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 2006 Sun Microsystems, Inc. All rights reserved. 23 * Use is subject to license terms. 24 */ 25 26 #ifndef _ZIO_H 27 #define _ZIO_H 28 29 #pragma ident "%Z%%M% %I% %E% SMI" 30 31 #include <sys/zfs_context.h> 32 #include <sys/spa.h> 33 #include <sys/txg.h> 34 #include <sys/avl.h> 35 #include <sys/dkio.h> 36 #include <sys/fs/zfs.h> 37 #include <sys/zio_impl.h> 38 39 #ifdef __cplusplus 40 extern "C" { 41 #endif 42 43 #define ZBT_MAGIC 0x210da7ab10c7a11ULL /* zio data bloc tail */ 44 45 typedef struct zio_block_tail { 46 uint64_t zbt_magic; /* for validation, endianness */ 47 zio_cksum_t zbt_cksum; /* 256-bit checksum */ 48 } zio_block_tail_t; 49 50 /* 51 * Gang block headers are self-checksumming and contain an array 52 * of block pointers. 53 */ 54 #define SPA_GANGBLOCKSIZE SPA_MINBLOCKSIZE 55 #define SPA_GBH_NBLKPTRS ((SPA_GANGBLOCKSIZE - \ 56 sizeof (zio_block_tail_t)) / sizeof (blkptr_t)) 57 #define SPA_GBH_FILLER ((SPA_GANGBLOCKSIZE - \ 58 sizeof (zio_block_tail_t) - \ 59 (SPA_GBH_NBLKPTRS * sizeof (blkptr_t))) /\ 60 sizeof (uint64_t)) 61 62 #define ZIO_GET_IOSIZE(zio) \ 63 (BP_IS_GANG((zio)->io_bp) ? \ 64 SPA_GANGBLOCKSIZE : BP_GET_PSIZE((zio)->io_bp)) 65 66 typedef struct zio_gbh { 67 blkptr_t zg_blkptr[SPA_GBH_NBLKPTRS]; 68 uint64_t zg_filler[SPA_GBH_FILLER]; 69 zio_block_tail_t zg_tail; 70 } zio_gbh_phys_t; 71 72 enum zio_checksum { 73 ZIO_CHECKSUM_INHERIT = 0, 74 ZIO_CHECKSUM_ON, 75 ZIO_CHECKSUM_OFF, 76 ZIO_CHECKSUM_LABEL, 77 ZIO_CHECKSUM_GANG_HEADER, 78 ZIO_CHECKSUM_ZILOG, 79 ZIO_CHECKSUM_FLETCHER_2, 80 ZIO_CHECKSUM_FLETCHER_4, 81 ZIO_CHECKSUM_SHA256, 82 ZIO_CHECKSUM_FUNCTIONS 83 }; 84 85 #define ZIO_CHECKSUM_ON_VALUE ZIO_CHECKSUM_FLETCHER_2 86 #define ZIO_CHECKSUM_DEFAULT ZIO_CHECKSUM_ON 87 88 enum zio_compress { 89 ZIO_COMPRESS_INHERIT = 0, 90 ZIO_COMPRESS_ON, 91 ZIO_COMPRESS_OFF, 92 ZIO_COMPRESS_LZJB, 93 ZIO_COMPRESS_FUNCTIONS 94 }; 95 96 #define ZIO_COMPRESS_ON_VALUE ZIO_COMPRESS_LZJB 97 #define ZIO_COMPRESS_DEFAULT ZIO_COMPRESS_OFF 98 99 #define ZIO_PRIORITY_NOW (zio_priority_table[0]) 100 #define ZIO_PRIORITY_SYNC_READ (zio_priority_table[1]) 101 #define ZIO_PRIORITY_SYNC_WRITE (zio_priority_table[2]) 102 #define ZIO_PRIORITY_ASYNC_READ (zio_priority_table[3]) 103 #define ZIO_PRIORITY_ASYNC_WRITE (zio_priority_table[4]) 104 #define ZIO_PRIORITY_FREE (zio_priority_table[5]) 105 #define ZIO_PRIORITY_CACHE_FILL (zio_priority_table[6]) 106 #define ZIO_PRIORITY_LOG_WRITE (zio_priority_table[7]) 107 #define ZIO_PRIORITY_RESILVER (zio_priority_table[8]) 108 #define ZIO_PRIORITY_SCRUB (zio_priority_table[9]) 109 #define ZIO_PRIORITY_TABLE_SIZE 10 110 111 #define ZIO_FLAG_MUSTSUCCEED 0x00000 112 #define ZIO_FLAG_CANFAIL 0x00001 113 #define ZIO_FLAG_FAILFAST 0x00002 114 #define ZIO_FLAG_CONFIG_HELD 0x00004 115 116 #define ZIO_FLAG_DONT_CACHE 0x00010 117 #define ZIO_FLAG_DONT_QUEUE 0x00020 118 #define ZIO_FLAG_DONT_PROPAGATE 0x00040 119 #define ZIO_FLAG_DONT_RETRY 0x00080 120 121 #define ZIO_FLAG_PHYSICAL 0x00100 122 #define ZIO_FLAG_IO_BYPASS 0x00200 123 #define ZIO_FLAG_IO_REPAIR 0x00400 124 #define ZIO_FLAG_SPECULATIVE 0x00800 125 126 #define ZIO_FLAG_RESILVER 0x01000 127 #define ZIO_FLAG_SCRUB 0x02000 128 #define ZIO_FLAG_SCRUB_THREAD 0x04000 129 #define ZIO_FLAG_SUBBLOCK 0x08000 130 131 #define ZIO_FLAG_NOBOOKMARK 0x10000 132 #define ZIO_FLAG_USER 0x20000 133 134 #define ZIO_FLAG_GANG_INHERIT \ 135 (ZIO_FLAG_CANFAIL | \ 136 ZIO_FLAG_FAILFAST | \ 137 ZIO_FLAG_CONFIG_HELD | \ 138 ZIO_FLAG_DONT_RETRY | \ 139 ZIO_FLAG_IO_REPAIR | \ 140 ZIO_FLAG_SPECULATIVE | \ 141 ZIO_FLAG_RESILVER | \ 142 ZIO_FLAG_SCRUB | \ 143 ZIO_FLAG_SCRUB_THREAD) 144 145 #define ZIO_FLAG_VDEV_INHERIT \ 146 (ZIO_FLAG_GANG_INHERIT | \ 147 ZIO_FLAG_DONT_CACHE | \ 148 ZIO_FLAG_PHYSICAL) 149 150 /* 151 * We'll take the unused errno 'EBADE' (from the Convergent graveyard) 152 * to indicate checksum errors. 