2 * Copyright (c) 2020 iXsystems, Inc.
5 * Redistribution and use in source and binary forms, with or without
6 * modification, are permitted provided that the following conditions
8 * 1. Redistributions of source code must retain the above copyright
9 * notice, this list of conditions and the following disclaimer.
10 * 2. Redistributions in binary form must reproduce the above copyright
11 * notice, this list of conditions and the following disclaimer in the
12 * documentation and/or other materials provided with the distribution.
14 * THIS SOFTWARE IS PROVIDED BY THE AUTHORS AND CONTRIBUTORS ``AS IS'' AND
15 * ANY EXPRESS OR IMPLIED WARRANTIES, INCLUDING, BUT NOT LIMITED TO, THE
16 * IMPLIED WARRANTIES OF MERCHANTABILITY AND FITNESS FOR A PARTICULAR PURPOSE
17 * ARE DISCLAIMED. IN NO EVENT SHALL THE AUTHORS OR CONTRIBUTORS BE LIABLE
18 * FOR ANY DIRECT, INDIRECT, INCIDENTAL, SPECIAL, EXEMPLARY, OR CONSEQUENTIAL
19 * DAMAGES (INCLUDING, BUT NOT LIMITED TO, PROCUREMENT OF SUBSTITUTE GOODS
20 * OR SERVICES; LOSS OF USE, DATA, OR PROFITS; OR BUSINESS INTERRUPTION)
21 * HOWEVER CAUSED AND ON ANY THEORY OF LIABILITY, WHETHER IN CONTRACT, STRICT
22 * LIABILITY, OR TORT (INCLUDING NEGLIGENCE OR OTHERWISE) ARISING IN ANY WAY
23 * OUT OF THE USE OF THIS SOFTWARE, EVEN IF ADVISED OF THE POSSIBILITY OF
28 #include <sys/cdefs.h>
29 __FBSDID("$FreeBSD$");
31 #include <sys/types.h>
32 #include <sys/param.h>
33 #include <sys/systm.h>
35 #include <sys/kernel.h>
37 #include <sys/malloc.h>
38 #include <sys/mutex.h>
40 #include <sys/errno.h>
46 #include <sys/cmn_err.h>
48 #include <sys/zfs_ioctl.h>
49 #include <sys/zfs_vfsops.h>
50 #include <sys/zfs_znode.h>
53 #include <sys/spa_impl.h>
55 #include <sys/vdev_impl.h>
57 #include <sys/dsl_dir.h>
58 #include <sys/dsl_dataset.h>
59 #include <sys/dsl_prop.h>
60 #include <sys/dsl_deleg.h>
61 #include <sys/dmu_objset.h>
62 #include <sys/dmu_impl.h>
63 #include <sys/dmu_tx.h>
64 #include <sys/sunddi.h>
65 #include <sys/policy.h>
67 #include <sys/nvpair.h>
68 #include <sys/mount.h>
69 #include <sys/taskqueue.h>
71 #include <sys/fs/zfs.h>
72 #include <sys/zfs_ctldir.h>
73 #include <sys/zfs_dir.h>
74 #include <sys/zfs_onexit.h>
76 #include <sys/dsl_scan.h>
77 #include <sys/dmu_objset.h>
78 #include <sys/dmu_send.h>
79 #include <sys/dsl_destroy.h>
80 #include <sys/dsl_bookmark.h>
81 #include <sys/dsl_userhold.h>
82 #include <sys/zfeature.h>
84 #include <sys/zio_checksum.h>
85 #include <sys/vdev_removal.h>
86 #include <sys/dsl_crypt.h>
88 #include <sys/zfs_ioctl_compat.h>
89 #include <sys/zfs_context.h>
91 #include <sys/arc_impl.h>
92 #include <sys/dsl_pool.h>
96 SYSCTL_DECL(_vfs_zfs);
97 SYSCTL_NODE(_vfs_zfs, OID_AUTO, arc, CTLFLAG_RW, 0, "ZFS adaptive replacement cache");
98 SYSCTL_NODE(_vfs_zfs, OID_AUTO, condense, CTLFLAG_RW, 0, "ZFS condense");
99 SYSCTL_NODE(_vfs_zfs, OID_AUTO, dbuf, CTLFLAG_RW, 0, "ZFS disk buf cache");
100 SYSCTL_NODE(_vfs_zfs, OID_AUTO, dbuf_cache, CTLFLAG_RW, 0, "ZFS disk buf cache");
101 SYSCTL_NODE(_vfs_zfs, OID_AUTO, deadman, CTLFLAG_RW, 0, "ZFS deadman");
102 SYSCTL_NODE(_vfs_zfs, OID_AUTO, dedup, CTLFLAG_RW, 0, "ZFS dedup");
103 SYSCTL_NODE(_vfs_zfs, OID_AUTO, l2arc, CTLFLAG_RW, 0, "ZFS l2arc");
104 SYSCTL_NODE(_vfs_zfs, OID_AUTO, livelist, CTLFLAG_RW, 0, "ZFS livelist");
