uio)); 824 825 } 826 /* ARGSUSED */ 827 int 828 raidwrite(dev_t dev, struct uio *uio, int flags) 829 { 830 int unit = raidunit(dev); 831 struct raid_softc *rs; 832 833 if (unit >= numraid) 834 return (ENXIO); 835 rs = &raid_softc[unit]; 836 837 if ((rs->sc_flags & RAIDF_INITED) == 0) 838 return (ENXIO); 839 840 return (physio(raidstrategy, NULL, dev, B_WRITE, minphys, uio)); 841 842 } 843 844 int 845 raidioctl(dev_t dev, u_long cmd, void *data, int flag, struct lwp *l) 846 { 847 int unit = raidunit(dev); 848 int error = 0; 849 int part, pmask; 850 struct cfdata *cf; 851 struct raid_softc *rs; 852 RF_Config_t *k_cfg, *u_cfg; 853 RF_Raid_t *raidPtr; 854 RF_RaidDisk_t *diskPtr; 855 RF_AccTotals_t *totals; 856 RF_DeviceConfig_t *d_cfg, **ucfgp; 857 u_char *specific_buf; 858 int retcode = 0; 859 int column; 860 int raidid; 861 struct rf_recon_req *rrcopy, *rr; 862 RF_ComponentLabel_t *clabel; 863 RF_ComponentLabel_t *ci_label; 864 RF_ComponentLabel_t **clabel_ptr; 865 RF_SingleComponent_t *sparePtr,*componentPtr; 866 RF_SingleComponent_t component; 867 RF_ProgressInfo_t progressInfo, **progressInfoPtr; 868 int i, j, d; 869 #ifdef __HAVE_OLD_DISKLABEL 870 struct disklabel newlabel; 871 #endif 872 struct dkwedge_info *dkw; 873 874 if (unit >= numraid) 875 return (ENXIO); 876 rs = &raid_softc[unit]; 877 raidPtr = raidPtrs[unit]; 878 879 db1_printf(("raidioctl: %d %d %d %d\n", (int) dev, 880 (int) DISKPART(dev), (int) unit, (int) cmd)); 881 882 /* Must be open for writes for these commands... */ 883 switch (cmd) { 884 #ifdef DIOCGSECTORSIZE 885 case DIOCGSECTORSIZE: 886 *(u_int *)data = raidPtr->bytesPerSector; 887 return 0; 888 case DIOCGMEDIASIZE: 889 *(off_t *)data = 890 (off_t)raidPtr->totalSectors * raidPtr->bytesPerSector; 891 return 0; 892 #endif 893 case DIOCSDINFO: 894 case DIOCWDINFO: 895 #ifdef __HAVE_OLD_DISKLABEL 896 case ODIOCWDINFO: 897 case ODIOCSDINFO: 898 #endif 899 case DIOCWLABEL: 900 case DIOCAWEDGE: 901 case DIOCDWEDGE: 902 if ((flag & FWRITE) == 0) 903 return (EBADF); 904 } 905 906 /* Must be initialized for these... */ 907 switch (cmd) { 908 case DIOCGDINFO: 909 case DIOCSDINFO: 910 case DIOCWDINFO: 911 #ifdef __HAVE_OLD_DISKLABEL 912 case ODIOCGDINFO: 913 case ODIOCWDINFO: 914 case ODIOCSDINFO: 915 case ODIOCGDEFLABEL: 916 #endif 917 case DIOCGPART: 918 case DIOCWLABEL: 919 case DIOCGDEFLABEL: 920 case DIOCAWEDGE: 921 case DIOCDWEDGE: 922 case DIOCLWEDGES: 923 case RAIDFRAME_SHUTDOWN: 924 case RAIDFRAME_REWRITEPARITY: 925 case RAIDFRAME_GET_INFO: 926 case RAIDFRAME_RESET_ACCTOTALS: 927 case RAIDFRAME_GET_ACCTOTALS: 928 case RAIDFRAME_KEEP_ACCTOTALS: 929 case RAIDFRAME_GET_SIZE: 930 case RAIDFRAME_FAIL_DISK: 931 case RAIDFRAME_COPYBACK: 932 case RAIDFRAME_CHECK_RECON_STATUS: 933 case RAIDFRAME_CHECK_RECON_STATUS_EXT: 934 case RAIDFRAME_GET_COMPONENT_LABEL: 935 case RAIDFRAME_SET_COMPONENT_LABEL: 936 case RAIDFRAME_ADD_HOT_SPARE: 937 case RAIDFRAME_REMOVE_HOT_SPARE: 938 case RAIDFRAME_INIT_LABELS: 939 case RAIDFRAME_REBUILD_IN_PLACE: 940 case RAIDFRAME_CHECK_PARITY: 941 case RAIDFRAME_CHECK_PARITYREWRITE_STATUS: 942 case RAIDFRAME_CHECK_PARITYREWRITE_STATUS_EXT: 943 case RAIDFRAME_CHECK_COPYBACK_STATUS: 944 case RAIDFRAME_CHECK_COPYBACK_STATUS_EXT: 945 case RAIDFRAME_SET_AUTOCONFIG: 946 case RAIDFRAME_SET_ROOT: 947 case RAIDFRAME_DELETE_COMPONENT: 948 case RAIDFRAME_INCORPORATE_HOT_SPARE: 949 if ((rs->sc_flags & RAIDF_INITED) == 0) 950 return (ENXIO); 951 } 952 953 switch (cmd) { 954 955 /* configure the system */ 956 case RAIDFRAME_CONFIGURE: 957 958 if (raidPtr->valid) { 959 /* There is a valid RAID set running on this unit! */ 960 printf("raid%d: Device already configured!\n",unit); 961 return(EINVAL); 962 } 963 964 /* copy-in the configuration information */ 965 /* data points to a pointer to the configuration structure */ 966 967 u_cfg = *((RF_Config_t **) data); 968 RF_Malloc(k_cfg, sizeof(RF_Config_t), (RF_Config_t *)); 969 if (k_cfg == NULL) { 970 return (ENOMEM); 971 } 972 retcode = copyin(u_cfg, k_cfg, sizeof(RF_Config_t)); 973 if (retcode) { 974 RF_Free(k_cfg, sizeof(RF_Config_t)); 975 db1_printf(("rf_ioctl: retcode=%d copyin.1\n", 976 retcode)); 977 return (retcode); 978 } 979 /* allocate a buffer for the layout-specific data, and copy it 980 * in */ 981 if (k_cfg->layoutSpecificSize) { 982 if (k_cfg->layoutSpecificSize > 10000) { 983 /* sanity check */ 984 RF_Free(k_cfg, sizeof(RF_Config_t)); 985 return (EINVAL); 986 } 987 RF_Malloc(specific_buf, k_cfg->layoutSpecificSize, 988 (u_char *)); 989 if (specific_buf == NULL) { 990 RF_Free(k_cfg, sizeof(RF_Config_t)); 991 return (ENOMEM); 992 } 993 retcode = copyin(k_cfg->layoutSpecific, specific_buf, 994 k_cfg->layoutSpecificSize); 995 if (retcode) { 996 RF_Free(k_cfg, sizeof(RF_Config_t)); 997 RF_Free(specific_buf, 998 k_cfg->layoutSpecificSize); 999 db1_printf(("rf_ioctl: retcode=%d copyin.2\n", 1000 retcode)); 1001 return (retcode); 1002 } 1003 } else 1004 specific_buf = NULL; 1005 k_cfg->layoutSpecific = specific_buf; 1006 1007 /* should do some kind of sanity check on the configuration. 1008 * Store the sum of all the bytes in the last byte? */ 1009 1010 /* configure the system */ 1011 1012 /* 1013 * Clear the entire RAID descriptor, just to make sure 1014 * there is no stale data left in the case of a 1015 * reconfiguration 1016 */ 1017 memset((char *) raidPtr, 0, sizeof(RF_Raid_t)); 1018 raidPtr->raidid = unit; 1019 1020 retcode = rf_Configure(raidPtr, k_cfg, NULL); 1021 1022 if (retcode == 0) { 1023 1024 /* allow this many simultaneous IO's to 1025 this RAID device */ 1026 raidPtr->openings = RAIDOUTSTANDING; 1027 1028 raidinit(raidPtr); 1029 rf_markalldirty(raidPtr); 1030 } 1031 /* free the buffers. No return code here. */ 1032 if (k_cfg->layoutSpecificSize) { 1033 RF_Free(specific_buf, k_cfg->layoutSpecificSize); 1034 } 1035 RF_Free(k_cfg, sizeof(RF_Config_t)); 1036 1037 return (retcode); 1038 1039 /* shutdown the system */ 1040 case RAIDFRAME_SHUTDOWN: 1041 1042 if ((error = raidlock(rs)) != 0) 1043 return (error); 1044 1045 /* 1046 * If somebody has a partition mounted, we shouldn't 1047 * shutdown. 1048 */ 1049 1050 part = DISKPART(dev); 1051 pmask = (1 << part); 1052 if ((rs->sc_dkdev.dk_openmask & ~pmask) || 1053 ((rs->sc_dkdev.dk_bopenmask & pmask) && 1054 (rs->sc_dkdev.dk_copenmask & pmask))) { 1055 raidunlock(rs); 1056 return (EBUSY); 1057 } 1058 1059 retcode = rf_Shutdown(raidPtr); 1060 1061 /* It's no longer initialized... */ 1062 rs->sc_flags &= ~RAIDF_INITED; 1063 1064 /* free the pseudo device attach bits */ 1065 1066 cf = device_cfdata(rs->sc_dev); 1067 /* XXX this causes us to not return any errors 1068 from the above call to rf_Shutdown() */ 1069 retcode = config_detach(rs->sc_dev, DETACH_QUIET); 1070 free(cf, M_RAIDFRAME); 1071 1072 /* Detach the disk. */ 1073 pseudo_disk_detach(&rs->sc_dkdev); 1074 1075 raidunlock(rs); 1076 1077 return (retcode); 1078 case RAIDFRAME_GET_COMPONENT_LABEL: 1079 clabel_ptr = (RF_ComponentLabel_t **) data; 1080 /* need to read the component label for the disk indicated 1081 by row,column in clabel */ 1082 1083 /* For practice, let's get it directly fromdisk, rather 1084 than from the in-core copy */ 1085 RF_Malloc( clabel, sizeof( RF_ComponentLabel_t ), 1086 (RF_ComponentLabel_t *)); 1087 if (clabel == NULL) 1088 return (ENOMEM); 1089 1090 retcode = copyin( *clabel_ptr, clabel, 1091 sizeof(RF_ComponentLabel_t)); 1092 1093 if (retcode) { 1094 RF_Free( clabel, sizeof(RF_ComponentLabel_t)); 1095 return(retcode); 1096 } 1097 1098 clabel->row = 0; /* Don't allow looking at anything else.*/ 1099 1100 column = clabel->column; 1101 1102 if ((column < 0) || (column >= raidPtr->numCol + 1103 raidPtr->numSpare)) { 1104 RF_Free( clabel, sizeof(RF_ComponentLabel_t)); 1105 return(EINVAL); 1106 } 1107 1108 retcode = raidread_component_label(raidPtr->Disks[column].dev, 1109 raidPtr->raid_cinfo[column].ci_vp, 1110 clabel ); 1111 1112 if (retcode == 0) { 1113 retcode = copyout(clabel, *clabel_ptr, 1114 sizeof(RF_ComponentLabel_t)); 1115 } 1116 RF_Free(clabel, sizeof(RF_ComponentLabel_t)); 1117 return (retcode); 1118 1119 case RAIDFRAME_SET_COMPONENT_LABEL: 1120 clabel = (RF_ComponentLabel_t *) data; 1121 1122 /* XXX check the label for valid stuff... */ 1123 /* Note that some things *should not* get modified -- 1124 the user should be re-initing the labels instead of 1125 trying to patch things. 