]> git.proxmox.com Git - mirror_zfs.git/blame - module/zfs/dsl_dataset.c
projects / mirror_zfs.git / blame
? search:
summary | shortlog | log | commit | commitdiff | tree
blob | blame (incremental) | history | HEAD
Fix kstat xuio
[mirror_zfs.git] / module / zfs / dsl_dataset.c
CommitLineData
34dc7c2f
BB		 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/*
428870ff 22 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved.
34dc7c2f
BB		 23 */
24
34dc7c2f
BB		 25#include <sys/dmu_objset.h>
26#include <sys/dsl_dataset.h>
27#include <sys/dsl_dir.h>
28#include <sys/dsl_prop.h>
29#include <sys/dsl_synctask.h>
30#include <sys/dmu_traverse.h>
31#include <sys/dmu_tx.h>
32#include <sys/arc.h>
33#include <sys/zio.h>
34#include <sys/zap.h>
35#include <sys/unique.h>
36#include <sys/zfs_context.h>
37#include <sys/zfs_ioctl.h>
38#include <sys/spa.h>
b128c09f 39#include <sys/zfs_znode.h>
572e2857 40#include <sys/zfs_onexit.h>
45d1cae3 41#include <sys/zvol.h>
428870ff
BB		 42#include <sys/dsl_scan.h>
43#include <sys/dsl_deadlist.h>
44
b128c09f
BB		 45static char *dsl_reaper = "the grim reaper";
46
34dc7c2f
BB		 47static dsl_checkfunc_t dsl_dataset_destroy_begin_check;
48static dsl_syncfunc_t dsl_dataset_destroy_begin_sync;
34dc7c2f
BB		 49static dsl_syncfunc_t dsl_dataset_set_reservation_sync;
50
428870ff
BB		 51#define SWITCH64(x, y) \
52 { \
53 uint64_t __tmp = (x); \
54 (x) = (y); \
55 (y) = __tmp; \
56 }
57
34dc7c2f
BB		 58#define DS_REF_MAX (1ULL << 62)
59
60#define DSL_DEADLIST_BLOCKSIZE SPA_MAXBLOCKSIZE
61
b128c09f
BB		 62#define DSL_DATASET_IS_DESTROYED(ds) ((ds)->ds_owner == dsl_reaper)
63
34dc7c2f
BB		 64
65/*
66 * Figure out how much of this delta should be propogated to the dsl_dir
67 * layer. If there's a refreservation, that space has already been
68 * partially accounted for in our ancestors.
69 */
70static int64_t
71parent_delta(dsl_dataset_t *ds, int64_t delta)
72{
73 uint64_t old_bytes, new_bytes;
74
75 if (ds->ds_reserved == 0)
76 return (delta);
77
78 old_bytes = MAX(ds->ds_phys->ds_unique_bytes, ds->ds_reserved);
79 new_bytes = MAX(ds->ds_phys->ds_unique_bytes + delta, ds->ds_reserved);
80
81 ASSERT3U(ABS((int64_t)(new_bytes - old_bytes)), <=, ABS(delta));
82 return (new_bytes - old_bytes);
83}
84
85void
428870ff 86dsl_dataset_block_born(dsl_dataset_t *ds, const blkptr_t *bp, dmu_tx_t *tx)
34dc7c2f 87{
d6320ddb 88 int used, compressed, uncompressed;
34dc7c2f
BB		 89 int64_t delta;
90
d6320ddb
BB		 91 used = bp_get_dsize_sync(tx->tx_pool->dp_spa, bp);
92 compressed = BP_GET_PSIZE(bp);
93 uncompressed = BP_GET_UCSIZE(bp);
94
428870ff 95 dprintf_bp(bp, "ds=%p", ds);
34dc7c2f
BB		 96
97 ASSERT(dmu_tx_is_syncing(tx));
98 /* It could have been compressed away to nothing */
99 if (BP_IS_HOLE(bp))
100 return;
101 ASSERT(BP_GET_TYPE(bp) != DMU_OT_NONE);
102 ASSERT3U(BP_GET_TYPE(bp), <, DMU_OT_NUMTYPES);
103 if (ds == NULL) {
104 /*
105 * Account for the meta-objset space in its placeholder
106 * dsl_dir.
107 */
108 ASSERT3U(compressed, ==, uncompressed); /* it's all metadata */
b128c09f 109 dsl_dir_diduse_space(tx->tx_pool->dp_mos_dir, DD_USED_HEAD,
34dc7c2f
BB		 110 used, compressed, uncompressed, tx);
111 dsl_dir_dirty(tx->tx_pool->dp_mos_dir, tx);
112 return;
113 }
114 dmu_buf_will_dirty(ds->ds_dbuf, tx);
428870ff 115
b128c09f 116 mutex_enter(&ds->ds_dir->dd_lock);
34dc7c2f
BB		 117 mutex_enter(&ds->ds_lock);
118 delta = parent_delta(ds, used);
119 ds->ds_phys->ds_used_bytes += used;
120 ds->ds_phys->ds_compressed_bytes += compressed;
121 ds->ds_phys->ds_uncompressed_bytes += uncompressed;
122 ds->ds_phys->ds_unique_bytes += used;
123 mutex_exit(&ds->ds_lock);
b128c09f
BB		 124 dsl_dir_diduse_space(ds->ds_dir, DD_USED_HEAD, delta,
125 compressed, uncompressed, tx);
126 dsl_dir_transfer_space(ds->ds_dir, used - delta,
127 DD_USED_REFRSRV, DD_USED_HEAD, tx);
128 mutex_exit(&ds->ds_dir->dd_lock);
34dc7c2f
BB		 129}
130
b128c09f 131int
428870ff
BB		 132dsl_dataset_block_kill(dsl_dataset_t *ds, const blkptr_t *bp, dmu_tx_t *tx,
133 boolean_t async)
34dc7c2f 134{
d6320ddb
BB		 135 int used, compressed, uncompressed;
136
34dc7c2f 137 if (BP_IS_HOLE(bp))
b128c09f 138 return (0);
34dc7c2f 139
428870ff
BB		 140 ASSERT(dmu_tx_is_syncing(tx));
141 ASSERT(bp->blk_birth <= tx->tx_txg);
142
d6320ddb
BB		 143 used = bp_get_dsize_sync(tx->tx_pool->dp_spa, bp);
144 compressed = BP_GET_PSIZE(bp);
145 uncompressed = BP_GET_UCSIZE(bp);
428870ff 146
34dc7c2f
BB		 147 ASSERT(used > 0);
148 if (ds == NULL) {
34dc7c2f
BB		 149 /*
150 * Account for the meta-objset space in its placeholder
151 * dataset.
152 */
428870ff 153 dsl_free(tx->tx_pool, tx->tx_txg, bp);
34dc7c2f 154
b128c09f 155 dsl_dir_diduse_space(tx->tx_pool->dp_mos_dir, DD_USED_HEAD,
34dc7c2f
BB		 156 -used, -compressed, -uncompressed, tx);
157 dsl_dir_dirty(tx->tx_pool->dp_mos_dir, tx);
b128c09f 158 return (used);
34dc7c2f
BB		 159 }
160 ASSERT3P(tx->tx_pool, ==, ds->ds_dir->dd_pool);
161
b128c09f 162 ASSERT(!dsl_dataset_is_snapshot(ds));
34dc7c2f
BB		 163 dmu_buf_will_dirty(ds->ds_dbuf, tx);
164
165 if (bp->blk_birth > ds->ds_phys->ds_prev_snap_txg) {
34dc7c2f
BB		 166 int64_t delta;
167
428870ff
BB		 168 dprintf_bp(bp, "freeing ds=%llu", ds->ds_object);
169 dsl_free(tx->tx_pool, tx->tx_txg, bp);
34dc7c2f 170
b128c09f 171 mutex_enter(&ds->ds_dir->dd_lock);
34dc7c2f
BB		 172 mutex_enter(&ds->ds_lock);
173 ASSERT(ds->ds_phys->ds_unique_bytes >= used ||
174 !DS_UNIQUE_IS_ACCURATE(ds));
175 delta = parent_delta(ds, -used);
176 ds->ds_phys->ds_unique_bytes -= used;
177 mutex_exit(&ds->ds_lock);
b128c09f 178 dsl_dir_diduse_space(ds->ds_dir, DD_USED_HEAD,
34dc7c2f 179 delta, -compressed, -uncompressed, tx);
b128c09f
BB		 180 dsl_dir_transfer_space(ds->ds_dir, -used - delta,
181 DD_USED_REFRSRV, DD_USED_HEAD, tx);
182 mutex_exit(&ds->ds_dir->dd_lock);
34dc7c2f
BB		 183 } else {
184 dprintf_bp(bp, "putting on dead list: %s", "");
428870ff
BB		 185 if (async) {
186 /*
187 * We are here as part of zio's write done callback,
188 * which means we're a zio interrupt thread. We can't
189 * call dsl_deadlist_insert() now because it may block
190 * waiting for I/O. Instead, put bp on the deferred
191 * queue and let dsl_pool_sync() finish the job.
192 */
193 bplist_append(&ds->ds_pending_deadlist, bp);
194 } else {
195 dsl_deadlist_insert(&ds->ds_deadlist, bp, tx);
196 }
34dc7c2f
BB		 197 ASSERT3U(ds->ds_prev->ds_object, ==,
198 ds->ds_phys->ds_prev_snap_obj);
199 ASSERT(ds->ds_prev->ds_phys->ds_num_children > 0);
200 /* if (bp->blk_birth > prev prev snap txg) prev unique += bs */
201 if (ds->ds_prev->ds_phys->ds_next_snap_obj ==
202 ds->ds_object && bp->blk_birth >
203 ds->ds_prev->ds_phys->ds_prev_snap_txg) {
204 dmu_buf_will_dirty(ds->ds_prev->ds_dbuf, tx);
205 mutex_enter(&ds->ds_prev->ds_lock);
206 ds->ds_prev->ds_phys->ds_unique_bytes += used;
207 mutex_exit(&ds->ds_prev->ds_lock);
208 }
428870ff 209 if (bp->blk_birth > ds->ds_dir->dd_origin_txg) {
b128c09f
BB		 210 dsl_dir_transfer_space(ds->ds_dir, used,
211 DD_USED_HEAD, DD_USED_SNAP, tx);
212 }
34dc7c2f
BB		 213 }
214 mutex_enter(&ds->ds_lock);
215 ASSERT3U(ds->ds_phys->ds_used_bytes, >=, used);
216 ds->ds_phys->ds_used_bytes -= used;
217 ASSERT3U(ds->ds_phys->ds_compressed_bytes, >=, compressed);
218 ds->ds_phys->ds_compressed_bytes -= compressed;
219 ASSERT3U(ds->ds_phys->ds_uncompressed_bytes, >=, uncompressed);
220 ds->ds_phys->ds_uncompressed_bytes -= uncompressed;
221 mutex_exit(&ds->ds_lock);
b128c09f
BB		 222
223 return (used);
34dc7c2f
BB		 224}
225
226uint64_t
227dsl_dataset_prev_snap_txg(dsl_dataset_t *ds)
228{
229 uint64_t trysnap = 0;
230
231 if (ds == NULL)
232 return (0);
233 /*
234 * The snapshot creation could fail, but that would cause an
235 * incorrect FALSE return, which would only result in an
236 * overestimation of the amount of space that an operation would
237 * consume, which is OK.
238 *
239 * There's also a small window where we could miss a pending
240 * snapshot, because we could set the sync task in the quiescing
241 * phase. So this should only be used as a guess.
242 */
243 if (ds->ds_trysnap_txg >
244 spa_last_synced_txg(ds->ds_dir->dd_pool->dp_spa))
245 trysnap = ds->ds_trysnap_txg;
246 return (MAX(ds->ds_phys->ds_prev_snap_txg, trysnap));
247}
248
9babb374 249boolean_t
428870ff
BB		 250dsl_dataset_block_freeable(dsl_dataset_t *ds, const blkptr_t *bp,
251 uint64_t blk_birth)
34dc7c2f 252{
428870ff
BB		 253 if (blk_birth <= dsl_dataset_prev_snap_txg(ds))
254 return (B_FALSE);
255
572e2857 256 ddt_prefetch(dsl_dataset_get_spa(ds), bp);
428870ff
BB		 257
258 return (B_TRUE);
34dc7c2f
BB		 259}
260
261/* ARGSUSED */
262static void
263dsl_dataset_evict(dmu_buf_t *db, void *dsv)
264{
265 dsl_dataset_t *ds = dsv;
266
b128c09f 267 ASSERT(ds->ds_owner == NULL || DSL_DATASET_IS_DESTROYED(ds));
34dc7c2f 268
34dc7c2f
BB		 269 unique_remove(ds->ds_fsid_guid);
270
428870ff
BB		 271 if (ds->ds_objset != NULL)
272 dmu_objset_evict(ds->ds_objset);
34dc7c2f
BB		 273
274 if (ds->ds_prev) {
b128c09f 275 dsl_dataset_drop_ref(ds->ds_prev, ds);
34dc7c2f
BB		 276 ds->ds_prev = NULL;
277 }
278
428870ff
BB		 279 bplist_destroy(&ds->ds_pending_deadlist);
280 if (db != NULL) {
281 dsl_deadlist_close(&ds->ds_deadlist);
282 } else {
283 ASSERT(ds->ds_deadlist.dl_dbuf == NULL);
284 ASSERT(!ds->ds_deadlist.dl_oldfmt);
285 }
b128c09f
BB		 286 if (ds->ds_dir)
287 dsl_dir_close(ds->ds_dir, ds);
34dc7c2f
BB		 288
289 ASSERT(!list_link_active(&ds->ds_synced_link));
290
291 mutex_destroy(&ds->ds_lock);
45d1cae3 292 mutex_destroy(&ds->ds_recvlock);
34dc7c2f 293 mutex_destroy(&ds->ds_opening_lock);
b128c09f
BB		 294 rw_destroy(&ds->ds_rwlock);
295 cv_destroy(&ds->ds_exclusive_cv);
34dc7c2f
BB		 296
297 kmem_free(ds, sizeof (dsl_dataset_t));
298}
299
300static int
301dsl_dataset_get_snapname(dsl_dataset_t *ds)
302{
303 dsl_dataset_phys_t *headphys;
304 int err;
305 dmu_buf_t *headdbuf;
306 dsl_pool_t *dp = ds->ds_dir->dd_pool;
307 objset_t *mos = dp->dp_meta_objset;
308
309 if (ds->ds_snapname[0])
310 return (0);
311 if (ds->ds_phys->ds_next_snap_obj == 0)
312 return (0);
313
314 err = dmu_bonus_hold(mos, ds->ds_dir->dd_phys->dd_head_dataset_obj,
315 FTAG, &headdbuf);
316 if (err)
317 return (err);
318 headphys = headdbuf->db_data;
319 err = zap_value_search(dp->dp_meta_objset,
320 headphys->ds_snapnames_zapobj, ds->ds_object, 0, ds->ds_snapname);
321 dmu_buf_rele(headdbuf, FTAG);
322 return (err);
323}
324
325static int
b128c09f 326dsl_dataset_snap_lookup(dsl_dataset_t *ds, const char *name, uint64_t *value)
34dc7c2f 327{
b128c09f
BB		 328 objset_t *mos = ds->ds_dir->dd_pool->dp_meta_objset;
329 uint64_t snapobj = ds->ds_phys->ds_snapnames_zapobj;
34dc7c2f
BB		 330 matchtype_t mt;
331 int err;
332
b128c09f 333 if (ds->ds_phys->ds_flags & DS_FLAG_CI_DATASET)
34dc7c2f
BB		 334 mt = MT_FIRST;
335 else
336 mt = MT_EXACT;
337
b128c09f 338 err = zap_lookup_norm(mos, snapobj, name, 8, 1,
34dc7c2f
BB		 339 value, mt, NULL, 0, NULL);
340 if (err == ENOTSUP && mt == MT_FIRST)
b128c09f 341 err = zap_lookup(mos, snapobj, name, 8, 1, value);
34dc7c2f
BB		 342 return (err);
343}
344
345static int
b128c09f 346dsl_dataset_snap_remove(dsl_dataset_t *ds, char *name, dmu_tx_t *tx)
34dc7c2f 347{
b128c09f
BB		 348 objset_t *mos = ds->ds_dir->dd_pool->dp_meta_objset;
349 uint64_t snapobj = ds->ds_phys->ds_snapnames_zapobj;
34dc7c2f
BB		 350 matchtype_t mt;
351 int err;
352
428870ff
BB		 353 dsl_dir_snap_cmtime_update(ds->ds_dir);
354
b128c09f 355 if (ds->ds_phys->ds_flags & DS_FLAG_CI_DATASET)
34dc7c2f
BB		 356 mt = MT_FIRST;
357 else
358 mt = MT_EXACT;
359
b128c09f 360 err = zap_remove_norm(mos, snapobj, name, mt, tx);
34dc7c2f 361 if (err == ENOTSUP && mt == MT_FIRST)
b128c09f 362 err = zap_remove(mos, snapobj, name, tx);
34dc7c2f
BB		 363 return (err);
364}
365
b128c09f
BB		 366static int
367dsl_dataset_get_ref(dsl_pool_t *dp, uint64_t dsobj, void *tag,
368 dsl_dataset_t **dsp)
34dc7c2f 369{
34dc7c2f
BB		 370 objset_t *mos = dp->dp_meta_objset;
371 dmu_buf_t *dbuf;
372 dsl_dataset_t *ds;
373 int err;
572e2857 374 dmu_object_info_t doi;
34dc7c2f
BB		 375
376 ASSERT(RW_LOCK_HELD(&dp->dp_config_rwlock) ||
377 dsl_pool_sync_context(dp));
378
379 err = dmu_bonus_hold(mos, dsobj, tag, &dbuf);
380 if (err)
381 return (err);
572e2857
BB		 382
383 /* Make sure dsobj has the correct object type. */
384 dmu_object_info_from_db(dbuf, &doi);
385 if (doi.doi_type != DMU_OT_DSL_DATASET)
386 return (EINVAL);
387
34dc7c2f
BB		 388 ds = dmu_buf_get_user(dbuf);
389 if (ds == NULL) {
d4ed6673 390 dsl_dataset_t *winner = NULL;
34dc7c2f
BB		 391
392 ds = kmem_zalloc(sizeof (dsl_dataset_t), KM_SLEEP);
393 ds->ds_dbuf = dbuf;
394 ds->ds_object = dsobj;
395 ds->ds_phys = dbuf->db_data;
396
397 mutex_init(&ds->ds_lock, NULL, MUTEX_DEFAULT, NULL);
45d1cae3 398 mutex_init(&ds->ds_recvlock, NULL, MUTEX_DEFAULT, NULL);
34dc7c2f 399 mutex_init(&ds->ds_opening_lock, NULL, MUTEX_DEFAULT, NULL);
b128c09f
BB		 400 rw_init(&ds->ds_rwlock, 0, 0, 0);
401 cv_init(&ds->ds_exclusive_cv, NULL, CV_DEFAULT, NULL);
34dc7c2f 402
428870ff
BB		 403 bplist_create(&ds->ds_pending_deadlist);
404 dsl_deadlist_open(&ds->ds_deadlist,
34dc7c2f 405 mos, ds->ds_phys->ds_deadlist_obj);
428870ff 406
34dc7c2f
BB		 407 if (err == 0) {
408 err = dsl_dir_open_obj(dp,
409 ds->ds_phys->ds_dir_obj, NULL, ds, &ds->ds_dir);
410 }
411 if (err) {
34dc7c2f 412 mutex_destroy(&ds->ds_lock);
45d1cae3 413 mutex_destroy(&ds->ds_recvlock);
34dc7c2f 414 mutex_destroy(&ds->ds_opening_lock);
b128c09f
BB		 415 rw_destroy(&ds->ds_rwlock);
416 cv_destroy(&ds->ds_exclusive_cv);
428870ff
BB		 417 bplist_destroy(&ds->ds_pending_deadlist);
418 dsl_deadlist_close(&ds->ds_deadlist);
34dc7c2f
BB		 419 kmem_free(ds, sizeof (dsl_dataset_t));
420 dmu_buf_rele(dbuf, tag);
421 return (err);
422 }
423
b128c09f 424 if (!dsl_dataset_is_snapshot(ds)) {
34dc7c2f
BB		 425 ds->ds_snapname[0] = '\0';
426 if (ds->ds_phys->ds_prev_snap_obj) {
b128c09f
BB		 427 err = dsl_dataset_get_ref(dp,
428 ds->ds_phys->ds_prev_snap_obj,
429 ds, &ds->ds_prev);
34dc7c2f 430 }
45d1cae3
BB		 431 } else {
432 if (zfs_flags & ZFS_DEBUG_SNAPNAMES)
433 err = dsl_dataset_get_snapname(ds);
434 if (err == 0 && ds->ds_phys->ds_userrefs_obj != 0) {
435 err = zap_count(
436 ds->ds_dir->dd_pool->dp_meta_objset,
437 ds->ds_phys->ds_userrefs_obj,
438 &ds->ds_userrefs);
439 }
34dc7c2f
BB		 440 }
441
b128c09f 442 if (err == 0 && !dsl_dataset_is_snapshot(ds)) {
34dc7c2f
BB		 443 /*
444 * In sync context, we're called with either no lock
445 * or with the write lock. If we're not syncing,
446 * we're always called with the read lock held.
447 */
448 boolean_t need_lock =
449 !RW_WRITE_HELD(&dp->dp_config_rwlock) &&
450 dsl_pool_sync_context(dp);
451
452 if (need_lock)
453 rw_enter(&dp->dp_config_rwlock, RW_READER);
454
b128c09f 455 err = dsl_prop_get_ds(ds,
34dc7c2f
BB		 456 "refreservation", sizeof (uint64_t), 1,
457 &ds->ds_reserved, NULL);
458 if (err == 0) {
b128c09f 459 err = dsl_prop_get_ds(ds,
34dc7c2f
BB		 460 "refquota", sizeof (uint64_t), 1,
461 &ds->ds_quota, NULL);
462 }
463
464 if (need_lock)
465 rw_exit(&dp->dp_config_rwlock);
466 } else {
467 ds->ds_reserved = ds->ds_quota = 0;
468 }
469
470 if (err == 0) {
471 winner = dmu_buf_set_user_ie(dbuf, ds, &ds->ds_phys,
472 dsl_dataset_evict);
473 }
474 if (err || winner) {
428870ff
BB		 475 bplist_destroy(&ds->ds_pending_deadlist);
476 dsl_deadlist_close(&ds->ds_deadlist);
b128c09f
BB		 477 if (ds->ds_prev)
478 dsl_dataset_drop_ref(ds->ds_prev, ds);
34dc7c2f
BB		 479 dsl_dir_close(ds->ds_dir, ds);
480 mutex_destroy(&ds->ds_lock);
45d1cae3 481 mutex_destroy(&ds->ds_recvlock);
34dc7c2f 482 mutex_destroy(&ds->ds_opening_lock);
b128c09f
BB		 483 rw_destroy(&ds->ds_rwlock);
484 cv_destroy(&ds->ds_exclusive_cv);
34dc7c2f
BB		 485 kmem_free(ds, sizeof (dsl_dataset_t));
486 if (err) {
487 dmu_buf_rele(dbuf, tag);
488 return (err);
489 }
490 ds = winner;
491 } else {
492 ds->ds_fsid_guid =
493 unique_insert(ds->ds_phys->ds_fsid_guid);
494 }
495 }
496 ASSERT3P(ds->ds_dbuf, ==, dbuf);
497 ASSERT3P(ds->ds_phys, ==, dbuf->db_data);
b128c09f
BB		 498 ASSERT(ds->ds_phys->ds_prev_snap_obj != 0 ||
499 spa_version(dp->dp_spa) < SPA_VERSION_ORIGIN ||
500 dp->dp_origin_snap == NULL || ds == dp->dp_origin_snap);
34dc7c2f 501 mutex_enter(&ds->ds_lock);
b128c09f 502 if (!dsl_pool_sync_context(dp) && DSL_DATASET_IS_DESTROYED(ds)) {
34dc7c2f 503 mutex_exit(&ds->ds_lock);
b128c09f
BB		 504 dmu_buf_rele(ds->ds_dbuf, tag);
505 return (ENOENT);
34dc7c2f 506 }
34dc7c2f 507 mutex_exit(&ds->ds_lock);
34dc7c2f
BB		 508 *dsp = ds;
509 return (0);
510}
511
b128c09f
BB		 512static int
513dsl_dataset_hold_ref(dsl_dataset_t *ds, void *tag)
514{
515 dsl_pool_t *dp = ds->ds_dir->dd_pool;
516
517 /*
518 * In syncing context we don't want the rwlock lock: there
519 * may be an existing writer waiting for sync phase to
520 * finish. We don't need to worry about such writers, since
521 * sync phase is single-threaded, so the writer can't be
522 * doing anything while we are active.