105 SYSCTL_NODE(_vfs_zfs, OID_AUTO, lua, CTLFLAG_RW, 0, "ZFS lua");
106 SYSCTL_NODE(_vfs_zfs, OID_AUTO, metaslab, CTLFLAG_RW, 0, "ZFS metaslab");
107 SYSCTL_NODE(_vfs_zfs, OID_AUTO, mg, CTLFLAG_RW, 0, "ZFS metaslab group");
108 SYSCTL_NODE(_vfs_zfs, OID_AUTO, multihost, CTLFLAG_RW, 0, "ZFS multihost protection");
109 SYSCTL_NODE(_vfs_zfs, OID_AUTO, prefetch, CTLFLAG_RW, 0, "ZFS prefetch");
110 SYSCTL_NODE(_vfs_zfs, OID_AUTO, reconstruct, CTLFLAG_RW, 0, "ZFS reconstruct");
111 SYSCTL_NODE(_vfs_zfs, OID_AUTO, recv, CTLFLAG_RW, 0, "ZFS receive");
112 SYSCTL_NODE(_vfs_zfs, OID_AUTO, send, CTLFLAG_RW, 0, "ZFS send");
113 SYSCTL_NODE(_vfs_zfs, OID_AUTO, spa, CTLFLAG_RW, 0, "ZFS space allocation");
114 SYSCTL_NODE(_vfs_zfs, OID_AUTO, trim, CTLFLAG_RW, 0, "ZFS TRIM");
115 SYSCTL_NODE(_vfs_zfs, OID_AUTO, txg, CTLFLAG_RW, 0, "ZFS transaction group");
116 SYSCTL_NODE(_vfs_zfs, OID_AUTO, vdev, CTLFLAG_RW, 0, "ZFS VDEV");
117 SYSCTL_NODE(_vfs_zfs, OID_AUTO, zevent, CTLFLAG_RW, 0, "ZFS event");
118 SYSCTL_NODE(_vfs_zfs, OID_AUTO, zil, CTLFLAG_RW, 0, "ZFS ZIL");
119 SYSCTL_NODE(_vfs_zfs, OID_AUTO, zio, CTLFLAG_RW, 0, "ZFS ZIO");
121 SYSCTL_NODE(_vfs_zfs_livelist, OID_AUTO, condense, CTLFLAG_RW, 0,
122 "ZFS livelist condense");
123 SYSCTL_NODE(_vfs_zfs_vdev, OID_AUTO, cache, CTLFLAG_RW, 0, "ZFS VDEV Cache");
124 SYSCTL_NODE(_vfs_zfs_vdev, OID_AUTO, mirror, CTLFLAG_RD, 0,
127 SYSCTL_DECL(_vfs_zfs_version);
128 SYSCTL_CONST_STRING(_vfs_zfs_version, OID_AUTO, module, CTLFLAG_RD,
129 (ZFS_META_VERSION "-" ZFS_META_RELEASE), "OpenZFS module version");
131 extern arc_state_t ARC_anon;
132 extern arc_state_t ARC_mru;
133 extern arc_state_t ARC_mru_ghost;
134 extern arc_state_t ARC_mfu;
135 extern arc_state_t ARC_mfu_ghost;
136 extern arc_state_t ARC_l2c_only;
139 * minimum lifespan of a prefetch block in clock ticks
140 * (initialized in arc_init())
146 extern uint64_t l2arc_write_max; /* def max write size */
147 extern uint64_t l2arc_write_boost; /* extra warmup write */
148 extern uint64_t l2arc_headroom; /* # of dev writes */
149 extern uint64_t l2arc_headroom_boost;
150 extern uint64_t l2arc_feed_secs; /* interval seconds */
151 extern uint64_t l2arc_feed_min_ms; /* min interval msecs */
152 extern int l2arc_noprefetch; /* don't cache prefetch bufs */
153 extern int l2arc_feed_again; /* turbo warmup */
154 extern int l2arc_norw; /* no reads during writes */
156 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_max, CTLFLAG_RW,
157 &l2arc_write_max, 0, "max write size (LEGACY)");
158 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_boost, CTLFLAG_RW,
159 &l2arc_write_boost, 0, "extra write during warmup (LEGACY)");
160 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_headroom, CTLFLAG_RW,
161 &l2arc_headroom, 0, "number of dev writes (LEGACY)");
162 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_secs, CTLFLAG_RW,
163 &l2arc_feed_secs, 0, "interval seconds (LEGACY)");
164 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_min_ms, CTLFLAG_RW,
165 &l2arc_feed_min_ms, 0, "min interval milliseconds (LEGACY)");
167 SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_noprefetch, CTLFLAG_RW,
168 &l2arc_noprefetch, 0, "don't cache prefetch bufs (LEGACY)");
169 SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_feed_again, CTLFLAG_RW,
