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, file, CTLFLAG_RW, 0, "ZFS VDEV file");
125 SYSCTL_NODE(_vfs_zfs_vdev, OID_AUTO, mirror, CTLFLAG_RD, 0,
128 SYSCTL_DECL(_vfs_zfs_version);
129 SYSCTL_CONST_STRING(_vfs_zfs_version, OID_AUTO, module, CTLFLAG_RD,
130 (ZFS_META_VERSION "-" ZFS_META_RELEASE), "OpenZFS module version");
132 extern arc_state_t ARC_anon;
133 extern arc_state_t ARC_mru;
134 extern arc_state_t ARC_mru_ghost;
135 extern arc_state_t ARC_mfu;
136 extern arc_state_t ARC_mfu_ghost;
137 extern arc_state_t ARC_l2c_only;
140 * minimum lifespan of a prefetch block in clock ticks
141 * (initialized in arc_init())
147 extern uint64_t l2arc_write_max; /* def max write size */
148 extern uint64_t l2arc_write_boost; /* extra warmup write */
149 extern uint64_t l2arc_headroom; /* # of dev writes */
150 extern uint64_t l2arc_headroom_boost;
151 extern uint64_t l2arc_feed_secs; /* interval seconds */
152 extern uint64_t l2arc_feed_min_ms; /* min interval msecs */
153 extern int l2arc_noprefetch; /* don't cache prefetch bufs */
154 extern int l2arc_feed_again; /* turbo warmup */
155 extern int l2arc_norw; /* no reads during writes */
157 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_max, CTLFLAG_RW,
158 &l2arc_write_max, 0, "max write size (LEGACY)");
159 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_write_boost, CTLFLAG_RW,
160 &l2arc_write_boost, 0, "extra write during warmup (LEGACY)");
161 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_headroom, CTLFLAG_RW,
162 &l2arc_headroom, 0, "number of dev writes (LEGACY)");
163 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_secs, CTLFLAG_RW,
164 &l2arc_feed_secs, 0, "interval seconds (LEGACY)");
165 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2arc_feed_min_ms, CTLFLAG_RW,
166 &l2arc_feed_min_ms, 0, "min interval milliseconds (LEGACY)");
168 SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_noprefetch, CTLFLAG_RW,
169 &l2arc_noprefetch, 0, "don't cache prefetch bufs (LEGACY)");
170 SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_feed_again, CTLFLAG_RW,
171 &l2arc_feed_again, 0, "turbo warmup (LEGACY)");
172 SYSCTL_INT(_vfs_zfs, OID_AUTO, l2arc_norw, CTLFLAG_RW,
173 &l2arc_norw, 0, "no reads during writes (LEGACY)");
175 extern int zfs_compressed_arc_enabled;
176 SYSCTL_INT(_vfs_zfs, OID_AUTO, compressed_arc_enabled, CTLFLAG_RW,
177 &zfs_compressed_arc_enabled, 1, "compressed arc buffers (LEGACY)");
180 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_size, CTLFLAG_RD,
181 &ARC_anon.arcs_size.rc_count, 0, "size of anonymous state");
182 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_metadata_esize, CTLFLAG_RD,
183 &ARC_anon.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
184 "size of anonymous state");
185 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, anon_data_esize, CTLFLAG_RD,
186 &ARC_anon.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
187 "size of anonymous state");
189 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_size, CTLFLAG_RD,
190 &ARC_mru.arcs_size.rc_count, 0, "size of mru state");
191 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_metadata_esize, CTLFLAG_RD,
192 &ARC_mru.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
193 "size of metadata in mru state");
194 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_data_esize, CTLFLAG_RD,
195 &ARC_mru.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
196 "size of data in mru state");
198 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_size, CTLFLAG_RD,
199 &ARC_mru_ghost.arcs_size.rc_count, 0, "size of mru ghost state");
200 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_metadata_esize, CTLFLAG_RD,
201 &ARC_mru_ghost.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
202 "size of metadata in mru ghost state");
203 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mru_ghost_data_esize, CTLFLAG_RD,
204 &ARC_mru_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
205 "size of data in mru ghost state");
207 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_size, CTLFLAG_RD,
