static uint64_t zpios_dmu_object_create(run_args_t *run_args, objset_t *os) { struct dmu_tx *tx; uint64_t obj = 0ULL; int rc; tx = dmu_tx_create(os); dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, OBJ_SIZE); rc = dmu_tx_assign(tx, TXG_WAIT); if (rc) { zpios_print(run_args->file, "dmu_tx_assign() failed: %d\n", rc); dmu_tx_abort(tx); return obj; } obj = dmu_object_alloc(os, DMU_OT_UINT64_OTHER, 0, DMU_OT_NONE, 0, tx); rc = dmu_object_set_blocksize(os, obj, 128ULL << 10, 0, tx); if (rc) { zpios_print(run_args->file, "dmu_object_set_blocksize() failed: %d\n", rc); dmu_tx_abort(tx); return obj; } dmu_tx_commit(tx); return obj; }
/* * Load the fuid table(s) into memory. */ static void zfs_fuid_init(zfsvfs_t *zfsvfs, dmu_tx_t *tx) { int error = 0; rw_enter(&zfsvfs->z_fuid_lock, RW_WRITER); if (zfsvfs->z_fuid_loaded) { rw_exit(&zfsvfs->z_fuid_lock); return; } if (zfsvfs->z_fuid_obj == 0) { /* first make sure we need to allocate object */ error = zap_lookup(zfsvfs->z_os, MASTER_NODE_OBJ, ZFS_FUID_TABLES, 8, 1, &zfsvfs->z_fuid_obj); if (error == ENOENT && tx != NULL) { zfsvfs->z_fuid_obj = dmu_object_alloc(zfsvfs->z_os, DMU_OT_FUID, 1 << 14, DMU_OT_FUID_SIZE, sizeof (uint64_t), tx); VERIFY(zap_add(zfsvfs->z_os, MASTER_NODE_OBJ, ZFS_FUID_TABLES, sizeof (uint64_t), 1, &zfsvfs->z_fuid_obj, tx) == 0); } } zfsvfs->z_fuid_size = zfs_fuid_table_load(zfsvfs->z_os, zfsvfs->z_fuid_obj, &zfsvfs->z_fuid_idx, &zfsvfs->z_fuid_domain); zfsvfs->z_fuid_loaded = B_TRUE; rw_exit(&zfsvfs->z_fuid_lock); }
void spa_history_create_obj(spa_t *spa, dmu_tx_t *tx) { dmu_buf_t *dbp; spa_history_phys_t *shpp; objset_t *mos = spa->spa_meta_objset; ASSERT(spa->spa_history == 0); spa->spa_history = dmu_object_alloc(mos, DMU_OT_SPA_HISTORY, SPA_MAXBLOCKSIZE, DMU_OT_SPA_HISTORY_OFFSETS, sizeof (spa_history_phys_t), tx); VERIFY(zap_add(mos, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_HISTORY, sizeof (uint64_t), 1, &spa->spa_history, tx) == 0); VERIFY(0 == dmu_bonus_hold(mos, spa->spa_history, FTAG, &dbp)); ASSERT(dbp->db_size >= sizeof (spa_history_phys_t)); shpp = dbp->db_data; dmu_buf_will_dirty(dbp, tx); /* * Figure out maximum size of history log. We set it at * 0.1% of pool size, with a max of 1G and min of 128KB. */ shpp->sh_phys_max_off = metaslab_class_get_dspace(spa_normal_class(spa)) / 1000; shpp->sh_phys_max_off = MIN(shpp->sh_phys_max_off, 1<<30); shpp->sh_phys_max_off = MAX(shpp->sh_phys_max_off, 128<<10); dmu_buf_rele(dbp, FTAG); }
/* * The transaction passed to this routine must have * dmu_tx_hold_bonus(tx, DMU_NEW_OBJECT) called and then assigned * to a transaction group. */ int __osd_object_create(const struct lu_env *env, udmu_objset_t *uos, dmu_buf_t **dbp, dmu_tx_t *tx, struct lu_attr *la, void *tag) { uint64_t oid; int rc; LASSERT(tag); spin_lock(&uos->lock); uos->objects++; spin_unlock(&uos->lock); /* Assert that the transaction has been assigned to a transaction group. */ LASSERT(tx->tx_txg != 0); /* Create a new DMU object. */ oid = dmu_object_alloc(uos->os, DMU_OT_PLAIN_FILE_CONTENTS, 0, DMU_OT_SA, DN_MAX_BONUSLEN, tx); rc = -sa_buf_hold(uos->os, oid, tag, dbp); if (rc) return rc; LASSERT(la->la_valid & LA_MODE); la->la_size = 0; la->la_nlink = 1; return __osd_attr_init(env, uos, oid, tx, la); }