170 &l2arc_feed_again, 0, "turbo warmup (LEGACY)");
171 SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_norw, CTLFLAG_RW,
172 &l2arc_norw, 0, "no reads during writes (LEGACY)");
174 extern int zfs_compressed_arc_enabled;
175 SYSCTL_INT(_vfs_zfs, OID_AUTO, compressed_arc_enabled, CTLFLAG_RW,
176 &zfs_compressed_arc_enabled, 1, "compressed arc buffers (LEGACY)");
179 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_size, CTLFLAG_RD,
180 &ARC_anon.arcs_size.rc_count, 0, "size of anonymous state");
181 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_metadata_esize, CTLFLAG_RD,
182 &ARC_anon.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
183 "size of anonymous state");
184 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_data_esize, CTLFLAG_RD,
185 &ARC_anon.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
186 "size of anonymous state");
188 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_size, CTLFLAG_RD,
189 &ARC_mru.arcs_size.rc_count, 0, "size of mru state");
190 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_metadata_esize, CTLFLAG_RD,
191 &ARC_mru.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
192 "size of metadata in mru state");
193 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_data_esize, CTLFLAG_RD,
194 &ARC_mru.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
195 "size of data in mru state");
197 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_size, CTLFLAG_RD,
198 &ARC_mru_ghost.arcs_size.rc_count, 0, "size of mru ghost state");
199 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_metadata_esize, CTLFLAG_RD,
200 &ARC_mru_ghost.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
201 "size of metadata in mru ghost state");
202 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_data_esize, CTLFLAG_RD,
203 &ARC_mru_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
204 "size of data in mru ghost state");
206 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_size, CTLFLAG_RD,
207 &ARC_mfu.arcs_size.rc_count, 0, "size of mfu state");
208 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_metadata_esize, CTLFLAG_RD,
209 &ARC_mfu.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
210 "size of metadata in mfu state");
211 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_data_esize, CTLFLAG_RD,
212 &ARC_mfu.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
213 "size of data in mfu state");
215 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_size, CTLFLAG_RD,
216 &ARC_mfu_ghost.arcs_size.rc_count, 0, "size of mfu ghost state");
217 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_metadata_esize, CTLFLAG_RD,
218 &ARC_mfu_ghost.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
219 "size of metadata in mfu ghost state");
220 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_data_esize, CTLFLAG_RD,
221 &ARC_mfu_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
222 "size of data in mfu ghost state");
224 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2c_only_size, CTLFLAG_RD,
225 &ARC_l2c_only.arcs_size.rc_count, 0, "size of mru state");
228 sysctl_vfs_zfs_arc_no_grow_shift(SYSCTL_HANDLER_ARGS)
233 val = arc_no_grow_shift;
234 err = sysctl_handle_32(oidp, &val, 0, req);
235 if (err != 0 || req->newptr == NULL)
238 if (val >= arc_shrink_shift)