208 &ARC_mfu.arcs_size.rc_count, 0, "size of mfu state");
209 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_metadata_esize, CTLFLAG_RD,
210 &ARC_mfu.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
211 "size of metadata in mfu state");
212 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_data_esize, CTLFLAG_RD,
213 &ARC_mfu.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
214 "size of data in mfu state");
216 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_size, CTLFLAG_RD,
217 &ARC_mfu_ghost.arcs_size.rc_count, 0, "size of mfu ghost state");
218 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_metadata_esize, CTLFLAG_RD,
219 &ARC_mfu_ghost.arcs_esize[ARC_BUFC_METADATA].rc_count, 0,
220 "size of metadata in mfu ghost state");
221 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, mfu_ghost_data_esize, CTLFLAG_RD,
222 &ARC_mfu_ghost.arcs_esize[ARC_BUFC_DATA].rc_count, 0,
223 "size of data in mfu ghost state");
225 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, l2c_only_size, CTLFLAG_RD,
226 &ARC_l2c_only.arcs_size.rc_count, 0, "size of mru state");
229 sysctl_vfs_zfs_arc_no_grow_shift(SYSCTL_HANDLER_ARGS)
234 val = arc_no_grow_shift;
235 err = sysctl_handle_32(oidp, &val, 0, req);
236 if (err != 0 || req->newptr == NULL)
239 if (val >= arc_shrink_shift)
242 arc_no_grow_shift = val;
246 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_no_grow_shift, CTLTYPE_U32 | CTLFLAG_RWTUN,
247 0, sizeof (uint32_t), sysctl_vfs_zfs_arc_no_grow_shift, "U",
248 "log2(fraction of ARC which must be free to allow growing)");
251 param_set_arc_long(SYSCTL_HANDLER_ARGS)
255 err = sysctl_handle_long(oidp, arg1, 0, req);
256 if (err != 0 || req->newptr == NULL)
259 arc_tuning_update(B_TRUE);
265 param_set_arc_int(SYSCTL_HANDLER_ARGS)
269 err = sysctl_handle_int(oidp, arg1, 0, req);
270 if (err != 0 || req->newptr == NULL)
273 arc_tuning_update(B_TRUE);
278 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_min, CTLTYPE_ULONG | CTLFLAG_RWTUN,
279 &zfs_arc_min, sizeof (zfs_arc_min), param_set_arc_long, "LU",
280 "min arc size (LEGACY)");
281 SYSCTL_PROC(_vfs_zfs, OID_AUTO, arc_max, CTLTYPE_ULONG | CTLFLAG_RWTUN,
282 &zfs_arc_max, sizeof (zfs_arc_max), param_set_arc_long, "LU",
283 "max arc size (LEGACY)");
291 SYSCTL_NODE(_vfs_zfs, OID_AUTO, zfetch, CTLFLAG_RW, 0, "ZFS ZFETCH (LEGACY)");
293 /* max bytes to prefetch per stream (default 8MB) */
294 extern uint32_t zfetch_max_distance;
295 SYSCTL_UINT(_vfs_zfs_zfetch, OID_AUTO, max_distance, CTLFLAG_RWTUN,
296 &zfetch_max_distance, 0, "Max bytes to prefetch per stream (LEGACY)");
298 /* max bytes to prefetch indirects for per stream (default 64MB) */
299 extern uint32_t zfetch_max_idistance;
300 SYSCTL_UINT(_vfs_zfs_prefetch, OID_AUTO, max_idistance, CTLFLAG_RWTUN,
301 &zfetch_max_idistance, 0, "Max bytes to prefetch indirects for per stream");
306 extern int zfs_default_bs;
307 SYSCTL_INT(_vfs_zfs, OID_AUTO, default_bs, CTLFLAG_RWTUN,
308 &zfs_default_bs, 0, "Default dnode block shift");
310 extern int zfs_default_ibs;
311 SYSCTL_INT(_vfs_zfs, OID_AUTO, default_ibs, CTLFLAG_RWTUN,
312 &zfs_default_ibs, 0, "Default dnode indirect block shift");
320 * In pools where the log space map feature is not enabled we touch
321 * multiple metaslabs (and their respective space maps) with each
322 * transaction group. Thus, we benefit from having a small space map
323 * block size since it allows us to issue more I/O operations scattered
324 * around the disk. So a sane default for the space map block size
327 extern int zfs_metaslab_sm_blksz_no_log;
328 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_no_log, CTLFLAG_RDTUN,
329 &zfs_metaslab_sm_blksz_no_log, 0,
330 "Block size for space map in pools with log space map disabled. "
331 "Power of 2 and greater than 4096.");
334 * When the log space map feature is enabled, we accumulate a lot of
335 * changes per metaslab that are flushed once in a while so we benefit
336 * from a bigger block size like 128K for the metaslab space maps.