uint64_t dsl_dir_create_sync(dsl_pool_t *dp, dsl_dir_t *pds, const char *name, dmu_tx_t *tx) { objset_t *mos = dp->dp_meta_objset; uint64_t ddobj; dsl_dir_phys_t *dsphys; dmu_buf_t *dbuf; ddobj = dmu_object_alloc(mos, DMU_OT_DSL_DIR, 0, DMU_OT_DSL_DIR, sizeof (dsl_dir_phys_t), tx); if (pds) { VERIFY(0 == zap_add(mos, pds->dd_phys->dd_child_dir_zapobj, name, sizeof (uint64_t), 1, &ddobj, tx)); } else { /* it's the root dir */ VERIFY(0 == zap_add(mos, DMU_POOL_DIRECTORY_OBJECT, DMU_POOL_ROOT_DATASET, sizeof (uint64_t), 1, &ddobj, tx)); } VERIFY(0 == dmu_bonus_hold(mos, ddobj, FTAG, &dbuf)); dmu_buf_will_dirty(dbuf, tx); dsphys = dbuf->db_data; dsphys->dd_creation_time = gethrestime_sec(); if (pds) dsphys->dd_parent_obj = pds->dd_object; dsphys->dd_props_zapobj = zap_create(mos, DMU_OT_DSL_PROPS, DMU_OT_NONE, 0, tx); dsphys->dd_child_dir_zapobj = zap_create(mos, DMU_OT_DSL_DIR_CHILD_MAP, DMU_OT_NONE, 0, tx); dmu_buf_rele(dbuf, FTAG); return (ddobj); }
uint64_t vdev_indirect_births_alloc(objset_t *os, dmu_tx_t *tx) { ASSERT(dmu_tx_is_syncing(tx)); return (dmu_object_alloc(os, DMU_OTN_UINT64_METADATA, SPA_OLD_MAXBLOCKSIZE, DMU_OTN_UINT64_METADATA, sizeof (vdev_indirect_birth_phys_t), tx)); }
uint64_t bplist_create(objset_t *mos, int blocksize, dmu_tx_t *tx) { int size; size = spa_version(dmu_objset_spa(mos)) < SPA_VERSION_BPLIST_ACCOUNT ? BPLIST_SIZE_V0 : sizeof (bplist_phys_t); return (dmu_object_alloc(mos, DMU_OT_BPLIST, blocksize, DMU_OT_BPLIST_HDR, size, tx)); }
static uint64_t zpios_dmu_object_create(run_args_t *run_args, objset_t *os) { struct dmu_tx *tx; uint64_t obj = 0ULL; uint64_t blksize = run_args->block_size; int rc; if (blksize < SPA_MINBLOCKSIZE || blksize > spa_maxblocksize(dmu_objset_spa(os)) || !ISP2(blksize)) { zpios_print(run_args->file, "invalid block size for pool: %d\n", (int)blksize); return (obj); } tx = dmu_tx_create(os); dmu_tx_hold_write(tx, DMU_NEW_OBJECT, 0, OBJ_SIZE); rc = dmu_tx_assign(tx, TXG_WAIT); if (rc) { zpios_print(run_args->file, "dmu_tx_assign() failed: %d\n", rc); dmu_tx_abort(tx); return (obj); } obj = dmu_object_alloc(os, DMU_OT_UINT64_OTHER, 0, DMU_OT_NONE, 0, tx); rc = dmu_object_set_blocksize(os, obj, blksize, 0, tx); if (rc) { zpios_print(run_args->file, "dmu_object_set_blocksize to %d failed: %d\n", (int)blksize, rc); dmu_tx_abort(tx); return (obj); } dmu_tx_commit(tx); return (obj); }
/* * The transaction passed to this routine must have * dmu_tx_hold_bonus(tx, DMU_NEW_OBJECT) called and then assigned * to a transaction group. */ int __osd_object_create(const struct lu_env *env, struct osd_object *obj, dmu_buf_t **dbp, dmu_tx_t *tx, struct lu_attr *la, uint64_t parent) { uint64_t oid; int rc; struct osd_device *osd = osd_obj2dev(obj); const struct lu_fid *fid = lu_object_fid(&obj->oo_dt.do_lu); dmu_object_type_t type = DMU_OT_PLAIN_FILE_CONTENTS; /* Assert that the transaction has been assigned to a transaction group. */ LASSERT(tx->tx_txg != 0); /* Use DMU_OTN_UINT8_METADATA for local objects so their data blocks * would get an additional ditto copy */ if (unlikely(S_ISREG(la->la_mode) && fid_seq_is_local_file(fid_seq(fid)))) type = DMU_OTN_UINT8_METADATA; /* Create a new DMU object. */ oid = dmu_object_alloc(osd->od_os, type, 0, DMU_OT_SA, DN_MAX_BONUSLEN, tx); rc = -sa_buf_hold(osd->od_os, oid, osd_obj_tag, dbp); LASSERTF(rc == 0, "sa_buf_hold "LPU64" failed: %d\n", oid, rc); LASSERT(la->la_valid & LA_MODE); la->la_size = 0; la->la_nlink = 1; rc = __osd_attr_init(env, osd, oid, tx, la, parent); if (rc != 0) { sa_buf_rele(*dbp, osd_obj_tag); *dbp = NULL; dmu_object_free(osd->od_os, oid, tx); return rc; } return 0; }