1126 */ 1127 1128 raidid = raidPtr->raidid; 1129 #ifdef DEBUG 1130 printf("raid%d: Got component label:\n", raidid); 1131 printf("raid%d: Version: %d\n", raidid, clabel->version); 1132 printf("raid%d: Serial Number: %d\n", raidid, clabel->serial_number); 1133 printf("raid%d: Mod counter: %d\n", raidid, clabel->mod_counter); 1134 printf("raid%d: Column: %d\n", raidid, clabel->column); 1135 printf("raid%d: Num Columns: %d\n", raidid, clabel->num_columns); 1136 printf("raid%d: Clean: %d\n", raidid, clabel->clean); 1137 printf("raid%d: Status: %d\n", raidid, clabel->status); 1138 #endif 1139 clabel->row = 0; 1140 column = clabel->column; 1141 1142 if ((column < 0) || (column >= raidPtr->numCol)) { 1143 return(EINVAL); 1144 } 1145 1146 /* XXX this isn't allowed to do anything for now :-) */ 1147 1148 /* XXX and before it is, we need to fill in the rest 1149 of the fields!?!?!?! */ 1150 #if 0 1151 raidwrite_component_label( 1152 raidPtr->Disks[column].dev, 1153 raidPtr->raid_cinfo[column].ci_vp, 1154 clabel ); 1155 #endif 1156 return (0); 1157 1158 case RAIDFRAME_INIT_LABELS: 1159 clabel = (RF_ComponentLabel_t *) data; 1160 /* 1161 we only want the serial number from 1162 the above. We get all the rest of the information 1163 from the config that was used to create this RAID 1164 set. 1165 */ 1166 1167 raidPtr->serial_number = clabel->serial_number; 1168 1169 RF_Malloc(ci_label, sizeof(RF_ComponentLabel_t), 1170 (RF_ComponentLabel_t *)); 1171 if (ci_label == NULL) 1172 return (ENOMEM); 1173 1174 raid_init_component_label(raidPtr, ci_label); 1175 ci_label->serial_number = clabel->serial_number; 1176 ci_label->row = 0; /* we dont' pretend to support more */ 1177 1178 for(column=0;column<raidPtr->numCol;column++) { 1179 diskPtr = &raidPtr->Disks[column]; 1180 if (!RF_DEAD_DISK(diskPtr->status)) { 1181 ci_label->partitionSize = diskPtr->partitionSize; 1182 ci_label->column = column; 1183 raidwrite_component_label( 1184 raidPtr->Disks[column].dev, 1185 raidPtr->raid_cinfo[column].ci_vp, 1186 ci_label ); 1187 } 1188 } 1189 RF_Free(ci_label, sizeof(RF_ComponentLabel_t)); 1190 1191 return (retcode); 1192 case RAIDFRAME_SET_AUTOCONFIG: 1193 d = rf_set_autoconfig(raidPtr, *(int *) data); 1194 printf("raid%d: New autoconfig value is: %d\n", 1195 raidPtr->raidid, d); 1196 *(int *) data = d; 1197 return (retcode); 1198 1199 case RAIDFRAME_SET_ROOT: 1200 d = rf_set_rootpartition(raidPtr, *(int *) data); 1201 printf("raid%d: New rootpartition value is: %d\n", 1202 raidPtr->raidid, d); 1203 *(int *) data = d; 1204 return (retcode); 1205 1206 /* initialize all parity */ 1207 case RAIDFRAME_REWRITEPARITY: 1208 1209 if (raidPtr->Layout.map->faultsTolerated == 0) { 1210 /* Parity for RAID 0 is trivially correct */ 1211 raidPtr->parity_good = RF_RAID_CLEAN; 1212 return(0); 1213 } 1214 1215 if (raidPtr->parity_rewrite_in_progress == 1) { 1216 /* Re-write is already in progress! */ 1217 return(EINVAL); 1218 } 1219 1220 retcode = RF_CREATE_THREAD(raidPtr->parity_rewrite_thread, 1221 rf_RewriteParityThread, 1222 raidPtr,"raid_parity"); 1223 return (retcode); 1224 1225 1226 case RAIDFRAME_ADD_HOT_SPARE: 1227 sparePtr = (RF_SingleComponent_t *) data; 1228 memcpy( &component, sparePtr, sizeof(RF_SingleComponent_t)); 1229 retcode = rf_add_hot_spare(raidPtr, &component); 1230 return(retcode); 1231 1232 case RAIDFRAME_REMOVE_HOT_SPARE: 1233 return(retcode); 1234 1235 case RAIDFRAME_DELETE_COMPONENT: 1236 componentPtr = (RF_SingleComponent_t *)data; 1237 memcpy( &component, componentPtr, 1238 sizeof(RF_SingleComponent_t)); 1239 retcode = rf_delete_component(raidPtr, &component); 1240 return(retcode); 1241 1242 case RAIDFRAME_INCORPORATE_HOT_SPARE: 1243 componentPtr = (RF_SingleComponent_t *)data; 1244 memcpy( &component, componentPtr, 1245 sizeof(RF_SingleComponent_t)); 1246 retcode = rf_incorporate_hot_spare(raidPtr, &component); 1247 return(retcode); 1248 1249 case RAIDFRAME_REBUILD_IN_PLACE: 1250 1251 if (raidPtr->Layout.map->faultsTolerated == 0) { 1252 /* Can't do this on a RAID 0!! */ 1253 return(EINVAL); 1254 } 1255 1256 if (raidPtr->recon_in_progress == 1) { 1257 /* a reconstruct is already in progress! */ 1258 return(EINVAL); 1259 } 1260 1261 componentPtr = (RF_SingleComponent_t *) data; 1262 memcpy( &component, componentPtr, 1263 sizeof(RF_SingleComponent_t)); 1264 component.row = 0; /* we don't support any more */ 1265 column = component.column; 1266 1267 if ((column < 0) || (column >= raidPtr->numCol)) { 1268 return(EINVAL); 1269 } 1270 1271 RF_LOCK_MUTEX(raidPtr->mutex); 1272 if ((raidPtr->Disks[column].status == rf_ds_optimal) && 1273 (raidPtr->numFailures > 0)) { 1274 /* XXX 0 above shouldn't be constant!!! */ 1275 /* some component other than this has failed. 1276 Let's not make things worse than they already 1277 are... */ 1278 printf("raid%d: Unable to reconstruct to disk at:\n", 1279 raidPtr->raidid); 1280 printf("raid%d: Col: %d Too many failures.\n", 1281 raidPtr->raidid, column); 1282 RF_UNLOCK_MUTEX(raidPtr->mutex); 1283 return (EINVAL); 1284 } 1285 if (raidPtr->Disks[column].status == 1286 rf_ds_reconstructing) { 1287 printf("raid%d: Unable to reconstruct to disk at:\n", 1288 raidPtr->raidid); 1289 printf("raid%d: Col: %d Reconstruction already occuring!\n", raidPtr->raidid, column); 1290 1291 RF_UNLOCK_MUTEX(raidPtr->mutex); 1292 return (EINVAL); 1293 } 1294 if (raidPtr->Disks[column].status == rf_ds_spared) { 1295 RF_UNLOCK_MUTEX(raidPtr->mutex); 1296 return (EINVAL); 1297 } 1298 RF_UNLOCK_MUTEX(raidPtr->mutex); 1299 1300 RF_Malloc(rrcopy, sizeof(*rrcopy), (struct rf_recon_req *)); 1301 if (rrcopy == NULL) 1302 return(ENOMEM); 1303 1304 rrcopy->raidPtr = (void *) raidPtr; 1305 rrcopy->col = column; 1306 1307 retcode = RF_CREATE_THREAD(raidPtr->recon_thread, 1308 rf_ReconstructInPlaceThread, 1309 rrcopy,"raid_reconip"); 1310 return(retcode); 1311 1312 case RAIDFRAME_GET_INFO: 1313 if (!raidPtr->valid) 1314 return (ENODEV); 1315 ucfgp = (RF_DeviceConfig_t **) data; 1316 RF_Malloc(d_cfg, sizeof(RF_DeviceConfig_t), 1317 (RF_DeviceConfig_t *)); 1318 if (d_cfg == NULL) 1319 return (ENOMEM); 1320 d_cfg->rows = 1; /* there is only 1 row now */ 1321 d_cfg->cols = raidPtr->numCol; 1322 d_cfg->ndevs = raidPtr->numCol; 1323 if (d_cfg->ndevs >= RF_MAX_DISKS) { 1324 RF_Free(d_cfg, sizeof(RF_DeviceConfig_t)); 1325 return (ENOMEM); 1326 } 1327 d_cfg->nspares = raidPtr->numSpare; 1328 if (d_cfg->nspares >= RF_MAX_DISKS) { 1329 RF_Free(d_cfg, sizeof(RF_DeviceConfig_t)); 1330 return (ENOMEM); 1331 } 1332 d_cfg->maxqdepth = raidPtr->maxQueueDepth; 1333 d = 0; 1334 for (j = 0; j < d_cfg->cols; j++) { 1335 d_cfg->devs[d] = raidPtr->Disks[j]; 1336 d++; 1337 } 1338 for (j = d_cfg->cols, i = 0; i < d_cfg->nspares; i++, j++) { 1339 d_cfg->spares[i] = raidPtr->Disks[j]; 1340 } 1341 retcode = copyout(d_cfg, *ucfgp, sizeof(RF_DeviceConfig_t)); 1342 RF_Free(d_cfg, sizeof(RF_DeviceConfig_t)); 1343 1344 return (retcode); 1345 1346 case RAIDFRAME_CHECK_PARITY: 1347 *(int *) data = raidPtr->parity_good; 1348 return (0); 1349 1350 case RAIDFRAME_RESET_ACCTOTALS: 1351 memset(&raidPtr->acc_totals, 0, sizeof(raidPtr->acc_totals)); 1352 return (0); 1353 1354 case RAIDFRAME_GET_ACCTOTALS: 1355 totals = (RF_AccTotals_t *) data; 1356 *totals = raidPtr->acc_totals; 1357 return (0); 1358 1359 case RAIDFRAME_KEEP_ACCTOTALS: 1360 raidPtr->keep_acc_totals = *(int *)data; 1361 return (0); 1362 1363 case RAIDFRAME_GET_SIZE: 1364 *(int *) data = raidPtr->totalSectors; 1365 return (0); 1366 1367 /* fail a disk & optionally start reconstruction */ 1368 case RAIDFRAME_FAIL_DISK: 1369 1370 if (raidPtr->Layout.map->faultsTolerated == 0) { 1371 /* Can't do this on a RAID 0!! */ 1372 return(EINVAL); 1373 } 1374 1375 rr = (struct rf_recon_req *) data; 1376 rr->row = 0; 1377 if (rr->col < 0 || rr->col >= raidPtr->numCol) 1378 return (EINVAL); 1379 1380 1381 RF_LOCK_MUTEX(raidPtr->mutex); 1382 if (raidPtr->status == rf_rs_reconstructing) { 1383 /* you can't fail a disk while we're reconstructing! */ 1384 /* XXX wrong for RAID6 */ 1385 RF_UNLOCK_MUTEX(raidPtr->mutex); 1386 return (EINVAL); 1387 } 1388 if ((raidPtr->Disks[rr->col].status == 1389 rf_ds_optimal) && (raidPtr->numFailures > 0)) { 1390 /* some other component has failed. Let's not make 1391 things worse. XXX wrong for RAID6 */ 1392 RF_UNLOCK_MUTEX(raidPtr->mutex); 1393 return (EINVAL); 1394 } 1395 if (raidPtr->Disks[rr->col].status == rf_ds_spared) { 1396 /* Can't fail a spared disk! */ 1397 RF_UNLOCK_MUTEX(raidPtr->mutex); 1398 return (EINVAL); 1399 } 1400 RF_UNLOCK_MUTEX(raidPtr->mutex); 1401 1402 /* make a copy of the recon request so that we don't rely on 1403 * the user's buffer */ 1404 RF_Malloc(rrcopy, sizeof(*rrcopy), (struct rf_recon_req *)); 1405 if (rrcopy == NULL) 1406 return(ENOMEM); 1407 memcpy(rrcopy, rr, sizeof(*rr)); 1408 rrcopy->raidPtr = (void *) raidPtr; 1409 1410 retcode = RF_CREATE_THREAD(raidPtr->recon_thread, 1411 rf_ReconThread, 1412 rrcopy,"raid_recon"); 1413 return (0); 1414 1415 /* invoke a copyback operation after recon on whatever disk 1416 * needs it, if any */ 1417 case RAIDFRAME_COPYBACK: 1418 1419 if (raidPtr->Layout.map->faultsTolerated == 0) { 1420 /* This makes no sense on a RAID 0!! */ 1421 return(EINVAL); 1422 } 1423 1424 if (raidPtr->copyback_in_progress == 1) { 1425 /* Copyback is already in progress! */ 1426 return(EINVAL); 1427 } 1428 1429 retcode = RF_CREATE_THREAD(raidPtr->copyback_thread, 1430 rf_CopybackThread, 1431 raidPtr,"raid_copyback"); 1432 return (retcode); 1433 1434 /* return the percentage completion of reconstruction */ 1435 case RAIDFRAME_CHECK_RECON_STATUS: 1436 if (raidPtr->Layout.map->faultsTolerated == 0) { 1437 /* This makes no sense on a RAID 0, so tell the 1438 user it's done. */ 1439 *(int *) data = 100; 1440 return(0); 1441 } 1442 if (raidPtr->status != rf_rs_reconstructing) 1443 *(int *) data = 100; 1444 else { 1445 if (raidPtr->reconControl->numRUsTotal > 0) { 1446 *(int *) data = (raidPtr->reconControl->numRUsComplete * 100 / raidPtr->reconControl->numRUsTotal); 1447 } else { 1448 *(int *) data = 0; 1449 } 1450 } 1451 return (0); 1452 case RAIDFRAME_CHECK_RECON_STATUS_EXT: 1453 progressInfoPtr = (RF_ProgressInfo_t **) data; 1454 if (raidPtr->status != rf_rs_reconstructing) { 1455 progressInfo.remaining = 0; 1456 progressInfo.completed = 100; 1457 progressInfo.total = 100; 1458 } else { 1459 progressInfo.total = 1460 raidPtr->reconControl->numRUsTotal; 1461 progressInfo.completed = 1462 raidPtr->reconControl->numRUsComplete; 1463 progressInfo.remaining = progressInfo.total - 1464 progressInfo.completed; 1465 } 1466 retcode = copyout(&progressInfo, *progressInfoPtr, 1467 sizeof(RF_ProgressInfo_t)); 1468 