523 */
524 if (dsl_pool_sync_context(dp)) {
525 ASSERT(!DSL_DATASET_IS_DESTROYED(ds));
526 return (0);
527 }
528
529 /*
530 * Normal users will hold the ds_rwlock as a READER until they
531 * are finished (i.e., call dsl_dataset_rele()). "Owners" will
532 * drop their READER lock after they set the ds_owner field.
533 *
534 * If the dataset is being destroyed, the destroy thread will
535 * obtain a WRITER lock for exclusive access after it's done its
536 * open-context work and then change the ds_owner to
537 * dsl_reaper once destruction is assured. So threads
538 * may block here temporarily, until the "destructability" of
539 * the dataset is determined.
540 */
541 ASSERT(!RW_WRITE_HELD(&dp->dp_config_rwlock));
542 mutex_enter(&ds->ds_lock);
543 while (!rw_tryenter(&ds->ds_rwlock, RW_READER)) {
544 rw_exit(&dp->dp_config_rwlock);
545 cv_wait(&ds->ds_exclusive_cv, &ds->ds_lock);
546 if (DSL_DATASET_IS_DESTROYED(ds)) {
547 mutex_exit(&ds->ds_lock);
548 dsl_dataset_drop_ref(ds, tag);
549 rw_enter(&dp->dp_config_rwlock, RW_READER);
550 return (ENOENT);
551 }
428870ff
BB		 552 /*
553 * The dp_config_rwlock lives above the ds_lock. And
554 * we need to check DSL_DATASET_IS_DESTROYED() while
555 * holding the ds_lock, so we have to drop and reacquire
556 * the ds_lock here.
557 */
558 mutex_exit(&ds->ds_lock);
b128c09f 559 rw_enter(&dp->dp_config_rwlock, RW_READER);
428870ff 560 mutex_enter(&ds->ds_lock);
b128c09f
BB		 561 }
562 mutex_exit(&ds->ds_lock);
563 return (0);
564}
565
34dc7c2f 566int
b128c09f
BB		 567dsl_dataset_hold_obj(dsl_pool_t *dp, uint64_t dsobj, void *tag,
568 dsl_dataset_t **dsp)
569{
570 int err = dsl_dataset_get_ref(dp, dsobj, tag, dsp);
571
572 if (err)
573 return (err);
574 return (dsl_dataset_hold_ref(*dsp, tag));
575}
576
577int
428870ff
BB		 578dsl_dataset_own_obj(dsl_pool_t *dp, uint64_t dsobj, boolean_t inconsistentok,
579 void *tag, dsl_dataset_t **dsp)
b128c09f 580{
428870ff 581 int err = dsl_dataset_hold_obj(dp, dsobj, tag, dsp);
b128c09f
BB		 582 if (err)
583 return (err);
428870ff
BB		 584 if (!dsl_dataset_tryown(*dsp, inconsistentok, tag)) {
585 dsl_dataset_rele(*dsp, tag);
9babb374 586 *dsp = NULL;
b128c09f
BB		 587 return (EBUSY);
588 }
589 return (0);
590}
591
592int
593dsl_dataset_hold(const char *name, void *tag, dsl_dataset_t **dsp)
34dc7c2f
BB		 594{
595 dsl_dir_t *dd;
596 dsl_pool_t *dp;
b128c09f 597 const char *snapname;
34dc7c2f 598 uint64_t obj;
34dc7c2f
BB		 599 int err = 0;
600
b128c09f 601 err = dsl_dir_open_spa(NULL, name, FTAG, &dd, &snapname);
34dc7c2f
BB		 602 if (err)
603 return (err);
604
605 dp = dd->dd_pool;
606 obj = dd->dd_phys->dd_head_dataset_obj;
607 rw_enter(&dp->dp_config_rwlock, RW_READER);
b128c09f
BB		 608 if (obj)
609 err = dsl_dataset_get_ref(dp, obj, tag, dsp);
610 else
34dc7c2f 611 err = ENOENT;
b128c09f 612 if (err)
34dc7c2f 613 goto out;
34dc7c2f 614
b128c09f 615 err = dsl_dataset_hold_ref(*dsp, tag);
34dc7c2f 616
b128c09f
BB		 617 /* we may be looking for a snapshot */
618 if (err == 0 && snapname != NULL) {
619 dsl_dataset_t *ds = NULL;
34dc7c2f 620
b128c09f
BB		 621 if (*snapname++ != '@') {
622 dsl_dataset_rele(*dsp, tag);
34dc7c2f
BB		 623 err = ENOENT;
624 goto out;
625 }
34dc7c2f 626
b128c09f
BB		 627 dprintf("looking for snapshot '%s'\n", snapname);
628 err = dsl_dataset_snap_lookup(*dsp, snapname, &obj);
629 if (err == 0)
630 err = dsl_dataset_get_ref(dp, obj, tag, &ds);
631 dsl_dataset_rele(*dsp, tag);
632
633 ASSERT3U((err == 0), ==, (ds != NULL));
634
635 if (ds) {
636 mutex_enter(&ds->ds_lock);
637 if (ds->ds_snapname[0] == 0)
638 (void) strlcpy(ds->ds_snapname, snapname,
639 sizeof (ds->ds_snapname));
640 mutex_exit(&ds->ds_lock);
641 err = dsl_dataset_hold_ref(ds, tag);
642 *dsp = err ? NULL : ds;
34dc7c2f 643 }
34dc7c2f 644 }
34dc7c2f
BB		 645out:
646 rw_exit(&dp->dp_config_rwlock);
647 dsl_dir_close(dd, FTAG);
34dc7c2f
BB		 648 return (err);
649}
650
651int
428870ff
BB		 652dsl_dataset_own(const char *name, boolean_t inconsistentok,
653 void *tag, dsl_dataset_t **dsp)
34dc7c2f 654{
428870ff 655 int err = dsl_dataset_hold(name, tag, dsp);
b128c09f
BB		 656 if (err)
657 return (err);
428870ff
BB		 658 if (!dsl_dataset_tryown(*dsp, inconsistentok, tag)) {
659 dsl_dataset_rele(*dsp, tag);
b128c09f
BB		 660 return (EBUSY);
661 }
662 return (0);
34dc7c2f
BB		 663}
664
665void
666dsl_dataset_name(dsl_dataset_t *ds, char *name)
667{
668 if (ds == NULL) {
669 (void) strcpy(name, "mos");
670 } else {
671 dsl_dir_name(ds->ds_dir, name);
672 VERIFY(0 == dsl_dataset_get_snapname(ds));
673 if (ds->ds_snapname[0]) {
674 (void) strcat(name, "@");
b128c09f
BB		 675 /*
676 * We use a "recursive" mutex so that we
677 * can call dprintf_ds() with ds_lock held.
678 */
34dc7c2f 679 if (!MUTEX_HELD(&ds->ds_lock)) {
34dc7c2f
BB		 680 mutex_enter(&ds->ds_lock);
681 (void) strcat(name, ds->ds_snapname);
682 mutex_exit(&ds->ds_lock);
683 } else {
684 (void) strcat(name, ds->ds_snapname);
685 }
686 }
687 }
688}
689
690static int
691dsl_dataset_namelen(dsl_dataset_t *ds)
692{
693 int result;
694
695 if (ds == NULL) {
696 result = 3; /* "mos" */
697 } else {
698 result = dsl_dir_namelen(ds->ds_dir);
699 VERIFY(0 == dsl_dataset_get_snapname(ds));
700 if (ds->ds_snapname[0]) {
701 ++result; /* adding one for the @-sign */
702 if (!MUTEX_HELD(&ds->ds_lock)) {
34dc7c2f
BB		 703 mutex_enter(&ds->ds_lock);
704 result += strlen(ds->ds_snapname);
705 mutex_exit(&ds->ds_lock);
706 } else {
707 result += strlen(ds->ds_snapname);
708 }
709 }
710 }
711
712 return (result);
713}
714
715void
b128c09f 716dsl_dataset_drop_ref(dsl_dataset_t *ds, void *tag)
34dc7c2f 717{
34dc7c2f
BB		 718 dmu_buf_rele(ds->ds_dbuf, tag);
719}
720
721void
b128c09f
BB		 722dsl_dataset_rele(dsl_dataset_t *ds, void *tag)
723{
724 if (!dsl_pool_sync_context(ds->ds_dir->dd_pool)) {
725 rw_exit(&ds->ds_rwlock);
726 }
727 dsl_dataset_drop_ref(ds, tag);
728}
729
730void
428870ff 731dsl_dataset_disown(dsl_dataset_t *ds, void *tag)
34dc7c2f 732{
428870ff 733 ASSERT((ds->ds_owner == tag && ds->ds_dbuf) ||
b128c09f
BB		 734 (DSL_DATASET_IS_DESTROYED(ds) && ds->ds_dbuf == NULL));
735
34dc7c2f 736 mutex_enter(&ds->ds_lock);
b128c09f
BB		 737 ds->ds_owner = NULL;
738 if (RW_WRITE_HELD(&ds->ds_rwlock)) {
739 rw_exit(&ds->ds_rwlock);
740 cv_broadcast(&ds->ds_exclusive_cv);
741 }
34dc7c2f 742 mutex_exit(&ds->ds_lock);
b128c09f 743 if (ds->ds_dbuf)
428870ff 744 dsl_dataset_drop_ref(ds, tag);
b128c09f 745 else
428870ff 746 dsl_dataset_evict(NULL, ds);
34dc7c2f
BB		 747}
748
749boolean_t
428870ff 750dsl_dataset_tryown(dsl_dataset_t *ds, boolean_t inconsistentok, void *tag)
34dc7c2f 751{
b128c09f
BB		 752 boolean_t gotit = FALSE;
753
34dc7c2f 754 mutex_enter(&ds->ds_lock);
b128c09f
BB		 755 if (ds->ds_owner == NULL &&
756 (!DS_IS_INCONSISTENT(ds) || inconsistentok)) {
428870ff 757 ds->ds_owner = tag;
b128c09f
BB		 758 if (!dsl_pool_sync_context(ds->ds_dir->dd_pool))
759 rw_exit(&ds->ds_rwlock);
760 gotit = TRUE;
34dc7c2f
BB		 761 }
762 mutex_exit(&ds->ds_lock);
b128c09f 763 return (gotit);
34dc7c2f
BB		 764}
765
766void
b128c09f 767dsl_dataset_make_exclusive(dsl_dataset_t *ds, void *owner)
34dc7c2f 768{
b128c09f
BB		 769 ASSERT3P(owner, ==, ds->ds_owner);
770 if (!RW_WRITE_HELD(&ds->ds_rwlock))
771 rw_enter(&ds->ds_rwlock, RW_WRITER);
34dc7c2f
BB		 772}
773
774uint64_t
b128c09f 775dsl_dataset_create_sync_dd(dsl_dir_t *dd, dsl_dataset_t *origin,
34dc7c2f
BB		 776 uint64_t flags, dmu_tx_t *tx)
777{
778 dsl_pool_t *dp = dd->dd_pool;
779 dmu_buf_t *dbuf;
780 dsl_dataset_phys_t *dsphys;
781 uint64_t dsobj;
782 objset_t *mos = dp->dp_meta_objset;
783
b128c09f
BB		 784 if (origin == NULL)
785 origin = dp->dp_origin_snap;
786
34dc7c2f
BB		 787 ASSERT(origin == NULL || origin->ds_dir->dd_pool == dp);
788 ASSERT(origin == NULL || origin->ds_phys->ds_num_children > 0);
789 ASSERT(dmu_tx_is_syncing(tx));
790 ASSERT(dd->dd_phys->dd_head_dataset_obj == 0);
791
792 dsobj = dmu_object_alloc(mos, DMU_OT_DSL_DATASET, 0,
793 DMU_OT_DSL_DATASET, sizeof (dsl_dataset_phys_t), tx);
794 VERIFY(0 == dmu_bonus_hold(mos, dsobj, FTAG, &dbuf));
795 dmu_buf_will_dirty(dbuf, tx);
796 dsphys = dbuf->db_data;
b128c09f 797 bzero(dsphys, sizeof (dsl_dataset_phys_t));
34dc7c2f
BB		 798 dsphys->ds_dir_obj = dd->dd_object;
799 dsphys->ds_flags = flags;
800 dsphys->ds_fsid_guid = unique_create();
801 (void) random_get_pseudo_bytes((void*)&dsphys->ds_guid,
802 sizeof (dsphys->ds_guid));
803 dsphys->ds_snapnames_zapobj =
804 zap_create_norm(mos, U8_TEXTPREP_TOUPPER, DMU_OT_DSL_DS_SNAP_MAP,
805 DMU_OT_NONE, 0, tx);
806 dsphys->ds_creation_time = gethrestime_sec();
b128c09f 807 dsphys->ds_creation_txg = tx->tx_txg == TXG_INITIAL ? 1 : tx->tx_txg;
34dc7c2f 808
428870ff
BB		 809 if (origin == NULL) {
810 dsphys->ds_deadlist_obj = dsl_deadlist_alloc(mos, tx);
811 } else {
812 dsl_dataset_t *ohds;
813
34dc7c2f
BB		 814 dsphys->ds_prev_snap_obj = origin->ds_object;
815 dsphys->ds_prev_snap_txg =
816 origin->ds_phys->ds_creation_txg;
817 dsphys->ds_used_bytes =
818 origin->ds_phys->ds_used_bytes;
819 dsphys->ds_compressed_bytes =
820 origin->ds_phys->ds_compressed_bytes;
821 dsphys->ds_uncompressed_bytes =
822 origin->ds_phys->ds_uncompressed_bytes;
823 dsphys->ds_bp = origin->ds_phys->ds_bp;
824 dsphys->ds_flags |= origin->ds_phys->ds_flags;
825
826 dmu_buf_will_dirty(origin->ds_dbuf, tx);
827 origin->ds_phys->ds_num_children++;
828
428870ff
BB		 829 VERIFY3U(0, ==, dsl_dataset_hold_obj(dp,
830 origin->ds_dir->dd_phys->dd_head_dataset_obj, FTAG, &ohds));
831 dsphys->ds_deadlist_obj = dsl_deadlist_clone(&ohds->ds_deadlist,
832 dsphys->ds_prev_snap_txg, dsphys->ds_prev_snap_obj, tx);
833 dsl_dataset_rele(ohds, FTAG);
834
b128c09f
BB		 835 if (spa_version(dp->dp_spa) >= SPA_VERSION_NEXT_CLONES) {
836 if (origin->ds_phys->ds_next_clones_obj == 0) {
837 origin->ds_phys->ds_next_clones_obj =
838 zap_create(mos,
839 DMU_OT_NEXT_CLONES, DMU_OT_NONE, 0, tx);
840 }
841 VERIFY(0 == zap_add_int(mos,
842 origin->ds_phys->ds_next_clones_obj,
843 dsobj, tx));
844 }
845
34dc7c2f
BB		 846 dmu_buf_will_dirty(dd->dd_dbuf, tx);
847 dd->dd_phys->dd_origin_obj = origin->ds_object;
428870ff
BB		 848 if (spa_version(dp->dp_spa) >= SPA_VERSION_DIR_CLONES) {
849 if (origin->ds_dir->dd_phys->dd_clones == 0) {
850 dmu_buf_will_dirty(origin->ds_dir->dd_dbuf, tx);
851 origin->ds_dir->dd_phys->dd_clones =
852 zap_create(mos,
853 DMU_OT_DSL_CLONES, DMU_OT_NONE, 0, tx);
854 }
855 VERIFY3U(0, ==, zap_add_int(mos,
856 origin->ds_dir->dd_phys->dd_clones, dsobj, tx));
857 }
34dc7c2f
BB		 858 }
859
860 if (spa_version(dp->dp_spa) >= SPA_VERSION_UNIQUE_ACCURATE)
861 dsphys->ds_flags |= DS_FLAG_UNIQUE_ACCURATE;
862
863 dmu_buf_rele(dbuf, FTAG);
864
865 dmu_buf_will_dirty(dd->dd_dbuf, tx);
866 dd->dd_phys->dd_head_dataset_obj = dsobj;
867
868 return (dsobj);
869}
870
871uint64_t
872dsl_dataset_create_sync(dsl_dir_t *pdd, const char *lastname,
873 dsl_dataset_t *origin, uint64_t flags, cred_t *cr, dmu_tx_t *tx)
874{
875 dsl_pool_t *dp = pdd->dd_pool;
876 uint64_t dsobj, ddobj;
877 dsl_dir_t *dd;
878
879 ASSERT(lastname[0] != '@');
880
b128c09f 881 ddobj = dsl_dir_create_sync(dp, pdd, lastname, tx);
34dc7c2f
BB		 882 VERIFY(0 == dsl_dir_open_obj(dp, ddobj, lastname, FTAG, &dd));
883
b128c09f 884 dsobj = dsl_dataset_create_sync_dd(dd, origin, flags, tx);
34dc7c2f
BB		 885
886 dsl_deleg_set_create_perms(dd, tx, cr);
887
888 dsl_dir_close(dd, FTAG);
889
572e2857
BB		 890 /*
891 * If we are creating a clone, make sure we zero out any stale
892 * data from the origin snapshots zil header.
893 */
894 if (origin != NULL) {
895 dsl_dataset_t *ds;
896 objset_t *os;
897
898 VERIFY3U(0, ==, dsl_dataset_hold_obj(dp, dsobj, FTAG, &ds));
899 VERIFY3U(0, ==, dmu_objset_from_ds(ds, &os));
900 bzero(&os->os_zil_header, sizeof (os->os_zil_header));
901 dsl_dataset_dirty(ds, tx);
902 dsl_dataset_rele(ds, FTAG);
903 }
904
34dc7c2f
BB		 905 return (dsobj);
906}
907
908struct destroyarg {
909 dsl_sync_task_group_t *dstg;
910 char *snapname;
911 char *failed;
45d1cae3 912 boolean_t defer;
34dc7c2f
BB		 913};
914
915static int
428870ff 916dsl_snapshot_destroy_one(const char *name, void *arg)
34dc7c2f
BB		 917{
918 struct destroyarg *da = arg;
919 dsl_dataset_t *ds;
34dc7c2f 920 int err;
45d1cae3 921 char *dsname;
428870ff
BB		 922
923 dsname = kmem_asprintf("%s@%s", name, da->snapname);
924 err = dsl_dataset_own(dsname, B_TRUE, da->dstg, &ds);
925 strfree(dsname);
b128c09f 926 if (err == 0) {
45d1cae3
BB		 927 struct dsl_ds_destroyarg *dsda;
928
b128c09f 929 dsl_dataset_make_exclusive(ds, da->dstg);
45d1cae3
BB		 930 dsda = kmem_zalloc(sizeof (struct dsl_ds_destroyarg), KM_SLEEP);
931 dsda->ds = ds;
932 dsda->defer = da->defer;
b128c09f 933 dsl_sync_task_create(da->dstg, dsl_dataset_destroy_check,
45d1cae3 934 dsl_dataset_destroy_sync, dsda, da->dstg, 0);
b128c09f
BB		 935 } else if (err == ENOENT) {
936 err = 0;
937 } else {
34dc7c2f 938 (void) strcpy(da->failed, name);
34dc7c2f 939 }
b128c09f 940 return (err);
34dc7c2f
BB		 941}
942
943/*
944 * Destroy 'snapname' in all descendants of 'fsname'.
945 */
946#pragma weak dmu_snapshots_destroy = dsl_snapshots_destroy
947int
45d1cae3 948dsl_snapshots_destroy(char *fsname, char *snapname, boolean_t defer)
34dc7c2f
BB		 949{
950 int err;
951 struct destroyarg da;
952 dsl_sync_task_t *dst;
953 spa_t *spa;
954
955 err = spa_open(fsname, &spa, FTAG);
956 if (err)
957 return (err);
958 da.dstg = dsl_sync_task_group_create(spa_get_dsl(spa));
959 da.snapname = snapname;
960 da.failed = fsname;
45d1cae3 961 da.defer = defer;
34dc7c2f
BB		 962
963 err = dmu_objset_find(fsname,
964 dsl_snapshot_destroy_one, &da, DS_FIND_CHILDREN);
965
966 if (err == 0)
967 err = dsl_sync_task_group_wait(da.dstg);
968
969 for (dst = list_head(&da.dstg->dstg_tasks); dst;
970 dst = list_next(&da.dstg->dstg_tasks, dst)) {
45d1cae3
BB		 971 struct dsl_ds_destroyarg *dsda = dst->dst_arg1;
972 dsl_dataset_t *ds = dsda->ds;
973
b128c09f
BB		 974 /*
975 * Return the file system name that triggered the error
976 */
34dc7c2f
BB		 977 if (dst->dst_err) {
978 dsl_dataset_name(ds, fsname);
979 *strchr(fsname, '@') = '\0';
980 }
45d1cae3 981 ASSERT3P(dsda->rm_origin, ==, NULL);
b128c09f 982 dsl_dataset_disown(ds, da.dstg);
45d1cae3 983 kmem_free(dsda, sizeof (struct dsl_ds_destroyarg));
34dc7c2f
BB		 984 }
985
986 dsl_sync_task_group_destroy(da.dstg);
987 spa_close(spa, FTAG);
988 return (err);
989}
990
45d1cae3
BB		 991static boolean_t
992dsl_dataset_might_destroy_origin(dsl_dataset_t *ds)
993{
994 boolean_t might_destroy = B_FALSE;
995
996 mutex_enter(&ds->ds_lock);
997 if (ds->ds_phys->ds_num_children == 2 && ds->ds_userrefs == 0 &&
998 DS_IS_DEFER_DESTROY(ds))
999 might_destroy = B_TRUE;
1000 mutex_exit(&ds->ds_lock);
1001
1002 return (might_destroy);
1003}
1004
45d1cae3
BB		 1005/*
1006 * If we're removing a clone, and these three conditions are true:
1007 * 1) the clone's origin has no other children
1008 * 2) the clone's origin has no user references
1009 * 3) the clone's origin has been marked for deferred destruction
1010 * Then, prepare to remove the origin as part of this sync task group.