153 */ 154 #define ECKSUM EBADE 155 156 typedef struct zio zio_t; 157 typedef void zio_done_func_t(zio_t *zio); 158 159 extern uint8_t zio_priority_table[ZIO_PRIORITY_TABLE_SIZE]; 160 extern char *zio_type_name[ZIO_TYPES]; 161 162 /* 163 * A bookmark is a four-tuple <objset, object, level, blkid> that uniquely 164 * identifies any block in the pool. By convention, the meta-objset (MOS) 165 * is objset 0, the meta-dnode is object 0, the root block (osphys_t) is 166 * level -1 of the meta-dnode, and intent log blocks (which are chained 167 * off the root block) have blkid == sequence number. In summary: 168 * 169 * mos is objset 0 170 * meta-dnode is object 0 171 * root block is <objset, 0, -1, 0> 172 * intent log is <objset, 0, -1, ZIL sequence number> 173 * 174 * Note: this structure is called a bookmark because its first purpose was 175 * to remember where to resume a pool-wide traverse. The absolute ordering 176 * for block visitation during traversal is defined in compare_bookmark(). 177 * 178 * Note: this structure is passed between userland and the kernel. 179 * Therefore it must not change size or alignment between 32/64 bit 180 * compilation options. 181 */ 182 typedef struct zbookmark { 183 uint64_t zb_objset; 184 uint64_t zb_object; 185 int64_t zb_level; 186 uint64_t zb_blkid; 187 } zbookmark_t; 188 189 struct zio { 190 /* Core information about this I/O */ 191 zio_t *io_parent; 192 zio_t *io_root; 193 spa_t *io_spa; 194 zbookmark_t io_bookmark; 195 enum zio_checksum io_checksum; 196 enum zio_compress io_compress; 197 int io_ndvas; 198 uint64_t io_txg; 199 blkptr_t *io_bp; 200 blkptr_t io_bp_copy; 201 zio_t *io_child; 202 zio_t *io_sibling_prev; 203 zio_t *io_sibling_next; 204 zio_transform_t *io_transform_stack; 205 zio_t *io_logical; 206 207 /* Callback info */ 208 zio_done_func_t *io_done; 209 void *io_private; 210 blkptr_t io_bp_orig; 211 212 /* Data represented by this I/O */ 213 void *io_data; 214 uint64_t io_size; 215 216 /* Stuff for the vdev stack */ 217 vdev_t *io_vd; 218 void *io_vsd; 219 uint64_t io_offset; 220 uint64_t io_deadline; 221 uint64_t io_timestamp; 222 avl_node_t io_offset_node; 223 avl_node_t io_deadline_node; 224 avl_tree_t *io_vdev_tree; 225 zio_t *io_delegate_list; 226 zio_t *io_delegate_next; 227 228 /* Internal pipeline state */ 229 int io_flags; 230 enum zio_type io_type; 231 enum zio_stage io_stage; 232 uint8_t io_stalled; 233 uint8_t io_priority; 234 struct dk_callback io_dk_callback; 235 int io_cmd; 236 int io_retries; 237 int io_error; 238 uint32_t io_numerrors; 239 uint32_t io_pipeline; 240 uint32_t io_async_stages; 241 uint64_t io_children_notready; 242 uint64_t io_children_notdone; 243 void *io_waiter; 244 kmutex_t io_lock; 245 kcondvar_t io_cv; 246 247 /* FMA state */ 248 uint64_t io_ena; 249 }; 250 251 extern zio_t *zio_null(zio_t *pio, spa_t *spa, 252 zio_done_func_t *done, void *private, int flags); 253 254 extern zio_t *zio_root(spa_t *spa, 255 zio_done_func_t *done, void *private, int flags); 256 257 extern zio_t *zio_read(zio_t *pio, spa_t *spa, blkptr_t *bp, void *data, 258 uint64_t size, zio_done_func_t *done, void *private, 259 int priority, int flags, zbookmark_t *zb); 260 261 extern zio_t *zio_write(zio_t *pio, spa_t *spa, int checksum, int compress, 262 int ncopies, uint64_t txg, blkptr_t *bp, void *data, uint64_t size, 263 zio_done_func_t *done, void *private, int priority, int flags, 264 zbookmark_t *zb); 265 266 extern zio_t *zio_rewrite(zio_t *pio, spa_t *spa, int checksum, 267 uint64_t txg, blkptr_t *bp, void *data, uint64_t size, 268 zio_done_func_t *done, void *private, int priority, int flags, 269 zbookmark_t *zb); 270 271 extern zio_t *zio_free(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, 272 zio_done_func_t *done, void *private); 273 274 extern zio_t *zio_claim(zio_t *pio, spa_t *spa, uint64_t txg, blkptr_t *bp, 275 zio_done_func_t *done, void *private); 276 277 extern zio_t *zio_ioctl(zio_t *pio, spa_t *spa, vdev_t *vd, int cmd, 278 zio_done_func_t *done, void *private, int priority, int flags); 279 280 extern zio_t *zio_read_phys(zio_t *pio, vdev_t *vd, uint64_t offset, 281 uint64_t size, void *data, int checksum, 282 zio_done_func_t *done, void *private, int priority, int flags); 283 284 extern zio_t *zio_write_phys(zio_t *pio, vdev_t *vd, uint64_t offset, 285 uint64_t size, void *data, int checksum, 286 zio_done_func_t *done, void *private, int priority, int flags); 287 288 extern int zio_alloc_blk(spa_t *spa, uint64_t size, blkptr_t *bp, uint64_t txg); 289 extern void zio_free_blk(spa_t *spa, blkptr_t *bp, uint64_t txg); 290 291 extern int zio_wait(zio_t *zio); 292 extern void zio_nowait(zio_t *zio); 293 294 extern void *zio_buf_alloc(size_t size); 295 extern void zio_buf_free(void *buf, size_t size); 296 297 /* 298 * Move an I/O to the next stage of the pipeline and execute that stage. 299 * There's no locking on io_stage because there's no legitimate way for 300 * multiple threads to be attempting to process the same I/O. 301 */ 302 extern void zio_next_stage(zio_t *zio); 303 extern void zio_next_stage_async(zio_t *zio); 304 extern void zio_wait_children_done(zio_t *zio); 305 306 /* 307 * Delegate I/O to a child vdev. 308 */ 309 extern zio_t *zio_vdev_child_io(zio_t *zio, blkptr_t *bp, vdev_t *vd, 310 uint64_t offset, void *data, uint64_t size, int type, int priority, 311 int flags, zio_done_func_t *done, void *private); 312 313 extern void zio_vdev_io_bypass(zio_t *zio); 314 extern void zio_vdev_io_reissue(zio_t *zio); 315 extern void zio_vdev_io_redone(zio_t *zio); 316 317 extern void zio_checksum_verified(zio_t *zio); 318 extern void zio_set_gang_verifier(zio_t *zio, zio_cksum_t *zcp); 319 320 extern uint8_t zio_checksum_select(uint8_t child, uint8_t parent); 321 extern uint8_t zio_compress_select(uint8_t child, uint8_t parent); 322 323 boolean_t zio_should_retry(zio_t *zio); 324 325 /* 326 * Initial setup and teardown. 327 */ 328 extern void zio_init(void); 329 extern void zio_fini(void); 330 331 /* 332 * Fault injection 333 */ 334 struct zinject_record; 335 extern uint32_t zio_injection_enabled; 336 extern int zio_inject_fault(char *name, int flags, int *id, 337 struct zinject_record *record); 338 extern int zio_inject_list_next(int *id, char *name, size_t buflen, 339 struct zinject_record *record); 340 extern int zio_clear_fault(int id); 341 extern int zio_handle_fault_injection(zio_t *zio, int error); 342 extern int zio_handle_device_injection(vdev_t *vd, int error); 343 344 #ifdef __cplusplus 345 } 346 #endif 347 348 #endif /* _ZIO_H */ 349