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