1011 */
1012static int
1013dsl_dataset_origin_rm_prep(struct dsl_ds_destroyarg *dsda, void *tag)
1014{
1015 dsl_dataset_t *ds = dsda->ds;
1016 dsl_dataset_t *origin = ds->ds_prev;
1017
1018 if (dsl_dataset_might_destroy_origin(origin)) {
1019 char *name;
1020 int namelen;
1021 int error;
1022
1023 namelen = dsl_dataset_namelen(origin) + 1;
1024 name = kmem_alloc(namelen, KM_SLEEP);
1025 dsl_dataset_name(origin, name);
1026#ifdef _KERNEL
1027 error = zfs_unmount_snap(name, NULL);
1028 if (error) {
1029 kmem_free(name, namelen);
1030 return (error);
1031 }
45d1cae3 1032#endif
428870ff 1033 error = dsl_dataset_own(name, B_TRUE, tag, &origin);
45d1cae3
BB		 1034 kmem_free(name, namelen);
1035 if (error)
1036 return (error);
1037 dsda->rm_origin = origin;
1038 dsl_dataset_make_exclusive(origin, tag);
1039 }
1040
1041 return (0);
1042}
1043
34dc7c2f 1044/*
b128c09f
BB		 1045 * ds must be opened as OWNER. On return (whether successful or not),
1046 * ds will be closed and caller can no longer dereference it.
34dc7c2f
BB		 1047 */
1048int
45d1cae3 1049dsl_dataset_destroy(dsl_dataset_t *ds, void *tag, boolean_t defer)
34dc7c2f
BB		 1050{
1051 int err;
1052 dsl_sync_task_group_t *dstg;
1053 objset_t *os;
1054 dsl_dir_t *dd;
1055 uint64_t obj;
428870ff
BB		 1056 struct dsl_ds_destroyarg dsda = { 0 };
1057 dsl_dataset_t dummy_ds = { 0 };
45d1cae3
BB		 1058
1059 dsda.ds = ds;
34dc7c2f 1060
34dc7c2f
BB		 1061 if (dsl_dataset_is_snapshot(ds)) {
1062 /* Destroying a snapshot is simpler */
b128c09f
BB		 1063 dsl_dataset_make_exclusive(ds, tag);
1064
45d1cae3 1065 dsda.defer = defer;
34dc7c2f
BB		 1066 err = dsl_sync_task_do(ds->ds_dir->dd_pool,
1067 dsl_dataset_destroy_check, dsl_dataset_destroy_sync,
45d1cae3
BB		 1068 &dsda, tag, 0);
1069 ASSERT3P(dsda.rm_origin, ==, NULL);
34dc7c2f 1070 goto out;
428870ff
BB		 1071 } else if (defer) {
1072 err = EINVAL;
1073 goto out;
34dc7c2f
BB		 1074 }
1075
1076 dd = ds->ds_dir;
428870ff
BB		 1077 dummy_ds.ds_dir = dd;
1078 dummy_ds.ds_object = ds->ds_object;
34dc7c2f
BB		 1079
1080 /*
1081 * Check for errors and mark this ds as inconsistent, in
1082 * case we crash while freeing the objects.
1083 */
1084 err = dsl_sync_task_do(dd->dd_pool, dsl_dataset_destroy_begin_check,
1085 dsl_dataset_destroy_begin_sync, ds, NULL, 0);
1086 if (err)
1087 goto out;
1088
428870ff 1089 err = dmu_objset_from_ds(ds, &os);
34dc7c2f
BB		 1090 if (err)
1091 goto out;
1092
1093 /*
1094 * remove the objects in open context, so that we won't
1095 * have too much to do in syncing context.
1096 */
1097 for (obj = 0; err == 0; err = dmu_object_next(os, &obj, FALSE,
1098 ds->ds_phys->ds_prev_snap_txg)) {
b128c09f
BB		 1099 /*
1100 * Ignore errors, if there is not enough disk space
1101 * we will deal with it in dsl_dataset_destroy_sync().
1102 */
1103 (void) dmu_free_object(os, obj);
34dc7c2f 1104 }
572e2857
BB		 1105 if (err != ESRCH)
1106 goto out;
34dc7c2f 1107
9babb374 1108 /*
572e2857
BB		 1109 * Only the ZIL knows how to free log blocks.
1110 */
1111 zil_destroy(dmu_objset_zil(os), B_FALSE);
1112
1113 /*
1114 * Sync out all in-flight IO.
9babb374
BB		 1115 */
1116 txg_wait_synced(dd->dd_pool, 0);
1117
1118 /*
1119 * If we managed to free all the objects in open
1120 * context, the user space accounting should be zero.
1121 */
1122 if (ds->ds_phys->ds_bp.blk_fill == 0 &&
428870ff 1123 dmu_objset_userused_enabled(os)) {
1fde1e37 1124 ASSERTV(uint64_t count);
9babb374
BB		 1125 ASSERT(zap_count(os, DMU_USERUSED_OBJECT, &count) != 0 ||
1126 count == 0);
1127 ASSERT(zap_count(os, DMU_GROUPUSED_OBJECT, &count) != 0 ||
1128 count == 0);
1129 }
1130
34dc7c2f
BB		 1131 rw_enter(&dd->dd_pool->dp_config_rwlock, RW_READER);
1132 err = dsl_dir_open_obj(dd->dd_pool, dd->dd_object, NULL, FTAG, &dd);
1133 rw_exit(&dd->dd_pool->dp_config_rwlock);
1134
1135 if (err)
1136 goto out;
1137
1138 /*
1139 * Blow away the dsl_dir + head dataset.
1140 */
b128c09f 1141 dsl_dataset_make_exclusive(ds, tag);
45d1cae3
BB		 1142 /*
1143 * If we're removing a clone, we might also need to remove its
1144 * origin.
1145 */
1146 do {
1147 dsda.need_prep = B_FALSE;
1148 if (dsl_dir_is_clone(dd)) {
1149 err = dsl_dataset_origin_rm_prep(&dsda, tag);
1150 if (err) {
1151 dsl_dir_close(dd, FTAG);
1152 goto out;
1153 }
1154 }
1155
1156 dstg = dsl_sync_task_group_create(ds->ds_dir->dd_pool);
1157 dsl_sync_task_create(dstg, dsl_dataset_destroy_check,
1158 dsl_dataset_destroy_sync, &dsda, tag, 0);
1159 dsl_sync_task_create(dstg, dsl_dir_destroy_check,
428870ff 1160 dsl_dir_destroy_sync, &dummy_ds, FTAG, 0);
45d1cae3
BB		 1161 err = dsl_sync_task_group_wait(dstg);
1162 dsl_sync_task_group_destroy(dstg);
1163
1164 /*
1165 * We could be racing against 'zfs release' or 'zfs destroy -d'
1166 * on the origin snap, in which case we can get EBUSY if we
1167 * needed to destroy the origin snap but were not ready to
1168 * do so.
1169 */
1170 if (dsda.need_prep) {
1171 ASSERT(err == EBUSY);
1172 ASSERT(dsl_dir_is_clone(dd));
1173 ASSERT(dsda.rm_origin == NULL);
1174 }
1175 } while (dsda.need_prep);
1176
1177 if (dsda.rm_origin != NULL)
1178 dsl_dataset_disown(dsda.rm_origin, tag);
1179
b128c09f 1180 /* if it is successful, dsl_dir_destroy_sync will close the dd */
34dc7c2f
BB		 1181 if (err)
1182 dsl_dir_close(dd, FTAG);
1183out:
b128c09f 1184 dsl_dataset_disown(ds, tag);
34dc7c2f
BB		 1185 return (err);
1186}
1187
34dc7c2f
BB		 1188blkptr_t *
1189dsl_dataset_get_blkptr(dsl_dataset_t *ds)
1190{
1191 return (&ds->ds_phys->ds_bp);
1192}
1193
1194void
1195dsl_dataset_set_blkptr(dsl_dataset_t *ds, blkptr_t *bp, dmu_tx_t *tx)
1196{
1197 ASSERT(dmu_tx_is_syncing(tx));
1198 /* If it's the meta-objset, set dp_meta_rootbp */
1199 if (ds == NULL) {
1200 tx->tx_pool->dp_meta_rootbp = *bp;
1201 } else {
1202 dmu_buf_will_dirty(ds->ds_dbuf, tx);
1203 ds->ds_phys->ds_bp = *bp;
1204 }
1205}
1206
1207spa_t *
1208dsl_dataset_get_spa(dsl_dataset_t *ds)
1209{
1210 return (ds->ds_dir->dd_pool->dp_spa);
1211}
1212
1213void
1214dsl_dataset_dirty(dsl_dataset_t *ds, dmu_tx_t *tx)
1215{
1216 dsl_pool_t *dp;
1217
1218 if (ds == NULL) /* this is the meta-objset */
1219 return;
1220
428870ff 1221 ASSERT(ds->ds_objset != NULL);
34dc7c2f
BB		 1222
1223 if (ds->ds_phys->ds_next_snap_obj != 0)
1224 panic("dirtying snapshot!");
1225
1226 dp = ds->ds_dir->dd_pool;
1227
1228 if (txg_list_add(&dp->dp_dirty_datasets, ds, tx->tx_txg) == 0) {
1229 /* up the hold count until we can be written out */
1230 dmu_buf_add_ref(ds->ds_dbuf, ds);
1231 }
1232}
1233
1234/*
1235 * The unique space in the head dataset can be calculated by subtracting
1236 * the space used in the most recent snapshot, that is still being used
1237 * in this file system, from the space currently in use. To figure out
1238 * the space in the most recent snapshot still in use, we need to take
1239 * the total space used in the snapshot and subtract out the space that
1240 * has been freed up since the snapshot was taken.
1241 */
1242static void
1243dsl_dataset_recalc_head_uniq(dsl_dataset_t *ds)
1244{
1245 uint64_t mrs_used;
1246 uint64_t dlused, dlcomp, dluncomp;
1247
428870ff 1248 ASSERT(!dsl_dataset_is_snapshot(ds));
34dc7c2f
BB		 1249
1250 if (ds->ds_phys->ds_prev_snap_obj != 0)
1251 mrs_used = ds->ds_prev->ds_phys->ds_used_bytes;
1252 else
1253 mrs_used = 0;
1254
428870ff 1255 dsl_deadlist_space(&ds->ds_deadlist, &dlused, &dlcomp, &dluncomp);
34dc7c2f
BB		 1256
1257 ASSERT3U(dlused, <=, mrs_used);
1258 ds->ds_phys->ds_unique_bytes =
1259 ds->ds_phys->ds_used_bytes - (mrs_used - dlused);
1260
428870ff 1261 if (spa_version(ds->ds_dir->dd_pool->dp_spa) >=
34dc7c2f
BB		 1262 SPA_VERSION_UNIQUE_ACCURATE)
1263 ds->ds_phys->ds_flags |= DS_FLAG_UNIQUE_ACCURATE;
1264}
1265
34dc7c2f 1266struct killarg {
b128c09f 1267 dsl_dataset_t *ds;
34dc7c2f
BB		 1268 dmu_tx_t *tx;
1269};
1270
b128c09f 1271/* ARGSUSED */
34dc7c2f 1272static int
428870ff
BB		 1273kill_blkptr(spa_t *spa, zilog_t *zilog, const blkptr_t *bp, arc_buf_t *pbuf,
1274 const zbookmark_t *zb, const dnode_phys_t *dnp, void *arg)
34dc7c2f
BB		 1275{
1276 struct killarg *ka = arg;
428870ff 1277 dmu_tx_t *tx = ka->tx;
34dc7c2f 1278
b128c09f
BB		 1279 if (bp == NULL)
1280 return (0);
1281
428870ff
BB		 1282 if (zb->zb_level == ZB_ZIL_LEVEL) {
1283 ASSERT(zilog != NULL);
9babb374
BB		 1284 /*
1285 * It's a block in the intent log. It has no
1286 * accounting, so just free it.
1287 */
428870ff 1288 dsl_free(ka->tx->tx_pool, ka->tx->tx_txg, bp);
9babb374 1289 } else {
428870ff 1290 ASSERT(zilog == NULL);
9babb374 1291 ASSERT3U(bp->blk_birth, >, ka->ds->ds_phys->ds_prev_snap_txg);
428870ff 1292 (void) dsl_dataset_block_kill(ka->ds, bp, tx, B_FALSE);
9babb374 1293 }
34dc7c2f 1294
34dc7c2f
BB		 1295 return (0);
1296}
1297
34dc7c2f
BB		 1298/* ARGSUSED */
1299static int
1300dsl_dataset_destroy_begin_check(void *arg1, void *arg2, dmu_tx_t *tx)
1301{
1302 dsl_dataset_t *ds = arg1;
1303 objset_t *mos = ds->ds_dir->dd_pool->dp_meta_objset;
1304 uint64_t count;
1305 int err;
1306
1307 /*
1308 * Can't delete a head dataset if there are snapshots of it.
1309 * (Except if the only snapshots are from the branch we cloned
1310 * from.)
1311 */
1312 if (ds->ds_prev != NULL &&
1313 ds->ds_prev->ds_phys->ds_next_snap_obj == ds->ds_object)
428870ff 1314 return (EBUSY);
34dc7c2f
BB		 1315
1316 /*
1317 * This is really a dsl_dir thing, but check it here so that
1318 * we'll be less likely to leave this dataset inconsistent &
1319 * nearly destroyed.
1320 */
1321 err = zap_count(mos, ds->ds_dir->dd_phys->dd_child_dir_zapobj, &count);
1322 if (err)
1323 return (err);
1324 if (count != 0)
1325 return (EEXIST);
1326
1327 return (0);
1328}
1329
1330/* ARGSUSED */
1331static void
428870ff 1332dsl_dataset_destroy_begin_sync(void *arg1, void *arg2, dmu_tx_t *tx)
34dc7c2f
BB		 1333{
1334 dsl_dataset_t *ds = arg1;
1335 dsl_pool_t *dp = ds->ds_dir->dd_pool;
1336
1337 /* Mark it as inconsistent on-disk, in case we crash */
1338 dmu_buf_will_dirty(ds->ds_dbuf, tx);
1339 ds->ds_phys->ds_flags |= DS_FLAG_INCONSISTENT;
1340
428870ff
BB		 1341 spa_history_log_internal(LOG_DS_DESTROY_BEGIN, dp->dp_spa, tx,
1342 "dataset = %llu", ds->ds_object);
34dc7c2f
BB		 1343}
1344
45d1cae3
BB		 1345static int
1346dsl_dataset_origin_check(struct dsl_ds_destroyarg *dsda, void *tag,
1347 dmu_tx_t *tx)
1348{
1349 dsl_dataset_t *ds = dsda->ds;
1350 dsl_dataset_t *ds_prev = ds->ds_prev;
1351
1352 if (dsl_dataset_might_destroy_origin(ds_prev)) {
1353 struct dsl_ds_destroyarg ndsda = {0};
1354
1355 /*
1356 * If we're not prepared to remove the origin, don't remove
1357 * the clone either.
1358 */
1359 if (dsda->rm_origin == NULL) {
1360 dsda->need_prep = B_TRUE;
1361 return (EBUSY);
1362 }
1363
1364 ndsda.ds = ds_prev;
1365 ndsda.is_origin_rm = B_TRUE;
1366 return (dsl_dataset_destroy_check(&ndsda, tag, tx));
1367 }
1368
1369 /*
1370 * If we're not going to remove the origin after all,
1371 * undo the open context setup.
1372 */
1373 if (dsda->rm_origin != NULL) {
1374 dsl_dataset_disown(dsda->rm_origin, tag);
1375 dsda->rm_origin = NULL;
1376 }
1377
1378 return (0);
1379}
1380
572e2857
BB		 1381/*
1382 * If you add new checks here, you may need to add
1383 * additional checks to the "temporary" case in
1384 * snapshot_check() in dmu_objset.c.
1385 */
34dc7c2f
BB		 1386/* ARGSUSED */
1387int
1388dsl_dataset_destroy_check(void *arg1, void *arg2, dmu_tx_t *tx)
1389{
45d1cae3
BB		 1390 struct dsl_ds_destroyarg *dsda = arg1;
1391 dsl_dataset_t *ds = dsda->ds;
34dc7c2f 1392
b128c09f
BB		 1393 /* we have an owner hold, so noone else can destroy us */
1394 ASSERT(!DSL_DATASET_IS_DESTROYED(ds));
1395
45d1cae3
BB		 1396 /*
1397 * Only allow deferred destroy on pools that support it.
1398 * NOTE: deferred destroy is only supported on snapshots.
1399 */
1400 if (dsda->defer) {
1401 if (spa_version(ds->ds_dir->dd_pool->dp_spa) <
1402 SPA_VERSION_USERREFS)
1403 return (ENOTSUP);
1404 ASSERT(dsl_dataset_is_snapshot(ds));
1405 return (0);
1406 }
34dc7c2f
BB		 1407
1408 /*
1409 * Can't delete a head dataset if there are snapshots of it.
1410 * (Except if the only snapshots are from the branch we cloned
1411 * from.)
1412 */
1413 if (ds->ds_prev != NULL &&
1414 ds->ds_prev->ds_phys->ds_next_snap_obj == ds->ds_object)
428870ff 1415 return (EBUSY);
34dc7c2f
BB		 1416
1417 /*
1418 * If we made changes this txg, traverse_dsl_dataset won't find
1419 * them. Try again.
1420 */
1421 if (ds->ds_phys->ds_bp.blk_birth >= tx->tx_txg)
1422 return (EAGAIN);
1423
45d1cae3
BB		 1424 if (dsl_dataset_is_snapshot(ds)) {
1425 /*
1426 * If this snapshot has an elevated user reference count,
1427 * we can't destroy it yet.
1428 */
1429 if (ds->ds_userrefs > 0 && !dsda->releasing)
1430 return (EBUSY);
1431
1432 mutex_enter(&ds->ds_lock);
1433 /*
1434 * Can't delete a branch point. However, if we're destroying
1435 * a clone and removing its origin due to it having a user
1436 * hold count of 0 and having been marked for deferred destroy,
1437 * it's OK for the origin to have a single clone.
1438 */
1439 if (ds->ds_phys->ds_num_children >
1440 (dsda->is_origin_rm ? 2 : 1)) {
1441 mutex_exit(&ds->ds_lock);
1442 return (EEXIST);
1443 }
1444 mutex_exit(&ds->ds_lock);
1445 } else if (dsl_dir_is_clone(ds->ds_dir)) {
1446 return (dsl_dataset_origin_check(dsda, arg2, tx));
1447 }
1448
34dc7c2f
BB		 1449 /* XXX we should do some i/o error checking... */
1450 return (0);
1451}
1452
b128c09f
BB		 1453struct refsarg {
1454 kmutex_t lock;
1455 boolean_t gone;
1456 kcondvar_t cv;
1457};
1458
1459/* ARGSUSED */
1460static void
1461dsl_dataset_refs_gone(dmu_buf_t *db, void *argv)
1462{
1463 struct refsarg *arg = argv;
1464
1465 mutex_enter(&arg->lock);
1466 arg->gone = TRUE;
1467 cv_signal(&arg->cv);
1468 mutex_exit(&arg->lock);
1469}
1470
1471static void
1472dsl_dataset_drain_refs(dsl_dataset_t *ds, void *tag)
1473{
1474 struct refsarg arg;
1475
1476 mutex_init(&arg.lock, NULL, MUTEX_DEFAULT, NULL);
1477 cv_init(&arg.cv, NULL, CV_DEFAULT, NULL);
1478 arg.gone = FALSE;
1479 (void) dmu_buf_update_user(ds->ds_dbuf, ds, &arg, &ds->ds_phys,
1480 dsl_dataset_refs_gone);
1481 dmu_buf_rele(ds->ds_dbuf, tag);
1482 mutex_enter(&arg.lock);
1483 while (!arg.gone)
1484 cv_wait(&arg.cv, &arg.lock);
1485 ASSERT(arg.gone);
1486 mutex_exit(&arg.lock);
1487 ds->ds_dbuf = NULL;
1488 ds->ds_phys = NULL;
1489 mutex_destroy(&arg.lock);
1490 cv_destroy(&arg.cv);
1491}
1492
428870ff
BB		 1493static void
1494remove_from_next_clones(dsl_dataset_t *ds, uint64_t obj, dmu_tx_t *tx)
1495{
1496 objset_t *mos = ds->ds_dir->dd_pool->dp_meta_objset;
428870ff 1497 int err;
1fde1e37 1498 ASSERTV(uint64_t count);
428870ff
BB		 1499
1500 ASSERT(ds->ds_phys->ds_num_children >= 2);
1501 err = zap_remove_int(mos, ds->ds_phys->ds_next_clones_obj, obj, tx);
1502 /*
1503 * The err should not be ENOENT, but a bug in a previous version
1504 * of the code could cause upgrade_clones_cb() to not set
1505 * ds_next_snap_obj when it should, leading to a missing entry.
1506 * If we knew that the pool was created after
1507 * SPA_VERSION_NEXT_CLONES, we could assert that it isn't
1508 * ENOENT. However, at least we can check that we don't have
1509 * too many entries in the next_clones_obj even after failing to
1510 * remove this one.
1511 */
1512 if (err != ENOENT) {
1513 VERIFY3U(err, ==, 0);
1514 }
1515 ASSERT3U(0, ==, zap_count(mos, ds->ds_phys->ds_next_clones_obj,
1516 &count));
1517 ASSERT3U(count, <=, ds->ds_phys->ds_num_children - 2);
1518}
1519
1520static void
1521dsl_dataset_remove_clones_key(dsl_dataset_t *ds, uint64_t mintxg, dmu_tx_t *tx)
1522{
1523 objset_t *mos = ds->ds_dir->dd_pool->dp_meta_objset;
1524 zap_cursor_t zc;
1525 zap_attribute_t za;
1526
1527 /*
1528 * If it is the old version, dd_clones doesn't exist so we can't
1529 * find the clones, but deadlist_remove_key() is a no-op so it
1530 * doesn't matter.
