1/* 2 * CDDL HEADER START 3 * 4 * The contents of this file are subject to the terms of the 5 * Common Development and Distribution License (the "License"). 6 * You may not use this file except in compliance with the License. 7 * 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 9 * or http://www.opensolaris.org/os/licensing. 10 * See the License for the specific language governing permissions 11 * and limitations under the License. 12 * 13 * When distributing Covered Code, include this CDDL HEADER in each 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE. 15 * If applicable, add the following below this CDDL HEADER, with the 16 * fields enclosed by brackets "[]" replaced with your own identifying 17 * information: Portions Copyright [yyyy] [name of copyright owner] 18 * 19 * CDDL HEADER END 20 */ 21/* 22 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. 23 * Copyright (c) 2012 by Delphix. All rights reserved. 24 * Copyright 2011 Nexenta Systems, Inc. All rights reserved. 25 */ 26 27#ifndef _SYS_SPA_H 28#define _SYS_SPA_H 29 30#include <sys/avl.h> 31#include <sys/zfs_context.h> 32#include <sys/nvpair.h> 33#include <sys/sysmacros.h> 34#include <sys/types.h> 35#include <sys/fs/zfs.h> 36 37#ifdef __cplusplus 38extern "C" { 39#endif 40 41/* 42 * Forward references that lots of things need. 43 */ 44typedef struct spa spa_t; 45typedef struct vdev vdev_t; 46typedef struct metaslab metaslab_t; 47typedef struct metaslab_group metaslab_group_t; 48typedef struct metaslab_class metaslab_class_t; 49typedef struct zio zio_t; 50typedef struct zilog zilog_t; 51typedef struct spa_aux_vdev spa_aux_vdev_t; 52typedef struct ddt ddt_t; 53typedef struct ddt_entry ddt_entry_t; 54struct dsl_pool; 55struct dsl_dataset; 56 57/* 58 * General-purpose 32-bit and 64-bit bitfield encodings. 59 */ 60#define BF32_DECODE(x, low, len) P2PHASE((x) >> (low), 1U << (len)) 61#define BF64_DECODE(x, low, len) P2PHASE((x) >> (low), 1ULL << (len)) 62#define BF32_ENCODE(x, low, len) (P2PHASE((x), 1U << (len)) << (low)) 63#define BF64_ENCODE(x, low, len) (P2PHASE((x), 1ULL << (len)) << (low)) 64 65#define BF32_GET(x, low, len) BF32_DECODE(x, low, len) 66#define BF64_GET(x, low, len) BF64_DECODE(x, low, len) 67 68#define BF32_SET(x, low, len, val) \ 69 ((x) ^= BF32_ENCODE((x >> low) ^ (val), low, len)) 70#define BF64_SET(x, low, len, val) \ 71 ((x) ^= BF64_ENCODE((x >> low) ^ (val), low, len)) 72 73#define BF32_GET_SB(x, low, len, shift, bias) \ 74 ((BF32_GET(x, low, len) + (bias)) << (shift)) 75#define BF64_GET_SB(x, low, len, shift, bias) \ 76 ((BF64_GET(x, low, len) + (bias)) << (shift)) 77 78#define BF32_SET_SB(x, low, len, shift, bias, val) \ 79 BF32_SET(x, low, len, ((val) >> (shift)) - (bias)) 80#define BF64_SET_SB(x, low, len, shift, bias, val) \ 81 BF64_SET(x, low, len, ((val) >> (shift)) - (bias)) 82 83/* 84 * We currently support nine block sizes, from 512 bytes to 128K. 85 * We could go higher, but the benefits are near-zero and the cost 86 * of COWing a giant block to modify one byte would become excessive. 87 */ 88#define SPA_MINBLOCKSHIFT 9 89#define SPA_MAXBLOCKSHIFT 17 90#define SPA_MINBLOCKSIZE (1ULL << SPA_MINBLOCKSHIFT) 91#define SPA_MAXBLOCKSIZE (1ULL << SPA_MAXBLOCKSHIFT) 92 93#define SPA_BLOCKSIZES (SPA_MAXBLOCKSHIFT - SPA_MINBLOCKSHIFT + 1) 94 95/* 96 * Maximum supported logical ashift. 