spa_impl.h revision 307108
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) 2011, 2015 by Delphix. All rights reserved. 24 * Copyright 2011 Nexenta Systems, Inc. All rights reserved. 25 * Copyright 2013 Martin Matuska <mm@FreeBSD.org>. All rights reserved. 26 * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. 27 * Copyright 2013 Saso Kiselkov. All rights reserved. 28 */ 29 30#ifndef _SYS_SPA_IMPL_H 31#define _SYS_SPA_IMPL_H 32 33#include <sys/spa.h> 34#include <sys/vdev.h> 35#include <sys/metaslab.h> 36#include <sys/dmu.h> 37#include <sys/dsl_pool.h> 38#include <sys/uberblock_impl.h> 39#include <sys/zfs_context.h> 40#include <sys/avl.h> 41#include <sys/refcount.h> 42#include <sys/bplist.h> 43#include <sys/bpobj.h> 44#include <sys/zfeature.h> 45#include <zfeature_common.h> 46 47#ifdef __cplusplus 48extern "C" { 49#endif 50 51typedef struct spa_error_entry { 52 zbookmark_phys_t se_bookmark; 53 char *se_name; 54 avl_node_t se_avl; 55} spa_error_entry_t; 56 57typedef struct spa_history_phys { 58 uint64_t sh_pool_create_len; /* ending offset of zpool create */ 59 uint64_t sh_phys_max_off; /* physical EOF */ 60 uint64_t sh_bof; /* logical BOF */ 61 uint64_t sh_eof; /* logical EOF */ 62 uint64_t sh_records_lost; /* num of records overwritten */ 63} spa_history_phys_t; 64 65struct spa_aux_vdev { 66 uint64_t sav_object; /* MOS object for device list */ 67 nvlist_t *sav_config; /* cached device config */ 68 vdev_t **sav_vdevs; /* devices */ 69 int sav_count; /* number devices */ 70 boolean_t sav_sync; /* sync the device list */ 71 nvlist_t **sav_pending; /* pending device additions */ 72 uint_t sav_npending; /* # pending devices */ 73}; 74 75typedef struct spa_config_lock { 76 kmutex_t scl_lock; 77 kthread_t *scl_writer; 78 int scl_write_wanted; 79 kcondvar_t scl_cv; 80 refcount_t scl_count; 81} spa_config_lock_t; 82 83typedef struct spa_config_dirent { 84 list_node_t scd_link; 85 char *scd_path; 86} spa_config_dirent_t; 87 88typedef enum zio_taskq_type { 89 ZIO_TASKQ_ISSUE = 0, 90 ZIO_TASKQ_ISSUE_HIGH, 91 ZIO_TASKQ_INTERRUPT, 92 ZIO_TASKQ_INTERRUPT_HIGH, 93 ZIO_TASKQ_TYPES 94} zio_taskq_type_t; 95 96/* 97 * State machine for the zpool-poolname process. The states transitions 98 * are done as follows: 99 * 100 * From To Routine 101 * PROC_NONE -> PROC_CREATED spa_activate() 102 * PROC_CREATED -> PROC_ACTIVE spa_thread() 103 * PROC_ACTIVE -> PROC_DEACTIVATE spa_deactivate() 104 * PROC_DEACTIVATE -> PROC_GONE spa_thread() 105 * PROC_GONE -> PROC_NONE spa_deactivate() 106 */ 107typedef enum spa_proc_state { 108 SPA_PROC_NONE, /* spa_proc = &p0, no process created */ 109 SPA_PROC_CREATED, /* spa_activate() has proc, is waiting */ 110 SPA_PROC_ACTIVE, /* taskqs created, spa_proc set */ 111 SPA_PROC_DEACTIVATE, /* spa_deactivate() requests process exit */ 112 SPA_PROC_GONE /* spa_thread() is exiting, spa_proc = &p0 */ 113} spa_proc_state_t; 114 115typedef struct spa_taskqs { 116 uint_t stqs_count; 117 taskq_t **stqs_taskq; 118} spa_taskqs_t; 119 120typedef enum spa_all_vdev_zap_action { 121 AVZ_ACTION_NONE = 0, 122 AVZ_ACTION_DESTROY, /* Destroy all per-vdev ZAPs and the AVZ. */ 123 AVZ_ACTION_REBUILD /* Populate the new AVZ, see spa_avz_rebuild */ 124} spa_avz_action_t; 125 126struct spa { 127 /* 128 * Fields protected by spa_namespace_lock. 129 */ 130 char spa_name[ZFS_MAX_DATASET_NAME_LEN]; /* pool name */ 131 char *spa_comment; /* comment */ 132 avl_node_t spa_avl; /* node in spa_namespace_avl */ 133 nvlist_t *spa_config; /* last synced config */ 134 nvlist_t *spa_config_syncing; /* currently syncing config */ 135 nvlist_t *spa_config_splitting; /* config for splitting */ 136 nvlist_t *spa_load_info; /* info and errors from load */ 137 uint64_t spa_config_txg; /* txg of last config change */ 138 int spa_sync_pass; /* iterate-to-convergence */ 139 pool_state_t spa_state; /* pool state */ 140 int spa_inject_ref; /* injection references */ 141 uint8_t spa_sync_on; /* sync threads are running */ 142 spa_load_state_t spa_load_state; /* current load operation */ 143 uint64_t spa_import_flags; /* import specific flags */ 144 spa_taskqs_t spa_zio_taskq[ZIO_TYPES][ZIO_TASKQ_TYPES]; 145 dsl_pool_t *spa_dsl_pool; 146 boolean_t spa_is_initializing; /* true while opening pool */ 147 metaslab_class_t *spa_normal_class; /* normal data class */ 148 metaslab_class_t *spa_log_class; /* intent log data class */ 149 uint64_t spa_first_txg; /* first txg after spa_open() */ 150 uint64_t spa_final_txg; /* txg of export/destroy */ 151 uint64_t spa_freeze_txg; /* freeze pool at this txg */ 152 uint64_t spa_load_max_txg; /* best initial ub_txg */ 153 uint64_t spa_claim_max_txg; /* highest claimed birth txg */ 154 timespec_t spa_loaded_ts; /* 1st successful open time */ 155 objset_t *spa_meta_objset; /* copy of dp->dp_meta_objset */ 156 kmutex_t spa_evicting_os_lock; /* Evicting objset list lock */ 157 list_t spa_evicting_os_list; /* Objsets being evicted. */ 158 kcondvar_t spa_evicting_os_cv; /* Objset Eviction Completion */ 159 txg_list_t spa_vdev_txg_list; /* per-txg dirty vdev list */ 160 vdev_t *spa_root_vdev; /* top-level vdev container */ 161 int spa_min_ashift; /* of vdevs in normal class */ 162 int spa_max_ashift; /* of vdevs in normal class */ 163 uint64_t spa_config_guid; /* config pool guid */ 164 uint64_t spa_load_guid; /* spa_load initialized guid */ 165 uint64_t spa_last_synced_guid; /* last synced guid */ 166 list_t spa_config_dirty_list; /* vdevs with dirty config */ 167 list_t spa_state_dirty_list; /* vdevs with dirty state */ 168 spa_aux_vdev_t spa_spares; /* hot spares */ 169 spa_aux_vdev_t spa_l2cache; /* L2ARC cache devices */ 170 nvlist_t *spa_label_features; /* Features for reading MOS */ 171 uint64_t spa_config_object; /* MOS object for pool config */ 172 uint64_t spa_config_generation; /* config generation number */ 173 uint64_t spa_syncing_txg; /* txg currently syncing */ 174 bpobj_t spa_deferred_bpobj; /* deferred-free bplist */ 175 bplist_t spa_free_bplist[TXG_SIZE]; /* bplist of stuff to free */ 176 zio_cksum_salt_t spa_cksum_salt; /* secret salt for cksum */ 177 /* checksum context templates */ 178 kmutex_t spa_cksum_tmpls_lock; 179 void *spa_cksum_tmpls[ZIO_CHECKSUM_FUNCTIONS]; 180 uberblock_t spa_ubsync; /* last synced uberblock */ 181 uberblock_t spa_uberblock; /* current uberblock */ 182 boolean_t spa_extreme_rewind; /* rewind past deferred frees */ 183 uint64_t spa_last_io; /* lbolt of last non-scan I/O */ 184 kmutex_t spa_scrub_lock; /* resilver/scrub lock */ 185 uint64_t spa_scrub_inflight; /* in-flight scrub I/Os */ 186 kcondvar_t spa_scrub_io_cv; /* scrub I/O completion */ 187 uint8_t spa_scrub_active; /* active or suspended? */ 188 uint8_t spa_scrub_type; /* type of scrub we're doing */ 189 uint8_t spa_scrub_finished; /* indicator to rotate logs */ 190 uint8_t spa_scrub_started; /* started since last boot */ 191 uint8_t spa_scrub_reopen; /* scrub doing vdev_reopen */ 192 uint64_t spa_scan_pass_start; /* start time per pass/reboot */ 193 uint64_t spa_scan_pass_exam; /* examined bytes per pass */ 194 kmutex_t spa_async_lock; /* protect async state */ 195 kthread_t *spa_async_thread; /* thread doing async task */ 196 kthread_t *spa_async_thread_vd; /* thread doing vd async task */ 197 int spa_async_suspended; /* async tasks suspended */ 198 kcondvar_t spa_async_cv; /* wait for thread_exit() */ 199 uint16_t spa_async_tasks; /* async task mask */ 200 char *spa_root; /* alternate root directory */ 201 uint64_t spa_ena; /* spa-wide ereport ENA */ 202 int spa_last_open_failed; /* error if last open failed */ 203 uint64_t spa_last_ubsync_txg; /* "best" uberblock txg */ 204 uint64_t spa_last_ubsync_txg_ts; /* timestamp from that ub */ 205 uint64_t spa_load_txg; /* ub txg that loaded */ 206 uint64_t spa_load_txg_ts; /* timestamp from that ub */ 207 uint64_t spa_load_meta_errors; /* verify metadata err count */ 208 uint64_t spa_load_data_errors; /* verify data err count */ 209 uint64_t spa_verify_min_txg; /* start txg of verify scrub */ 210 kmutex_t spa_errlog_lock; /* error log lock */ 211 uint64_t spa_errlog_last; /* last error log object */ 212 uint64_t spa_errlog_scrub; /* scrub error log object */ 213 kmutex_t spa_errlist_lock; /* error list/ereport lock */ 214 avl_tree_t spa_errlist_last; /* last error list */ 215 avl_tree_t spa_errlist_scrub; /* scrub error list */ 216 uint64_t spa_deflate; /* should we deflate? */ 217 uint64_t spa_history; /* history object */ 218 kmutex_t spa_history_lock; /* history lock */ 219 vdev_t *spa_pending_vdev; /* pending vdev additions */ 220 kmutex_t spa_props_lock; /* property lock */ 221 uint64_t spa_pool_props_object; /* object for properties */ 222 uint64_t spa_bootfs; /* default boot filesystem */ 223 uint64_t spa_failmode; /* failure mode for the pool */ 224 uint64_t spa_delegation; /* delegation on/off */ 225 list_t spa_config_list; /* previous cache file(s) */ 226 /* per-CPU array of root of async I/O: */ 227 zio_t **spa_async_zio_root; 228 zio_t *spa_suspend_zio_root; /* root of all suspended I/O */ 229 kmutex_t spa_suspend_lock; /* protects suspend_zio_root */ 230 kcondvar_t spa_suspend_cv; /* notification of resume */ 231 uint8_t spa_suspended; /* pool is suspended */ 232 uint8_t spa_claiming; /* pool is doing zil_claim() */ 233 boolean_t spa_debug; /* debug enabled? */ 234 boolean_t spa_is_root; /* pool is root */ 235 int spa_minref; /* num refs when first opened */ 236 int spa_mode; /* FREAD | FWRITE */ 237 spa_log_state_t