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 https://opensource.org/licenses/CDDL-1.0. 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 2009 Sun Microsystems, Inc. All rights reserved. 23 * Use is subject to license terms. 24 */ 25 26/* 27 * Copyright (c) 2012, 2015 by Delphix. All rights reserved. 28 * Copyright (c) 2024, Klara Inc. 29 */ 30 31#ifndef _ZIO_IMPL_H 32#define _ZIO_IMPL_H 33 34#ifdef __cplusplus 35extern "C" { 36#endif 37 38/* 39 * XXX -- Describe ZFS I/O pipeline here. Fill in as needed. 40 * 41 * The ZFS I/O pipeline is comprised of various stages which are defined 42 * in the zio_stage enum below. The individual stages are used to construct 43 * these basic I/O operations: Read, Write, Free, Claim, Flush and Trim. 44 * 45 * I/O operations: (XXX - provide detail for each of the operations) 46 * 47 * Read: 48 * Write: 49 * Free: 50 * Claim: 51 * Flush: 52 * Trim: 53 * 54 * Although the most common pipeline are used by the basic I/O operations 55 * above, there are some helper pipelines (one could consider them 56 * sub-pipelines) which are used internally by the ZIO module and are 57 * explained below: 58 * 59 * Interlock Pipeline: 60 * The interlock pipeline is the most basic pipeline and is used by all 61 * of the I/O operations. The interlock pipeline does not perform any I/O 62 * and is used to coordinate the dependencies between I/Os that are being 63 * issued (i.e. the parent/child relationship). 64 * 65 * Vdev child Pipeline: 66 * The vdev child pipeline is responsible for performing the physical I/O. 67 * It is in this pipeline where the I/O are queued and possibly cached. 68 * 69 * In addition to performing I/O, the pipeline is also responsible for 70 * data transformations. The transformations performed are based on the 71 * specific properties that user may have selected and modify the 72 * behavior of the pipeline. Examples of supported transformations are 73 * compression, dedup, and nop writes. Transformations will either modify 74 * the data or the pipeline. This list below further describes each of 75 * the supported transformations: 76 * 77 * Compression: 78 * ZFS supports five different flavors of compression -- gzip, lzjb, lz4, zle, 79 * and zstd. Compression occurs as part of the write pipeline and is 80 * performed in the ZIO_STAGE_WRITE_BP_INIT stage. 81 * 82 * Block cloning: 83 * The block cloning functionality introduces ZIO_STAGE_BRT_FREE stage which 84 * is called during a free pipeline. If the block is referenced in the 85 * Block Cloning Table (BRT) we will just decrease its reference counter 86 * instead of actually freeing the block. 87 * 88 * Dedup: 89 * Dedup reads are handled by the ZIO_STAGE_DDT_READ_START and 90 * ZIO_STAGE_DDT_READ_DONE stages. These stages are added to an existing 91 * read pipeline if the dedup bit is set on the block pointer. 92 * Writing a dedup block is performed by the ZIO_STAGE_DDT_WRITE stage 93 * and added to a write pipeline if a user has enabled dedup on that 94 * particular dataset. 