97 * 98 * The current 8k allocation block size limit is due to the 8k 99 * aligned/sized operations performed by vdev_probe() on 100 * vdev_label->vl_pad2. Using another "safe region" for these tests 101 * would allow the limit to be raised to 16k, at the expense of 102 * only having 8 available uberblocks in the label area. 103 */ 104#define SPA_MAXASHIFT 13 105 106/* 107 * Size of block to hold the configuration data (a packed nvlist) 108 */ 109#define SPA_CONFIG_BLOCKSIZE (1ULL << 14) 110 111/* 112 * The DVA size encodings for LSIZE and PSIZE support blocks up to 32MB. 113 * The ASIZE encoding should be at least 64 times larger (6 more bits) 114 * to support up to 4-way RAID-Z mirror mode with worst-case gang block 115 * overhead, three DVAs per bp, plus one more bit in case we do anything 116 * else that expands the ASIZE. 117 */ 118#define SPA_LSIZEBITS 16 /* LSIZE up to 32M (2^16 * 512) */ 119#define SPA_PSIZEBITS 16 /* PSIZE up to 32M (2^16 * 512) */ 120#define SPA_ASIZEBITS 24 /* ASIZE up to 64 times larger */ 121 122/* 123 * All SPA data is represented by 128-bit data virtual addresses (DVAs). 124 * The members of the dva_t should be considered opaque outside the SPA. 125 */ 126typedef struct dva { 127 uint64_t dva_word[2]; 128} dva_t; 129 130/* 131 * Each block has a 256-bit checksum -- strong enough for cryptographic hashes. 132 */ 133typedef struct zio_cksum { 134 uint64_t zc_word[4]; 135} zio_cksum_t; 136 137/* 138 * Each block is described by its DVAs, time of birth, checksum, etc. 139 * The word-by-word, bit-by-bit layout of the blkptr is as follows: 140 * 141 * 64 56 48 40 32 24 16 8 0 142 * +-------+-------+-------+-------+-------+-------+-------+-------+ 143 * 0 | vdev1 | GRID | ASIZE | 144 * +-------+-------+-------+-------+-------+-------+-------+-------+ 145 * 1 |G| offset1 | 146 * +-------+-------+-------+-------+-------+-------+-------+-------+ 147 * 2 | vdev2 | GRID | ASIZE | 148 * +-------+-------+-------+-------+-------+-------+-------+-------+ 149 * 3 |G| offset2 | 150 * +-------+-------+-------+-------+-------+-------+-------+-------+ 151 * 4 | vdev3 | GRID | ASIZE | 152 * +-------+-------+-------+-------+-------+-------+-------+-------+ 153 * 5 |G| offset3 | 154 * +-------+-------+-------+-------+-------+-------+-------+-------+ 155 * 6 |BDX|lvl| type | cksum | comp | PSIZE | LSIZE | 156 * +-------+-------+-------+-------+-------+-------+-------+-------+ 157 * 7 | padding | 158 * +-------+-------+-------+-------+-------+-------+-------+-------+ 159 * 8 | padding | 160 * +-------+-------+-------+-------+-------+-------+-------+-------+ 161 * 9 | physical birth txg | 162 * +-------+-------+-------+-------+-------+-------+-------+-------+ 163 * a | logical birth txg | 164 * +-------+-------+-------+-------+-------+-------+-------+-------+ 165 * b | fill count | 166 * +-------+-------+-------+-------+-------+-------+-------+-------+ 167 * c | checksum[0] | 168 * +-------+-------+-------+-------+-------+-------+-------+-------+ 169 * d | checksum[1] | 170 * +-------+-------+-------+-------+-------+-------+-------+-------+ 171 * e | checksum[2] | 172 * +-------+-------+-------+-------+-------+-------+-------+-------+ 173 * f | checksum[3] | 174 * +-------+-------+-------+-------+-------+-------+-------+-------+ 175 * 176 * Legend: 177 * 178 * vdev virtual device ID 179 * offset offset into virtual device 180 * LSIZE logical size 181 * PSIZE physical size (after compression) 182 * ASIZE allocated size (including RAID-Z parity and gang block headers) 183 * GRID