1531 */
1532 if (ds->ds_dir->dd_phys->dd_clones == 0)
1533 return;
1534
1535 for (zap_cursor_init(&zc, mos, ds->ds_dir->dd_phys->dd_clones);
1536 zap_cursor_retrieve(&zc, &za) == 0;
1537 zap_cursor_advance(&zc)) {
1538 dsl_dataset_t *clone;
1539
1540 VERIFY3U(0, ==, dsl_dataset_hold_obj(ds->ds_dir->dd_pool,
1541 za.za_first_integer, FTAG, &clone));
1542 if (clone->ds_dir->dd_origin_txg > mintxg) {
1543 dsl_deadlist_remove_key(&clone->ds_deadlist,
1544 mintxg, tx);
1545 dsl_dataset_remove_clones_key(clone, mintxg, tx);
1546 }
1547 dsl_dataset_rele(clone, FTAG);
1548 }
1549 zap_cursor_fini(&zc);
1550}
1551
1552struct process_old_arg {
1553 dsl_dataset_t *ds;
1554 dsl_dataset_t *ds_prev;
1555 boolean_t after_branch_point;
1556 zio_t *pio;
1557 uint64_t used, comp, uncomp;
1558};
1559
1560static int
1561process_old_cb(void *arg, const blkptr_t *bp, dmu_tx_t *tx)
1562{
1563 struct process_old_arg *poa = arg;
1564 dsl_pool_t *dp = poa->ds->ds_dir->dd_pool;
1565
1566 if (bp->blk_birth <= poa->ds->ds_phys->ds_prev_snap_txg) {
1567 dsl_deadlist_insert(&poa->ds->ds_deadlist, bp, tx);
1568 if (poa->ds_prev && !poa->after_branch_point &&
1569 bp->blk_birth >
1570 poa->ds_prev->ds_phys->ds_prev_snap_txg) {
1571 poa->ds_prev->ds_phys->ds_unique_bytes +=
1572 bp_get_dsize_sync(dp->dp_spa, bp);
1573 }
1574 } else {
1575 poa->used += bp_get_dsize_sync(dp->dp_spa, bp);
1576 poa->comp += BP_GET_PSIZE(bp);
1577 poa->uncomp += BP_GET_UCSIZE(bp);
1578 dsl_free_sync(poa->pio, dp, tx->tx_txg, bp);
1579 }
1580 return (0);
1581}
1582
1583static void
1584process_old_deadlist(dsl_dataset_t *ds, dsl_dataset_t *ds_prev,
1585 dsl_dataset_t *ds_next, boolean_t after_branch_point, dmu_tx_t *tx)
1586{
1587 struct process_old_arg poa = { 0 };
1588 dsl_pool_t *dp = ds->ds_dir->dd_pool;
1589 objset_t *mos = dp->dp_meta_objset;
1590
1591 ASSERT(ds->ds_deadlist.dl_oldfmt);
1592 ASSERT(ds_next->ds_deadlist.dl_oldfmt);
1593
1594 poa.ds = ds;
1595 poa.ds_prev = ds_prev;
1596 poa.after_branch_point = after_branch_point;
1597 poa.pio = zio_root(dp->dp_spa, NULL, NULL, ZIO_FLAG_MUSTSUCCEED);
1598 VERIFY3U(0, ==, bpobj_iterate(&ds_next->ds_deadlist.dl_bpobj,
1599 process_old_cb, &poa, tx));
1600 VERIFY3U(zio_wait(poa.pio), ==, 0);
1601 ASSERT3U(poa.used, ==, ds->ds_phys->ds_unique_bytes);
1602
1603 /* change snapused */
1604 dsl_dir_diduse_space(ds->ds_dir, DD_USED_SNAP,
1605 -poa.used, -poa.comp, -poa.uncomp, tx);
1606
1607 /* swap next's deadlist to our deadlist */
1608 dsl_deadlist_close(&ds->ds_deadlist);
1609 dsl_deadlist_close(&ds_next->ds_deadlist);
1610 SWITCH64(ds_next->ds_phys->ds_deadlist_obj,
1611 ds->ds_phys->ds_deadlist_obj);
1612 dsl_deadlist_open(&ds->ds_deadlist, mos, ds->ds_phys->ds_deadlist_obj);
1613 dsl_deadlist_open(&ds_next->ds_deadlist, mos,
1614 ds_next->ds_phys->ds_deadlist_obj);
1615}
1616
34dc7c2f 1617void
428870ff 1618dsl_dataset_destroy_sync(void *arg1, void *tag, dmu_tx_t *tx)
34dc7c2f 1619{
45d1cae3
BB		 1620 struct dsl_ds_destroyarg *dsda = arg1;
1621 dsl_dataset_t *ds = dsda->ds;
34dc7c2f
BB		 1622 int err;
1623 int after_branch_point = FALSE;
1624 dsl_pool_t *dp = ds->ds_dir->dd_pool;
1625 objset_t *mos = dp->dp_meta_objset;
1626 dsl_dataset_t *ds_prev = NULL;
572e2857 1627 boolean_t wont_destroy;
34dc7c2f
BB		 1628 uint64_t obj;
1629
572e2857
BB		 1630 wont_destroy = (dsda->defer &&
1631 (ds->ds_userrefs > 0 || ds->ds_phys->ds_num_children > 1));
1632
1633 ASSERT(ds->ds_owner || wont_destroy);
45d1cae3 1634 ASSERT(dsda->defer || ds->ds_phys->ds_num_children <= 1);
34dc7c2f
BB		 1635 ASSERT(ds->ds_prev == NULL ||
1636 ds->ds_prev->ds_phys->ds_next_snap_obj != ds->ds_object);
1637 ASSERT3U(ds->ds_phys->ds_bp.blk_birth, <=, tx->tx_txg);
1638
572e2857 1639 if (wont_destroy) {
45d1cae3 1640 ASSERT(spa_version(dp->dp_spa) >= SPA_VERSION_USERREFS);
572e2857
BB		 1641 dmu_buf_will_dirty(ds->ds_dbuf, tx);
1642 ds->ds_phys->ds_flags |= DS_FLAG_DEFER_DESTROY;
1643 return;
45d1cae3
BB		 1644 }
1645
b128c09f
BB		 1646 /* signal any waiters that this dataset is going away */
1647 mutex_enter(&ds->ds_lock);
1648 ds->ds_owner = dsl_reaper;
1649 cv_broadcast(&ds->ds_exclusive_cv);
1650 mutex_exit(&ds->ds_lock);
1651
34dc7c2f
BB		 1652 /* Remove our reservation */
1653 if (ds->ds_reserved != 0) {
428870ff
BB		 1654 dsl_prop_setarg_t psa;
1655 uint64_t value = 0;
1656
1657 dsl_prop_setarg_init_uint64(&psa, "refreservation",
1658 (ZPROP_SRC_NONE | ZPROP_SRC_LOCAL | ZPROP_SRC_RECEIVED),
1659 &value);
1660 psa.psa_effective_value = 0; /* predict default value */
1661
1662 dsl_dataset_set_reservation_sync(ds, &psa, tx);
34dc7c2f
BB		 1663 ASSERT3U(ds->ds_reserved, ==, 0);
1664 }
1665
1666 ASSERT(RW_WRITE_HELD(&dp->dp_config_rwlock));
1667
428870ff 1668 dsl_scan_ds_destroyed(ds, tx);
b128c09f 1669
34dc7c2f
BB		 1670 obj = ds->ds_object;
1671
1672 if (ds->ds_phys->ds_prev_snap_obj != 0) {
1673 if (ds->ds_prev) {
1674 ds_prev = ds->ds_prev;
1675 } else {
b128c09f
BB		 1676 VERIFY(0 == dsl_dataset_hold_obj(dp,
1677 ds->ds_phys->ds_prev_snap_obj, FTAG, &ds_prev));
34dc7c2f
BB		 1678 }
1679 after_branch_point =
1680 (ds_prev->ds_phys->ds_next_snap_obj != obj);
1681
1682 dmu_buf_will_dirty(ds_prev->ds_dbuf, tx);
b128c09f
BB		 1683 if (after_branch_point &&
1684 ds_prev->ds_phys->ds_next_clones_obj != 0) {
428870ff 1685 remove_from_next_clones(ds_prev, obj, tx);
b128c09f
BB		 1686 if (ds->ds_phys->ds_next_snap_obj != 0) {
1687 VERIFY(0 == zap_add_int(mos,
1688 ds_prev->ds_phys->ds_next_clones_obj,
1689 ds->ds_phys->ds_next_snap_obj, tx));
1690 }
1691 }
34dc7c2f
BB		 1692 if (after_branch_point &&
1693 ds->ds_phys->ds_next_snap_obj == 0) {
1694 /* This clone is toast. */
1695 ASSERT(ds_prev->ds_phys->ds_num_children > 1);
1696 ds_prev->ds_phys->ds_num_children--;
45d1cae3
BB		 1697
1698 /*
1699 * If the clone's origin has no other clones, no
1700 * user holds, and has been marked for deferred
1701 * deletion, then we should have done the necessary
1702 * destroy setup for it.
1703 */
1704 if (ds_prev->ds_phys->ds_num_children == 1 &&
1705 ds_prev->ds_userrefs == 0 &&
1706 DS_IS_DEFER_DESTROY(ds_prev)) {
1707 ASSERT3P(dsda->rm_origin, !=, NULL);
1708 } else {
1709 ASSERT3P(dsda->rm_origin, ==, NULL);
1710 }
34dc7c2f
BB		 1711 } else if (!after_branch_point) {
1712 ds_prev->ds_phys->ds_next_snap_obj =
1713 ds->ds_phys->ds_next_snap_obj;
1714 }
1715 }
1716
428870ff 1717 if (dsl_dataset_is_snapshot(ds)) {
34dc7c2f 1718 dsl_dataset_t *ds_next;
34dc7c2f 1719 uint64_t old_unique;
428870ff 1720 uint64_t used = 0, comp = 0, uncomp = 0;
34dc7c2f 1721
b128c09f
BB		 1722 VERIFY(0 == dsl_dataset_hold_obj(dp,
1723 ds->ds_phys->ds_next_snap_obj, FTAG, &ds_next));
34dc7c2f
BB		 1724 ASSERT3U(ds_next->ds_phys->ds_prev_snap_obj, ==, obj);
1725
428870ff 1726 old_unique = ds_next->ds_phys->ds_unique_bytes;
34dc7c2f
BB		 1727
1728 dmu_buf_will_dirty(ds_next->ds_dbuf, tx);
1729 ds_next->ds_phys->ds_prev_snap_obj =
1730 ds->ds_phys->ds_prev_snap_obj;
1731 ds_next->ds_phys->ds_prev_snap_txg =
1732 ds->ds_phys->ds_prev_snap_txg;
1733 ASSERT3U(ds->ds_phys->ds_prev_snap_txg, ==,
1734 ds_prev ? ds_prev->ds_phys->ds_creation_txg : 0);
1735
34dc7c2f 1736
428870ff
BB		 1737 if (ds_next->ds_deadlist.dl_oldfmt) {
1738 process_old_deadlist(ds, ds_prev, ds_next,
1739 after_branch_point, tx);
1740 } else {
1741 /* Adjust prev's unique space. */
1742 if (ds_prev && !after_branch_point) {
1743 dsl_deadlist_space_range(&ds_next->ds_deadlist,
1744 ds_prev->ds_phys->ds_prev_snap_txg,
1745 ds->ds_phys->ds_prev_snap_txg,
1746 &used, &comp, &uncomp);
1747 ds_prev->ds_phys->ds_unique_bytes += used;
1748 }
b128c09f 1749
428870ff
BB		 1750 /* Adjust snapused. */
1751 dsl_deadlist_space_range(&ds_next->ds_deadlist,
1752 ds->ds_phys->ds_prev_snap_txg, UINT64_MAX,
1753 &used, &comp, &uncomp);
1754 dsl_dir_diduse_space(ds->ds_dir, DD_USED_SNAP,
1755 -used, -comp, -uncomp, tx);
1756
1757 /* Move blocks to be freed to pool's free list. */
1758 dsl_deadlist_move_bpobj(&ds_next->ds_deadlist,
1759 &dp->dp_free_bpobj, ds->ds_phys->ds_prev_snap_txg,
1760 tx);
1761 dsl_dir_diduse_space(tx->tx_pool->dp_free_dir,
1762 DD_USED_HEAD, used, comp, uncomp, tx);
1763 dsl_dir_dirty(tx->tx_pool->dp_free_dir, tx);
1764
1765 /* Merge our deadlist into next's and free it. */
1766 dsl_deadlist_merge(&ds_next->ds_deadlist,
1767 ds->ds_phys->ds_deadlist_obj, tx);
1768 }
1769 dsl_deadlist_close(&ds->ds_deadlist);
1770 dsl_deadlist_free(mos, ds->ds_phys->ds_deadlist_obj, tx);
b128c09f 1771
428870ff
BB		 1772 /* Collapse range in clone heads */
1773 dsl_dataset_remove_clones_key(ds,
1774 ds->ds_phys->ds_creation_txg, tx);
34dc7c2f 1775
428870ff
BB		 1776 if (dsl_dataset_is_snapshot(ds_next)) {
1777 dsl_dataset_t *ds_nextnext;
d6320ddb 1778 dsl_dataset_t *hds;
34dc7c2f 1779
34dc7c2f
BB		 1780 /*
1781 * Update next's unique to include blocks which
1782 * were previously shared by only this snapshot
1783 * and it. Those blocks will be born after the
1784 * prev snap and before this snap, and will have
1785 * died after the next snap and before the one
1786 * after that (ie. be on the snap after next's
1787 * deadlist).
34dc7c2f 1788 */
b128c09f
BB		 1789 VERIFY(0 == dsl_dataset_hold_obj(dp,
1790 ds_next->ds_phys->ds_next_snap_obj,
428870ff
BB		 1791 FTAG, &ds_nextnext));
1792 dsl_deadlist_space_range(&ds_nextnext->ds_deadlist,
b128c09f 1793 ds->ds_phys->ds_prev_snap_txg,
428870ff
BB		 1794 ds->ds_phys->ds_creation_txg,
1795 &used, &comp, &uncomp);
1796 ds_next->ds_phys->ds_unique_bytes += used;
1797 dsl_dataset_rele(ds_nextnext, FTAG);
34dc7c2f 1798 ASSERT3P(ds_next->ds_prev, ==, NULL);
428870ff
BB		 1799
1800 /* Collapse range in this head. */
428870ff
BB		 1801 VERIFY3U(0, ==, dsl_dataset_hold_obj(dp,
1802 ds->ds_dir->dd_phys->dd_head_dataset_obj,
1803 FTAG, &hds));
1804 dsl_deadlist_remove_key(&hds->ds_deadlist,
1805 ds->ds_phys->ds_creation_txg, tx);
1806 dsl_dataset_rele(hds, FTAG);
1807
34dc7c2f
BB		 1808 } else {
1809 ASSERT3P(ds_next->ds_prev, ==, ds);
b128c09f
BB		 1810 dsl_dataset_drop_ref(ds_next->ds_prev, ds_next);
1811 ds_next->ds_prev = NULL;
34dc7c2f 1812 if (ds_prev) {
b128c09f
BB		 1813 VERIFY(0 == dsl_dataset_get_ref(dp,
1814 ds->ds_phys->ds_prev_snap_obj,
1815 ds_next, &ds_next->ds_prev));
34dc7c2f
BB		 1816 }
1817
1818 dsl_dataset_recalc_head_uniq(ds_next);
1819
1820 /*
1821 * Reduce the amount of our unconsmed refreservation
1822 * being charged to our parent by the amount of
1823 * new unique data we have gained.
1824 */
1825 if (old_unique < ds_next->ds_reserved) {
1826 int64_t mrsdelta;
1827 uint64_t new_unique =
1828 ds_next->ds_phys->ds_unique_bytes;
1829
1830 ASSERT(old_unique <= new_unique);
1831 mrsdelta = MIN(new_unique - old_unique,
1832 ds_next->ds_reserved - old_unique);
b128c09f
BB		 1833 dsl_dir_diduse_space(ds->ds_dir,
1834 DD_USED_REFRSRV, -mrsdelta, 0, 0, tx);
34dc7c2f
BB		 1835 }
1836 }
b128c09f 1837 dsl_dataset_rele(ds_next, FTAG);
34dc7c2f
BB		 1838 } else {
1839 /*
1840 * There's no next snapshot, so this is a head dataset.
1841 * Destroy the deadlist. Unless it's a clone, the
1842 * deadlist should be empty. (If it's a clone, it's
1843 * safe to ignore the deadlist contents.)
1844 */
1845 struct killarg ka;
1846
428870ff
BB		 1847 dsl_deadlist_close(&ds->ds_deadlist);
1848 dsl_deadlist_free(mos, ds->ds_phys->ds_deadlist_obj, tx);
34dc7c2f
BB		 1849 ds->ds_phys->ds_deadlist_obj = 0;
1850
1851 /*
1852 * Free everything that we point to (that's born after
1853 * the previous snapshot, if we are a clone)
1854 *
b128c09f
BB		 1855 * NB: this should be very quick, because we already
1856 * freed all the objects in open context.
34dc7c2f 1857 */
b128c09f 1858 ka.ds = ds;
34dc7c2f 1859 ka.tx = tx;
b128c09f
BB		 1860 err = traverse_dataset(ds, ds->ds_phys->ds_prev_snap_txg,
1861 TRAVERSE_POST, kill_blkptr, &ka);
34dc7c2f 1862 ASSERT3U(err, ==, 0);
9babb374 1863 ASSERT(!DS_UNIQUE_IS_ACCURATE(ds) ||
b128c09f 1864 ds->ds_phys->ds_unique_bytes == 0);
34dc7c2f 1865
428870ff
BB		 1866 if (ds->ds_prev != NULL) {
1867 if (spa_version(dp->dp_spa) >= SPA_VERSION_DIR_CLONES) {
1868 VERIFY3U(0, ==, zap_remove_int(mos,
1869 ds->ds_prev->ds_dir->dd_phys->dd_clones,
1870 ds->ds_object, tx));
1871 }
1872 dsl_dataset_rele(ds->ds_prev, ds);
1873 ds->ds_prev = ds_prev = NULL;
1874 }
1875 }
34dc7c2f 1876
572e2857
BB		 1877 /*
1878 * This must be done after the dsl_traverse(), because it will
1879 * re-open the objset.
1880 */
1881 if (ds->ds_objset) {
1882 dmu_objset_evict(ds->ds_objset);
1883 ds->ds_objset = NULL;
1884 }
1885
34dc7c2f 1886 if (ds->ds_dir->dd_phys->dd_head_dataset_obj == ds->ds_object) {
b128c09f 1887 /* Erase the link in the dir */
34dc7c2f
BB		 1888 dmu_buf_will_dirty(ds->ds_dir->dd_dbuf, tx);
1889 ds->ds_dir->dd_phys->dd_head_dataset_obj = 0;
b128c09f
BB		 1890 ASSERT(ds->ds_phys->ds_snapnames_zapobj != 0);
1891 err = zap_destroy(mos, ds->ds_phys->ds_snapnames_zapobj, tx);
1892 ASSERT(err == 0);
34dc7c2f
BB		 1893 } else {
1894 /* remove from snapshot namespace */
1895 dsl_dataset_t *ds_head;
b128c09f
BB		 1896 ASSERT(ds->ds_phys->ds_snapnames_zapobj == 0);
1897 VERIFY(0 == dsl_dataset_hold_obj(dp,
1898 ds->ds_dir->dd_phys->dd_head_dataset_obj, FTAG, &ds_head));
34dc7c2f
BB		 1899 VERIFY(0 == dsl_dataset_get_snapname(ds));
1900#ifdef ZFS_DEBUG
1901 {
1902 uint64_t val;
1903
b128c09f 1904 err = dsl_dataset_snap_lookup(ds_head,
34dc7c2f
BB		 1905 ds->ds_snapname, &val);
1906 ASSERT3U(err, ==, 0);
1907 ASSERT3U(val, ==, obj);
1908 }
1909#endif
b128c09f 1910 err = dsl_dataset_snap_remove(ds_head, ds->ds_snapname, tx);
34dc7c2f 1911 ASSERT(err == 0);
b128c09f 1912 dsl_dataset_rele(ds_head, FTAG);
34dc7c2f
BB		 1913 }
1914
1915 if (ds_prev && ds->ds_prev != ds_prev)
b128c09f 1916 dsl_dataset_rele(ds_prev, FTAG);
34dc7c2f
BB		 1917
1918 spa_prop_clear_bootfs(dp->dp_spa, ds->ds_object, tx);
428870ff
BB		 1919 spa_history_log_internal(LOG_DS_DESTROY, dp->dp_spa, tx,
1920 "dataset = %llu", ds->ds_object);
34dc7c2f 1921
b128c09f 1922 if (ds->ds_phys->ds_next_clones_obj != 0) {
1fde1e37 1923 ASSERTV(uint64_t count);
b128c09f
BB		 1924 ASSERT(0 == zap_count(mos,
1925 ds->ds_phys->ds_next_clones_obj, &count) && count == 0);
1926 VERIFY(0 == dmu_object_free(mos,
1927 ds->ds_phys->ds_next_clones_obj, tx));
1928 }
1929 if (ds->ds_phys->ds_props_obj != 0)
1930 VERIFY(0 == zap_destroy(mos, ds->ds_phys->ds_props_obj, tx));
45d1cae3
BB		 1931 if (ds->ds_phys->ds_userrefs_obj != 0)
1932 VERIFY(0 == zap_destroy(mos, ds->ds_phys->ds_userrefs_obj, tx));
b128c09f
BB		 1933 dsl_dir_close(ds->ds_dir, ds);
1934 ds->ds_dir = NULL;
1935 dsl_dataset_drain_refs(ds, tag);
34dc7c2f 1936 VERIFY(0 == dmu_object_free(mos, obj, tx));
45d1cae3
BB		 1937
1938 if (dsda->rm_origin) {
1939 /*
1940 * Remove the origin of the clone we just destroyed.
1941 */
45d1cae3
BB		 1942 struct dsl_ds_destroyarg ndsda = {0};
1943
428870ff
BB		 1944 ndsda.ds = dsda->rm_origin;
1945 dsl_dataset_destroy_sync(&ndsda, tag, tx);
45d1cae3 1946 }
34dc7c2f
BB		 1947}
1948
1949static int
1950dsl_dataset_snapshot_reserve_space(dsl_dataset_t *ds, dmu_tx_t *tx)
1951{
1952 uint64_t asize;
1953
1954 if (!dmu_tx_is_syncing(tx))
1955 return (0);
1956
1957 /*
1958 * If there's an fs-only reservation, any blocks that might become
1959 * owned by the snapshot dataset must be accommodated by space
1960 * outside of the reservation.
1961 */
428870ff
BB		 1962 ASSERT(ds->ds_reserved == 0 || DS_UNIQUE_IS_ACCURATE(ds));
1963 asize = MIN(ds->ds_phys->ds_unique_bytes, ds->ds_reserved);
572e2857 1964 if (asize > dsl_dir_space_available(ds->ds_dir, NULL, 0, TRUE))
34dc7c2f
BB		 1965 return (ENOSPC);
1966
1967 /*
1968 * Propogate any reserved space for this snapshot to other
1969 * snapshot checks in this sync group.
1970 */
1971 if (asize > 0)
1972 dsl_dir_willuse_space(ds->ds_dir, asize, tx);
1973
1974 return (0);
1975}
1976
34dc7c2f
BB		 1977int
1978dsl_dataset_snapshot_check(void *arg1, void *arg2, dmu_tx_t *tx)
1979{
1980 dsl_dataset_t *ds = arg1;
1981 const char *snapname = arg2;
34dc7c2f
BB		 1982 int err;
1983 uint64_t value;
1984
1985 /*
1986 * We don't allow multiple snapshots of the same txg. If there
1987 * is already one, try again.
1988 */
1989 if (ds->ds_phys->ds_prev_snap_txg >= tx->tx_txg)
1990 return (EAGAIN);
1991
1992 /*
1993 * Check for conflicting name snapshot name.
