1 /* 2 * CDDL HEADER START 3 * 4 * The contents of this file are subject to the terms of the 5 * Common Development and Distribution License (the "License"). 6 * You may not use this file except in compliance with the License. 7 * 8 * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE 9 * or http://www.opensolaris.org/os/licensing. 10 * See the License for the specific language governing permissions 11 * and limitations under the License. 12 * 13 * When distributing Covered Code, include this CDDL HEADER in each 14 * file and include the License file at usr/src/OPENSOLARIS.LICENSE. 15 * If applicable, add the following below this CDDL HEADER, with the 16 * fields enclosed by brackets "[]" replaced with your own identifying 17 * information: Portions Copyright [yyyy] [name of copyright owner] 18 * 19 * CDDL HEADER END 20 */ 21 /* 22 * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. 23 * Copyright (c) 2012, 2014 by Delphix. All rights reserved. 24 * Copyright (c) 2013 by Saso Kiselkov. All rights reserved. 25 * Copyright (c) 2013, Joyent, Inc. All rights reserved. 26 * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. 27 * Copyright 2015 Nexenta Systems, Inc. All rights reserved. 28 * Copyright (c) 2015, STRATO AG, Inc. All rights reserved. 29 */ 30 31 /* Portions Copyright 2010 Robert Milkowski */ 32 33 #include <sys/cred.h> 34 #include <sys/zfs_context.h> 35 #include <sys/dmu_objset.h> 36 #include <sys/dsl_dir.h> 37 #include <sys/dsl_dataset.h> 38 #include <sys/dsl_prop.h> 39 #include <sys/dsl_pool.h> 40 #include <sys/dsl_synctask.h> 41 #include <sys/dsl_deleg.h> 42 #include <sys/dnode.h> 43 #include <sys/dbuf.h> 44 #include <sys/zvol.h> 45 #include <sys/dmu_tx.h> 46 #include <sys/zap.h> 47 #include <sys/zil.h> 48 #include <sys/dmu_impl.h> 49 #include <sys/zfs_ioctl.h> 50 #include <sys/sa.h> 51 #include <sys/zfs_onexit.h> 52 #include <sys/dsl_destroy.h> 53 #include <sys/vdev.h> 54 55 /* 56 * Needed to close a window in dnode_move() that allows the objset to be freed 57 * before it can be safely accessed. 58 */ 59 krwlock_t os_lock; 60 61 /* 62 * Tunable to overwrite the maximum number of threads for the parallization 63 * of dmu_objset_find_dp, needed to speed up the import of pools with many 64 * datasets. 65 * Default is 4 times the number of leaf vdevs. 66 */ 67 int dmu_find_threads = 0; 68 69 static void dmu_objset_find_dp_cb(void *arg); 70 71 void 72 dmu_objset_init(void) 73 { 74 rw_init(&os_lock, NULL, RW_DEFAULT, NULL); 75 } 76 77 void 78 dmu_objset_fini(void) 79 { 80 rw_destroy(&os_lock); 81 } 82 83 spa_t * 84 dmu_objset_spa(objset_t *os) 85 { 86 return (os->os_spa); 87 } 88 89 zilog_t * 90 dmu_objset_zil(objset_t *os) 91 { 92 return (os->os_zil); 93 } 94 95 dsl_pool_t * 96 dmu_objset_pool(objset_t *os) 97 { 98 dsl_dataset_t *ds; 99 100 if ((ds = os->os_dsl_dataset) != NULL && ds->ds_dir) 101 return (ds->ds_dir->dd_pool); 102 else 103 return (spa_get_dsl(os->os_spa)); 104 } 105 106 dsl_dataset_t * 107 dmu_objset_ds(objset_t *os) 108 { 109 return (os->os_dsl_dataset); 110 } 111 112 dmu_objset_type_t 113 dmu_objset_type(objset_t *os) 114 { 115 return (os->os_phys->os_type); 116 } 117 118 void 119 dmu_objset_name(objset_t *os, char *buf) 120 { 121 dsl_dataset_name(os->os_dsl_dataset, buf); 122 } 123 124 uint64_t 125 dmu_objset_id(objset_t *os) 126 { 127 dsl_dataset_t *ds = os->os_dsl_dataset; 128 129 return (ds ? ds->ds_object : 0); 130 } 131 132 zfs_sync_type_t 133 dmu_objset_syncprop(objset_t *os) 134 { 135 return (os->os_sync); 136 } 137 138 zfs_logbias_op_t 139 dmu_objset_logbias(objset_t *os) 140 { 141 return (os->os_logbias); 142 } 143 144 static void 145 checksum_changed_cb(void *arg, uint64_t newval) 146 { 147 objset_t *os = arg; 148 149 /* 150 * Inheritance should have been done by now. 151 */ 152 ASSERT(newval != ZIO_CHECKSUM_INHERIT); 153 154 os->os_checksum = zio_checksum_select(newval, ZIO_CHECKSUM_ON_VALUE); 155 } 156 157 static void 158 compression_changed_cb(void *arg, uint64_t newval) 159 { 160 objset_t *os = arg; 161 162 /* 163 * Inheritance and range checking should have been done by now. 164 */ 165 ASSERT(newval != ZIO_COMPRESS_INHERIT); 166 167 os->os_compress = zio_compress_select(os->os_spa, newval, 168 ZIO_COMPRESS_ON); 169 } 170 171 static void 172 copies_changed_cb(void *arg, uint64_t newval) 173 { 174 objset_t *os = arg; 175 176 /* 177 * Inheritance and range checking should have been done by now. 178 */ 179 ASSERT(newval > 0); 180 ASSERT(newval <= spa_max_replication(os->os_spa)); 181 182 os->os_copies = newval; 183 } 184 185 static void 186 dedup_changed_cb(void *arg, uint64_t newval) 187 { 188 objset_t *os = arg; 189 spa_t *spa = os->os_spa; 190 enum zio_checksum checksum; 191 192 /* 193 * Inheritance should have been done by now. 194 */ 195 ASSERT(newval != ZIO_CHECKSUM_INHERIT); 196 197 checksum = zio_checksum_dedup_select(spa, newval, ZIO_CHECKSUM_OFF); 198 199 os->os_dedup_checksum = checksum & ZIO_CHECKSUM_MASK; 200 os->os_dedup_verify = !!(checksum & ZIO_CHECKSUM_VERIFY); 201 } 202 203 static void 204 primary_cache_changed_cb(void *arg, uint64_t newval) 205 { 206 objset_t *os = arg; 207 208 /* 209 * Inheritance and range checking should have been done by now. 210 */ 211 ASSERT(newval == ZFS_CACHE_ALL || newval == ZFS_CACHE_NONE || 212 newval == ZFS_CACHE_METADATA); 213 214 os->os_primary_cache = newval; 215 } 216 217 static void 218 secondary_cache_changed_cb(void *arg, uint64_t newval) 219 { 220 objset_t *os = arg; 221 222 /* 223 * Inheritance and range checking should have been done by now. 224 */ 225 ASSERT(newval == ZFS_CACHE_ALL || newval == ZFS_CACHE_NONE || 226 newval == ZFS_CACHE_METADATA); 227 228 os->os_secondary_cache = newval; 229 } 230 231 static void 232 sync_changed_cb(void *arg, uint64_t newval) 233 { 234 objset_t *os = arg; 235 236 /* 237 * Inheritance and range checking should have been done by now. 238 */ 239 ASSERT(newval == ZFS_SYNC_STANDARD || newval == ZFS_SYNC_ALWAYS || 240 newval == ZFS_SYNC_DISABLED); 241 242 os->os_sync = newval; 243 if (os->os_zil) 244 zil_set_sync(os->os_zil, newval); 245 } 246 247 static void 248 redundant_metadata_changed_cb(void *arg, uint64_t newval) 249 { 250 objset_t *os = arg; 251 252 /* 253 * Inheritance and range checking should have been done by now. 254 */ 255 ASSERT(newval == ZFS_REDUNDANT_METADATA_ALL || 256 newval == ZFS_REDUNDANT_METADATA_MOST); 257 258 os->os_redundant_metadata = newval; 259 } 260 261 static void 262 logbias_changed_cb(void *arg, uint64_t newval) 263 { 264 objset_t *os = arg; 265 266 ASSERT(newval == ZFS_LOGBIAS_LATENCY || 267 newval == ZFS_LOGBIAS_THROUGHPUT); 268 os->os_logbias = newval; 269 if (os->os_zil) 270 zil_set_logbias(os->os_zil, newval); 271 } 272 273 static void 274 recordsize_changed_cb(void *arg, uint64_t newval) 275 { 276 objset_t *os = arg; 277 278 os->os_recordsize = newval; 279 } 280 281 void 282 dmu_objset_byteswap(void *buf, size_t size) 283 { 284 objset_phys_t *osp = buf; 285 286 ASSERT(size == OBJSET_OLD_PHYS_SIZE || size == sizeof (objset_phys_t)); 287 dnode_byteswap(&osp->os_meta_dnode); 288 byteswap_uint64_array(&osp->os_zil_header, sizeof (zil_header_t)); 289 osp->os_type = BSWAP_64(osp->os_type); 290 osp->os_flags = BSWAP_64(osp->os_flags); 291 if (size == sizeof (objset_phys_t)) { 292 dnode_byteswap(&osp->os_userused_dnode); 293 dnode_byteswap(&osp->os_groupused_dnode); 294 } 295 } 296 297 int 298 dmu_objset_open_impl(spa_t *spa, dsl_dataset_t *ds, blkptr_t *bp, 299 objset_t **osp) 300 { 301 objset_t *os; 302 int i, err; 303 304 ASSERT(ds == NULL || MUTEX_HELD(&ds->ds_opening_lock)); 305 306 os = kmem_zalloc(sizeof (objset_t), KM_SLEEP); 307 os->os_dsl_dataset = ds; 308 os->os_spa = spa; 309 os->os_rootbp = bp; 310 if (!BP_IS_HOLE(os->os_rootbp)) { 311 arc_flags_t aflags = ARC_FLAG_WAIT; 312 zbookmark_phys_t zb; 313 SET_BOOKMARK(&zb, ds ? ds->ds_object : DMU_META_OBJSET, 314 ZB_ROOT_OBJECT, ZB_ROOT_LEVEL, ZB_ROOT_BLKID); 315 316 if (DMU_OS_IS_L2CACHEABLE(os)) 317 aflags |= ARC_FLAG_L2CACHE; 318 if (DMU_OS_IS_L2COMPRESSIBLE(os)) 319 aflags |= ARC_FLAG_L2COMPRESS; 320 321 dprintf_bp(os->os_rootbp, "reading %s", ""); 322 err = arc_read(NULL, spa, os->os_rootbp, 323 arc_getbuf_func, &os->os_phys_buf, 324 ZIO_PRIORITY_SYNC_READ, ZIO_FLAG_CANFAIL, &aflags, &zb); 325 if (err != 0) { 326 kmem_free(os, sizeof (objset_t)); 327 /* convert checksum errors into IO errors */ 328 if (err == ECKSUM) 329 err = SET_ERROR(EIO); 330 return (err); 331 } 332 333 /* Increase the blocksize if we are permitted. */ 334 if (spa_version(spa) >= SPA_VERSION_USERSPACE && 335 arc_buf_size(os->os_phys_buf) < sizeof (objset_phys_t)) { 336 arc_buf_t *buf = arc_buf_alloc(spa, 337 sizeof (objset_phys_t), &os->os_phys_buf, 338 ARC_BUFC_METADATA); 339 bzero(buf->b_data, sizeof (objset_phys_t)); 340 bcopy(os->os_phys_buf->b_data, buf->b_data, 341 arc_buf_size(os->os_phys_buf)); 342 (void) arc_buf_remove_ref(os->os_phys_buf, 343 &os->os_phys_buf); 344 os->os_phys_buf = buf; 345 } 346 347 os->os_phys = os->os_phys_buf->b_data; 348 os->os_flags = os->os_phys->os_flags; 349 } else { 350 int size = spa_version(spa) >= SPA_VERSION_USERSPACE ? 351 sizeof (objset_phys_t) : OBJSET_OLD_PHYS_SIZE; 352 os->os_phys_buf = arc_buf_alloc(spa, size, 353 &os->os_phys_buf, ARC_BUFC_METADATA); 354 os->os_phys = os->os_phys_buf->b_data; 355 bzero(os->os_phys, size); 356 } 357 358 /* 359 * Note: the changed_cb will be called once before the register 360 * func returns, thus changing the checksum/compression from the 361 * default (fletcher2/off). Snapshots don't need to know about 362 * checksum/compression/copies. 363 */ 364 if (ds != NULL) { 365 boolean_t needlock = B_FALSE; 366 367 /* 368 * Note: it's valid to open the objset if the dataset is 369 * long-held, in which case the pool_config lock will not 370 * be held. 371 */ 372 if (!dsl_pool_config_held(dmu_objset_pool(os))) { 373 needlock = B_TRUE; 374 dsl_pool_config_enter(dmu_objset_pool(os), FTAG); 375 } 376 err = dsl_prop_register(ds, 377 zfs_prop_to_name(ZFS_PROP_PRIMARYCACHE), 378 primary_cache_changed_cb, os); 379 if (err == 0) { 380 err = dsl_prop_register(ds, 381 zfs_prop_to_name(ZFS_PROP_SECONDARYCACHE), 382 secondary_cache_changed_cb, os); 383 } 384 if (!ds->ds_is_snapshot) { 385 if (err == 0) { 386 err = dsl_prop_register(ds, 387 zfs_prop_to_name(ZFS_PROP_CHECKSUM), 388 checksum_changed_cb, os); 389 } 390 if (err == 0) { 391 err = dsl_prop_register(ds, 392 zfs_prop_to_name(ZFS_PROP_COMPRESSION), 393 compression_changed_cb, os); 394 } 395 if (err == 0) { 396 err = dsl_prop_register(ds, 397 zfs_prop_to_name(ZFS_PROP_COPIES), 398 copies_changed_cb, os); 399 } 400 if (err == 0) { 401 err = dsl_prop_register(ds, 402 zfs_prop_to_name(ZFS_PROP_DEDUP), 403 dedup_changed_cb, os); 404 } 405 if (err == 0) { 406 err = dsl_prop_register(ds, 407 zfs_prop_to_name(ZFS_PROP_LOGBIAS), 408 logbias_changed_cb, os); 409 } 410 if (err == 0) { 411 err = dsl_prop_register(ds, 412 zfs_prop_to_name(ZFS_PROP_SYNC), 413 sync_changed_cb, os); 414 } 415 if (err == 0) { 416 err = dsl_prop_register(ds, 417 zfs_prop_to_name( 418 ZFS_PROP_REDUNDANT_METADATA), 419 redundant_metadata_changed_cb, os); 420 } 421 if (err == 0) { 422 err = dsl_prop_register(ds, 423 zfs_prop_to_name(ZFS_PROP_RECORDSIZE), 424 recordsize_changed_cb, os); 425 } 426 } 427 if (needlock) 428 dsl_pool_config_exit(dmu_objset_pool(os), FTAG); 429 if (err != 0) { 430 VERIFY(arc_buf_remove_ref(os->os_phys_buf, 431 &os->os_phys_buf)); 432 kmem_free(os, sizeof (objset_t)); 433 return (err); 434 } 435 } else { 436 /* It's the meta-objset. */ 437 os->os_checksum = ZIO_CHECKSUM_FLETCHER_4; 438 os->os_compress = ZIO_COMPRESS_ON; 439 os->os_copies = spa_max_replication(spa); 440 os->os_dedup_checksum = ZIO_CHECKSUM_OFF; 441 os->os_dedup_verify = B_FALSE; 442 os->os_logbias = ZFS_LOGBIAS_LATENCY; 443 os->os_sync = ZFS_SYNC_STANDARD; 444 os->os_primary_cache = ZFS_CACHE_ALL; 445 os->os_secondary_cache = ZFS_CACHE_ALL; 446 } 447 448 if (ds == NULL || !ds->ds_is_snapshot) 449 os->os_zil_header = os->os_phys->os_zil_header; 450 os->os_zil = zil_alloc(os, &os->os_zil_header); 451 452 for (i = 0; i < TXG_SIZE; i++) { 453 list_create(&os->os_dirty_dnodes[i], sizeof (dnode_t), 454 offsetof(dnode_t, dn_dirty_link[i])); 455 list_create(&os->os_free_dnodes[i], sizeof (dnode_t), 456 offsetof(dnode_t, dn_dirty_link[i])); 457 } 458 list_create(&os->os_dnodes, sizeof (dnode_t), 459 offsetof(dnode_t, dn_link)); 460 list_create(&os->os_downgraded_dbufs, sizeof (dmu_buf_impl_t), 461 offsetof(dmu_buf_impl_t, db_link)); 462 463 mutex_init(&os->os_lock, NULL, MUTEX_DEFAULT, NULL); 464 mutex_init(&os->os_obj_lock, NULL, MUTEX_DEFAULT, NULL); 465 mutex_init(&os->os_user_ptr_lock, NULL, MUTEX_DEFAULT, NULL); 466 467 dnode_special_open(os, &os->os_phys->os_meta_dnode, 468 DMU_META_DNODE_OBJECT, &os->os_meta_dnode); 469 if (arc_buf_size(os->os_phys_buf) >= sizeof (objset_phys_t)) { 470 dnode_special_open(os, &os->os_phys->os_userused_dnode, 471 DMU_USERUSED_OBJECT, &os->os_userused_dnode); 472 dnode_special_open(os, &os->os_phys->os_groupused_dnode, 473 DMU_GROUPUSED_OBJECT, &os->os_groupused_dnode); 474 } 475 476 *osp = os; 477 return (0); 478 } 479 480 int 481 dmu_objset_from_ds(dsl_dataset_t *ds, objset_t **osp) 482 { 483 int err = 0; 484 485 /* 486 * We shouldn't be doing anything with dsl_dataset_t's unless the 487 * pool_config lock is held, or the dataset is long-held. 488 */ 489 ASSERT(dsl_pool_config_held(ds->ds_dir->dd_pool) || 490 dsl_dataset_long_held(ds)); 491 492 mutex_enter(&ds->ds_opening_lock); 493 if (ds->ds_objset == NULL) { 494 objset_t *os; 495 err = dmu_objset_open_impl(dsl_dataset_get_spa(ds), 496 ds, dsl_dataset_get_blkptr(ds), &os); 497 498 if (err == 0) { 499 mutex_enter(&ds->ds_lock); 500 ASSERT(ds->ds_objset == NULL); 501 ds->ds_objset = os; 502 mutex_exit(&ds->ds_lock); 503 } 504 } 505 *osp = ds->ds_objset; 506 mutex_exit(&ds->ds_opening_lock); 507 return (err); 508 } 509 510 /* 511 * Holds the pool while the objset is held. Therefore only one objset 512 * can be held at a time. 513 */ 514 int 515 dmu_objset_hold(const char *name, void *tag, objset_t **osp) 516 { 517 dsl_pool_t *dp; 518 dsl_dataset_t *ds; 519 int err; 520 521 err = dsl_pool_hold(name, tag, &dp); 522 if (err != 0) 523 return (err); 524 err = dsl_dataset_hold(dp, name, tag, &ds); 525 if (err != 0) { 526 dsl_pool_rele(dp, tag); 527 return (err); 528 } 529 530 err = dmu_objset_from_ds(ds, osp); 531 if (err != 0) { 532 dsl_dataset_rele(ds, tag); 533 dsl_pool_rele(dp, tag); 534 } 535 536 return (err); 537 } 538 539 static int 540 dmu_objset_own_impl(dsl_dataset_t *ds, dmu_objset_type_t type, 541 boolean_t readonly, void *tag, objset_t **osp) 542 { 543 int err; 544 545 err = dmu_objset_from_ds(ds, osp); 546 if (err != 0) { 547 dsl_dataset_disown(ds, tag); 548 } else if (type != DMU_OST_ANY && type != (*osp)->os_phys->os_type) { 549 dsl_dataset_disown(ds, tag); 550 return (SET_ERROR(EINVAL)); 551 } else if (!readonly && dsl_dataset_is_snapshot(ds)) { 552 dsl_dataset_disown(ds, tag); 553 return (SET_ERROR(EROFS)); 554 } 555 return (err); 556 } 557 558 /* 559 * dsl_pool must not be held when this is called. 560 * Upon successful return, there will be a longhold on the dataset, 561 * and the dsl_pool will not be held. 562 */ 563 int 564 dmu_objset_own(const char *name, dmu_objset_type_t type, 565 boolean_t readonly, void *tag, objset_t **osp) 566 { 567 dsl_pool_t *dp; 568 dsl_dataset_t *ds; 569 int err; 570 571 err = dsl_pool_hold(name, FTAG, &dp); 572 if (err != 0) 573 return (err); 574 err = dsl_dataset_own(dp, name, tag, &ds); 575 if (err != 0) { 576 dsl_pool_rele(dp, FTAG); 577 return (err); 578 } 579 err = dmu_objset_own_impl(ds, type, readonly, tag, osp); 580 dsl_pool_rele(dp, FTAG); 581 582 return (err); 583 } 584 585 int 586 dmu_objset_own_obj(dsl_pool_t *dp, uint64_t obj, dmu_objset_type_t type, 587 boolean_t readonly, void *tag, objset_t **osp) 588 { 589 dsl_dataset_t *ds; 590 int err; 591 592 err = dsl_dataset_own_obj(dp, obj, tag, &ds); 593 if (err != 0) 594 return (err); 595 596 return (dmu_objset_own_impl(ds, type, readonly, tag, osp)); 597 } 598 599 void 600 dmu_objset_rele(objset_t *os, void *tag) 601 { 602 dsl_pool_t *dp = dmu_objset_pool(os); 603 dsl_dataset_rele(os->os_dsl_dataset, tag); 604 dsl_pool_rele(dp, tag); 605 } 606 607 /* 608 * When we are called, os MUST refer to an objset associated with a dataset 609 * that is owned by 'tag'; that is, is held and long held by 'tag' and ds_owner 610 * == tag. We will then release and reacquire ownership of the dataset while 611 * holding the pool config_rwlock to avoid intervening namespace or ownership 612 * changes may occur. 613 * 614 * This exists solely to accommodate zfs_ioc_userspace_upgrade()'s desire to 615 * release the hold on its dataset and acquire a new one on the dataset of the 616 * same name so that it can be partially torn down and reconstructed. 617 */ 618 void 619 dmu_objset_refresh_ownership(objset_t *os, void *tag) 620 { 621 dsl_pool_t *dp; 622 dsl_dataset_t *ds, *newds; 623 char name[MAXNAMELEN]; 624 625 ds = os->os_dsl_dataset; 626 VERIFY3P(ds, !=, NULL); 627 VERIFY3P(ds->ds_owner, ==, tag); 628 VERIFY(dsl_dataset_long_held(ds)); 629 630 dsl_dataset_name(ds, name); 631 dp = dmu_objset_pool(os); 632 dsl_pool_config_enter(dp, FTAG); 633 dmu_objset_disown(os, tag); 634 VERIFY0(dsl_dataset_own(dp, name, tag, &newds)); 635 VERIFY3P(newds, ==, os->os_dsl_dataset); 636 dsl_pool_config_exit(dp, FTAG); 637 } 638 639 void 640 dmu_objset_disown(objset_t *os, void *tag) 641 { 642 dsl_dataset_disown(os->os_dsl_dataset, tag); 643 } 644 645 void 646 dmu_objset_evict_dbufs(objset_t *os) 647 { 648 dnode_t dn_marker; 649 dnode_t *dn; 650 651 mutex_enter(&os->os_lock); 652 dn = list_head(&os->os_dnodes); 653 while (dn != NULL) { 654 /* 655 * Skip dnodes without holds. We have to do this dance 656 * because dnode_add_ref() only works if there is already a 657 * hold. If the dnode has no holds, then it has no dbufs. 658 */ 659 if (dnode_add_ref(dn, FTAG)) { 660 list_insert_after(&os->os_dnodes, dn, &dn_marker); 661 mutex_exit(&os->os_lock); 662 663 dnode_evict_dbufs(dn); 664 dnode_rele(dn, FTAG); 665 666 mutex_enter(&os->os_lock); 667 dn = list_next(&os->os_dnodes, &dn_marker); 668 list_remove(&os->os_dnodes, &dn_marker); 669 } else { 670 dn = list_next(&os->os_dnodes, dn); 671 } 672 } 673 mutex_exit(&os->os_lock); 674 675 if (DMU_USERUSED_DNODE(os) != NULL) { 676 dnode_evict_dbufs(DMU_GROUPUSED_DNODE(os)); 677 dnode_evict_dbufs(DMU_USERUSED_DNODE(os)); 678 } 679 dnode_evict_dbufs(DMU_META_DNODE(os)); 680 } 681 682 /* 683 * Objset eviction processing is split into into two pieces. 684 * The first marks the objset as evicting, evicts any dbufs that 685 * have a refcount of zero, and then queues up the objset for the 686 * second phase of eviction. Once os->os_dnodes has been cleared by 687 * dnode_buf_pageout()->dnode_destroy(), the second phase is executed. 688 * The second phase closes the special dnodes, dequeues the objset from 689 * the list of those undergoing eviction, and finally frees the objset. 690 * 691 * NOTE: Due to asynchronous eviction processing (invocation of 692 * dnode_buf_pageout()), it is possible for the meta dnode for the 693 * objset to have no holds even though os->os_dnodes is not empty. 694 */ 695 void 696 dmu_objset_evict(objset_t *os) 697 { 698 dsl_dataset_t *ds = os->os_dsl_dataset; 699 700 for (int t = 0; t < TXG_SIZE; t++) 701 ASSERT(!dmu_objset_is_dirty(os, t)); 702 703 if (ds) 704 dsl_prop_unregister_all(ds, os); 705 706 if (os->os_sa) 707 sa_tear_down(os); 708 709 dmu_objset_evict_dbufs(os); 710 711 mutex_enter(&os->os_lock); 712 spa_evicting_os_register(os->os_spa, os); 713 if (list_is_empty(&os->os_dnodes)) { 714 mutex_exit(&os->os_lock); 715 dmu_objset_evict_done(os); 716 } else { 717 mutex_exit(&os->os_lock); 718 } 719 } 720 721 void 722 dmu_objset_evict_done(objset_t *os) 723 { 724 ASSERT3P(list_head(&os->os_dnodes), ==, NULL); 725 726 dnode_special_close(&os->os_meta_dnode); 727 if (DMU_USERUSED_DNODE(os)) { 728 dnode_special_close(&os->os_userused_dnode); 729 dnode_special_close(&os->os_groupused_dnode); 730 } 731 zil_free(os->os_zil); 732 733 VERIFY(arc_buf_remove_ref(os->os_phys_buf, &os->os_phys_buf)); 734 735 /* 736 * This is a barrier to prevent the objset from going away in 737 * dnode_move() until we can safely ensure that the objset is still in 738 * use. We consider the objset valid before the barrier and invalid 739 * after the barrier. 