RAID-Z layout information (reserved for future use) 184 * cksum checksum function 185 * comp compression function 186 * G gang block indicator 187 * B byteorder (endianness) 188 * D dedup 189 * X unused 190 * lvl level of indirection 191 * type DMU object type 192 * phys birth txg of block allocation; zero if same as logical birth txg 193 * log. birth transaction group in which the block was logically born 194 * fill count number of non-zero blocks under this bp 195 * checksum[4] 256-bit checksum of the data this bp describes 196 */ 197#define SPA_BLKPTRSHIFT 7 /* blkptr_t is 128 bytes */ 198#define SPA_DVAS_PER_BP 3 /* Number of DVAs in a bp */ 199 200typedef struct blkptr { 201 dva_t blk_dva[SPA_DVAS_PER_BP]; /* Data Virtual Addresses */ 202 uint64_t blk_prop; /* size, compression, type, etc */ 203 uint64_t blk_pad[2]; /* Extra space for the future */ 204 uint64_t blk_phys_birth; /* txg when block was allocated */ 205 uint64_t blk_birth; /* transaction group at birth */ 206 uint64_t blk_fill; /* fill count */ 207 zio_cksum_t blk_cksum; /* 256-bit checksum */ 208} blkptr_t; 209 210/* 211 * Macros to get and set fields in a bp or DVA. 212 */ 213#define DVA_GET_ASIZE(dva) \ 214 BF64_GET_SB((dva)->dva_word[0], 0, 24, SPA_MINBLOCKSHIFT, 0) 215#define DVA_SET_ASIZE(dva, x) \ 216 BF64_SET_SB((dva)->dva_word[0], 0, 24, SPA_MINBLOCKSHIFT, 0, x) 217 218#define DVA_GET_GRID(dva) BF64_GET((dva)->dva_word[0], 24, 8) 219#define DVA_SET_GRID(dva, x) BF64_SET((dva)->dva_word[0], 24, 8, x) 220 221#define DVA_GET_VDEV(dva) BF64_GET((dva)->dva_word[0], 32, 32) 222#define DVA_SET_VDEV(dva, x) BF64_SET((dva)->dva_word[0], 32, 32, x) 223 224#define DVA_GET_OFFSET(dva) \ 225 BF64_GET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0) 226#define DVA_SET_OFFSET(dva, x) \ 227 BF64_SET_SB((dva)->dva_word[1], 0, 63, SPA_MINBLOCKSHIFT, 0, x) 228 229#define DVA_GET_GANG(dva) BF64_GET((dva)->dva_word[1], 63, 1) 230#define DVA_SET_GANG(dva, x) BF64_SET((dva)->dva_word[1], 63, 1, x) 231 232#define BP_GET_LSIZE(bp) \ 233 BF64_GET_SB((bp)->blk_prop, 0, 16, SPA_MINBLOCKSHIFT, 1) 234#define BP_SET_LSIZE(bp, x) \ 235 BF64_SET_SB((bp)->blk_prop, 0, 16, SPA_MINBLOCKSHIFT, 1, x) 236 237#define BP_GET_PSIZE(bp) \ 238 BF64_GET_SB((bp)->blk_prop, 16, 16, SPA_MINBLOCKSHIFT, 1) 239#define BP_SET_PSIZE(bp, x) \ 240 BF64_SET_SB((bp)->blk_prop, 16, 16, SPA_MINBLOCKSHIFT, 1, x) 241 242#define BP_GET_COMPRESS(bp) BF64_GET((bp)->blk_prop, 32, 8) 243#define BP_SET_COMPRESS(bp, x) BF64_SET((bp)->blk_prop, 32, 8, x) 244 245#define BP_GET_CHECKSUM(bp) BF64_GET((bp)->blk_prop, 40, 8) 246#define BP_SET_CHECKSUM(bp, x) BF64_SET((bp)->blk_prop, 40, 8, x) 247 248#define BP_GET_TYPE(bp) BF64_GET((bp)->blk_prop, 48, 8) 249#define BP_SET_TYPE(bp, x) BF64_SET((bp)->blk_prop, 48, 8, x) 250 251#define BP_GET_LEVEL(bp) BF64_GET((bp)->blk_prop, 56, 5) 252#define BP_SET_LEVEL(bp, x) BF64_SET((bp)->blk_prop, 56, 5, x) 253 254#define BP_GET_PROP_BIT_61(bp) BF64_GET((bp)->blk_prop, 61, 1) 255#define BP_SET_PROP_BIT_61(bp, x) BF64_SET((bp)->blk_prop, 61, 1, x) 256 257#define BP_GET_DEDUP(bp) BF64_GET((bp)->blk_prop, 62, 1) 258#define BP_SET_DEDUP(bp, x) BF64_SET((bp)->blk_prop, 62, 1, x) 259 260#define BP_GET_BYTEORDER(bp) (0 - BF64_GET((bp)->blk_prop, 63, 1)) 261#define BP_SET_BYTEORDER(bp, x) BF64_SET((bp)->blk_prop, 63, 1, x) 262 263#define BP_PHYSICAL_BIRTH(bp) \ 264 ((bp)->blk_phys_birth ? (bp)->blk_phys_birth : (bp)->blk_birth) 265 266#define BP_SET_BIRTH(bp, logical, physical) \ 267{ \ 268 (bp)->blk_birth = (logical); \ 269 (bp)->blk_phys_birth = ((logical) == (physical) ? 