241 arc_no_grow_shift = val;
245 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_no_grow_shift, CTLTYPE_U32 | CTLFLAG_RWTUN,
246 0, sizeof (uint32_t), sysctl_vfs_zfs_arc_no_grow_shift, "U",
247 "log2(fraction of ARC which must be free to allow growing)");
250 param_set_arc_long(SYSCTL_HANDLER_ARGS)
254 err = sysctl_handle_long(oidp, arg1, 0, req);
255 if (err != 0 || req->newptr == NULL)
258 arc_tuning_update(B_TRUE);
264 param_set_arc_int(SYSCTL_HANDLER_ARGS)
268 err = sysctl_handle_int(oidp, arg1, 0, req);
269 if (err != 0 || req->newptr == NULL)
272 arc_tuning_update(B_TRUE);
277 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_min, CTLTYPE_ULONG | CTLFLAG_RWTUN,
278 &zfs_arc_min, sizeof (zfs_arc_min), param_set_arc_long, "LU",
279 "min arc size (LEGACY)");
280 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_max, CTLTYPE_ULONG | CTLFLAG_RWTUN,
281 &zfs_arc_max, sizeof (zfs_arc_max), param_set_arc_long, "LU",
282 "max arc size (LEGACY)");
290 SYSCTL_NODE(_vfs_zfs, OID_AUTO, zfetch, CTLFLAG_RW, 0, "ZFS ZFETCH (LEGACY)");
292 /* max bytes to prefetch per stream (default 8MB) */
293 extern uint32_t zfetch_max_distance;
294 SYSCTL_UINT(_vfs_zfs_zfetch, OID_AUTO, max_distance, CTLFLAG_RWTUN,
295 &zfetch_max_distance, 0, "Max bytes to prefetch per stream (LEGACY)");
297 /* max bytes to prefetch indirects for per stream (default 64MB) */
298 extern uint32_t zfetch_max_idistance;
299 SYSCTL_UINT(_vfs_zfs_prefetch, OID_AUTO, max_idistance, CTLFLAG_RWTUN,
300 &zfetch_max_idistance, 0, "Max bytes to prefetch indirects for per stream");
305 extern int zfs_default_bs;
306 SYSCTL_INT(_vfs_zfs, OID_AUTO, default_bs, CTLFLAG_RWTUN,
307 &zfs_default_bs, 0, "Default dnode block shift");
309 extern int zfs_default_ibs;
310 SYSCTL_INT(_vfs_zfs, OID_AUTO, default_ibs, CTLFLAG_RWTUN,
311 &zfs_default_ibs, 0, "Default dnode indirect block shift");
319 * In pools where the log space map feature is not enabled we touch
320 * multiple metaslabs (and their respective space maps) with each
321 * transaction group. Thus, we benefit from having a small space map
322 * block size since it allows us to issue more I/O operations scattered
323 * around the disk. So a sane default for the space map block size
326 extern int zfs_metaslab_sm_blksz_no_log;
327 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_no_log, CTLFLAG_RDTUN,
328 &zfs_metaslab_sm_blksz_no_log, 0,
329 "Block size for space map in pools with log space map disabled. "
330 "Power of 2 and greater than 4096.");
333 * When the log space map feature is enabled, we accumulate a lot of
334 * changes per metaslab that are flushed once in a while so we benefit
335 * from a bigger block size like 128K for the metaslab space maps.
337 extern int zfs_metaslab_sm_blksz_with_log;
338 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_with_log, CTLFLAG_RDTUN,
339 &zfs_metaslab_sm_blksz_with_log, 0,
340 "Block size for space map in pools with log space map enabled. "
341 "Power of 2 and greater than 4096.");
344 * The in-core space map representation is more compact than its on-disk form.
345 * The zfs_condense_pct determines how much more compact the in-core
346 * space map representation must be before we compact it on-disk.
347 * Values should be greater than or equal to 100.