1994 */
b128c09f 1995 err = dsl_dataset_snap_lookup(ds, snapname, &value);
34dc7c2f
BB		 1996 if (err == 0)
1997 return (EEXIST);
1998 if (err != ENOENT)
1999 return (err);
2000
2001 /*
2002 * Check that the dataset's name is not too long. Name consists
2003 * of the dataset's length + 1 for the @-sign + snapshot name's length
2004 */
2005 if (dsl_dataset_namelen(ds) + 1 + strlen(snapname) >= MAXNAMELEN)
2006 return (ENAMETOOLONG);
2007
2008 err = dsl_dataset_snapshot_reserve_space(ds, tx);
2009 if (err)
2010 return (err);
2011
2012 ds->ds_trysnap_txg = tx->tx_txg;
2013 return (0);
2014}
2015
2016void
428870ff 2017dsl_dataset_snapshot_sync(void *arg1, void *arg2, dmu_tx_t *tx)
34dc7c2f
BB		 2018{
2019 dsl_dataset_t *ds = arg1;
2020 const char *snapname = arg2;
2021 dsl_pool_t *dp = ds->ds_dir->dd_pool;
2022 dmu_buf_t *dbuf;
2023 dsl_dataset_phys_t *dsphys;
b128c09f 2024 uint64_t dsobj, crtxg;
34dc7c2f
BB		 2025 objset_t *mos = dp->dp_meta_objset;
2026 int err;
2027
34dc7c2f
BB		 2028 ASSERT(RW_WRITE_HELD(&dp->dp_config_rwlock));
2029
b128c09f
BB		 2030 /*
2031 * The origin's ds_creation_txg has to be < TXG_INITIAL
2032 */
2033 if (strcmp(snapname, ORIGIN_DIR_NAME) == 0)
2034 crtxg = 1;
2035 else
2036 crtxg = tx->tx_txg;
2037
34dc7c2f
BB		 2038 dsobj = dmu_object_alloc(mos, DMU_OT_DSL_DATASET, 0,
2039 DMU_OT_DSL_DATASET, sizeof (dsl_dataset_phys_t), tx);
2040 VERIFY(0 == dmu_bonus_hold(mos, dsobj, FTAG, &dbuf));
2041 dmu_buf_will_dirty(dbuf, tx);
2042 dsphys = dbuf->db_data;
b128c09f 2043 bzero(dsphys, sizeof (dsl_dataset_phys_t));
34dc7c2f
BB		 2044 dsphys->ds_dir_obj = ds->ds_dir->dd_object;
2045 dsphys->ds_fsid_guid = unique_create();
2046 (void) random_get_pseudo_bytes((void*)&dsphys->ds_guid,
2047 sizeof (dsphys->ds_guid));
2048 dsphys->ds_prev_snap_obj = ds->ds_phys->ds_prev_snap_obj;
2049 dsphys->ds_prev_snap_txg = ds->ds_phys->ds_prev_snap_txg;
2050 dsphys->ds_next_snap_obj = ds->ds_object;
2051 dsphys->ds_num_children = 1;
2052 dsphys->ds_creation_time = gethrestime_sec();
b128c09f 2053 dsphys->ds_creation_txg = crtxg;
34dc7c2f
BB		 2054 dsphys->ds_deadlist_obj = ds->ds_phys->ds_deadlist_obj;
2055 dsphys->ds_used_bytes = ds->ds_phys->ds_used_bytes;
2056 dsphys->ds_compressed_bytes = ds->ds_phys->ds_compressed_bytes;
2057 dsphys->ds_uncompressed_bytes = ds->ds_phys->ds_uncompressed_bytes;
2058 dsphys->ds_flags = ds->ds_phys->ds_flags;
2059 dsphys->ds_bp = ds->ds_phys->ds_bp;
2060 dmu_buf_rele(dbuf, FTAG);
2061
2062 ASSERT3U(ds->ds_prev != 0, ==, ds->ds_phys->ds_prev_snap_obj != 0);
2063 if (ds->ds_prev) {
b128c09f
BB		 2064 uint64_t next_clones_obj =
2065 ds->ds_prev->ds_phys->ds_next_clones_obj;
34dc7c2f
BB		 2066 ASSERT(ds->ds_prev->ds_phys->ds_next_snap_obj ==
2067 ds->ds_object ||
2068 ds->ds_prev->ds_phys->ds_num_children > 1);
2069 if (ds->ds_prev->ds_phys->ds_next_snap_obj == ds->ds_object) {
2070 dmu_buf_will_dirty(ds->ds_prev->ds_dbuf, tx);
2071 ASSERT3U(ds->ds_phys->ds_prev_snap_txg, ==,
2072 ds->ds_prev->ds_phys->ds_creation_txg);
2073 ds->ds_prev->ds_phys->ds_next_snap_obj = dsobj;
b128c09f 2074 } else if (next_clones_obj != 0) {
428870ff
BB		 2075 remove_from_next_clones(ds->ds_prev,
2076 dsphys->ds_next_snap_obj, tx);
b128c09f
BB		 2077 VERIFY3U(0, ==, zap_add_int(mos,
2078 next_clones_obj, dsobj, tx));
34dc7c2f
BB		 2079 }
2080 }
2081
2082 /*
2083 * If we have a reference-reservation on this dataset, we will
2084 * need to increase the amount of refreservation being charged
2085 * since our unique space is going to zero.
2086 */
2087 if (ds->ds_reserved) {
428870ff
BB		 2088 int64_t delta;
2089 ASSERT(DS_UNIQUE_IS_ACCURATE(ds));
2090 delta = MIN(ds->ds_phys->ds_unique_bytes, ds->ds_reserved);
b128c09f 2091 dsl_dir_diduse_space(ds->ds_dir, DD_USED_REFRSRV,
428870ff 2092 delta, 0, 0, tx);
34dc7c2f
BB		 2093 }
2094
34dc7c2f 2095 dmu_buf_will_dirty(ds->ds_dbuf, tx);
428870ff
BB		 2096 zfs_dbgmsg("taking snapshot %s@%s/%llu; newkey=%llu",
2097 ds->ds_dir->dd_myname, snapname, dsobj,
2098 ds->ds_phys->ds_prev_snap_txg);
2099 ds->ds_phys->ds_deadlist_obj = dsl_deadlist_clone(&ds->ds_deadlist,
2100 UINT64_MAX, ds->ds_phys->ds_prev_snap_obj, tx);
2101 dsl_deadlist_close(&ds->ds_deadlist);
2102 dsl_deadlist_open(&ds->ds_deadlist, mos, ds->ds_phys->ds_deadlist_obj);
2103 dsl_deadlist_add_key(&ds->ds_deadlist,
2104 ds->ds_phys->ds_prev_snap_txg, tx);
2105
34dc7c2f
BB		 2106 ASSERT3U(ds->ds_phys->ds_prev_snap_txg, <, tx->tx_txg);
2107 ds->ds_phys->ds_prev_snap_obj = dsobj;
b128c09f 2108 ds->ds_phys->ds_prev_snap_txg = crtxg;
34dc7c2f
BB		 2109 ds->ds_phys->ds_unique_bytes = 0;
2110 if (spa_version(dp->dp_spa) >= SPA_VERSION_UNIQUE_ACCURATE)
2111 ds->ds_phys->ds_flags |= DS_FLAG_UNIQUE_ACCURATE;
34dc7c2f 2112
34dc7c2f
BB		 2113 err = zap_add(mos, ds->ds_phys->ds_snapnames_zapobj,
2114 snapname, 8, 1, &dsobj, tx);
2115 ASSERT(err == 0);
2116
2117 if (ds->ds_prev)
b128c09f
BB		 2118 dsl_dataset_drop_ref(ds->ds_prev, ds);
2119 VERIFY(0 == dsl_dataset_get_ref(dp,
2120 ds->ds_phys->ds_prev_snap_obj, ds, &ds->ds_prev));
2121
428870ff
BB		 2122 dsl_scan_ds_snapshotted(ds, tx);
2123
2124 dsl_dir_snap_cmtime_update(ds->ds_dir);
34dc7c2f 2125
428870ff 2126 spa_history_log_internal(LOG_DS_SNAPSHOT, dp->dp_spa, tx,
34dc7c2f
BB		 2127 "dataset = %llu", dsobj);
2128}
2129
2130void
2131dsl_dataset_sync(dsl_dataset_t *ds, zio_t *zio, dmu_tx_t *tx)
2132{
2133 ASSERT(dmu_tx_is_syncing(tx));
428870ff 2134 ASSERT(ds->ds_objset != NULL);
34dc7c2f
BB		 2135 ASSERT(ds->ds_phys->ds_next_snap_obj == 0);
2136
2137 /*
2138 * in case we had to change ds_fsid_guid when we opened it,
2139 * sync it out now.
2140 */
2141 dmu_buf_will_dirty(ds->ds_dbuf, tx);
2142 ds->ds_phys->ds_fsid_guid = ds->ds_fsid_guid;
2143
2144 dsl_dir_dirty(ds->ds_dir, tx);
428870ff 2145 dmu_objset_sync(ds->ds_objset, zio, tx);
34dc7c2f
BB		 2146}
2147
2148void
2149dsl_dataset_stats(dsl_dataset_t *ds, nvlist_t *nv)
2150{
2151 uint64_t refd, avail, uobjs, aobjs;
2152
2153 dsl_dir_stats(ds->ds_dir, nv);
2154
2155 dsl_dataset_space(ds, &refd, &avail, &uobjs, &aobjs);
2156 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_AVAILABLE, avail);
2157 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_REFERENCED, refd);
2158
2159 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_CREATION,
2160 ds->ds_phys->ds_creation_time);
2161 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_CREATETXG,
2162 ds->ds_phys->ds_creation_txg);
2163 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_REFQUOTA,
2164 ds->ds_quota);
2165 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_REFRESERVATION,
2166 ds->ds_reserved);
b128c09f
BB		 2167 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_GUID,
2168 ds->ds_phys->ds_guid);
428870ff
BB		 2169 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_UNIQUE,
2170 ds->ds_phys->ds_unique_bytes);
2171 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_OBJSETID,
2172 ds->ds_object);
2173 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_USERREFS,
2174 ds->ds_userrefs);
45d1cae3
BB		 2175 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_DEFER_DESTROY,
2176 DS_IS_DEFER_DESTROY(ds) ? 1 : 0);
34dc7c2f
BB		 2177
2178 if (ds->ds_phys->ds_next_snap_obj) {
2179 /*
2180 * This is a snapshot; override the dd's space used with
2181 * our unique space and compression ratio.
2182 */
2183 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_USED,
2184 ds->ds_phys->ds_unique_bytes);
2185 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_COMPRESSRATIO,
2186 ds->ds_phys->ds_compressed_bytes == 0 ? 100 :
2187 (ds->ds_phys->ds_uncompressed_bytes * 100 /
2188 ds->ds_phys->ds_compressed_bytes));
2189 }
2190}
2191
2192void
2193dsl_dataset_fast_stat(dsl_dataset_t *ds, dmu_objset_stats_t *stat)
2194{
2195 stat->dds_creation_txg = ds->ds_phys->ds_creation_txg;
2196 stat->dds_inconsistent = ds->ds_phys->ds_flags & DS_FLAG_INCONSISTENT;
2197 stat->dds_guid = ds->ds_phys->ds_guid;
2198 if (ds->ds_phys->ds_next_snap_obj) {
2199 stat->dds_is_snapshot = B_TRUE;
2200 stat->dds_num_clones = ds->ds_phys->ds_num_children - 1;
fb5f0bc8
BB		 2201 } else {
2202 stat->dds_is_snapshot = B_FALSE;
2203 stat->dds_num_clones = 0;
34dc7c2f
BB		 2204 }
2205
2206 /* clone origin is really a dsl_dir thing... */
2207 rw_enter(&ds->ds_dir->dd_pool->dp_config_rwlock, RW_READER);
b128c09f 2208 if (dsl_dir_is_clone(ds->ds_dir)) {
34dc7c2f
BB		 2209 dsl_dataset_t *ods;
2210
b128c09f
BB		 2211 VERIFY(0 == dsl_dataset_get_ref(ds->ds_dir->dd_pool,
2212 ds->ds_dir->dd_phys->dd_origin_obj, FTAG, &ods));
34dc7c2f 2213 dsl_dataset_name(ods, stat->dds_origin);
b128c09f 2214 dsl_dataset_drop_ref(ods, FTAG);
fb5f0bc8
BB		 2215 } else {
2216 stat->dds_origin[0] = '\0';
34dc7c2f
BB		 2217 }
2218 rw_exit(&ds->ds_dir->dd_pool->dp_config_rwlock);
2219}
2220
2221uint64_t
2222dsl_dataset_fsid_guid(dsl_dataset_t *ds)
2223{
2224 return (ds->ds_fsid_guid);
2225}
2226
2227void
2228dsl_dataset_space(dsl_dataset_t *ds,
2229 uint64_t *refdbytesp, uint64_t *availbytesp,
2230 uint64_t *usedobjsp, uint64_t *availobjsp)
2231{
2232 *refdbytesp = ds->ds_phys->ds_used_bytes;
2233 *availbytesp = dsl_dir_space_available(ds->ds_dir, NULL, 0, TRUE);
2234 if (ds->ds_reserved > ds->ds_phys->ds_unique_bytes)
2235 *availbytesp += ds->ds_reserved - ds->ds_phys->ds_unique_bytes;
2236 if (ds->ds_quota != 0) {
2237 /*
2238 * Adjust available bytes according to refquota
2239 */
2240 if (*refdbytesp < ds->ds_quota)
2241 *availbytesp = MIN(*availbytesp,
2242 ds->ds_quota - *refdbytesp);
2243 else
2244 *availbytesp = 0;
2245 }
2246 *usedobjsp = ds->ds_phys->ds_bp.blk_fill;
2247 *availobjsp = DN_MAX_OBJECT - *usedobjsp;
2248}
2249
2250boolean_t
2251dsl_dataset_modified_since_lastsnap(dsl_dataset_t *ds)
2252{
1fde1e37 2253 ASSERTV(dsl_pool_t *dp = ds->ds_dir->dd_pool);
34dc7c2f
BB		 2254
2255 ASSERT(RW_LOCK_HELD(&dp->dp_config_rwlock) ||
2256 dsl_pool_sync_context(dp));
2257 if (ds->ds_prev == NULL)
2258 return (B_FALSE);
2259 if (ds->ds_phys->ds_bp.blk_birth >
572e2857
BB		 2260 ds->ds_prev->ds_phys->ds_creation_txg) {
2261 objset_t *os, *os_prev;
2262 /*
2263 * It may be that only the ZIL differs, because it was
2264 * reset in the head. Don't count that as being
2265 * modified.
2266 */
2267 if (dmu_objset_from_ds(ds, &os) != 0)
2268 return (B_TRUE);
2269 if (dmu_objset_from_ds(ds->ds_prev, &os_prev) != 0)
2270 return (B_TRUE);
2271 return (bcmp(&os->os_phys->os_meta_dnode,
2272 &os_prev->os_phys->os_meta_dnode,
2273 sizeof (os->os_phys->os_meta_dnode)) != 0);
2274 }
34dc7c2f
BB		 2275 return (B_FALSE);
2276}
2277
2278/* ARGSUSED */
2279static int
2280dsl_dataset_snapshot_rename_check(void *arg1, void *arg2, dmu_tx_t *tx)
2281{
2282 dsl_dataset_t *ds = arg1;
2283 char *newsnapname = arg2;
2284 dsl_dir_t *dd = ds->ds_dir;
34dc7c2f
BB		 2285 dsl_dataset_t *hds;
2286 uint64_t val;
2287 int err;
2288
b128c09f
BB		 2289 err = dsl_dataset_hold_obj(dd->dd_pool,
2290 dd->dd_phys->dd_head_dataset_obj, FTAG, &hds);
34dc7c2f
BB		 2291 if (err)
2292 return (err);
2293
2294 /* new name better not be in use */
b128c09f
BB		 2295 err = dsl_dataset_snap_lookup(hds, newsnapname, &val);
2296 dsl_dataset_rele(hds, FTAG);
34dc7c2f
BB		 2297
2298 if (err == 0)
2299 err = EEXIST;
2300 else if (err == ENOENT)
2301 err = 0;
2302
2303 /* dataset name + 1 for the "@" + the new snapshot name must fit */
2304 if (dsl_dir_namelen(ds->ds_dir) + 1 + strlen(newsnapname) >= MAXNAMELEN)
2305 err = ENAMETOOLONG;
2306
2307 return (err);
2308}
2309
2310static void
428870ff 2311dsl_dataset_snapshot_rename_sync(void *arg1, void *arg2, dmu_tx_t *tx)
34dc7c2f
BB		 2312{
2313 dsl_dataset_t *ds = arg1;
2314 const char *newsnapname = arg2;
2315 dsl_dir_t *dd = ds->ds_dir;
2316 objset_t *mos = dd->dd_pool->dp_meta_objset;
2317 dsl_dataset_t *hds;
2318 int err;
2319
2320 ASSERT(ds->ds_phys->ds_next_snap_obj != 0);
2321
b128c09f
BB		 2322 VERIFY(0 == dsl_dataset_hold_obj(dd->dd_pool,
2323 dd->dd_phys->dd_head_dataset_obj, FTAG, &hds));
34dc7c2f
BB		 2324
2325 VERIFY(0 == dsl_dataset_get_snapname(ds));
b128c09f 2326 err = dsl_dataset_snap_remove(hds, ds->ds_snapname, tx);
34dc7c2f
BB		 2327 ASSERT3U(err, ==, 0);
2328 mutex_enter(&ds->ds_lock);
2329 (void) strcpy(ds->ds_snapname, newsnapname);
2330 mutex_exit(&ds->ds_lock);
2331 err = zap_add(mos, hds->ds_phys->ds_snapnames_zapobj,
2332 ds->ds_snapname, 8, 1, &ds->ds_object, tx);
2333 ASSERT3U(err, ==, 0);
2334
428870ff
BB		 2335 spa_history_log_internal(LOG_DS_RENAME, dd->dd_pool->dp_spa, tx,
2336 "dataset = %llu", ds->ds_object);
b128c09f 2337 dsl_dataset_rele(hds, FTAG);
34dc7c2f
BB		 2338}
2339
2340struct renamesnaparg {
2341 dsl_sync_task_group_t *dstg;
2342 char failed[MAXPATHLEN];
2343 char *oldsnap;
2344 char *newsnap;
2345};
2346
2347static int
428870ff 2348dsl_snapshot_rename_one(const char *name, void *arg)
34dc7c2f
BB		 2349{
2350 struct renamesnaparg *ra = arg;
2351 dsl_dataset_t *ds = NULL;
428870ff 2352 char *snapname;
34dc7c2f
BB		 2353 int err;
2354
428870ff
BB		 2355 snapname = kmem_asprintf("%s@%s", name, ra->oldsnap);
2356 (void) strlcpy(ra->failed, snapname, sizeof (ra->failed));
34dc7c2f
BB		 2357
2358 /*
2359 * For recursive snapshot renames the parent won't be changing
2360 * so we just pass name for both the to/from argument.
2361 */
428870ff
BB		 2362 err = zfs_secpolicy_rename_perms(snapname, snapname, CRED());
2363 if (err != 0) {
2364 strfree(snapname);
2365 return (err == ENOENT ? 0 : err);
34dc7c2f
BB		 2366 }
2367
2368#ifdef _KERNEL
b128c09f
BB		 2369 /*
2370 * For all filesystems undergoing rename, we'll need to unmount it.
2371 */
428870ff 2372 (void) zfs_unmount_snap(snapname, NULL);
34dc7c2f 2373#endif
428870ff
BB		 2374 err = dsl_dataset_hold(snapname, ra->dstg, &ds);
2375 strfree(snapname);
2376 if (err != 0)
2377 return (err == ENOENT ? 0 : err);
34dc7c2f
BB		 2378
2379 dsl_sync_task_create(ra->dstg, dsl_dataset_snapshot_rename_check,
2380 dsl_dataset_snapshot_rename_sync, ds, ra->newsnap, 0);
2381
2382 return (0);
2383}
2384
2385static int
2386dsl_recursive_rename(char *oldname, const char *newname)
2387{
2388 int err;
2389 struct renamesnaparg *ra;
2390 dsl_sync_task_t *dst;
2391 spa_t *spa;
2392 char *cp, *fsname = spa_strdup(oldname);
428870ff 2393 int len = strlen(oldname) + 1;
34dc7c2f
BB		 2394
2395 /* truncate the snapshot name to get the fsname */
2396 cp = strchr(fsname, '@');
2397 *cp = '\0';
2398
2399 err = spa_open(fsname, &spa, FTAG);
2400 if (err) {
428870ff 2401 kmem_free(fsname, len);
34dc7c2f
BB		 2402 return (err);
2403 }
2404 ra = kmem_alloc(sizeof (struct renamesnaparg), KM_SLEEP);
2405 ra->dstg = dsl_sync_task_group_create(spa_get_dsl(spa));
2406
2407 ra->oldsnap = strchr(oldname, '@') + 1;
2408 ra->newsnap = strchr(newname, '@') + 1;
2409 *ra->failed = '\0';
2410
2411 err = dmu_objset_find(fsname, dsl_snapshot_rename_one, ra,
2412 DS_FIND_CHILDREN);
428870ff 2413 kmem_free(fsname, len);
34dc7c2f
BB		 2414
2415 if (err == 0) {
2416 err = dsl_sync_task_group_wait(ra->dstg);
2417 }
2418
2419 for (dst = list_head(&ra->dstg->dstg_tasks); dst;
2420 dst = list_next(&ra->dstg->dstg_tasks, dst)) {
2421 dsl_dataset_t *ds = dst->dst_arg1;
2422 if (dst->dst_err) {
2423 dsl_dir_name(ds->ds_dir, ra->failed);
428870ff
BB		 2424 (void) strlcat(ra->failed, "@", sizeof (ra->failed));
2425 (void) strlcat(ra->failed, ra->newsnap,
2426 sizeof (ra->failed));
34dc7c2f 2427 }
b128c09f 2428 dsl_dataset_rele(ds, ra->dstg);
34dc7c2f
BB		 2429 }
2430
2431 if (err)
428870ff 2432 (void) strlcpy(oldname, ra->failed, sizeof (ra->failed));
34dc7c2f
BB		 2433
2434 dsl_sync_task_group_destroy(ra->dstg);
2435 kmem_free(ra, sizeof (struct renamesnaparg));
2436 spa_close(spa, FTAG);
2437 return (err);
2438}
2439
2440static int
428870ff 2441dsl_valid_rename(const char *oldname, void *arg)
34dc7c2f
BB		 2442{
2443 int delta = *(int *)arg;
2444
2445 if (strlen(oldname) + delta >= MAXNAMELEN)
2446 return (ENAMETOOLONG);
2447
2448 return (0);
2449}
2450
2451#pragma weak dmu_objset_rename = dsl_dataset_rename
2452int
b128c09f 2453dsl_dataset_rename(char *oldname, const char *newname, boolean_t recursive)
34dc7c2f
BB		 2454{
2455 dsl_dir_t *dd;
2456 dsl_dataset_t *ds;
2457 const char *tail;
2458 int err;
2459
2460 err = dsl_dir_open(oldname, FTAG, &dd, &tail);
2461 if (err)
2462 return (err);
428870ff 2463
34dc7c2f
BB		 2464 if (tail == NULL) {
2465 int delta = strlen(newname) - strlen(oldname);
2466
b128c09f 2467 /* if we're growing, validate child name lengths */
34dc7c2f
BB		 2468 if (delta > 0)
2469 err = dmu_objset_find(oldname, dsl_valid_rename,
2470 &delta, DS_FIND_CHILDREN | DS_FIND_SNAPSHOTS);
2471
428870ff 2472 if (err == 0)
34dc7c2f
BB		 2473 err = dsl_dir_rename(dd, newname);
2474 dsl_dir_close(dd, FTAG);
2475 return (err);
2476 }
428870ff 2477
34dc7c2f 2478 if (tail[0] != '@') {
428870ff 2479 /* the name ended in a nonexistent component */
34dc7c2f
BB		 2480 dsl_dir_close(dd, FTAG);
2481 return (ENOENT);
2482 }
2483
2484 dsl_dir_close(dd, FTAG);
2485
2486 /* new name must be snapshot in same filesystem */
2487 tail = strchr(newname, '@');
2488 if (tail == NULL)
2489 return (EINVAL);
2490 tail++;
2491 if (strncmp(oldname, newname, tail - newname) != 0)
2492 return (EXDEV);
2493
2494 if (recursive) {
2495 err = dsl_recursive_rename(oldname, newname);
2496 } else {
b128c09f 2497 err = dsl_dataset_hold(oldname, FTAG, &ds);
34dc7c2f
BB		 2498 if (err)
2499 return (err);
2500
2501 err = dsl_sync_task_do(ds->ds_dir->dd_pool,
2502 dsl_dataset_snapshot_rename_check,
2503 dsl_dataset_snapshot_rename_sync, ds, (char *)tail, 1);
2504
b128c09f 2505 dsl_dataset_rele(ds, FTAG);
34dc7c2f
BB		 2506 }
2507
2508 return (err);
2509}
2510
b128c09f
BB		 2511struct promotenode {
2512 list_node_t link;
2513 dsl_dataset_t *ds;
2514};
2515
34dc7c2f 2516struct promotearg {
b128c09f 2517 list_t shared_snaps, origin_snaps, clone_snaps;
428870ff 2518 dsl_dataset_t *origin_origin;
b128c09f 2519 uint64_t used, comp, uncomp, unique, cloneusedsnap, originusedsnap;
428870ff 2520 char *err_ds;
34dc7c2f
BB		 2521};
2522
b128c09f
BB		 2523static int snaplist_space(list_t *l, uint64_t mintxg, uint64_t *spacep);
2524
34dc7c2f
BB		 2525static int
2526dsl_dataset_promote_check(void *arg1, void *arg2, dmu_tx_t *tx)
2527{
2528 dsl_dataset_t *hds = arg1;
2529 struct promotearg *pa = arg2;
b128c09f
BB		 2530 struct promotenode *snap = list_head(&pa->shared_snaps);
2531 dsl_dataset_t *origin_ds = snap->ds;
34dc7c2f 2532 int err;
428870ff 2533 uint64_t unused;
34dc7c2f 2534
b128c09f
BB		 2535 /* Check that it is a real clone */
2536 if (!dsl_dir_is_clone(hds->ds_dir))
34dc7c2f
BB		 2537 return (EINVAL);
2538
2539 /* Since this is so expensive, don't do the preliminary check */
2540 if (!dmu_tx_is_syncing(tx))
2541 return (0);
2542
b128c09f
BB		 2543 if (hds->ds_phys->ds_flags & DS_FLAG_NOPROMOTE)
2544 return (EXDEV);
34dc7c2f
BB		 2545
2546 /* compute origin's new unique space */
b128c09f
BB		 2547 snap = list_tail(&pa->clone_snaps);
2548 ASSERT3U(snap->ds->ds_phys->ds_prev_snap_obj, ==, origin_ds->ds_object);
428870ff
BB		 2549 dsl_deadlist_space_range(&snap->ds->ds_deadlist,
2550 origin_ds->ds_phys->ds_prev_snap_txg, UINT64_MAX,
2551 &pa->unique, &unused, &unused);
34dc7c2f 2552
b128c09f
BB		 2553 /*
2554 * Walk the snapshots that we are moving
2555 *
2556 * Compute space to transfer. Consider the incremental changes
2557 * to used for each snapshot:
2558 * (my used) = (prev's used) + (blocks born) - (blocks killed)
2559 * So each snapshot gave birth to:
2560 * (blocks born) = (my used) - (prev's used) + (blocks killed)
2561 * So a sequence would look like:
2562 * (uN - u(N-1) + kN) + ... + (u1 - u0 + k1) + (u0 - 0 + k0)
2563 * Which simplifies to:
2564 * uN + kN + kN-1 + ... + k1 + k0
2565 * Note however, if we stop before we reach the ORIGIN we get:
2566 * uN + kN + kN-1 + ... + kM - uM-1
2567 */
2568 pa->used = origin_ds->ds_phys->ds_used_bytes;
2569 pa->comp = origin_ds->ds_phys->ds_compressed_bytes;
2570 pa->uncomp = origin_ds->ds_phys->ds_uncompressed_bytes;
2571 for (snap = list_head(&pa->shared_snaps); snap;
2572 snap = list_next(&pa->shared_snaps, snap)) {
34dc7c2f 2573 uint64_t val, dlused, dlcomp, dluncomp;
b128c09f 2574 dsl_dataset_t *ds = snap->ds;
34dc7c2f
BB		 2575
2576 /* Check that the snapshot name does not conflict */
b128c09f
BB		 2577 VERIFY(0 == dsl_dataset_get_snapname(ds));
2578 err = dsl_dataset_snap_lookup(hds, ds->ds_snapname, &val);
428870ff
BB		 2579 if (err == 0) {
2580 err = EEXIST;
2581 goto out;
2582 }
b128c09f 2583 if (err != ENOENT)
428870ff 2584 goto out;
34dc7c2f 2585
b128c09f 2586 /* The very first snapshot does not have a deadlist */
34dc7c2f 2587 if (ds->ds_phys->ds_prev_snap_obj == 0)
b128c09f 2588 continue;
34dc7c2f 2589
428870ff
BB		 2590 dsl_deadlist_space(&ds->ds_deadlist,
2591 &dlused, &dlcomp, &dluncomp);
b128c09f
BB		 2592 pa->used += dlused;
2593 pa->comp += dlcomp;
2594 pa->uncomp += dluncomp;
2595 }
34dc7c2f 2596
b128c09f
BB		 2597 /*
2598 * If we are a clone of a clone then we never reached ORIGIN,
2599 * so we need to subtract out the clone origin's used space.