spa_log_state; /* log state */ 238 uint64_t spa_autoexpand; /* lun expansion on/off */ 239 ddt_t *spa_ddt[ZIO_CHECKSUM_FUNCTIONS]; /* in-core DDTs */ 240 uint64_t spa_ddt_stat_object; /* DDT statistics */ 241 uint64_t spa_dedup_ditto; /* dedup ditto threshold */ 242 uint64_t spa_dedup_checksum; /* default dedup checksum */ 243 uint64_t spa_dspace; /* dspace in normal class */ 244 kmutex_t spa_vdev_top_lock; /* dueling offline/remove */ 245 kmutex_t spa_proc_lock; /* protects spa_proc* */ 246 kcondvar_t spa_proc_cv; /* spa_proc_state transitions */ 247 spa_proc_state_t spa_proc_state; /* see definition */ 248 struct proc *spa_proc; /* "zpool-poolname" process */ 249 uint64_t spa_did; /* if procp != p0, did of t1 */ 250 kthread_t *spa_trim_thread; /* thread sending TRIM I/Os */ 251 kmutex_t spa_trim_lock; /* protects spa_trim_cv */ 252 kcondvar_t spa_trim_cv; /* used to notify TRIM thread */ 253 boolean_t spa_autoreplace; /* autoreplace set in open */ 254 int spa_vdev_locks; /* locks grabbed */ 255 uint64_t spa_creation_version; /* version at pool creation */ 256 uint64_t spa_prev_software_version; /* See ub_software_version */ 257 uint64_t spa_feat_for_write_obj; /* required to write to pool */ 258 uint64_t spa_feat_for_read_obj; /* required to read from pool */ 259 uint64_t spa_feat_desc_obj; /* Feature descriptions */ 260 uint64_t spa_feat_enabled_txg_obj; /* Feature enabled txg */ 261 /* cache feature refcounts */ 262 uint64_t spa_feat_refcount_cache[SPA_FEATURES]; 263#ifdef illumos 264 cyclic_id_t spa_deadman_cycid; /* cyclic id */ 265#else /* !illumos */ 266#ifdef _KERNEL 267 struct callout spa_deadman_cycid; /* callout id */ 268#endif 269#endif /* illumos */ 270 uint64_t spa_deadman_calls; /* number of deadman calls */ 271 hrtime_t spa_sync_starttime; /* starting time fo spa_sync */ 272 uint64_t spa_deadman_synctime; /* deadman expiration timer */ 273 uint64_t spa_all_vdev_zaps; /* ZAP of per-vd ZAP obj #s */ 274 spa_avz_action_t spa_avz_action; /* destroy/rebuild AVZ? */ 275 276#ifdef illumos 277 /* 278 * spa_iokstat_lock protects spa_iokstat and 279 * spa_queue_stats[]. 280 */ 281 kmutex_t spa_iokstat_lock; 282 struct kstat *spa_iokstat; /* kstat of io to this pool */ 283 struct { 284 int spa_active; 285 int spa_queued; 286 } spa_queue_stats[ZIO_PRIORITY_NUM_QUEUEABLE]; 287#endif 288 hrtime_t spa_ccw_fail_time; /* Conf cache write fail time */ 289 290 /* 291 * spa_refcount & spa_config_lock must be the last elements 292 * because refcount_t changes size based on compilation options. 293 * In order for the MDB module to function correctly, the other 294 * fields must remain in the same location. 295 */ 296 spa_config_lock_t spa_config_lock[SCL_LOCKS]; /* config changes */ 297 refcount_t spa_refcount; /* number of opens */ 298#ifndef illumos 299 boolean_t spa_splitting_newspa; /* creating new spa in split */ 300#endif 301}; 302 303extern const char *spa_config_path; 304 305extern void spa_taskq_dispatch_ent(spa_t *spa, zio_type_t t, zio_taskq_type_t q, 306 task_func_t *func, void *arg, uint_t flags, taskq_ent_t *ent); 307 308#ifdef __cplusplus 309} 310#endif 311 312#endif /* _SYS_SPA_IMPL_H */ 313