338 extern int zfs_metaslab_sm_blksz_with_log;
339 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, sm_blksz_with_log, CTLFLAG_RDTUN,
340 &zfs_metaslab_sm_blksz_with_log, 0,
341 "Block size for space map in pools with log space map enabled. "
342 "Power of 2 and greater than 4096.");
345 * The in-core space map representation is more compact than its on-disk form.
346 * The zfs_condense_pct determines how much more compact the in-core
347 * space map representation must be before we compact it on-disk.
348 * Values should be greater than or equal to 100.
350 extern int zfs_condense_pct;
351 SYSCTL_INT(_vfs_zfs, OID_AUTO, condense_pct, CTLFLAG_RWTUN,
352 &zfs_condense_pct, 0,
353 "Condense on-disk spacemap when it is more than this many percents"
354 " of in-memory counterpart");
356 extern int zfs_remove_max_segment;
357 SYSCTL_INT(_vfs_zfs, OID_AUTO, remove_max_segment, CTLFLAG_RWTUN,
358 &zfs_remove_max_segment, 0, "Largest contiguous segment ZFS will attempt to"
359 " allocate when removing a device");
361 extern int zfs_removal_suspend_progress;
362 SYSCTL_INT(_vfs_zfs, OID_AUTO, removal_suspend_progress, CTLFLAG_RWTUN,
363 &zfs_removal_suspend_progress, 0, "Ensures certain actions can happen while"
364 " in the middle of a removal");
368 * Minimum size which forces the dynamic allocator to change
369 * it's allocation strategy. Once the space map cannot satisfy
370 * an allocation of this size then it switches to using more
371 * aggressive strategy (i.e search by size rather than offset).
373 extern uint64_t metaslab_df_alloc_threshold;
374 SYSCTL_QUAD(_vfs_zfs_metaslab, OID_AUTO, df_alloc_threshold, CTLFLAG_RWTUN,
375 &metaslab_df_alloc_threshold, 0,
376 "Minimum size which forces the dynamic allocator to change it's allocation strategy");
379 * The minimum free space, in percent, which must be available
380 * in a space map to continue allocations in a first-fit fashion.
381 * Once the space map's free space drops below this level we dynamically
382 * switch to using best-fit allocations.
384 extern int metaslab_df_free_pct;
385 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, df_free_pct, CTLFLAG_RWTUN,
386 &metaslab_df_free_pct, 0,
387 "The minimum free space, in percent, which must be available in a "
388 "space map to continue allocations in a first-fit fashion");
391 * Percentage of all cpus that can be used by the metaslab taskq.
393 extern int metaslab_load_pct;
394 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, load_pct, CTLFLAG_RWTUN,
395 &metaslab_load_pct, 0,
396 "Percentage of cpus that can be used by the metaslab taskq");
399 * Max number of metaslabs per group to preload.
401 extern int metaslab_preload_limit;
402 SYSCTL_INT(_vfs_zfs_metaslab, OID_AUTO, preload_limit, CTLFLAG_RWTUN,
403 &metaslab_preload_limit, 0,
404 "Max number of metaslabs per group to preload");
407 extern int reference_tracking_enable;
408 SYSCTL_INT(_vfs_zfs, OID_AUTO, reference_tracking_enable, CTLFLAG_RDTUN,
409 &reference_tracking_enable, 0,
410 "Track reference holders to refcount_t objects, used mostly by ZFS");
413 extern int zfs_ccw_retry_interval;
414 SYSCTL_INT(_vfs_zfs, OID_AUTO, ccw_retry_interval, CTLFLAG_RWTUN,
415 &zfs_ccw_retry_interval, 0,
416 "Configuration cache file write, retry after failure, interval (seconds)");
418 extern uint64_t zfs_max_missing_tvds_cachefile;
419 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, max_missing_tvds_cachefile, CTLFLAG_RWTUN,
420 &zfs_max_missing_tvds_cachefile, 0,