return (retcode); 1469 1470 case RAIDFRAME_CHECK_PARITYREWRITE_STATUS: 1471 if (raidPtr->Layout.map->faultsTolerated == 0) { 1472 /* This makes no sense on a RAID 0, so tell the 1473 user it's done. */ 1474 *(int *) data = 100; 1475 return(0); 1476 } 1477 if (raidPtr->parity_rewrite_in_progress == 1) { 1478 *(int *) data = 100 * 1479 raidPtr->parity_rewrite_stripes_done / 1480 raidPtr->Layout.numStripe; 1481 } else { 1482 *(int *) data = 100; 1483 } 1484 return (0); 1485 1486 case RAIDFRAME_CHECK_PARITYREWRITE_STATUS_EXT: 1487 progressInfoPtr = (RF_ProgressInfo_t **) data; 1488 if (raidPtr->parity_rewrite_in_progress == 1) { 1489 progressInfo.total = raidPtr->Layout.numStripe; 1490 progressInfo.completed = 1491 raidPtr->parity_rewrite_stripes_done; 1492 progressInfo.remaining = progressInfo.total - 1493 progressInfo.completed; 1494 } else { 1495 progressInfo.remaining = 0; 1496 progressInfo.completed = 100; 1497 progressInfo.total = 100; 1498 } 1499 retcode = copyout(&progressInfo, *progressInfoPtr, 1500 sizeof(RF_ProgressInfo_t)); 1501 return (retcode); 1502 1503 case RAIDFRAME_CHECK_COPYBACK_STATUS: 1504 if (raidPtr->Layout.map->faultsTolerated == 0) { 1505 /* This makes no sense on a RAID 0 */ 1506 *(int *) data = 100; 1507 return(0); 1508 } 1509 if (raidPtr->copyback_in_progress == 1) { 1510 *(int *) data = 100 * raidPtr->copyback_stripes_done / 1511 raidPtr->Layout.numStripe; 1512 } else { 1513 *(int *) data = 100; 1514 } 1515 return (0); 1516 1517 case RAIDFRAME_CHECK_COPYBACK_STATUS_EXT: 1518 progressInfoPtr = (RF_ProgressInfo_t **) data; 1519 if (raidPtr->copyback_in_progress == 1) { 1520 progressInfo.total = raidPtr->Layout.numStripe; 1521 progressInfo.completed = 1522 raidPtr->copyback_stripes_done; 1523 progressInfo.remaining = progressInfo.total - 1524 progressInfo.completed; 1525 } else { 1526 progressInfo.remaining = 0; 1527 progressInfo.completed = 100; 1528 progressInfo.total = 100; 1529 } 1530 retcode = copyout(&progressInfo, *progressInfoPtr, 1531 sizeof(RF_ProgressInfo_t)); 1532 return (retcode); 1533 1534 /* the sparetable daemon calls this to wait for the kernel to 1535 * need a spare table. this ioctl does not return until a 1536 * spare table is needed. XXX -- calling mpsleep here in the 1537 * ioctl code is almost certainly wrong and evil. -- XXX XXX 1538 * -- I should either compute the spare table in the kernel, 1539 * or have a different -- XXX XXX -- interface (a different 1540 * character device) for delivering the table -- XXX */ 1541 #if 0 1542 case RAIDFRAME_SPARET_WAIT: 1543 RF_LOCK_MUTEX(rf_sparet_wait_mutex); 1544 while (!rf_sparet_wait_queue) 1545 mpsleep(&rf_sparet_wait_queue, (PZERO + 1) | PCATCH, "sparet wait", 0, (void *) simple_lock_addr(rf_sparet_wait_mutex), MS_LOCK_SIMPLE); 1546 waitreq = rf_sparet_wait_queue; 1547 rf_sparet_wait_queue = rf_sparet_wait_queue->next; 1548 RF_UNLOCK_MUTEX(rf_sparet_wait_mutex); 1549 1550 /* structure assignment */ 1551 *((RF_SparetWait_t *) data) = *waitreq; 1552 1553 RF_Free(waitreq, sizeof(*waitreq)); 1554 return (0); 1555 1556 /* wakes up a process waiting on SPARET_WAIT and puts an error 1557 * code in it that will cause the dameon to exit */ 1558 case RAIDFRAME_ABORT_SPARET_WAIT: 1559 RF_Malloc(waitreq, sizeof(*waitreq), (RF_SparetWait_t *)); 1560 waitreq->fcol = -1; 1561 RF_LOCK_MUTEX(rf_sparet_wait_mutex); 1562 waitreq->next = rf_sparet_wait_queue; 1563 rf_sparet_wait_queue = waitreq; 1564 RF_UNLOCK_MUTEX(rf_sparet_wait_mutex); 1565 wakeup(&rf_sparet_wait_queue); 1566 return (0); 1567 1568 /* used by the spare table daemon to deliver a spare table 1569 * into the kernel */ 1570 case RAIDFRAME_SEND_SPARET: 1571 1572 /* install the spare table */ 1573 retcode = rf_SetSpareTable(raidPtr, *(void **) data); 1574 1575 /* respond to the requestor. the return status of the spare 1576 * table installation is passed in the "fcol" field */ 1577 RF_Malloc(waitreq, sizeof(*waitreq), (RF_SparetWait_t *)); 1578 waitreq->fcol = retcode; 1579 RF_LOCK_MUTEX(rf_sparet_wait_mutex); 1580 waitreq->next = rf_sparet_resp_queue; 1581 rf_sparet_resp_queue = waitreq; 1582 wakeup(&rf_sparet_resp_queue); 1583 RF_UNLOCK_MUTEX(rf_sparet_wait_mutex); 1584 1585 return (retcode); 1586 #endif 1587 1588 default: 1589 break; /* fall through to the os-specific code below */ 1590 1591 } 1592 1593 if (!raidPtr->valid) 1594 return (EINVAL); 1595 1596 /* 1597 * Add support for "regular" device ioctls here. 1598 */ 1599 1600 switch (cmd) { 1601 case DIOCGDINFO: 1602 *(struct disklabel *) data = *(rs->sc_dkdev.dk_label); 1603 break; 1604 #ifdef __HAVE_OLD_DISKLABEL 1605 case ODIOCGDINFO: 1606 newlabel = *(rs->sc_dkdev.dk_label); 1607 if (newlabel.d_npartitions > OLDMAXPARTITIONS) 1608 return ENOTTY; 1609 memcpy(data, &newlabel, sizeof (struct olddisklabel)); 1610 break; 1611 #endif 1612 1613 case DIOCGPART: 1614 ((struct partinfo *) data)->disklab = rs->sc_dkdev.dk_label; 1615 ((struct partinfo *) data)->part = 1616 &rs->sc_dkdev.dk_label->d_partitions[DISKPART(dev)]; 1617 break; 1618 1619 case DIOCWDINFO: 1620 case DIOCSDINFO: 1621 #ifdef __HAVE_OLD_DISKLABEL 1622 case ODIOCWDINFO: 1623 case ODIOCSDINFO: 1624 #endif 1625 { 1626 struct disklabel *lp; 1627 #ifdef __HAVE_OLD_DISKLABEL 1628 if (cmd == ODIOCSDINFO || cmd == ODIOCWDINFO) { 1629 memset(&newlabel, 0, sizeof newlabel); 1630 memcpy(&newlabel, data, sizeof (struct olddisklabel)); 1631 lp = &newlabel; 1632 } else 1633 #endif 1634 lp = (struct disklabel *)data; 1635 1636 if ((error = raidlock(rs)) != 0) 1637 return (error); 1638 1639 rs->sc_flags |= RAIDF_LABELLING; 1640 1641 error = setdisklabel(rs->sc_dkdev.dk_label, 1642 lp, 0, rs->sc_dkdev.dk_cpulabel); 1643 if (error == 0) { 1644 if (cmd == DIOCWDINFO 1645 #ifdef __HAVE_OLD_DISKLABEL 1646 || cmd == ODIOCWDINFO 1647 #endif 1648 ) 1649 error = writedisklabel(RAIDLABELDEV(dev), 1650 raidstrategy, rs->sc_dkdev.dk_label, 1651 rs->sc_dkdev.dk_cpulabel); 1652 } 1653 rs->sc_flags &= ~RAIDF_LABELLING; 1654 1655 raidunlock(rs); 1656 1657 if (error) 1658 return (error); 1659 break; 1660 } 1661 1662 case DIOCWLABEL: 1663 if (*(int *) data != 0) 1664 rs->sc_flags |= RAIDF_WLABEL; 1665 else 1666 rs->sc_flags &= ~RAIDF_WLABEL; 1667 break; 1668 1669 case DIOCGDEFLABEL: 1670 raidgetdefaultlabel(raidPtr, rs, (struct disklabel *) data); 1671 break; 1672 1673 #ifdef __HAVE_OLD_DISKLABEL 1674 case ODIOCGDEFLABEL: 1675 raidgetdefaultlabel(raidPtr, rs, &newlabel); 1676 if (newlabel.d_npartitions > OLDMAXPARTITIONS) 1677 return ENOTTY; 1678 memcpy(data, &newlabel, sizeof (struct olddisklabel)); 1679 break; 1680 #endif 1681 1682 case DIOCAWEDGE: 1683 case DIOCDWEDGE: 1684 dkw = (void *)data; 1685 1686 /* If the ioctl happens here, the parent is us. */ 1687 (void)strcpy(dkw->dkw_parent, rs->sc_xname); 1688 return cmd == DIOCAWEDGE ? dkwedge_add(dkw) : dkwedge_del(dkw); 1689 1690 case DIOCLWEDGES: 1691 return dkwedge_list(&rs->sc_dkdev, 1692 (struct dkwedge_list *)data, l); 1693 1694 default: 1695 retcode = ENOTTY; 1696 } 1697 return (retcode); 1698 1699 } 1700 1701 1702 /* raidinit -- complete the rest of the initialization for the 1703 RAIDframe device. */ 1704 1705 1706 static void 1707 raidinit(RF_Raid_t *raidPtr) 1708 { 1709 struct cfdata *cf; 1710 struct raid_softc *rs; 1711 int unit; 1712 1713 unit = raidPtr->raidid; 1714 1715 rs = &raid_softc[unit]; 1716 1717 /* XXX should check return code first... */ 1718 rs->sc_flags |= RAIDF_INITED; 1719 1720 /* XXX doesn't check bounds. */ 1721 snprintf(rs->sc_xname, sizeof(rs->sc_xname), "raid%d", unit); 1722 1723 rs->sc_dkdev.dk_name = rs->sc_xname; 1724 1725 /* attach the pseudo device */ 1726 cf = malloc(sizeof(*cf), M_RAIDFRAME, M_WAITOK); 1727 cf->cf_name = raid_cd.cd_name; 1728 cf->cf_atname = raid_cd.cd_name; 1729 cf->cf_unit = unit; 1730 cf->cf_fstate = FSTATE_STAR; 1731 1732 rs->sc_dev = config_attach_pseudo(cf); 1733 1734 if (rs->sc_dev==NULL) { 1735 printf("raid%d: config_attach_pseudo failed\n", 1736 raidPtr->raidid); 1737 } 1738 1739 /* disk_attach actually creates space for the CPU disklabel, among 1740 * other things, so it's critical to call this *BEFORE* we try putzing 1741 * with disklabels. */ 1742 1743 disk_attach(&rs->sc_dkdev); 1744 1745 /* XXX There may be a weird interaction here between this, and 1746 * protectedSectors, as used in RAIDframe. */ 1747 1748 rs->sc_size = raidPtr->totalSectors; 1749 } 1750 #if (RF_INCLUDE_PARITY_DECLUSTERING_DS > 0) 1751 /* wake up the daemon & tell it to get us a spare table 1752 * XXX 1753 * the entries in the queues should be tagged with the raidPtr 1754 * so that in the extremely rare case that two recons happen at once, 1755 * we know for which device were requesting a spare table 1756 * XXX 1757 * 1758 * XXX This code is not currently used. GO 1759 */ 1760 int 1761 rf_GetSpareTableFromDaemon(RF_SparetWait_t *req) 1762 { 1763 int retcode; 1764 1765 RF_LOCK_MUTEX(rf_sparet_wait_mutex); 1766 req->next = rf_sparet_wait_queue; 1767 rf_sparet_wait_queue = req; 1768 wakeup(&rf_sparet_wait_queue); 1769 1770 /* mpsleep unlocks the mutex */ 1771 while (!rf_sparet_resp_queue) { 1772 tsleep(&rf_sparet_resp_queue, PRIBIO, 1773 "raidframe getsparetable", 0); 1774 } 1775 req = rf_sparet_resp_queue; 1776 rf_sparet_resp_queue = req->next; 1777 RF_UNLOCK_MUTEX(rf_sparet_wait_mutex); 1778 1779 retcode = req->fcol; 1780 RF_Free(req, sizeof(*req)); /* this is not the same req as we 1781 * alloc'd */ 1782 return (retcode); 1783 } 1784 #endif 1785 1786 /* a wrapper around rf_DoAccess that extracts appropriate info from the 