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 * Copyright 2013 Saso Kiselkov. All rights reserved.
26 */
27
28 #ifndef _SYS_SPA_IMPL_H
29 #define _SYS_SPA_IMPL_H
30
31 #include <sys/spa.h>
32 #include <sys/vdev.h>
33 #include <sys/metaslab.h>
34 #include <sys/dmu.h>
35 #include <sys/dsl_pool.h>
36 #include <sys/uberblock_impl.h>
37 #include <sys/zfs_context.h>
38 #include <sys/avl.h>
39 #include <sys/refcount.h>
40 #include <sys/bplist.h>
41 #include <sys/bpobj.h>
42
43 #ifdef __cplusplus
44 extern "C" {
45 #endif
46
47 typedef struct spa_error_entry {
48 zbookmark_t se_bookmark;
49 char *se_name;
50 avl_node_t se_avl;
51 } spa_error_entry_t;
52
53 typedef struct spa_history_phys {
54 uint64_t sh_pool_create_len; /* ending offset of zpool create */
55 uint64_t sh_phys_max_off; /* physical EOF */
56 uint64_t sh_bof; /* logical BOF */
57 uint64_t sh_eof; /* logical EOF */
58 uint64_t sh_records_lost; /* num of records overwritten */
59 } spa_history_phys_t;
60
61 struct spa_aux_vdev {
62 uint64_t sav_object; /* MOS object for device list */
63 nvlist_t *sav_config; /* cached device config */
64 vdev_t **sav_vdevs; /* devices */
65 int sav_count; /* number devices */
66 boolean_t sav_sync; /* sync the device list */
67 nvlist_t **sav_pending; /* pending device additions */
68 uint_t sav_npending; /* # pending devices */
69 };
70
71 typedef struct spa_config_lock {
72 kmutex_t scl_lock;
73 kthread_t *scl_writer;
74 int scl_write_wanted;
75 kcondvar_t scl_cv;
76 refcount_t scl_count;
77 } spa_config_lock_t;
78
79 typedef struct spa_config_dirent {
80 list_node_t scd_link;
81 char *scd_path;
82 } spa_config_dirent_t;
83
84 typedef enum zio_taskq_type {
85 ZIO_TASKQ_ISSUE = 0,
86 ZIO_TASKQ_ISSUE_HIGH,
87 ZIO_TASKQ_INTERRUPT,
88 ZIO_TASKQ_INTERRUPT_HIGH,
89 ZIO_TASKQ_TYPES
90 } zio_taskq_type_t;
91
92 /*
93 * State machine for the zpool-poolname process. The states transitions
94 * are done as follows:
95 *
96 * From To Routine
97 * PROC_NONE -> PROC_CREATED spa_activate()
98 * PROC_CREATED -> PROC_ACTIVE spa_thread()
99 * PROC_ACTIVE -> PROC_DEACTIVATE spa_deactivate()
100 * PROC_DEACTIVATE -> PROC_GONE spa_thread()
101 * PROC_GONE -> PROC_NONE spa_deactivate()
102 */
103 typedef enum spa_proc_state {
104 SPA_PROC_NONE, /* spa_proc = &p0, no process created */
105 SPA_PROC_CREATED, /* spa_activate() has proc, is waiting */
106 SPA_PROC_ACTIVE, /* taskqs created, spa_proc set */
107 SPA_PROC_DEACTIVATE, /* spa_deactivate() requests process exit */
108 SPA_PROC_GONE /* spa_thread() is exiting, spa_proc = &p0 */
109 } spa_proc_state_t;
110
111 typedef struct spa_taskqs {
112 uint_t stqs_count;
113 taskq_t **stqs_taskq;
114 } spa_taskqs_t;
115
116 struct spa {
117 /*
118 * Fields protected by spa_namespace_lock.
