Index: stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_object.c =================================================================== --- stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_object.c (revision 339135) +++ stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dmu_object.c (revision 339136) @@ -1,251 +1,255 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2013, 2017 by Delphix. All rights reserved. * Copyright 2014 HybridCluster. All rights reserved. */ #include #include #include #include #include #include uint64_t dmu_object_alloc_ibs(objset_t *os, dmu_object_type_t ot, int blocksize, int indirect_blockshift, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx) { uint64_t object; uint64_t L1_dnode_count = DNODES_PER_BLOCK << (DMU_META_DNODE(os)->dn_indblkshift - SPA_BLKPTRSHIFT); dnode_t *dn = NULL; mutex_enter(&os->os_obj_lock); for (;;) { object = os->os_obj_next; /* * Each time we polish off a L1 bp worth of dnodes (2^12 * objects), move to another L1 bp that's still reasonably * sparse (at most 1/4 full). Look from the beginning at most * once per txg, but after that keep looking from here. * os_scan_dnodes is set during txg sync if enough objects * have been freed since the previous rescan to justify * backfilling again. If we can't find a suitable block, just * keep going from here. * * Note that dmu_traverse depends on the behavior that we use * multiple blocks of the dnode object before going back to * reuse objects. Any change to this algorithm should preserve * that property or find another solution to the issues * described in traverse_visitbp. */ if (P2PHASE(object, L1_dnode_count) == 0) { uint64_t offset; int error; if (os->os_rescan_dnodes) { offset = 0; os->os_rescan_dnodes = B_FALSE; } else { offset = object << DNODE_SHIFT; } error = dnode_next_offset(DMU_META_DNODE(os), DNODE_FIND_HOLE, &offset, 2, DNODES_PER_BLOCK >> 2, 0); if (error == 0) object = offset >> DNODE_SHIFT; } os->os_obj_next = ++object; /* * XXX We should check for an i/o error here and return * up to our caller. Actually we should pre-read it in * dmu_tx_assign(), but there is currently no mechanism * to do so. */ (void) dnode_hold_impl(os, object, DNODE_MUST_BE_FREE, FTAG, &dn); if (dn) break; if (dmu_object_next(os, &object, B_TRUE, 0) == 0) os->os_obj_next = object - 1; } dnode_allocate(dn, ot, blocksize, indirect_blockshift, bonustype, bonuslen, tx); mutex_exit(&os->os_obj_lock); dmu_tx_add_new_object(tx, dn); dnode_rele(dn, FTAG); return (object); } uint64_t dmu_object_alloc(objset_t *os, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx) { return (dmu_object_alloc_ibs(os, ot, blocksize, 0, bonustype, bonuslen, tx)); } int dmu_object_claim(objset_t *os, uint64_t object, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx) { dnode_t *dn; int err; if (object == DMU_META_DNODE_OBJECT && !dmu_tx_private_ok(tx)) return (SET_ERROR(EBADF)); err = dnode_hold_impl(os, object, DNODE_MUST_BE_FREE, FTAG, &dn); if (err) return (err); dnode_allocate(dn, ot, blocksize, 0, bonustype, bonuslen, tx); dmu_tx_add_new_object(tx, dn); dnode_rele(dn, FTAG); return (0); } int dmu_object_reclaim(objset_t *os, uint64_t object, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx) { dnode_t *dn; int err; if (object == DMU_META_DNODE_OBJECT) return (SET_ERROR(EBADF)); err = dnode_hold_impl(os, object, DNODE_MUST_BE_ALLOCATED, FTAG, &dn); if (err) return (err); dnode_reallocate(dn, ot, blocksize, bonustype, bonuslen, tx); dnode_rele(dn, FTAG); return (err); } int dmu_object_free(objset_t *os, uint64_t object, dmu_tx_t *tx) { dnode_t *dn; int err; ASSERT(object != DMU_META_DNODE_OBJECT || dmu_tx_private_ok(tx)); err = dnode_hold_impl(os, object, DNODE_MUST_BE_ALLOCATED, FTAG, &dn); if (err) return (err); ASSERT(dn->dn_type != DMU_OT_NONE); + /* + * If we don't create this free range, we'll leak indirect blocks when + * we get to freeing the dnode in syncing context. + */ dnode_free_range(dn, 0, DMU_OBJECT_END, tx); dnode_free(dn, tx); dnode_rele(dn, FTAG); return (0); } /* * Return (in *objectp) the next object which is allocated (or a hole) * after *object, taking into account only objects that may have been modified * after the specified txg. */ int dmu_object_next(objset_t *os, uint64_t *objectp, boolean_t hole, uint64_t txg) { uint64_t offset = (*objectp + 1) << DNODE_SHIFT; int error; error = dnode_next_offset(DMU_META_DNODE(os), (hole ? DNODE_FIND_HOLE : 0), &offset, 0, DNODES_PER_BLOCK, txg); *objectp = offset >> DNODE_SHIFT; return (error); } /* * Turn this object from old_type into DMU_OTN_ZAP_METADATA, and bump the * refcount on SPA_FEATURE_EXTENSIBLE_DATASET. * * Only for use from syncing context, on MOS objects. */ void dmu_object_zapify(objset_t *mos, uint64_t object, dmu_object_type_t old_type, dmu_tx_t *tx) { dnode_t *dn; ASSERT(dmu_tx_is_syncing(tx)); VERIFY0(dnode_hold(mos, object, FTAG, &dn)); if (dn->dn_type == DMU_OTN_ZAP_METADATA) { dnode_rele(dn, FTAG); return; } ASSERT3U(dn->dn_type, ==, old_type); ASSERT0(dn->dn_maxblkid); /* * We must initialize the ZAP data before changing the type, * so that concurrent calls to *_is_zapified() can determine if * the object has been completely zapified by checking the type. */ mzap_create_impl(mos, object, 0, 0, tx); dn->dn_next_type[tx->tx_txg & TXG_MASK] = dn->dn_type = DMU_OTN_ZAP_METADATA; dnode_setdirty(dn, tx); dnode_rele(dn, FTAG); spa_feature_incr(dmu_objset_spa(mos), SPA_FEATURE_EXTENSIBLE_DATASET, tx); } void dmu_object_free_zapified(objset_t *mos, uint64_t object, dmu_tx_t *tx) { dnode_t *dn; dmu_object_type_t t; ASSERT(dmu_tx_is_syncing(tx)); VERIFY0(dnode_hold(mos, object, FTAG, &dn)); t = dn->dn_type; dnode_rele(dn, FTAG); if (t == DMU_OTN_ZAP_METADATA) { spa_feature_decr(dmu_objset_spa(mos), SPA_FEATURE_EXTENSIBLE_DATASET, tx); } VERIFY0(dmu_object_free(mos, object, tx)); } Index: stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode.c =================================================================== --- stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode.c (revision 339135) +++ stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode.c (revision 339136) @@ -1,2012 +1,2080 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2017 by Delphix. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. * Copyright (c) 2014 Integros [integros.com] * Copyright 2017 RackTop Systems. */ #include #include #include #include #include #include #include #include #include #include #include #include #include static kmem_cache_t *dnode_cache; /* * Define DNODE_STATS to turn on statistic gathering. By default, it is only * turned on when DEBUG is also defined. */ #ifdef DEBUG #define DNODE_STATS #endif /* DEBUG */ #ifdef DNODE_STATS #define DNODE_STAT_ADD(stat) ((stat)++) #else #define DNODE_STAT_ADD(stat) /* nothing */ #endif /* DNODE_STATS */ static dnode_phys_t dnode_phys_zero; int zfs_default_bs = SPA_MINBLOCKSHIFT; int zfs_default_ibs = DN_MAX_INDBLKSHIFT; SYSCTL_DECL(_vfs_zfs); SYSCTL_INT(_vfs_zfs, OID_AUTO, default_bs, CTLFLAG_RWTUN, &zfs_default_bs, 0, "Default dnode block shift"); SYSCTL_INT(_vfs_zfs, OID_AUTO, default_ibs, CTLFLAG_RWTUN, &zfs_default_ibs, 0, "Default dnode indirect block shift"); #ifdef illumos #ifdef _KERNEL static kmem_cbrc_t dnode_move(void *, void *, size_t, void *); #endif /* _KERNEL */ #endif static int dbuf_compare(const void *x1, const void *x2) { const dmu_buf_impl_t *d1 = x1; const dmu_buf_impl_t *d2 = x2; if (d1->db_level < d2->db_level) { return (-1); } if (d1->db_level > d2->db_level) { return (1); } if (d1->db_blkid < d2->db_blkid) { return (-1); } if (d1->db_blkid > d2->db_blkid) { return (1); } if (d1->db_state == DB_SEARCH) { ASSERT3S(d2->db_state, !=, DB_SEARCH); return (-1); } else if (d2->db_state == DB_SEARCH) { ASSERT3S(d1->db_state, !=, DB_SEARCH); return (1); } if ((uintptr_t)d1 < (uintptr_t)d2) { return (-1); } if ((uintptr_t)d1 > (uintptr_t)d2) { return (1); } return (0); } /* ARGSUSED */ static int dnode_cons(void *arg, void *unused, int kmflag) { dnode_t *dn = arg; int i; rw_init(&dn->dn_struct_rwlock, NULL, RW_DEFAULT, NULL); mutex_init(&dn->dn_mtx, NULL, MUTEX_DEFAULT, NULL); mutex_init(&dn->dn_dbufs_mtx, NULL, MUTEX_DEFAULT, NULL); cv_init(&dn->dn_notxholds, NULL, CV_DEFAULT, NULL); /* * Every dbuf has a reference, and dropping a tracked reference is * O(number of references), so don't track dn_holds. */ refcount_create_untracked(&dn->dn_holds); refcount_create(&dn->dn_tx_holds); list_link_init(&dn->dn_link); bzero(&dn->dn_next_nblkptr[0], sizeof (dn->dn_next_nblkptr)); bzero(&dn->dn_next_nlevels[0], sizeof (dn->dn_next_nlevels)); bzero(&dn->dn_next_indblkshift[0], sizeof (dn->dn_next_indblkshift)); bzero(&dn->dn_next_bonustype[0], sizeof (dn->dn_next_bonustype)); bzero(&dn->dn_rm_spillblk[0], sizeof (dn->dn_rm_spillblk)); bzero(&dn->dn_next_bonuslen[0], sizeof (dn->dn_next_bonuslen)); bzero(&dn->dn_next_blksz[0], sizeof (dn->dn_next_blksz)); for (i = 0; i < TXG_SIZE; i++) { list_link_init(&dn->dn_dirty_link[i]); dn->dn_free_ranges[i] = NULL; list_create(&dn->dn_dirty_records[i], sizeof (dbuf_dirty_record_t), offsetof(dbuf_dirty_record_t, dr_dirty_node)); } dn->dn_allocated_txg = 0; dn->dn_free_txg = 0; dn->dn_assigned_txg = 0; dn->dn_dirtyctx = 0; dn->dn_dirtyctx_firstset = NULL; dn->dn_bonus = NULL; dn->dn_have_spill = B_FALSE; dn->dn_zio = NULL; dn->dn_oldused = 0; dn->dn_oldflags = 0; dn->dn_olduid = 0; dn->dn_oldgid = 0; dn->dn_newuid = 0; dn->dn_newgid = 0; dn->dn_id_flags = 0; dn->dn_dbufs_count = 0; avl_create(&dn->dn_dbufs, dbuf_compare, sizeof (dmu_buf_impl_t), offsetof(dmu_buf_impl_t, db_link)); dn->dn_moved = 0; POINTER_INVALIDATE(&dn->dn_objset); return (0); } /* ARGSUSED */ static void dnode_dest(void *arg, void *unused) { int i; dnode_t *dn = arg; rw_destroy(&dn->dn_struct_rwlock); mutex_destroy(&dn->dn_mtx); mutex_destroy(&dn->dn_dbufs_mtx); cv_destroy(&dn->dn_notxholds); refcount_destroy(&dn->dn_holds); refcount_destroy(&dn->dn_tx_holds); ASSERT(!list_link_active(&dn->dn_link)); for (i = 0; i < TXG_SIZE; i++) { ASSERT(!list_link_active(&dn->dn_dirty_link[i])); ASSERT3P(dn->dn_free_ranges[i], ==, NULL); list_destroy(&dn->dn_dirty_records[i]); ASSERT0(dn->dn_next_nblkptr[i]); ASSERT0(dn->dn_next_nlevels[i]); ASSERT0(dn->dn_next_indblkshift[i]); ASSERT0(dn->dn_next_bonustype[i]); ASSERT0(dn->dn_rm_spillblk[i]); ASSERT0(dn->dn_next_bonuslen[i]); ASSERT0(dn->dn_next_blksz[i]); } ASSERT0(dn->dn_allocated_txg); ASSERT0(dn->dn_free_txg); ASSERT0(dn->dn_assigned_txg); ASSERT0(dn->dn_dirtyctx); ASSERT3P(dn->dn_dirtyctx_firstset, ==, NULL); ASSERT3P(dn->dn_bonus, ==, NULL); ASSERT(!dn->dn_have_spill); ASSERT3P(dn->dn_zio, ==, NULL); ASSERT0(dn->dn_oldused); ASSERT0(dn->dn_oldflags); ASSERT0(dn->dn_olduid); ASSERT0(dn->dn_oldgid); ASSERT0(dn->dn_newuid); ASSERT0(dn->dn_newgid); ASSERT0(dn->dn_id_flags); ASSERT0(dn->dn_dbufs_count); avl_destroy(&dn->dn_dbufs); } void dnode_init(void) { ASSERT(dnode_cache == NULL); dnode_cache = kmem_cache_create("dnode_t", sizeof (dnode_t), 0, dnode_cons, dnode_dest, NULL, NULL, NULL, 0); #ifdef _KERNEL kmem_cache_set_move(dnode_cache, dnode_move); #endif /* _KERNEL */ } void dnode_fini(void) { kmem_cache_destroy(dnode_cache); dnode_cache = NULL; } #ifdef ZFS_DEBUG void dnode_verify(dnode_t *dn) { int drop_struct_lock = FALSE; ASSERT(dn->dn_phys); ASSERT(dn->dn_objset); ASSERT(dn->dn_handle->dnh_dnode == dn); ASSERT(DMU_OT_IS_VALID(dn->dn_phys->dn_type)); if (!(zfs_flags & ZFS_DEBUG_DNODE_VERIFY)) return; if (!RW_WRITE_HELD(&dn->dn_struct_rwlock)) { rw_enter(&dn->dn_struct_rwlock, RW_READER); drop_struct_lock = TRUE; } if (dn->dn_phys->dn_type != DMU_OT_NONE || dn->dn_allocated_txg != 0) { int i; ASSERT3U(dn->dn_indblkshift, >=, 0); ASSERT3U(dn->dn_indblkshift, <=, SPA_MAXBLOCKSHIFT); if (dn->dn_datablkshift) { ASSERT3U(dn->dn_datablkshift, >=, SPA_MINBLOCKSHIFT); ASSERT3U(dn->dn_datablkshift, <=, SPA_MAXBLOCKSHIFT); ASSERT3U(1<dn_datablkshift, ==, dn->dn_datablksz); } ASSERT3U(dn->dn_nlevels, <=, 30); ASSERT(DMU_OT_IS_VALID(dn->dn_type)); ASSERT3U(dn->dn_nblkptr, >=, 1); ASSERT3U(dn->dn_nblkptr, <=, DN_MAX_NBLKPTR); ASSERT3U(dn->dn_bonuslen, <=, DN_MAX_BONUSLEN); ASSERT3U(dn->dn_datablksz, ==, dn->dn_datablkszsec << SPA_MINBLOCKSHIFT); ASSERT3U(ISP2(dn->dn_datablksz), ==, dn->dn_datablkshift != 0); ASSERT3U((dn->dn_nblkptr - 1) * sizeof (blkptr_t) + dn->dn_bonuslen, <=, DN_MAX_BONUSLEN); for (i = 0; i < TXG_SIZE; i++) { ASSERT3U(dn->dn_next_nlevels[i], <=, dn->dn_nlevels); } } if (dn->dn_phys->dn_type != DMU_OT_NONE) ASSERT3U(dn->dn_phys->dn_nlevels, <=, dn->dn_nlevels); ASSERT(DMU_OBJECT_IS_SPECIAL(dn->dn_object) || dn->dn_dbuf != NULL); if (dn->dn_dbuf != NULL) { ASSERT3P(dn->dn_phys, ==, (dnode_phys_t *)dn->dn_dbuf->db.db_data + (dn->dn_object % (dn->dn_dbuf->db.db_size >> DNODE_SHIFT))); } if (drop_struct_lock) rw_exit(&dn->dn_struct_rwlock); } #endif void dnode_byteswap(dnode_phys_t *dnp) { uint64_t *buf64 = (void*)&dnp->dn_blkptr; int i; if (dnp->dn_type == DMU_OT_NONE) { bzero(dnp, sizeof (dnode_phys_t)); return; } dnp->dn_datablkszsec = BSWAP_16(dnp->dn_datablkszsec); dnp->dn_bonuslen = BSWAP_16(dnp->dn_bonuslen); dnp->dn_maxblkid = BSWAP_64(dnp->dn_maxblkid); dnp->dn_used = BSWAP_64(dnp->dn_used); /* * dn_nblkptr is only one byte, so it's OK to read it in either * byte order. We can't read dn_bouslen. */ ASSERT(dnp->dn_indblkshift <= SPA_MAXBLOCKSHIFT); ASSERT(dnp->dn_nblkptr <= DN_MAX_NBLKPTR); for (i = 0; i < dnp->dn_nblkptr * sizeof (blkptr_t)/8; i++) buf64[i] = BSWAP_64(buf64[i]); /* * OK to check dn_bonuslen for zero, because it won't matter if * we have the wrong byte order. This is necessary because the * dnode dnode is smaller than a regular dnode. */ if (dnp->dn_bonuslen != 0) { /* * Note that the bonus length calculated here may be * longer than the actual bonus buffer. This is because * we always put the bonus buffer after the last block * pointer (instead of packing it against the end of the * dnode buffer). */ int off = (dnp->dn_nblkptr-1) * sizeof (blkptr_t); size_t len = DN_MAX_BONUSLEN - off; ASSERT(DMU_OT_IS_VALID(dnp->dn_bonustype)); dmu_object_byteswap_t byteswap = DMU_OT_BYTESWAP(dnp->dn_bonustype); dmu_ot_byteswap[byteswap].ob_func(dnp->dn_bonus + off, len); } /* Swap SPILL block if we have one */ if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) byteswap_uint64_array(&dnp->dn_spill, sizeof (blkptr_t)); } void dnode_buf_byteswap(void *vbuf, size_t size) { dnode_phys_t *buf = vbuf; int i; ASSERT3U(sizeof (dnode_phys_t), ==, (1<>= DNODE_SHIFT; for (i = 0; i < size; i++) { dnode_byteswap(buf); buf++; } } void dnode_setbonuslen(dnode_t *dn, int newsize, dmu_tx_t *tx) { ASSERT3U(refcount_count(&dn->dn_holds), >=, 1); dnode_setdirty(dn, tx); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); ASSERT3U(newsize, <=, DN_MAX_BONUSLEN - (dn->dn_nblkptr-1) * sizeof (blkptr_t)); dn->dn_bonuslen = newsize; if (newsize == 0) dn->dn_next_bonuslen[tx->tx_txg & TXG_MASK] = DN_ZERO_BONUSLEN; else dn->dn_next_bonuslen[tx->tx_txg & TXG_MASK] = dn->dn_bonuslen; rw_exit(&dn->dn_struct_rwlock); } void dnode_setbonus_type(dnode_t *dn, dmu_object_type_t newtype, dmu_tx_t *tx) { ASSERT3U(refcount_count(&dn->dn_holds), >=, 1); dnode_setdirty(dn, tx); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); dn->dn_bonustype = newtype; dn->dn_next_bonustype[tx->tx_txg & TXG_MASK] = dn->dn_bonustype; rw_exit(&dn->dn_struct_rwlock); } void dnode_rm_spill(dnode_t *dn, dmu_tx_t *tx) { ASSERT3U(refcount_count(&dn->dn_holds), >=, 1); ASSERT(RW_WRITE_HELD(&dn->dn_struct_rwlock)); dnode_setdirty(dn, tx); dn->dn_rm_spillblk[tx->tx_txg&TXG_MASK] = DN_KILL_SPILLBLK; dn->dn_have_spill = B_FALSE; } static void dnode_setdblksz(dnode_t *dn, int size) { ASSERT0(P2PHASE(size, SPA_MINBLOCKSIZE)); ASSERT3U(size, <=, SPA_MAXBLOCKSIZE); ASSERT3U(size, >=, SPA_MINBLOCKSIZE); ASSERT3U(size >> SPA_MINBLOCKSHIFT, <, 1<<(sizeof (dn->dn_phys->dn_datablkszsec) * 8)); dn->dn_datablksz = size; dn->dn_datablkszsec = size >> SPA_MINBLOCKSHIFT; dn->dn_datablkshift = ISP2(size) ? highbit64(size - 1) : 0; } static dnode_t * dnode_create(objset_t *os, dnode_phys_t *dnp, dmu_buf_impl_t *db, uint64_t object, dnode_handle_t *dnh) { dnode_t *dn; dn = kmem_cache_alloc(dnode_cache, KM_SLEEP); #ifdef _KERNEL ASSERT(!POINTER_IS_VALID(dn->dn_objset)); #endif /* _KERNEL */ dn->dn_moved = 0; /* * Defer setting dn_objset until the dnode is ready to be a candidate * for the dnode_move() callback. */ dn->dn_object = object; dn->dn_dbuf = db; dn->dn_handle = dnh; dn->dn_phys = dnp; if (dnp->dn_datablkszsec) { dnode_setdblksz(dn, dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT); } else { dn->dn_datablksz = 0; dn->dn_datablkszsec = 0; dn->dn_datablkshift = 0; } dn->dn_indblkshift = dnp->dn_indblkshift; dn->dn_nlevels = dnp->dn_nlevels; dn->dn_type = dnp->dn_type; dn->dn_nblkptr = dnp->dn_nblkptr; dn->dn_checksum = dnp->dn_checksum; dn->dn_compress = dnp->dn_compress; dn->dn_bonustype = dnp->dn_bonustype; dn->dn_bonuslen = dnp->dn_bonuslen; dn->dn_maxblkid = dnp->dn_maxblkid; dn->dn_have_spill = ((dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) != 0); dn->dn_id_flags = 0; dmu_zfetch_init(&dn->dn_zfetch, dn); ASSERT(DMU_OT_IS_VALID(dn->dn_phys->dn_type)); mutex_enter(&os->os_lock); if (dnh->dnh_dnode != NULL) { /* Lost the allocation race. */ mutex_exit(&os->os_lock); kmem_cache_free(dnode_cache, dn); return (dnh->dnh_dnode); } /* * Exclude special dnodes from os_dnodes so an empty os_dnodes * signifies that the special dnodes have no references from * their children (the entries in os_dnodes). This allows * dnode_destroy() to easily determine if the last child has * been removed and then complete eviction of the objset. */ if (!DMU_OBJECT_IS_SPECIAL(object)) list_insert_head(&os->os_dnodes, dn); membar_producer(); /* * Everything else must be valid before assigning dn_objset * makes the dnode eligible for dnode_move(). */ dn->dn_objset = os; dnh->dnh_dnode = dn; mutex_exit(&os->os_lock); arc_space_consume(sizeof (dnode_t), ARC_SPACE_OTHER); return (dn); } /* * Caller must be holding the dnode handle, which is released upon return. */ static void dnode_destroy(dnode_t *dn) { objset_t *os = dn->dn_objset; boolean_t complete_os_eviction = B_FALSE; ASSERT((dn->dn_id_flags & DN_ID_NEW_EXIST) == 0); mutex_enter(&os->os_lock); POINTER_INVALIDATE(&dn->dn_objset); if (!DMU_OBJECT_IS_SPECIAL(dn->dn_object)) { list_remove(&os->os_dnodes, dn); complete_os_eviction = list_is_empty(&os->os_dnodes) && list_link_active(&os->os_evicting_node); } mutex_exit(&os->os_lock); /* the dnode can no longer move, so we can release the handle */ zrl_remove(&dn->dn_handle->dnh_zrlock); dn->dn_allocated_txg = 0; dn->dn_free_txg = 0; dn->dn_assigned_txg = 0; dn->dn_dirtyctx = 0; if (dn->dn_dirtyctx_firstset != NULL) { kmem_free(dn->dn_dirtyctx_firstset, 1); dn->dn_dirtyctx_firstset = NULL; } if (dn->dn_bonus != NULL) { mutex_enter(&dn->dn_bonus->db_mtx); dbuf_destroy(dn->dn_bonus); dn->dn_bonus = NULL; } dn->dn_zio = NULL; dn->dn_have_spill = B_FALSE; dn->dn_oldused = 0; dn->dn_oldflags = 0; dn->dn_olduid = 0; dn->dn_oldgid = 0; dn->dn_newuid = 0; dn->dn_newgid = 0; dn->dn_id_flags = 0; dmu_zfetch_fini(&dn->dn_zfetch); kmem_cache_free(dnode_cache, dn); arc_space_return(sizeof (dnode_t), ARC_SPACE_OTHER); if (complete_os_eviction) dmu_objset_evict_done(os); } void dnode_allocate(dnode_t *dn, dmu_object_type_t ot, int blocksize, int ibs, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx) { int i; ASSERT3U(blocksize, <=, spa_maxblocksize(dmu_objset_spa(dn->dn_objset))); if (blocksize == 0) blocksize = 1 << zfs_default_bs; else blocksize = P2ROUNDUP(blocksize, SPA_MINBLOCKSIZE); if (ibs == 0) ibs = zfs_default_ibs; ibs = MIN(MAX(ibs, DN_MIN_INDBLKSHIFT), DN_MAX_INDBLKSHIFT); dprintf("os=%p obj=%llu txg=%llu blocksize=%d ibs=%d\n", dn->dn_objset, dn->dn_object, tx->tx_txg, blocksize, ibs); ASSERT(dn->dn_type == DMU_OT_NONE); ASSERT(bcmp(dn->dn_phys, &dnode_phys_zero, sizeof (dnode_phys_t)) == 0); ASSERT(dn->dn_phys->dn_type == DMU_OT_NONE); ASSERT(ot != DMU_OT_NONE); ASSERT(DMU_OT_IS_VALID(ot)); ASSERT((bonustype == DMU_OT_NONE && bonuslen == 0) || (bonustype == DMU_OT_SA && bonuslen == 0) || (bonustype != DMU_OT_NONE && bonuslen != 0)); ASSERT(DMU_OT_IS_VALID(bonustype)); ASSERT3U(bonuslen, <=, DN_MAX_BONUSLEN); ASSERT(dn->dn_type == DMU_OT_NONE); ASSERT0(dn->dn_maxblkid); ASSERT0(dn->dn_allocated_txg); ASSERT0(dn->dn_assigned_txg); ASSERT(refcount_is_zero(&dn->dn_tx_holds)); ASSERT3U(refcount_count(&dn->dn_holds), <=, 1); ASSERT(avl_is_empty(&dn->dn_dbufs)); for (i = 0; i < TXG_SIZE; i++) { ASSERT0(dn->dn_next_nblkptr[i]); ASSERT0(dn->dn_next_nlevels[i]); ASSERT0(dn->dn_next_indblkshift[i]); ASSERT0(dn->dn_next_bonuslen[i]); ASSERT0(dn->dn_next_bonustype[i]); ASSERT0(dn->dn_rm_spillblk[i]); ASSERT0(dn->dn_next_blksz[i]); ASSERT(!list_link_active(&dn->dn_dirty_link[i])); ASSERT3P(list_head(&dn->dn_dirty_records[i]), ==, NULL); ASSERT3P(dn->dn_free_ranges[i], ==, NULL); } dn->dn_type = ot; dnode_setdblksz(dn, blocksize); dn->dn_indblkshift = ibs; dn->dn_nlevels = 1; if (bonustype == DMU_OT_SA) /* Maximize bonus space for SA */ dn->dn_nblkptr = 1; else dn->dn_nblkptr = 1 + ((DN_MAX_BONUSLEN - bonuslen) >> SPA_BLKPTRSHIFT); dn->dn_bonustype = bonustype; dn->dn_bonuslen = bonuslen; dn->dn_checksum = ZIO_CHECKSUM_INHERIT; dn->dn_compress = ZIO_COMPRESS_INHERIT; dn->dn_dirtyctx = 0; dn->dn_free_txg = 0; if (dn->dn_dirtyctx_firstset) { kmem_free(dn->dn_dirtyctx_firstset, 1); dn->dn_dirtyctx_firstset = NULL; } dn->dn_allocated_txg = tx->tx_txg; dn->dn_id_flags = 0; dnode_setdirty(dn, tx); dn->dn_next_indblkshift[tx->tx_txg & TXG_MASK] = ibs; dn->dn_next_bonuslen[tx->tx_txg & TXG_MASK] = dn->dn_bonuslen; dn->dn_next_bonustype[tx->tx_txg & TXG_MASK] = dn->dn_bonustype; dn->dn_next_blksz[tx->tx_txg & TXG_MASK] = dn->dn_datablksz; } void dnode_reallocate(dnode_t *dn, dmu_object_type_t ot, int blocksize, dmu_object_type_t bonustype, int bonuslen, dmu_tx_t *tx) { int nblkptr; ASSERT3U(blocksize, >=, SPA_MINBLOCKSIZE); ASSERT3U(blocksize, <=, spa_maxblocksize(dmu_objset_spa(dn->dn_objset))); ASSERT0(blocksize % SPA_MINBLOCKSIZE); ASSERT(dn->dn_object != DMU_META_DNODE_OBJECT || dmu_tx_private_ok(tx)); ASSERT(tx->tx_txg != 0); ASSERT((bonustype == DMU_OT_NONE && bonuslen == 0) || (bonustype != DMU_OT_NONE && bonuslen != 0) || (bonustype == DMU_OT_SA && bonuslen == 0)); ASSERT(DMU_OT_IS_VALID(bonustype)); ASSERT3U(bonuslen, <=, DN_MAX_BONUSLEN); /* clean up any unreferenced dbufs */ dnode_evict_dbufs(dn); dn->dn_id_flags = 0; rw_enter(&dn->dn_struct_rwlock, RW_WRITER); dnode_setdirty(dn, tx); if (dn->dn_datablksz != blocksize) { /* change blocksize */ ASSERT(dn->dn_maxblkid == 0 && (BP_IS_HOLE(&dn->dn_phys->dn_blkptr[0]) || dnode_block_freed(dn, 0))); dnode_setdblksz(dn, blocksize); dn->dn_next_blksz[tx->tx_txg&TXG_MASK] = blocksize; } if (dn->dn_bonuslen != bonuslen) dn->dn_next_bonuslen[tx->tx_txg&TXG_MASK] = bonuslen; if (bonustype == DMU_OT_SA) /* Maximize bonus space for SA */ nblkptr = 1; else nblkptr = 1 + ((DN_MAX_BONUSLEN - bonuslen) >> SPA_BLKPTRSHIFT); if (dn->dn_bonustype != bonustype) dn->dn_next_bonustype[tx->tx_txg&TXG_MASK] = bonustype; if (dn->dn_nblkptr != nblkptr) dn->dn_next_nblkptr[tx->tx_txg&TXG_MASK] = nblkptr; if (dn->dn_phys->dn_flags & DNODE_FLAG_SPILL_BLKPTR) { dbuf_rm_spill(dn, tx); dnode_rm_spill(dn, tx); } rw_exit(&dn->dn_struct_rwlock); /* change type */ dn->dn_type = ot; /* change bonus size and type */ mutex_enter(&dn->dn_mtx); dn->dn_bonustype = bonustype; dn->dn_bonuslen = bonuslen; dn->dn_nblkptr = nblkptr; dn->dn_checksum = ZIO_CHECKSUM_INHERIT; dn->dn_compress = ZIO_COMPRESS_INHERIT; ASSERT3U(dn->dn_nblkptr, <=, DN_MAX_NBLKPTR); /* fix up the bonus db_size */ if (dn->dn_bonus) { dn->dn_bonus->db.db_size = DN_MAX_BONUSLEN - (dn->dn_nblkptr-1) * sizeof (blkptr_t); ASSERT(dn->dn_bonuslen <= dn->dn_bonus->db.db_size); } dn->dn_allocated_txg = tx->tx_txg; mutex_exit(&dn->dn_mtx); } #ifdef DNODE_STATS static struct { uint64_t dms_dnode_invalid; uint64_t dms_dnode_recheck1; uint64_t dms_dnode_recheck2; uint64_t dms_dnode_special; uint64_t dms_dnode_handle; uint64_t dms_dnode_rwlock; uint64_t dms_dnode_active; } dnode_move_stats; #endif /* DNODE_STATS */ #ifdef _KERNEL static void dnode_move_impl(dnode_t *odn, dnode_t *ndn) { int i; ASSERT(!RW_LOCK_HELD(&odn->dn_struct_rwlock)); ASSERT(MUTEX_NOT_HELD(&odn->dn_mtx)); ASSERT(MUTEX_NOT_HELD(&odn->dn_dbufs_mtx)); ASSERT(!RW_LOCK_HELD(&odn->dn_zfetch.zf_rwlock)); /* Copy fields. */ ndn->dn_objset = odn->dn_objset; ndn->dn_object = odn->dn_object; ndn->dn_dbuf = odn->dn_dbuf; ndn->dn_handle = odn->dn_handle; ndn->dn_phys = odn->dn_phys; ndn->dn_type = odn->dn_type; ndn->dn_bonuslen = odn->dn_bonuslen; ndn->dn_bonustype = odn->dn_bonustype; ndn->dn_nblkptr = odn->dn_nblkptr; ndn->dn_checksum = odn->dn_checksum; ndn->dn_compress = odn->dn_compress; ndn->dn_nlevels = odn->dn_nlevels; ndn->dn_indblkshift = odn->dn_indblkshift; ndn->dn_datablkshift = odn->dn_datablkshift; ndn->dn_datablkszsec = odn->dn_datablkszsec; ndn->dn_datablksz = odn->dn_datablksz; ndn->dn_maxblkid = odn->dn_maxblkid; bcopy(&odn->dn_next_nblkptr[0], &ndn->dn_next_nblkptr[0], sizeof (odn->dn_next_nblkptr)); bcopy(&odn->dn_next_nlevels[0], &ndn->dn_next_nlevels[0], sizeof (odn->dn_next_nlevels)); bcopy(&odn->dn_next_indblkshift[0], &ndn->dn_next_indblkshift[0], sizeof (odn->dn_next_indblkshift)); bcopy(&odn->dn_next_bonustype[0], &ndn->dn_next_bonustype[0], sizeof (odn->dn_next_bonustype)); bcopy(&odn->dn_rm_spillblk[0], &ndn->dn_rm_spillblk[0], sizeof (odn->dn_rm_spillblk)); bcopy(&odn->dn_next_bonuslen[0], &ndn->dn_next_bonuslen[0], sizeof (odn->dn_next_bonuslen)); bcopy(&odn->dn_next_blksz[0], &ndn->dn_next_blksz[0], sizeof (odn->dn_next_blksz)); for (i = 0; i < TXG_SIZE; i++) { list_move_tail(&ndn->dn_dirty_records[i], &odn->dn_dirty_records[i]); } bcopy(&odn->dn_free_ranges[0], &ndn->dn_free_ranges[0], sizeof (odn->dn_free_ranges)); ndn->dn_allocated_txg = odn->dn_allocated_txg; ndn->dn_free_txg = odn->dn_free_txg; ndn->dn_assigned_txg = odn->dn_assigned_txg; ndn->dn_dirtyctx = odn->dn_dirtyctx; ndn->dn_dirtyctx_firstset = odn->dn_dirtyctx_firstset; ASSERT(refcount_count(&odn->dn_tx_holds) == 0); refcount_transfer(&ndn->dn_holds, &odn->dn_holds); ASSERT(avl_is_empty(&ndn->dn_dbufs)); avl_swap(&ndn->dn_dbufs, &odn->dn_dbufs); ndn->dn_dbufs_count = odn->dn_dbufs_count; ndn->dn_bonus = odn->dn_bonus; ndn->dn_have_spill = odn->dn_have_spill; ndn->dn_zio = odn->dn_zio; ndn->dn_oldused = odn->dn_oldused; ndn->dn_oldflags = odn->dn_oldflags; ndn->dn_olduid = odn->dn_olduid; ndn->dn_oldgid = odn->dn_oldgid; ndn->dn_newuid = odn->dn_newuid; ndn->dn_newgid = odn->dn_newgid; ndn->dn_id_flags = odn->dn_id_flags; dmu_zfetch_init(&ndn->dn_zfetch, NULL); list_move_tail(&ndn->dn_zfetch.zf_stream, &odn->dn_zfetch.zf_stream); ndn->dn_zfetch.zf_dnode = odn->dn_zfetch.zf_dnode; /* * Update back pointers. Updating the handle fixes the back pointer of * every descendant dbuf as well as the bonus dbuf. */ ASSERT(ndn->dn_handle->dnh_dnode == odn); ndn->dn_handle->dnh_dnode = ndn; if (ndn->dn_zfetch.zf_dnode == odn) { ndn->dn_zfetch.zf_dnode = ndn; } /* * Invalidate the original dnode by clearing all of its back pointers. */ odn->dn_dbuf = NULL; odn->dn_handle = NULL; avl_create(&odn->dn_dbufs, dbuf_compare, sizeof (dmu_buf_impl_t), offsetof(dmu_buf_impl_t, db_link)); odn->dn_dbufs_count = 0; odn->dn_bonus = NULL; odn->dn_zfetch.zf_dnode = NULL; /* * Set the low bit of the objset pointer to ensure that dnode_move() * recognizes the dnode as invalid in any subsequent callback. */ POINTER_INVALIDATE(&odn->dn_objset); /* * Satisfy the destructor. */ for (i = 0; i < TXG_SIZE; i++) { list_create(&odn->dn_dirty_records[i], sizeof (dbuf_dirty_record_t), offsetof(dbuf_dirty_record_t, dr_dirty_node)); odn->dn_free_ranges[i] = NULL; odn->dn_next_nlevels[i] = 0; odn->dn_next_indblkshift[i] = 0; odn->dn_next_bonustype[i] = 0; odn->dn_rm_spillblk[i] = 0; odn->dn_next_bonuslen[i] = 0; odn->dn_next_blksz[i] = 0; } odn->dn_allocated_txg = 0; odn->dn_free_txg = 0; odn->dn_assigned_txg = 0; odn->dn_dirtyctx = 0; odn->dn_dirtyctx_firstset = NULL; odn->dn_have_spill = B_FALSE; odn->dn_zio = NULL; odn->dn_oldused = 0; odn->dn_oldflags = 0; odn->dn_olduid = 0; odn->dn_oldgid = 0; odn->dn_newuid = 0; odn->dn_newgid = 0; odn->dn_id_flags = 0; /* * Mark the dnode. */ ndn->dn_moved = 1; odn->dn_moved = (uint8_t)-1; } #ifdef illumos /*ARGSUSED*/ static kmem_cbrc_t dnode_move(void *buf, void *newbuf, size_t size, void *arg) { dnode_t *odn = buf, *ndn = newbuf; objset_t *os; int64_t refcount; uint32_t dbufs; /* * The dnode is on the objset's list of known dnodes if the objset * pointer is valid. We set the low bit of the objset pointer when * freeing the dnode to invalidate it, and the memory patterns written * by kmem (baddcafe and deadbeef) set at least one of the two low bits. * A newly created dnode sets the objset pointer last of all to indicate * that the dnode is known and in a valid state to be moved by this * function. */ os = odn->dn_objset; if (!POINTER_IS_VALID(os)) { DNODE_STAT_ADD(dnode_move_stats.dms_dnode_invalid); return (KMEM_CBRC_DONT_KNOW); } /* * Ensure that the objset does not go away during the move. */ rw_enter(&os_lock, RW_WRITER); if (os != odn->dn_objset) { rw_exit(&os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_recheck1); return (KMEM_CBRC_DONT_KNOW); } /* * If the dnode is still valid, then so is the objset. We know that no * valid objset can be freed while we hold os_lock, so we can safely * ensure that the objset remains in use. */ mutex_enter(&os->os_lock); /* * Recheck the objset pointer in case the dnode was removed just before * acquiring the lock. */ if (os != odn->dn_objset) { mutex_exit(&os->os_lock); rw_exit(&os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_recheck2); return (KMEM_CBRC_DONT_KNOW); } /* * At this point we know that as long as we hold os->os_lock, the dnode * cannot be freed and fields within the dnode can be safely accessed. * The objset listing this dnode cannot go away as long as this dnode is * on its list. */ rw_exit(&os_lock); if (DMU_OBJECT_IS_SPECIAL(odn->dn_object)) { mutex_exit(&os->os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_special); return (KMEM_CBRC_NO); } ASSERT(odn->dn_dbuf != NULL); /* only "special" dnodes have no parent */ /* * Lock the dnode handle to prevent the dnode from obtaining any new * holds. This also prevents the descendant dbufs and the bonus dbuf * from accessing the dnode, so that we can discount their holds. The * handle is safe to access because we know that while the dnode cannot * go away, neither can its handle. Once we hold dnh_zrlock, we can * safely move any dnode referenced only by dbufs. */ if (!zrl_tryenter(&odn->dn_handle->dnh_zrlock)) { mutex_exit(&os->os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_handle); return (KMEM_CBRC_LATER); } /* * Ensure a consistent view of the dnode's holds and the dnode's dbufs. * We need to guarantee that there is a hold for every dbuf in order to * determine whether the dnode is actively referenced. Falsely matching * a dbuf to an active hold would lead to an unsafe move. It's possible * that a thread already having an active dnode hold is about to add a * dbuf, and we can't compare hold and dbuf counts while the add is in * progress. */ if (!rw_tryenter(&odn->dn_struct_rwlock, RW_WRITER)) { zrl_exit(&odn->dn_handle->dnh_zrlock); mutex_exit(&os->os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_rwlock); return (KMEM_CBRC_LATER); } /* * A dbuf may be removed (evicted) without an active dnode hold. In that * case, the dbuf count is decremented under the handle lock before the * dbuf's hold is released. This order ensures that if we count the hold * after the dbuf is removed but before its hold is released, we will * treat the unmatched hold as active and exit safely. If we count the * hold before the dbuf is removed, the hold is discounted, and the * removal is blocked until the move completes. */ refcount = refcount_count(&odn->dn_holds); ASSERT(refcount >= 0); dbufs = odn->dn_dbufs_count; /* We can't have more dbufs than dnode holds. */ ASSERT3U(dbufs, <=, refcount); DTRACE_PROBE3(dnode__move, dnode_t *, odn, int64_t, refcount, uint32_t, dbufs); if (refcount > dbufs) { rw_exit(&odn->dn_struct_rwlock); zrl_exit(&odn->dn_handle->dnh_zrlock); mutex_exit(&os->os_lock); DNODE_STAT_ADD(dnode_move_stats.dms_dnode_active); return (KMEM_CBRC_LATER); } rw_exit(&odn->dn_struct_rwlock); /* * At this point we know that anyone with a hold on the dnode is not * actively referencing it. The dnode is known and in a valid state to * move. We're holding the locks needed to execute the critical section. */ dnode_move_impl(odn, ndn); list_link_replace(&odn->dn_link, &ndn->dn_link); /* If the dnode was safe to move, the refcount cannot have changed. */ ASSERT(refcount == refcount_count(&ndn->dn_holds)); ASSERT(dbufs == ndn->dn_dbufs_count); zrl_exit(&ndn->dn_handle->dnh_zrlock); /* handle has moved */ mutex_exit(&os->os_lock); return (KMEM_CBRC_YES); } #endif /* illumos */ #endif /* _KERNEL */ void dnode_special_close(dnode_handle_t *dnh) { dnode_t *dn = dnh->dnh_dnode; /* * Wait for final references to the dnode to clear. This can * only happen if the arc is asyncronously evicting state that * has a hold on this dnode while we are trying to evict this * dnode. */ while (refcount_count(&dn->dn_holds) > 0) delay(1); ASSERT(dn->dn_dbuf == NULL || dmu_buf_get_user(&dn->dn_dbuf->db) == NULL); zrl_add(&dnh->dnh_zrlock); dnode_destroy(dn); /* implicit zrl_remove() */ zrl_destroy(&dnh->dnh_zrlock); dnh->dnh_dnode = NULL; } void dnode_special_open(objset_t *os, dnode_phys_t *dnp, uint64_t object, dnode_handle_t *dnh) { dnode_t *dn; dn = dnode_create(os, dnp, NULL, object, dnh); zrl_init(&dnh->dnh_zrlock); DNODE_VERIFY(dn); } static void dnode_buf_evict_async(void *dbu) { dnode_children_t *children_dnodes = dbu; int i; for (i = 0; i < children_dnodes->dnc_count; i++) { dnode_handle_t *dnh = &children_dnodes->dnc_children[i]; dnode_t *dn; /* * The dnode handle lock guards against the dnode moving to * another valid address, so there is no need here to guard * against changes to or from NULL. */ if (dnh->dnh_dnode == NULL) { zrl_destroy(&dnh->dnh_zrlock); continue; } zrl_add(&dnh->dnh_zrlock); dn = dnh->dnh_dnode; /* * If there are holds on this dnode, then there should * be holds on the dnode's containing dbuf as well; thus * it wouldn't be eligible for eviction and this function * would not have been called. */ ASSERT(refcount_is_zero(&dn->dn_holds)); ASSERT(refcount_is_zero(&dn->dn_tx_holds)); dnode_destroy(dn); /* implicit zrl_remove() */ zrl_destroy(&dnh->dnh_zrlock); dnh->dnh_dnode = NULL; } kmem_free(children_dnodes, sizeof (dnode_children_t) + children_dnodes->dnc_count * sizeof (dnode_handle_t)); } /* * errors: * EINVAL - invalid object number. * EIO - i/o error. * succeeds even for free dnodes. */ int dnode_hold_impl(objset_t *os, uint64_t object, int flag, void *tag, dnode_t **dnp) { int epb, idx, err; int drop_struct_lock = FALSE; int type; uint64_t blk; dnode_t *mdn, *dn; dmu_buf_impl_t *db; dnode_children_t *children_dnodes; dnode_handle_t *dnh; /* * If you are holding the spa config lock as writer, you shouldn't * be asking the DMU to do *anything* unless it's the root pool * which may require us to read from the root filesystem while * holding some (not all) of the locks as writer. */ ASSERT(spa_config_held(os->os_spa, SCL_ALL, RW_WRITER) == 0 || (spa_is_root(os->os_spa) && spa_config_held(os->os_spa, SCL_STATE, RW_WRITER))); ASSERT((flag & DNODE_MUST_BE_ALLOCATED) || (flag & DNODE_MUST_BE_FREE)); if (object == DMU_USERUSED_OBJECT || object == DMU_GROUPUSED_OBJECT) { dn = (object == DMU_USERUSED_OBJECT) ? DMU_USERUSED_DNODE(os) : DMU_GROUPUSED_DNODE(os); if (dn == NULL) return (SET_ERROR(ENOENT)); type = dn->dn_type; if ((flag & DNODE_MUST_BE_ALLOCATED) && type == DMU_OT_NONE) return (SET_ERROR(ENOENT)); if ((flag & DNODE_MUST_BE_FREE) && type != DMU_OT_NONE) return (SET_ERROR(EEXIST)); DNODE_VERIFY(dn); (void) refcount_add(&dn->dn_holds, tag); *dnp = dn; return (0); } if (object == 0 || object >= DN_MAX_OBJECT) return (SET_ERROR(EINVAL)); mdn = DMU_META_DNODE(os); ASSERT(mdn->dn_object == DMU_META_DNODE_OBJECT); DNODE_VERIFY(mdn); if (!RW_WRITE_HELD(&mdn->dn_struct_rwlock)) { rw_enter(&mdn->dn_struct_rwlock, RW_READER); drop_struct_lock = TRUE; } blk = dbuf_whichblock(mdn, 0, object * sizeof (dnode_phys_t)); db = dbuf_hold(mdn, blk, FTAG); if (drop_struct_lock) rw_exit(&mdn->dn_struct_rwlock); if (db == NULL) return (SET_ERROR(EIO)); err = dbuf_read(db, NULL, DB_RF_CANFAIL); if (err) { dbuf_rele(db, FTAG); return (err); } ASSERT3U(db->db.db_size, >=, 1<db.db_size >> DNODE_SHIFT; idx = object & (epb-1); ASSERT(DB_DNODE(db)->dn_type == DMU_OT_DNODE); children_dnodes = dmu_buf_get_user(&db->db); if (children_dnodes == NULL) { int i; dnode_children_t *winner; children_dnodes = kmem_zalloc(sizeof (dnode_children_t) + epb * sizeof (dnode_handle_t), KM_SLEEP); children_dnodes->dnc_count = epb; dnh = &children_dnodes->dnc_children[0]; for (i = 0; i < epb; i++) { zrl_init(&dnh[i].dnh_zrlock); } dmu_buf_init_user(&children_dnodes->dnc_dbu, NULL, dnode_buf_evict_async, NULL); winner = dmu_buf_set_user(&db->db, &children_dnodes->dnc_dbu); if (winner != NULL) { for (i = 0; i < epb; i++) { zrl_destroy(&dnh[i].dnh_zrlock); } kmem_free(children_dnodes, sizeof (dnode_children_t) + epb * sizeof (dnode_handle_t)); children_dnodes = winner; } } ASSERT(children_dnodes->dnc_count == epb); dnh = &children_dnodes->dnc_children[idx]; zrl_add(&dnh->dnh_zrlock); dn = dnh->dnh_dnode; if (dn == NULL) { dnode_phys_t *phys = (dnode_phys_t *)db->db.db_data+idx; dn = dnode_create(os, phys, db, object, dnh); } mutex_enter(&dn->dn_mtx); type = dn->dn_type; if (dn->dn_free_txg || ((flag & DNODE_MUST_BE_ALLOCATED) && type == DMU_OT_NONE) || ((flag & DNODE_MUST_BE_FREE) && (type != DMU_OT_NONE || !refcount_is_zero(&dn->dn_holds)))) { mutex_exit(&dn->dn_mtx); zrl_remove(&dnh->dnh_zrlock); dbuf_rele(db, FTAG); return ((flag & DNODE_MUST_BE_ALLOCATED) ? ENOENT : EEXIST); } if (refcount_add(&dn->dn_holds, tag) == 1) dbuf_add_ref(db, dnh); mutex_exit(&dn->dn_mtx); /* Now we can rely on the hold to prevent the dnode from moving. */ zrl_remove(&dnh->dnh_zrlock); DNODE_VERIFY(dn); ASSERT3P(dn->dn_dbuf, ==, db); ASSERT3U(dn->dn_object, ==, object); dbuf_rele(db, FTAG); *dnp = dn; return (0); } /* * Return held dnode if the object is allocated, NULL if not. */ int dnode_hold(objset_t *os, uint64_t object, void *tag, dnode_t **dnp) { return (dnode_hold_impl(os, object, DNODE_MUST_BE_ALLOCATED, tag, dnp)); } /* * Can only add a reference if there is already at least one * reference on the dnode. Returns FALSE if unable to add a * new reference. */ boolean_t dnode_add_ref(dnode_t *dn, void *tag) { mutex_enter(&dn->dn_mtx); if (refcount_is_zero(&dn->dn_holds)) { mutex_exit(&dn->dn_mtx); return (FALSE); } VERIFY(1 < refcount_add(&dn->dn_holds, tag)); mutex_exit(&dn->dn_mtx); return (TRUE); } void dnode_rele(dnode_t *dn, void *tag) { mutex_enter(&dn->dn_mtx); dnode_rele_and_unlock(dn, tag); } void dnode_rele_and_unlock(dnode_t *dn, void *tag) { uint64_t refs; /* Get while the hold prevents the dnode from moving. */ dmu_buf_impl_t *db = dn->dn_dbuf; dnode_handle_t *dnh = dn->dn_handle; refs = refcount_remove(&dn->dn_holds, tag); mutex_exit(&dn->dn_mtx); /* * It's unsafe to release the last hold on a dnode by dnode_rele() or * indirectly by dbuf_rele() while relying on the dnode handle to * prevent the dnode from moving, since releasing the last hold could * result in the dnode's parent dbuf evicting its dnode handles. For * that reason anyone calling dnode_rele() or dbuf_rele() without some * other direct or indirect hold on the dnode must first drop the dnode * handle. */ ASSERT(refs > 0 || dnh->dnh_zrlock.zr_owner != curthread); /* NOTE: the DNODE_DNODE does not have a dn_dbuf */ if (refs == 0 && db != NULL) { /* * Another thread could add a hold to the dnode handle in * dnode_hold_impl() while holding the parent dbuf. Since the * hold on the parent dbuf prevents the handle from being * destroyed, the hold on the handle is OK. We can't yet assert * that the handle has zero references, but that will be * asserted anyway when the handle gets destroyed. */ dbuf_rele(db, dnh); } } void dnode_setdirty(dnode_t *dn, dmu_tx_t *tx) { objset_t *os = dn->dn_objset; uint64_t txg = tx->tx_txg; if (DMU_OBJECT_IS_SPECIAL(dn->dn_object)) { dsl_dataset_dirty(os->os_dsl_dataset, tx); return; } DNODE_VERIFY(dn); #ifdef ZFS_DEBUG mutex_enter(&dn->dn_mtx); ASSERT(dn->dn_phys->dn_type || dn->dn_allocated_txg); ASSERT(dn->dn_free_txg == 0 || dn->dn_free_txg >= txg); mutex_exit(&dn->dn_mtx); #endif /* * Determine old uid/gid when necessary */ dmu_objset_userquota_get_ids(dn, B_TRUE, tx); multilist_t *dirtylist = os->os_dirty_dnodes[txg & TXG_MASK]; multilist_sublist_t *mls = multilist_sublist_lock_obj(dirtylist, dn); /* * If we are already marked dirty, we're done. */ if (list_link_active(&dn->dn_dirty_link[txg & TXG_MASK])) { multilist_sublist_unlock(mls); return; } ASSERT(!refcount_is_zero(&dn->dn_holds) || !avl_is_empty(&dn->dn_dbufs)); ASSERT(dn->dn_datablksz != 0); ASSERT0(dn->dn_next_bonuslen[txg&TXG_MASK]); ASSERT0(dn->dn_next_blksz[txg&TXG_MASK]); ASSERT0(dn->dn_next_bonustype[txg&TXG_MASK]); dprintf_ds(os->os_dsl_dataset, "obj=%llu txg=%llu\n", dn->dn_object, txg); multilist_sublist_insert_head(mls, dn); multilist_sublist_unlock(mls); /* * The dnode maintains a hold on its containing dbuf as * long as there are holds on it. Each instantiated child * dbuf maintains a hold on the dnode. When the last child * drops its hold, the dnode will drop its hold on the * containing dbuf. We add a "dirty hold" here so that the * dnode will hang around after we finish processing its * children. */ VERIFY(dnode_add_ref(dn, (void *)(uintptr_t)tx->tx_txg)); (void) dbuf_dirty(dn->dn_dbuf, tx); dsl_dataset_dirty(os->os_dsl_dataset, tx); } void dnode_free(dnode_t *dn, dmu_tx_t *tx) { mutex_enter(&dn->dn_mtx); if (dn->dn_type == DMU_OT_NONE || dn->dn_free_txg) { mutex_exit(&dn->dn_mtx); return; } dn->dn_free_txg = tx->tx_txg; mutex_exit(&dn->dn_mtx); dnode_setdirty(dn, tx); } /* * Try to change the block size for the indicated dnode. This can only * succeed if there are no blocks allocated or dirty beyond first block */ int dnode_set_blksz(dnode_t *dn, uint64_t size, int ibs, dmu_tx_t *tx) { dmu_buf_impl_t *db; int err; ASSERT3U(size, <=, spa_maxblocksize(dmu_objset_spa(dn->dn_objset))); if (size == 0) size = SPA_MINBLOCKSIZE; else size = P2ROUNDUP(size, SPA_MINBLOCKSIZE); if (ibs == dn->dn_indblkshift) ibs = 0; if (size >> SPA_MINBLOCKSHIFT == dn->dn_datablkszsec && ibs == 0) return (0); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); /* Check for any allocated blocks beyond the first */ if (dn->dn_maxblkid != 0) goto fail; mutex_enter(&dn->dn_dbufs_mtx); for (db = avl_first(&dn->dn_dbufs); db != NULL; db = AVL_NEXT(&dn->dn_dbufs, db)) { if (db->db_blkid != 0 && db->db_blkid != DMU_BONUS_BLKID && db->db_blkid != DMU_SPILL_BLKID) { mutex_exit(&dn->dn_dbufs_mtx); goto fail; } } mutex_exit(&dn->dn_dbufs_mtx); if (ibs && dn->dn_nlevels != 1) goto fail; /* resize the old block */ err = dbuf_hold_impl(dn, 0, 0, TRUE, FALSE, FTAG, &db); if (err == 0) dbuf_new_size(db, size, tx); else if (err != ENOENT) goto fail; dnode_setdblksz(dn, size); dnode_setdirty(dn, tx); dn->dn_next_blksz[tx->tx_txg&TXG_MASK] = size; if (ibs) { dn->dn_indblkshift = ibs; dn->dn_next_indblkshift[tx->tx_txg&TXG_MASK] = ibs; } /* rele after we have fixed the blocksize in the dnode */ if (db) dbuf_rele(db, FTAG); rw_exit(&dn->dn_struct_rwlock); return (0); fail: rw_exit(&dn->dn_struct_rwlock); return (SET_ERROR(ENOTSUP)); } /* read-holding callers must not rely on the lock being continuously held */ void dnode_new_blkid(dnode_t *dn, uint64_t blkid, dmu_tx_t *tx, boolean_t have_read) { uint64_t txgoff = tx->tx_txg & TXG_MASK; int epbs, new_nlevels; uint64_t sz; ASSERT(blkid != DMU_BONUS_BLKID); ASSERT(have_read ? RW_READ_HELD(&dn->dn_struct_rwlock) : RW_WRITE_HELD(&dn->dn_struct_rwlock)); /* * if we have a read-lock, check to see if we need to do any work * before upgrading to a write-lock. */ if (have_read) { if (blkid <= dn->dn_maxblkid) return; if (!rw_tryupgrade(&dn->dn_struct_rwlock)) { rw_exit(&dn->dn_struct_rwlock); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); } } if (blkid <= dn->dn_maxblkid) goto out; dn->dn_maxblkid = blkid; /* * Compute the number of levels necessary to support the new maxblkid. */ new_nlevels = 1; epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; for (sz = dn->dn_nblkptr; sz <= blkid && sz >= dn->dn_nblkptr; sz <<= epbs) new_nlevels++; if (new_nlevels > dn->dn_nlevels) { int old_nlevels = dn->dn_nlevels; dmu_buf_impl_t *db; list_t *list; dbuf_dirty_record_t *new, *dr, *dr_next; dn->dn_nlevels = new_nlevels; ASSERT3U(new_nlevels, >, dn->dn_next_nlevels[txgoff]); dn->dn_next_nlevels[txgoff] = new_nlevels; /* dirty the left indirects */ db = dbuf_hold_level(dn, old_nlevels, 0, FTAG); ASSERT(db != NULL); new = dbuf_dirty(db, tx); dbuf_rele(db, FTAG); /* transfer the dirty records to the new indirect */ mutex_enter(&dn->dn_mtx); mutex_enter(&new->dt.di.dr_mtx); list = &dn->dn_dirty_records[txgoff]; for (dr = list_head(list); dr; dr = dr_next) { dr_next = list_next(&dn->dn_dirty_records[txgoff], dr); if (dr->dr_dbuf->db_level != new_nlevels-1 && dr->dr_dbuf->db_blkid != DMU_BONUS_BLKID && dr->dr_dbuf->db_blkid != DMU_SPILL_BLKID) { ASSERT(dr->dr_dbuf->db_level == old_nlevels-1); list_remove(&dn->dn_dirty_records[txgoff], dr); list_insert_tail(&new->dt.di.dr_children, dr); dr->dr_parent = new; } } mutex_exit(&new->dt.di.dr_mtx); mutex_exit(&dn->dn_mtx); } out: if (have_read) rw_downgrade(&dn->dn_struct_rwlock); } static void dnode_dirty_l1(dnode_t *dn, uint64_t l1blkid, dmu_tx_t *tx) { dmu_buf_impl_t *db = dbuf_hold_level(dn, 1, l1blkid, FTAG); if (db != NULL) { dmu_buf_will_dirty(&db->db, tx); dbuf_rele(db, FTAG); } } +/* + * Dirty all the in-core level-1 dbufs in the range specified by start_blkid + * and end_blkid. + */ +static void +dnode_dirty_l1range(dnode_t *dn, uint64_t start_blkid, uint64_t end_blkid, + dmu_tx_t *tx) +{ + dmu_buf_impl_t db_search; + dmu_buf_impl_t *db; + avl_index_t where; + + mutex_enter(&dn->dn_dbufs_mtx); + + db_search.db_level = 1; + db_search.db_blkid = start_blkid + 1; + db_search.db_state = DB_SEARCH; + for (;;) { + + db = avl_find(&dn->dn_dbufs, &db_search, &where); + if (db == NULL) + db = avl_nearest(&dn->dn_dbufs, where, AVL_AFTER); + + if (db == NULL || db->db_level != 1 || + db->db_blkid >= end_blkid) { + break; + } + + /* + * Setup the next blkid we want to search for. + */ + db_search.db_blkid = db->db_blkid + 1; + ASSERT3U(db->db_blkid, >=, start_blkid); + + /* + * If the dbuf transitions to DB_EVICTING while we're trying + * to dirty it, then we will be unable to discover it in + * the dbuf hash table. This will result in a call to + * dbuf_create() which needs to acquire the dn_dbufs_mtx + * lock. To avoid a deadlock, we drop the lock before + * dirtying the level-1 dbuf. + */ + mutex_exit(&dn->dn_dbufs_mtx); + dnode_dirty_l1(dn, db->db_blkid, tx); + mutex_enter(&dn->dn_dbufs_mtx); + } + +#ifdef ZFS_DEBUG + /* + * Walk all the in-core level-1 dbufs and verify they have been dirtied. + */ + db_search.db_level = 1; + db_search.db_blkid = start_blkid + 1; + db_search.db_state = DB_SEARCH; + db = avl_find(&dn->dn_dbufs, &db_search, &where); + if (db == NULL) + db = avl_nearest(&dn->dn_dbufs, where, AVL_AFTER); + for (; db != NULL; db = AVL_NEXT(&dn->dn_dbufs, db)) { + if (db->db_level != 1 || db->db_blkid >= end_blkid) + break; + ASSERT(db->db_dirtycnt > 0); + } +#endif + mutex_exit(&dn->dn_dbufs_mtx); +} + void dnode_free_range(dnode_t *dn, uint64_t off, uint64_t len, dmu_tx_t *tx) { dmu_buf_impl_t *db; uint64_t blkoff, blkid, nblks; int blksz, blkshift, head, tail; int trunc = FALSE; int epbs; rw_enter(&dn->dn_struct_rwlock, RW_WRITER); blksz = dn->dn_datablksz; blkshift = dn->dn_datablkshift; epbs = dn->dn_indblkshift - SPA_BLKPTRSHIFT; if (len == DMU_OBJECT_END) { len = UINT64_MAX - off; trunc = TRUE; } /* * First, block align the region to free: */ if (ISP2(blksz)) { head = P2NPHASE(off, blksz); blkoff = P2PHASE(off, blksz); if ((off >> blkshift) > dn->dn_maxblkid) goto out; } else { ASSERT(dn->dn_maxblkid == 0); if (off == 0 && len >= blksz) { /* * Freeing the whole block; fast-track this request. * Note that we won't dirty any indirect blocks, * which is fine because we will be freeing the entire * file and thus all indirect blocks will be freed * by free_children(). */ blkid = 0; nblks = 1; goto done; } else if (off >= blksz) { /* Freeing past end-of-data */ goto out; } else { /* Freeing part of the block. */ head = blksz - off; ASSERT3U(head, >, 0); } blkoff = off; } /* zero out any partial block data at the start of the range */ if (head) { ASSERT3U(blkoff + head, ==, blksz); if (len < head) head = len; if (dbuf_hold_impl(dn, 0, dbuf_whichblock(dn, 0, off), TRUE, FALSE, FTAG, &db) == 0) { caddr_t data; /* don't dirty if it isn't on disk and isn't dirty */ if (db->db_last_dirty || (db->db_blkptr && !BP_IS_HOLE(db->db_blkptr))) { rw_exit(&dn->dn_struct_rwlock); dmu_buf_will_dirty(&db->db, tx); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); data = db->db.db_data; bzero(data + blkoff, head); } dbuf_rele(db, FTAG); } off += head; len -= head; } /* If the range was less than one block, we're done */ if (len == 0) goto out; /* If the remaining range is past end of file, we're done */ if ((off >> blkshift) > dn->dn_maxblkid) goto out; ASSERT(ISP2(blksz)); if (trunc) tail = 0; else tail = P2PHASE(len, blksz); ASSERT0(P2PHASE(off, blksz)); /* zero out any partial block data at the end of the range */ if (tail) { if (len < tail) tail = len; if (dbuf_hold_impl(dn, 0, dbuf_whichblock(dn, 0, off+len), TRUE, FALSE, FTAG, &db) == 0) { /* don't dirty if not on disk and not dirty */ if (db->db_last_dirty || (db->db_blkptr && !BP_IS_HOLE(db->db_blkptr))) { rw_exit(&dn->dn_struct_rwlock); dmu_buf_will_dirty(&db->db, tx); rw_enter(&dn->dn_struct_rwlock, RW_WRITER); bzero(db->db.db_data, tail); } dbuf_rele(db, FTAG); } len -= tail; } /* If the range did not include a full block, we are done */ if (len == 0) goto out; ASSERT(IS_P2ALIGNED(off, blksz)); ASSERT(trunc || IS_P2ALIGNED(len, blksz)); blkid = off >> blkshift; nblks = len >> blkshift; if (trunc) nblks += 1; /* * Dirty all the indirect blocks in this range. Note that only * the first and last indirect blocks can actually be written * (if they were partially freed) -- they must be dirtied, even if * they do not exist on disk yet. The interior blocks will * be freed by free_children(), so they will not actually be written. * Even though these interior blocks will not be written, we * dirty them for two reasons: * * - It ensures that the indirect blocks remain in memory until * syncing context. (They have already been prefetched by * dmu_tx_hold_free(), so we don't have to worry about reading * them serially here.) * * - The dirty space accounting will put pressure on the txg sync * mechanism to begin syncing, and to delay transactions if there * is a large amount of freeing. Even though these indirect * blocks will not be written, we could need to write the same * amount of space if we copy the freed BPs into deadlists. */ if (dn->dn_nlevels > 1) { uint64_t first, last; first = blkid >> epbs; dnode_dirty_l1(dn, first, tx); if (trunc) last = dn->dn_maxblkid >> epbs; else last = (blkid + nblks - 1) >> epbs; if (last != first) dnode_dirty_l1(dn, last, tx); + + dnode_dirty_l1range(dn, first, last, tx); int shift = dn->dn_datablkshift + dn->dn_indblkshift - SPA_BLKPTRSHIFT; for (uint64_t i = first + 1; i < last; i++) { /* * Set i to the blockid of the next non-hole * level-1 indirect block at or after i. Note * that dnode_next_offset() operates in terms of * level-0-equivalent bytes. */ uint64_t ibyte = i << shift; int err = dnode_next_offset(dn, DNODE_FIND_HAVELOCK, &ibyte, 2, 1, 0); i = ibyte >> shift; if (i >= last) break; /* * Normally we should not see an error, either * from dnode_next_offset() or dbuf_hold_level() * (except for ESRCH from dnode_next_offset). * If there is an i/o error, then when we read * this block in syncing context, it will use * ZIO_FLAG_MUSTSUCCEED, and thus hang/panic according * to the "failmode" property. dnode_next_offset() * doesn't have a flag to indicate MUSTSUCCEED. */ if (err != 0) break; dnode_dirty_l1(dn, i, tx); } } done: /* * Add this range to the dnode range list. * We will finish up this free operation in the syncing phase. */ mutex_enter(&dn->dn_mtx); int txgoff = tx->tx_txg & TXG_MASK; if (dn->dn_free_ranges[txgoff] == NULL) { dn->dn_free_ranges[txgoff] = range_tree_create(NULL, NULL); } range_tree_clear(dn->dn_free_ranges[txgoff], blkid, nblks); range_tree_add(dn->dn_free_ranges[txgoff], blkid, nblks); dprintf_dnode(dn, "blkid=%llu nblks=%llu txg=%llu\n", blkid, nblks, tx->tx_txg); mutex_exit(&dn->dn_mtx); dbuf_free_range(dn, blkid, blkid + nblks - 1, tx); dnode_setdirty(dn, tx); out: rw_exit(&dn->dn_struct_rwlock); } static boolean_t dnode_spill_freed(dnode_t *dn) { int i; mutex_enter(&dn->dn_mtx); for (i = 0; i < TXG_SIZE; i++) { if (dn->dn_rm_spillblk[i] == DN_KILL_SPILLBLK) break; } mutex_exit(&dn->dn_mtx); return (i < TXG_SIZE); } /* return TRUE if this blkid was freed in a recent txg, or FALSE if it wasn't */ uint64_t dnode_block_freed(dnode_t *dn, uint64_t blkid) { void *dp = spa_get_dsl(dn->dn_objset->os_spa); int i; if (blkid == DMU_BONUS_BLKID) return (FALSE); /* * If we're in the process of opening the pool, dp will not be * set yet, but there shouldn't be anything dirty. */ if (dp == NULL) return (FALSE); if (dn->dn_free_txg) return (TRUE); if (blkid == DMU_SPILL_BLKID) return (dnode_spill_freed(dn)); mutex_enter(&dn->dn_mtx); for (i = 0; i < TXG_SIZE; i++) { if (dn->dn_free_ranges[i] != NULL && range_tree_contains(dn->dn_free_ranges[i], blkid, 1)) break; } mutex_exit(&dn->dn_mtx); return (i < TXG_SIZE); } /* call from syncing context when we actually write/free space for this dnode */ void dnode_diduse_space(dnode_t *dn, int64_t delta) { uint64_t space; dprintf_dnode(dn, "dn=%p dnp=%p used=%llu delta=%lld\n", dn, dn->dn_phys, (u_longlong_t)dn->dn_phys->dn_used, (longlong_t)delta); mutex_enter(&dn->dn_mtx); space = DN_USED_BYTES(dn->dn_phys); if (delta > 0) { ASSERT3U(space + delta, >=, space); /* no overflow */ } else { ASSERT3U(space, >=, -delta); /* no underflow */ } space += delta; if (spa_version(dn->dn_objset->os_spa) < SPA_VERSION_DNODE_BYTES) { ASSERT((dn->dn_phys->dn_flags & DNODE_FLAG_USED_BYTES) == 0); ASSERT0(P2PHASE(space, 1<dn_phys->dn_used = space >> DEV_BSHIFT; } else { dn->dn_phys->dn_used = space; dn->dn_phys->dn_flags |= DNODE_FLAG_USED_BYTES; } mutex_exit(&dn->dn_mtx); } /* * Scans a block at the indicated "level" looking for a hole or data, * depending on 'flags'. * * If level > 0, then we are scanning an indirect block looking at its * pointers. If level == 0, then we are looking at a block of dnodes. * * If we don't find what we are looking for in the block, we return ESRCH. * Otherwise, return with *offset pointing to the beginning (if searching * forwards) or end (if searching backwards) of the range covered by the * block pointer we matched on (or dnode). * * The basic search algorithm used below by dnode_next_offset() is to * use this function to search up the block tree (widen the search) until * we find something (i.e., we don't return ESRCH) and then search back * down the tree (narrow the search) until we reach our original search * level. */ static int dnode_next_offset_level(dnode_t *dn, int flags, uint64_t *offset, int lvl, uint64_t blkfill, uint64_t txg) { dmu_buf_impl_t *db = NULL; void *data = NULL; uint64_t epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; uint64_t epb = 1ULL << epbs; uint64_t minfill, maxfill; boolean_t hole; int i, inc, error, span; dprintf("probing object %llu offset %llx level %d of %u\n", dn->dn_object, *offset, lvl, dn->dn_phys->dn_nlevels); hole = ((flags & DNODE_FIND_HOLE) != 0); inc = (flags & DNODE_FIND_BACKWARDS) ? -1 : 1; ASSERT(txg == 0 || !hole); if (lvl == dn->dn_phys->dn_nlevels) { error = 0; epb = dn->dn_phys->dn_nblkptr; data = dn->dn_phys->dn_blkptr; } else { uint64_t blkid = dbuf_whichblock(dn, lvl, *offset); error = dbuf_hold_impl(dn, lvl, blkid, TRUE, FALSE, FTAG, &db); if (error) { if (error != ENOENT) return (error); if (hole) return (0); /* * This can only happen when we are searching up * the block tree for data. We don't really need to * adjust the offset, as we will just end up looking * at the pointer to this block in its parent, and its * going to be unallocated, so we will skip over it. */ return (SET_ERROR(ESRCH)); } error = dbuf_read(db, NULL, DB_RF_CANFAIL | DB_RF_HAVESTRUCT); if (error) { dbuf_rele(db, FTAG); return (error); } data = db->db.db_data; } if (db != NULL && txg != 0 && (db->db_blkptr == NULL || db->db_blkptr->blk_birth <= txg || BP_IS_HOLE(db->db_blkptr))) { /* * This can only happen when we are searching up the tree * and these conditions mean that we need to keep climbing. */ error = SET_ERROR(ESRCH); } else if (lvl == 0) { dnode_phys_t *dnp = data; span = DNODE_SHIFT; ASSERT(dn->dn_type == DMU_OT_DNODE); for (i = (*offset >> span) & (blkfill - 1); i >= 0 && i < blkfill; i += inc) { if ((dnp[i].dn_type == DMU_OT_NONE) == hole) break; *offset += (1ULL << span) * inc; } if (i < 0 || i == blkfill) error = SET_ERROR(ESRCH); } else { blkptr_t *bp = data; uint64_t start = *offset; span = (lvl - 1) * epbs + dn->dn_datablkshift; minfill = 0; maxfill = blkfill << ((lvl - 1) * epbs); if (hole) maxfill--; else minfill++; *offset = *offset >> span; for (i = BF64_GET(*offset, 0, epbs); i >= 0 && i < epb; i += inc) { if (BP_GET_FILL(&bp[i]) >= minfill && BP_GET_FILL(&bp[i]) <= maxfill && (hole || bp[i].blk_birth > txg)) break; if (inc > 0 || *offset > 0) *offset += inc; } *offset = *offset << span; if (inc < 0) { /* traversing backwards; position offset at the end */ ASSERT3U(*offset, <=, start); *offset = MIN(*offset + (1ULL << span) - 1, start); } else if (*offset < start) { *offset = start; } if (i < 0 || i >= epb) error = SET_ERROR(ESRCH); } if (db) dbuf_rele(db, FTAG); return (error); } /* * Find the next hole, data, or sparse region at or after *offset. * The value 'blkfill' tells us how many items we expect to find * in an L0 data block; this value is 1 for normal objects, * DNODES_PER_BLOCK for the meta dnode, and some fraction of * DNODES_PER_BLOCK when searching for sparse regions thereof. * * Examples: * * dnode_next_offset(dn, flags, offset, 1, 1, 0); * Finds the next/previous hole/data in a file. * Used in dmu_offset_next(). * * dnode_next_offset(mdn, flags, offset, 0, DNODES_PER_BLOCK, txg); * Finds the next free/allocated dnode an objset's meta-dnode. * Only finds objects that have new contents since txg (ie. * bonus buffer changes and content removal are ignored). * Used in dmu_object_next(). * * dnode_next_offset(mdn, DNODE_FIND_HOLE, offset, 2, DNODES_PER_BLOCK >> 2, 0); * Finds the next L2 meta-dnode bp that's at most 1/4 full. * Used in dmu_object_alloc(). */ int dnode_next_offset(dnode_t *dn, int flags, uint64_t *offset, int minlvl, uint64_t blkfill, uint64_t txg) { uint64_t initial_offset = *offset; int lvl, maxlvl; int error = 0; if (!(flags & DNODE_FIND_HAVELOCK)) rw_enter(&dn->dn_struct_rwlock, RW_READER); if (dn->dn_phys->dn_nlevels == 0) { error = SET_ERROR(ESRCH); goto out; } if (dn->dn_datablkshift == 0) { if (*offset < dn->dn_datablksz) { if (flags & DNODE_FIND_HOLE) *offset = dn->dn_datablksz; } else { error = SET_ERROR(ESRCH); } goto out; } maxlvl = dn->dn_phys->dn_nlevels; for (lvl = minlvl; lvl <= maxlvl; lvl++) { error = dnode_next_offset_level(dn, flags, offset, lvl, blkfill, txg); if (error != ESRCH) break; } while (error == 0 && --lvl >= minlvl) { error = dnode_next_offset_level(dn, flags, offset, lvl, blkfill, txg); } /* * There's always a "virtual hole" at the end of the object, even * if all BP's which physically exist are non-holes. */ if ((flags & DNODE_FIND_HOLE) && error == ESRCH && txg == 0 && minlvl == 1 && blkfill == 1 && !(flags & DNODE_FIND_BACKWARDS)) { error = 0; } if (error == 0 && (flags & DNODE_FIND_BACKWARDS ? initial_offset < *offset : initial_offset > *offset)) error = SET_ERROR(ESRCH); out: if (!(flags & DNODE_FIND_HAVELOCK)) rw_exit(&dn->dn_struct_rwlock); return (error); } Index: stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode_sync.c =================================================================== --- stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode_sync.c (revision 339135) +++ stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/dnode_sync.c (revision 339136) @@ -1,728 +1,734 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2017 by Delphix. All rights reserved. * Copyright (c) 2014 Spectra Logic Corporation, All rights reserved. */ #include #include #include #include #include #include #include #include #include #include static void dnode_increase_indirection(dnode_t *dn, dmu_tx_t *tx) { dmu_buf_impl_t *db; int txgoff = tx->tx_txg & TXG_MASK; int nblkptr = dn->dn_phys->dn_nblkptr; int old_toplvl = dn->dn_phys->dn_nlevels - 1; int new_level = dn->dn_next_nlevels[txgoff]; int i; rw_enter(&dn->dn_struct_rwlock, RW_WRITER); /* this dnode can't be paged out because it's dirty */ ASSERT(dn->dn_phys->dn_type != DMU_OT_NONE); ASSERT(RW_WRITE_HELD(&dn->dn_struct_rwlock)); ASSERT(new_level > 1 && dn->dn_phys->dn_nlevels > 0); db = dbuf_hold_level(dn, dn->dn_phys->dn_nlevels, 0, FTAG); ASSERT(db != NULL); dn->dn_phys->dn_nlevels = new_level; dprintf("os=%p obj=%llu, increase to %d\n", dn->dn_objset, dn->dn_object, dn->dn_phys->dn_nlevels); /* transfer dnode's block pointers to new indirect block */ (void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED|DB_RF_HAVESTRUCT); ASSERT(db->db.db_data); ASSERT(arc_released(db->db_buf)); ASSERT3U(sizeof (blkptr_t) * nblkptr, <=, db->db.db_size); bcopy(dn->dn_phys->dn_blkptr, db->db.db_data, sizeof (blkptr_t) * nblkptr); arc_buf_freeze(db->db_buf); /* set dbuf's parent pointers to new indirect buf */ for (i = 0; i < nblkptr; i++) { dmu_buf_impl_t *child = dbuf_find(dn->dn_objset, dn->dn_object, old_toplvl, i); if (child == NULL) continue; #ifdef DEBUG DB_DNODE_ENTER(child); ASSERT3P(DB_DNODE(child), ==, dn); DB_DNODE_EXIT(child); #endif /* DEBUG */ if (child->db_parent && child->db_parent != dn->dn_dbuf) { ASSERT(child->db_parent->db_level == db->db_level); ASSERT(child->db_blkptr != &dn->dn_phys->dn_blkptr[child->db_blkid]); mutex_exit(&child->db_mtx); continue; } ASSERT(child->db_parent == NULL || child->db_parent == dn->dn_dbuf); child->db_parent = db; dbuf_add_ref(db, child); if (db->db.db_data) child->db_blkptr = (blkptr_t *)db->db.db_data + i; else child->db_blkptr = NULL; dprintf_dbuf_bp(child, child->db_blkptr, "changed db_blkptr to new indirect %s", ""); mutex_exit(&child->db_mtx); } bzero(dn->dn_phys->dn_blkptr, sizeof (blkptr_t) * nblkptr); dbuf_rele(db, FTAG); rw_exit(&dn->dn_struct_rwlock); } static void free_blocks(dnode_t *dn, blkptr_t *bp, int num, dmu_tx_t *tx) { dsl_dataset_t *ds = dn->dn_objset->os_dsl_dataset; uint64_t bytesfreed = 0; dprintf("ds=%p obj=%llx num=%d\n", ds, dn->dn_object, num); for (int i = 0; i < num; i++, bp++) { if (BP_IS_HOLE(bp)) continue; bytesfreed += dsl_dataset_block_kill(ds, bp, tx, B_FALSE); ASSERT3U(bytesfreed, <=, DN_USED_BYTES(dn->dn_phys)); /* * Save some useful information on the holes being * punched, including logical size, type, and indirection * level. Retaining birth time enables detection of when * holes are punched for reducing the number of free * records transmitted during a zfs send. */ uint64_t lsize = BP_GET_LSIZE(bp); dmu_object_type_t type = BP_GET_TYPE(bp); uint64_t lvl = BP_GET_LEVEL(bp); bzero(bp, sizeof (blkptr_t)); if (spa_feature_is_active(dn->dn_objset->os_spa, SPA_FEATURE_HOLE_BIRTH)) { BP_SET_LSIZE(bp, lsize); BP_SET_TYPE(bp, type); BP_SET_LEVEL(bp, lvl); BP_SET_BIRTH(bp, dmu_tx_get_txg(tx), 0); } } dnode_diduse_space(dn, -bytesfreed); } #ifdef ZFS_DEBUG static void free_verify(dmu_buf_impl_t *db, uint64_t start, uint64_t end, dmu_tx_t *tx) { int off, num; int i, err, epbs; uint64_t txg = tx->tx_txg; dnode_t *dn; DB_DNODE_ENTER(db); dn = DB_DNODE(db); epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; off = start - (db->db_blkid * 1<=, 0); ASSERT3U(num, >=, 0); ASSERT3U(db->db_level, >, 0); ASSERT3U(db->db.db_size, ==, 1 << dn->dn_phys->dn_indblkshift); ASSERT3U(off+num, <=, db->db.db_size >> SPA_BLKPTRSHIFT); ASSERT(db->db_blkptr != NULL); for (i = off; i < off+num; i++) { uint64_t *buf; dmu_buf_impl_t *child; dbuf_dirty_record_t *dr; int j; ASSERT(db->db_level == 1); rw_enter(&dn->dn_struct_rwlock, RW_READER); err = dbuf_hold_impl(dn, db->db_level-1, (db->db_blkid << epbs) + i, TRUE, FALSE, FTAG, &child); rw_exit(&dn->dn_struct_rwlock); if (err == ENOENT) continue; ASSERT(err == 0); ASSERT(child->db_level == 0); dr = child->db_last_dirty; while (dr && dr->dr_txg > txg) dr = dr->dr_next; ASSERT(dr == NULL || dr->dr_txg == txg); /* data_old better be zeroed */ if (dr) { buf = dr->dt.dl.dr_data->b_data; for (j = 0; j < child->db.db_size >> 3; j++) { if (buf[j] != 0) { panic("freed data not zero: " "child=%p i=%d off=%d num=%d\n", (void *)child, i, off, num); } } } /* * db_data better be zeroed unless it's dirty in a * future txg. */ mutex_enter(&child->db_mtx); buf = child->db.db_data; if (buf != NULL && child->db_state != DB_FILL && child->db_last_dirty == NULL) { for (j = 0; j < child->db.db_size >> 3; j++) { if (buf[j] != 0) { panic("freed data not zero: " "child=%p i=%d off=%d num=%d\n", (void *)child, i, off, num); } } } mutex_exit(&child->db_mtx); dbuf_rele(child, FTAG); } DB_DNODE_EXIT(db); } #endif +/* + * We don't usually free the indirect blocks here. If in one txg we have a + * free_range and a write to the same indirect block, it's important that we + * preserve the hole's birth times. Therefore, we don't free any any indirect + * blocks in free_children(). If an indirect block happens to turn into all + * holes, it will be freed by dbuf_write_children_ready, which happens at a + * point in the syncing process where we know for certain the contents of the + * indirect block. + * + * However, if we're freeing a dnode, its space accounting must go to zero + * before we actually try to free the dnode, or we will trip an assertion. In + * addition, we know the case described above cannot occur, because the dnode is + * being freed. Therefore, we free the indirect blocks immediately in that + * case. + */ static void free_children(dmu_buf_impl_t *db, uint64_t blkid, uint64_t nblks, - dmu_tx_t *tx) + boolean_t free_indirects, dmu_tx_t *tx) { dnode_t *dn; blkptr_t *bp; dmu_buf_impl_t *subdb; uint64_t start, end, dbstart, dbend; unsigned int epbs, shift, i; /* * There is a small possibility that this block will not be cached: * 1 - if level > 1 and there are no children with level <= 1 * 2 - if this block was evicted since we read it from * dmu_tx_hold_free(). */ if (db->db_state != DB_CACHED) (void) dbuf_read(db, NULL, DB_RF_MUST_SUCCEED); dbuf_release_bp(db); bp = db->db.db_data; DB_DNODE_ENTER(db); dn = DB_DNODE(db); epbs = dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT; ASSERT3U(epbs, <, 31); shift = (db->db_level - 1) * epbs; dbstart = db->db_blkid << epbs; start = blkid >> shift; if (dbstart < start) { bp += start - dbstart; } else { start = dbstart; } dbend = ((db->db_blkid + 1) << epbs) - 1; end = (blkid + nblks - 1) >> shift; if (dbend <= end) end = dbend; ASSERT3U(start, <=, end); if (db->db_level == 1) { FREE_VERIFY(db, start, end, tx); free_blocks(dn, bp, end-start+1, tx); } else { for (uint64_t id = start; id <= end; id++, bp++) { if (BP_IS_HOLE(bp)) continue; rw_enter(&dn->dn_struct_rwlock, RW_READER); VERIFY0(dbuf_hold_impl(dn, db->db_level - 1, id, TRUE, FALSE, FTAG, &subdb)); rw_exit(&dn->dn_struct_rwlock); ASSERT3P(bp, ==, subdb->db_blkptr); - free_children(subdb, blkid, nblks, tx); + free_children(subdb, blkid, nblks, free_indirects, tx); dbuf_rele(subdb, FTAG); } } - /* If this whole block is free, free ourself too. */ - for (i = 0, bp = db->db.db_data; i < 1 << epbs; i++, bp++) { - if (!BP_IS_HOLE(bp)) - break; - } - if (i == 1 << epbs) { - /* - * We only found holes. Grab the rwlock to prevent - * anybody from reading the blocks we're about to - * zero out. - */ - rw_enter(&dn->dn_struct_rwlock, RW_WRITER); + if (free_indirects) { + for (i = 0, bp = db->db.db_data; i < 1 << epbs; i++, bp++) + ASSERT(BP_IS_HOLE(bp)); bzero(db->db.db_data, db->db.db_size); - rw_exit(&dn->dn_struct_rwlock); free_blocks(dn, db->db_blkptr, 1, tx); - } else { - /* - * Partial block free; must be marked dirty so that it - * will be written out. - */ - ASSERT(db->db_dirtycnt > 0); } DB_DNODE_EXIT(db); arc_buf_freeze(db->db_buf); } /* * Traverse the indicated range of the provided file * and "free" all the blocks contained there. */ static void dnode_sync_free_range_impl(dnode_t *dn, uint64_t blkid, uint64_t nblks, - dmu_tx_t *tx) + boolean_t free_indirects, dmu_tx_t *tx) { blkptr_t *bp = dn->dn_phys->dn_blkptr; int dnlevel = dn->dn_phys->dn_nlevels; boolean_t trunc = B_FALSE; if (blkid > dn->dn_phys->dn_maxblkid) return; ASSERT(dn->dn_phys->dn_maxblkid < UINT64_MAX); if (blkid + nblks > dn->dn_phys->dn_maxblkid) { nblks = dn->dn_phys->dn_maxblkid - blkid + 1; trunc = B_TRUE; } /* There are no indirect blocks in the object */ if (dnlevel == 1) { if (blkid >= dn->dn_phys->dn_nblkptr) { /* this range was never made persistent */ return; } ASSERT3U(blkid + nblks, <=, dn->dn_phys->dn_nblkptr); free_blocks(dn, bp + blkid, nblks, tx); } else { int shift = (dnlevel - 1) * (dn->dn_phys->dn_indblkshift - SPA_BLKPTRSHIFT); int start = blkid >> shift; int end = (blkid + nblks - 1) >> shift; dmu_buf_impl_t *db; ASSERT(start < dn->dn_phys->dn_nblkptr); bp += start; for (int i = start; i <= end; i++, bp++) { if (BP_IS_HOLE(bp)) continue; rw_enter(&dn->dn_struct_rwlock, RW_READER); VERIFY0(dbuf_hold_impl(dn, dnlevel - 1, i, TRUE, FALSE, FTAG, &db)); rw_exit(&dn->dn_struct_rwlock); - free_children(db, blkid, nblks, tx); + free_children(db, blkid, nblks, free_indirects, tx); dbuf_rele(db, FTAG); } } if (trunc) { dn->dn_phys->dn_maxblkid = blkid == 0 ? 0 : blkid - 1; uint64_t off = (dn->dn_phys->dn_maxblkid + 1) * (dn->dn_phys->dn_datablkszsec << SPA_MINBLOCKSHIFT); ASSERT(off < dn->dn_phys->dn_maxblkid || dn->dn_phys->dn_maxblkid == 0 || dnode_next_offset(dn, 0, &off, 1, 1, 0) != 0); } } typedef struct dnode_sync_free_range_arg { dnode_t *dsfra_dnode; dmu_tx_t *dsfra_tx; + boolean_t dsfra_free_indirects; } dnode_sync_free_range_arg_t; static void dnode_sync_free_range(void *arg, uint64_t blkid, uint64_t nblks) { dnode_sync_free_range_arg_t *dsfra = arg; dnode_t *dn = dsfra->dsfra_dnode; mutex_exit(&dn->dn_mtx); - dnode_sync_free_range_impl(dn, blkid, nblks, dsfra->dsfra_tx); + dnode_sync_free_range_impl(dn, blkid, nblks, + dsfra->dsfra_free_indirects, dsfra->dsfra_tx); mutex_enter(&dn->dn_mtx); } /* * Try to kick all the dnode's dbufs out of the cache... */ void dnode_evict_dbufs(dnode_t *dn) { dmu_buf_impl_t db_marker; dmu_buf_impl_t *db, *db_next; mutex_enter(&dn->dn_dbufs_mtx); for (db = avl_first(&dn->dn_dbufs); db != NULL; db = db_next) { #ifdef DEBUG DB_DNODE_ENTER(db); ASSERT3P(DB_DNODE(db), ==, dn); DB_DNODE_EXIT(db); #endif /* DEBUG */ mutex_enter(&db->db_mtx); if (db->db_state != DB_EVICTING && refcount_is_zero(&db->db_holds)) { db_marker.db_level = db->db_level; db_marker.db_blkid = db->db_blkid; db_marker.db_state = DB_SEARCH; avl_insert_here(&dn->dn_dbufs, &db_marker, db, AVL_BEFORE); dbuf_destroy(db); db_next = AVL_NEXT(&dn->dn_dbufs, &db_marker); avl_remove(&dn->dn_dbufs, &db_marker); } else { db->db_pending_evict = TRUE; mutex_exit(&db->db_mtx); db_next = AVL_NEXT(&dn->dn_dbufs, db); } } mutex_exit(&dn->dn_dbufs_mtx); dnode_evict_bonus(dn); } void dnode_evict_bonus(dnode_t *dn) { rw_enter(&dn->dn_struct_rwlock, RW_WRITER); if (dn->dn_bonus != NULL) { if (refcount_is_zero(&dn->dn_bonus->db_holds)) { mutex_enter(&dn->dn_bonus->db_mtx); dbuf_destroy(dn->dn_bonus); dn->dn_bonus = NULL; } else { dn->dn_bonus->db_pending_evict = TRUE; } } rw_exit(&dn->dn_struct_rwlock); } static void dnode_undirty_dbufs(list_t *list) { dbuf_dirty_record_t *dr; while (dr = list_head(list)) { dmu_buf_impl_t *db = dr->dr_dbuf; uint64_t txg = dr->dr_txg; if (db->db_level != 0) dnode_undirty_dbufs(&dr->dt.di.dr_children); mutex_enter(&db->db_mtx); /* XXX - use dbuf_undirty()? */ list_remove(list, dr); ASSERT(db->db_last_dirty == dr); db->db_last_dirty = NULL; db->db_dirtycnt -= 1; if (db->db_level == 0) { ASSERT(db->db_blkid == DMU_BONUS_BLKID || dr->dt.dl.dr_data == db->db_buf); dbuf_unoverride(dr); } else { mutex_destroy(&dr->dt.di.dr_mtx); list_destroy(&dr->dt.di.dr_children); } kmem_free(dr, sizeof (dbuf_dirty_record_t)); dbuf_rele_and_unlock(db, (void *)(uintptr_t)txg); } } static void dnode_sync_free(dnode_t *dn, dmu_tx_t *tx) { int txgoff = tx->tx_txg & TXG_MASK; ASSERT(dmu_tx_is_syncing(tx)); /* * Our contents should have been freed in dnode_sync() by the * free range record inserted by the caller of dnode_free(). */ ASSERT0(DN_USED_BYTES(dn->dn_phys)); ASSERT(BP_IS_HOLE(dn->dn_phys->dn_blkptr)); dnode_undirty_dbufs(&dn->dn_dirty_records[txgoff]); dnode_evict_dbufs(dn); /* * XXX - It would be nice to assert this, but we may still * have residual holds from async evictions from the arc... * * zfs_obj_to_path() also depends on this being * commented out. * * ASSERT3U(refcount_count(&dn->dn_holds), ==, 1); */ /* Undirty next bits */ dn->dn_next_nlevels[txgoff] = 0; dn->dn_next_indblkshift[txgoff] = 0; dn->dn_next_blksz[txgoff] = 0; /* ASSERT(blkptrs are zero); */ ASSERT(dn->dn_phys->dn_type != DMU_OT_NONE); ASSERT(dn->dn_type != DMU_OT_NONE); ASSERT(dn->dn_free_txg > 0); if (dn->dn_allocated_txg != dn->dn_free_txg) dmu_buf_will_dirty(&dn->dn_dbuf->db, tx); bzero(dn->dn_phys, sizeof (dnode_phys_t)); mutex_enter(&dn->dn_mtx); dn->dn_type = DMU_OT_NONE; dn->dn_maxblkid = 0; dn->dn_allocated_txg = 0; dn->dn_free_txg = 0; dn->dn_have_spill = B_FALSE; mutex_exit(&dn->dn_mtx); ASSERT(dn->dn_object != DMU_META_DNODE_OBJECT); dnode_rele(dn, (void *)(uintptr_t)tx->tx_txg); /* * Now that we've released our hold, the dnode may * be evicted, so we musn't access it. */ } /* * Write out the dnode's dirty buffers. */ void dnode_sync(dnode_t *dn, dmu_tx_t *tx) { dnode_phys_t *dnp = dn->dn_phys; int txgoff = tx->tx_txg & TXG_MASK; list_t *list = &dn->dn_dirty_records[txgoff]; static const dnode_phys_t zerodn = { 0 }; boolean_t kill_spill = B_FALSE; ASSERT(dmu_tx_is_syncing(tx)); ASSERT(dnp->dn_type != DMU_OT_NONE || dn->dn_allocated_txg); ASSERT(dnp->dn_type != DMU_OT_NONE || bcmp(dnp, &zerodn, DNODE_SIZE) == 0); DNODE_VERIFY(dn); ASSERT(dn->dn_dbuf == NULL || arc_released(dn->dn_dbuf->db_buf)); if (dmu_objset_userused_enabled(dn->dn_objset) && !DMU_OBJECT_IS_SPECIAL(dn->dn_object)) { mutex_enter(&dn->dn_mtx); dn->dn_oldused = DN_USED_BYTES(dn->dn_phys); dn->dn_oldflags = dn->dn_phys->dn_flags; dn->dn_phys->dn_flags |= DNODE_FLAG_USERUSED_ACCOUNTED; mutex_exit(&dn->dn_mtx); dmu_objset_userquota_get_ids(dn, B_FALSE, tx); } else { /* Once we account for it, we should always account for it. */ ASSERT(!(dn->dn_phys->dn_flags & DNODE_FLAG_USERUSED_ACCOUNTED)); } mutex_enter(&dn->dn_mtx); if (dn->dn_allocated_txg == tx->tx_txg) { /* The dnode is newly allocated or reallocated */ if (dnp->dn_type == DMU_OT_NONE) { /* this is a first alloc, not a realloc */ dnp->dn_nlevels = 1; dnp->dn_nblkptr = dn->dn_nblkptr; } dnp->dn_type = dn->dn_type; dnp->dn_bonustype = dn->dn_bonustype; dnp->dn_bonuslen = dn->dn_bonuslen; } ASSERT(dnp->dn_nlevels > 1 || BP_IS_HOLE(&dnp->dn_blkptr[0]) || BP_IS_EMBEDDED(&dnp->dn_blkptr[0]) || BP_GET_LSIZE(&dnp->dn_blkptr[0]) == dnp->dn_datablkszsec << SPA_MINBLOCKSHIFT); ASSERT(dnp->dn_nlevels < 2 || BP_IS_HOLE(&dnp->dn_blkptr[0]) || BP_GET_LSIZE(&dnp->dn_blkptr[0]) == 1 << dnp->dn_indblkshift); if (dn->dn_next_type[txgoff] != 0) { dnp->dn_type = dn->dn_type; dn->dn_next_type[txgoff] = 0; } if (dn->dn_next_blksz[txgoff] != 0) { ASSERT(P2PHASE(dn->dn_next_blksz[txgoff], SPA_MINBLOCKSIZE) == 0); ASSERT(BP_IS_HOLE(&dnp->dn_blkptr[0]) || dn->dn_maxblkid == 0 || list_head(list) != NULL || dn->dn_next_blksz[txgoff] >> SPA_MINBLOCKSHIFT == dnp->dn_datablkszsec || !range_tree_is_empty(dn->dn_free_ranges[txgoff])); dnp->dn_datablkszsec = dn->dn_next_blksz[txgoff] >> SPA_MINBLOCKSHIFT; dn->dn_next_blksz[txgoff] = 0; } if (dn->dn_next_bonuslen[txgoff] != 0) { if (dn->dn_next_bonuslen[txgoff] == DN_ZERO_BONUSLEN) dnp->dn_bonuslen = 0; else dnp->dn_bonuslen = dn->dn_next_bonuslen[txgoff]; ASSERT(dnp->dn_bonuslen <= DN_MAX_BONUSLEN); dn->dn_next_bonuslen[txgoff] = 0; } if (dn->dn_next_bonustype[txgoff] != 0) { ASSERT(DMU_OT_IS_VALID(dn->dn_next_bonustype[txgoff])); dnp->dn_bonustype = dn->dn_next_bonustype[txgoff]; dn->dn_next_bonustype[txgoff] = 0; } boolean_t freeing_dnode = dn->dn_free_txg > 0 && dn->dn_free_txg <= tx->tx_txg; /* * Remove the spill block if we have been explicitly asked to * remove it, or if the object is being removed. */ if (dn->dn_rm_spillblk[txgoff] || freeing_dnode) { if (dnp->dn_flags & DNODE_FLAG_SPILL_BLKPTR) kill_spill = B_TRUE; dn->dn_rm_spillblk[txgoff] = 0; } if (dn->dn_next_indblkshift[txgoff] != 0) { ASSERT(dnp->dn_nlevels == 1); dnp->dn_indblkshift = dn->dn_next_indblkshift[txgoff]; dn->dn_next_indblkshift[txgoff] = 0; } /* * Just take the live (open-context) values for checksum and compress. * Strictly speaking it's a future leak, but nothing bad happens if we * start using the new checksum or compress algorithm a little early. */ dnp->dn_checksum = dn->dn_checksum; dnp->dn_compress = dn->dn_compress; mutex_exit(&dn->dn_mtx); if (kill_spill) { free_blocks(dn, &dn->dn_phys->dn_spill, 1, tx); mutex_enter(&dn->dn_mtx); dnp->dn_flags &= ~DNODE_FLAG_SPILL_BLKPTR; mutex_exit(&dn->dn_mtx); } /* process all the "freed" ranges in the file */ if (dn->dn_free_ranges[txgoff] != NULL) { dnode_sync_free_range_arg_t dsfra; dsfra.dsfra_dnode = dn; dsfra.dsfra_tx = tx; + dsfra.dsfra_free_indirects = freeing_dnode; + if (freeing_dnode) { + ASSERT(range_tree_contains(dn->dn_free_ranges[txgoff], + 0, dn->dn_maxblkid + 1)); + } mutex_enter(&dn->dn_mtx); range_tree_vacate(dn->dn_free_ranges[txgoff], dnode_sync_free_range, &dsfra); range_tree_destroy(dn->dn_free_ranges[txgoff]); dn->dn_free_ranges[txgoff] = NULL; mutex_exit(&dn->dn_mtx); } if (freeing_dnode) { dn->dn_objset->os_freed_dnodes++; dnode_sync_free(dn, tx); return; } if (dn->dn_next_nlevels[txgoff]) { dnode_increase_indirection(dn, tx); dn->dn_next_nlevels[txgoff] = 0; } if (dn->dn_next_nblkptr[txgoff]) { /* this should only happen on a realloc */ ASSERT(dn->dn_allocated_txg == tx->tx_txg); if (dn->dn_next_nblkptr[txgoff] > dnp->dn_nblkptr) { /* zero the new blkptrs we are gaining */ bzero(dnp->dn_blkptr + dnp->dn_nblkptr, sizeof (blkptr_t) * (dn->dn_next_nblkptr[txgoff] - dnp->dn_nblkptr)); #ifdef ZFS_DEBUG } else { int i; ASSERT(dn->dn_next_nblkptr[txgoff] < dnp->dn_nblkptr); /* the blkptrs we are losing better be unallocated */ for (i = dn->dn_next_nblkptr[txgoff]; i < dnp->dn_nblkptr; i++) ASSERT(BP_IS_HOLE(&dnp->dn_blkptr[i])); #endif } mutex_enter(&dn->dn_mtx); dnp->dn_nblkptr = dn->dn_next_nblkptr[txgoff]; dn->dn_next_nblkptr[txgoff] = 0; mutex_exit(&dn->dn_mtx); } dbuf_sync_list(list, dn->dn_phys->dn_nlevels - 1, tx); if (!DMU_OBJECT_IS_SPECIAL(dn->dn_object)) { ASSERT3P(list_head(list), ==, NULL); dnode_rele(dn, (void *)(uintptr_t)tx->tx_txg); } /* * Although we have dropped our reference to the dnode, it * can't be evicted until its written, and we haven't yet * initiated the IO for the dnode's dbuf. */ } Index: stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c =================================================================== --- stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c (revision 339135) +++ stable/11/sys/cddl/contrib/opensolaris/uts/common/fs/zfs/zfs_znode.c (revision 339136) @@ -1,2258 +1,2259 @@ /* * CDDL HEADER START * * The contents of this file are subject to the terms of the * Common Development and Distribution License (the "License"). * You may not use this file except in compliance with the License. * * You can obtain a copy of the license at usr/src/OPENSOLARIS.LICENSE * or http://www.opensolaris.org/os/licensing. * See the License for the specific language governing permissions * and limitations under the License. * * When distributing Covered Code, include this CDDL HEADER in each * file and include the License file at usr/src/OPENSOLARIS.LICENSE. * If applicable, add the following below this CDDL HEADER, with the * fields enclosed by brackets "[]" replaced with your own identifying * information: Portions Copyright [yyyy] [name of copyright owner] * * CDDL HEADER END */ /* * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. * Copyright (c) 2012, 2014 by Delphix. All rights reserved. * Copyright (c) 2014 Integros [integros.com] */ /* Portions Copyright 2007 Jeremy Teo */ /* Portions Copyright 2011 Martin Matuska */ #ifdef _KERNEL #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #include #endif /* _KERNEL */ #include #include #include #include #include #include #include #include #include #include #include "zfs_prop.h" #include "zfs_comutil.h" /* Used by fstat(1). */ SYSCTL_INT(_debug_sizeof, OID_AUTO, znode, CTLFLAG_RD, SYSCTL_NULL_INT_PTR, sizeof(znode_t), "sizeof(znode_t)"); /* * Define ZNODE_STATS to turn on statistic gathering. By default, it is only * turned on when DEBUG is also defined. */ #ifdef DEBUG #define ZNODE_STATS #endif /* DEBUG */ #ifdef ZNODE_STATS #define ZNODE_STAT_ADD(stat) ((stat)++) #else #define ZNODE_STAT_ADD(stat) /* nothing */ #endif /* ZNODE_STATS */ /* * Functions needed for userland (ie: libzpool) are not put under * #ifdef_KERNEL; the rest of the functions have dependencies * (such as VFS logic) that will not compile easily in userland. */ #ifdef _KERNEL /* * Needed to close a small window in zfs_znode_move() that allows the zfsvfs to * be freed before it can be safely accessed. */ krwlock_t zfsvfs_lock; static kmem_cache_t *znode_cache = NULL; /*ARGSUSED*/ static void znode_evict_error(dmu_buf_t *dbuf, void *user_ptr) { /* * We should never drop all dbuf refs without first clearing * the eviction callback. */ panic("evicting znode %p\n", user_ptr); } extern struct vop_vector zfs_vnodeops; extern struct vop_vector zfs_fifoops; extern struct vop_vector zfs_shareops; static int zfs_znode_cache_constructor(void *buf, void *arg, int kmflags) { znode_t *zp = buf; POINTER_INVALIDATE(&zp->z_zfsvfs); list_link_init(&zp->z_link_node); mutex_init(&zp->z_acl_lock, NULL, MUTEX_DEFAULT, NULL); mutex_init(&zp->z_range_lock, NULL, MUTEX_DEFAULT, NULL); avl_create(&zp->z_range_avl, zfs_range_compare, sizeof (rl_t), offsetof(rl_t, r_node)); zp->z_acl_cached = NULL; zp->z_vnode = NULL; zp->z_moved = 0; return (0); } /*ARGSUSED*/ static void zfs_znode_cache_destructor(void *buf, void *arg) { znode_t *zp = buf; ASSERT(!POINTER_IS_VALID(zp->z_zfsvfs)); ASSERT3P(zp->z_vnode, ==, NULL); ASSERT(!list_link_active(&zp->z_link_node)); mutex_destroy(&zp->z_acl_lock); avl_destroy(&zp->z_range_avl); mutex_destroy(&zp->z_range_lock); ASSERT(zp->z_acl_cached == NULL); } #ifdef ZNODE_STATS static struct { uint64_t zms_zfsvfs_invalid; uint64_t zms_zfsvfs_recheck1; uint64_t zms_zfsvfs_unmounted; uint64_t zms_zfsvfs_recheck2; uint64_t zms_obj_held; uint64_t zms_vnode_locked; uint64_t zms_not_only_dnlc; } znode_move_stats; #endif /* ZNODE_STATS */ #ifdef illumos static void zfs_znode_move_impl(znode_t *ozp, znode_t *nzp) { vnode_t *vp; /* Copy fields. */ nzp->z_zfsvfs = ozp->z_zfsvfs; /* Swap