2600 */
2601 if (pa->origin_origin) {
2602 pa->used -= pa->origin_origin->ds_phys->ds_used_bytes;
2603 pa->comp -= pa->origin_origin->ds_phys->ds_compressed_bytes;
2604 pa->uncomp -= pa->origin_origin->ds_phys->ds_uncompressed_bytes;
34dc7c2f
BB		 2605 }
2606
2607 /* Check that there is enough space here */
b128c09f
BB		 2608 err = dsl_dir_transfer_possible(origin_ds->ds_dir, hds->ds_dir,
2609 pa->used);
2610 if (err)
2611 return (err);
34dc7c2f 2612
b128c09f
BB		 2613 /*
2614 * Compute the amounts of space that will be used by snapshots
2615 * after the promotion (for both origin and clone). For each,
2616 * it is the amount of space that will be on all of their
2617 * deadlists (that was not born before their new origin).
2618 */
2619 if (hds->ds_dir->dd_phys->dd_flags & DD_FLAG_USED_BREAKDOWN) {
2620 uint64_t space;
2621
2622 /*
2623 * Note, typically this will not be a clone of a clone,
428870ff
BB		 2624 * so dd_origin_txg will be < TXG_INITIAL, so
2625 * these snaplist_space() -> dsl_deadlist_space_range()
b128c09f
BB		 2626 * calls will be fast because they do not have to
2627 * iterate over all bps.
2628 */
2629 snap = list_head(&pa->origin_snaps);
2630 err = snaplist_space(&pa->shared_snaps,
428870ff 2631 snap->ds->ds_dir->dd_origin_txg, &pa->cloneusedsnap);
b128c09f
BB		 2632 if (err)
2633 return (err);
2634
2635 err = snaplist_space(&pa->clone_snaps,
428870ff 2636 snap->ds->ds_dir->dd_origin_txg, &space);
b128c09f
BB		 2637 if (err)
2638 return (err);
2639 pa->cloneusedsnap += space;
2640 }
2641 if (origin_ds->ds_dir->dd_phys->dd_flags & DD_FLAG_USED_BREAKDOWN) {
2642 err = snaplist_space(&pa->origin_snaps,
2643 origin_ds->ds_phys->ds_creation_txg, &pa->originusedsnap);
2644 if (err)
2645 return (err);
2646 }
2647
2648 return (0);
428870ff
BB		 2649out:
2650 pa->err_ds = snap->ds->ds_snapname;
2651 return (err);
34dc7c2f
BB		 2652}
2653
2654static void
428870ff 2655dsl_dataset_promote_sync(void *arg1, void *arg2, dmu_tx_t *tx)
34dc7c2f
BB		 2656{
2657 dsl_dataset_t *hds = arg1;
2658 struct promotearg *pa = arg2;
b128c09f
BB		 2659 struct promotenode *snap = list_head(&pa->shared_snaps);
2660 dsl_dataset_t *origin_ds = snap->ds;
2661 dsl_dataset_t *origin_head;
34dc7c2f
BB		 2662 dsl_dir_t *dd = hds->ds_dir;
2663 dsl_pool_t *dp = hds->ds_dir->dd_pool;
2664 dsl_dir_t *odd = NULL;
b128c09f
BB		 2665 uint64_t oldnext_obj;
2666 int64_t delta;
34dc7c2f 2667
34dc7c2f
BB		 2668 ASSERT(0 == (hds->ds_phys->ds_flags & DS_FLAG_NOPROMOTE));
2669
b128c09f
BB		 2670 snap = list_head(&pa->origin_snaps);
2671 origin_head = snap->ds;
2672
34dc7c2f
BB		 2673 /*
2674 * We need to explicitly open odd, since origin_ds's dd will be
2675 * changing.
2676 */
2677 VERIFY(0 == dsl_dir_open_obj(dp, origin_ds->ds_dir->dd_object,
2678 NULL, FTAG, &odd));
2679
b128c09f
BB		 2680 /* change origin's next snap */
2681 dmu_buf_will_dirty(origin_ds->ds_dbuf, tx);
2682 oldnext_obj = origin_ds->ds_phys->ds_next_snap_obj;
2683 snap = list_tail(&pa->clone_snaps);
2684 ASSERT3U(snap->ds->ds_phys->ds_prev_snap_obj, ==, origin_ds->ds_object);
2685 origin_ds->ds_phys->ds_next_snap_obj = snap->ds->ds_object;
2686
2687 /* change the origin's next clone */
2688 if (origin_ds->ds_phys->ds_next_clones_obj) {
428870ff 2689 remove_from_next_clones(origin_ds, snap->ds->ds_object, tx);
b128c09f
BB		 2690 VERIFY3U(0, ==, zap_add_int(dp->dp_meta_objset,
2691 origin_ds->ds_phys->ds_next_clones_obj,
2692 oldnext_obj, tx));
2693 }
2694
2695 /* change origin */
2696 dmu_buf_will_dirty(dd->dd_dbuf, tx);
2697 ASSERT3U(dd->dd_phys->dd_origin_obj, ==, origin_ds->ds_object);
2698 dd->dd_phys->dd_origin_obj = odd->dd_phys->dd_origin_obj;
428870ff 2699 dd->dd_origin_txg = origin_head->ds_dir->dd_origin_txg;
b128c09f
BB		 2700 dmu_buf_will_dirty(odd->dd_dbuf, tx);
2701 odd->dd_phys->dd_origin_obj = origin_ds->ds_object;
428870ff
BB		 2702 origin_head->ds_dir->dd_origin_txg =
2703 origin_ds->ds_phys->ds_creation_txg;
2704
2705 /* change dd_clone entries */
2706 if (spa_version(dp->dp_spa) >= SPA_VERSION_DIR_CLONES) {
2707 VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset,
2708 odd->dd_phys->dd_clones, hds->ds_object, tx));
2709 VERIFY3U(0, ==, zap_add_int(dp->dp_meta_objset,
2710 pa->origin_origin->ds_dir->dd_phys->dd_clones,
2711 hds->ds_object, tx));
2712
2713 VERIFY3U(0, ==, zap_remove_int(dp->dp_meta_objset,
2714 pa->origin_origin->ds_dir->dd_phys->dd_clones,
2715 origin_head->ds_object, tx));
2716 if (dd->dd_phys->dd_clones == 0) {
2717 dd->dd_phys->dd_clones = zap_create(dp->dp_meta_objset,
2718 DMU_OT_DSL_CLONES, DMU_OT_NONE, 0, tx);
2719 }
2720 VERIFY3U(0, ==, zap_add_int(dp->dp_meta_objset,
2721 dd->dd_phys->dd_clones, origin_head->ds_object, tx));
2722
2723 }
34dc7c2f 2724
b128c09f
BB		 2725 /* move snapshots to this dir */
2726 for (snap = list_head(&pa->shared_snaps); snap;
2727 snap = list_next(&pa->shared_snaps, snap)) {
2728 dsl_dataset_t *ds = snap->ds;
2729
2730 /* unregister props as dsl_dir is changing */
428870ff
BB		 2731 if (ds->ds_objset) {
2732 dmu_objset_evict(ds->ds_objset);
2733 ds->ds_objset = NULL;
b128c09f 2734 }
34dc7c2f 2735 /* move snap name entry */
b128c09f
BB		 2736 VERIFY(0 == dsl_dataset_get_snapname(ds));
2737 VERIFY(0 == dsl_dataset_snap_remove(origin_head,
2738 ds->ds_snapname, tx));
34dc7c2f
BB		 2739 VERIFY(0 == zap_add(dp->dp_meta_objset,
2740 hds->ds_phys->ds_snapnames_zapobj, ds->ds_snapname,
2741 8, 1, &ds->ds_object, tx));
428870ff 2742
34dc7c2f
BB		 2743 /* change containing dsl_dir */
2744 dmu_buf_will_dirty(ds->ds_dbuf, tx);
2745 ASSERT3U(ds->ds_phys->ds_dir_obj, ==, odd->dd_object);
2746 ds->ds_phys->ds_dir_obj = dd->dd_object;
2747 ASSERT3P(ds->ds_dir, ==, odd);
2748 dsl_dir_close(ds->ds_dir, ds);
2749 VERIFY(0 == dsl_dir_open_obj(dp, dd->dd_object,
2750 NULL, ds, &ds->ds_dir));
2751
428870ff
BB		 2752 /* move any clone references */
2753 if (ds->ds_phys->ds_next_clones_obj &&
2754 spa_version(dp->dp_spa) >= SPA_VERSION_DIR_CLONES) {
2755 zap_cursor_t zc;
2756 zap_attribute_t za;
2757
2758 for (zap_cursor_init(&zc, dp->dp_meta_objset,
2759 ds->ds_phys->ds_next_clones_obj);
2760 zap_cursor_retrieve(&zc, &za) == 0;
2761 zap_cursor_advance(&zc)) {
2762 dsl_dataset_t *cnds;
2763 uint64_t o;
2764
2765 if (za.za_first_integer == oldnext_obj) {
2766 /*
2767 * We've already moved the
2768 * origin's reference.
2769 */
2770 continue;
2771 }
2772
2773 VERIFY3U(0, ==, dsl_dataset_hold_obj(dp,
2774 za.za_first_integer, FTAG, &cnds));
2775 o = cnds->ds_dir->dd_phys->dd_head_dataset_obj;
2776
2777 VERIFY3U(zap_remove_int(dp->dp_meta_objset,
2778 odd->dd_phys->dd_clones, o, tx), ==, 0);
2779 VERIFY3U(zap_add_int(dp->dp_meta_objset,
2780 dd->dd_phys->dd_clones, o, tx), ==, 0);
2781 dsl_dataset_rele(cnds, FTAG);
2782 }
2783 zap_cursor_fini(&zc);
2784 }
2785
34dc7c2f 2786 ASSERT3U(dsl_prop_numcb(ds), ==, 0);
34dc7c2f 2787 }
34dc7c2f 2788
b128c09f
BB		 2789 /*
2790 * Change space accounting.
2791 * Note, pa->*usedsnap and dd_used_breakdown[SNAP] will either
2792 * both be valid, or both be 0 (resulting in delta == 0). This
2793 * is true for each of {clone,origin} independently.
2794 */
34dc7c2f 2795
b128c09f
BB		 2796 delta = pa->cloneusedsnap -
2797 dd->dd_phys->dd_used_breakdown[DD_USED_SNAP];
2798 ASSERT3S(delta, >=, 0);
2799 ASSERT3U(pa->used, >=, delta);
2800 dsl_dir_diduse_space(dd, DD_USED_SNAP, delta, 0, 0, tx);
2801 dsl_dir_diduse_space(dd, DD_USED_HEAD,
2802 pa->used - delta, pa->comp, pa->uncomp, tx);
2803
2804 delta = pa->originusedsnap -
2805 odd->dd_phys->dd_used_breakdown[DD_USED_SNAP];
2806 ASSERT3S(delta, <=, 0);
2807 ASSERT3U(pa->used, >=, -delta);
2808 dsl_dir_diduse_space(odd, DD_USED_SNAP, delta, 0, 0, tx);
2809 dsl_dir_diduse_space(odd, DD_USED_HEAD,
2810 -pa->used - delta, -pa->comp, -pa->uncomp, tx);
34dc7c2f 2811
34dc7c2f
BB		 2812 origin_ds->ds_phys->ds_unique_bytes = pa->unique;
2813
2814 /* log history record */
428870ff
BB		 2815 spa_history_log_internal(LOG_DS_PROMOTE, dd->dd_pool->dp_spa, tx,
2816 "dataset = %llu", hds->ds_object);
34dc7c2f
BB		 2817
2818 dsl_dir_close(odd, FTAG);
34dc7c2f
BB		 2819}
2820
b128c09f
BB		 2821static char *snaplist_tag = "snaplist";
2822/*
2823 * Make a list of dsl_dataset_t's for the snapshots between first_obj
2824 * (exclusive) and last_obj (inclusive). The list will be in reverse
2825 * order (last_obj will be the list_head()). If first_obj == 0, do all
2826 * snapshots back to this dataset's origin.
2827 */
2828static int
2829snaplist_make(dsl_pool_t *dp, boolean_t own,
2830 uint64_t first_obj, uint64_t last_obj, list_t *l)
2831{
2832 uint64_t obj = last_obj;
2833
2834 ASSERT(RW_LOCK_HELD(&dp->dp_config_rwlock));
2835
2836 list_create(l, sizeof (struct promotenode),
2837 offsetof(struct promotenode, link));
2838
2839 while (obj != first_obj) {
2840 dsl_dataset_t *ds;
2841 struct promotenode *snap;
2842 int err;
2843
2844 if (own) {
2845 err = dsl_dataset_own_obj(dp, obj,
2846 0, snaplist_tag, &ds);
2847 if (err == 0)
2848 dsl_dataset_make_exclusive(ds, snaplist_tag);
2849 } else {
2850 err = dsl_dataset_hold_obj(dp, obj, snaplist_tag, &ds);
2851 }
2852 if (err == ENOENT) {
2853 /* lost race with snapshot destroy */
2854 struct promotenode *last = list_tail(l);
2855 ASSERT(obj != last->ds->ds_phys->ds_prev_snap_obj);
2856 obj = last->ds->ds_phys->ds_prev_snap_obj;
2857 continue;
2858 } else if (err) {
2859 return (err);
2860 }
2861
2862 if (first_obj == 0)
2863 first_obj = ds->ds_dir->dd_phys->dd_origin_obj;
2864
2865 snap = kmem_alloc(sizeof (struct promotenode), KM_SLEEP);
2866 snap->ds = ds;
2867 list_insert_tail(l, snap);
2868 obj = ds->ds_phys->ds_prev_snap_obj;
2869 }
2870
2871 return (0);
2872}
2873
2874static int
2875snaplist_space(list_t *l, uint64_t mintxg, uint64_t *spacep)
2876{
2877 struct promotenode *snap;
2878
2879 *spacep = 0;
2880 for (snap = list_head(l); snap; snap = list_next(l, snap)) {
428870ff
BB		 2881 uint64_t used, comp, uncomp;
2882 dsl_deadlist_space_range(&snap->ds->ds_deadlist,
2883 mintxg, UINT64_MAX, &used, &comp, &uncomp);
b128c09f
BB		 2884 *spacep += used;
2885 }
2886 return (0);
2887}
2888
2889static void
2890snaplist_destroy(list_t *l, boolean_t own)
2891{
2892 struct promotenode *snap;
2893
9babb374 2894 if (!l || !list_link_active(&l->list_head))
b128c09f
BB		 2895 return;
2896
2897 while ((snap = list_tail(l)) != NULL) {
2898 list_remove(l, snap);
2899 if (own)
2900 dsl_dataset_disown(snap->ds, snaplist_tag);
2901 else
2902 dsl_dataset_rele(snap->ds, snaplist_tag);
2903 kmem_free(snap, sizeof (struct promotenode));
2904 }
2905 list_destroy(l);
2906}
2907
2908/*
2909 * Promote a clone. Nomenclature note:
2910 * "clone" or "cds": the original clone which is being promoted
2911 * "origin" or "ods": the snapshot which is originally clone's origin
2912 * "origin head" or "ohds": the dataset which is the head
2913 * (filesystem/volume) for the origin
2914 * "origin origin": the origin of the origin's filesystem (typically
2915 * NULL, indicating that the clone is not a clone of a clone).
2916 */
34dc7c2f 2917int
428870ff 2918dsl_dataset_promote(const char *name, char *conflsnap)
34dc7c2f
BB		 2919{
2920 dsl_dataset_t *ds;
b128c09f
BB		 2921 dsl_dir_t *dd;
2922 dsl_pool_t *dp;
34dc7c2f 2923 dmu_object_info_t doi;
2598c001 2924 struct promotearg pa;
b128c09f
BB		 2925 struct promotenode *snap;
2926 int err;
34dc7c2f 2927
2598c001 2928 bzero(&pa, sizeof(struct promotearg));
b128c09f 2929 err = dsl_dataset_hold(name, FTAG, &ds);
34dc7c2f
BB		 2930 if (err)
2931 return (err);
b128c09f
BB		 2932 dd = ds->ds_dir;
2933 dp = dd->dd_pool;
34dc7c2f 2934
b128c09f 2935 err = dmu_object_info(dp->dp_meta_objset,
34dc7c2f
BB		 2936 ds->ds_phys->ds_snapnames_zapobj, &doi);
2937 if (err) {
b128c09f 2938 dsl_dataset_rele(ds, FTAG);
34dc7c2f
BB		 2939 return (err);
2940 }
2941
b128c09f
BB		 2942 if (dsl_dataset_is_snapshot(ds) || dd->dd_phys->dd_origin_obj == 0) {
2943 dsl_dataset_rele(ds, FTAG);
2944 return (EINVAL);
2945 }
2946
2947 /*
2948 * We are going to inherit all the snapshots taken before our
2949 * origin (i.e., our new origin will be our parent's origin).
2950 * Take ownership of them so that we can rename them into our
2951 * namespace.
2952 */
2953 rw_enter(&dp->dp_config_rwlock, RW_READER);
2954
2955 err = snaplist_make(dp, B_TRUE, 0, dd->dd_phys->dd_origin_obj,
2956 &pa.shared_snaps);
2957 if (err != 0)
2958 goto out;
2959
2960 err = snaplist_make(dp, B_FALSE, 0, ds->ds_object, &pa.clone_snaps);
2961 if (err != 0)
2962 goto out;
2963
2964 snap = list_head(&pa.shared_snaps);
2965 ASSERT3U(snap->ds->ds_object, ==, dd->dd_phys->dd_origin_obj);
2966 err = snaplist_make(dp, B_FALSE, dd->dd_phys->dd_origin_obj,
2967 snap->ds->ds_dir->dd_phys->dd_head_dataset_obj, &pa.origin_snaps);
2968 if (err != 0)
2969 goto out;
2970
428870ff
BB		 2971 if (snap->ds->ds_dir->dd_phys->dd_origin_obj != 0) {
2972 err = dsl_dataset_hold_obj(dp,
b128c09f 2973 snap->ds->ds_dir->dd_phys->dd_origin_obj,
428870ff 2974 FTAG, &pa.origin_origin);
b128c09f
BB		 2975 if (err != 0)
2976 goto out;
2977 }
2978
2979out:
2980 rw_exit(&dp->dp_config_rwlock);
2981
34dc7c2f
BB		 2982 /*
2983 * Add in 128x the snapnames zapobj size, since we will be moving
2984 * a bunch of snapnames to the promoted ds, and dirtying their
2985 * bonus buffers.