1787 * bp & passes it down. 1788 * any calls originating in the kernel must use non-blocking I/O 1789 * do some extra sanity checking to return "appropriate" error values for 1790 * certain conditions (to make some standard utilities work) 1791 * 1792 * Formerly known as: rf_DoAccessKernel 1793 */ 1794 void 1795 raidstart(RF_Raid_t *raidPtr) 1796 { 1797 RF_SectorCount_t num_blocks, pb, sum; 1798 RF_RaidAddr_t raid_addr; 1799 struct partition *pp; 1800 daddr_t blocknum; 1801 int unit; 1802 struct raid_softc *rs; 1803 int do_async; 1804 struct buf *bp; 1805 int rc; 1806 1807 unit = raidPtr->raidid; 1808 rs = &raid_softc[unit]; 1809 1810 /* quick check to see if anything has died recently */ 1811 RF_LOCK_MUTEX(raidPtr->mutex); 1812 if (raidPtr->numNewFailures > 0) { 1813 RF_UNLOCK_MUTEX(raidPtr->mutex); 1814 rf_update_component_labels(raidPtr, 1815 RF_NORMAL_COMPONENT_UPDATE); 1816 RF_LOCK_MUTEX(raidPtr->mutex); 1817 raidPtr->numNewFailures--; 1818 } 1819 1820 /* Check to see if we're at the limit... */ 1821 while (raidPtr->openings > 0) { 1822 RF_UNLOCK_MUTEX(raidPtr->mutex); 1823 1824 /* get the next item, if any, from the queue */ 1825 if ((bp = BUFQ_GET(rs->buf_queue)) == NULL) { 1826 /* nothing more to do */ 1827 return; 1828 } 1829 1830 /* Ok, for the bp we have here, bp->b_blkno is relative to the 1831 * partition.. Need to make it absolute to the underlying 1832 * device.. */ 1833 1834 blocknum = bp->b_blkno; 1835 if (DISKPART(bp->b_dev) != RAW_PART) { 1836 pp = &rs->sc_dkdev.dk_label->d_partitions[DISKPART(bp->b_dev)]; 1837 blocknum += pp->p_offset; 1838 } 1839 1840 db1_printf(("Blocks: %d, %d\n", (int) bp->b_blkno, 1841 (int) blocknum)); 1842 1843 db1_printf(("bp->b_bcount = %d\n", (int) bp->b_bcount)); 1844 db1_printf(("bp->b_resid = %d\n", (int) bp->b_resid)); 1845 1846 /* *THIS* is where we adjust what block we're going to... 1847 * but DO NOT TOUCH bp->b_blkno!!! */ 1848 raid_addr = blocknum; 1849 1850 num_blocks = bp->b_bcount >> raidPtr->logBytesPerSector; 1851 pb = (bp->b_bcount & raidPtr->sectorMask) ? 1 : 0; 1852 sum = raid_addr + num_blocks + pb; 1853 if (1 || rf_debugKernelAccess) { 1854 db1_printf(("raid_addr=%d sum=%d num_blocks=%d(+%d) (%d)\n", 1855 (int) raid_addr, (int) sum, (int) num_blocks, 1856 (int) pb, (int) bp->b_resid)); 1857 } 1858 if ((sum > raidPtr->totalSectors) || (sum < raid_addr) 1859 || (sum < num_blocks) || (sum < pb)) { 1860 bp->b_error = ENOSPC; 1861 bp->b_resid = bp->b_bcount; 1862 biodone(bp); 1863 RF_LOCK_MUTEX(raidPtr->mutex); 1864 continue; 1865 } 1866 /* 1867 * XXX rf_DoAccess() should do this, not just DoAccessKernel() 1868 */ 1869 1870 if (bp->b_bcount & raidPtr->sectorMask) { 1871 bp->b_error = EINVAL; 1872 bp->b_resid = bp->b_bcount; 1873 biodone(bp); 1874 RF_LOCK_MUTEX(raidPtr->mutex); 1875 continue; 1876 1877 } 1878 db1_printf(("Calling DoAccess..\n")); 1879 1880 1881 RF_LOCK_MUTEX(raidPtr->mutex); 1882 raidPtr->openings--; 1883 RF_UNLOCK_MUTEX(raidPtr->mutex); 1884 1885 /* 1886 * Everything is async. 1887 */ 1888 do_async = 1; 1889 1890 disk_busy(&rs->sc_dkdev); 1891 1892 /* XXX we're still at splbio() here... do we *really* 1893 need to be? */ 1894 1895 /* don't ever condition on bp->b_flags & B_WRITE. 1896 * always condition on B_READ instead */ 1897 1898 rc = rf_DoAccess(raidPtr, (bp->b_flags & B_READ) ? 1899 RF_IO_TYPE_READ : RF_IO_TYPE_WRITE, 1900 do_async, raid_addr, num_blocks, 1901 bp->b_data, bp, RF_DAG_NONBLOCKING_IO); 1902 1903 if (rc) { 1904 bp->b_error = rc; 1905 bp->b_resid = bp->b_bcount; 1906 biodone(bp); 1907 /* continue loop */ 1908 } 1909 1910 RF_LOCK_MUTEX(raidPtr->mutex); 1911 } 1912 RF_UNLOCK_MUTEX(raidPtr->mutex); 1913 } 1914 1915 1916 1917 1918 /* invoke an I/O from kernel mode. Disk queue should be locked upon entry */ 1919 1920 int 1921 rf_DispatchKernelIO(RF_DiskQueue_t *queue, RF_DiskQueueData_t *req) 1922 { 1923 int op = (req->type == RF_IO_TYPE_READ) ? B_READ : B_WRITE; 1924 struct buf *bp; 1925 1926 req->queue = queue; 1927 1928 #if DIAGNOSTIC 1929 if (queue->raidPtr->raidid >= numraid) { 1930 printf("Invalid unit number: %d %d\n", queue->raidPtr->raidid, 1931 numraid); 1932 panic("Invalid Unit number in rf_DispatchKernelIO"); 1933 } 1934 #endif 1935 1936 bp = req->bp; 1937 1938 switch (req->type) { 1939 case RF_IO_TYPE_NOP: /* used primarily to unlock a locked queue */ 1940 /* XXX need to do something extra here.. */ 1941 /* I'm leaving this in, as I've never actually seen it used, 1942 * and I'd like folks to report it... GO */ 1943 printf(("WAKEUP CALLED\n")); 1944 queue->numOutstanding++; 1945 1946 bp->b_flags = 0; 1947 bp->b_private = req; 1948 1949 KernelWakeupFunc(bp); 1950 break; 1951 1952 case RF_IO_TYPE_READ: 1953 case RF_IO_TYPE_WRITE: 1954 #if RF_ACC_TRACE > 0 1955 if (req->tracerec) { 1956 RF_ETIMER_START(req->tracerec->timer); 1957 } 1958 #endif 1959 InitBP(bp, queue->rf_cinfo->ci_vp, 1960 op, queue->rf_cinfo->ci_dev, 1961 req->sectorOffset, req->numSector, 1962 req->buf, KernelWakeupFunc, (void *) req, 1963 queue->raidPtr->logBytesPerSector, req->b_proc); 1964 1965 if (rf_debugKernelAccess) { 1966 db1_printf(("dispatch: bp->b_blkno = %ld\n", 1967 (long) bp->b_blkno)); 1968 } 1969 queue->numOutstanding++; 1970 queue->last_deq_sector = req->sectorOffset; 1971 /* acc wouldn't have been let in if there were any pending 1972 * reqs at any other priority */ 1973 queue->curPriority = req->priority; 1974 1975 db1_printf(("Going for %c to unit %d col %d\n", 1976 req->type, queue->raidPtr->raidid, 1977 queue->col)); 1978 db1_printf(("sector %d count %d (%d bytes) %d\n", 1979 (int) req->sectorOffset, (int) req->numSector, 1980 (int) (req->numSector << 1981 queue->raidPtr->logBytesPerSector), 1982 (int) queue->raidPtr->logBytesPerSector)); 1983 VOP_STRATEGY(bp->b_vp, bp); 1984 1985 break; 1986 1987 default: 1988 panic("bad req->type in rf_DispatchKernelIO"); 1989 } 1990 db1_printf(("Exiting from DispatchKernelIO\n")); 1991 1992 return (0); 1993 } 1994 /* this is the callback function associated with a I/O invoked from 1995 kernel code. 1996 */ 1997 static void 1998 KernelWakeupFunc(struct buf *bp) 1999 { 2000 RF_DiskQueueData_t *req = NULL; 2001 RF_DiskQueue_t *queue; 2002 int s; 2003 2004 s = splbio(); 2005 db1_printf(("recovering the request queue:\n")); 2006 req = bp->b_private; 2007 2008 queue = (RF_DiskQueue_t *) req->queue; 2009 2010 #if RF_ACC_TRACE > 0 2011 if (req->tracerec) { 2012 RF_ETIMER_STOP(req->tracerec->timer); 2013 RF_ETIMER_EVAL(req->tracerec->timer); 2014 RF_LOCK_MUTEX(rf_tracing_mutex); 2015 req->tracerec->diskwait_us += RF_ETIMER_VAL_US(req->tracerec->timer); 2016 req->tracerec->phys_io_us += RF_ETIMER_VAL_US(req->tracerec->timer); 2017 req->tracerec->num_phys_ios++; 2018 RF_UNLOCK_MUTEX(rf_tracing_mutex); 2019 } 2020 #endif 2021 2022 /* XXX Ok, let's get aggressive... If b_error is set, let's go 2023 * ballistic, and mark the component as hosed... */ 2024 2025 if (bp->b_error != 0) { 2026 /* Mark the disk as dead */ 2027 /* but only mark it once... */ 2028 /* and only if it wouldn't leave this RAID set 2029 completely broken */ 2030 if (((queue->raidPtr->Disks[queue->col].status == 2031 rf_ds_optimal) || 2032 (queue->raidPtr->Disks[queue->col].status == 2033 rf_ds_used_spare)) && 2034 (queue->raidPtr->numFailures < 2035 queue->raidPtr->Layout.map->faultsTolerated)) { 2036 printf("raid%d: IO Error. Marking %s as failed.\n", 2037 queue->raidPtr->raidid, 2038 queue->raidPtr->Disks[queue->col].devname); 2039 queue->raidPtr->Disks[queue->col].status = 2040 rf_ds_failed; 2041 queue->raidPtr->status = rf_rs_degraded; 2042 queue->raidPtr->numFailures++; 2043 queue->raidPtr->numNewFailures++; 2044 } else { /* Disk is already dead... */ 2045 /* printf("Disk already marked as dead!\n"); */ 2046 } 2047 2048 } 2049 2050 /* Fill in the error value */ 2051 2052 req->error = bp->b_error; 2053 2054 simple_lock(&queue->raidPtr->iodone_lock); 2055 2056 /* Drop this one on the "finished" queue... */ 2057 TAILQ_INSERT_TAIL(&(queue->raidPtr->iodone), req, iodone_entries); 2058 2059 /* Let the raidio thread know there is work to be done. */ 2060 wakeup(&(queue->raidPtr->iodone)); 2061 2062 simple_unlock(&queue->raidPtr->iodone_lock); 2063 2064 splx(s); 2065 } 2066 2067 2068 2069 /* 2070 * initialize a buf structure for doing an I/O in the kernel. 