/* * sync out AVL trees to persistent storage. */ void zfs_fuid_sync(zfsvfs_t *zfsvfs, dmu_tx_t *tx) { #ifdef HAVE_ZPL nvlist_t *nvp; nvlist_t **fuids; size_t nvsize = 0; char *packed; dmu_buf_t *db; fuid_domain_t *domnode; int numnodes; int i; if (!zfsvfs->z_fuid_dirty) { return; } rw_enter(&zfsvfs->z_fuid_lock, RW_WRITER); /* * First see if table needs to be created? */ if (zfsvfs->z_fuid_obj == 0) { zfsvfs->z_fuid_obj = dmu_object_alloc(zfsvfs->z_os, DMU_OT_FUID, 1 << 14, DMU_OT_FUID_SIZE, sizeof (uint64_t), tx); VERIFY(zap_add(zfsvfs->z_os, MASTER_NODE_OBJ, ZFS_FUID_TABLES, sizeof (uint64_t), 1, &zfsvfs->z_fuid_obj, tx) == 0); } VERIFY(nvlist_alloc(&nvp, NV_UNIQUE_NAME, KM_SLEEP) == 0); numnodes = avl_numnodes(&zfsvfs->z_fuid_idx); fuids = kmem_alloc(numnodes * sizeof (void *), KM_SLEEP); for (i = 0, domnode = avl_first(&zfsvfs->z_fuid_domain); domnode; i++, domnode = AVL_NEXT(&zfsvfs->z_fuid_domain, domnode)) { VERIFY(nvlist_alloc(&fuids[i], NV_UNIQUE_NAME, KM_SLEEP) == 0); VERIFY(nvlist_add_uint64(fuids[i], FUID_IDX, domnode->f_idx) == 0); VERIFY(nvlist_add_uint64(fuids[i], FUID_OFFSET, 0) == 0); VERIFY(nvlist_add_string(fuids[i], FUID_DOMAIN, domnode->f_ksid->kd_name) == 0); } VERIFY(nvlist_add_nvlist_array(nvp, FUID_NVP_ARRAY, fuids, numnodes) == 0); for (i = 0; i != numnodes; i++) nvlist_free(fuids[i]); kmem_free(fuids, numnodes * sizeof (void *)); VERIFY(nvlist_size(nvp, &nvsize, NV_ENCODE_XDR) == 0); packed = kmem_alloc(nvsize, KM_SLEEP); VERIFY(nvlist_pack(nvp, &packed, &nvsize, NV_ENCODE_XDR, KM_SLEEP) == 0); nvlist_free(nvp); zfsvfs->z_fuid_size = nvsize; dmu_write(zfsvfs->z_os, zfsvfs->z_fuid_obj, 0, zfsvfs->z_fuid_size, packed, tx); kmem_free(packed, zfsvfs->z_fuid_size); VERIFY(0 == dmu_bonus_hold(zfsvfs->z_os, zfsvfs->z_fuid_obj, FTAG, &db)); dmu_buf_will_dirty(db, tx); *(uint64_t *)db->db_data = zfsvfs->z_fuid_size; dmu_buf_rele(db, FTAG); zfsvfs->z_fuid_dirty = B_FALSE; rw_exit(&zfsvfs->z_fuid_lock); #endif /* HAVE_ZPL */ }
/* * 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; zfs_sb_t *zsb = ZTOZSB(dzp); dmu_buf_t *db; timestruc_t now; uint64_t gen, obj; int bonuslen; sa_handle_t *sa_hdl; dmu_object_type_t obj_type; sa_bulk_attr_t *sa_attrs; int cnt = 0; zfs_acl_locator_cb_t locate = { 0 }; if (zsb->z_replay) { obj = vap->va_nodeid; now = vap->va_ctime; /* see zfs_replay_create() */ gen = vap->va_nblocks; /* ditto */ } else { obj = 0; gethrestime(&now); gen = dmu_tx_get_txg(tx); } obj_type = zsb->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 (S_ISDIR(vap->va_mode)) { if (zsb->z_replay) { VERIFY0(zap_create_claim_norm(zsb->z_os, obj, zsb->z_norm, DMU_OT_DIRECTORY_CONTENTS, obj_type, bonuslen, tx)); } else { obj = zap_create_norm(zsb->z_os, zsb->z_norm, DMU_OT_DIRECTORY_CONTENTS, obj_type, bonuslen, tx); } } else { if (zsb->z_replay) { VERIFY0(dmu_object_claim(zsb->z_os, obj, DMU_OT_PLAIN_FILE_CONTENTS, 0, obj_type, bonuslen, tx)); } else { obj = dmu_object_alloc(zsb->z_os, DMU_OT_PLAIN_FILE_CONTENTS, 0, obj_type, bonuslen, tx); } } ZFS_OBJ_HOLD_ENTER(zsb, obj); VERIFY(0 == sa_buf_hold(zsb->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 (zsb->z_use_fuids) pflags = ZFS_ARCHIVE | ZFS_AV_MODIFIED; else pflags = 0; if (S_ISDIR(vap->va_mode)) { size = 2; /* contents ("." and "..") */ links = (flag & (IS_ROOT_NODE | IS_XATTR)) ? 