740 */ 741 rw_enter(&os_lock, RW_READER); 742 rw_exit(&os_lock); 743 744 mutex_destroy(&os->os_lock); 745 mutex_destroy(&os->os_obj_lock); 746 mutex_destroy(&os->os_user_ptr_lock); 747 spa_evicting_os_deregister(os->os_spa, os); 748 kmem_free(os, sizeof (objset_t)); 749 } 750 751 timestruc_t 752 dmu_objset_snap_cmtime(objset_t *os) 753 { 754 return (dsl_dir_snap_cmtime(os->os_dsl_dataset->ds_dir)); 755 } 756 757 /* called from dsl for meta-objset */ 758 objset_t * 759 dmu_objset_create_impl(spa_t *spa, dsl_dataset_t *ds, blkptr_t *bp, 760 dmu_objset_type_t type, dmu_tx_t *tx) 761 { 762 objset_t *os; 763 dnode_t *mdn; 764 765 ASSERT(dmu_tx_is_syncing(tx)); 766 767 if (ds != NULL) 768 VERIFY0(dmu_objset_from_ds(ds, &os)); 769 else 770 VERIFY0(dmu_objset_open_impl(spa, NULL, bp, &os)); 771 772 mdn = DMU_META_DNODE(os); 773 774 dnode_allocate(mdn, DMU_OT_DNODE, 1 << DNODE_BLOCK_SHIFT, 775 DN_MAX_INDBLKSHIFT, DMU_OT_NONE, 0, tx); 776 777 /* 778 * We don't want to have to increase the meta-dnode's nlevels 779 * later, because then we could do it in quescing context while 780 * we are also accessing it in open context. 781 * 782 * This precaution is not necessary for the MOS (ds == NULL), 783 * because the MOS is only updated in syncing context. 784 * This is most fortunate: the MOS is the only objset that 785 * needs to be synced multiple times as spa_sync() iterates 786 * to convergence, so minimizing its dn_nlevels matters. 787 */ 788 if (ds != NULL) { 789 int levels = 1; 790 791 /* 792 * Determine the number of levels necessary for the meta-dnode 793 * to contain DN_MAX_OBJECT dnodes. 794 */ 795 while ((uint64_t)mdn->dn_nblkptr << (mdn->dn_datablkshift + 796 (levels - 1) * (mdn->dn_indblkshift - SPA_BLKPTRSHIFT)) < 797 DN_MAX_OBJECT * sizeof (dnode_phys_t)) 798 levels++; 799 800 mdn->dn_next_nlevels[tx->tx_txg & TXG_MASK] = 801 mdn->dn_nlevels = levels; 802 } 803 804 ASSERT(type != DMU_OST_NONE); 805 ASSERT(type != DMU_OST_ANY); 806 ASSERT(type < DMU_OST_NUMTYPES); 807 os->os_phys->os_type = type; 808 if (dmu_objset_userused_enabled(os)) { 809 os->os_phys->os_flags |= OBJSET_FLAG_USERACCOUNTING_COMPLETE; 810 os->os_flags = os->os_phys->os_flags; 811 } 812 813 dsl_dataset_dirty(ds, tx); 814 815 return (os); 816 } 817 818 typedef struct dmu_objset_create_arg { 819 const char *doca_name; 820 cred_t *doca_cred; 821 void (*doca_userfunc)(objset_t *os, void *arg, 822 cred_t *cr, dmu_tx_t *tx); 823 void *doca_userarg; 824 dmu_objset_type_t doca_type; 825 uint64_t doca_flags; 826 } dmu_objset_create_arg_t; 827 828 /*ARGSUSED*/ 829 static int 830 dmu_objset_create_check(void *arg, dmu_tx_t *tx) 831 { 832 dmu_objset_create_arg_t *doca = arg; 833 dsl_pool_t *dp = dmu_tx_pool(tx); 834 dsl_dir_t *pdd; 835 const char *tail; 836 int error; 837 838 if (strchr(doca->doca_name, '@') != NULL) 839 return (SET_ERROR(EINVAL)); 840 841 error = dsl_dir_hold(dp, doca->doca_name, FTAG, &pdd, &tail); 842 if (error != 0) 843 return (error); 844 if (tail == NULL) { 845 dsl_dir_rele(pdd, FTAG); 846 return (SET_ERROR(EEXIST)); 847 } 848 error = dsl_fs_ss_limit_check(pdd, 1, ZFS_PROP_FILESYSTEM_LIMIT, NULL, 849 doca->doca_cred); 850 dsl_dir_rele(pdd, FTAG); 851 852 return (error); 853 } 854 855 static void 856 dmu_objset_create_sync(void *arg, dmu_tx_t *tx) 857 { 858 dmu_objset_create_arg_t *doca = arg; 859 dsl_pool_t *dp = dmu_tx_pool(tx); 860 dsl_dir_t *pdd; 861 const char *tail; 862 dsl_dataset_t *ds; 863 uint64_t obj; 864 blkptr_t *bp; 865 objset_t *os; 866 867 VERIFY0(dsl_dir_hold(dp, doca->doca_name, FTAG, &pdd, &tail)); 868 869 obj = dsl_dataset_create_sync(pdd, tail, NULL, doca->doca_flags, 870 doca->doca_cred, tx); 871 872 VERIFY0(dsl_dataset_hold_obj(pdd->dd_pool, obj, FTAG, &ds)); 873 bp = dsl_dataset_get_blkptr(ds); 874 os = dmu_objset_create_impl(pdd->dd_pool->dp_spa, 875 ds, bp, doca->doca_type, tx); 876 877 if (doca->doca_userfunc != NULL) { 878 doca->doca_userfunc(os, doca->doca_userarg, 879 doca->doca_cred, tx); 880 } 881 882 spa_history_log_internal_ds(ds, "create", tx, ""); 883 dsl_dataset_rele(ds, FTAG); 884 dsl_dir_rele(pdd, FTAG); 885 } 886 887 int 888 dmu_objset_create(const char *name, dmu_objset_type_t type, uint64_t flags, 889 void (*func)(objset_t *os, void *arg, cred_t *cr, dmu_tx_t *tx), void *arg) 890 { 891 dmu_objset_create_arg_t doca; 892 893 doca.doca_name = name; 894 doca.doca_cred = CRED(); 895 doca.doca_flags = flags; 896 doca.doca_userfunc = func; 897 doca.doca_userarg = arg; 898 doca.doca_type = type; 899 900 return (dsl_sync_task(name, 901 dmu_objset_create_check, dmu_objset_create_sync, &doca, 902 5, ZFS_SPACE_CHECK_NORMAL)); 903 } 904 905 typedef struct dmu_objset_clone_arg { 906 const char *doca_clone; 907 const char *doca_origin; 908 cred_t *doca_cred; 909 } dmu_objset_clone_arg_t; 910 911 /*ARGSUSED*/ 912 static int 913 dmu_objset_clone_check(void *arg, dmu_tx_t *tx) 914 { 915 dmu_objset_clone_arg_t *doca = arg; 916 dsl_dir_t *pdd; 917 const char *tail; 918 int error; 919 dsl_dataset_t *origin; 920 dsl_pool_t *dp = dmu_tx_pool(tx); 921 922 if (strchr(doca->doca_clone, '@') != NULL) 923 return (SET_ERROR(EINVAL)); 924 925 error = dsl_dir_hold(dp, doca->doca_clone, FTAG, &pdd, &tail); 926 if (error != 0) 927 return (error); 928 if (tail == NULL) { 929 dsl_dir_rele(pdd, FTAG); 930 return (SET_ERROR(EEXIST)); 931 } 932 933 error = dsl_fs_ss_limit_check(pdd, 1, ZFS_PROP_FILESYSTEM_LIMIT, NULL, 934 doca->doca_cred); 935 if (error != 0) { 936 dsl_dir_rele(pdd, FTAG); 937 return (SET_ERROR(EDQUOT)); 938 } 939 dsl_dir_rele(pdd, FTAG); 940 941 error = dsl_dataset_hold(dp, doca->doca_origin, FTAG, &origin); 942 if (error != 0) 943 return (error); 944 945 /* You can only clone snapshots, not the head datasets. */ 946 if (!origin->ds_is_snapshot) { 947 dsl_dataset_rele(origin, FTAG); 948 return (SET_ERROR(EINVAL)); 949 } 950 dsl_dataset_rele(origin, FTAG); 951 952 return (0); 953 } 954 955 static void 956 dmu_objset_clone_sync(void *arg, dmu_tx_t *tx) 957 { 958 dmu_objset_clone_arg_t *doca = arg; 959 dsl_pool_t *dp = dmu_tx_pool(tx); 960 dsl_dir_t *pdd; 961 const char *tail; 962 dsl_dataset_t *origin, *ds; 963 uint64_t obj; 964 char namebuf[MAXNAMELEN]; 965 966 VERIFY0(dsl_dir_hold(dp, doca->doca_clone, FTAG, &pdd, &tail)); 967 VERIFY0(dsl_dataset_hold(dp, doca->doca_origin, FTAG, &origin)); 968 969 obj = dsl_dataset_create_sync(pdd, tail, origin, 0, 970 doca->doca_cred, tx); 971 972 VERIFY0(dsl_dataset_hold_obj(pdd->dd_pool, obj, FTAG, &ds)); 973 dsl_dataset_name(origin, namebuf); 974 spa_history_log_internal_ds(ds, "clone", tx, 975 "origin=%s (%llu)", namebuf, origin->ds_object); 976 dsl_dataset_rele(ds, FTAG); 977 dsl_dataset_rele(origin, FTAG); 978 dsl_dir_rele(pdd, FTAG); 979 } 980 981 int 982 dmu_objset_clone(const char *clone, const char *origin) 983 { 984 dmu_objset_clone_arg_t doca; 985 986 doca.doca_clone = clone; 987 doca.doca_origin = origin; 988 doca.doca_cred = CRED(); 989 990 return (dsl_sync_task(clone, 991 dmu_objset_clone_check, dmu_objset_clone_sync, &doca, 992 5, ZFS_SPACE_CHECK_NORMAL)); 993 } 994 995 int 996 dmu_objset_snapshot_one(const char *fsname, const char *snapname) 997 { 998 int err; 999 char *longsnap = kmem_asprintf("%s@%s", fsname, snapname); 1000 nvlist_t *snaps = fnvlist_alloc(); 1001 1002 fnvlist_add_boolean(snaps, longsnap); 1003 strfree(longsnap); 1004 err = dsl_dataset_snapshot(snaps, NULL, NULL); 1005 fnvlist_free(snaps); 1006 return (err); 1007 } 1008 1009 static void 1010 dmu_objset_sync_dnodes(list_t *list, list_t *newlist, dmu_tx_t *tx) 1011 { 1012 dnode_t *dn; 1013 1014 while (dn = list_head(list)) { 1015 ASSERT(dn->dn_object != DMU_META_DNODE_OBJECT); 1016 ASSERT(dn->dn_dbuf->db_data_pending); 1017 /* 1018 * Initialize dn_zio outside dnode_sync() because the 1019 * meta-dnode needs to set it ouside dnode_sync(). 