2071 */ 2072 static void 2073 InitBP(struct buf *bp, struct vnode *b_vp, unsigned rw_flag, dev_t dev, 2074 RF_SectorNum_t startSect, RF_SectorCount_t numSect, void *bf, 2075 void (*cbFunc) (struct buf *), void *cbArg, int logBytesPerSector, 2076 struct proc *b_proc) 2077 { 2078 /* bp->b_flags = B_PHYS | rw_flag; */ 2079 bp->b_flags = B_CALL | rw_flag; /* XXX need B_PHYS here too??? */ 2080 bp->b_bcount = numSect << logBytesPerSector; 2081 bp->b_bufsize = bp->b_bcount; 2082 bp->b_error = 0; 2083 bp->b_dev = dev; 2084 bp->b_data = bf; 2085 bp->b_blkno = startSect; 2086 bp->b_resid = bp->b_bcount; /* XXX is this right!??!?!! */ 2087 if (bp->b_bcount == 0) { 2088 panic("bp->b_bcount is zero in InitBP!!"); 2089 } 2090 bp->b_proc = b_proc; 2091 bp->b_iodone = cbFunc; 2092 bp->b_private = cbArg; 2093 bp->b_vp = b_vp; 2094 if ((bp->b_flags & B_READ) == 0) { 2095 bp->b_vp->v_numoutput++; 2096 } 2097 2098 } 2099 2100 static void 2101 raidgetdefaultlabel(RF_Raid_t *raidPtr, struct raid_softc *rs, 2102 struct disklabel *lp) 2103 { 2104 memset(lp, 0, sizeof(*lp)); 2105 2106 /* fabricate a label... */ 2107 lp->d_secperunit = raidPtr->totalSectors; 2108 lp->d_secsize = raidPtr->bytesPerSector; 2109 lp->d_nsectors = raidPtr->Layout.dataSectorsPerStripe; 2110 lp->d_ntracks = 4 * raidPtr->numCol; 2111 lp->d_ncylinders = raidPtr->totalSectors / 2112 (lp->d_nsectors * lp->d_ntracks); 2113 lp->d_secpercyl = lp->d_ntracks * lp->d_nsectors; 2114 2115 strncpy(lp->d_typename, "raid", sizeof(lp->d_typename)); 2116 lp->d_type = DTYPE_RAID; 2117 strncpy(lp->d_packname, "fictitious", sizeof(lp->d_packname)); 2118 lp->d_rpm = 3600; 2119 lp->d_interleave = 1; 2120 lp->d_flags = 0; 2121 2122 lp->d_partitions[RAW_PART].p_offset = 0; 2123 lp->d_partitions[RAW_PART].p_size = raidPtr->totalSectors; 2124 lp->d_partitions[RAW_PART].p_fstype = FS_UNUSED; 2125 lp->d_npartitions = RAW_PART + 1; 2126 2127 lp->d_magic = DISKMAGIC; 2128 lp->d_magic2 = DISKMAGIC; 2129 lp->d_checksum = dkcksum(rs->sc_dkdev.dk_label); 2130 2131 } 2132 /* 2133 * Read the disklabel from the raid device. If one is not present, fake one 2134 * up. 2135 */ 2136 static void 2137 raidgetdisklabel(dev_t dev) 2138 { 2139 int unit = raidunit(dev); 2140 struct raid_softc *rs = &raid_softc[unit]; 2141 const char *errstring; 2142 struct disklabel *lp = rs->sc_dkdev.dk_label; 2143 struct cpu_disklabel *clp = rs->sc_dkdev.dk_cpulabel; 2144 RF_Raid_t *raidPtr; 2145 2146 db1_printf(("Getting the disklabel...\n")); 2147 2148 memset(clp, 0, sizeof(*clp)); 2149 2150 raidPtr = raidPtrs[unit]; 2151 2152 raidgetdefaultlabel(raidPtr, rs, lp); 2153 2154 /* 2155 * Call the generic disklabel extraction routine. 2156 */ 2157 errstring = readdisklabel(RAIDLABELDEV(dev), raidstrategy, 2158 rs->sc_dkdev.dk_label, rs->sc_dkdev.dk_cpulabel); 2159 if (errstring) 2160 raidmakedisklabel(rs); 2161 else { 2162 int i; 2163 struct partition *pp; 2164 2165 /* 2166 * Sanity check whether the found disklabel is valid. 2167 * 2168 * This is necessary since total size of the raid device 2169 * may vary when an interleave is changed even though exactly 2170 * same components are used, and old disklabel may used 2171 * if that is found. 2172 */ 2173 if (lp->d_secperunit != rs->sc_size) 2174 printf("raid%d: WARNING: %s: " 2175 "total sector size in disklabel (%d) != " 2176 "the size of raid (%ld)\n", unit, rs->sc_xname, 2177 lp->d_secperunit, (long) rs->sc_size); 2178 for (i = 0; i < lp->d_npartitions; i++) { 2179 pp = &lp->d_partitions[i]; 2180 if (pp->p_offset + pp->p_size > rs->sc_size) 2181 printf("raid%d: WARNING: %s: end of partition `%c' " 2182 "exceeds the size of raid (%ld)\n", 2183 unit, rs->sc_xname, 'a' + i, (long) rs->sc_size); 2184 } 2185 } 2186 2187 } 2188 /* 2189 * Take care of things one might want to take care of in the event 2190 * that a disklabel isn't present. 2191 */ 2192 static void 2193 raidmakedisklabel(struct raid_softc *rs) 2194 { 2195 struct disklabel *lp = rs->sc_dkdev.dk_label; 2196 db1_printf(("Making a label..\n")); 2197 2198 /* 2199 * For historical reasons, if there's no disklabel present 2200 * the raw partition must be marked FS_BSDFFS. 2201 */ 2202 2203 lp->d_partitions[RAW_PART].p_fstype = FS_BSDFFS; 2204 2205 strncpy(lp->d_packname, "default label", sizeof(lp->d_packname)); 2206 2207 lp->d_checksum = dkcksum(lp); 2208 } 2209 /* 2210 * Wait interruptibly for an exclusive lock. 2211 * 2212 * XXX 2213 * Several drivers do this; it should be abstracted and made MP-safe. 2214 * (Hmm... where have we seen this warning before :-> GO ) 2215 */ 2216 static int 2217 raidlock(struct raid_softc *rs) 2218 { 2219 int error; 2220 2221 while ((rs->sc_flags & RAIDF_LOCKED) != 0) { 2222 rs->sc_flags |= RAIDF_WANTED; 2223 if ((error = 2224 tsleep(rs, PRIBIO | PCATCH, "raidlck", 0)) != 0) 2225 return (error); 2226 } 2227 rs->sc_flags |= RAIDF_LOCKED; 2228 return (0); 2229 } 2230 /* 2231 * Unlock and wake up any waiters. 2232 */ 2233 static void 2234 raidunlock(struct raid_softc *rs) 2235 { 2236 2237 rs->sc_flags &= ~RAIDF_LOCKED; 2238 if ((rs->sc_flags & RAIDF_WANTED) != 0) { 2239 rs->sc_flags &= ~RAIDF_WANTED; 2240 wakeup(rs); 2241 } 2242 } 2243 2244 2245 #define RF_COMPONENT_INFO_OFFSET 16384 /* bytes */ 2246 #define RF_COMPONENT_INFO_SIZE 1024 /* bytes */ 2247 2248 int 2249 raidmarkclean(dev_t dev, struct vnode *b_vp, int mod_counter) 2250 { 2251 RF_ComponentLabel_t clabel; 2252 raidread_component_label(dev, b_vp, &clabel); 2253 clabel.mod_counter = mod_counter; 2254 clabel.clean = RF_RAID_CLEAN; 2255 raidwrite_component_label(dev, b_vp, &clabel); 2256 return(0); 2257 } 2258 2259 2260 int 2261 raidmarkdirty(dev_t dev, struct vnode *b_vp, int mod_counter) 2262 { 2263 RF_ComponentLabel_t clabel; 2264 raidread_component_label(dev, b_vp, &clabel); 2265 clabel.mod_counter = mod_counter; 2266 clabel.clean = RF_RAID_DIRTY; 2267 raidwrite_component_label(dev, b_vp, &clabel); 2268 return(0); 2269 } 2270 2271 /* ARGSUSED */ 2272 int 2273 raidread_component_label(dev_t dev, struct vnode *b_vp, 2274 RF_ComponentLabel_t *clabel) 2275 { 2276 struct buf *bp; 2277 const struct bdevsw *bdev; 2278 int error; 2279 2280 /* XXX should probably ensure that we don't try to do this if 2281 someone has changed rf_protected_sectors. */ 2282 2283 if (b_vp == NULL) { 2284 /* For whatever reason, this component is not valid. 2285 Don't try to read a component label from it. */ 2286 return(EINVAL); 2287 } 2288 2289 /* get a block of the appropriate size... */ 2290 bp = geteblk((int)RF_COMPONENT_INFO_SIZE); 2291 bp->b_dev = dev; 2292 2293 /* get our ducks in a row for the read */ 2294 bp->b_blkno = RF_COMPONENT_INFO_OFFSET / DEV_BSIZE; 2295 bp->b_bcount = RF_COMPONENT_INFO_SIZE; 2296 bp->b_flags |= B_READ; 2297 bp->b_resid = RF_COMPONENT_INFO_SIZE / DEV_BSIZE; 2298 2299 bdev = bdevsw_lookup(bp->b_dev); 2300 if (bdev == NULL) 2301 return (ENXIO); 2302 (*bdev->d_strategy)(bp); 2303 2304 error = biowait(bp); 2305 2306 if (!error) { 2307 memcpy(clabel, bp->b_data, 2308 sizeof(RF_ComponentLabel_t)); 2309 } 2310 2311 brelse(bp, 0); 2312 return(error); 2313 } 2314 /* ARGSUSED */ 2315 int 2316 raidwrite_component_label(dev_t dev, struct vnode *b_vp, 2317 RF_ComponentLabel_t *clabel) 2318 { 2319 struct buf *bp; 2320 const struct bdevsw *bdev; 2321 int error; 2322 2323 /* get a block of the appropriate size... */ 2324 bp = geteblk((int)RF_COMPONENT_INFO_SIZE); 2325 bp->b_dev = dev; 2326 2327 /* get our ducks in a row for the write */ 2328 bp->b_blkno = RF_COMPONENT_INFO_OFFSET / DEV_BSIZE; 2329 bp->b_bcount = RF_COMPONENT_INFO_SIZE; 2330 bp->b_flags |= B_WRITE; 2331 bp->b_resid = RF_COMPONENT_INFO_SIZE / DEV_BSIZE; 2332 2333 memset(bp->b_data, 0, RF_COMPONENT_INFO_SIZE ); 2334 2335 memcpy(bp->b_data, clabel, sizeof(RF_ComponentLabel_t)); 2336 2337 bdev = bdevsw_lookup(bp->b_dev); 2338 if (bdev == NULL) 2339 return (ENXIO); 2340 (*bdev->d_strategy)(bp); 2341 error = biowait(bp); 2342 brelse(bp, 0); 2343 if (error) { 2344 #if 1 2345 printf("Failed to write RAID component info!\n"); 2346 #endif 2347 } 2348 2349 return(error); 2350 } 2351 2352 void 2353 rf_markalldirty(RF_Raid_t *raidPtr) 2354 { 2355 RF_ComponentLabel_t clabel; 2356 int sparecol; 2357 int c; 2358 int j; 2359 int scol = -1; 2360 2361 raidPtr->mod_counter++; 2362 for (c = 0; c < raidPtr->numCol; c++) { 2363 /* we don't want to touch (at all) a disk that has 2364 failed */ 2365 if (!RF_DEAD_DISK(raidPtr->Disks[c].status)) { 2366 raidread_component_label( 2367 raidPtr->Disks[c].dev, 2368 raidPtr->raid_cinfo[c].ci_vp, 2369 &clabel); 2370 if (clabel.status == rf_ds_spared) { 2371 /* XXX do something special... 2372 but whatever you do, don't 2373 try to access it!! */ 2374 } else { 2375 raidmarkdirty( 2376 raidPtr->Disks[c].dev, 2377 raidPtr->raid_cinfo[c].ci_vp, 2378 raidPtr->mod_counter); 2379 } 2380 } 2381 } 2382 2383 for( c = 0; c < raidPtr->numSpare ; c++) { 2384 sparecol = raidPtr->numCol + c; 2385 if (raidPtr->Disks[sparecol].status == rf_ds_used_spare) { 2386 /* 2387 2388 we claim this disk is "optimal" if it's 2389 rf_ds_used_spare, as that means it should be 2390 directly substitutable for the disk it replaced. 2391 We note that too... 2392 2393 */ 2394 2395 for(j=0;j<raidPtr->numCol;j++) { 2396 if (raidPtr->Disks[j].spareCol == sparecol) { 2397 scol = j; 2398 break; 2399 } 2400 } 2401 2402 raidread_component_label( 2403 raidPtr->Disks[sparecol].dev, 2404 raidPtr->raid_cinfo[sparecol].ci_vp, 2405 &clabel); 2406 /* make sure status is noted */ 2407 2408 raid_init_component_label(raidPtr, &clabel); 2409 2410 clabel.row = 0; 2411 clabel.column = scol; 2412 /* Note: we *don't* change status from rf_ds_used_spare 2413 to rf_ds_optimal */ 2414 /* clabel.status = rf_ds_optimal; */ 2415 2416 raidmarkdirty(raidPtr->Disks[sparecol].dev, 2417 raidPtr->raid_cinfo[sparecol].ci_vp, 2418 raidPtr->mod_counter); 2419 } 2420 } 2421 } 2422 2423 2424 void 2425 rf_update_component_labels(RF_Raid_t *raidPtr, int final) 2426 { 2427 RF_ComponentLabel_t clabel; 2428 int sparecol; 2429 int c; 2430 int j; 2431 int scol; 2432 2433 scol = -1; 2434 2435 /* XXX should do extra checks to make sure things really are clean, 2436 rather than blindly setting the clean bit... */ 2437 2438 raidPtr->mod_counter++; 2439 2440 for (c = 0; c < raidPtr->numCol; c++) { 2441 if (raidPtr->Disks[c].status == rf_ds_optimal) { 2442 raidread_component_label( 2443 raidPtr->Disks[c].dev, 2444 raidPtr->raid_cinfo[c].ci_vp, 2445 &clabel); 2446 /* make sure status is noted */ 2447 clabel.status = rf_ds_optimal; 2448 2449 /* bump the counter */ 2450 clabel.mod_counter = raidPtr->mod_counter; 2451 2452 /* note what unit we are configured as */ 2453 clabel.last_unit = raidPtr->raidid; 2454 2455 raidwrite_component_label( 2456 raidPtr->Disks[c].dev, 2457 raidPtr->raid_cinfo[c].ci_vp, 2458 &clabel); 2459 if (final == RF_FINAL_COMPONENT_UPDATE) { 2460 if (raidPtr->parity_good == RF_RAID_CLEAN) { 2461 raidmarkclean( 2462 raidPtr->Disks[c].dev, 2463 raidPtr->raid_cinfo[c].ci_vp, 2464 raidPtr->mod_counter); 2465 } 2466 } 2467 } 2468 /* else we don't touch it.. */ 2469 } 2470 2471 for( c = 0; c < raidPtr->numSpare ; c++) { 2472 sparecol = raidPtr->numCol + c; 2473 /* Need to ensure that the reconstruct actually completed! */ 2474 if (raidPtr->Disks[sparecol].status == rf_ds_used_spare) { 2475 /* 2476 2477 we claim this disk is "optimal" if it's 2478 rf_ds_used_spare, as that means it should be 2479 directly substitutable for the disk it replaced. 