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