1020 */ 1021 dn->dn_zio = dn->dn_dbuf->db_data_pending->dr_zio; 1022 ASSERT(dn->dn_zio); 1023 1024 ASSERT3U(dn->dn_nlevels, <=, DN_MAX_LEVELS); 1025 list_remove(list, dn); 1026 1027 if (newlist) { 1028 (void) dnode_add_ref(dn, newlist); 1029 list_insert_tail(newlist, dn); 1030 } 1031 1032 dnode_sync(dn, tx); 1033 } 1034 } 1035 1036 /* ARGSUSED */ 1037 static void 1038 dmu_objset_write_ready(zio_t *zio, arc_buf_t *abuf, void *arg) 1039 { 1040 blkptr_t *bp = zio->io_bp; 1041 objset_t *os = arg; 1042 dnode_phys_t *dnp = &os->os_phys->os_meta_dnode; 1043 1044 ASSERT(!BP_IS_EMBEDDED(bp)); 1045 ASSERT3P(bp, ==, os->os_rootbp); 1046 ASSERT3U(BP_GET_TYPE(bp), ==, DMU_OT_OBJSET); 1047 ASSERT0(BP_GET_LEVEL(bp)); 1048 1049 /* 1050 * Update rootbp fill count: it should be the number of objects 1051 * allocated in the object set (not counting the "special" 1052 * objects that are stored in the objset_phys_t -- the meta 1053 * dnode and user/group accounting objects). 1054 */ 1055 bp->blk_fill = 0; 1056 for (int i = 0; i < dnp->dn_nblkptr; i++) 1057 bp->blk_fill += BP_GET_FILL(&dnp->dn_blkptr[i]); 1058 } 1059 1060 /* ARGSUSED */ 1061 static void 1062 dmu_objset_write_done(zio_t *zio, arc_buf_t *abuf, void *arg) 1063 { 1064 blkptr_t *bp = zio->io_bp; 1065 blkptr_t *bp_orig = &zio->io_bp_orig; 1066 objset_t *os = arg; 1067 1068 if (zio->io_flags & ZIO_FLAG_IO_REWRITE) { 1069 ASSERT(BP_EQUAL(bp, bp_orig)); 1070 } else { 1071 dsl_dataset_t *ds = os->os_dsl_dataset; 1072 dmu_tx_t *tx = os->os_synctx; 1073 1074 (void) dsl_dataset_block_kill(ds, bp_orig, tx, B_TRUE); 1075 dsl_dataset_block_born(ds, bp, tx); 1076 } 1077 } 1078 1079 /* called from dsl */ 1080 void 1081 dmu_objset_sync(objset_t *os, zio_t *pio, dmu_tx_t *tx) 1082 { 1083 int txgoff; 1084 zbookmark_phys_t zb; 1085 zio_prop_t zp; 1086 zio_t *zio; 1087 list_t *list; 1088 list_t *newlist = NULL; 1089 dbuf_dirty_record_t *dr; 1090 1091 dprintf_ds(os->os_dsl_dataset, "txg=%llu\n", tx->tx_txg); 1092 1093 ASSERT(dmu_tx_is_syncing(tx)); 1094 /* XXX the write_done callback should really give us the tx... */ 1095 os->os_synctx = tx; 1096 1097 if (os->os_dsl_dataset == NULL) { 1098 /* 1099 * This is the MOS. If we have upgraded, 1100 * spa_max_replication() could change, so reset 1101 * os_copies here. 1102 */ 1103 os->os_copies = spa_max_replication(os->os_spa); 1104 } 1105 1106 /* 1107 * Create the root block IO 1108 */ 1109 SET_BOOKMARK(&zb, os->os_dsl_dataset ? 1110 os->os_dsl_dataset->ds_object : DMU_META_OBJSET, 1111 ZB_ROOT_OBJECT, ZB_ROOT_LEVEL, ZB_ROOT_BLKID); 1112 arc_release(os->os_phys_buf, &os->os_phys_buf); 1113 1114 dmu_write_policy(os, NULL, 0, 0, &zp); 1115 1116 zio = arc_write(pio, os->os_spa, tx->tx_txg, 1117 os->os_rootbp, os->os_phys_buf, DMU_OS_IS_L2CACHEABLE(os), 1118 DMU_OS_IS_L2COMPRESSIBLE(os), &zp, dmu_objset_write_ready, 1119 NULL, dmu_objset_write_done, os, ZIO_PRIORITY_ASYNC_WRITE, 1120 ZIO_FLAG_MUSTSUCCEED, &zb); 1121 1122 /* 1123 * Sync special dnodes - the parent IO for the sync is the root block 1124 */ 1125 DMU_META_DNODE(os)->dn_zio = zio; 1126 dnode_sync(DMU_META_DNODE(os), tx); 1127 1128 os->os_phys->os_flags = os->os_flags; 1129 1130 if (DMU_USERUSED_DNODE(os) && 1131 DMU_USERUSED_DNODE(os)->dn_type != DMU_OT_NONE) { 1132 DMU_USERUSED_DNODE(os)->dn_zio = zio; 1133 dnode_sync(DMU_USERUSED_DNODE(os), tx); 1134 DMU_GROUPUSED_DNODE(os)->dn_zio = zio; 1135 dnode_sync(DMU_GROUPUSED_DNODE(os), tx); 1136 } 1137 1138 txgoff = tx->tx_txg & TXG_MASK; 1139 1140 if (dmu_objset_userused_enabled(os)) { 1141 newlist = &os->os_synced_dnodes; 1142 /* 1143 * We must create the list here because it uses the 1144 * dn_dirty_link[] of this txg. 1145 */ 1146 list_create(newlist, sizeof (dnode_t), 1147 offsetof(dnode_t, dn_dirty_link[txgoff])); 1148 } 1149 1150 dmu_objset_sync_dnodes(&os->os_free_dnodes[txgoff], newlist, tx); 1151 dmu_objset_sync_dnodes(&os->os_dirty_dnodes[txgoff], newlist, tx); 1152 1153 list = &DMU_META_DNODE(os)->dn_dirty_records[txgoff]; 1154 while (dr = list_head(list)) { 1155 ASSERT0(dr->dr_dbuf->db_level); 1156 list_remove(list, dr); 1157 if (dr->dr_zio) 1158 zio_nowait(dr->dr_zio); 1159 } 1160 /* 1161 * Free intent log blocks up to this tx. 1162 */ 1163 zil_sync(os->os_zil, tx); 1164 os->os_phys->os_zil_header = os->os_zil_header; 1165 zio_nowait(zio); 1166 } 1167 1168 boolean_t 1169 dmu_objset_is_dirty(objset_t *os, uint64_t txg) 1170 { 1171 return (!list_is_empty(&os->os_dirty_dnodes[txg & TXG_MASK]) || 1172 !list_is_empty(&os->os_free_dnodes[txg & TXG_MASK])); 1173 } 1174 1175 static objset_used_cb_t *used_cbs[DMU_OST_NUMTYPES]; 1176 1177 void 1178 dmu_objset_register_type(dmu_objset_type_t ost, objset_used_cb_t *cb) 1179 { 1180 used_cbs[ost] = cb; 1181 } 1182 1183 boolean_t 1184 dmu_objset_userused_enabled(objset_t *os) 1185 { 1186 return (spa_version(os->os_spa) >= SPA_VERSION_USERSPACE && 1187 used_cbs[os->os_phys->os_type] != NULL && 1188 DMU_USERUSED_DNODE(os) != NULL); 1189 } 1190 1191 static void 1192 do_userquota_update(objset_t *os, uint64_t used, uint64_t flags, 1193 uint64_t user, uint64_t group, boolean_t subtract, dmu_tx_t *tx) 1194 { 1195 if ((flags & DNODE_FLAG_USERUSED_ACCOUNTED)) { 1196 int64_t delta = DNODE_SIZE + used; 1197 if (subtract) 1198 delta = -delta; 1199 VERIFY3U(0, ==, zap_increment_int(os, DMU_USERUSED_OBJECT, 1200 user, delta, tx)); 1201 VERIFY3U(0, ==, zap_increment_int(os, DMU_GROUPUSED_OBJECT, 1202 group, delta, tx)); 1203 } 1204 } 1205 1206 void 1207 dmu_objset_do_userquota_updates(objset_t *os, dmu_tx_t *tx) 1208 { 1209 dnode_t *dn; 1210 list_t *list = &os->os_synced_dnodes; 1211 1212 ASSERT(list_head(list) == NULL || dmu_objset_userused_enabled(os)); 1213 1214 while (dn = list_head(list)) { 1215 int flags; 1216 ASSERT(!DMU_OBJECT_IS_SPECIAL(dn->dn_object)); 1217 ASSERT(dn->dn_phys->dn_type == DMU_OT_NONE || 1218 dn->dn_phys->dn_flags & 1219 DNODE_FLAG_USERUSED_ACCOUNTED); 1220 1221 /* Allocate the user/groupused objects if necessary. */ 1222 if (DMU_USERUSED_DNODE(os)->dn_type == DMU_OT_NONE) { 1223 VERIFY(0 == zap_create_claim(os, 1224 DMU_USERUSED_OBJECT, 1225 DMU_OT_USERGROUP_USED, DMU_OT_NONE, 0, tx)); 1226 VERIFY(0 == zap_create_claim(os, 1227 DMU_GROUPUSED_OBJECT, 1228 DMU_OT_USERGROUP_USED, DMU_OT_NONE, 0, tx)); 1229 } 1230 1231 /* 1232 * We intentionally modify the zap object even if the 1233 * net delta is zero. Otherwise 1234 * the block of the zap obj could be shared between 1235 * datasets but need to be different between them after 1236 * a bprewrite. 