2480 We note that too... 2481 2482 */ 2483 2484 for(j=0;j<raidPtr->numCol;j++) { 2485 if (raidPtr->Disks[j].spareCol == sparecol) { 2486 scol = j; 2487 break; 2488 } 2489 } 2490 2491 /* XXX shouldn't *really* need this... */ 2492 raidread_component_label( 2493 raidPtr->Disks[sparecol].dev, 2494 raidPtr->raid_cinfo[sparecol].ci_vp, 2495 &clabel); 2496 /* make sure status is noted */ 2497 2498 raid_init_component_label(raidPtr, &clabel); 2499 2500 clabel.mod_counter = raidPtr->mod_counter; 2501 clabel.column = scol; 2502 clabel.status = rf_ds_optimal; 2503 clabel.last_unit = raidPtr->raidid; 2504 2505 raidwrite_component_label( 2506 raidPtr->Disks[sparecol].dev, 2507 raidPtr->raid_cinfo[sparecol].ci_vp, 2508 &clabel); 2509 if (final == RF_FINAL_COMPONENT_UPDATE) { 2510 if (raidPtr->parity_good == RF_RAID_CLEAN) { 2511 raidmarkclean( raidPtr->Disks[sparecol].dev, 2512 raidPtr->raid_cinfo[sparecol].ci_vp, 2513 raidPtr->mod_counter); 2514 } 2515 } 2516 } 2517 } 2518 } 2519 2520 void 2521 rf_close_component(RF_Raid_t *raidPtr, struct vnode *vp, int auto_configured) 2522 { 2523 struct proc *p; 2524 struct lwp *l; 2525 2526 p = raidPtr->engine_thread; 2527 l = LIST_FIRST(&p->p_lwps); 2528 2529 if (vp != NULL) { 2530 if (auto_configured == 1) { 2531 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 2532 VOP_CLOSE(vp, FREAD | FWRITE, NOCRED, 0); 2533 vput(vp); 2534 2535 } else { 2536 (void) vn_close(vp, FREAD | FWRITE, p->p_cred, l); 2537 } 2538 } 2539 } 2540 2541 2542 void 2543 rf_UnconfigureVnodes(RF_Raid_t *raidPtr) 2544 { 2545 int r,c; 2546 struct vnode *vp; 2547 int acd; 2548 2549 2550 /* We take this opportunity to close the vnodes like we should.. */ 2551 2552 for (c = 0; c < raidPtr->numCol; c++) { 2553 vp = raidPtr->raid_cinfo[c].ci_vp; 2554 acd = raidPtr->Disks[c].auto_configured; 2555 rf_close_component(raidPtr, vp, acd); 2556 raidPtr->raid_cinfo[c].ci_vp = NULL; 2557 raidPtr->Disks[c].auto_configured = 0; 2558 } 2559 2560 for (r = 0; r < raidPtr->numSpare; r++) { 2561 vp = raidPtr->raid_cinfo[raidPtr->numCol + r].ci_vp; 2562 acd = raidPtr->Disks[raidPtr->numCol + r].auto_configured; 2563 rf_close_component(raidPtr, vp, acd); 2564 raidPtr->raid_cinfo[raidPtr->numCol + r].ci_vp = NULL; 2565 raidPtr->Disks[raidPtr->numCol + r].auto_configured = 0; 2566 } 2567 } 2568 2569 2570 void 2571 rf_ReconThread(struct rf_recon_req *req) 2572 { 2573 int s; 2574 RF_Raid_t *raidPtr; 2575 2576 s = splbio(); 2577 raidPtr = (RF_Raid_t *) req->raidPtr; 2578 raidPtr->recon_in_progress = 1; 2579 2580 rf_FailDisk((RF_Raid_t *) req->raidPtr, req->col, 2581 ((req->flags & RF_FDFLAGS_RECON) ? 1 : 0)); 2582 2583 RF_Free(req, sizeof(*req)); 2584 2585 raidPtr->recon_in_progress = 0; 2586 splx(s); 2587 2588 /* That's all... */ 2589 kthread_exit(0); /* does not return */ 2590 } 2591 2592 void 2593 rf_RewriteParityThread(RF_Raid_t *raidPtr) 2594 { 2595 int retcode; 2596 int s; 2597 2598 raidPtr->parity_rewrite_stripes_done = 0; 2599 raidPtr->parity_rewrite_in_progress = 1; 2600 s = splbio(); 2601 retcode = rf_RewriteParity(raidPtr); 2602 splx(s); 2603 if (retcode) { 2604 printf("raid%d: Error re-writing parity!\n",raidPtr->raidid); 2605 } else { 2606 /* set the clean bit! If we shutdown correctly, 2607 the clean bit on each component label will get 2608 set */ 2609 raidPtr->parity_good = RF_RAID_CLEAN; 2610 } 2611 raidPtr->parity_rewrite_in_progress = 0; 2612 2613 /* Anyone waiting for us to stop? If so, inform them... */ 2614 if (raidPtr->waitShutdown) { 2615 wakeup(&raidPtr->parity_rewrite_in_progress); 2616 } 2617 2618 /* That's all... */ 2619 kthread_exit(0); /* does not return */ 2620 } 2621 2622 2623 void 2624 rf_CopybackThread(RF_Raid_t *raidPtr) 2625 { 2626 int s; 2627 2628 raidPtr->copyback_in_progress = 1; 2629 s = splbio(); 2630 rf_CopybackReconstructedData(raidPtr); 2631 splx(s); 2632 raidPtr->copyback_in_progress = 0; 2633 2634 /* That's all... */ 2635 kthread_exit(0); /* does not return */ 2636 } 2637 2638 2639 void 2640 rf_ReconstructInPlaceThread(struct rf_recon_req *req) 2641 { 2642 int s; 2643 RF_Raid_t *raidPtr; 2644 2645 s = splbio(); 2646 raidPtr = req->raidPtr; 2647 raidPtr->recon_in_progress = 1; 2648 rf_ReconstructInPlace(raidPtr, req->col); 2649 RF_Free(req, sizeof(*req)); 2650 raidPtr->recon_in_progress = 0; 2651 splx(s); 2652 2653 /* That's all... */ 2654 kthread_exit(0); /* does not return */ 2655 } 2656 2657 static RF_AutoConfig_t * 2658 rf_get_component(RF_AutoConfig_t *ac_list, dev_t dev, struct vnode *vp, 2659 const char *cname, RF_SectorCount_t size) 2660 { 2661 int good_one = 0; 2662 RF_ComponentLabel_t *clabel; 2663 RF_AutoConfig_t *ac; 2664 2665 clabel = malloc(sizeof(RF_ComponentLabel_t), M_RAIDFRAME, M_NOWAIT); 2666 if (clabel == NULL) { 2667 oomem: 2668 while(ac_list) { 2669 ac = ac_list; 2670 if (ac->clabel) 2671 free(ac->clabel, M_RAIDFRAME); 2672 ac_list = ac_list->next; 2673 free(ac, M_RAIDFRAME); 2674 } 2675 printf("RAID auto config: out of memory!\n"); 2676 return NULL; /* XXX probably should panic? */ 2677 } 2678 2679 if (!raidread_component_label(dev, vp, clabel)) { 2680 /* Got the label. Does it look reasonable? */ 2681 if (rf_reasonable_label(clabel) && 2682 (clabel->partitionSize <= size)) { 2683 #ifdef DEBUG 2684 printf("Component on: %s: %llu\n", 2685 cname, (unsigned long long)size); 2686 rf_print_component_label(clabel); 2687 #endif 2688 /* if it's reasonable, add it, else ignore it. */ 2689 ac = malloc(sizeof(RF_AutoConfig_t), M_RAIDFRAME, 2690 M_NOWAIT); 2691 if (ac == NULL) { 2692 free(clabel, M_RAIDFRAME); 2693 goto oomem; 2694 } 2695 strlcpy(ac->devname, cname, sizeof(ac->devname)); 2696 ac->dev = dev; 2697 ac->vp = vp; 2698 ac->clabel = clabel; 2699 ac->next = ac_list; 2700 ac_list = ac; 2701 good_one = 1; 2702 } 2703 } 2704 if (!good_one) { 2705 /* cleanup */ 2706 free(clabel, M_RAIDFRAME); 2707 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 2708 VOP_CLOSE(vp, FREAD | FWRITE, NOCRED, 0); 2709 vput(vp); 2710 } 2711 return ac_list; 2712 } 2713 2714 RF_AutoConfig_t * 2715 rf_find_raid_components() 2716 { 2717 struct vnode *vp; 2718 struct disklabel label; 2719 struct device *dv; 2720 dev_t dev; 2721 int bmajor, bminor, wedge; 2722 int error; 2723 int i; 2724 RF_AutoConfig_t *ac_list; 2725 2726 2727 /* initialize the AutoConfig list */ 2728 ac_list = NULL; 2729 2730 /* we begin by trolling through *all* the devices on the system */ 2731 2732 for (dv = alldevs.tqh_first; dv != NULL; 2733 dv = dv->dv_list.tqe_next) { 2734 2735 /* we are only interested in disks... */ 2736 if (device_class(dv) != DV_DISK) 2737 continue; 2738 2739 /* we don't care about floppies... */ 2740 if (device_is_a(dv, "fd")) { 2741 continue; 2742 } 2743 2744 /* we don't care about CD's... */ 2745 if (device_is_a(dv, "cd")) { 2746 continue; 2747 } 2748 2749 /* hdfd is the Atari/Hades floppy driver */ 2750 if (device_is_a(dv, "hdfd")) { 2751 continue; 2752 } 2753 2754 /* fdisa is the Atari/Milan floppy driver */ 2755 if (device_is_a(dv, "fdisa")) { 2756 continue; 2757 } 2758 2759 /* need to find the device_name_to_block_device_major stuff */ 2760 bmajor = devsw_name2blk(dv->dv_xname, NULL, 0); 2761 2762 /* get a vnode for the raw partition of this disk */ 2763 2764 wedge = device_is_a(dv, "dk"); 2765 bminor = minor(device_unit(dv)); 2766 dev = wedge ? makedev(bmajor, bminor) : 2767 MAKEDISKDEV(bmajor, bminor, RAW_PART); 2768 if (bdevvp(dev, &vp)) 2769 panic("RAID can't alloc vnode"); 2770 2771 error = VOP_OPEN(vp, FREAD, NOCRED, 0); 2772 2773 if (error) { 2774 /* "Who cares." Continue looking 2775 for something that exists*/ 2776 vput(vp); 2777 continue; 2778 } 2779 2780 if (wedge) { 2781 struct dkwedge_info dkw; 2782 error = VOP_IOCTL(vp, DIOCGWEDGEINFO, &dkw, FREAD, 2783 NOCRED, 0); 2784 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 2785 VOP_CLOSE(vp, FREAD | FWRITE, NOCRED, 0); 2786 vput(vp); 2787 if (error) { 2788 printf("RAIDframe: can't get wedge info for " 2789 "dev %s (%d)\n", dv->dv_xname, error); 2790 continue; 2791 } 2792 2793 if (strcmp(dkw.dkw_ptype, DKW_PTYPE_RAIDFRAME) != 0) 2794 continue; 2795 2796 ac_list = rf_get_component(ac_list, dev, vp, 2797 dv->dv_xname, dkw.dkw_size); 2798 continue; 2799 } 2800 2801 /* Ok, the disk exists. Go get the disklabel. */ 2802 error = VOP_IOCTL(vp, DIOCGDINFO, &label, FREAD, NOCRED, 0); 2803 if (error) { 2804 /* 2805 * XXX can't happen - open() would 2806 * have errored out (or faked up one) 2807 */ 2808 if (error != ENOTTY) 2809 printf("RAIDframe: can't get label for dev " 2810 "%s (%d)\n", dv->dv_xname, error); 2811 } 2812 2813 /* don't need this any more. We'll allocate it again 2814 a little later if we really do... */ 2815 vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); 2816 VOP_CLOSE(vp, FREAD | FWRITE, NOCRED, 0); 2817 vput(vp); 2818 2819 if (error) 2820 continue; 