2986 */
b128c09f
BB		 2987 if (err == 0) {
2988 err = dsl_sync_task_do(dp, dsl_dataset_promote_check,
2989 dsl_dataset_promote_sync, ds, &pa,
428870ff
BB		 2990 2 + 2 * doi.doi_physical_blocks_512);
2991 if (err && pa.err_ds && conflsnap)
2992 (void) strncpy(conflsnap, pa.err_ds, MAXNAMELEN);
b128c09f
BB		 2993 }
2994
2995 snaplist_destroy(&pa.shared_snaps, B_TRUE);
2996 snaplist_destroy(&pa.clone_snaps, B_FALSE);
2997 snaplist_destroy(&pa.origin_snaps, B_FALSE);
2998 if (pa.origin_origin)
428870ff 2999 dsl_dataset_rele(pa.origin_origin, FTAG);
b128c09f 3000 dsl_dataset_rele(ds, FTAG);
34dc7c2f
BB		 3001 return (err);
3002}
3003
3004struct cloneswaparg {
3005 dsl_dataset_t *cds; /* clone dataset */
3006 dsl_dataset_t *ohds; /* origin's head dataset */
3007 boolean_t force;
3008 int64_t unused_refres_delta; /* change in unconsumed refreservation */
3009};
3010
3011/* ARGSUSED */
3012static int
3013dsl_dataset_clone_swap_check(void *arg1, void *arg2, dmu_tx_t *tx)
3014{
3015 struct cloneswaparg *csa = arg1;
3016
3017 /* they should both be heads */
3018 if (dsl_dataset_is_snapshot(csa->cds) ||
3019 dsl_dataset_is_snapshot(csa->ohds))
3020 return (EINVAL);
3021
3022 /* the branch point should be just before them */
3023 if (csa->cds->ds_prev != csa->ohds->ds_prev)
3024 return (EINVAL);
3025
428870ff
BB		 3026 /* cds should be the clone (unless they are unrelated) */
3027 if (csa->cds->ds_prev != NULL &&
3028 csa->cds->ds_prev != csa->cds->ds_dir->dd_pool->dp_origin_snap &&
3029 csa->ohds->ds_object !=
3030 csa->cds->ds_prev->ds_phys->ds_next_snap_obj)
34dc7c2f
BB		 3031 return (EINVAL);
3032
3033 /* the clone should be a child of the origin */
3034 if (csa->cds->ds_dir->dd_parent != csa->ohds->ds_dir)
3035 return (EINVAL);
3036
3037 /* ohds shouldn't be modified unless 'force' */
3038 if (!csa->force && dsl_dataset_modified_since_lastsnap(csa->ohds))
3039 return (ETXTBSY);
3040
3041 /* adjust amount of any unconsumed refreservation */
3042 csa->unused_refres_delta =
3043 (int64_t)MIN(csa->ohds->ds_reserved,
3044 csa->ohds->ds_phys->ds_unique_bytes) -
3045 (int64_t)MIN(csa->ohds->ds_reserved,
3046 csa->cds->ds_phys->ds_unique_bytes);
3047
3048 if (csa->unused_refres_delta > 0 &&
3049 csa->unused_refres_delta >
3050 dsl_dir_space_available(csa->ohds->ds_dir, NULL, 0, TRUE))
3051 return (ENOSPC);
3052
428870ff
BB		 3053 if (csa->ohds->ds_quota != 0 &&
3054 csa->cds->ds_phys->ds_unique_bytes > csa->ohds->ds_quota)
3055 return (EDQUOT);
3056
34dc7c2f
BB		 3057 return (0);
3058}
3059
3060/* ARGSUSED */
3061static void
428870ff 3062dsl_dataset_clone_swap_sync(void *arg1, void *arg2, dmu_tx_t *tx)
34dc7c2f
BB		 3063{
3064 struct cloneswaparg *csa = arg1;
3065 dsl_pool_t *dp = csa->cds->ds_dir->dd_pool;
34dc7c2f
BB		 3066
3067 ASSERT(csa->cds->ds_reserved == 0);
428870ff
BB		 3068 ASSERT(csa->ohds->ds_quota == 0 ||
3069 csa->cds->ds_phys->ds_unique_bytes <= csa->ohds->ds_quota);
34dc7c2f
BB		 3070
3071 dmu_buf_will_dirty(csa->cds->ds_dbuf, tx);
3072 dmu_buf_will_dirty(csa->ohds->ds_dbuf, tx);
34dc7c2f 3073
428870ff
BB		 3074 if (csa->cds->ds_objset != NULL) {
3075 dmu_objset_evict(csa->cds->ds_objset);
3076 csa->cds->ds_objset = NULL;
34dc7c2f
BB		 3077 }
3078
428870ff
BB		 3079 if (csa->ohds->ds_objset != NULL) {
3080 dmu_objset_evict(csa->ohds->ds_objset);
3081 csa->ohds->ds_objset = NULL;
34dc7c2f
BB		 3082 }
3083
428870ff
BB		 3084 /*
3085 * Reset origin's unique bytes, if it exists.
3086 */
3087 if (csa->cds->ds_prev) {
3088 dsl_dataset_t *origin = csa->cds->ds_prev;
3089 uint64_t comp, uncomp;
3090
3091 dmu_buf_will_dirty(origin->ds_dbuf, tx);
3092 dsl_deadlist_space_range(&csa->cds->ds_deadlist,
3093 origin->ds_phys->ds_prev_snap_txg, UINT64_MAX,
3094 &origin->ds_phys->ds_unique_bytes, &comp, &uncomp);
3095 }
34dc7c2f
BB		 3096
3097 /* swap blkptrs */
3098 {
3099 blkptr_t tmp;
3100 tmp = csa->ohds->ds_phys->ds_bp;
3101 csa->ohds->ds_phys->ds_bp = csa->cds->ds_phys->ds_bp;
3102 csa->cds->ds_phys->ds_bp = tmp;
3103 }
3104
3105 /* set dd_*_bytes */
3106 {
3107 int64_t dused, dcomp, duncomp;
3108 uint64_t cdl_used, cdl_comp, cdl_uncomp;
3109 uint64_t odl_used, odl_comp, odl_uncomp;
3110
b128c09f
BB		 3111 ASSERT3U(csa->cds->ds_dir->dd_phys->
3112 dd_used_breakdown[DD_USED_SNAP], ==, 0);
3113
428870ff
BB		 3114 dsl_deadlist_space(&csa->cds->ds_deadlist,
3115 &cdl_used, &cdl_comp, &cdl_uncomp);
3116 dsl_deadlist_space(&csa->ohds->ds_deadlist,
3117 &odl_used, &odl_comp, &odl_uncomp);
b128c09f 3118
34dc7c2f
BB		 3119 dused = csa->cds->ds_phys->ds_used_bytes + cdl_used -
3120 (csa->ohds->ds_phys->ds_used_bytes + odl_used);
3121 dcomp = csa->cds->ds_phys->ds_compressed_bytes + cdl_comp -
3122 (csa->ohds->ds_phys->ds_compressed_bytes + odl_comp);
3123 duncomp = csa->cds->ds_phys->ds_uncompressed_bytes +
3124 cdl_uncomp -
3125 (csa->ohds->ds_phys->ds_uncompressed_bytes + odl_uncomp);
3126
b128c09f 3127 dsl_dir_diduse_space(csa->ohds->ds_dir, DD_USED_HEAD,
34dc7c2f 3128 dused, dcomp, duncomp, tx);
b128c09f 3129 dsl_dir_diduse_space(csa->cds->ds_dir, DD_USED_HEAD,
34dc7c2f 3130 -dused, -dcomp, -duncomp, tx);
b128c09f
BB		 3131
3132 /*
3133 * The difference in the space used by snapshots is the
3134 * difference in snapshot space due to the head's
3135 * deadlist (since that's the only thing that's
3136 * changing that affects the snapused).
3137 */
428870ff
BB		 3138 dsl_deadlist_space_range(&csa->cds->ds_deadlist,
3139 csa->ohds->ds_dir->dd_origin_txg, UINT64_MAX,
3140 &cdl_used, &cdl_comp, &cdl_uncomp);
3141 dsl_deadlist_space_range(&csa->ohds->ds_deadlist,
3142 csa->ohds->ds_dir->dd_origin_txg, UINT64_MAX,
3143 &odl_used, &odl_comp, &odl_uncomp);
b128c09f
BB		 3144 dsl_dir_transfer_space(csa->ohds->ds_dir, cdl_used - odl_used,
3145 DD_USED_HEAD, DD_USED_SNAP, tx);
34dc7c2f
BB		 3146 }
3147
34dc7c2f
BB		 3148 /* swap ds_*_bytes */
3149 SWITCH64(csa->ohds->ds_phys->ds_used_bytes,
3150 csa->cds->ds_phys->ds_used_bytes);
3151 SWITCH64(csa->ohds->ds_phys->ds_compressed_bytes,
3152 csa->cds->ds_phys->ds_compressed_bytes);
3153 SWITCH64(csa->ohds->ds_phys->ds_uncompressed_bytes,
3154 csa->cds->ds_phys->ds_uncompressed_bytes);
3155 SWITCH64(csa->ohds->ds_phys->ds_unique_bytes,
3156 csa->cds->ds_phys->ds_unique_bytes);
3157
3158 /* apply any parent delta for change in unconsumed refreservation */
b128c09f
BB		 3159 dsl_dir_diduse_space(csa->ohds->ds_dir, DD_USED_REFRSRV,
3160 csa->unused_refres_delta, 0, 0, tx);
34dc7c2f 3161
428870ff
BB		 3162 /*
3163 * Swap deadlists.
3164 */
3165 dsl_deadlist_close(&csa->cds->ds_deadlist);
3166 dsl_deadlist_close(&csa->ohds->ds_deadlist);
34dc7c2f
BB		 3167 SWITCH64(csa->ohds->ds_phys->ds_deadlist_obj,
3168 csa->cds->ds_phys->ds_deadlist_obj);
428870ff
BB		 3169 dsl_deadlist_open(&csa->cds->ds_deadlist, dp->dp_meta_objset,
3170 csa->cds->ds_phys->ds_deadlist_obj);
3171 dsl_deadlist_open(&csa->ohds->ds_deadlist, dp->dp_meta_objset,
3172 csa->ohds->ds_phys->ds_deadlist_obj);
b128c09f 3173
428870ff 3174 dsl_scan_ds_clone_swapped(csa->ohds, csa->cds, tx);
34dc7c2f
BB		 3175}
3176
3177/*
428870ff
BB		 3178 * Swap 'clone' with its origin head datasets. Used at the end of "zfs
3179 * recv" into an existing fs to swizzle the file system to the new
3180 * version, and by "zfs rollback". Can also be used to swap two
3181 * independent head datasets if neither has any snapshots.
34dc7c2f
BB		 3182 */
3183int
3184dsl_dataset_clone_swap(dsl_dataset_t *clone, dsl_dataset_t *origin_head,
3185 boolean_t force)
3186{
3187 struct cloneswaparg csa;
b128c09f 3188 int error;
34dc7c2f 3189
b128c09f
BB		 3190 ASSERT(clone->ds_owner);
3191 ASSERT(origin_head->ds_owner);
3192retry:
572e2857
BB		 3193 /*
3194 * Need exclusive access for the swap. If we're swapping these
3195 * datasets back after an error, we already hold the locks.
3196 */
3197 if (!RW_WRITE_HELD(&clone->ds_rwlock))
3198 rw_enter(&clone->ds_rwlock, RW_WRITER);
3199 if (!RW_WRITE_HELD(&origin_head->ds_rwlock) &&
3200 !rw_tryenter(&origin_head->ds_rwlock, RW_WRITER)) {
b128c09f
BB		 3201 rw_exit(&clone->ds_rwlock);
3202 rw_enter(&origin_head->ds_rwlock, RW_WRITER);
3203 if (!rw_tryenter(&clone->ds_rwlock, RW_WRITER)) {
3204 rw_exit(&origin_head->ds_rwlock);
3205 goto retry;
3206 }
3207 }
34dc7c2f
BB		 3208 csa.cds = clone;
3209 csa.ohds = origin_head;
3210 csa.force = force;
b128c09f 3211 error = dsl_sync_task_do(clone->ds_dir->dd_pool,
34dc7c2f 3212 dsl_dataset_clone_swap_check,
b128c09f
BB		 3213 dsl_dataset_clone_swap_sync, &csa, NULL, 9);
3214 return (error);
34dc7c2f
BB		 3215}
3216
3217/*
3218 * Given a pool name and a dataset object number in that pool,
3219 * return the name of that dataset.
3220 */
3221int
3222dsl_dsobj_to_dsname(char *pname, uint64_t obj, char *buf)
3223{
3224 spa_t *spa;
3225 dsl_pool_t *dp;
b128c09f 3226 dsl_dataset_t *ds;
34dc7c2f
BB		 3227 int error;
3228
3229 if ((error = spa_open(pname, &spa, FTAG)) != 0)
3230 return (error);
3231 dp = spa_get_dsl(spa);
3232 rw_enter(&dp->dp_config_rwlock, RW_READER);
b128c09f
BB		 3233 if ((error = dsl_dataset_hold_obj(dp, obj, FTAG, &ds)) == 0) {
3234 dsl_dataset_name(ds, buf);
3235 dsl_dataset_rele(ds, FTAG);
34dc7c2f 3236 }
34dc7c2f
BB		 3237 rw_exit(&dp->dp_config_rwlock);
3238 spa_close(spa, FTAG);
3239
b128c09f 3240 return (error);
34dc7c2f
BB		 3241}
3242
3243int
3244dsl_dataset_check_quota(dsl_dataset_t *ds, boolean_t check_quota,
b128c09f 3245 uint64_t asize, uint64_t inflight, uint64_t *used, uint64_t *ref_rsrv)
34dc7c2f
BB		 3246{
3247 int error = 0;
3248
3249 ASSERT3S(asize, >, 0);
3250
3251 /*
3252 * *ref_rsrv is the portion of asize that will come from any
3253 * unconsumed refreservation space.
3254 */
3255 *ref_rsrv = 0;
3256
3257 mutex_enter(&ds->ds_lock);
3258 /*
3259 * Make a space adjustment for reserved bytes.
3260 */
3261 if (ds->ds_reserved > ds->ds_phys->ds_unique_bytes) {
3262 ASSERT3U(*used, >=,
3263 ds->ds_reserved - ds->ds_phys->ds_unique_bytes);
3264 *used -= (ds->ds_reserved - ds->ds_phys->ds_unique_bytes);
3265 *ref_rsrv =
3266 asize - MIN(asize, parent_delta(ds, asize + inflight));
3267 }
3268
3269 if (!check_quota || ds->ds_quota == 0) {
3270 mutex_exit(&ds->ds_lock);
3271 return (0);
3272 }
3273 /*
3274 * If they are requesting more space, and our current estimate
3275 * is over quota, they get to try again unless the actual
3276 * on-disk is over quota and there are no pending changes (which
3277 * may free up space for us).
3278 */
3279 if (ds->ds_phys->ds_used_bytes + inflight >= ds->ds_quota) {
3280 if (inflight > 0 || ds->ds_phys->ds_used_bytes < ds->ds_quota)
3281 error = ERESTART;
3282 else
3283 error = EDQUOT;
3284 }
3285 mutex_exit(&ds->ds_lock);
3286
3287 return (error);
3288}
3289
3290/* ARGSUSED */
3291static int
3292dsl_dataset_set_quota_check(void *arg1, void *arg2, dmu_tx_t *tx)
3293{
3294 dsl_dataset_t *ds = arg1;
428870ff
BB		 3295 dsl_prop_setarg_t *psa = arg2;
3296 int err;
34dc7c2f
BB		 3297
3298 if (spa_version(ds->ds_dir->dd_pool->dp_spa) < SPA_VERSION_REFQUOTA)
3299 return (ENOTSUP);
3300
428870ff
BB		 3301 if ((err = dsl_prop_predict_sync(ds->ds_dir, psa)) != 0)
3302 return (err);
3303
3304 if (psa->psa_effective_value == 0)
34dc7c2f
BB		 3305 return (0);
3306
428870ff
BB		 3307 if (psa->psa_effective_value < ds->ds_phys->ds_used_bytes ||
3308 psa->psa_effective_value < ds->ds_reserved)
34dc7c2f
BB		 3309 return (ENOSPC);
3310
3311 return (0);
3312}
3313
428870ff
BB		 3314extern void dsl_prop_set_sync(void *, void *, dmu_tx_t *);
3315
34dc7c2f 3316void
428870ff 3317dsl_dataset_set_quota_sync(void *arg1, void *arg2, dmu_tx_t *tx)
34dc7c2f
BB		 3318{
3319 dsl_dataset_t *ds = arg1;
428870ff
BB		 3320 dsl_prop_setarg_t *psa = arg2;
3321 uint64_t effective_value = psa->psa_effective_value;
34dc7c2f 3322
428870ff
BB		 3323 dsl_prop_set_sync(ds, psa, tx);
3324 DSL_PROP_CHECK_PREDICTION(ds->ds_dir, psa);
34dc7c2f 3325
428870ff
BB		 3326 if (ds->ds_quota != effective_value) {
3327 dmu_buf_will_dirty(ds->ds_dbuf, tx);
3328 ds->ds_quota = effective_value;
34dc7c2f 3329
428870ff
BB		 3330 spa_history_log_internal(LOG_DS_REFQUOTA,
3331 ds->ds_dir->dd_pool->dp_spa, tx, "%lld dataset = %llu ",
3332 (longlong_t)ds->ds_quota, ds->ds_object);
3333 }
34dc7c2f
BB		 3334}
3335
3336int
428870ff 3337dsl_dataset_set_quota(const char *dsname, zprop_source_t source, uint64_t quota)
34dc7c2f
BB		 3338{
3339 dsl_dataset_t *ds;
428870ff 3340 dsl_prop_setarg_t psa;
34dc7c2f
BB		 3341 int err;
3342
428870ff
BB		 3343 dsl_prop_setarg_init_uint64(&psa, "refquota", source, &quota);
3344
b128c09f 3345 err = dsl_dataset_hold(dsname, FTAG, &ds);
34dc7c2f
BB		 3346 if (err)
3347 return (err);
3348
428870ff
BB		 3349 /*
3350 * If someone removes a file, then tries to set the quota, we
3351 * want to make sure the file freeing takes effect.
3352 */
3353 txg_wait_open(ds->ds_dir->dd_pool, 0);
3354
3355 err = dsl_sync_task_do(ds->ds_dir->dd_pool,
3356 dsl_dataset_set_quota_check, dsl_dataset_set_quota_sync,
3357 ds, &psa, 0);
34dc7c2f 3358
b128c09f 3359 dsl_dataset_rele(ds, FTAG);
34dc7c2f
BB		 3360 return (err);
3361}
3362
3363static int
3364dsl_dataset_set_reservation_check(void *arg1, void *arg2, dmu_tx_t *tx)
3365{
3366 dsl_dataset_t *ds = arg1;
428870ff
BB		 3367 dsl_prop_setarg_t *psa = arg2;
3368 uint64_t effective_value;
34dc7c2f 3369 uint64_t unique;
428870ff 3370 int err;
34dc7c2f 3371
34dc7c2f
BB		 3372 if (spa_version(ds->ds_dir->dd_pool->dp_spa) <
3373 SPA_VERSION_REFRESERVATION)
3374 return (ENOTSUP);
3375
3376 if (dsl_dataset_is_snapshot(ds))
3377 return (EINVAL);
3378
428870ff
BB		 3379 if ((err = dsl_prop_predict_sync(ds->ds_dir, psa)) != 0)
3380 return (err);
3381
3382 effective_value = psa->psa_effective_value;
3383
34dc7c2f
BB		 3384 /*
3385 * If we are doing the preliminary check in open context, the
3386 * space estimates may be inaccurate.
3387 */
3388 if (!dmu_tx_is_syncing(tx))
3389 return (0);
3390
3391 mutex_enter(&ds->ds_lock);
428870ff
BB		 3392 if (!DS_UNIQUE_IS_ACCURATE(ds))
3393 dsl_dataset_recalc_head_uniq(ds);
3394 unique = ds->ds_phys->ds_unique_bytes;
34dc7c2f
BB		 3395 mutex_exit(&ds->ds_lock);
3396
428870ff
BB		 3397 if (MAX(unique, effective_value) > MAX(unique, ds->ds_reserved)) {
3398 uint64_t delta = MAX(unique, effective_value) -
d164b209
BB		 3399 MAX(unique, ds->ds_reserved);
3400
3401 if (delta > dsl_dir_space_available(ds->ds_dir, NULL, 0, TRUE))
3402 return (ENOSPC);
3403 if (ds->ds_quota > 0 &&
428870ff 3404 effective_value > ds->ds_quota)
d164b209
BB		 3405 return (ENOSPC);
3406 }
34dc7c2f
BB		 3407
3408 return (0);
3409}
3410
34dc7c2f 3411static void
428870ff 3412dsl_dataset_set_reservation_sync(void *arg1, void *arg2, dmu_tx_t *tx)
34dc7c2f
BB		 3413{
3414 dsl_dataset_t *ds = arg1;
428870ff
BB		 3415 dsl_prop_setarg_t *psa = arg2;
3416 uint64_t effective_value = psa->psa_effective_value;
34dc7c2f
BB		 3417 uint64_t unique;
3418 int64_t delta;
3419
428870ff
BB		 3420 dsl_prop_set_sync(ds, psa, tx);
3421 DSL_PROP_CHECK_PREDICTION(ds->ds_dir, psa);
3422
34dc7c2f
BB		 3423 dmu_buf_will_dirty(ds->ds_dbuf, tx);
3424
b128c09f 3425 mutex_enter(&ds->ds_dir->dd_lock);
34dc7c2f 3426 mutex_enter(&ds->ds_lock);
428870ff
BB		 3427 ASSERT(DS_UNIQUE_IS_ACCURATE(ds));
3428 unique = ds->ds_phys->ds_unique_bytes;
3429 delta = MAX(0, (int64_t)(effective_value - unique)) -
34dc7c2f 3430 MAX(0, (int64_t)(ds->ds_reserved - unique));
428870ff 3431 ds->ds_reserved = effective_value;
34dc7c2f
BB		 3432 mutex_exit(&ds->ds_lock);
3433
b128c09f
BB		 3434 dsl_dir_diduse_space(ds->ds_dir, DD_USED_REFRSRV, delta, 0, 0, tx);
3435 mutex_exit(&ds->ds_dir->dd_lock);
34dc7c2f 3436
428870ff
BB		 3437 spa_history_log_internal(LOG_DS_REFRESERV,
3438 ds->ds_dir->dd_pool->dp_spa, tx, "%lld dataset = %llu",
3439 (longlong_t)effective_value, ds->ds_object);
34dc7c2f
BB		 3440}
3441
3442int
428870ff
BB		 3443dsl_dataset_set_reservation(const char *dsname, zprop_source_t source,
3444 uint64_t reservation)
34dc7c2f
BB		 3445{
3446 dsl_dataset_t *ds;
428870ff 3447 dsl_prop_setarg_t psa;
34dc7c2f
BB		 3448 int err;
3449
428870ff
BB		 3450 dsl_prop_setarg_init_uint64(&psa, "refreservation", source,
3451 &reservation);
3452
b128c09f 3453 err = dsl_dataset_hold(dsname, FTAG, &ds);
34dc7c2f
BB		 3454 if (err)
3455 return (err);
3456
3457 err = dsl_sync_task_do(ds->ds_dir->dd_pool,
3458 dsl_dataset_set_reservation_check,
428870ff
BB		 3459 dsl_dataset_set_reservation_sync, ds, &psa, 0);
3460
b128c09f 3461 dsl_dataset_rele(ds, FTAG);
34dc7c2f
BB		 3462 return (err);
3463}
45d1cae3 3464
572e2857
BB		 3465typedef struct zfs_hold_cleanup_arg {
3466 dsl_pool_t *dp;
3467 uint64_t dsobj;
3468 char htag[MAXNAMELEN];
3469} zfs_hold_cleanup_arg_t;
3470
3471static void
3472dsl_dataset_user_release_onexit(void *arg)
3473{
3474 zfs_hold_cleanup_arg_t *ca = arg;
3475
3476 (void) dsl_dataset_user_release_tmp(ca->dp, ca->dsobj, ca->htag,
3477 B_TRUE);
3478 kmem_free(ca, sizeof (zfs_hold_cleanup_arg_t));
3479}
3480
3481void
3482dsl_register_onexit_hold_cleanup(dsl_dataset_t *ds, const char *htag,
3483 minor_t minor)
3484{
3485 zfs_hold_cleanup_arg_t *ca;
3486
3487 ca = kmem_alloc(sizeof (zfs_hold_cleanup_arg_t), KM_SLEEP);
3488 ca->dp = ds->ds_dir->dd_pool;
3489 ca->dsobj = ds->ds_object;
3490 (void) strlcpy(ca->htag, htag, sizeof (ca->htag));
3491 VERIFY3U(0, ==, zfs_onexit_add_cb(minor,
3492 dsl_dataset_user_release_onexit, ca, NULL));
3493}
428870ff
BB		 3494
3495/*
572e2857
BB		 3496 * If you add new checks here, you may need to add
3497 * additional checks to the "temporary" case in
3498 * snapshot_check() in dmu_objset.c.