1237 */ 1238 1239 flags = dn->dn_id_flags; 1240 ASSERT(flags); 1241 if (flags & DN_ID_OLD_EXIST) { 1242 do_userquota_update(os, dn->dn_oldused, dn->dn_oldflags, 1243 dn->dn_olduid, dn->dn_oldgid, B_TRUE, tx); 1244 } 1245 if (flags & DN_ID_NEW_EXIST) { 1246 do_userquota_update(os, DN_USED_BYTES(dn->dn_phys), 1247 dn->dn_phys->dn_flags, dn->dn_newuid, 1248 dn->dn_newgid, B_FALSE, tx); 1249 } 1250 1251 mutex_enter(&dn->dn_mtx); 1252 dn->dn_oldused = 0; 1253 dn->dn_oldflags = 0; 1254 if (dn->dn_id_flags & DN_ID_NEW_EXIST) { 1255 dn->dn_olduid = dn->dn_newuid; 1256 dn->dn_oldgid = dn->dn_newgid; 1257 dn->dn_id_flags |= DN_ID_OLD_EXIST; 1258 if (dn->dn_bonuslen == 0) 1259 dn->dn_id_flags |= DN_ID_CHKED_SPILL; 1260 else 1261 dn->dn_id_flags |= DN_ID_CHKED_BONUS; 1262 } 1263 dn->dn_id_flags &= ~(DN_ID_NEW_EXIST); 1264 mutex_exit(&dn->dn_mtx); 1265 1266 list_remove(list, dn); 1267 dnode_rele(dn, list); 1268 } 1269 } 1270 1271 /* 1272 * Returns a pointer to data to find uid/gid from 1273 * 1274 * If a dirty record for transaction group that is syncing can't 1275 * be found then NULL is returned. In the NULL case it is assumed 1276 * the uid/gid aren't changing. 1277 */ 1278 static void * 1279 dmu_objset_userquota_find_data(dmu_buf_impl_t *db, dmu_tx_t *tx) 1280 { 1281 dbuf_dirty_record_t *dr, **drp; 1282 void *data; 1283 1284 if (db->db_dirtycnt == 0) 1285 return (db->db.db_data); /* Nothing is changing */ 1286 1287 for (drp = &db->db_last_dirty; (dr = *drp) != NULL; drp = &dr->dr_next) 1288 if (dr->dr_txg == tx->tx_txg) 1289 break; 1290 1291 if (dr == NULL) { 1292 data = NULL; 1293 } else { 1294 dnode_t *dn; 1295 1296 DB_DNODE_ENTER(dr->dr_dbuf); 1297 dn = DB_DNODE(dr->dr_dbuf); 1298 1299 if (dn->dn_bonuslen == 0 && 1300 dr->dr_dbuf->db_blkid == DMU_SPILL_BLKID) 1301 data = dr->dt.dl.dr_data->b_data; 1302 else 1303 data = dr->dt.dl.dr_data; 1304 1305 DB_DNODE_EXIT(dr->dr_dbuf); 1306 } 1307 1308 return (data); 1309 } 1310 1311 void 1312 dmu_objset_userquota_get_ids(dnode_t *dn, boolean_t before, dmu_tx_t *tx) 1313 { 1314 objset_t *os = dn->dn_objset; 1315 void *data = NULL; 1316 dmu_buf_impl_t *db = NULL; 1317 uint64_t *user = NULL; 1318 uint64_t *group = NULL; 1319 int flags = dn->dn_id_flags; 1320 int error; 1321 boolean_t have_spill = B_FALSE; 1322 1323 if (!dmu_objset_userused_enabled(dn->dn_objset)) 1324 return; 1325 1326 if (before && (flags & (DN_ID_CHKED_BONUS|DN_ID_OLD_EXIST| 1327 DN_ID_CHKED_SPILL))) 1328 return; 1329 1330 if (before && dn->dn_bonuslen != 0) 1331 data = DN_BONUS(dn->dn_phys); 1332 else if (!before && dn->dn_bonuslen != 0) { 1333 if (dn->dn_bonus) { 1334 db = dn->dn_bonus; 1335 mutex_enter(&db->db_mtx); 1336 data = dmu_objset_userquota_find_data(db, tx); 1337 } else { 1338 data = DN_BONUS(dn->dn_phys); 1339 } 1340 } else if (dn->dn_bonuslen == 0 && dn->dn_bonustype == DMU_OT_SA) { 1341 int rf = 0; 1342 1343 if (RW_WRITE_HELD(&dn->dn_struct_rwlock)) 1344 rf |= DB_RF_HAVESTRUCT; 1345 error = dmu_spill_hold_by_dnode(dn, 1346 rf | DB_RF_MUST_SUCCEED, 1347 FTAG, (dmu_buf_t **)&db); 1348 ASSERT(error == 0); 1349 mutex_enter(&db->db_mtx); 1350 data = (before) ? db->db.db_data : 1351 dmu_objset_userquota_find_data(db, tx); 1352 have_spill = B_TRUE; 1353 } else { 1354 mutex_enter(&dn->dn_mtx); 1355 dn->dn_id_flags |= DN_ID_CHKED_BONUS; 1356 mutex_exit(&dn->dn_mtx); 1357 return; 1358 } 1359 1360 if (before) { 1361 ASSERT(data); 1362 user = &dn->dn_olduid; 1363 group = &dn->dn_oldgid; 1364 } else if (data) { 1365 user = &dn->dn_newuid; 1366 group = &dn->dn_newgid; 1367 } 1368 1369 /* 1370 * Must always call the callback in case the object 1371 * type has changed and that type isn't an object type to track 1372 */ 1373 error = used_cbs[os->os_phys->os_type](dn->dn_bonustype, data, 1374 user, group); 1375 1376 /* 1377 * Preserve existing uid/gid when the callback can't determine 1378 * what the new uid/gid are and the callback returned EEXIST. 1379 * The EEXIST error tells us to just use the existing uid/gid. 1380 * If we don't know what the old values are then just assign 1381 * them to 0, since that is a new file being created. 1382 */ 1383 if (!before && data == NULL && error == EEXIST) { 1384 if (flags & DN_ID_OLD_EXIST) { 1385 dn->dn_newuid = dn->dn_olduid; 1386 dn->dn_newgid = dn->dn_oldgid; 1387 } else { 1388 dn->dn_newuid = 0; 1389 dn->dn_newgid = 0; 1390 } 1391 error = 0; 1392 } 1393 1394 if (db) 1395 mutex_exit(&db->db_mtx); 1396 1397 mutex_enter(&dn->dn_mtx); 1398 if (error == 0 && before) 1399 dn->dn_id_flags |= DN_ID_OLD_EXIST; 1400 if (error == 0 && !before) 1401 dn->dn_id_flags |= DN_ID_NEW_EXIST; 1402 1403 if (have_spill) { 1404 dn->dn_id_flags |= DN_ID_CHKED_SPILL; 1405 } else { 1406 dn->dn_id_flags |= DN_ID_CHKED_BONUS; 1407 } 1408 mutex_exit(&dn->dn_mtx); 1409 if (have_spill) 1410 dmu_buf_rele((dmu_buf_t *)db, FTAG); 1411 } 1412 1413 boolean_t 1414 dmu_objset_userspace_present(objset_t *os) 1415 { 1416 return (os->os_phys->os_flags & 1417 OBJSET_FLAG_USERACCOUNTING_COMPLETE); 1418 } 1419 1420 int 1421 dmu_objset_userspace_upgrade(objset_t *os) 1422 { 1423 uint64_t obj; 1424 int err = 0; 1425 1426 if (dmu_objset_userspace_present(os)) 1427 return (0); 1428 if (!dmu_objset_userused_enabled(os)) 1429 return (SET_ERROR(ENOTSUP)); 1430 if (dmu_objset_is_snapshot(os)) 1431 return (SET_ERROR(EINVAL)); 1432 1433 /* 1434 * We simply need to mark every object dirty, so that it will be 1435 * synced out and now accounted. If this is called 1436 * concurrently, or if we already did some work before crashing, 1437 * that's fine, since we track each object's accounted state 1438 * independently. 1439 */ 1440 1441 for (obj = 0; err == 0; err = dmu_object_next(os, &obj, FALSE, 0)) { 1442 dmu_tx_t *tx; 1443 dmu_buf_t *db; 1444 int objerr; 1445 1446 if (issig(JUSTLOOKING) && issig(FORREAL)) 1447 return (SET_ERROR(EINTR)); 1448 1449 objerr = dmu_bonus_hold(os, obj, FTAG, &db); 1450 if (objerr != 0) 1451 continue; 1452 tx = dmu_tx_create(os); 1453 dmu_tx_hold_bonus(tx, obj); 1454 objerr = dmu_tx_assign(tx, TXG_WAIT); 1455 if (objerr != 0) { 1456 dmu_tx_abort(tx); 1457 continue; 1458 } 1459 dmu_buf_will_dirty(db, tx); 1460 dmu_buf_rele(db, FTAG); 1461 dmu_tx_commit(tx); 1462 } 1463 1464 os->os_flags |= OBJSET_FLAG_USERACCOUNTING_COMPLETE; 1465 txg_wait_synced(dmu_objset_pool(os), 0); 1466 return (0); 1467 } 1468 1469 void 1470 dmu_objset_space(objset_t *os, uint64_t *refdbytesp, uint64_t *availbytesp, 1471 uint64_t *usedobjsp, uint64_t *availobjsp) 1472 { 1473 dsl_dataset_space(os->os_dsl_dataset, refdbytesp, availbytesp, 1474 usedobjsp, availobjsp); 1475 } 1476 1477 uint64_t 1478 dmu_objset_fsid_guid(objset_t *os) 1479 { 1480 return (dsl_dataset_fsid_guid(os->os_dsl_dataset)); 1481 } 1482 1483 void 1484 dmu_objset_fast_stat(objset_t *os, dmu_objset_stats_t *stat) 1485 { 1486 stat->dds_type = os->os_phys->os_type; 1487 if (os->os_dsl_dataset) 1488 dsl_dataset_fast_stat(os->os_dsl_dataset, stat); 1489 } 1490 1491 void 1492 dmu_objset_stats(objset_t *os, nvlist_t *nv) 1493 { 1494 ASSERT(os->os_dsl_dataset || 