0 : (physical)); \ 270} 271 272#define BP_GET_ASIZE(bp) \ 273 (DVA_GET_ASIZE(&(bp)->blk_dva[0]) + DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \ 274 DVA_GET_ASIZE(&(bp)->blk_dva[2])) 275 276#define BP_GET_UCSIZE(bp) \ 277 ((BP_GET_LEVEL(bp) > 0 || DMU_OT_IS_METADATA(BP_GET_TYPE(bp))) ? \ 278 BP_GET_PSIZE(bp) : BP_GET_LSIZE(bp)) 279 280#define BP_GET_NDVAS(bp) \ 281 (!!DVA_GET_ASIZE(&(bp)->blk_dva[0]) + \ 282 !!DVA_GET_ASIZE(&(bp)->blk_dva[1]) + \ 283 !!DVA_GET_ASIZE(&(bp)->blk_dva[2])) 284 285#define BP_COUNT_GANG(bp) \ 286 (DVA_GET_GANG(&(bp)->blk_dva[0]) + \ 287 DVA_GET_GANG(&(bp)->blk_dva[1]) + \ 288 DVA_GET_GANG(&(bp)->blk_dva[2])) 289 290#define DVA_EQUAL(dva1, dva2) \ 291 ((dva1)->dva_word[1] == (dva2)->dva_word[1] && \ 292 (dva1)->dva_word[0] == (dva2)->dva_word[0]) 293 294#define BP_EQUAL(bp1, bp2) \ 295 (BP_PHYSICAL_BIRTH(bp1) == BP_PHYSICAL_BIRTH(bp2) && \ 296 DVA_EQUAL(&(bp1)->blk_dva[0], &(bp2)->blk_dva[0]) && \ 297 DVA_EQUAL(&(bp1)->blk_dva[1], &(bp2)->blk_dva[1]) && \ 298 DVA_EQUAL(&(bp1)->blk_dva[2], &(bp2)->blk_dva[2])) 299 300#define ZIO_CHECKSUM_EQUAL(zc1, zc2) \ 301 (0 == (((zc1).zc_word[0] - (zc2).zc_word[0]) | \ 302 ((zc1).zc_word[1] - (zc2).zc_word[1]) | \ 303 ((zc1).zc_word[2] - (zc2).zc_word[2]) | \ 304 ((zc1).zc_word[3] - (zc2).zc_word[3]))) 305 306#define DVA_IS_VALID(dva) (DVA_GET_ASIZE(dva) != 0) 307 308#define ZIO_SET_CHECKSUM(zcp, w0, w1, w2, w3) \ 309{ \ 310 (zcp)->zc_word[0] = w0; \ 311 (zcp)->zc_word[1] = w1; \ 312 (zcp)->zc_word[2] = w2; \ 313 (zcp)->zc_word[3] = w3; \ 314} 315 316#define BP_IDENTITY(bp) (&(bp)->blk_dva[0]) 317#define BP_IS_GANG(bp) DVA_GET_GANG(BP_IDENTITY(bp)) 318#define BP_IS_HOLE(bp) ((bp)->blk_birth == 0) 319 320/* BP_IS_RAIDZ(bp) assumes no block compression */ 321#define BP_IS_RAIDZ(bp) (DVA_GET_ASIZE(&(bp)->blk_dva[0]) > \ 322 BP_GET_PSIZE(bp)) 323 324#define BP_ZERO(bp) \ 325{ \ 326 (bp)->blk_dva[0].dva_word[0] = 0; \ 327 (bp)->blk_dva[0].dva_word[1] = 0; \ 328 (bp)->blk_dva[1].dva_word[0] = 0; \ 329 (bp)->blk_dva[1].dva_word[1] = 0; \ 330 (bp)->blk_dva[2].dva_word[0] = 0; \ 331 (bp)->blk_dva[2].dva_word[1] = 0; \ 332 (bp)->blk_prop = 0; \ 333 (bp)->blk_pad[0] = 0; \ 334 (bp)->blk_pad[1] = 0; \ 335 (bp)->blk_phys_birth = 0; \ 336 (bp)->blk_birth = 0; \ 337 (bp)->blk_fill = 0; \ 338 ZIO_SET_CHECKSUM(&(bp)->blk_cksum, 0, 0, 0, 0); \ 339} 340 341/* 342 * Note: the byteorder is either 0 or -1, both of which are palindromes. 343 * This simplifies the endianness handling a bit. 344 */ 345#if BYTE_ORDER == _BIG_ENDIAN 346#define ZFS_HOST_BYTEORDER (0ULL) 347#else 348#define ZFS_HOST_BYTEORDER (-1ULL) 349#endif 350 351#define BP_SHOULD_BYTESWAP(bp) (BP_GET_BYTEORDER(bp) != ZFS_HOST_BYTEORDER) 352 353#define BP_SPRINTF_LEN 320 354 355/* 356 * This macro allows code sharing between zfs, libzpool, and mdb. 357 * 'func' is either snprintf() or mdb_snprintf(). 