428870ff 3499 */
45d1cae3
BB		 3500static int
3501dsl_dataset_user_hold_check(void *arg1, void *arg2, dmu_tx_t *tx)
3502{
3503 dsl_dataset_t *ds = arg1;
428870ff
BB		 3504 struct dsl_ds_holdarg *ha = arg2;
3505 char *htag = ha->htag;
45d1cae3
BB		 3506 objset_t *mos = ds->ds_dir->dd_pool->dp_meta_objset;
3507 int error = 0;
3508
3509 if (spa_version(ds->ds_dir->dd_pool->dp_spa) < SPA_VERSION_USERREFS)
3510 return (ENOTSUP);
3511
3512 if (!dsl_dataset_is_snapshot(ds))
3513 return (EINVAL);
3514
45d1cae3
BB		 3515 /* tags must be unique */
3516 mutex_enter(&ds->ds_lock);
3517 if (ds->ds_phys->ds_userrefs_obj) {
3518 error = zap_lookup(mos, ds->ds_phys->ds_userrefs_obj, htag,
3519 8, 1, tx);
3520 if (error == 0)
3521 error = EEXIST;
3522 else if (error == ENOENT)
3523 error = 0;
3524 }
3525 mutex_exit(&ds->ds_lock);
3526
428870ff
BB		 3527 if (error == 0 && ha->temphold &&
3528 strlen(htag) + MAX_TAG_PREFIX_LEN >= MAXNAMELEN)
3529 error = E2BIG;
3530
45d1cae3
BB		 3531 return (error);
3532}
3533
572e2857 3534void
428870ff 3535dsl_dataset_user_hold_sync(void *arg1, void *arg2, dmu_tx_t *tx)
45d1cae3
BB		 3536{
3537 dsl_dataset_t *ds = arg1;
428870ff
BB		 3538 struct dsl_ds_holdarg *ha = arg2;
3539 char *htag = ha->htag;
3540 dsl_pool_t *dp = ds->ds_dir->dd_pool;
3541 objset_t *mos = dp->dp_meta_objset;
3542 uint64_t now = gethrestime_sec();
45d1cae3
BB		 3543 uint64_t zapobj;
3544
3545 mutex_enter(&ds->ds_lock);
3546 if (ds->ds_phys->ds_userrefs_obj == 0) {
3547 /*
3548 * This is the first user hold for this dataset. Create
3549 * the userrefs zap object.
3550 */
3551 dmu_buf_will_dirty(ds->ds_dbuf, tx);
3552 zapobj = ds->ds_phys->ds_userrefs_obj =
3553 zap_create(mos, DMU_OT_USERREFS, DMU_OT_NONE, 0, tx);
3554 } else {
3555 zapobj = ds->ds_phys->ds_userrefs_obj;
3556 }
3557 ds->ds_userrefs++;
3558 mutex_exit(&ds->ds_lock);
3559
3560 VERIFY(0 == zap_add(mos, zapobj, htag, 8, 1, &now, tx));
3561
428870ff
BB		 3562 if (ha->temphold) {
3563 VERIFY(0 == dsl_pool_user_hold(dp, ds->ds_object,
3564 htag, &now, tx));
3565 }
45d1cae3 3566
428870ff
BB		 3567 spa_history_log_internal(LOG_DS_USER_HOLD,
3568 dp->dp_spa, tx, "<%s> temp = %d dataset = %llu", htag,
3569 (int)ha->temphold, ds->ds_object);
3570}
45d1cae3
BB		 3571
3572static int
428870ff 3573dsl_dataset_user_hold_one(const char *dsname, void *arg)
45d1cae3
BB		 3574{
3575 struct dsl_ds_holdarg *ha = arg;
3576 dsl_dataset_t *ds;
3577 int error;
3578 char *name;
45d1cae3
BB		 3579
3580 /* alloc a buffer to hold dsname@snapname plus terminating NULL */
428870ff 3581 name = kmem_asprintf("%s@%s", dsname, ha->snapname);
45d1cae3 3582 error = dsl_dataset_hold(name, ha->dstg, &ds);
428870ff 3583 strfree(name);
45d1cae3 3584 if (error == 0) {
428870ff 3585 ha->gotone = B_TRUE;
45d1cae3 3586 dsl_sync_task_create(ha->dstg, dsl_dataset_user_hold_check,
428870ff 3587 dsl_dataset_user_hold_sync, ds, ha, 0);
45d1cae3
BB		 3588 } else if (error == ENOENT && ha->recursive) {
3589 error = 0;
3590 } else {
428870ff 3591 (void) strlcpy(ha->failed, dsname, sizeof (ha->failed));
45d1cae3
BB		 3592 }
3593 return (error);
3594}
3595
572e2857
BB		 3596int
3597dsl_dataset_user_hold_for_send(dsl_dataset_t *ds, char *htag,
3598 boolean_t temphold)
3599{
3600 struct dsl_ds_holdarg *ha;
3601 int error;
3602
3603 ha = kmem_zalloc(sizeof (struct dsl_ds_holdarg), KM_SLEEP);
3604 ha->htag = htag;
3605 ha->temphold = temphold;
3606 error = dsl_sync_task_do(ds->ds_dir->dd_pool,
3607 dsl_dataset_user_hold_check, dsl_dataset_user_hold_sync,
3608 ds, ha, 0);
3609 kmem_free(ha, sizeof (struct dsl_ds_holdarg));
3610
3611 return (error);
3612}
3613
45d1cae3
BB		 3614int
3615dsl_dataset_user_hold(char *dsname, char *snapname, char *htag,
572e2857 3616 boolean_t recursive, boolean_t temphold, int cleanup_fd)
45d1cae3
BB		 3617{
3618 struct dsl_ds_holdarg *ha;
3619 dsl_sync_task_t *dst;
3620 spa_t *spa;
3621 int error;
572e2857
BB		 3622 minor_t minor = 0;
3623
3624 if (cleanup_fd != -1) {
3625 /* Currently we only support cleanup-on-exit of tempholds. */
3626 if (!temphold)
3627 return (EINVAL);
3628 error = zfs_onexit_fd_hold(cleanup_fd, &minor);
3629 if (error)
3630 return (error);
3631 }
45d1cae3
BB		 3632
3633 ha = kmem_zalloc(sizeof (struct dsl_ds_holdarg), KM_SLEEP);
3634
3635 (void) strlcpy(ha->failed, dsname, sizeof (ha->failed));
3636
3637 error = spa_open(dsname, &spa, FTAG);
3638 if (error) {
3639 kmem_free(ha, sizeof (struct dsl_ds_holdarg));
572e2857
BB		 3640 if (cleanup_fd != -1)
3641 zfs_onexit_fd_rele(cleanup_fd);
45d1cae3
BB		 3642 return (error);
3643 }
3644
3645 ha->dstg = dsl_sync_task_group_create(spa_get_dsl(spa));
3646 ha->htag = htag;
3647 ha->snapname = snapname;
3648 ha->recursive = recursive;
428870ff 3649 ha->temphold = temphold;
572e2857 3650
45d1cae3
BB		 3651 if (recursive) {
3652 error = dmu_objset_find(dsname, dsl_dataset_user_hold_one,
3653 ha, DS_FIND_CHILDREN);
3654 } else {
3655 error = dsl_dataset_user_hold_one(dsname, ha);
3656 }
3657 if (error == 0)
3658 error = dsl_sync_task_group_wait(ha->dstg);
3659
3660 for (dst = list_head(&ha->dstg->dstg_tasks); dst;
3661 dst = list_next(&ha->dstg->dstg_tasks, dst)) {
3662 dsl_dataset_t *ds = dst->dst_arg1;
3663
3664 if (dst->dst_err) {
3665 dsl_dataset_name(ds, ha->failed);
3666 *strchr(ha->failed, '@') = '\0';
572e2857
BB		 3667 } else if (error == 0 && minor != 0 && temphold) {
3668 /*
3669 * If this hold is to be released upon process exit,
3670 * register that action now.
3671 */
3672 dsl_register_onexit_hold_cleanup(ds, htag, minor);
45d1cae3
BB		 3673 }
3674 dsl_dataset_rele(ds, ha->dstg);
3675 }
3676
428870ff
BB		 3677 if (error == 0 && recursive && !ha->gotone)
3678 error = ENOENT;
3679
45d1cae3 3680 if (error)
428870ff 3681 (void) strlcpy(dsname, ha->failed, sizeof (ha->failed));
45d1cae3
BB		 3682
3683 dsl_sync_task_group_destroy(ha->dstg);
572e2857 3684
45d1cae3
BB		 3685 kmem_free(ha, sizeof (struct dsl_ds_holdarg));
3686 spa_close(spa, FTAG);
572e2857
BB		 3687 if (cleanup_fd != -1)
3688 zfs_onexit_fd_rele(cleanup_fd);
45d1cae3
BB		 3689 return (error);
3690}
3691
3692struct dsl_ds_releasearg {
3693 dsl_dataset_t *ds;
3694 const char *htag;
3695 boolean_t own; /* do we own or just hold ds? */
3696};
3697
3698static int
3699dsl_dataset_release_might_destroy(dsl_dataset_t *ds, const char *htag,
3700 boolean_t *might_destroy)
3701{
3702 objset_t *mos = ds->ds_dir->dd_pool->dp_meta_objset;
3703 uint64_t zapobj;
3704 uint64_t tmp;
3705 int error;
3706
3707 *might_destroy = B_FALSE;
3708
3709 mutex_enter(&ds->ds_lock);
3710 zapobj = ds->ds_phys->ds_userrefs_obj;
3711 if (zapobj == 0) {
3712 /* The tag can't possibly exist */
3713 mutex_exit(&ds->ds_lock);
3714 return (ESRCH);
3715 }
3716
3717 /* Make sure the tag exists */
3718 error = zap_lookup(mos, zapobj, htag, 8, 1, &tmp);
3719 if (error) {
3720 mutex_exit(&ds->ds_lock);
3721 if (error == ENOENT)
3722 error = ESRCH;
3723 return (error);
3724 }
3725
3726 if (ds->ds_userrefs == 1 && ds->ds_phys->ds_num_children == 1 &&
3727 DS_IS_DEFER_DESTROY(ds))
3728 *might_destroy = B_TRUE;
3729
3730 mutex_exit(&ds->ds_lock);
3731 return (0);
3732}
3733
3734static int
3735dsl_dataset_user_release_check(void *arg1, void *tag, dmu_tx_t *tx)
3736{
3737 struct dsl_ds_releasearg *ra = arg1;
3738 dsl_dataset_t *ds = ra->ds;
3739 boolean_t might_destroy;
3740 int error;
3741
3742 if (spa_version(ds->ds_dir->dd_pool->dp_spa) < SPA_VERSION_USERREFS)
3743 return (ENOTSUP);
3744
3745 error = dsl_dataset_release_might_destroy(ds, ra->htag, &might_destroy);
3746 if (error)
3747 return (error);
3748
3749 if (might_destroy) {
3750 struct dsl_ds_destroyarg dsda = {0};
3751
3752 if (dmu_tx_is_syncing(tx)) {
3753 /*
3754 * If we're not prepared to remove the snapshot,
3755 * we can't allow the release to happen right now.
3756 */
3757 if (!ra->own)
3758 return (EBUSY);
45d1cae3
BB		 3759 }
3760 dsda.ds = ds;
3761 dsda.releasing = B_TRUE;
3762 return (dsl_dataset_destroy_check(&dsda, tag, tx));
3763 }
3764
3765 return (0);
3766}
3767
3768static void
428870ff 3769dsl_dataset_user_release_sync(void *arg1, void *tag, dmu_tx_t *tx)
45d1cae3
BB		 3770{
3771 struct dsl_ds_releasearg *ra = arg1;
3772 dsl_dataset_t *ds = ra->ds;
428870ff
BB		 3773 dsl_pool_t *dp = ds->ds_dir->dd_pool;
3774 objset_t *mos = dp->dp_meta_objset;
45d1cae3
BB		 3775 uint64_t zapobj;
3776 uint64_t dsobj = ds->ds_object;
3777 uint64_t refs;
428870ff
BB		 3778 int error;
3779
45d1cae3
BB		 3780 mutex_enter(&ds->ds_lock);
3781 ds->ds_userrefs--;
3782 refs = ds->ds_userrefs;
3783 mutex_exit(&ds->ds_lock);
428870ff
BB		 3784 error = dsl_pool_user_release(dp, ds->ds_object, ra->htag, tx);
3785 VERIFY(error == 0 || error == ENOENT);
45d1cae3
BB		 3786 zapobj = ds->ds_phys->ds_userrefs_obj;
3787 VERIFY(0 == zap_remove(mos, zapobj, ra->htag, tx));
3788 if (ds->ds_userrefs == 0 && ds->ds_phys->ds_num_children == 1 &&
3789 DS_IS_DEFER_DESTROY(ds)) {
3790 struct dsl_ds_destroyarg dsda = {0};
3791
3792 ASSERT(ra->own);
3793 dsda.ds = ds;
3794 dsda.releasing = B_TRUE;
3795 /* We already did the destroy_check */
428870ff 3796 dsl_dataset_destroy_sync(&dsda, tag, tx);
45d1cae3
BB		 3797 }
3798
428870ff
BB		 3799 spa_history_log_internal(LOG_DS_USER_RELEASE,
3800 dp->dp_spa, tx, "<%s> %lld dataset = %llu",
45d1cae3
BB		 3801 ra->htag, (longlong_t)refs, dsobj);
3802}
3803
3804static int
428870ff 3805dsl_dataset_user_release_one(const char *dsname, void *arg)
45d1cae3
BB		 3806{
3807 struct dsl_ds_holdarg *ha = arg;
3808 struct dsl_ds_releasearg *ra;
3809 dsl_dataset_t *ds;
3810 int error;
3811 void *dtag = ha->dstg;
3812 char *name;
45d1cae3
BB		 3813 boolean_t own = B_FALSE;
3814 boolean_t might_destroy;
3815
45d1cae3 3816 /* alloc a buffer to hold dsname@snapname, plus the terminating NULL */
428870ff 3817 name = kmem_asprintf("%s@%s", dsname, ha->snapname);
45d1cae3 3818 error = dsl_dataset_hold(name, dtag, &ds);
428870ff 3819 strfree(name);
45d1cae3
BB		 3820 if (error == ENOENT && ha->recursive)
3821 return (0);
428870ff 3822 (void) strlcpy(ha->failed, dsname, sizeof (ha->failed));
45d1cae3
BB		 3823 if (error)
3824 return (error);
3825
428870ff
BB		 3826 ha->gotone = B_TRUE;
3827
45d1cae3
BB		 3828 ASSERT(dsl_dataset_is_snapshot(ds));
3829
3830 error = dsl_dataset_release_might_destroy(ds, ha->htag, &might_destroy);
3831 if (error) {
3832 dsl_dataset_rele(ds, dtag);
3833 return (error);
3834 }
3835
3836 if (might_destroy) {
3837#ifdef _KERNEL
428870ff 3838 name = kmem_asprintf("%s@%s", dsname, ha->snapname);
45d1cae3 3839 error = zfs_unmount_snap(name, NULL);
428870ff 3840 strfree(name);
45d1cae3
BB		 3841 if (error) {
3842 dsl_dataset_rele(ds, dtag);
3843 return (error);
3844 }
3845#endif
428870ff 3846 if (!dsl_dataset_tryown(ds, B_TRUE, dtag)) {
45d1cae3
BB		 3847 dsl_dataset_rele(ds, dtag);
3848 return (EBUSY);
3849 } else {
3850 own = B_TRUE;
3851 dsl_dataset_make_exclusive(ds, dtag);
3852 }
3853 }
3854
3855 ra = kmem_alloc(sizeof (struct dsl_ds_releasearg), KM_SLEEP);
3856 ra->ds = ds;
3857 ra->htag = ha->htag;
3858 ra->own = own;
3859 dsl_sync_task_create(ha->dstg, dsl_dataset_user_release_check,
3860 dsl_dataset_user_release_sync, ra, dtag, 0);
3861
3862 return (0);
3863}
3864
3865int
3866dsl_dataset_user_release(char *dsname, char *snapname, char *htag,
3867 boolean_t recursive)
3868{
3869 struct dsl_ds_holdarg *ha;
3870 dsl_sync_task_t *dst;
3871 spa_t *spa;
3872 int error;
3873
428870ff 3874top:
45d1cae3
BB		 3875 ha = kmem_zalloc(sizeof (struct dsl_ds_holdarg), KM_SLEEP);
3876
3877 (void) strlcpy(ha->failed, dsname, sizeof (ha->failed));
3878
3879 error = spa_open(dsname, &spa, FTAG);
3880 if (error) {
3881 kmem_free(ha, sizeof (struct dsl_ds_holdarg));
3882 return (error);
3883 }
3884
3885 ha->dstg = dsl_sync_task_group_create(spa_get_dsl(spa));
3886 ha->htag = htag;
3887 ha->snapname = snapname;
3888 ha->recursive = recursive;
3889 if (recursive) {
3890 error = dmu_objset_find(dsname, dsl_dataset_user_release_one,
3891 ha, DS_FIND_CHILDREN);
3892 } else {
3893 error = dsl_dataset_user_release_one(dsname, ha);
3894 }
3895 if (error == 0)
3896 error = dsl_sync_task_group_wait(ha->dstg);
3897
3898 for (dst = list_head(&ha->dstg->dstg_tasks); dst;
3899 dst = list_next(&ha->dstg->dstg_tasks, dst)) {
3900 struct dsl_ds_releasearg *ra = dst->dst_arg1;
3901 dsl_dataset_t *ds = ra->ds;
3902
3903 if (dst->dst_err)
3904 dsl_dataset_name(ds, ha->failed);
3905
3906 if (ra->own)
3907 dsl_dataset_disown(ds, ha->dstg);
3908 else
3909 dsl_dataset_rele(ds, ha->dstg);
3910
3911 kmem_free(ra, sizeof (struct dsl_ds_releasearg));
3912 }
3913
428870ff
BB		 3914 if (error == 0 && recursive && !ha->gotone)
3915 error = ENOENT;
3916
3917 if (error && error != EBUSY)
3918 (void) strlcpy(dsname, ha->failed, sizeof (ha->failed));
45d1cae3
BB		 3919
3920 dsl_sync_task_group_destroy(ha->dstg);
3921 kmem_free(ha, sizeof (struct dsl_ds_holdarg));
3922 spa_close(spa, FTAG);
428870ff
BB		 3923
3924 /*
3925 * We can get EBUSY if we were racing with deferred destroy and
3926 * dsl_dataset_user_release_check() hadn't done the necessary
3927 * open context setup. We can also get EBUSY if we're racing
3928 * with destroy and that thread is the ds_owner. Either way
3929 * the busy condition should be transient, and we should retry
3930 * the release operation.
3931 */
3932 if (error == EBUSY)
3933 goto top;
3934
45d1cae3
BB		 3935 return (error);
3936}
3937
428870ff 3938/*
572e2857
BB		 3939 * Called at spa_load time (with retry == B_FALSE) to release a stale
3940 * temporary user hold. Also called by the onexit code (with retry == B_TRUE).
428870ff
BB		 3941 */
3942int
572e2857
BB		 3943dsl_dataset_user_release_tmp(dsl_pool_t *dp, uint64_t dsobj, char *htag,
3944 boolean_t retry)
428870ff
BB		 3945{
3946 dsl_dataset_t *ds;
3947 char *snap;
3948 char *name;
3949 int namelen;
3950 int error;
3951
572e2857
BB		 3952 do {
3953 rw_enter(&dp->dp_config_rwlock, RW_READER);
3954 error = dsl_dataset_hold_obj(dp, dsobj, FTAG, &ds);
3955 rw_exit(&dp->dp_config_rwlock);
3956 if (error)
3957 return (error);
3958 namelen = dsl_dataset_namelen(ds)+1;
3959 name = kmem_alloc(namelen, KM_SLEEP);
3960 dsl_dataset_name(ds, name);
3961 dsl_dataset_rele(ds, FTAG);
428870ff 3962
572e2857
BB		 3963 snap = strchr(name, '@');
3964 *snap = '\0';
3965 ++snap;
3966 error = dsl_dataset_user_release(name, snap, htag, B_FALSE);
3967 kmem_free(name, namelen);
3968
3969 /*
3970 * The object can't have been destroyed because we have a hold,
3971 * but it might have been renamed, resulting in ENOENT. Retry
3972 * if we've been requested to do so.
3973 *
3974 * It would be nice if we could use the dsobj all the way
3975 * through and avoid ENOENT entirely. But we might need to
3976 * unmount the snapshot, and there's currently no way to lookup
3977 * a vfsp using a ZFS object id.
3978 */
3979 } while ((error == ENOENT) && retry);
3980
3981 return (error);
428870ff
BB		 3982}
3983
45d1cae3
BB		 3984int
3985dsl_dataset_get_holds(const char *dsname, nvlist_t **nvp)
3986{
3987 dsl_dataset_t *ds;
3988 int err;
3989
3990 err = dsl_dataset_hold(dsname, FTAG, &ds);
3991 if (err)
3992 return (err);
3993
3994 VERIFY(0 == nvlist_alloc(nvp, NV_UNIQUE_NAME, KM_SLEEP));
3995 if (ds->ds_phys->ds_userrefs_obj != 0) {
3996 zap_attribute_t *za;
3997 zap_cursor_t zc;
3998
3999 za = kmem_alloc(sizeof (zap_attribute_t), KM_SLEEP);
4000 for (zap_cursor_init(&zc, ds->ds_dir->dd_pool->dp_meta_objset,
4001 ds->ds_phys->ds_userrefs_obj);
4002 zap_cursor_retrieve(&zc, za) == 0;
4003 zap_cursor_advance(&zc)) {
4004 VERIFY(0 == nvlist_add_uint64(*nvp, za->za_name,
4005 za->za_first_integer));
4006 }
4007 zap_cursor_fini(&zc);
4008 kmem_free(za, sizeof (zap_attribute_t));
4009 }
4010 dsl_dataset_rele(ds, FTAG);
4011 return (0);
4012}
428870ff
BB		 4013
4014/*
4015 * Note, this fuction is used as the callback for dmu_objset_find(). We
4016 * always return 0 so that we will continue to find and process
4017 * inconsistent datasets, even if we encounter an error trying to
4018 * process one of them.
4019 */
4020/* ARGSUSED */
4021int
4022dsl_destroy_inconsistent(const char *dsname, void *arg)
4023{
4024 dsl_dataset_t *ds;
4025
4026 if (dsl_dataset_own(dsname, B_TRUE, FTAG, &ds) == 0) {
4027 if (DS_IS_INCONSISTENT(ds))
4028 (void) dsl_dataset_destroy(ds, FTAG, B_FALSE);
4029 else
4030 dsl_dataset_disown(ds, FTAG);
4031 }
4032 return (0);
4033}
ZFS On Linux mirror for PVE