1495 os->os_phys->os_type == DMU_OST_META); 1496 1497 if (os->os_dsl_dataset != NULL) 1498 dsl_dataset_stats(os->os_dsl_dataset, nv); 1499 1500 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_TYPE, 1501 os->os_phys->os_type); 1502 dsl_prop_nvlist_add_uint64(nv, ZFS_PROP_USERACCOUNTING, 1503 dmu_objset_userspace_present(os)); 1504 } 1505 1506 int 1507 dmu_objset_is_snapshot(objset_t *os) 1508 { 1509 if (os->os_dsl_dataset != NULL) 1510 return (os->os_dsl_dataset->ds_is_snapshot); 1511 else 1512 return (B_FALSE); 1513 } 1514 1515 int 1516 dmu_snapshot_realname(objset_t *os, char *name, char *real, int maxlen, 1517 boolean_t *conflict) 1518 { 1519 dsl_dataset_t *ds = os->os_dsl_dataset; 1520 uint64_t ignored; 1521 1522 if (dsl_dataset_phys(ds)->ds_snapnames_zapobj == 0) 1523 return (SET_ERROR(ENOENT)); 1524 1525 return (zap_lookup_norm(ds->ds_dir->dd_pool->dp_meta_objset, 1526 dsl_dataset_phys(ds)->ds_snapnames_zapobj, name, 8, 1, &ignored, 1527 MT_FIRST, real, maxlen, conflict)); 1528 } 1529 1530 int 1531 dmu_snapshot_list_next(objset_t *os, int namelen, char *name, 1532 uint64_t *idp, uint64_t *offp, boolean_t *case_conflict) 1533 { 1534 dsl_dataset_t *ds = os->os_dsl_dataset; 1535 zap_cursor_t cursor; 1536 zap_attribute_t attr; 1537 1538 ASSERT(dsl_pool_config_held(dmu_objset_pool(os))); 1539 1540 if (dsl_dataset_phys(ds)->ds_snapnames_zapobj == 0) 1541 return (SET_ERROR(ENOENT)); 1542 1543 zap_cursor_init_serialized(&cursor, 1544 ds->ds_dir->dd_pool->dp_meta_objset, 1545 dsl_dataset_phys(ds)->ds_snapnames_zapobj, *offp); 1546 1547 if (zap_cursor_retrieve(&cursor, &attr) != 0) { 1548 zap_cursor_fini(&cursor); 1549 return (SET_ERROR(ENOENT)); 1550 } 1551 1552 if (strlen(attr.za_name) + 1 > namelen) { 1553 zap_cursor_fini(&cursor); 1554 return (SET_ERROR(ENAMETOOLONG)); 1555 } 1556 1557 (void) strcpy(name, attr.za_name); 1558 if (idp) 1559 *idp = attr.za_first_integer; 1560 if (case_conflict) 1561 *case_conflict = attr.za_normalization_conflict; 1562 zap_cursor_advance(&cursor); 1563 *offp = zap_cursor_serialize(&cursor); 1564 zap_cursor_fini(&cursor); 1565 1566 return (0); 1567 } 1568 1569 int 1570 dmu_dir_list_next(objset_t *os, int namelen, char *name, 1571 uint64_t *idp, uint64_t *offp) 1572 { 1573 dsl_dir_t *dd = os->os_dsl_dataset->ds_dir; 1574 zap_cursor_t cursor; 1575 zap_attribute_t attr; 1576 1577 /* there is no next dir on a snapshot! */ 1578 if (os->os_dsl_dataset->ds_object != 1579 dsl_dir_phys(dd)->dd_head_dataset_obj) 1580 return (SET_ERROR(ENOENT)); 1581 1582 zap_cursor_init_serialized(&cursor, 1583 dd->dd_pool->dp_meta_objset, 1584 dsl_dir_phys(dd)->dd_child_dir_zapobj, *offp); 1585 1586 if (zap_cursor_retrieve(&cursor, &attr) != 0) { 1587 zap_cursor_fini(&cursor); 1588 return (SET_ERROR(ENOENT)); 1589 } 1590 1591 if (strlen(attr.za_name) + 1 > namelen) { 1592 zap_cursor_fini(&cursor); 1593 return (SET_ERROR(ENAMETOOLONG)); 1594 } 1595 1596 (void) strcpy(name, attr.za_name); 1597 if (idp) 1598 *idp = attr.za_first_integer; 1599 zap_cursor_advance(&cursor); 1600 *offp = zap_cursor_serialize(&cursor); 1601 zap_cursor_fini(&cursor); 1602 1603 return (0); 1604 } 1605 1606 typedef struct dmu_objset_find_ctx { 1607 taskq_t *dc_tq; 1608 dsl_pool_t *dc_dp; 1609 uint64_t dc_ddobj; 1610 int (*dc_func)(dsl_pool_t *, dsl_dataset_t *, void *); 1611 void *dc_arg; 1612 int dc_flags; 1613 kmutex_t *dc_error_lock; 1614 int *dc_error; 1615 } dmu_objset_find_ctx_t; 1616 1617 static void 1618 dmu_objset_find_dp_impl(dmu_objset_find_ctx_t *dcp) 1619 { 1620 dsl_pool_t *dp = dcp->dc_dp; 1621 dmu_objset_find_ctx_t *child_dcp; 1622 dsl_dir_t *dd; 1623 dsl_dataset_t *ds; 1624 zap_cursor_t zc; 1625 zap_attribute_t *attr; 1626 uint64_t thisobj; 1627 int err = 0; 1628 1629 /* don't process if there already was an error */ 1630 if (*dcp->dc_error != 0) 1631 goto out; 1632 1633 err = dsl_dir_hold_obj(dp, dcp->dc_ddobj, NULL, FTAG, &dd); 1634 if (err != 0) 1635 goto out; 1636 1637 /* Don't visit hidden ($MOS & $ORIGIN) objsets. */ 1638 if (dd->dd_myname[0] == '$') { 1639 dsl_dir_rele(dd, FTAG); 1640 goto out; 1641 } 1642 1643 thisobj = dsl_dir_phys(dd)->dd_head_dataset_obj; 1644 attr = kmem_alloc(sizeof (zap_attribute_t), KM_SLEEP); 1645 1646 /* 1647 * Iterate over all children. 1648 */ 1649 if (dcp->dc_flags & DS_FIND_CHILDREN) { 1650 for (zap_cursor_init(&zc, dp->dp_meta_objset, 1651 dsl_dir_phys(dd)->dd_child_dir_zapobj); 1652 zap_cursor_retrieve(&zc, attr) == 0; 1653 (void) zap_cursor_advance(&zc)) { 1654 ASSERT3U(attr->za_integer_length, ==, 1655 sizeof (uint64_t)); 1656 ASSERT3U(attr->za_num_integers, ==, 1); 1657 1658 child_dcp = kmem_alloc(sizeof (*child_dcp), KM_SLEEP); 1659 *child_dcp = *dcp; 1660 child_dcp->dc_ddobj = attr->za_first_integer; 1661 if (dcp->dc_tq != NULL) 1662 (void) taskq_dispatch(dcp->dc_tq, 1663 dmu_objset_find_dp_cb, child_dcp, TQ_SLEEP); 1664 else 1665 dmu_objset_find_dp_impl(child_dcp); 1666 } 1667 zap_cursor_fini(&zc); 1668 } 1669 1670 /* 1671 * Iterate over all snapshots. 1672 */ 1673 if (dcp->dc_flags & DS_FIND_SNAPSHOTS) { 1674 dsl_dataset_t *ds; 1675 err = dsl_dataset_hold_obj(dp, thisobj, FTAG, &ds); 1676 1677 if (err == 0) { 1678 uint64_t snapobj; 1679 1680 snapobj = dsl_dataset_phys(ds)->ds_snapnames_zapobj; 1681 dsl_dataset_rele(ds, FTAG); 1682 1683 for (zap_cursor_init(&zc, dp->dp_meta_objset, snapobj); 1684 zap_cursor_retrieve(&zc, attr) == 0; 1685 (void) zap_cursor_advance(&zc)) { 1686 ASSERT3U(attr->za_integer_length, ==, 1687 sizeof (uint64_t)); 1688 ASSERT3U(attr->za_num_integers, ==, 1); 1689 1690 err = dsl_dataset_hold_obj(dp, 1691 attr->za_first_integer, FTAG, &ds); 1692 if (err != 0) 1693 break; 1694 err = dcp->dc_func(dp, ds, dcp->dc_arg); 1695 dsl_dataset_rele(ds, FTAG); 1696 if (err != 0) 1697 break; 1698 } 1699 zap_cursor_fini(&zc); 1700 } 1701 } 1702 1703 dsl_dir_rele(dd, FTAG); 1704 kmem_free(attr, sizeof (zap_attribute_t)); 1705 1706 if (err != 0) 1707 goto out; 1708 1709 /* 1710 * Apply to self. 1711 */ 1712 err = dsl_dataset_hold_obj(dp, thisobj, FTAG, &ds); 1713 if (err != 0) 1714 goto out; 1715 err = dcp->dc_func(dp, ds, dcp->dc_arg); 1716 dsl_dataset_rele(ds, FTAG); 1717 1718 out: 1719 if (err != 0) { 1720 mutex_enter(dcp->dc_error_lock); 1721 /* only keep first error */ 1722 if (*dcp->dc_error == 0) 1723 *dcp->dc_error = err; 1724 mutex_exit(dcp->dc_error_lock); 1725 } 1726 1727 kmem_free(dcp, sizeof (*dcp)); 1728 } 1729 1730 static void 1731 dmu_objset_find_dp_cb(void *arg) 1732 { 1733 dmu_objset_find_ctx_t *dcp = arg; 1734 dsl_pool_t *dp = dcp->dc_dp; 1735 1736 /* 1737 * We need to get a pool_config_lock here, as there are several 1738 * asssert(pool_config_held) down the stack. Getting a lock via 1739 * dsl_pool_config_enter is risky, as it might be stalled by a 1740 * pending writer. This would deadlock, as the write lock can 1741 * only be granted when our parent thread gives up the lock. 1742 * The _prio interface gives us priority over a pending writer. 