119 */
120 char spa_name[MAXNAMELEN]; /* pool name */
121 char *spa_comment; /* comment */
122 avl_node_t spa_avl; /* node in spa_namespace_avl */
123 nvlist_t *spa_config; /* last synced config */
124 nvlist_t *spa_config_syncing; /* currently syncing config */
125 nvlist_t *spa_config_splitting; /* config for splitting */
126 nvlist_t *spa_load_info; /* info and errors from load */
127 uint64_t spa_config_txg; /* txg of last config change */
128 int spa_sync_pass; /* iterate-to-convergence */
129 pool_state_t spa_state; /* pool state */
130 int spa_inject_ref; /* injection references */
131 uint8_t spa_sync_on; /* sync threads are running */
132 spa_load_state_t spa_load_state; /* current load operation */
133 uint64_t spa_import_flags; /* import specific flags */
134 spa_taskqs_t spa_zio_taskq[ZIO_TYPES][ZIO_TASKQ_TYPES];
135 dsl_pool_t *spa_dsl_pool;
136 boolean_t spa_is_initializing; /* true while opening pool */
137 metaslab_class_t *spa_normal_class; /* normal data class */
138 metaslab_class_t *spa_log_class; /* intent log data class */
139 uint64_t spa_first_txg; /* first txg after spa_open() */
140 uint64_t spa_final_txg; /* txg of export/destroy */
141 uint64_t spa_freeze_txg; /* freeze pool at this txg */
142 uint64_t spa_load_max_txg; /* best initial ub_txg */
143 uint64_t spa_claim_max_txg; /* highest claimed birth txg */
144 timespec_t spa_loaded_ts; /* 1st successful open time */
145 objset_t *spa_meta_objset; /* copy of dp->dp_meta_objset */
146 txg_list_t spa_vdev_txg_list; /* per-txg dirty vdev list */
147 vdev_t *spa_root_vdev; /* top-level vdev container */
148 uint64_t spa_config_guid; /* config pool guid */
149 uint64_t spa_load_guid; /* spa_load initialized guid */
150 uint64_t spa_last_synced_guid; /* last synced guid */
151 list_t spa_config_dirty_list; /* vdevs with dirty config */
152 list_t spa_state_dirty_list; /* vdevs with dirty state */
153 spa_aux_vdev_t spa_spares; /* hot spares */
154 spa_aux_vdev_t spa_l2cache; /* L2ARC cache devices */
155 nvlist_t *spa_label_features; /* Features for reading MOS */
156 uint64_t spa_config_object; /* MOS object for pool config */
157 uint64_t spa_config_generation; /* config generation number */
158 uint64_t spa_syncing_txg; /* txg currently syncing */
159 bpobj_t spa_deferred_bpobj; /* deferred-free bplist */
160 bplist_t spa_free_bplist[TXG_SIZE]; /* bplist of stuff to free */
161 zio_cksum_salt_t spa_cksum_salt; /* secret salt for cksum */
162 uint64_t spa_cksum_salt_obj; /* persistent salt object */
163 /* checksum context templates */
164 kmutex_t spa_cksum_tmpls_lock;
165 void *spa_cksum_tmpls[ZIO_CHECKSUM_FUNCTIONS];
166 uberblock_t spa_ubsync; /* last synced uberblock */
167 uberblock_t spa_uberblock; /* current uberblock */
168 boolean_t spa_extreme_rewind; /* rewind past deferred frees */
169 uint64_t spa_last_io; /* lbolt of last non-scan I/O */
170 kmutex_t spa_scrub_lock; /* resilver/scrub lock */
171 uint64_t spa_scrub_inflight; /* in-flight scrub I/Os */
172 kcondvar_t spa_scrub_io_cv; /* scrub I/O completion */
173 uint8_t spa_scrub_active; /* active or suspended? */
174 uint8_t spa_scrub_type; /* type of scrub we're doing */
175 uint8_t spa_scrub_finished; /* indicator to rotate logs */
176 uint8_t spa_scrub_started; /* started since last boot */
177 uint8_t spa_scrub_reopen; /* scrub doing vdev_reopen */
178 uint64_t spa_scan_pass_start; /* start time per pass/reboot */
179 uint64_t spa_scan_pass_exam; /* examined bytes per pass */
180 kmutex_t spa_async_lock; /* protect async state */