vnodes. */ vp = nzp->z_vnode; nzp->z_vnode = ozp->z_vnode; ozp->z_vnode = vp; /* let destructor free the overwritten vnode */ ZTOV(ozp)->v_data = ozp; ZTOV(nzp)->v_data = nzp; nzp->z_id = ozp->z_id; ASSERT(ozp->z_dirlocks == NULL); /* znode not in use */ ASSERT(avl_numnodes(&ozp->z_range_avl) == 0); nzp->z_unlinked = ozp->z_unlinked; nzp->z_atime_dirty = ozp->z_atime_dirty; nzp->z_zn_prefetch = ozp->z_zn_prefetch; nzp->z_blksz = ozp->z_blksz; nzp->z_seq = ozp->z_seq; nzp->z_mapcnt = ozp->z_mapcnt; nzp->z_gen = ozp->z_gen; nzp->z_sync_cnt = ozp->z_sync_cnt; nzp->z_is_sa = ozp->z_is_sa; nzp->z_sa_hdl = ozp->z_sa_hdl; bcopy(ozp->z_atime, nzp->z_atime, sizeof (uint64_t) * 2); nzp->z_links = ozp->z_links; nzp->z_size = ozp->z_size; nzp->z_pflags = ozp->z_pflags; nzp->z_uid = ozp->z_uid; nzp->z_gid = ozp->z_gid; nzp->z_mode = ozp->z_mode; /* * Since this is just an idle znode and kmem is already dealing with * memory pressure, release any cached ACL. */ if (ozp->z_acl_cached) { zfs_acl_free(ozp->z_acl_cached); ozp->z_acl_cached = NULL; } sa_set_userp(nzp->z_sa_hdl, nzp); /* * Invalidate the original znode by clearing fields that provide a * pointer back to the znode. Set the low bit of the vfs pointer to * ensure that zfs_znode_move() recognizes the znode as invalid in any * subsequent callback. */ ozp->z_sa_hdl = NULL; POINTER_INVALIDATE(&ozp->z_zfsvfs); /* * Mark the znode. */ nzp->z_moved = 1; ozp->z_moved = (uint8_t)-1; } /*ARGSUSED*/ static kmem_cbrc_t zfs_znode_move(void *buf, void *newbuf, size_t size, void *arg) { znode_t *ozp = buf, *nzp = newbuf; zfsvfs_t *zfsvfs; vnode_t *vp; /* * The znode is on the file system's list of known znodes if the vfs * pointer is valid. We set the low bit of the vfs pointer when freeing * the znode to invalidate it, and the memory patterns written by kmem * (baddcafe and deadbeef) set at least one of the two low bits. A newly * created znode sets the vfs pointer last of all to indicate that the * znode is known and in a valid state to be moved by this function. */ zfsvfs = ozp->z_zfsvfs; if (!POINTER_IS_VALID(zfsvfs)) { ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_invalid); return (KMEM_CBRC_DONT_KNOW); } /* * Close a small window in which it's possible that the filesystem could * be unmounted and freed, and zfsvfs, though valid in the previous * statement, could point to unrelated memory by the time we try to * prevent the filesystem from being unmounted. */ rw_enter(&zfsvfs_lock, RW_WRITER); if (zfsvfs != ozp->z_zfsvfs) { rw_exit(&zfsvfs_lock); ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_recheck1); return (KMEM_CBRC_DONT_KNOW); } /* * If the znode is still valid, then so is the file system. We know that * no valid file system can be freed while we hold zfsvfs_lock, so we * can safely ensure that the filesystem is not and will not be * unmounted. The next statement is equivalent to ZFS_ENTER(). */ rrm_enter(&zfsvfs->z_teardown_lock, RW_READER, FTAG); if (zfsvfs->z_unmounted) { ZFS_EXIT(zfsvfs); rw_exit(&zfsvfs_lock); ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_unmounted); return (KMEM_CBRC_DONT_KNOW); } rw_exit(&zfsvfs_lock); mutex_enter(&zfsvfs->z_znodes_lock); /* * Recheck the vfs pointer in case the znode was removed just before * acquiring the lock. */ if (zfsvfs != ozp->z_zfsvfs) { mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_recheck2); return (KMEM_CBRC_DONT_KNOW); } /* * At this point we know that as long as we hold z_znodes_lock, the * znode cannot be freed and fields within the znode can be safely * accessed. Now, prevent a race with zfs_zget(). */ if (ZFS_OBJ_HOLD_TRYENTER(zfsvfs, ozp->z_id) == 0) { mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); ZNODE_STAT_ADD(znode_move_stats.zms_obj_held); return (KMEM_CBRC_LATER); } vp = ZTOV(ozp); if (mutex_tryenter(&vp->v_lock) == 0) { ZFS_OBJ_HOLD_EXIT(zfsvfs, ozp->z_id); mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); ZNODE_STAT_ADD(znode_move_stats.zms_vnode_locked); return (KMEM_CBRC_LATER); } /* Only move znodes that are referenced _only_ by the DNLC. */ if (vp->v_count != 1 || !vn_in_dnlc(vp)) { mutex_exit(&vp->v_lock); ZFS_OBJ_HOLD_EXIT(zfsvfs, ozp->z_id); mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); ZNODE_STAT_ADD(znode_move_stats.zms_not_only_dnlc); return (KMEM_CBRC_LATER); } /* * The znode is known and in a valid state to move. We're holding the * locks needed to execute the critical section. */ zfs_znode_move_impl(ozp, nzp); mutex_exit(&vp->v_lock); ZFS_OBJ_HOLD_EXIT(zfsvfs, ozp->z_id); list_link_replace(&ozp->z_link_node, &nzp->z_link_node); mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); return (KMEM_CBRC_YES); } #endif /* illumos */ void zfs_znode_init(void) { /* * Initialize zcache */ rw_init(&zfsvfs_lock, NULL, RW_DEFAULT, NULL); ASSERT(znode_cache == NULL); znode_cache = kmem_cache_create("zfs_znode_cache", sizeof (znode_t), 0, zfs_znode_cache_constructor, zfs_znode_cache_destructor, NULL, NULL, NULL, 0); kmem_cache_set_move(znode_cache, zfs_znode_move); } void zfs_znode_fini(void) { #ifdef illumos /* * Cleanup vfs & vnode ops */ zfs_remove_op_tables(); #endif /* * Cleanup zcache */ if (znode_cache) kmem_cache_destroy(znode_cache); znode_cache = NULL; rw_destroy(&zfsvfs_lock); } #ifdef illumos struct vnodeops *zfs_dvnodeops; struct vnodeops *zfs_fvnodeops; struct vnodeops *zfs_symvnodeops; struct vnodeops *zfs_xdvnodeops; struct vnodeops *zfs_evnodeops; struct vnodeops *zfs_sharevnodeops; void zfs_remove_op_tables() { /* * Remove vfs ops */ ASSERT(zfsfstype); (void) vfs_freevfsops_by_type(zfsfstype); zfsfstype = 0; /* * Remove vnode ops */ if (zfs_dvnodeops) vn_freevnodeops(zfs_dvnodeops); if (zfs_fvnodeops) vn_freevnodeops(zfs_fvnodeops); if (zfs_symvnodeops) vn_freevnodeops(zfs_symvnodeops); if (zfs_xdvnodeops) vn_freevnodeops(zfs_xdvnodeops); if (zfs_evnodeops) vn_freevnodeops(zfs_evnodeops); if (zfs_sharevnodeops) vn_freevnodeops(zfs_sharevnodeops); zfs_dvnodeops = NULL; zfs_fvnodeops = NULL; zfs_symvnodeops = NULL; zfs_xdvnodeops = NULL; zfs_evnodeops = NULL; zfs_sharevnodeops = NULL; } extern const fs_operation_def_t zfs_dvnodeops_template[]; extern const fs_operation_def_t zfs_fvnodeops_template[]; extern const fs_operation_def_t zfs_xdvnodeops_template[]; extern const fs_operation_def_t zfs_symvnodeops_template[]; extern const fs_operation_def_t zfs_evnodeops_template[]; extern const fs_operation_def_t zfs_sharevnodeops_template[]; int zfs_create_op_tables() { int error; /* * zfs_dvnodeops can be set if mod_remove() calls mod_installfs() * due to a failure to remove the the 2nd modlinkage (zfs_modldrv). * In this case we just return as the ops vectors are already set up. */ if (zfs_dvnodeops) return (0); error = vn_make_ops(MNTTYPE_ZFS, zfs_dvnodeops_template, &zfs_dvnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_fvnodeops_template, &zfs_fvnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_symvnodeops_template, &zfs_symvnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_xdvnodeops_template, &zfs_xdvnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_evnodeops_template, &zfs_evnodeops); if (error) return (error); error = vn_make_ops(MNTTYPE_ZFS, zfs_sharevnodeops_template, &zfs_sharevnodeops); return (error); } #endif /* illumos */ int zfs_create_share_dir(zfsvfs_t *zfsvfs, dmu_tx_t *tx) { zfs_acl_ids_t acl_ids; vattr_t vattr; znode_t *sharezp; znode_t *zp; int error; vattr.va_mask = AT_MODE|AT_UID|AT_GID|AT_TYPE; vattr.va_type = VDIR; vattr.va_mode = S_IFDIR|0555; vattr.va_uid = crgetuid(kcred); vattr.va_gid = crgetgid(kcred); sharezp = kmem_cache_alloc(znode_cache, KM_SLEEP); ASSERT(!POINTER_IS_VALID(sharezp->z_zfsvfs)); sharezp->z_moved = 0; sharezp->z_unlinked = 0; sharezp->z_atime_dirty = 0; sharezp->z_zfsvfs = zfsvfs; sharezp->z_is_sa = zfsvfs->z_use_sa; VERIFY(0 == zfs_acl_ids_create(sharezp, IS_ROOT_NODE, &vattr, kcred, NULL, &acl_ids)); zfs_mknode(sharezp, &vattr, tx, kcred, IS_ROOT_NODE, &zp, &acl_ids); ASSERT3P(zp, ==, sharezp); POINTER_INVALIDATE(&sharezp->z_zfsvfs); error = zap_add(zfsvfs->z_os, MASTER_NODE_OBJ, ZFS_SHARES_DIR, 8, 1, &sharezp->z_id, tx); zfsvfs->z_shares_dir = sharezp->z_id; zfs_acl_ids_free(&acl_ids); sa_handle_destroy(sharezp->z_sa_hdl); kmem_cache_free(znode_cache, sharezp); return (error); } /* * define a couple of values we need available * for both 64 and 32 bit environments. */ #ifndef NBITSMINOR64 #define NBITSMINOR64 32 #endif #ifndef MAXMAJ64 #define MAXMAJ64 0xffffffffUL #endif #ifndef MAXMIN64 #define MAXMIN64 0xffffffffUL #endif /* * Create special expldev for ZFS private use. * Can't use standard expldev since it doesn't do * what we want. The standard expldev() takes a * dev32_t in LP64 and expands it to a long dev_t. * We need an interface that takes a dev32_t in ILP32 * and expands it to a long dev_t. */ static uint64_t zfs_expldev(dev_t dev) { return (((uint64_t)major(dev) << NBITSMINOR64) | minor(dev)); } /* * Special cmpldev for ZFS private use. * Can't use standard cmpldev since it takes * a long dev_t and compresses it to dev32_t in * LP64. We need to do a compaction of a long dev_t * to a dev32_t in ILP32. */ dev_t zfs_cmpldev(uint64_t dev) { return (makedev((dev >> NBITSMINOR64), (dev & MAXMIN64))); } static void zfs_znode_sa_init(zfsvfs_t *zfsvfs, znode_t *zp, dmu_buf_t *db, dmu_object_type_t obj_type, sa_handle_t *sa_hdl) { ASSERT(!POINTER_IS_VALID(zp->z_zfsvfs) || (zfsvfs == zp->z_zfsvfs)); ASSERT(MUTEX_HELD(ZFS_OBJ_MUTEX(zfsvfs, zp->z_id))); ASSERT(zp->z_sa_hdl == NULL); ASSERT(zp->z_acl_cached == NULL); if (sa_hdl == NULL) { VERIFY(0 == sa_handle_get_from_db(zfsvfs->z_os, db, zp, SA_HDL_SHARED, &zp->z_sa_hdl)); } else { zp->z_sa_hdl = sa_hdl; sa_set_userp(sa_hdl, zp); } zp->z_is_sa = (obj_type == DMU_OT_SA) ? B_TRUE : B_FALSE; /* * Slap on VROOT if we are the root znode unless we are the root * node of a snapshot mounted under .zfs. */ if (zp->z_id == zfsvfs->z_root && zfsvfs->z_parent == zfsvfs) ZTOV(zp)->v_flag |= VROOT; vn_exists(ZTOV(zp)); } void zfs_znode_dmu_fini(znode_t *zp) { ASSERT(MUTEX_HELD(ZFS_OBJ_MUTEX(zp->z_zfsvfs, zp->z_id)) || zp->z_unlinked || RW_WRITE_HELD(&zp->z_zfsvfs->z_teardown_inactive_lock)); sa_handle_destroy(zp->z_sa_hdl); zp->z_sa_hdl = NULL; } static void zfs_vnode_forget(vnode_t *vp) { /* copied from insmntque_stddtr */ vp->v_data = NULL; vp->v_op = &dead_vnodeops; vgone(vp); vput(vp); } /* * Construct a new znode/vnode and intialize. * * This does not do a call to dmu_set_user() that is * up to the caller to do, in case you don't want to * return the znode */ static znode_t * zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz, dmu_object_type_t obj_type, sa_handle_t *hdl) { znode_t *zp; vnode_t *vp; uint64_t mode; uint64_t parent; sa_bulk_attr_t bulk[9]; int count = 0; int error; zp = kmem_cache_alloc(znode_cache, KM_SLEEP); KASSERT(curthread->td_vp_reserv > 0, ("zfs_znode_alloc: getnewvnode without any vnodes reserved")); error = getnewvnode("zfs", zfsvfs->z_parent->z_vfs, &zfs_vnodeops, &vp); if (error != 0) { kmem_cache_free(znode_cache, zp); return (NULL); } zp->z_vnode = vp; vp->v_data = zp; ASSERT(!POINTER_IS_VALID(zp->z_zfsvfs)); zp->z_moved = 0; /* * Defer setting z_zfsvfs until the znode is ready to be a candidate for * the zfs_znode_move() callback. */ zp->z_sa_hdl = NULL; zp->z_unlinked = 0; zp->z_atime_dirty = 0; zp->z_mapcnt = 0; zp->z_id = db->db_object; zp->z_blksz = blksz; zp->z_seq = 0x7A4653; zp->z_sync_cnt = 0; vp = ZTOV(zp); zfs_znode_sa_init(zfsvfs, zp, db, obj_type, hdl); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GEN(zfsvfs), NULL, &zp->z_gen, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL, &zp->z_size, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs), NULL, &zp->z_links, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_PARENT(zfsvfs), NULL, &parent, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL, &zp->z_atime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL, &zp->z_uid, 8); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL, &zp->z_gid, 8); if (sa_bulk_lookup(zp->z_sa_hdl, bulk, count) != 0 || zp->z_gen == 0) { if (hdl == NULL) sa_handle_destroy(zp->z_sa_hdl); zfs_vnode_forget(vp); zp->z_vnode = NULL; kmem_cache_free(znode_cache, zp); return (NULL); } zp->z_mode = mode; vp->v_type = IFTOVT((mode_t)mode); switch (vp->v_type) { case VDIR: zp->z_zn_prefetch = B_TRUE; /* z_prefetch default is enabled */ break; #ifdef illumos case VBLK: case VCHR: { uint64_t rdev; VERIFY(sa_lookup(zp->z_sa_hdl, SA_ZPL_RDEV(zfsvfs), &rdev, sizeof (rdev)) == 0); vp->v_rdev = zfs_cmpldev(rdev); } break; #endif case VFIFO: #ifdef illumos case VSOCK: case VDOOR: #endif vp->v_op = &zfs_fifoops; break; case VREG: if (parent == zfsvfs->z_shares_dir) { ASSERT(zp->z_uid == 0 && zp->z_gid == 0); vp->v_op = &zfs_shareops; } break; #ifdef illumos case VLNK: vn_setops(vp, zfs_symvnodeops); break; default: vn_setops(vp, zfs_evnodeops); break; #endif } mutex_enter(&zfsvfs->z_znodes_lock); list_insert_tail(&zfsvfs->z_all_znodes, zp); membar_producer(); /* * Everything else must be valid before assigning z_zfsvfs makes the * znode eligible for zfs_znode_move(). */ zp->z_zfsvfs = zfsvfs; mutex_exit(&zfsvfs->z_znodes_lock); /* * Acquire vnode lock before making it available to the world. */ vn_lock(vp, LK_EXCLUSIVE | LK_RETRY); VN_LOCK_AREC(vp); if (vp->v_type != VFIFO) VN_LOCK_ASHARE(vp); #ifdef illumos VFS_HOLD(zfsvfs->z_vfs); #endif return (zp); } static uint64_t empty_xattr; static uint64_t pad[4]; static zfs_acl_phys_t acl_phys; /* * Create a new DMU object to hold a zfs znode. * * IN: dzp - parent directory for new znode * vap - file attributes for new znode * tx - dmu transaction id for zap operations * cr - credentials of caller * flag - flags: * IS_ROOT_NODE - new object will be root * IS_XATTR - new object is an attribute * bonuslen - length of bonus buffer * setaclp - File/Dir initial ACL * fuidp - Tracks fuid allocation. * * OUT: zpp - allocated znode * */ void zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr, uint_t flag, znode_t **zpp, zfs_acl_ids_t *acl_ids) { uint64_t crtime[2], atime[2], mtime[2], ctime[2]; uint64_t mode, size, links, parent, pflags; uint64_t dzp_pflags = 0; uint64_t rdev = 0; zfsvfs_t *zfsvfs = dzp->z_zfsvfs; dmu_buf_t *db; timestruc_t now; uint64_t gen, obj; int err; int bonuslen; sa_handle_t *sa_hdl; dmu_object_type_t obj_type; sa_bulk_attr_t sa_attrs[ZPL_END]; int cnt = 0; zfs_acl_locator_cb_t locate = { 0 }; ASSERT(vap && (vap->va_mask & (AT_TYPE|AT_MODE)) == (AT_TYPE|AT_MODE)); if (zfsvfs->z_replay) { obj = vap->va_nodeid; now = vap->va_ctime; /* see zfs_replay_create() */ gen = vap->va_nblocks; /* ditto */ } else { obj = 0; vfs_timestamp(&now); gen = dmu_tx_get_txg(tx); } obj_type = zfsvfs->z_use_sa ? DMU_OT_SA : DMU_OT_ZNODE; bonuslen = (obj_type == DMU_OT_SA) ? DN_MAX_BONUSLEN : ZFS_OLD_ZNODE_PHYS_SIZE; /* * Create a new DMU object. */ /* * There's currently no mechanism for pre-reading the blocks that will * be needed to allocate a new object, so we accept the small chance * that there will be an i/o error and we will fail one of the * assertions below. */ if (vap->va_type == VDIR) { if (zfsvfs->z_replay) { VERIFY0(zap_create_claim_norm(zfsvfs->z_os, obj, zfsvfs->z_norm, DMU_OT_DIRECTORY_CONTENTS, obj_type, bonuslen, tx)); } else { obj = zap_create_norm(zfsvfs->z_os, zfsvfs->z_norm, DMU_OT_DIRECTORY_CONTENTS, obj_type, bonuslen, tx); } } else { if (zfsvfs->z_replay) { VERIFY0(dmu_object_claim(zfsvfs->z_os, obj, DMU_OT_PLAIN_FILE_CONTENTS, 0, obj_type, bonuslen, tx)); } else { obj = dmu_object_alloc(zfsvfs->z_os, DMU_OT_PLAIN_FILE_CONTENTS, 0, obj_type, bonuslen, tx); } } ZFS_OBJ_HOLD_ENTER(zfsvfs, obj); VERIFY(0 == sa_buf_hold(zfsvfs->z_os, obj, NULL, &db)); /* * If this is the root, fix up the half-initialized parent pointer * to reference the just-allocated physical data area. */ if (flag & IS_ROOT_NODE) { dzp->z_id = obj; } else { dzp_pflags = dzp->z_pflags; } /* * If parent is an xattr, so am I. */ if (dzp_pflags & ZFS_XATTR) { flag |= IS_XATTR; } if (zfsvfs->z_use_fuids) pflags = ZFS_ARCHIVE | ZFS_AV_MODIFIED; else pflags = 0; if (vap->va_type == VDIR) { size = 2; /* contents ("." and "..") */ links = (flag & (IS_ROOT_NODE | IS_XATTR)) ? 