1743 */ 1744 dsl_pool_config_enter_prio(dp, FTAG); 1745 1746 dmu_objset_find_dp_impl(dcp); 1747 1748 dsl_pool_config_exit(dp, FTAG); 1749 } 1750 1751 /* 1752 * Find objsets under and including ddobj, call func(ds) on each. 1753 * The order for the enumeration is completely undefined. 1754 * func is called with dsl_pool_config held. 1755 */ 1756 int 1757 dmu_objset_find_dp(dsl_pool_t *dp, uint64_t ddobj, 1758 int func(dsl_pool_t *, dsl_dataset_t *, void *), void *arg, int flags) 1759 { 1760 int error = 0; 1761 taskq_t *tq = NULL; 1762 int ntasks; 1763 dmu_objset_find_ctx_t *dcp; 1764 kmutex_t err_lock; 1765 1766 mutex_init(&err_lock, NULL, MUTEX_DEFAULT, NULL); 1767 dcp = kmem_alloc(sizeof (*dcp), KM_SLEEP); 1768 dcp->dc_tq = NULL; 1769 dcp->dc_dp = dp; 1770 dcp->dc_ddobj = ddobj; 1771 dcp->dc_func = func; 1772 dcp->dc_arg = arg; 1773 dcp->dc_flags = flags; 1774 dcp->dc_error_lock = &err_lock; 1775 dcp->dc_error = &error; 1776 1777 if ((flags & DS_FIND_SERIALIZE) || dsl_pool_config_held_writer(dp)) { 1778 /* 1779 * In case a write lock is held we can't make use of 1780 * parallelism, as down the stack of the worker threads 1781 * the lock is asserted via dsl_pool_config_held. 1782 * In case of a read lock this is solved by getting a read 1783 * lock in each worker thread, which isn't possible in case 1784 * of a writer lock. So we fall back to the synchronous path 1785 * here. 1786 * In the future it might be possible to get some magic into 1787 * dsl_pool_config_held in a way that it returns true for 1788 * the worker threads so that a single lock held from this 1789 * thread suffices. For now, stay single threaded. 1790 */ 1791 dmu_objset_find_dp_impl(dcp); 1792 1793 return (error); 1794 } 1795 1796 ntasks = dmu_find_threads; 1797 if (ntasks == 0) 1798 ntasks = vdev_count_leaves(dp->dp_spa) * 4; 1799 tq = taskq_create("dmu_objset_find", ntasks, minclsyspri, ntasks, 1800 INT_MAX, 0); 1801 if (tq == NULL) { 1802 kmem_free(dcp, sizeof (*dcp)); 1803 return (SET_ERROR(ENOMEM)); 1804 } 1805 dcp->dc_tq = tq; 1806 1807 /* dcp will be freed by task */ 1808 (void) taskq_dispatch(tq, dmu_objset_find_dp_cb, dcp, TQ_SLEEP); 1809 1810 /* 1811 * PORTING: this code relies on the property of taskq_wait to wait 1812 * until no more tasks are queued and no more tasks are active. As 1813 * we always queue new tasks from within other tasks, task_wait 1814 * reliably waits for the full recursion to finish, even though we 1815 * enqueue new tasks after taskq_wait has been called. 1816 * On platforms other than illumos, taskq_wait may not have this 1817 * property. 1818 */ 1819 taskq_wait(tq); 1820 taskq_destroy(tq); 1821 mutex_destroy(&err_lock); 1822 1823 return (error); 1824 } 1825 1826 /* 1827 * Find all objsets under name, and for each, call 'func(child_name, arg)'. 1828 * The dp_config_rwlock must not be held when this is called, and it 1829 * will not be held when the callback is called. 1830 * Therefore this function should only be used when the pool is not changing 1831 * (e.g. in syncing context), or the callback can deal with the possible races. 1832 */ 1833 static int 1834 dmu_objset_find_impl(spa_t *spa, const char *name, 1835 int func(const char *, void *), void *arg, int flags) 1836 { 1837 dsl_dir_t *dd; 1838 dsl_pool_t *dp = spa_get_dsl(spa); 1839 dsl_dataset_t *ds; 1840 zap_cursor_t zc; 1841 zap_attribute_t *attr; 1842 char *child; 1843 uint64_t thisobj; 1844 int err; 1845 1846 dsl_pool_config_enter(dp, FTAG); 1847 1848 err = dsl_dir_hold(dp, name, FTAG, &dd, NULL); 1849 if (err != 0) { 1850 dsl_pool_config_exit(dp, FTAG); 1851 return (err); 1852 } 1853 1854 /* Don't visit hidden ($MOS & $ORIGIN) objsets. */ 1855 if (dd->dd_myname[0] == '$') { 1856 dsl_dir_rele(dd, FTAG); 1857 dsl_pool_config_exit(dp, FTAG); 1858 return (0); 1859 } 1860 1861 thisobj = dsl_dir_phys(dd)->dd_head_dataset_obj; 1862 attr = kmem_alloc(sizeof (zap_attribute_t), KM_SLEEP); 1863 1864 /* 1865 * Iterate over all children. 1866 */ 1867 if (flags & DS_FIND_CHILDREN) { 1868 for (zap_cursor_init(&zc, dp->dp_meta_objset, 1869 dsl_dir_phys(dd)->dd_child_dir_zapobj); 1870 zap_cursor_retrieve(&zc, attr) == 0; 1871 (void) zap_cursor_advance(&zc)) { 1872 ASSERT3U(attr->za_integer_length, ==, 1873 sizeof (uint64_t)); 1874 ASSERT3U(attr->za_num_integers, ==, 1); 1875 1876 child = kmem_asprintf("%s/%s", name, attr->za_name); 1877 dsl_pool_config_exit(dp, FTAG); 1878 err = dmu_objset_find_impl(spa, child, 1879 func, arg, flags); 1880 dsl_pool_config_enter(dp, FTAG); 1881 strfree(child); 1882 if (err != 0) 1883 break; 1884 } 1885 zap_cursor_fini(&zc); 1886 1887 if (err != 0) { 1888 dsl_dir_rele(dd, FTAG); 1889 dsl_pool_config_exit(dp, FTAG); 1890 kmem_free(attr, sizeof (zap_attribute_t)); 1891 return (err); 1892 } 1893 } 1894 1895 /* 1896 * Iterate over all snapshots. 1897 */ 1898 if (flags & DS_FIND_SNAPSHOTS) { 1899 err = dsl_dataset_hold_obj(dp, thisobj, FTAG, &ds); 1900 1901 if (err == 0) { 1902 uint64_t snapobj; 1903 1904 snapobj = dsl_dataset_phys(ds)->ds_snapnames_zapobj; 1905 dsl_dataset_rele(ds, FTAG); 1906 1907 for (zap_cursor_init(&zc, dp->dp_meta_objset, snapobj); 1908 zap_cursor_retrieve(&zc, attr) == 0; 1909 (void) zap_cursor_advance(&zc)) { 1910 ASSERT3U(attr->za_integer_length, ==, 1911 sizeof (uint64_t)); 1912 ASSERT3U(attr->za_num_integers, ==, 1); 1913 1914 child = kmem_asprintf("%s@%s", 1915 name, attr->za_name); 1916 dsl_pool_config_exit(dp, FTAG); 1917 err = func(child, arg); 1918 dsl_pool_config_enter(dp, FTAG); 1919 strfree(child); 1920 if (err != 0) 1921 break; 1922 } 1923 zap_cursor_fini(&zc); 1924 } 1925 } 1926 1927 dsl_dir_rele(dd, FTAG); 1928 kmem_free(attr, sizeof (zap_attribute_t)); 1929 dsl_pool_config_exit(dp, FTAG); 1930 1931 if (err != 0) 1932 return (err); 1933 1934 /* Apply to self. */ 1935 return (func(name, arg)); 1936 } 1937 1938 /* 1939 * See comment above dmu_objset_find_impl(). 1940 */ 1941 int 1942 dmu_objset_find(char *name, int func(const char *, void *), void *arg, 1943 int flags) 1944 { 1945 spa_t *spa; 1946 int error; 1947 1948 error = spa_open(name, &spa, FTAG); 1949 if (error != 0) 1950 return (error); 1951 error = dmu_objset_find_impl(spa, name, func, arg, flags); 1952 spa_close(spa, FTAG); 1953 return (error); 1954 } 1955 1956 void 1957 dmu_objset_set_user(objset_t *os, void *user_ptr) 1958 { 1959 ASSERT(MUTEX_HELD(&os->os_user_ptr_lock)); 1960 os->os_user_ptr = user_ptr; 1961 } 1962 1963 void * 1964 dmu_objset_get_user(objset_t *os) 1965 { 1966 ASSERT(MUTEX_HELD(&os->os_user_ptr_lock)); 1967 return (os->os_user_ptr); 1968 } 1969 1970 /* 1971 * Determine name of filesystem, given name of snapshot. 1972 * buf must be at least MAXNAMELEN bytes 1973 */ 1974 int 1975 dmu_fsname(const char *snapname, char *buf) 1976 { 1977 char *atp = strchr(snapname, '@'); 1978 if (atp == NULL) 1979 return (SET_ERROR(EINVAL)); 1980 if (atp - snapname >= MAXNAMELEN) 1981 return (SET_ERROR(ENAMETOOLONG)); 1982 (void) strlcpy(buf, snapname, atp - snapname + 1); 1983 return (0); 1984 } 1985