349 extern int zfs_condense_pct;
350 SYSCTL_INT(_vfs_zfs, OID_AUTO, condense_pct, CTLFLAG_RWTUN,
351 &zfs_condense_pct, 0,
352 "Condense on-disk spacemap when it is more than this many percents"
353 " of in-memory counterpart");
355 extern int zfs_remove_max_segment;
356 SYSCTL_INT(_vfs_zfs, OID_AUTO, remove_max_segment, CTLFLAG_RWTUN,
357 &zfs_remove_max_segment, 0, "Largest contiguous segment ZFS will attempt to"
358 " allocate when removing a device");
360 extern int zfs_removal_suspend_progress;
361 SYSCTL_INT(_vfs_zfs, OID_AUTO, removal_suspend_progress, CTLFLAG_RWTUN,
362 &zfs_removal_suspend_progress, 0, "Ensures certain actions can happen while"
363 " in the middle of a removal");
367 * Minimum size which forces the dynamic allocator to change
368 * it's allocation strategy. Once the space map cannot satisfy
369 * an allocation of this size then it switches to using more
370 * aggressive strategy (i.e search by size rather than offset).
372 extern uint64_t metaslab_df_alloc_threshold;
373 SYSCTL_QUAD(_vfs_zfs_metaslab, OID_AUTO, df_alloc_threshold, CTLFLAG_RWTUN,
374 &metaslab_df_alloc_threshold, 0,
375 "Minimum size which forces the dynamic allocator to change it's allocation strategy");
378 * The minimum free space, in percent, which must be available
379 * in a space map to continue allocations in a first-fit fashion.
380 * Once the space map's free space drops below this level we dynamically
381 * switch to using best-fit allocations.
383 extern int metaslab_df_free_pct;
384 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, df_free_pct, CTLFLAG_RWTUN,
385 &metaslab_df_free_pct, 0,
386 "The minimum free space, in percent, which must be available in a "
387 "space map to continue allocations in a first-fit fashion");
390 * Percentage of all cpus that can be used by the metaslab taskq.
392 extern int metaslab_load_pct;
393 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, load_pct, CTLFLAG_RWTUN,
394 &metaslab_load_pct, 0,
395 "Percentage of cpus that can be used by the metaslab taskq");
398 * Max number of metaslabs per group to preload.
400 extern int metaslab_preload_limit;
401 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, preload_limit, CTLFLAG_RWTUN,
402 &metaslab_preload_limit, 0,
403 "Max number of metaslabs per group to preload");
406 extern int reference_tracking_enable;
407 SYSCTL_INT(_vfs_zfs, OID_AUTO, reference_tracking_enable, CTLFLAG_RDTUN,
408 &reference_tracking_enable, 0,
409 "Track reference holders to refcount_t objects, used mostly by ZFS");
412 extern int zfs_ccw_retry_interval;
413 SYSCTL_INT(_vfs_zfs, OID_AUTO, ccw_retry_interval, CTLFLAG_RWTUN,
414 &zfs_ccw_retry_interval, 0,
415 "Configuration cache file write, retry after failure, interval (seconds)");
417 extern uint64_t zfs_max_missing_tvds_cachefile;
418 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, max_missing_tvds_cachefile, CTLFLAG_RWTUN,
419 &zfs_max_missing_tvds_cachefile, 0,
420 "allow importing pools with missing top-level vdevs in cache file");
422 extern uint64_t zfs_max_missing_tvds_scan;
423 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, max_missing_tvds_scan, CTLFLAG_RWTUN,
424 &zfs_max_missing_tvds_scan, 0,