421 "allow importing pools with missing top-level vdevs in cache file");
423 extern uint64_t zfs_max_missing_tvds_scan;
424 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, max_missing_tvds_scan, CTLFLAG_RWTUN,
425 &zfs_max_missing_tvds_scan, 0,
426 "allow importing pools with missing top-level vdevs during scan");
429 extern int zfs_flags;
431 sysctl_vfs_zfs_debug_flags(SYSCTL_HANDLER_ARGS)
436 err = sysctl_handle_int(oidp, &val, 0, req);
437 if (err != 0 || req->newptr == NULL)
441 * ZFS_DEBUG_MODIFY must be enabled prior to boot so all
442 * arc buffers in the system have the necessary additional
443 * checksum data. However, it is safe to disable at any
446 if (!(zfs_flags & ZFS_DEBUG_MODIFY))
447 val &= ~ZFS_DEBUG_MODIFY;
453 SYSCTL_PROC(_vfs_zfs, OID_AUTO, debugflags,
454 CTLTYPE_UINT | CTLFLAG_MPSAFE | CTLFLAG_RWTUN, NULL, 0,
455 sysctl_vfs_zfs_debug_flags, "IU", "Debug flags for ZFS testing.");
458 param_set_deadman_synctime(SYSCTL_HANDLER_ARGS)
463 val = zfs_deadman_synctime_ms;
464 err = sysctl_handle_long(oidp, &val, 0, req);
465 if (err != 0 || req->newptr == NULL)
467 zfs_deadman_synctime_ms = val;
469 spa_set_deadman_synctime(MSEC2NSEC(zfs_deadman_synctime_ms));
475 param_set_deadman_ziotime(SYSCTL_HANDLER_ARGS)
480 val = zfs_deadman_ziotime_ms;
481 err = sysctl_handle_long(oidp, &val, 0, req);
482 if (err != 0 || req->newptr == NULL)
484 zfs_deadman_ziotime_ms = val;
486 spa_set_deadman_ziotime(MSEC2NSEC(zfs_deadman_synctime_ms));
492 param_set_deadman_failmode(SYSCTL_HANDLER_ARGS)
497 if (req->newptr == NULL)
498 strlcpy(buf, zfs_deadman_failmode, sizeof (buf));
500 rc = sysctl_handle_string(oidp, buf, sizeof (buf), req);
501 if (rc || req->newptr == NULL)
503 if (strcmp(buf, zfs_deadman_failmode) == 0)
505 if (!strcmp(buf, "wait"))
506 zfs_deadman_failmode = "wait";
507 if (!strcmp(buf, "continue"))
508 zfs_deadman_failmode = "continue";
509 if (!strcmp(buf, "panic"))
510 zfs_deadman_failmode = "panic";
512 return (-param_set_deadman_failmode_common(buf));
517 extern int space_map_ibs;
518 SYSCTL_INT(_vfs_zfs, OID_AUTO, space_map_ibs, CTLFLAG_RWTUN,
519 &space_map_ibs, 0, "Space map indirect block shift");
524 param_set_min_auto_ashift(SYSCTL_HANDLER_ARGS)
529 val = zfs_vdev_min_auto_ashift;
530 err = sysctl_handle_64(oidp, &val, 0, req);
531 if (err != 0 || req->newptr == NULL)
532 return (SET_ERROR(err));
534 if (val < ASHIFT_MIN || val > zfs_vdev_max_auto_ashift)
535 return (SET_ERROR(EINVAL));
537 zfs_vdev_min_auto_ashift = val;
543 param_set_max_auto_ashift(SYSCTL_HANDLER_ARGS)
548 val = zfs_vdev_max_auto_ashift;
549 err = sysctl_handle_64(oidp, &val, 0, req);
550 if (err != 0 || req->newptr == NULL)
551 return (SET_ERROR(err));
553 if (val > ASHIFT_MAX || val < zfs_vdev_min_auto_ashift)
554 return (SET_ERROR(EINVAL));
556 zfs_vdev_max_auto_ashift = val;
561 SYSCTL_PROC(_vfs_zfs, OID_AUTO, min_auto_ashift, CTLTYPE_U64 | CTLFLAG_RWTUN,
562 &zfs_vdev_min_auto_ashift, sizeof (zfs_vdev_min_auto_ashift),
563 param_set_min_auto_ashift, "QU",
564 "Min ashift used when creating new top-level vdev. (LEGACY)");
565 SYSCTL_PROC(_vfs_zfs, OID_AUTO, max_auto_ashift, CTLTYPE_U64 | CTLFLAG_RWTUN,
566 &zfs_vdev_max_auto_ashift, sizeof (zfs_vdev_max_auto_ashift),
567 param_set_max_auto_ashift, "QU",
568 "Max ashift used when optimizing for logical -> physical sector size on "
569 "new top-level vdevs. (LEGACY)");
572 * Since the DTL space map of a vdev is not expected to have a lot of
573 * entries, we default its block size to 4K.