2821 2822 for (i = 0; i < label.d_npartitions; i++) { 2823 char cname[sizeof(ac_list->devname)]; 2824 2825 /* We only support partitions marked as RAID */ 2826 if (label.d_partitions[i].p_fstype != FS_RAID) 2827 continue; 2828 2829 dev = MAKEDISKDEV(bmajor, device_unit(dv), i); 2830 if (bdevvp(dev, &vp)) 2831 panic("RAID can't alloc vnode"); 2832 2833 error = VOP_OPEN(vp, FREAD, NOCRED, 0); 2834 if (error) { 2835 /* Whatever... */ 2836 vput(vp); 2837 continue; 2838 } 2839 snprintf(cname, sizeof(cname), "%s%c", 2840 dv->dv_xname, 'a' + i); 2841 ac_list = rf_get_component(ac_list, dev, vp, cname, 2842 label.d_partitions[i].p_size); 2843 } 2844 } 2845 return ac_list; 2846 } 2847 2848 2849 static int 2850 rf_reasonable_label(RF_ComponentLabel_t *clabel) 2851 { 2852 2853 if (((clabel->version==RF_COMPONENT_LABEL_VERSION_1) || 2854 (clabel->version==RF_COMPONENT_LABEL_VERSION)) && 2855 ((clabel->clean == RF_RAID_CLEAN) || 2856 (clabel->clean == RF_RAID_DIRTY)) && 2857 clabel->row >=0 && 2858 clabel->column >= 0 && 2859 clabel->num_rows > 0 && 2860 clabel->num_columns > 0 && 2861 clabel->row < clabel->num_rows && 2862 clabel->column < clabel->num_columns && 2863 clabel->blockSize > 0 && 2864 clabel->numBlocks > 0) { 2865 /* label looks reasonable enough... */ 2866 return(1); 2867 } 2868 return(0); 2869 } 2870 2871 2872 #ifdef DEBUG 2873 void 2874 rf_print_component_label(RF_ComponentLabel_t *clabel) 2875 { 2876 printf(" Row: %d Column: %d Num Rows: %d Num Columns: %d\n", 2877 clabel->row, clabel->column, 2878 clabel->num_rows, clabel->num_columns); 2879 printf(" Version: %d Serial Number: %d Mod Counter: %d\n", 2880 clabel->version, clabel->serial_number, 2881 clabel->mod_counter); 2882 printf(" Clean: %s Status: %d\n", 2883 clabel->clean ? "Yes" : "No", clabel->status ); 2884 printf(" sectPerSU: %d SUsPerPU: %d SUsPerRU: %d\n", 2885 clabel->sectPerSU, clabel->SUsPerPU, clabel->SUsPerRU); 2886 printf(" RAID Level: %c blocksize: %d numBlocks: %d\n", 2887 (char) clabel->parityConfig, clabel->blockSize, 2888 clabel->numBlocks); 2889 printf(" Autoconfig: %s\n", clabel->autoconfigure ? "Yes" : "No" ); 2890 printf(" Contains root partition: %s\n", 2891 clabel->root_partition ? "Yes" : "No" ); 2892 printf(" Last configured as: raid%d\n", clabel->last_unit ); 2893 #if 0 2894 printf(" Config order: %d\n", clabel->config_order); 2895 #endif 2896 2897 } 2898 #endif 2899 2900 RF_ConfigSet_t * 2901 rf_create_auto_sets(RF_AutoConfig_t *ac_list) 2902 { 2903 RF_AutoConfig_t *ac; 2904 RF_ConfigSet_t *config_sets; 2905 RF_ConfigSet_t *cset; 2906 RF_AutoConfig_t *ac_next; 2907 2908 2909 config_sets = NULL; 2910 2911 /* Go through the AutoConfig list, and figure out which components 2912 belong to what sets. */ 2913 ac = ac_list; 2914 while(ac!=NULL) { 2915 /* we're going to putz with ac->next, so save it here 2916 for use at the end of the loop */ 2917 ac_next = ac->next; 2918 2919 if (config_sets == NULL) { 2920 /* will need at least this one... */ 2921 config_sets = (RF_ConfigSet_t *) 2922 malloc(sizeof(RF_ConfigSet_t), 2923 M_RAIDFRAME, M_NOWAIT); 2924 if (config_sets == NULL) { 2925 panic("rf_create_auto_sets: No memory!"); 2926 } 2927 /* this one is easy :) */ 2928 config_sets->ac = ac; 2929 config_sets->next = NULL; 2930 config_sets->rootable = 0; 2931 ac->next = NULL; 2932 } else { 2933 /* which set does this component fit into? */ 2934 cset = config_sets; 2935 while(cset!=NULL) { 2936 if (rf_does_it_fit(cset, ac)) { 2937 /* looks like it matches... */ 2938 ac->next = cset->ac; 2939 cset->ac = ac; 2940 break; 2941 } 2942 cset = cset->next; 2943 } 2944 if (cset==NULL) { 2945 /* didn't find a match above... new set..*/ 2946 cset = (RF_ConfigSet_t *) 2947 malloc(sizeof(RF_ConfigSet_t), 2948 M_RAIDFRAME, M_NOWAIT); 2949 if (cset == NULL) { 2950 panic("rf_create_auto_sets: No memory!"); 2951 } 2952 cset->ac = ac; 2953 ac->next = NULL; 2954 cset->next = config_sets; 2955 cset->rootable = 0; 2956 config_sets = cset; 2957 } 2958 } 2959 ac = ac_next; 2960 } 2961 2962 2963 return(config_sets); 2964 } 2965 2966 static int 2967 rf_does_it_fit(RF_ConfigSet_t *cset, RF_AutoConfig_t *ac) 2968 { 2969 RF_ComponentLabel_t *clabel1, *clabel2; 2970 2971 /* If this one matches the *first* one in the set, that's good 2972 enough, since the other members of the set would have been 2973 through here too... */ 2974 /* note that we are not checking partitionSize here.. 2975 2976 Note that we are also not checking the mod_counters here. 2977 If everything else matches execpt the mod_counter, that's 2978 good enough for this test. We will deal with the mod_counters 2979 a little later in the autoconfiguration process. 2980 2981 (clabel1->mod_counter == clabel2->mod_counter) && 2982 2983 The reason we don't check for this is that failed disks 2984 will have lower modification counts. If those disks are 2985 not added to the set they used to belong to, then they will 2986 form their own set, which may result in 2 different sets, 2987 for example, competing to be configured at raid0, and 2988 perhaps competing to be the root filesystem set. If the 2989 wrong ones get configured, or both attempt to become /, 2990 weird behaviour and or serious lossage will occur. Thus we 2991 need to bring them into the fold here, and kick them out at 2992 a later point. 2993 2994 */ 2995 2996 clabel1 = cset->ac->clabel; 2997 clabel2 = ac->clabel; 2998 if ((clabel1->version == clabel2->version) && 2999 (clabel1->serial_number == clabel2->serial_number) && 3000 (clabel1->num_rows == clabel2->num_rows) && 3001 (clabel1->num_columns == clabel2->num_columns) && 3002 (clabel1->sectPerSU == clabel2->sectPerSU) && 3003 (clabel1->SUsPerPU == clabel2->SUsPerPU) && 3004 (clabel1->SUsPerRU == clabel2->SUsPerRU) && 3005 (clabel1->parityConfig == clabel2->parityConfig) && 3006 (clabel1->maxOutstanding == clabel2->maxOutstanding) && 3007 (clabel1->blockSize == clabel2->blockSize) && 3008 (clabel1->numBlocks == clabel2->numBlocks) && 3009 (clabel1->autoconfigure == clabel2->autoconfigure) && 3010 (clabel1->root_partition == clabel2->root_partition) && 3011 (clabel1->last_unit == clabel2->last_unit) && 3012 (clabel1->config_order == clabel2->config_order)) { 3013 /* if it get's here, it almost *has* to be a match */ 3014 } else { 3015 /* it's not consistent with somebody in the set.. 3016 punt */ 3017 return(0); 3018 } 3019 /* all was fine.. it must fit... */ 3020 return(1); 3021 } 3022 3023 int 3024 rf_have_enough_components(RF_ConfigSet_t *cset) 3025 { 3026 RF_AutoConfig_t *ac; 3027 RF_AutoConfig_t *auto_config; 3028 RF_ComponentLabel_t *clabel; 3029 int c; 3030 int num_cols; 3031 int num_missing; 3032 int mod_counter; 3033 int mod_counter_found; 3034 int even_pair_failed; 3035 char parity_type; 3036 3037 3038 /* check to see that we have enough 'live' components 3039 of this set. If so, we can configure it if necessary */ 3040 3041 num_cols = cset->ac->clabel->num_columns; 3042 parity_type = cset->ac->clabel->parityConfig; 3043 3044 /* XXX Check for duplicate components!?!?!? */ 3045 3046 /* Determine what the mod_counter is supposed to be for this set. */ 3047 3048 mod_counter_found = 0; 3049 mod_counter = 0; 3050 ac = cset->ac; 3051 while(ac!=NULL) { 3052 if (mod_counter_found==0) { 3053 mod_counter = ac->clabel->mod_counter; 3054 mod_counter_found = 1; 3055 } else { 3056 if (ac->clabel->mod_counter > mod_counter) { 3057 mod_counter = ac->clabel->mod_counter; 3058 } 3059 } 3060 ac = ac->next; 3061 } 3062 3063 num_missing = 0; 3064 auto_config = cset->ac; 3065 3066 even_pair_failed = 0; 3067 for(c=0; c<num_cols; c++) { 3068 ac = auto_config; 3069 while(ac!=NULL) { 3070 if ((ac->clabel->column == c) && 3071 (ac->clabel->mod_counter == mod_counter)) { 3072 /* it's this one... */ 3073 #ifdef DEBUG 3074 printf("Found: %s at %d\n", 3075 ac->devname,c); 3076 #endif 3077 break; 3078 } 3079 ac=ac->next; 3080 } 3081 if (ac==NULL) { 3082 /* Didn't find one here! */ 3083 /* special case for RAID 1, especially 3084 where there are more than 2 3085 components (where RAIDframe treats 3086 things a little differently :( ) */ 3087 if (parity_type == '1') { 3088 if (c%2 == 0) { /* even component */ 3089 even_pair_failed = 1; 3090 } else { /* odd component. If 3091 we're failed, and 3092 so is the even 3093 component, it's 3094 "Good Night, Charlie" */ 3095 if (even_pair_failed == 1) { 3096 return(0); 3097 } 3098 } 3099 } else { 3100 /* normal accounting */ 3101 num_missing++; 3102 } 3103 } 3104 if ((parity_type == '1') && (c%2 == 1)) { 3105 /* Just did an even component, and we didn't 3106 bail.. reset the even_pair_failed flag, 3107 and go on to the next component.... */ 3108 even_pair_failed = 0; 3109 } 3110 } 3111 3112 clabel = cset->ac->clabel; 3113 3114 if (((clabel->parityConfig == '0') && (num_missing > 0)) || 3115 ((clabel->parityConfig == '4') && (num_missing > 1)) || 3116 ((clabel->parityConfig == '5') && (num_missing > 1))) { 3117 /* XXX this needs to be made *much* more general */ 3118 /* Too many failures */ 3119 return(0); 3120 } 3121 /* otherwise, all is well, and we've got enough to take a kick 3122 at autoconfiguring this set */ 3123 return(1); 3124 } 3125 3126 void 3127 rf_create_configuration(RF_AutoConfig_t *ac, RF_Config_t *config, 3128 RF_Raid_t *raidPtr) 3129 { 3130 RF_ComponentLabel_t *clabel; 3131 int i; 3132 3133 clabel = ac->clabel; 3134 3135 /* 1. Fill in the common stuff */ 3136 config->numRow = clabel->num_rows = 1; 3137 config->numCol = clabel->num_columns; 3138 config->numSpare = 0; /* XXX should this be set here? */ 3139 config->sectPerSU = clabel->sectPerSU; 3140 config->SUsPerPU = clabel->SUsPerPU; 3141 config->SUsPerRU = clabel->SUsPerRU; 3142 config->parityConfig = clabel->parityConfig; 3143 /* XXX... */ 3144 strcpy(config->diskQueueType,"fifo"); 3145 config->maxOutstandingDiskReqs = clabel->maxOutstanding; 3146 config->layoutSpecificSize = 0; /* XXX ?? */ 3147 3148 while(ac!