425 "allow importing pools with missing top-level vdevs during scan");
428 extern int zfs_flags;
430 sysctl_vfs_zfs_debug_flags(SYSCTL_HANDLER_ARGS)
435 err = sysctl_handle_int(oidp, &val, 0, req);
436 if (err != 0 || req->newptr == NULL)
440 * ZFS_DEBUG_MODIFY must be enabled prior to boot so all
441 * arc buffers in the system have the necessary additional
442 * checksum data. However, it is safe to disable at any
445 if (!(zfs_flags & ZFS_DEBUG_MODIFY))
446 val &= ~ZFS_DEBUG_MODIFY;
452 SYSCTL_PROC(_vfs_zfs, OID_AUTO, debugflags,
453 CTLTYPE_UINT | CTLFLAG_MPSAFE | CTLFLAG_RWTUN, NULL, 0,
454 sysctl_vfs_zfs_debug_flags, "IU", "Debug flags for ZFS testing.");
457 param_set_deadman_synctime(SYSCTL_HANDLER_ARGS)
462 val = zfs_deadman_synctime_ms;
463 err = sysctl_handle_long(oidp, &val, 0, req);
464 if (err != 0 || req->newptr == NULL)
466 zfs_deadman_synctime_ms = val;
468 spa_set_deadman_synctime(MSEC2NSEC(zfs_deadman_synctime_ms));
474 param_set_deadman_ziotime(SYSCTL_HANDLER_ARGS)
479 val = zfs_deadman_ziotime_ms;
480 err = sysctl_handle_long(oidp, &val, 0, req);
481 if (err != 0 || req->newptr == NULL)
483 zfs_deadman_ziotime_ms = val;
485 spa_set_deadman_ziotime(MSEC2NSEC(zfs_deadman_synctime_ms));
491 param_set_deadman_failmode(SYSCTL_HANDLER_ARGS)
496 if (req->newptr == NULL)
497 strlcpy(buf, zfs_deadman_failmode, sizeof (buf));
499 rc = sysctl_handle_string(oidp, buf, sizeof (buf), req);
500 if (rc || req->newptr == NULL)
502 if (strcmp(buf, zfs_deadman_failmode) == 0)
504 if (!strcmp(buf, "wait"))
505 zfs_deadman_failmode = "wait";
506 if (!strcmp(buf, "continue"))
507 zfs_deadman_failmode = "continue";
508 if (!strcmp(buf, "panic"))
509 zfs_deadman_failmode = "panic";
511 return (-param_set_deadman_failmode_common(buf));
516 extern int space_map_ibs;
517 SYSCTL_INT(_vfs_zfs, OID_AUTO, space_map_ibs, CTLFLAG_RWTUN,
518 &space_map_ibs, 0, "Space map indirect block shift");
523 param_set_min_auto_ashift(SYSCTL_HANDLER_ARGS)
528 val = zfs_vdev_min_auto_ashift;
529 err = sysctl_handle_64(oidp, &val, 0, req);
530 if (err != 0 || req->newptr == NULL)
531 return (SET_ERROR(err));
533 if (val < ASHIFT_MIN || val > zfs_vdev_max_auto_ashift)
534 return (SET_ERROR(EINVAL));
536 zfs_vdev_min_auto_ashift = val;
542 param_set_max_auto_ashift(SYSCTL_HANDLER_ARGS)
547 val = zfs_vdev_max_auto_ashift;
548 err = sysctl_handle_64(oidp, &val, 0, req);
549 if (err != 0 || req->newptr == NULL)
550 return (SET_ERROR(err));
552 if (val > ASHIFT_MAX || val < zfs_vdev_min_auto_ashift)
553 return (SET_ERROR(EINVAL));
555 zfs_vdev_max_auto_ashift = val;
560 SYSCTL_PROC(_vfs_zfs, OID_AUTO, min_auto_ashift, CTLTYPE_U64 | CTLFLAG_RWTUN,
561 &zfs_vdev_min_auto_ashift, sizeof (zfs_vdev_min_auto_ashift),
562 param_set_min_auto_ashift, "QU",
563 "Min ashift used when creating new top-level vdev. (LEGACY)");
564 SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift, CTLTYPE_U64 | CTLFLAG_RWTUN,
565 &zfs_vdev_max_auto_ashift, sizeof (zfs_vdev_max_auto_ashift),
566 param_set_max_auto_ashift, "QU",
567 "Max ashift used when optimizing for logical -> physical sector size on "
568 "new top-level vdevs. (LEGACY)");
571 * Since the DTL space map of a vdev is not expected to have a lot of
572 * entries, we default its block size to 4K.