181 kthread_t *spa_async_thread; /* thread doing async task */
182 int spa_async_suspended; /* async tasks suspended */
183 kcondvar_t spa_async_cv; /* wait for thread_exit() */
184 uint16_t spa_async_tasks; /* async task mask */
185 char *spa_root; /* alternate root directory */
186 uint64_t spa_ena; /* spa-wide ereport ENA */
187 int spa_last_open_failed; /* error if last open failed */
188 uint64_t spa_last_ubsync_txg; /* "best" uberblock txg */
189 uint64_t spa_last_ubsync_txg_ts; /* timestamp from that ub */
190 uint64_t spa_load_txg; /* ub txg that loaded */
191 uint64_t spa_load_txg_ts; /* timestamp from that ub */
192 uint64_t spa_load_meta_errors; /* verify metadata err count */
193 uint64_t spa_load_data_errors; /* verify data err count */
194 uint64_t spa_verify_min_txg; /* start txg of verify scrub */
195 kmutex_t spa_errlog_lock; /* error log lock */
196 uint64_t spa_errlog_last; /* last error log object */
197 uint64_t spa_errlog_scrub; /* scrub error log object */
198 kmutex_t spa_errlist_lock; /* error list/ereport lock */
199 avl_tree_t spa_errlist_last; /* last error list */
200 avl_tree_t spa_errlist_scrub; /* scrub error list */
201 uint64_t spa_deflate; /* should we deflate? */
202 uint64_t spa_history; /* history object */
203 kmutex_t spa_history_lock; /* history lock */
204 vdev_t *spa_pending_vdev; /* pending vdev additions */
205 kmutex_t spa_props_lock; /* property lock */
206 uint64_t spa_pool_props_object; /* object for properties */
207 uint64_t spa_bootfs; /* default boot filesystem */
208 uint64_t spa_failmode; /* failure mode for the pool */
209 uint64_t spa_delegation; /* delegation on/off */
210 list_t spa_config_list; /* previous cache file(s) */
211 zio_t *spa_async_zio_root; /* root of all async I/O */
212 zio_t *spa_suspend_zio_root; /* root of all suspended I/O */
213 kmutex_t spa_suspend_lock; /* protects suspend_zio_root */
214 kcondvar_t spa_suspend_cv; /* notification of resume */
215 uint8_t spa_suspended; /* pool is suspended */
216 uint8_t spa_claiming; /* pool is doing zil_claim() */
217 boolean_t spa_debug; /* debug enabled? */
218 boolean_t spa_is_root; /* pool is root */
219 int spa_minref; /* num refs when first opened */
220 int spa_mode; /* FREAD | FWRITE */
221 spa_log_state_t spa_log_state; /* log state */
222 uint64_t spa_autoexpand; /* lun expansion on/off */
223 ddt_t *spa_ddt[ZIO_CHECKSUM_FUNCTIONS]; /* in-core DDTs */
224 uint64_t spa_ddt_stat_object; /* DDT statistics */
225 uint64_t spa_dedup_ditto; /* dedup ditto threshold */
226 uint64_t spa_dedup_checksum; /* default dedup checksum */
227 uint64_t spa_dspace; /* dspace in normal class */
228 kmutex_t spa_vdev_top_lock; /* dueling offline/remove */
229 kmutex_t spa_proc_lock; /* protects spa_proc* */
230 kcondvar_t spa_proc_cv; /* spa_proc_state transitions */
231 spa_proc_state_t spa_proc_state; /* see definition */
232 struct proc *spa_proc; /* "zpool-poolname" process */
233 uint64_t spa_did; /* if procp != p0, did of t1 */
234 boolean_t spa_autoreplace; /* autoreplace set in open */
235 int spa_vdev_locks; /* locks grabbed */
236 uint64_t spa_creation_version; /* version at pool creation */
237 uint64_t spa_prev_software_version; /* See ub_software_version */
238 uint64_t spa_feat_for_write_obj; /* required to write to pool */
239 uint64_t spa_feat_for_read_obj; /* required to read from pool */
240 uint64_t spa_feat_desc_obj; /* Feature descriptions */
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 */