358 * 'ws' (whitespace) can be ' ' for single-line format, '\n' for multi-line. 359 */ 360#define SPRINTF_BLKPTR(func, ws, buf, bp, type, checksum, compress) \ 361{ \ 362 static const char *copyname[] = \ 363 { "zero", "single", "double", "triple" }; \ 364 int size = BP_SPRINTF_LEN; \ 365 int len = 0; \ 366 int copies = 0; \ 367 \ 368 if (bp == NULL) { \ 369 len = func(buf + len, size - len, "<NULL>"); \ 370 } else if (BP_IS_HOLE(bp)) { \ 371 len = func(buf + len, size - len, "<hole>"); \ 372 } else { \ 373 for (int d = 0; d < BP_GET_NDVAS(bp); d++) { \ 374 const dva_t *dva = &bp->blk_dva[d]; \ 375 if (DVA_IS_VALID(dva)) \ 376 copies++; \ 377 len += func(buf + len, size - len, \ 378 "DVA[%d]=<%llu:%llx:%llx>%c", d, \ 379 (u_longlong_t)DVA_GET_VDEV(dva), \ 380 (u_longlong_t)DVA_GET_OFFSET(dva), \ 381 (u_longlong_t)DVA_GET_ASIZE(dva), \ 382 ws); \ 383 } \ 384 if (BP_IS_GANG(bp) && \ 385 DVA_GET_ASIZE(&bp->blk_dva[2]) <= \ 386 DVA_GET_ASIZE(&bp->blk_dva[1]) / 2) \ 387 copies--; \ 388 len += func(buf + len, size - len, \ 389 "[L%llu %s] %s %s %s %s %s %s%c" \ 390 "size=%llxL/%llxP birth=%lluL/%lluP fill=%llu%c" \ 391 "cksum=%llx:%llx:%llx:%llx", \ 392 (u_longlong_t)BP_GET_LEVEL(bp), \ 393 type, \ 394 checksum, \ 395 compress, \ 396 BP_GET_BYTEORDER(bp) == 0 ? "BE" : "LE", \ 397 BP_IS_GANG(bp) ? "gang" : "contiguous", \ 398 BP_GET_DEDUP(bp) ? "dedup" : "unique", \ 399 copyname[copies], \ 400 ws, \ 401 (u_longlong_t)BP_GET_LSIZE(bp), \ 402 (u_longlong_t)BP_GET_PSIZE(bp), \ 403 (u_longlong_t)bp->blk_birth, \ 404 (u_longlong_t)BP_PHYSICAL_BIRTH(bp), \ 405 (u_longlong_t)bp->blk_fill, \ 406 ws, \ 407 (u_longlong_t)bp->blk_cksum.zc_word[0], \ 408 (u_longlong_t)bp->blk_cksum.zc_word[1], \ 409 (u_longlong_t)bp->blk_cksum.zc_word[2], \ 410 (u_longlong_t)bp->blk_cksum.zc_word[3]); \ 411 } \ 412 ASSERT(len < size); \ 413} 414 415#include <sys/dmu.h> 416 417#define BP_GET_BUFC_TYPE(bp) \ 418 (((BP_GET_LEVEL(bp) > 0) || (DMU_OT_IS_METADATA(BP_GET_TYPE(bp)))) ? \ 419 ARC_BUFC_METADATA : ARC_BUFC_DATA) 420 421typedef enum spa_import_type { 422 SPA_IMPORT_EXISTING, 423 SPA_IMPORT_ASSEMBLE 424} spa_import_type_t; 425 426/* state manipulation functions */ 427extern int spa_open(const char *pool, spa_t **, void *tag); 428extern int spa_open_rewind(const char *pool, spa_t **, void *tag, 429 nvlist_t *policy, nvlist_t **config); 430extern int spa_get_stats(const char *pool, nvlist_t **config, char *altroot, 431 size_t buflen); 432extern int spa_create(const char *pool, nvlist_t *config, nvlist_t *props, 433 nvlist_t *zplprops); 434#if defined(sun) 435extern int spa_import_rootpool(char *devpath, char *devid); 436#else 437extern int spa_import_rootpool(const char *name); 438#endif 439extern int spa_import(const char *pool, nvlist_t *config, nvlist_t *props, 440 uint64_t flags); 441extern nvlist_t *spa_tryimport(nvlist_t *tryconfig); 442extern int spa_destroy(char *pool); 443extern int spa_export(char *pool, nvlist_t **oldconfig, boolean_t force, 444 boolean_t hardforce); 445extern int spa_reset(char *pool); 446extern void spa_async_request(spa_t *spa, int flag); 447extern void spa_async_unrequest(spa_t *spa, int flag); 448extern void spa_async_suspend(spa_t *spa); 449extern void spa_async_resume(spa_t *spa); 450extern spa_t *spa_inject_addref(char *pool); 451extern void spa_inject_delref(spa_t *spa); 452extern void spa_scan_stat_init(spa_t *spa); 453extern int spa_scan_get_stats(spa_t *spa, pool_scan_stat_t *ps); 454 455#define SPA_ASYNC_CONFIG_UPDATE 0x01 456#define SPA_ASYNC_REMOVE 0x02 457#define SPA_ASYNC_PROBE 0x04 458#define SPA_ASYNC_RESILVER_DONE 0x08 459#define SPA_ASYNC_RESILVER 0x10 460#define SPA_ASYNC_AUTOEXPAND 0x20 461#define SPA_ASYNC_REMOVE_DONE 0x40 462#define SPA_ASYNC_REMOVE_STOP 0x80 463 464/* 465 * Controls the behavior of spa_vdev_remove(). 