95 * 96 * NOP Write: 97 * The NOP write feature is performed by the ZIO_STAGE_NOP_WRITE stage 98 * and is added to an existing write pipeline if a cryptographically 99 * secure checksum (i.e. SHA256) is enabled and compression is turned on. 100 * The NOP write stage will compare the checksums of the current data 101 * on-disk (level-0 blocks only) and the data that is currently being written. 102 * If the checksum values are identical then the pipeline is converted to 103 * an interlock pipeline skipping block allocation and bypassing the 104 * physical I/O. The nop write feature can handle writes in either 105 * syncing or open context (i.e. zil writes) and as a result is mutually 106 * exclusive with dedup. 107 * 108 * Encryption: 109 * Encryption and authentication is handled by the ZIO_STAGE_ENCRYPT stage. 110 * This stage determines how the encryption metadata is stored in the bp. 111 * Decryption and MAC verification is performed during zio_decrypt() as a 112 * transform callback. Encryption is mutually exclusive with nopwrite, because 113 * blocks with the same plaintext will be encrypted with different salts and 114 * IV's (if dedup is off), and therefore have different ciphertexts. For dedup 115 * blocks we deterministically generate the IV and salt by performing an HMAC 116 * of the plaintext, which is computationally expensive, but allows us to keep 117 * support for encrypted dedup. See the block comment in zio_crypt.c for 118 * details. 119 */ 120 121/* 122 * zio pipeline stage definitions 123 */ 124enum zio_stage { 125 ZIO_STAGE_OPEN = 1 << 0, /* RWFCXT */ 126 127 ZIO_STAGE_READ_BP_INIT = 1 << 1, /* R----- */ 128 ZIO_STAGE_WRITE_BP_INIT = 1 << 2, /* -W---- */ 129 ZIO_STAGE_FREE_BP_INIT = 1 << 3, /* --F--- */ 130 ZIO_STAGE_ISSUE_ASYNC = 1 << 4, /* -WF--T */ 131 ZIO_STAGE_WRITE_COMPRESS = 1 << 5, /* -W---- */ 132 133 ZIO_STAGE_ENCRYPT = 1 << 6, /* -W---- */ 134 ZIO_STAGE_CHECKSUM_GENERATE = 1 << 7, /* -W---- */ 135 136 ZIO_STAGE_NOP_WRITE = 1 << 8, /* -W---- */ 137 138 ZIO_STAGE_BRT_FREE = 1 << 9, /* --F--- */ 139 140 ZIO_STAGE_DDT_READ_START = 1 << 10, /* R----- */ 141 ZIO_STAGE_DDT_READ_DONE = 1 << 11, /* R----- */ 142 ZIO_STAGE_DDT_WRITE = 1 << 12, /* -W---- */ 143 ZIO_STAGE_DDT_FREE = 1 << 13, /* --F--- */ 144 145 ZIO_STAGE_GANG_ASSEMBLE = 1 << 14, /* RWFC-- */ 146 ZIO_STAGE_GANG_ISSUE = 1 << 15, /* RWFC-- */ 147 148 ZIO_STAGE_DVA_THROTTLE = 1 << 16, /* -W---- */ 149 ZIO_STAGE_DVA_ALLOCATE = 1 << 17, /* -W---- */ 150 ZIO_STAGE_DVA_FREE = 1 << 18, /* --F--- */ 151 ZIO_STAGE_DVA_CLAIM = 1 << 19, /* ---C-- */ 152 153 ZIO_STAGE_READY = 1 << 20, /* RWFCXT */ 154 155 ZIO_STAGE_VDEV_IO_START = 1 << 21, /* RW--XT */ 156 ZIO_STAGE_VDEV_IO_DONE = 1 << 22, /* RW--XT */ 157 ZIO_STAGE_VDEV_IO_ASSESS = 1 << 23, /* RW--XT */ 158 159 ZIO_STAGE_CHECKSUM_VERIFY = 1 << 24, /* R----- */ 160 161 ZIO_STAGE_DONE = 1 << 25 /* RWFCXT */ 162}; 163 164#define ZIO_ROOT_PIPELINE \ 165 ZIO_STAGE_DONE 166 167#define ZIO_INTERLOCK_STAGES \ 168 (ZIO_STAGE_READY | \ 169 ZIO_STAGE_DONE) 170 171#define ZIO_INTERLOCK_PIPELINE \ 172 ZIO_INTERLOCK_STAGES 173 174#define ZIO_VDEV_IO_STAGES \ 175 (ZIO_STAGE_VDEV_IO_START | \ 176 ZIO_STAGE_VDEV_IO_DONE | \ 177 ZIO_STAGE_VDEV_IO_ASSESS) 178 179#define ZIO_VDEV_CHILD_PIPELINE \ 180 (ZIO_VDEV_IO_STAGES | \ 181 ZIO_STAGE_DONE) 182 183#define ZIO_READ_COMMON_STAGES \ 184 (ZIO_INTERLOCK_STAGES | \ 185 ZIO_VDEV_IO_STAGES | \ 186 ZIO_STAGE_CHECKSUM_VERIFY) 187 188#define ZIO_READ_PHYS_PIPELINE \ 189 ZIO_READ_COMMON_STAGES 190 191#define ZIO_READ_PIPELINE \ 192 (ZIO_READ_COMMON_STAGES | \ 193 ZIO_STAGE_READ_BP_INIT) 194 195#define ZIO_DDT_CHILD_READ_PIPELINE \ 196 ZIO_READ_COMMON_STAGES 197 198#define ZIO_DDT_READ_PIPELINE \ 199 (ZIO_INTERLOCK_STAGES | \ 200 ZIO_STAGE_READ_BP_INIT | \ 201 ZIO_STAGE_DDT_READ_START | \ 202 ZIO_STAGE_DDT_READ_DONE) 203 204#define ZIO_WRITE_COMMON_STAGES \ 205 (ZIO_INTERLOCK_STAGES | \ 206 ZIO_VDEV_IO_STAGES | \ 207 ZIO_STAGE_ISSUE_ASYNC | \ 208 ZIO_STAGE_CHECKSUM_GENERATE) 209 210#define ZIO_WRITE_PHYS_PIPELINE \ 211 ZIO_WRITE_COMMON_STAGES 212 213#define ZIO_REWRITE_PIPELINE \ 214 (ZIO_WRITE_COMMON_STAGES | \ 215 ZIO_STAGE_WRITE_COMPRESS | \ 216 ZIO_STAGE_ENCRYPT | \ 217 ZIO_STAGE_WRITE_BP_INIT) 218 219#define ZIO_WRITE_PIPELINE \ 220 (ZIO_WRITE_COMMON_STAGES | \ 221 ZIO_STAGE_WRITE_BP_INIT | \ 222 ZIO_STAGE_WRITE_COMPRESS | \ 223 ZIO_STAGE_ENCRYPT | \ 224 ZIO_STAGE_DVA_THROTTLE | \ 225 ZIO_STAGE_DVA_ALLOCATE) 226 227#define ZIO_DDT_CHILD_WRITE_PIPELINE \ 228 (ZIO_INTERLOCK_STAGES | \ 229 ZIO_VDEV_IO_STAGES | \ 230 ZIO_STAGE_DVA_THROTTLE | \ 231 ZIO_STAGE_DVA_ALLOCATE) 232 233#define ZIO_DDT_WRITE_PIPELINE \ 234 (ZIO_INTERLOCK_STAGES | \ 235 ZIO_STAGE_WRITE_BP_INIT | \ 236 ZIO_STAGE_ISSUE_ASYNC | \ 237 ZIO_STAGE_WRITE_COMPRESS | \ 238 ZIO_STAGE_ENCRYPT | \ 239 ZIO_STAGE_CHECKSUM_GENERATE | \ 240 ZIO_STAGE_DDT_WRITE) 241 242#define ZIO_GANG_STAGES \ 243 (ZIO_STAGE_GANG_ASSEMBLE | \ 244 ZIO_STAGE_GANG_ISSUE) 245 246#define ZIO_FREE_PIPELINE \ 247 (ZIO_INTERLOCK_STAGES | \ 248 ZIO_STAGE_FREE_BP_INIT | \ 249 ZIO_STAGE_BRT_FREE | \ 250 ZIO_STAGE_DVA_FREE) 251 252#define ZIO_DDT_FREE_PIPELINE \ 253 (ZIO_INTERLOCK_STAGES | \ 254 ZIO_STAGE_FREE_BP_INIT | \ 255 ZIO_STAGE_ISSUE_ASYNC | \ 256 ZIO_STAGE_DDT_FREE) 257 258#define ZIO_CLAIM_PIPELINE \ 259 (ZIO_INTERLOCK_STAGES | \ 260 ZIO_STAGE_DVA_CLAIM) 261 262#define ZIO_FLUSH_PIPELINE \ 263 (ZIO_INTERLOCK_STAGES | \ 264 ZIO_VDEV_IO_STAGES) 265 266#define ZIO_TRIM_PIPELINE \ 267 (ZIO_INTERLOCK_STAGES | \ 268 ZIO_STAGE_ISSUE_ASYNC | \ 269 ZIO_VDEV_IO_STAGES) 270 271#define ZIO_BLOCKING_STAGES \ 272 (ZIO_STAGE_DVA_ALLOCATE | \ 273 ZIO_STAGE_DVA_CLAIM | \ 274 ZIO_STAGE_VDEV_IO_START) 275 276extern void zio_inject_init(void); 277extern void zio_inject_fini(void); 278 279#ifdef __cplusplus 280} 281#endif 282 283#endif /* _ZIO_IMPL_H */ 284