575 extern int zfs_vdev_dtl_sm_blksz;
576 SYSCTL_INT(_vfs_zfs, OID_AUTO, dtl_sm_blksz, CTLFLAG_RDTUN,
577 &zfs_vdev_dtl_sm_blksz, 0,
578 "Block size for DTL space map. Power of 2 and greater than 4096.");
581 * vdev-wide space maps that have lots of entries written to them at
582 * the end of each transaction can benefit from a higher I/O bandwidth
583 * (e.g. vdev_obsolete_sm), thus we default their block size to 128K.
585 extern int zfs_vdev_standard_sm_blksz;
586 SYSCTL_INT(_vfs_zfs, OID_AUTO, standard_sm_blksz, CTLFLAG_RDTUN,
587 &zfs_vdev_standard_sm_blksz, 0,
588 "Block size for standard space map. Power of 2 and greater than 4096.");
590 extern int vdev_validate_skip;
591 SYSCTL_INT(_vfs_zfs, OID_AUTO, validate_skip, CTLFLAG_RDTUN,
592 &vdev_validate_skip, 0,
593 "Enable to bypass vdev_validate().");
600 * The load configuration settings below are tuned by default for
601 * the case where all devices are of the same rotational type.
603 * If there is a mixture of rotating and non-rotating media, setting
604 * non_rotating_seek_inc to 0 may well provide better results as it
605 * will direct more reads to the non-rotating vdevs which are more
606 * likely to have a higher performance.
611 #define ZFS_VDEV_QUEUE_KNOB_MIN(name) \
612 extern uint32_t zfs_vdev_ ## name ## _min_active; \
613 SYSCTL_UINT(_vfs_zfs_vdev, OID_AUTO, name ## _min_active, CTLFLAG_RWTUN,\
614 &zfs_vdev_ ## name ## _min_active, 0, \
615 "Initial number of I/O requests of type " #name \
616 " active for each device");
618 #define ZFS_VDEV_QUEUE_KNOB_MAX(name) \
619 extern uint32_t zfs_vdev_ ## name ## _max_active; \
620 SYSCTL_UINT(_vfs_zfs_vdev, OID_AUTO, name ## _max_active, CTLFLAG_RWTUN, \
621 &zfs_vdev_ ## name ## _max_active, 0, \
622 "Maximum number of I/O requests of type " #name \
623 " active for each device");
626 #undef ZFS_VDEV_QUEUE_KNOB
628 extern uint32_t zfs_vdev_max_active;
629 SYSCTL_UINT(_vfs_zfs, OID_AUTO, top_maxinflight, CTLFLAG_RWTUN,
630 &zfs_vdev_max_active, 0,
631 "The maximum number of I/Os of all types active for each device. (LEGACY)");
633 extern int zfs_vdev_def_queue_depth;
634 SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, def_queue_depth, CTLFLAG_RWTUN,
635 &zfs_vdev_def_queue_depth, 0,
636 "Default queue depth for each allocator");
638 /*extern uint64_t zfs_multihost_history;
639 SYSCTL_UQUAD(_vfs_zfs, OID_AUTO, multihost_history, CTLFLAG_RWTUN,
640 &zfs_multihost_history, 0,
641 "Historical staticists for the last N multihost updates");*/
644 SYSCTL_INT(_vfs_zfs_vdev, OID_AUTO, trim_on_init, CTLFLAG_RW,
645 &vdev_trim_on_init, 0, "Enable/disable full vdev trim on initialisation");
650 #if defined(__LP64__)
656 SYSCTL_INT(_vfs_zfs_zio, OID_AUTO, use_uma, CTLFLAG_RDTUN, &zio_use_uma, 0,
657 "Use uma(9) for ZIO allocations");
658 SYSCTL_INT(_vfs_zfs_zio, OID_AUTO, exclude_metadata, CTLFLAG_RDTUN, &zio_exclude_metadata, 0,
659 "Exclude metadata buffers from dumps as well");
662 param_set_slop_shift(SYSCTL_HANDLER_ARGS)
669 err = sysctl_handle_int(oidp, &val, 0, req);
670 if (err != 0 || req->newptr == NULL)
673 if (val < 1 || val > 31)
682 param_set_multihost_interval(SYSCTL_HANDLER_ARGS)
686 err = sysctl_handle_long(oidp, arg1, 0, req);
687 if (err != 0 || req->newptr == NULL)
690 if (spa_mode_global != SPA_MODE_UNINIT)
691 mmp_signal_all_threads();