466 */ 467#define SPA_REMOVE_UNSPARE 0x01 468#define SPA_REMOVE_DONE 0x02 469 470/* device manipulation */ 471extern int spa_vdev_add(spa_t *spa, nvlist_t *nvroot); 472extern int spa_vdev_attach(spa_t *spa, uint64_t guid, nvlist_t *nvroot, 473 int replacing); 474extern int spa_vdev_detach(spa_t *spa, uint64_t guid, uint64_t pguid, 475 int replace_done); 476extern int spa_vdev_remove(spa_t *spa, uint64_t guid, boolean_t unspare); 477extern boolean_t spa_vdev_remove_active(spa_t *spa); 478extern int spa_vdev_setpath(spa_t *spa, uint64_t guid, const char *newpath); 479extern int spa_vdev_setfru(spa_t *spa, uint64_t guid, const char *newfru); 480extern int spa_vdev_split_mirror(spa_t *spa, char *newname, nvlist_t *config, 481 nvlist_t *props, boolean_t exp); 482 483/* spare state (which is global across all pools) */ 484extern void spa_spare_add(vdev_t *vd); 485extern void spa_spare_remove(vdev_t *vd); 486extern boolean_t spa_spare_exists(uint64_t guid, uint64_t *pool, int *refcnt); 487extern void spa_spare_activate(vdev_t *vd); 488 489/* L2ARC state (which is global across all pools) */ 490extern void spa_l2cache_add(vdev_t *vd); 491extern void spa_l2cache_remove(vdev_t *vd); 492extern boolean_t spa_l2cache_exists(uint64_t guid, uint64_t *pool); 493extern void spa_l2cache_activate(vdev_t *vd); 494extern void spa_l2cache_drop(spa_t *spa); 495 496/* scanning */ 497extern int spa_scan(spa_t *spa, pool_scan_func_t func); 498extern int spa_scan_stop(spa_t *spa); 499 500/* spa syncing */ 501extern void spa_sync(spa_t *spa, uint64_t txg); /* only for DMU use */ 502extern void spa_sync_allpools(void); 503 504/* spa namespace global mutex */ 505extern kmutex_t spa_namespace_lock; 506 507/* 508 * SPA configuration functions in spa_config.c 509 */ 510 511#define SPA_CONFIG_UPDATE_POOL 0 512#define SPA_CONFIG_UPDATE_VDEVS 1 513 514extern void spa_config_sync(spa_t *, boolean_t, boolean_t); 515extern void spa_config_load(void); 516extern nvlist_t *spa_all_configs(uint64_t *); 517extern void spa_config_set(spa_t *spa, nvlist_t *config); 518extern nvlist_t *spa_config_generate(spa_t *spa, vdev_t *vd, uint64_t txg, 519 int getstats); 520extern void spa_config_update(spa_t *spa, int what); 521 522/* 523 * Miscellaneous SPA routines in spa_misc.c 524 */ 525 526/* Namespace manipulation */ 527extern spa_t *spa_lookup(const char *name); 528extern spa_t *spa_add(const char *name, nvlist_t *config, const char *altroot); 529extern void spa_remove(spa_t *spa); 530extern spa_t *spa_next(spa_t *prev); 531 532/* Refcount functions */ 533extern void spa_open_ref(spa_t *spa, void *tag); 534extern void spa_close(spa_t *spa, void *tag); 535extern boolean_t spa_refcount_zero(spa_t *spa); 536 537#define SCL_NONE 0x00 538#define SCL_CONFIG 0x01 539#define SCL_STATE 0x02 540#define SCL_L2ARC 0x04 /* hack until L2ARC 2.0 */ 541#define SCL_ALLOC 0x08 542#define SCL_ZIO 0x10 543#define SCL_FREE 0x20 544#define SCL_VDEV 0x40 545#define SCL_LOCKS 7 546#define SCL_ALL ((1 << SCL_LOCKS) - 1) 547#define SCL_STATE_ALL (SCL_STATE | SCL_L2ARC | SCL_ZIO) 548 549/* Pool configuration locks */ 550extern int spa_config_tryenter(spa_t *spa, int locks, void *tag, krw_t rw); 551extern void spa_config_enter(spa_t *spa, int locks, void *tag, krw_t rw); 552extern void spa_config_exit(spa_t *spa, int locks, void *tag); 553extern int spa_config_held(spa_t *spa, int locks, krw_t rw); 554 555/* Pool vdev add/remove lock */ 556extern uint64_t spa_vdev_enter(spa_t *spa); 557extern uint64_t spa_vdev_config_enter(spa_t *spa); 558extern void spa_vdev_config_exit(spa_t *spa, vdev_t *vd, uint64_t txg, 559 int error, char *tag); 560extern int spa_vdev_exit(spa_t *spa, vdev_t *vd, uint64_t txg, int error); 561 562/* Pool vdev state change lock */ 563extern void spa_vdev_state_enter(spa_t *spa, int oplock); 564extern int spa_vdev_state_exit(spa_t *spa, vdev_t *vd, int error); 565 566/* Log state */ 567typedef enum spa_log_state { 568 SPA_LOG_UNKNOWN = 0, /* unknown log state */ 569 SPA_LOG_MISSING, /* missing log(s) */ 570 SPA_LOG_CLEAR, /* clear the log(s) */ 571 SPA_LOG_GOOD, /* log(s) are good */ 572} spa_log_state_t; 573 574extern spa_log_state_t spa_get_log_state(spa_t *spa); 575extern void spa_set_log_state(spa_t *spa, spa_log_state_t state); 576extern int spa_offline_log(spa_t *spa); 577 578/* Log claim callback */ 579extern void spa_claim_notify(zio_t *zio); 580 581/* Accessor functions */ 582extern boolean_t spa_shutting_down(spa_t *spa); 583extern struct dsl_pool *spa_get_dsl(spa_t *spa); 584extern boolean_t spa_is_initializing(spa_t *spa); 585extern blkptr_t *spa_get_rootblkptr(spa_t *spa); 586extern void spa_set_rootblkptr(spa_t *spa, const blkptr_t *bp); 587extern void spa_altroot(spa_t *, char *, size_t); 588extern int spa_sync_pass(spa_t *spa); 589extern char *spa_name(spa_t *spa); 590extern uint64_t spa_guid(spa_t *spa); 591extern uint64_t spa_load_guid(spa_t *spa); 592extern uint64_t spa_last_synced_txg(spa_t *spa); 593extern uint64_t spa_first_txg(spa_t *spa); 594extern uint64_t spa_syncing_txg(spa_t *spa); 595extern uint64_t spa_version(spa_t *spa); 596extern pool_state_t spa_state(spa_t *spa); 597extern spa_load_state_t spa_load_state(spa_t *spa); 598extern uint64_t spa_freeze_txg(spa_t *spa); 599extern uint64_t spa_get_asize(spa_t *spa, uint64_t lsize); 600extern uint64_t spa_get_dspace(spa_t *spa); 601extern void spa_update_dspace(spa_t *spa); 602extern uint64_t spa_version(spa_t *spa); 603extern boolean_t spa_deflate(spa_t *spa); 604extern metaslab_class_t *spa_normal_class(spa_t *spa); 605extern metaslab_class_t *spa_log_class(spa_t *spa); 606extern int spa_max_replication(spa_t *spa); 607extern int spa_prev_software_version(spa_t *spa); 608extern int spa_busy(void); 609extern uint8_t spa_get_failmode(spa_t *spa); 610extern boolean_t spa_suspended(spa_t *spa); 611extern uint64_t spa_bootfs(spa_t *spa); 612extern uint64_t spa_delegation(spa_t *spa); 613extern objset_t *spa_meta_objset(spa_t *spa); 614extern uint64_t spa_deadman_synctime(spa_t *spa); 615 616/* Miscellaneous support routines */ 617extern void spa_activate_mos_feature(spa_t *spa, const char *feature); 618extern void spa_deactivate_mos_feature(spa_t *spa, const char *feature); 619extern int spa_rename(const char *oldname, const char *newname); 620extern spa_t *spa_by_guid(uint64_t pool_guid, uint64_t device_guid); 621extern boolean_t spa_guid_exists(uint64_t pool_guid, uint64_t device_guid); 622extern char *spa_strdup(const char *); 623extern void spa_strfree(char *); 624extern uint64_t spa_get_random(uint64_t range); 625extern uint64_t spa_generate_guid(spa_t *spa); 626extern void sprintf_blkptr(char *buf, const blkptr_t *bp); 627extern void spa_freeze(spa_t *spa); 628extern int spa_change_guid(spa_t *spa); 629extern void spa_upgrade(spa_t *spa, uint64_t version); 630extern void spa_evict_all(void); 631extern vdev_t *spa_lookup_by_guid(spa_t *spa, uint64_t guid, 632 boolean_t l2cache); 633extern boolean_t