2 : 1; } else { size = links = 0; } if (vap->va_type == VBLK || vap->va_type == VCHR) { rdev = zfs_expldev(vap->va_rdev); } parent = dzp->z_id; mode = acl_ids->z_mode; if (flag & IS_XATTR) pflags |= ZFS_XATTR; /* * No execs denied will be deterimed when zfs_mode_compute() is called. */ pflags |= acl_ids->z_aclp->z_hints & (ZFS_ACL_TRIVIAL|ZFS_INHERIT_ACE|ZFS_ACL_AUTO_INHERIT| ZFS_ACL_DEFAULTED|ZFS_ACL_PROTECTED); ZFS_TIME_ENCODE(&now, crtime); ZFS_TIME_ENCODE(&now, ctime); if (vap->va_mask & AT_ATIME) { ZFS_TIME_ENCODE(&vap->va_atime, atime); } else { ZFS_TIME_ENCODE(&now, atime); } if (vap->va_mask & AT_MTIME) { ZFS_TIME_ENCODE(&vap->va_mtime, mtime); } else { ZFS_TIME_ENCODE(&now, mtime); } /* Now add in all of the "SA" attributes */ VERIFY(0 == sa_handle_get_from_db(zfsvfs->z_os, db, NULL, SA_HDL_SHARED, &sa_hdl)); /* * Setup the array of attributes to be replaced/set on the new file * * order for DMU_OT_ZNODE is critical since it needs to be constructed * in the old znode_phys_t format. Don't change this ordering */ if (obj_type == DMU_OT_ZNODE) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ATIME(zfsvfs), NULL, &atime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CRTIME(zfsvfs), NULL, &crtime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GEN(zfsvfs), NULL, &gen, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_SIZE(zfsvfs), NULL, &size, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PARENT(zfsvfs), NULL, &parent, 8); } else { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MODE(zfsvfs), NULL, &mode, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_SIZE(zfsvfs), NULL, &size, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GEN(zfsvfs), NULL, &gen, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_UID(zfsvfs), NULL, &acl_ids->z_fuid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GID(zfsvfs), NULL, &acl_ids->z_fgid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PARENT(zfsvfs), NULL, &parent, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_FLAGS(zfsvfs), NULL, &pflags, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ATIME(zfsvfs), NULL, &atime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MTIME(zfsvfs), NULL, &mtime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CTIME(zfsvfs), NULL, &ctime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CRTIME(zfsvfs), NULL, &crtime, 16); } SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_LINKS(zfsvfs), NULL, &links, 8); if (obj_type == DMU_OT_ZNODE) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_XATTR(zfsvfs), NULL, &empty_xattr, 8); } if (obj_type == DMU_OT_ZNODE || (vap->va_type == VBLK || vap->va_type == VCHR)) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_RDEV(zfsvfs), NULL, &rdev, 8); } if (obj_type == DMU_OT_ZNODE) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_FLAGS(zfsvfs), NULL, &pflags, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_UID(zfsvfs), NULL, &acl_ids->z_fuid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GID(zfsvfs), NULL, &acl_ids->z_fgid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PAD(zfsvfs), NULL, pad, sizeof (uint64_t) * 4); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ZNODE_ACL(zfsvfs), NULL, &acl_phys, sizeof (zfs_acl_phys_t)); } else if (acl_ids->z_aclp->z_version >= ZFS_ACL_VERSION_FUID) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_DACL_COUNT(zfsvfs), NULL, &acl_ids->z_aclp->z_acl_count, 8); locate.cb_aclp = acl_ids->z_aclp; SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_DACL_ACES(zfsvfs), zfs_acl_data_locator, &locate, acl_ids->z_aclp->z_acl_bytes); mode = zfs_mode_compute(mode, acl_ids->z_aclp, &pflags, acl_ids->z_fuid, acl_ids->z_fgid); } VERIFY(sa_replace_all_by_template(sa_hdl, sa_attrs, cnt, tx) == 0); if (!(flag & IS_ROOT_NODE)) { *zpp = zfs_znode_alloc(zfsvfs, db, 0, obj_type, sa_hdl); ASSERT(*zpp != NULL); } else { /* * If we are creating the root node, the "parent" we * passed in is the znode for the root. */ *zpp = dzp; (*zpp)->z_sa_hdl = sa_hdl; } (*zpp)->z_pflags = pflags; (*zpp)->z_mode = mode; if (vap->va_mask & AT_XVATTR) zfs_xvattr_set(*zpp, (xvattr_t *)vap, tx); if (obj_type == DMU_OT_ZNODE || acl_ids->z_aclp->z_version < ZFS_ACL_VERSION_FUID) { VERIFY0(zfs_aclset_common(*zpp, acl_ids->z_aclp, cr, tx)); } if (!(flag & IS_ROOT_NODE)) { vnode_t *vp; vp = ZTOV(*zpp); vp->v_vflag |= VV_FORCEINSMQ; err = insmntque(vp, zfsvfs->z_vfs); vp->v_vflag &= ~VV_FORCEINSMQ; KASSERT(err == 0, ("insmntque() failed: error %d", err)); } ZFS_OBJ_HOLD_EXIT(zfsvfs, obj); } /* * Update in-core attributes. It is assumed the caller will be doing an * sa_bulk_update to push the changes out. */ void zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx) { xoptattr_t *xoap; xoap = xva_getxoptattr(xvap); ASSERT(xoap); if (XVA_ISSET_REQ(xvap, XAT_CREATETIME)) { uint64_t times[2]; ZFS_TIME_ENCODE(&xoap->xoa_createtime, times); (void) sa_update(zp->z_sa_hdl, SA_ZPL_CRTIME(zp->z_zfsvfs), ×, sizeof (times), tx); XVA_SET_RTN(xvap, XAT_CREATETIME); } if (XVA_ISSET_REQ(xvap, XAT_READONLY)) { ZFS_ATTR_SET(zp, ZFS_READONLY, xoap->xoa_readonly, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_READONLY); } if (XVA_ISSET_REQ(xvap, XAT_HIDDEN)) { ZFS_ATTR_SET(zp, ZFS_HIDDEN, xoap->xoa_hidden, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_HIDDEN); } if (XVA_ISSET_REQ(xvap, XAT_SYSTEM)) { ZFS_ATTR_SET(zp, ZFS_SYSTEM, xoap->xoa_system, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_SYSTEM); } if (XVA_ISSET_REQ(xvap, XAT_ARCHIVE)) { ZFS_ATTR_SET(zp, ZFS_ARCHIVE, xoap->xoa_archive, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_ARCHIVE); } if (XVA_ISSET_REQ(xvap, XAT_IMMUTABLE)) { ZFS_ATTR_SET(zp, ZFS_IMMUTABLE, xoap->xoa_immutable, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_IMMUTABLE); } if (XVA_ISSET_REQ(xvap, XAT_NOUNLINK)) { ZFS_ATTR_SET(zp, ZFS_NOUNLINK, xoap->xoa_nounlink, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_NOUNLINK); } if (XVA_ISSET_REQ(xvap, XAT_APPENDONLY)) { ZFS_ATTR_SET(zp, ZFS_APPENDONLY, xoap->xoa_appendonly, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_APPENDONLY); } if (XVA_ISSET_REQ(xvap, XAT_NODUMP)) { ZFS_ATTR_SET(zp, ZFS_NODUMP, xoap->xoa_nodump, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_NODUMP); } if (XVA_ISSET_REQ(xvap, XAT_OPAQUE)) { ZFS_ATTR_SET(zp, ZFS_OPAQUE, xoap->xoa_opaque, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_OPAQUE); } if (XVA_ISSET_REQ(xvap, XAT_AV_QUARANTINED)) { ZFS_ATTR_SET(zp, ZFS_AV_QUARANTINED, xoap->xoa_av_quarantined, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_AV_QUARANTINED); } if (XVA_ISSET_REQ(xvap, XAT_AV_MODIFIED)) { ZFS_ATTR_SET(zp, ZFS_AV_MODIFIED, xoap->xoa_av_modified, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_AV_MODIFIED); } if (XVA_ISSET_REQ(xvap, XAT_AV_SCANSTAMP)) { zfs_sa_set_scanstamp(zp, xvap, tx); XVA_SET_RTN(xvap, XAT_AV_SCANSTAMP); } if (XVA_ISSET_REQ(xvap, XAT_REPARSE)) { ZFS_ATTR_SET(zp, ZFS_REPARSE, xoap->xoa_reparse, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_REPARSE); } if (XVA_ISSET_REQ(xvap, XAT_OFFLINE)) { ZFS_ATTR_SET(zp, ZFS_OFFLINE, xoap->xoa_offline, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_OFFLINE); } if (XVA_ISSET_REQ(xvap, XAT_SPARSE)) { ZFS_ATTR_SET(zp, ZFS_SPARSE, xoap->xoa_sparse, zp->z_pflags, tx); XVA_SET_RTN(xvap, XAT_SPARSE); } } int zfs_zget(zfsvfs_t *zfsvfs, uint64_t obj_num, znode_t **zpp) { dmu_object_info_t doi; dmu_buf_t *db; znode_t *zp; vnode_t *vp; sa_handle_t *hdl; struct thread *td; int locked; int err; td = curthread; getnewvnode_reserve(1); again: *zpp = NULL; ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num); err = sa_buf_hold(zfsvfs->z_os, obj_num, NULL, &db); if (err) { ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); getnewvnode_drop_reserve(); return (err); } dmu_object_info_from_db(db, &doi); if (doi.doi_bonus_type != DMU_OT_SA && (doi.doi_bonus_type != DMU_OT_ZNODE || (doi.doi_bonus_type == DMU_OT_ZNODE && doi.doi_bonus_size < sizeof (znode_phys_t)))) { sa_buf_rele(db, NULL); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); #ifdef __FreeBSD__ getnewvnode_drop_reserve(); #endif return (SET_ERROR(EINVAL)); } hdl = dmu_buf_get_user(db); if (hdl != NULL) { zp = sa_get_userdata(hdl); /* * Since "SA" does immediate eviction we * should never find a sa handle that doesn't * know about the znode. */ ASSERT3P(zp, !=, NULL); ASSERT3U(zp->z_id, ==, obj_num); *zpp = zp; vp = ZTOV(zp); /* Don't let the vnode disappear after ZFS_OBJ_HOLD_EXIT. */ VN_HOLD(vp); sa_buf_rele(db, NULL); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); locked = VOP_ISLOCKED(vp); VI_LOCK(vp); if ((vp->v_iflag & VI_DOOMED) != 0 && locked != LK_EXCLUSIVE) { /* * The vnode is doomed and this thread doesn't * hold the exclusive lock on it, so the vnode * must be being reclaimed by another thread. * Otherwise the doomed vnode is being reclaimed * by this thread and zfs_zget is called from * ZIL internals. */ VI_UNLOCK(vp); /* * XXX vrele() locks the vnode when the last reference * is dropped. Although in this case the vnode is * doomed / dead and so no inactivation is required, * the vnode lock is still acquired. That could result * in a LOR with z_teardown_lock if another thread holds * the vnode's lock and tries to take z_teardown_lock. * But that is only possible if the other thread peforms * a ZFS vnode operation on the vnode. That either * should not happen if the vnode is dead or the thread * should also have a refrence to the vnode and thus * our reference is not last. */ VN_RELE(vp); goto again; } VI_UNLOCK(vp); getnewvnode_drop_reserve(); return (0); } /* * Not found create new znode/vnode * but only if file exists. * * There is a small window where zfs_vget() could * find this object while a file create is still in * progress. This is checked for in zfs_znode_alloc() * * if zfs_znode_alloc() fails it will drop the hold on the * bonus buffer. */ zp = zfs_znode_alloc(zfsvfs, db, doi.doi_data_block_size, doi.doi_bonus_type, NULL); if (zp == NULL) { err = SET_ERROR(ENOENT); } else { *zpp = zp; } if (err == 0) { vnode_t *vp = ZTOV(zp); err = insmntque(vp, zfsvfs->z_vfs); if (err == 0) { vp->v_hash = obj_num; VOP_UNLOCK(vp, 0); } else { zp->z_vnode = NULL; zfs_znode_dmu_fini(zp); zfs_znode_free(zp); *zpp = NULL; } } ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); getnewvnode_drop_reserve(); return (err); } int zfs_rezget(znode_t *zp) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; dmu_object_info_t doi; dmu_buf_t *db; vnode_t *vp; uint64_t obj_num = zp->z_id; uint64_t mode, size; sa_bulk_attr_t bulk[8]; int err; int count = 0; uint64_t gen; /* * Remove cached pages before reloading the znode, so that they are not * lingering after we run into any error. Ideally, we should vgone() * the vnode in case of error, but currently we cannot do that * because of the LOR between the vnode lock and z_teardown_lock. * So, instead, we have to "doom" the znode in the illumos style. */ vp = ZTOV(zp); vn_pages_remove(vp, 0, 0); ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num); mutex_enter(&zp->z_acl_lock); if (zp->z_acl_cached) { zfs_acl_free(zp->z_acl_cached); zp->z_acl_cached = NULL; } mutex_exit(&zp->z_acl_lock); ASSERT(zp->z_sa_hdl == NULL); err = sa_buf_hold(zfsvfs->z_os, obj_num, NULL, &db); if (err) { ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (err); } dmu_object_info_from_db(db, &doi); if (doi.doi_bonus_type != DMU_OT_SA && (doi.doi_bonus_type != DMU_OT_ZNODE || (doi.doi_bonus_type == DMU_OT_ZNODE && doi.doi_bonus_size < sizeof (znode_phys_t)))) { sa_buf_rele(db, NULL); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (SET_ERROR(EINVAL)); } zfs_znode_sa_init(zfsvfs, zp, db, doi.doi_bonus_type, NULL); size = zp->z_size; /* reload cached values */ SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GEN(zfsvfs), NULL, &gen, sizeof (gen)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL, &zp->z_size, sizeof (zp->z_size)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_LINKS(zfsvfs), NULL, &zp->z_links, sizeof (zp->z_links)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, sizeof (zp->z_pflags)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_ATIME(zfsvfs), NULL, &zp->z_atime, sizeof (zp->z_atime)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_UID(zfsvfs), NULL, &zp->z_uid, sizeof (zp->z_uid)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_GID(zfsvfs), NULL, &zp->z_gid, sizeof (zp->z_gid)); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MODE(zfsvfs), NULL, &mode, sizeof (mode)); if (sa_bulk_lookup(zp->z_sa_hdl, bulk, count)) { zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (SET_ERROR(EIO)); } zp->z_mode = mode; if (gen != zp->z_gen) { zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (SET_ERROR(EIO)); } /* * It is highly improbable but still quite possible that two * objects in different datasets are created with the same * object numbers and in transaction groups with the same * numbers. znodes corresponding to those objects would * have the same z_id and z_gen, but their other attributes * may be different. * zfs recv -F may replace one of such objects with the other. * As a result file properties recorded in the replaced * object's vnode may no longer match the received object's * properties. At present the only cached property is the * files type recorded in v_type. * So, handle this case by leaving the old vnode and znode * disassociated from the actual object. A new vnode and a * znode will be created if the object is accessed * (e.g. via a look-up). The old vnode and znode will be * recycled when the last vnode reference is dropped. */ if (vp->v_type != IFTOVT((mode_t)zp->z_mode)) { zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (SET_ERROR(EIO)); } /* * If the file has zero links, then it has been unlinked on the send * side and it must be in the received unlinked set. * We call zfs_znode_dmu_fini() now to prevent any accesses to the * stale data and to prevent automatical removal of the file in * zfs_zinactive(). The file will be removed either when it is removed * on the send side and the next incremental stream is received or * when the unlinked set gets processed. */ zp->z_unlinked = (zp->z_links == 0); if (zp->z_unlinked) { zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (0); } zp->z_blksz = doi.doi_data_block_size; if (zp->z_size != size) vnode_pager_setsize(vp, zp->z_size); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (0); } void zfs_znode_delete(znode_t *zp, dmu_tx_t *tx) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; objset_t *os = zfsvfs->z_os; uint64_t obj = zp->z_id; uint64_t acl_obj = zfs_external_acl(zp); ZFS_OBJ_HOLD_ENTER(zfsvfs, obj); if (acl_obj) { VERIFY(!zp->z_is_sa); VERIFY(0 == dmu_object_free(os, acl_obj, tx)); } VERIFY(0 == dmu_object_free(os, obj, tx)); zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj); zfs_znode_free(zp); } void zfs_zinactive(znode_t *zp) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; uint64_t z_id = zp->z_id; ASSERT(zp->z_sa_hdl); /* * Don't allow a zfs_zget() while were trying to release this znode */ ZFS_OBJ_HOLD_ENTER(zfsvfs, z_id); /* * If this was the last reference to a file with no links, remove * the file from the file system unless the file system is mounted * read-only. That can happen, for example, if the file system was * originally read-write, the file was opened, then unlinked and * the file system was made read-only before the file was finally * closed. The file will remain in the unlinked set. */ if (zp->z_unlinked) { ASSERT(!zfsvfs->z_issnap); if ((zfsvfs->z_vfs->vfs_flag & VFS_RDONLY) == 0) { ZFS_OBJ_HOLD_EXIT(zfsvfs, z_id); zfs_rmnode(zp); return; } } zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, z_id); zfs_znode_free(zp); } void zfs_znode_free(znode_t *zp) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; ASSERT(zp->z_sa_hdl == NULL); zp->z_vnode = NULL; mutex_enter(&zfsvfs->z_znodes_lock); POINTER_INVALIDATE(&zp->z_zfsvfs); list_remove(&zfsvfs->z_all_znodes, zp); mutex_exit(&zfsvfs->z_znodes_lock); if (zp->z_acl_cached) { zfs_acl_free(zp->z_acl_cached); zp->z_acl_cached = NULL; } kmem_cache_free(znode_cache, zp); #ifdef illumos VFS_RELE(zfsvfs->z_vfs); #endif } void zfs_tstamp_update_setup(znode_t *zp, uint_t flag, uint64_t mtime[2], uint64_t ctime[2], boolean_t have_tx) { timestruc_t now; vfs_timestamp(&now); if (have_tx) { /* will sa_bulk_update happen really soon? */ zp->z_atime_dirty = 0; zp->z_seq++; } else { zp->z_atime_dirty = 1; } if (flag & AT_ATIME) { ZFS_TIME_ENCODE(&now, zp->z_atime); } if (flag & AT_MTIME) { ZFS_TIME_ENCODE(&now, mtime); if (zp->z_zfsvfs->z_use_fuids) { zp->z_pflags |= (ZFS_ARCHIVE | ZFS_AV_MODIFIED); } } if (flag & AT_CTIME) { ZFS_TIME_ENCODE(&now, ctime); if (zp->z_zfsvfs->z_use_fuids) zp->z_pflags |= ZFS_ARCHIVE; } } /* * Grow the block size for a file. * * IN: zp - znode of file to free data in. * size - requested block size * tx - open transaction. * * NOTE: this function assumes that the znode is write locked. */ void zfs_grow_blocksize(znode_t *zp, uint64_t size, dmu_tx_t *tx) { int error; u_longlong_t dummy; if (size <= zp->z_blksz) return; /* * If the file size is already greater than the current blocksize, * we will not grow. If there is more than one block in a file, * the blocksize cannot change. */ if (zp->z_blksz && zp->z_size > zp->z_blksz) return; error = dmu_object_set_blocksize(zp->z_zfsvfs->z_os, zp->z_id, size, 0, tx); if (error == ENOTSUP) return; ASSERT0(error); /* What blocksize did we actually get? */ dmu_object_size_from_db(sa_get_db(zp->z_sa_hdl), &zp->z_blksz, &dummy); } #ifdef illumos /* * This is a dummy interface used when pvn_vplist_dirty() should *not* * be calling back into the fs for a putpage(). E.g.: when truncating * a file, the pages being "thrown away* don't need to be written out. */ /* ARGSUSED */ static int zfs_no_putpage(vnode_t *vp, page_t *pp, u_offset_t *offp, size_t *lenp, int flags, cred_t *cr) { ASSERT(0); return (0); } #endif /* * Increase the file length * * IN: zp - znode of file to free data in. * end - new end-of-file * * RETURN: 0 on success, error code on failure */ static int zfs_extend(znode_t *zp, uint64_t end) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; dmu_tx_t *tx; rl_t *rl; uint64_t newblksz; int error; /* * We will change zp_size, lock the whole file. */ rl = zfs_range_lock(zp, 0, UINT64_MAX, RL_WRITER); /* * Nothing to do if file already at desired length. */ if (end <= zp->z_size) { zfs_range_unlock(rl); return (0); } tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); if (end > zp->z_blksz && (!ISP2(zp->z_blksz) || zp->z_blksz < zfsvfs->z_max_blksz)) { /* * We are growing the file past the current block size. */ if (zp->z_blksz > zp->z_zfsvfs->z_max_blksz) { /* * File's blocksize is already larger than the * "recordsize" property. Only let it grow to * the next power of 2. */ ASSERT(!ISP2(zp->z_blksz)); newblksz = MIN(end, 1 << highbit64(zp->z_blksz)); } else { newblksz = MIN(end, zp->z_zfsvfs->z_max_blksz); } dmu_tx_hold_write(tx, zp->z_id, 0, newblksz); } else { newblksz = 0; } error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); zfs_range_unlock(rl); return (error); } if (newblksz) zfs_grow_blocksize(zp, newblksz, tx); zp->z_size = end; VERIFY(0 == sa_update(zp->z_sa_hdl, SA_ZPL_SIZE(zp->z_zfsvfs), &zp->z_size, sizeof (zp->z_size), tx)); vnode_pager_setsize(ZTOV(zp), end); zfs_range_unlock(rl); dmu_tx_commit(tx); return (0); } /* * Free space in a file. * * IN: zp - znode of file to free data in. * off - start of section to free. * len - length of section to free. * * RETURN: 0 on success, error code on failure */ static int zfs_free_range(znode_t *zp, uint64_t off, uint64_t len) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; rl_t *rl; int error; /* * Lock the range being freed. */ rl = zfs_range_lock(zp, off, len, RL_WRITER); /* * Nothing to do if file already at desired length. */ if (off >= zp->z_size) { zfs_range_unlock(rl); return (0); } if (off + len > zp->z_size) len = zp->z_size - off; error = dmu_free_long_range(zfsvfs->z_os, zp->z_id, off, len); if (error == 0) { /* * In FreeBSD we cannot free block in the middle of a file, * but only at the end of a file, so this code path should * never happen. */ vnode_pager_setsize(ZTOV(zp), off); } zfs_range_unlock(rl); return (error); } /* * Truncate a file * * IN: zp - znode of file to free data in. * end - new end-of-file. * * RETURN: 0 on success, error code on failure */ static int zfs_trunc(znode_t *zp, uint64_t end) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; vnode_t *vp = ZTOV(zp); dmu_tx_t *tx; rl_t *rl; int error; sa_bulk_attr_t bulk[2]; int count = 0; /* * We will change zp_size, lock the whole file. */ rl = zfs_range_lock(zp, 0, UINT64_MAX, RL_WRITER); /* * Nothing to do if file already at desired length. */ if (end >= zp->z_size) { zfs_range_unlock(rl); return (0); } - error = dmu_free_long_range(zfsvfs->z_os, zp->z_id, end, -1); + error = dmu_free_long_range(zfsvfs->z_os, zp->z_id, end, + DMU_OBJECT_END); if (error) { zfs_range_unlock(rl); return (error); } tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); dmu_tx_mark_netfree(tx); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); zfs_range_unlock(rl); return (error); } zp->z_size = end; SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_SIZE(zfsvfs), NULL, &zp->z_size, sizeof (zp->z_size)); if (end == 0) { zp->z_pflags &= ~ZFS_SPARSE; SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, 8); } VERIFY(sa_bulk_update(zp->z_sa_hdl, bulk, count, tx) == 0); dmu_tx_commit(tx); /* * Clear any mapped pages in the truncated region. This has to * happen outside of the transaction to avoid the possibility of * a deadlock with someone trying to push a page that we are * about to invalidate. */ vnode_pager_setsize(vp, end); zfs_range_unlock(rl); return (0); } /* * Free space in a file * * IN: zp - znode of file to free data in. * off - start of range * len - end of range (0 => EOF) * flag - current file open mode flags. * log - TRUE if this action should be logged * * RETURN: 0 on success, error code on failure */ int zfs_freesp(znode_t *zp, uint64_t off, uint64_t len, int flag, boolean_t log) { vnode_t *vp = ZTOV(zp); dmu_tx_t *tx; zfsvfs_t *zfsvfs = zp->z_zfsvfs; zilog_t *zilog = zfsvfs->z_log; uint64_t mode; uint64_t mtime[2], ctime[2]; sa_bulk_attr_t bulk[3]; int count = 0; int error; if ((error = sa_lookup(zp->z_sa_hdl, SA_ZPL_MODE(zfsvfs), &mode, sizeof (mode))) != 0) return (error); if (off > zp->z_size) { error = zfs_extend(zp, off+len); if (error == 0 && log) goto log; else return (error); } /* * Check for any locks in the region to be freed. */ if (MANDLOCK(vp, (mode_t)mode)) { uint64_t length = (len ? len : zp->z_size - off); if (error = chklock(vp, FWRITE, off, length, flag, NULL)) return (error); } if (len == 0) { error = zfs_trunc(zp, off); } else { if ((error = zfs_free_range(zp, off, len)) == 0 && off + len > zp->z_size) error = zfs_extend(zp, off+len); } if (error || !log) return (error); log: tx = dmu_tx_create(zfsvfs->z_os); dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); zfs_sa_upgrade_txholds(tx, zp); error = dmu_tx_assign(tx, TXG_WAIT); if (error) { dmu_tx_abort(tx); return (error); } SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_MTIME(zfsvfs), NULL, mtime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_CTIME(zfsvfs), NULL, ctime, 16); SA_ADD_BULK_ATTR(bulk, count, SA_ZPL_FLAGS(zfsvfs), NULL, &zp->z_pflags, 8); zfs_tstamp_update_setup(zp, CONTENT_MODIFIED, mtime, ctime, B_TRUE); error = sa_bulk_update(zp->z_sa_hdl, bulk, count, tx); ASSERT(error == 0); zfs_log_truncate(zilog, tx, TX_TRUNCATE, zp, off, len); dmu_tx_commit(tx); return (0); } void zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx) { uint64_t moid, obj, sa_obj, version; uint64_t sense = ZFS_CASE_SENSITIVE; uint64_t norm = 0; nvpair_t *elem; int error; int i; znode_t *rootzp = NULL; zfsvfs_t *zfsvfs; vattr_t vattr; znode_t *zp; zfs_acl_ids_t acl_ids; /* * First attempt to create master node. */ /* * In an empty objset, there are no blocks to read and thus * there can be no i/o errors (which we assert below). */ moid = MASTER_NODE_OBJ; error = zap_create_claim(os, moid, DMU_OT_MASTER_NODE, DMU_OT_NONE, 0, tx); ASSERT(error == 0); /* * Set starting attributes. */ version = zfs_zpl_version_map(spa_version(dmu_objset_spa(os))); elem = NULL; while ((elem = nvlist_next_nvpair(zplprops, elem)) != NULL) { /* For the moment we expect all zpl props to be uint64_ts */ uint64_t val; char *name; ASSERT(nvpair_type(elem) == DATA_TYPE_UINT64); VERIFY(nvpair_value_uint64(elem, &val) == 0); name = nvpair_name(elem); if (strcmp(name, zfs_prop_to_name(ZFS_PROP_VERSION)) == 0) { if (val < version) version = val; } else { error = zap_update(os, moid, name, 8, 1, &val, tx); } ASSERT(error == 0); if (strcmp(name, zfs_prop_to_name(ZFS_PROP_NORMALIZE)) == 0) norm = val; else if (strcmp(name, zfs_prop_to_name(ZFS_PROP_CASE)) == 0) sense = val; } ASSERT(version != 0); error = zap_update(os, moid, ZPL_VERSION_STR, 8, 1, &version, tx); /* * Create zap object used for SA attribute registration */ if (version >= ZPL_VERSION_SA) { sa_obj = zap_create(os, DMU_OT_SA_MASTER_NODE, DMU_OT_NONE, 0, tx); error = zap_add(os, moid, ZFS_SA_ATTRS, 8, 1, &sa_obj, tx); ASSERT(error == 0); } else { sa_obj = 0; } /* * Create a delete queue. */ obj = zap_create(os, DMU_OT_UNLINKED_SET, DMU_OT_NONE, 0, tx); error = zap_add(os, moid, ZFS_UNLINKED_SET, 8, 1, &obj, tx); ASSERT(error == 0); /* * Create root znode. Create minimal znode/vnode/zfsvfs * to allow zfs_mknode to work. */ VATTR_NULL(&vattr); vattr.va_mask = AT_MODE|AT_UID|AT_GID|AT_TYPE; vattr.va_type = VDIR; vattr.va_mode = S_IFDIR|0755; vattr.va_uid = crgetuid(cr); vattr.va_gid = crgetgid(cr); zfsvfs = kmem_zalloc(sizeof (zfsvfs_t), KM_SLEEP); rootzp = kmem_cache_alloc(znode_cache, KM_SLEEP); ASSERT(!POINTER_IS_VALID(rootzp->z_zfsvfs)); rootzp->z_moved = 0; rootzp->z_unlinked = 0; rootzp->z_atime_dirty = 0; rootzp->z_is_sa = USE_SA(version, os); zfsvfs->z_os = os; zfsvfs->z_parent = zfsvfs; zfsvfs->z_version = version; zfsvfs->z_use_fuids = USE_FUIDS(version, os); zfsvfs->z_use_sa = USE_SA(version, os); zfsvfs->z_norm = norm; error = sa_setup(os, sa_obj, zfs_attr_table, ZPL_END, &zfsvfs->z_attr_table); ASSERT(error == 0); /* * Fold case on file systems that are always or sometimes case * insensitive. */ if (sense == ZFS_CASE_INSENSITIVE || sense == ZFS_CASE_MIXED) zfsvfs->z_norm |= U8_TEXTPREP_TOUPPER; mutex_init(&zfsvfs->z_znodes_lock, NULL, MUTEX_DEFAULT, NULL); list_create(&zfsvfs->z_all_znodes, sizeof (znode_t), offsetof(znode_t, z_link_node)); for (i = 0; i != ZFS_OBJ_MTX_SZ; i++) mutex_init(&zfsvfs->z_hold_mtx[i], NULL, MUTEX_DEFAULT, NULL); rootzp->z_zfsvfs = zfsvfs; VERIFY(0 == zfs_acl_ids_create(rootzp, IS_ROOT_NODE, &vattr, cr, NULL, &acl_ids)); zfs_mknode(rootzp, &vattr, tx, cr, IS_ROOT_NODE, &zp, &acl_ids); ASSERT3P(zp, ==, rootzp); error = zap_add(os, moid, ZFS_ROOT_OBJ, 8, 1, &rootzp->z_id, tx); ASSERT(error == 0); zfs_acl_ids_free(&acl_ids); POINTER_INVALIDATE(&rootzp->z_zfsvfs); sa_handle_destroy(rootzp->z_sa_hdl); kmem_cache_free(znode_cache, rootzp); /* * Create shares directory */ error = zfs_create_share_dir(zfsvfs, tx); ASSERT(error == 0); for (i = 0; i != ZFS_OBJ_MTX_SZ; i++) mutex_destroy(&zfsvfs->z_hold_mtx[i]); kmem_free(zfsvfs, sizeof (zfsvfs_t)); } #endif /* _KERNEL */ static int zfs_sa_setup(objset_t *osp, sa_attr_type_t **sa_table) { uint64_t sa_obj = 0; int error; error = zap_lookup(osp, MASTER_NODE_OBJ, ZFS_SA_ATTRS, 8, 1, &sa_obj); if (error != 0 && error != ENOENT) return (error); error = sa_setup(osp, sa_obj, zfs_attr_table, ZPL_END, sa_table); return (error); } static int zfs_grab_sa_handle(objset_t *osp, uint64_t obj, sa_handle_t **hdlp, dmu_buf_t **db, void *tag) { dmu_object_info_t doi; int error; if ((error = sa_buf_hold(osp, obj, tag, db)) != 0) return (error); dmu_object_info_from_db(*db, &doi); if ((doi.doi_bonus_type != DMU_OT_SA && doi.doi_bonus_type != DMU_OT_ZNODE) || doi.doi_bonus_type == DMU_OT_ZNODE && doi.doi_bonus_size < sizeof (znode_phys_t)) { sa_buf_rele(*db, tag); return (SET_ERROR(ENOTSUP)); } error = sa_handle_get(osp, obj, NULL, SA_HDL_PRIVATE, hdlp); if (error != 0) { sa_buf_rele(*db, tag); return (error); } return (0); } void zfs_release_sa_handle(sa_handle_t *hdl, dmu_buf_t *db, void *tag) { sa_handle_destroy(hdl); sa_buf_rele(db, tag); } /* * Given an object number, return its parent object number and whether * or not the object is an extended attribute directory. */ static int zfs_obj_to_pobj(objset_t *osp, sa_handle_t *hdl, sa_attr_type_t *sa_table, uint64_t *pobjp, int *is_xattrdir) { uint64_t parent; uint64_t pflags; uint64_t mode; uint64_t parent_mode; sa_bulk_attr_t bulk[3]; sa_handle_t *sa_hdl; dmu_buf_t *sa_db; int count = 0; int error; SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_PARENT], NULL, &parent, sizeof (parent)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_FLAGS], NULL, &pflags, sizeof (pflags)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_MODE], NULL, &mode, sizeof (mode)); if ((error = sa_bulk_lookup(hdl, bulk, count)) != 0) return (error); /* * When a link is removed its parent pointer is not changed and will * be invalid. There are two cases where a link is removed but the * file stays around, when it goes to the delete queue and when there * are additional links. */ error = zfs_grab_sa_handle(osp, parent, &sa_hdl, &sa_db, FTAG); if (error != 0) return (error); error = sa_lookup(sa_hdl, ZPL_MODE, &parent_mode, sizeof (parent_mode)); zfs_release_sa_handle(sa_hdl, sa_db, FTAG); if (error != 0) return (error); *is_xattrdir = ((pflags & ZFS_XATTR) != 0) && S_ISDIR(mode); /* * Extended attributes can be applied to files, directories, etc. * Otherwise the parent must be a directory. */ if (!*is_xattrdir && !S_ISDIR(parent_mode)) return (SET_ERROR(EINVAL)); *pobjp = parent; return (0); } /* * Given an object number, return some zpl level statistics */ static int zfs_obj_to_stats_impl(sa_handle_t *hdl, sa_attr_type_t *sa_table, zfs_stat_t *sb) { sa_bulk_attr_t bulk[4]; int count = 0; SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_MODE], NULL, &sb->zs_mode, sizeof (sb->zs_mode)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_GEN], NULL, &sb->zs_gen, sizeof (sb->zs_gen)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_LINKS], NULL, &sb->zs_links, sizeof (sb->zs_links)); SA_ADD_BULK_ATTR(bulk, count, sa_table[ZPL_CTIME], NULL, &sb->zs_ctime, sizeof (sb->zs_ctime)); return (sa_bulk_lookup(hdl, bulk, count)); } static int zfs_obj_to_path_impl(objset_t *osp, uint64_t obj, sa_handle_t *hdl, sa_attr_type_t *sa_table, char *buf, int len) { sa_handle_t *sa_hdl; sa_handle_t *prevhdl = NULL; dmu_buf_t *prevdb = NULL; dmu_buf_t *sa_db = NULL; char *path = buf + len - 1; int error; *path = '\0'; sa_hdl = hdl; uint64_t deleteq_obj; VERIFY0(zap_lookup(osp, MASTER_NODE_OBJ, ZFS_UNLINKED_SET, sizeof (uint64_t), 1, &deleteq_obj)); error = zap_lookup_int(osp, deleteq_obj, obj); if (error == 0) { return (ESTALE); } else if (error != ENOENT) { return (error); } error = 0; for (;;) { uint64_t pobj; char component[MAXNAMELEN + 2]; size_t complen; int is_xattrdir; if (prevdb) zfs_release_sa_handle(prevhdl, prevdb, FTAG); if ((error = zfs_obj_to_pobj(osp, sa_hdl, sa_table, &pobj, &is_xattrdir)) != 0) break; if (pobj == obj) { if (path[0] != '/') *--path = '/'; break; } component[0] = '/'; if (is_xattrdir) { (void) sprintf(component + 1, ""); } else { error = zap_value_search(osp, pobj, obj, ZFS_DIRENT_OBJ(-1ULL), component + 1); if (error != 0) break; } complen = strlen(component); path -= complen; ASSERT(path >= buf); bcopy(component, path, complen); obj = pobj; if (sa_hdl != hdl) { prevhdl = sa_hdl; prevdb = sa_db; } error = zfs_grab_sa_handle(osp, obj, &sa_hdl, &sa_db, FTAG); if (error != 0) { sa_hdl = prevhdl; sa_db = prevdb; break; } } if (sa_hdl != NULL && sa_hdl != hdl) { ASSERT(sa_db != NULL); zfs_release_sa_handle(sa_hdl, sa_db, FTAG); } if (error == 0) (void) memmove(buf, path, buf + len - path); return (error); } int zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len) { sa_attr_type_t *sa_table; sa_handle_t *hdl; dmu_buf_t *db; int error; error = zfs_sa_setup(osp, &sa_table); if (error != 0) return (error); error = zfs_grab_sa_handle(osp, obj, &hdl, &db, FTAG); if (error != 0) return (error); error = zfs_obj_to_path_impl(osp, obj, hdl, sa_table, buf, len); zfs_release_sa_handle(hdl, db, FTAG); return (error); } int zfs_obj_to_stats(objset_t *osp, uint64_t obj, zfs_stat_t *sb, char *buf, int len) { char *path = buf + len - 1; sa_attr_type_t *sa_table; sa_handle_t *hdl; dmu_buf_t *db; int error; *path = '\0'; error = zfs_sa_setup(osp, &sa_table); if (error != 0) return (error); error = zfs_grab_sa_handle(osp, obj, &hdl, &db, FTAG); if (error != 0) return (error); error = zfs_obj_to_stats_impl(hdl, sa_table, sb); if (error != 0) { zfs_release_sa_handle(hdl, db, FTAG); return (error); } error = zfs_obj_to_path_impl(osp, obj, hdl, sa_table, buf, len); zfs_release_sa_handle(hdl, db, FTAG); return (error); } #ifdef _KERNEL int zfs_znode_parent_and_name(znode_t *zp, znode_t **dzpp, char *buf) { zfsvfs_t *zfsvfs = zp->z_zfsvfs; uint64_t parent; int is_xattrdir; int err; /* Extended attributes should not be visible as regular files. */ if ((zp->z_pflags & ZFS_XATTR) != 0) return (SET_ERROR(EINVAL)); err = zfs_obj_to_pobj(zfsvfs->z_os, zp->z_sa_hdl, zfsvfs->z_attr_table, &parent, &is_xattrdir); if (err != 0) return (err); ASSERT0(is_xattrdir); /* No name as this is a root object. */ if (parent == zp->z_id) return (SET_ERROR(EINVAL)); err = zap_value_search(zfsvfs->z_os, parent, zp->z_id, ZFS_DIRENT_OBJ(-1ULL), buf); if (err != 0) return (err); err = zfs_zget(zfsvfs, parent, dzpp); return (err); } #endif /* _KERNEL */ Index: stable/11 =================================================================== --- stable/11 (revision 339135) +++ stable/11 (revision 339136) Property changes on: stable/11 ___________________________________________________________________ Modified: svn:mergeinfo ## -0,0 +0,1 ## Merged /head:r337202