2 : 1; } else { size = links = 0; } if (S_ISBLK(vap->va_mode) || S_ISCHR(vap->va_mode)) rdev = 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 & ATTR_ATIME) { ZFS_TIME_ENCODE(&vap->va_atime, atime); } else { ZFS_TIME_ENCODE(&now, atime); } if (vap->va_mask & ATTR_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(zsb->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 */ sa_attrs = kmem_alloc(sizeof (sa_bulk_attr_t) * ZPL_END, KM_SLEEP); if (obj_type == DMU_OT_ZNODE) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ATIME(zsb), NULL, &atime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MTIME(zsb), NULL, &mtime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CTIME(zsb), NULL, &ctime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CRTIME(zsb), NULL, &crtime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GEN(zsb), NULL, &gen, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MODE(zsb), NULL, &mode, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_SIZE(zsb), NULL, &size, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PARENT(zsb), NULL, &parent, 8); } else { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MODE(zsb), NULL, &mode, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_SIZE(zsb), NULL, &size, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GEN(zsb), NULL, &gen, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_UID(zsb), NULL, &acl_ids->z_fuid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GID(zsb), NULL, &acl_ids->z_fgid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PARENT(zsb), NULL, &parent, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_FLAGS(zsb), NULL, &pflags, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ATIME(zsb), NULL, &atime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_MTIME(zsb), NULL, &mtime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CTIME(zsb), NULL, &ctime, 16); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_CRTIME(zsb), NULL, &crtime, 16); } SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_LINKS(zsb), NULL, &links, 8); if (obj_type == DMU_OT_ZNODE) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_XATTR(zsb), NULL, &empty_xattr, 8); } if (obj_type == DMU_OT_ZNODE || (S_ISBLK(vap->va_mode) || S_ISCHR(vap->va_mode))) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_RDEV(zsb), NULL, &rdev, 8); } if (obj_type == DMU_OT_ZNODE) { SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_FLAGS(zsb), NULL, &pflags, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_UID(zsb), NULL, &acl_ids->z_fuid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_GID(zsb), NULL, &acl_ids->z_fgid, 8); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_PAD(zsb), NULL, pad, sizeof (uint64_t) * 4); SA_ADD_BULK_ATTR(sa_attrs, cnt, SA_ZPL_ZNODE_ACL(zsb), 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(zsb), 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(zsb), 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(zsb, db, 0, obj_type, obj, sa_hdl, ZTOI(dzp)); VERIFY(*zpp != NULL); VERIFY(dzp != 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 (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)); } kmem_free(sa_attrs, sizeof (sa_bulk_attr_t) * ZPL_END); ZFS_OBJ_HOLD_EXIT(zsb, obj); }