spa_has_spare(spa_t *, uint64_t guid); 634extern uint64_t dva_get_dsize_sync(spa_t *spa, const dva_t *dva); 635extern uint64_t bp_get_dsize_sync(spa_t *spa, const blkptr_t *bp); 636extern uint64_t bp_get_dsize(spa_t *spa, const blkptr_t *bp); 637extern boolean_t spa_has_slogs(spa_t *spa); 638extern boolean_t spa_is_root(spa_t *spa); 639extern boolean_t spa_writeable(spa_t *spa); 640 641extern int spa_mode(spa_t *spa); 642extern uint64_t zfs_strtonum(const char *str, char **nptr); 643#define strtonum(str, nptr) zfs_strtonum((str), (nptr)) 644 645extern char *spa_his_ievent_table[]; 646 647extern void spa_history_create_obj(spa_t *spa, dmu_tx_t *tx); 648extern int spa_history_get(spa_t *spa, uint64_t *offset, uint64_t *len_read, 649 char *his_buf); 650extern int spa_history_log(spa_t *spa, const char *his_buf); 651extern int spa_history_log_nvl(spa_t *spa, nvlist_t *nvl); 652extern void spa_history_log_version(spa_t *spa, const char *operation); 653extern void spa_history_log_internal(spa_t *spa, const char *operation, 654 dmu_tx_t *tx, const char *fmt, ...); 655extern void spa_history_log_internal_ds(struct dsl_dataset *ds, const char *op, 656 dmu_tx_t *tx, const char *fmt, ...); 657extern void spa_history_log_internal_dd(dsl_dir_t *dd, const char *operation, 658 dmu_tx_t *tx, const char *fmt, ...); 659 660/* error handling */ 661struct zbookmark; 662extern void spa_log_error(spa_t *spa, zio_t *zio); 663extern void zfs_ereport_post(const char *cls, spa_t *spa, vdev_t *vd, 664 zio_t *zio, uint64_t stateoroffset, uint64_t length); 665extern void zfs_post_remove(spa_t *spa, vdev_t *vd); 666extern void zfs_post_state_change(spa_t *spa, vdev_t *vd); 667extern void zfs_post_autoreplace(spa_t *spa, vdev_t *vd); 668extern uint64_t spa_get_errlog_size(spa_t *spa); 669extern int spa_get_errlog(spa_t *spa, void *uaddr, size_t *count); 670extern void spa_errlog_rotate(spa_t *spa); 671extern void spa_errlog_drain(spa_t *spa); 672extern void spa_errlog_sync(spa_t *spa, uint64_t txg); 673extern void spa_get_errlists(spa_t *spa, avl_tree_t *last, avl_tree_t *scrub); 674 675/* vdev cache */ 676extern void vdev_cache_stat_init(void); 677extern void vdev_cache_stat_fini(void); 678 679/* Initialization and termination */ 680extern void spa_init(int flags); 681extern void spa_fini(void); 682extern void spa_boot_init(); 683 684/* properties */ 685extern int spa_prop_set(spa_t *spa, nvlist_t *nvp); 686extern int spa_prop_get(spa_t *spa, nvlist_t **nvp); 687extern void spa_prop_clear_bootfs(spa_t *spa, uint64_t obj, dmu_tx_t *tx); 688extern void spa_configfile_set(spa_t *, nvlist_t *, boolean_t); 689 690/* asynchronous event notification */ 691extern void spa_event_notify(spa_t *spa, vdev_t *vdev, const char *name); 692 693#ifdef ZFS_DEBUG 694#define dprintf_bp(bp, fmt, ...) do { \ 695 if (zfs_flags & ZFS_DEBUG_DPRINTF) { \ 696 char *__blkbuf = kmem_alloc(BP_SPRINTF_LEN, KM_SLEEP); \ 697 sprintf_blkptr(__blkbuf, (bp)); \ 698 dprintf(fmt " %s\n", __VA_ARGS__, __blkbuf); \ 699 kmem_free(__blkbuf, BP_SPRINTF_LEN); \ 700 } \ 701_NOTE(CONSTCOND) } while (0) 702#else 703#define dprintf_bp(bp, fmt, ...) 704#endif 705 706extern boolean_t spa_debug_enabled(spa_t *spa); 707#define spa_dbgmsg(spa, ...) \ 708{ \ 709 if (spa_debug_enabled(spa)) \ 710 zfs_dbgmsg(__VA_ARGS__); \ 711} 712 713extern int spa_mode_global; /* mode, e.g. FREAD | FWRITE */ 714 715#ifdef __cplusplus 716} 717#endif 718 719#endif /* _SYS_SPA_H */ 720