=NULL) { 3149 /* row/col values will be in range due to the checks 3150 in reasonable_label() */ 3151 strcpy(config->devnames[0][ac->clabel->column], 3152 ac->devname); 3153 ac = ac->next; 3154 } 3155 3156 for(i=0;i<RF_MAXDBGV;i++) { 3157 config->debugVars[i][0] = 0; 3158 } 3159 } 3160 3161 int 3162 rf_set_autoconfig(RF_Raid_t *raidPtr, int new_value) 3163 { 3164 RF_ComponentLabel_t clabel; 3165 struct vnode *vp; 3166 dev_t dev; 3167 int column; 3168 int sparecol; 3169 3170 raidPtr->autoconfigure = new_value; 3171 3172 for(column=0; column<raidPtr->numCol; column++) { 3173 if (raidPtr->Disks[column].status == rf_ds_optimal) { 3174 dev = raidPtr->Disks[column].dev; 3175 vp = raidPtr->raid_cinfo[column].ci_vp; 3176 raidread_component_label(dev, vp, &clabel); 3177 clabel.autoconfigure = new_value; 3178 raidwrite_component_label(dev, vp, &clabel); 3179 } 3180 } 3181 for(column = 0; column < raidPtr->numSpare ; column++) { 3182 sparecol = raidPtr->numCol + column; 3183 if (raidPtr->Disks[sparecol].status == rf_ds_used_spare) { 3184 dev = raidPtr->Disks[sparecol].dev; 3185 vp = raidPtr->raid_cinfo[sparecol].ci_vp; 3186 raidread_component_label(dev, vp, &clabel); 3187 clabel.autoconfigure = new_value; 3188 raidwrite_component_label(dev, vp, &clabel); 3189 } 3190 } 3191 return(new_value); 3192 } 3193 3194 int 3195 rf_set_rootpartition(RF_Raid_t *raidPtr, int new_value) 3196 { 3197 RF_ComponentLabel_t clabel; 3198 struct vnode *vp; 3199 dev_t dev; 3200 int column; 3201 int sparecol; 3202 3203 raidPtr->root_partition = new_value; 3204 for(column=0; column<raidPtr->numCol; column++) { 3205 if (raidPtr->Disks[column].status == rf_ds_optimal) { 3206 dev = raidPtr->Disks[column].dev; 3207 vp = raidPtr->raid_cinfo[column].ci_vp; 3208 raidread_component_label(dev, vp, &clabel); 3209 clabel.root_partition = new_value; 3210 raidwrite_component_label(dev, vp, &clabel); 3211 } 3212 } 3213 for(column = 0; column < raidPtr->numSpare ; column++) { 3214 sparecol = raidPtr->numCol + column; 3215 if (raidPtr->Disks[sparecol].status == rf_ds_used_spare) { 3216 dev = raidPtr->Disks[sparecol].dev; 3217 vp = raidPtr->raid_cinfo[sparecol].ci_vp; 3218 raidread_component_label(dev, vp, &clabel); 3219 clabel.root_partition = new_value; 3220 raidwrite_component_label(dev, vp, &clabel); 3221 } 3222 } 3223 return(new_value); 3224 } 3225 3226 void 3227 rf_release_all_vps(RF_ConfigSet_t *cset) 3228 { 3229 RF_AutoConfig_t *ac; 3230 3231 ac = cset->ac; 3232 while(ac!=NULL) { 3233 /* Close the vp, and give it back */ 3234 if (ac->vp) { 3235 vn_lock(ac->vp, LK_EXCLUSIVE | LK_RETRY); 3236 VOP_CLOSE(ac->vp, FREAD, NOCRED, 0); 3237 vput(ac->vp); 3238 ac->vp = NULL; 3239 } 3240 ac = ac->next; 3241 } 3242 } 3243 3244 3245 void 3246 rf_cleanup_config_set(RF_ConfigSet_t *cset) 3247 { 3248 RF_AutoConfig_t *ac; 3249 RF_AutoConfig_t *next_ac; 3250 3251 ac = cset->ac; 3252 while(ac!=NULL) { 3253 next_ac = ac->next; 3254 /* nuke the label */ 3255 free(ac->clabel, M_RAIDFRAME); 3256 /* cleanup the config structure */ 3257 free(ac, M_RAIDFRAME); 3258 /* "next.." */ 3259 ac = next_ac; 3260 } 3261 /* and, finally, nuke the config set */ 3262 free(cset, M_RAIDFRAME); 3263 } 3264 3265 3266 void 3267 raid_init_component_label(RF_Raid_t *raidPtr, RF_ComponentLabel_t *clabel) 3268 { 3269 /* current version number */ 3270 clabel->version = RF_COMPONENT_LABEL_VERSION; 3271 clabel->serial_number = raidPtr->serial_number; 3272 clabel->mod_counter = raidPtr->mod_counter; 3273 clabel->num_rows = 1; 3274 clabel->num_columns = raidPtr->numCol; 3275 clabel->clean = RF_RAID_DIRTY; /* not clean */ 3276 clabel->status = rf_ds_optimal; /* "It's good!" */ 3277 3278 clabel->sectPerSU = raidPtr->Layout.sectorsPerStripeUnit; 3279 clabel->SUsPerPU = raidPtr->Layout.SUsPerPU; 3280 clabel->SUsPerRU = raidPtr->Layout.SUsPerRU; 3281 3282 clabel->blockSize = raidPtr->bytesPerSector; 3283 clabel->numBlocks = raidPtr->sectorsPerDisk; 3284 3285 /* XXX not portable */ 3286 clabel->parityConfig = raidPtr->Layout.map->parityConfig; 3287 clabel->maxOutstanding = raidPtr->maxOutstanding; 3288 clabel->autoconfigure = raidPtr->autoconfigure; 3289 clabel->root_partition = raidPtr->root_partition; 3290 clabel->last_unit = raidPtr->raidid; 3291 clabel->config_order = raidPtr->config_order; 3292 } 3293 3294 int 3295 rf_auto_config_set(RF_ConfigSet_t *cset, int *unit) 3296 { 3297 RF_Raid_t *raidPtr; 3298 RF_Config_t *config; 3299 int raidID; 3300 int retcode; 3301 3302 #ifdef DEBUG 3303 printf("RAID autoconfigure\n"); 3304 #endif 3305 3306 retcode = 0; 3307 *unit = -1; 3308 3309 /* 1. Create a config structure */ 3310 3311 config = (RF_Config_t *)malloc(sizeof(RF_Config_t), 3312 M_RAIDFRAME, 3313 M_NOWAIT); 3314 if (config==NULL) { 3315 printf("Out of mem!?!?\n"); 3316 /* XXX do something more intelligent here. */ 3317 return(1); 3318 } 3319 3320 memset(config, 0, sizeof(RF_Config_t)); 3321 3322 /* 3323 2. Figure out what RAID ID this one is supposed to live at 3324 See if we can get the same RAID dev that it was configured 3325 on last time.. 3326 */ 3327 3328 raidID = cset->ac->clabel->last_unit; 3329 if ((raidID < 0) || (raidID >= numraid)) { 3330 /* let's not wander off into lala land. */ 3331 raidID = numraid - 1; 3332 } 3333 if (raidPtrs[raidID]->valid != 0) { 3334 3335 /* 3336 Nope... Go looking for an alternative... 3337 Start high so we don't immediately use raid0 if that's 3338 not taken. 3339 */ 3340 3341 for(raidID = numraid - 1; raidID >= 0; raidID--) { 3342 if (raidPtrs[raidID]->valid == 0) { 3343 /* can use this one! */ 3344 break; 3345 } 3346 } 3347 } 3348 3349 if (raidID < 0) { 3350 /* punt... */ 3351 printf("Unable to auto configure this set!\n"); 3352 printf("(Out of RAID devs!)\n"); 3353 free(config, M_RAIDFRAME); 3354 return(1); 3355 } 3356 3357 #ifdef DEBUG 3358 printf("Configuring raid%d:\n",raidID); 3359 #endif 3360 3361 raidPtr = raidPtrs[raidID]; 3362 3363 /* XXX all this stuff should be done SOMEWHERE ELSE! */ 3364 raidPtr->raidid = raidID; 3365 raidPtr->openings = RAIDOUTSTANDING; 3366 3367 /* 3. Build the configuration structure */ 3368 rf_create_configuration(cset->ac, config, raidPtr); 3369 3370 /* 4. Do the configuration */ 3371 retcode = rf_Configure(raidPtr, config, cset->ac); 3372 3373 if (retcode == 0) { 3374 3375 raidinit(raidPtrs[raidID]); 3376 3377 rf_markalldirty(raidPtrs[raidID]); 3378 raidPtrs[raidID]->autoconfigure = 1; /* XXX do this here? */ 3379 if (cset->ac->clabel->root_partition==1) { 3380 /* everything configured just fine. Make a note 3381 that this set is eligible to be root. */ 3382 cset->rootable = 1; 3383 /* XXX do this here? */ 3384 raidPtrs[raidID]->root_partition = 1; 3385 } 3386 } 3387 3388 /* 5. Cleanup */ 3389 free(config, M_RAIDFRAME); 3390 3391 *unit = raidID; 3392 return(retcode); 3393 } 3394 3395 void 3396 rf_disk_unbusy(RF_RaidAccessDesc_t *desc) 3397 { 3398 struct buf *bp; 3399 3400 bp = (struct buf *)desc->bp; 3401 disk_unbusy(&raid_softc[desc->raidPtr->raidid].sc_dkdev, 3402 (bp->b_bcount - bp->b_resid), (bp->b_flags & B_READ)); 3403 } 3404 3405 void 3406 rf_pool_init(struct pool *p, size_t size, const char *w_chan, 3407 size_t xmin, size_t xmax) 3408 { 3409 pool_init(p, size, 0, 0, 0, w_chan, NULL, IPL_BIO); 3410 pool_sethiwat(p, xmax); 3411 pool_prime(p, xmin); 3412 pool_setlowat(p, xmin); 3413 } 3414 3415 /* 3416 * rf_buf_queue_check(int raidid) -- looks into the buf_queue to see 3417 * if there is IO pending and if that IO could possibly be done for a 3418 * given RAID set. Returns 0 if IO is waiting and can be done, 1 3419 * otherwise. 3420 * 3421 */ 3422 3423 int 3424 rf_buf_queue_check(int raidid) 3425 { 3426 if ((BUFQ_PEEK(raid_softc[raidid].buf_queue) != NULL) && 3427 raidPtrs[raidid]->openings > 0) { 3428 /* there is work to do */ 3429 return 0; 3430 } 3431 /* default is nothing to do */ 3432 return 1; 3433 } 3434 3435 int 3436 rf_getdisksize(struct vnode *vp, struct lwp *l, RF_RaidDisk_t *diskPtr) 3437 { 3438 struct partinfo dpart; 3439 struct dkwedge_info dkw; 3440 int error; 3441 3442 error = VOP_IOCTL(vp, DIOCGPART, &dpart, FREAD, l->l_cred, l); 3443 if (error == 0) { 3444 diskPtr->blockSize = dpart.disklab->d_secsize; 3445 diskPtr->numBlocks = dpart.part->p_size - rf_protectedSectors; 3446 diskPtr->partitionSize = dpart.part->p_size; 3447 return 0; 3448 } 3449 3450 error = VOP_IOCTL(vp, DIOCGWEDGEINFO, &dkw, FREAD, l->l_cred, l); 3451 if (error == 0) { 3452 diskPtr->blockSize = 512; /* XXX */ 3453 diskPtr->numBlocks = dkw.dkw_size - rf_protectedSectors; 3454 diskPtr->partitionSize = dkw.dkw_size; 3455 return 0; 3456 } 3457 return error; 3458 } 3459 3460 static int 3461 raid_match(struct device *self, struct cfdata *cfdata, 3462 void *aux) 3463 { 3464 return 1; 3465 } 3466 3467 static void 3468 raid_attach(struct device *parent, struct device *self, 3469 void *aux) 3470 { 3471 3472 } 3473 3474 3475 static int 3476 raid_detach(struct device *self, int flags) 3477 { 3478 struct raid_softc *rs = (struct raid_softc *)self; 3479 3480 if (rs->sc_flags & RAIDF_INITED) 3481 return EBUSY; 3482 3483 return 0; 3484 } 3485 3486 3487
Indexes created Tue Sep 30 11:09:46 GMT 2025