574 extern int zfs_vdev_dtl_sm_blksz;
575 SYSCTL_INT(_vfs_zfs, OID_AUTO, dtl_sm_blksz, CTLFLAG_RDTUN,
576 &zfs_vdev_dtl_sm_blksz, 0,
577 "Block size for DTL space map. Power of 2 and greater than 4096.");
580 * vdev-wide space maps that have lots of entries written to them at
581 * the end of each transaction can benefit from a higher I/O bandwidth
582 * (e.g. vdev_obsolete_sm), thus we default their block size to 128K.
584 extern int zfs_vdev_standard_sm_blksz;
585 SYSCTL_INT(_vfs_zfs, OID_AUTO, standard_sm_blksz, CTLFLAG_RDTUN,
586 &zfs_vdev_standard_sm_blksz, 0,
587 "Block size for standard space map. Power of 2 and greater than 4096.");
589 extern int vdev_validate_skip;
590 SYSCTL_INT(_vfs_zfs, OID_AUTO, validate_skip, CTLFLAG_RDTUN,
591 &vdev_validate_skip, 0,
592 "Enable to bypass vdev_validate().");
599 * The load configuration settings below are tuned by default for
600 * the case where all devices are of the same rotational type.
602 * If there is a mixture of rotating and non-rotating media, setting
603 * non_rotating_seek_inc to 0 may well provide better results as it
604 * will direct more reads to the non-rotating vdevs which are more
605 * likely to have a higher performance.
610 #define ZFS_VDEV_QUEUE_KNOB_MIN(name) \
611 extern uint32_t zfs_vdev_ ## name ## _min_active; \
612 SYSCTL_UINT(_vfs_zfs_vdev, OID_AUTO, name ## _min_active, CTLFLAG_RWTUN,\
613 &zfs_vdev_ ## name ## _min_active, 0, \
614 "Initial number of I/O requests of type " #name \
615 " active for each device");
617 #define ZFS_VDEV_QUEUE_KNOB_MAX(name) \
618 extern uint32_t zfs_vdev_ ## name ## _max_active; \
619 SYSCTL_UINT(_vfs_zfs_vdev, OID_AUTO, name ## _max_active, CTLFLAG_RWTUN, \
620 &zfs_vdev_ ## name ## _max_active, 0, \
621 "Maximum number of I/O requests of type " #name \
622 " active for each device");
625 #undef ZFS_VDEV_QUEUE_KNOB
627 extern uint32_t zfs_vdev_max_active;
628 SYSCTL_UINT(_vfs_zfs, OID_AUTO, top_maxinflight, CTLFLAG_RWTUN,
629 &zfs_vdev_max_active, 0,
630 "The maximum number of I/Os of all types active for each device. (LEGACY)");
632 extern int zfs_vdev_def_queue_depth;
633 SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, def_queue_depth, CTLFLAG_RWTUN,
634 &zfs_vdev_def_queue_depth, 0,
635 "Default queue depth for each allocator");
637 /*extern uint64_t zfs_multihost_history;
638 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, multihost_history, CTLFLAG_RWTUN,
639 &zfs_multihost_history, 0,
640 "Historical staticists for the last N multihost updates");*/
643 SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, trim_on_init, CTLFLAG_RW,
644 &vdev_trim_on_init, 0, "Enable/disable full vdev trim on initialisation");
649 #if defined(__LP64__)
655 SYSCTL_INT(_vfs_zfs_zio, OID_AUTO, use_uma, CTLFLAG_RDTUN, &zio_use_uma, 0,
656 "Use uma(9) for ZIO allocations");
657 SYSCTL_INT(_vfs_zfs_zio, OID_AUTO, exclude_metadata, CTLFLAG_RDTUN, &zio_exclude_metadata, 0,
658 "Exclude metadata buffers from dumps as well");
661 param_set_slop_shift(SYSCTL_HANDLER_ARGS)
668 err = sysctl_handle_int(oidp, &val, 0, req);
669 if (err != 0 || req->newptr == NULL)
672 if (val < 1 || val > 31)
681 param_set_multihost_interval(SYSCTL_HANDLER_ARGS)
685 err = sysctl_handle_long(oidp, arg1, 0, req);
686 if (err != 0 || req->newptr == NULL)
689 if (spa_mode_global != SPA_MODE_UNINIT)
690 mmp_signal_all_threads();