X-Git-Url: https://git.camperquake.de/gitweb.cgi?a=blobdiff_plain;f=module%2Fzfs%2Fzfs_znode.c;h=e1e4e9e03ac26666820e3fcf467ef00693d005ed;hb=572e285762521df27fe5b026f409ba1a21abb7ac;hp=74983cdc579181b3304056a4ed551b5a6b56f038;hpb=d164b2093561a9771db07346e6fffc9ca19427a2;p=zfs.git diff --git a/module/zfs/zfs_znode.c b/module/zfs/zfs_znode.c index 74983cd..e1e4e9e 100644 --- a/module/zfs/zfs_znode.c +++ b/module/zfs/zfs_znode.c @@ -19,8 +19,7 @@ * CDDL HEADER END */ /* - * Copyright 2009 Sun Microsystems, Inc. All rights reserved. - * Use is subject to license terms. + * Copyright (c) 2005, 2010, Oracle and/or its affiliates. All rights reserved. */ /* Portions Copyright 2007 Jeremy Teo */ @@ -52,6 +51,7 @@ #include #include #include +#include #include #include #endif /* _KERNEL */ @@ -61,8 +61,12 @@ #include #include #include +#include +#include +#include #include "zfs_prop.h" +#include "zfs_comutil.h" /* * Define ZNODE_STATS to turn on statistic gathering. By default, it is only @@ -78,15 +82,18 @@ #define ZNODE_STAT_ADD(stat) /* nothing */ #endif /* ZNODE_STATS */ -#define POINTER_IS_VALID(p) (!((uintptr_t)(p) & 0x3)) -#define POINTER_INVALIDATE(pp) (*(pp) = (void *)((uintptr_t)(*(pp)) | 0x1)) - /* * 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*/ @@ -125,8 +132,9 @@ zfs_znode_cache_constructor(void *buf, void *arg, int kmflags) avl_create(&zp->z_range_avl, zfs_range_compare, sizeof (rl_t), offsetof(rl_t, r_node)); - zp->z_dbuf = NULL; zp->z_dirlocks = NULL; + zp->z_acl_cached = NULL; + zp->z_moved = 0; return (0); } @@ -147,15 +155,16 @@ zfs_znode_cache_destructor(void *buf, void *arg) avl_destroy(&zp->z_range_avl); mutex_destroy(&zp->z_range_lock); - ASSERT(zp->z_dbuf == NULL); ASSERT(zp->z_dirlocks == NULL); + 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_recheck_invalid; + uint64_t zms_zfsvfs_recheck2; uint64_t zms_obj_held; uint64_t zms_vnode_locked; uint64_t zms_not_only_dnlc; @@ -186,15 +195,28 @@ zfs_znode_move_impl(znode_t *ozp, znode_t *nzp) nzp->z_blksz = ozp->z_blksz; nzp->z_seq = ozp->z_seq; nzp->z_mapcnt = ozp->z_mapcnt; - nzp->z_last_itx = ozp->z_last_itx; nzp->z_gen = ozp->z_gen; nzp->z_sync_cnt = ozp->z_sync_cnt; - nzp->z_phys = ozp->z_phys; - nzp->z_dbuf = ozp->z_dbuf; + 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; - /* Update back pointers. */ - (void) dmu_buf_update_user(nzp->z_dbuf, ozp, nzp, &nzp->z_phys, - znode_evict_error); + /* + * 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 @@ -202,19 +224,14 @@ zfs_znode_move_impl(znode_t *ozp, znode_t *nzp) * ensure that zfs_znode_move() recognizes the znode as invalid in any * subsequent callback. */ - ozp->z_dbuf = NULL; + ozp->z_sa_hdl = NULL; POINTER_INVALIDATE(&ozp->z_zfsvfs); -} -/* - * Wrapper function for ZFS_ENTER that returns 0 if successful and otherwise - * returns a non-zero error code. - */ -static int -zfs_enter(zfsvfs_t *zfsvfs) -{ - ZFS_ENTER(zfsvfs); - return (0); + /* + * Mark the znode. + */ + nzp->z_moved = 1; + ozp->z_moved = (uint8_t)-1; } /*ARGSUSED*/ @@ -240,12 +257,32 @@ zfs_znode_move(void *buf, void *newbuf, size_t size, void *arg) } /* - * Ensure that the filesystem is not unmounted during the move. + * 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. */ - if (zfs_enter(zfsvfs) != 0) { /* ZFS_ENTER */ + 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(). + */ + rrw_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); /* @@ -255,7 +292,7 @@ zfs_znode_move(void *buf, void *newbuf, size_t size, void *arg) if (zfsvfs != ozp->z_zfsvfs) { mutex_exit(&zfsvfs->z_znodes_lock); ZFS_EXIT(zfsvfs); - ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_recheck_invalid); + ZNODE_STAT_ADD(znode_move_stats.zms_zfsvfs_recheck2); return (KMEM_CBRC_DONT_KNOW); } @@ -311,6 +348,7 @@ 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, @@ -332,6 +370,7 @@ zfs_znode_fini(void) if (znode_cache) kmem_cache_destroy(znode_cache); znode_cache = NULL; + rw_destroy(&zfsvfs_lock); } struct vnodeops *zfs_dvnodeops; @@ -339,6 +378,7 @@ 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() @@ -363,12 +403,15 @@ zfs_remove_op_tables() 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[]; @@ -376,6 +419,7 @@ 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() @@ -412,103 +456,59 @@ zfs_create_op_tables() 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); } -/* - * zfs_init_fs - Initialize the zfsvfs struct and the file system - * incore "master" object. Verify version compatibility. - */ int -zfs_init_fs(zfsvfs_t *zfsvfs, znode_t **zpp) +zfs_create_share_dir(zfsvfs_t *zfsvfs, dmu_tx_t *tx) { - extern int zfsfstype; - - objset_t *os = zfsvfs->z_os; - int i, error; - uint64_t fsid_guid; - uint64_t zval; - - *zpp = NULL; - - error = zfs_get_zplprop(os, ZFS_PROP_VERSION, &zfsvfs->z_version); - if (error) { - return (error); - } else if (zfsvfs->z_version > ZPL_VERSION) { - (void) printf("Mismatched versions: File system " - "is version %llu on-disk format, which is " - "incompatible with this software version %lld!", - (u_longlong_t)zfsvfs->z_version, ZPL_VERSION); - return (ENOTSUP); - } - - if ((error = zfs_get_zplprop(os, ZFS_PROP_NORMALIZE, &zval)) != 0) - return (error); - zfsvfs->z_norm = (int)zval; - if ((error = zfs_get_zplprop(os, ZFS_PROP_UTF8ONLY, &zval)) != 0) - return (error); - zfsvfs->z_utf8 = (zval != 0); - if ((error = zfs_get_zplprop(os, ZFS_PROP_CASE, &zval)) != 0) - return (error); - zfsvfs->z_case = (uint_t)zval; - /* - * Fold case on file systems that are always or sometimes case - * insensitive. - */ - if (zfsvfs->z_case == ZFS_CASE_INSENSITIVE || - zfsvfs->z_case == ZFS_CASE_MIXED) - zfsvfs->z_norm |= U8_TEXTPREP_TOUPPER; - - /* - * The fsid is 64 bits, composed of an 8-bit fs type, which - * separates our fsid from any other filesystem types, and a - * 56-bit objset unique ID. The objset unique ID is unique to - * all objsets open on this system, provided by unique_create(). - * The 8-bit fs type must be put in the low bits of fsid[1] - * because that's where other Solaris filesystems put it. - */ - fsid_guid = dmu_objset_fsid_guid(os); - ASSERT((fsid_guid & ~((1ULL<<56)-1)) == 0); - zfsvfs->z_vfs->vfs_fsid.val[0] = fsid_guid; - zfsvfs->z_vfs->vfs_fsid.val[1] = ((fsid_guid>>32) << 8) | - zfsfstype & 0xFF; - - error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_ROOT_OBJ, 8, 1, - &zfsvfs->z_root); - if (error) - return (error); - ASSERT(zfsvfs->z_root != 0); - - error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_UNLINKED_SET, 8, 1, - &zfsvfs->z_unlinkedobj); - if (error) - return (error); + zfs_acl_ids_t acl_ids; + vattr_t vattr; + znode_t *sharezp; + vnode_t *vp; + znode_t *zp; + int error; - /* - * Initialize zget mutex's - */ - for (i = 0; i != ZFS_OBJ_MTX_SZ; i++) - mutex_init(&zfsvfs->z_hold_mtx[i], NULL, MUTEX_DEFAULT, NULL); + 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; + + vp = ZTOV(sharezp); + vn_reinit(vp); + vp->v_type = VDIR; - error = zfs_zget(zfsvfs, zfsvfs->z_root, zpp); - if (error) { - /* - * On error, we destroy the mutexes here since it's not - * possible for the caller to determine if the mutexes were - * initialized properly. - */ - for (i = 0; i != ZFS_OBJ_MTX_SZ; i++) - mutex_destroy(&zfsvfs->z_hold_mtx[i]); - return (error); - } - ASSERT3U((*zpp)->z_id, ==, zfsvfs->z_root); - error = zap_lookup(os, MASTER_NODE_OBJ, ZFS_FUID_TABLES, 8, 1, - &zfsvfs->z_fuid_obj); - if (error == ENOENT) - error = 0; + 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); + ASSERT(!vn_in_dnlc(ZTOV(sharezp))); /* not valid to move */ + 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); + ZTOV(sharezp)->v_count = 0; + sa_handle_destroy(sharezp->z_sa_hdl); + kmem_cache_free(znode_cache, sharezp); - return (0); + return (error); } /* @@ -569,25 +569,25 @@ zfs_cmpldev(uint64_t dev) } static void -zfs_znode_dmu_init(zfsvfs_t *zfsvfs, znode_t *zp, dmu_buf_t *db) +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) { - znode_t *nzp; - ASSERT(!POINTER_IS_VALID(zp->z_zfsvfs) || (zfsvfs == zp->z_zfsvfs)); ASSERT(MUTEX_HELD(ZFS_OBJ_MUTEX(zfsvfs, zp->z_id))); mutex_enter(&zp->z_lock); - ASSERT(zp->z_dbuf == NULL); - zp->z_dbuf = db; - nzp = dmu_buf_set_user_ie(db, zp, &zp->z_phys, znode_evict_error); + 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); + } - /* - * there should be no - * concurrent zgets on this object. - */ - if (nzp != NULL) - panic("existing znode %p for dbuf %p", (void *)nzp, (void *)db); + zp->z_is_sa = (obj_type == DMU_OT_SA) ? B_TRUE : B_FALSE; /* * Slap on VROOT if we are the root znode @@ -602,14 +602,12 @@ zfs_znode_dmu_init(zfsvfs_t *zfsvfs, znode_t *zp, dmu_buf_t *db) void zfs_znode_dmu_fini(znode_t *zp) { - dmu_buf_t *db = zp->z_dbuf; 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)); - ASSERT(zp->z_dbuf != NULL); - zp->z_dbuf = NULL; - VERIFY(zp == dmu_buf_update_user(db, zp, NULL, NULL, NULL)); - dmu_buf_rele(db, NULL); + + sa_handle_destroy(zp->z_sa_hdl); + zp->z_sa_hdl = NULL; } /* @@ -620,26 +618,30 @@ zfs_znode_dmu_fini(znode_t *zp) * return the znode */ static znode_t * -zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz) +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; zp = kmem_cache_alloc(znode_cache, KM_SLEEP); ASSERT(zp->z_dirlocks == NULL); - ASSERT(zp->z_dbuf == NULL); 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_phys = NULL; + zp->z_sa_hdl = NULL; zp->z_unlinked = 0; zp->z_atime_dirty = 0; zp->z_mapcnt = 0; - zp->z_last_itx = 0; zp->z_id = db->db_object; zp->z_blksz = blksz; zp->z_seq = 0x7A4653; @@ -648,16 +650,39 @@ zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz) vp = ZTOV(zp); vn_reinit(vp); - zfs_znode_dmu_init(zfsvfs, zp, db); - - zp->z_gen = zp->z_phys->zp_gen; + 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); + kmem_cache_free(znode_cache, zp); + return (NULL); + } + zp->z_mode = mode; vp->v_vfsp = zfsvfs->z_parent->z_vfs; - vp->v_type = IFTOVT((mode_t)zp->z_phys->zp_mode); + + vp->v_type = IFTOVT((mode_t)mode); switch (vp->v_type) { case VDIR: - if (zp->z_phys->zp_flags & ZFS_XATTR) { + if (zp->z_pflags & ZFS_XATTR) { vn_setops(vp, zfs_xdvnodeops); vp->v_flag |= V_XATTRDIR; } else { @@ -667,7 +692,13 @@ zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz) break; case VBLK: case VCHR: - vp->v_rdev = zfs_cmpldev(zp->z_phys->zp_rdev); + { + uint64_t rdev; + VERIFY(sa_lookup(zp->z_sa_hdl, SA_ZPL_RDEV(zfsvfs), + &rdev, sizeof (rdev)) == 0); + + vp->v_rdev = zfs_cmpldev(rdev); + } /*FALLTHROUGH*/ case VFIFO: case VSOCK: @@ -676,7 +707,12 @@ zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz) break; case VREG: vp->v_flag |= VMODSORT; - vn_setops(vp, zfs_fvnodeops); + if (parent == zfsvfs->z_shares_dir) { + ASSERT(zp->z_uid == 0 && zp->z_gid == 0); + vn_setops(vp, zfs_sharevnodeops); + } else { + vn_setops(vp, zfs_fvnodeops); + } break; case VLNK: vn_setops(vp, zfs_symvnodeops); @@ -700,6 +736,9 @@ zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz) 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. * @@ -710,7 +749,6 @@ zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz) * flag - flags: * IS_ROOT_NODE - new object will be root * IS_XATTR - new object is an attribute - * IS_REPLAY - intent log replay * bonuslen - length of bonus buffer * setaclp - File/Dir initial ACL * fuidp - Tracks fuid allocation. @@ -720,21 +758,28 @@ zfs_znode_alloc(zfsvfs_t *zfsvfs, dmu_buf_t *db, int blksz) */ void zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr, - uint_t flag, znode_t **zpp, int bonuslen, zfs_acl_t *setaclp, - zfs_fuid_info_t **fuidp) + uint_t flag, znode_t **zpp, zfs_acl_ids_t *acl_ids) { - dmu_buf_t *db; - znode_phys_t *pzp; + 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; - flag |= IS_REPLAY; now = vap->va_ctime; /* see zfs_replay_create() */ gen = vap->va_nblocks; /* ditto */ } else { @@ -743,114 +788,229 @@ zfs_mknode(znode_t *dzp, vattr_t *vap, dmu_tx_t *tx, cred_t *cr, 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 to needed allocate a new object, so we accept the small chance + * 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 (flag & IS_REPLAY) { + if (zfsvfs->z_replay) { err = zap_create_claim_norm(zfsvfs->z_os, obj, zfsvfs->z_norm, DMU_OT_DIRECTORY_CONTENTS, - DMU_OT_ZNODE, sizeof (znode_phys_t) + bonuslen, tx); + obj_type, bonuslen, tx); ASSERT3U(err, ==, 0); } else { obj = zap_create_norm(zfsvfs->z_os, zfsvfs->z_norm, DMU_OT_DIRECTORY_CONTENTS, - DMU_OT_ZNODE, sizeof (znode_phys_t) + bonuslen, tx); + obj_type, bonuslen, tx); } } else { - if (flag & IS_REPLAY) { + if (zfsvfs->z_replay) { err = dmu_object_claim(zfsvfs->z_os, obj, DMU_OT_PLAIN_FILE_CONTENTS, 0, - DMU_OT_ZNODE, sizeof (znode_phys_t) + bonuslen, tx); + obj_type, bonuslen, tx); ASSERT3U(err, ==, 0); } else { obj = dmu_object_alloc(zfsvfs->z_os, DMU_OT_PLAIN_FILE_CONTENTS, 0, - DMU_OT_ZNODE, sizeof (znode_phys_t) + bonuslen, tx); + obj_type, bonuslen, tx); } } - VERIFY(0 == dmu_bonus_hold(zfsvfs->z_os, obj, NULL, &db)); - dmu_buf_will_dirty(db, tx); - /* - * Initialize the znode physical data to zero. - */ - ASSERT(db->db_size >= sizeof (znode_phys_t)); - bzero(db->db_data, db->db_size); - pzp = db->db_data; + 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_dbuf = db; - dzp->z_phys = pzp; dzp->z_id = obj; + } else { + dzp_pflags = dzp->z_pflags; } /* * If parent is an xattr, so am I. */ - if (dzp->z_phys->zp_flags & ZFS_XATTR) + if (dzp_pflags & ZFS_XATTR) { flag |= IS_XATTR; - - if (vap->va_type == VBLK || vap->va_type == VCHR) { - pzp->zp_rdev = zfs_expldev(vap->va_rdev); } if (zfsvfs->z_use_fuids) - pzp->zp_flags = ZFS_ARCHIVE | ZFS_AV_MODIFIED; + pflags = ZFS_ARCHIVE | ZFS_AV_MODIFIED; + else + pflags = 0; if (vap->va_type == VDIR) { - pzp->zp_size = 2; /* contents ("." and "..") */ - pzp->zp_links = (flag & (IS_ROOT_NODE | IS_XATTR)) ? 2 : 1; + size = 2; /* contents ("." and "..") */ + links = (flag & (IS_ROOT_NODE | IS_XATTR)) ? 2 : 1; + } else { + size = links = 0; } - pzp->zp_parent = dzp->z_id; + 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) - pzp->zp_flags |= ZFS_XATTR; + pflags |= ZFS_XATTR; - pzp->zp_gen = gen; + /* + * 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, pzp->zp_crtime); - ZFS_TIME_ENCODE(&now, pzp->zp_ctime); + ZFS_TIME_ENCODE(&now, crtime); + ZFS_TIME_ENCODE(&now, ctime); if (vap->va_mask & AT_ATIME) { - ZFS_TIME_ENCODE(&vap->va_atime, pzp->zp_atime); + ZFS_TIME_ENCODE(&vap->va_atime, atime); } else { - ZFS_TIME_ENCODE(&now, pzp->zp_atime); + ZFS_TIME_ENCODE(&now, atime); } if (vap->va_mask & AT_MTIME) { - ZFS_TIME_ENCODE(&vap->va_mtime, pzp->zp_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 { - ZFS_TIME_ENCODE(&now, pzp->zp_mtime); + 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); } - pzp->zp_mode = MAKEIMODE(vap->va_type, vap->va_mode); + 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)) { - ZFS_OBJ_HOLD_ENTER(zfsvfs, obj); - *zpp = zfs_znode_alloc(zfsvfs, db, 0); - ZFS_OBJ_HOLD_EXIT(zfsvfs, obj); + *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; } - zfs_perm_init(*zpp, dzp, flag, vap, tx, cr, setaclp, fuidp); + + (*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) { + err = zfs_aclset_common(*zpp, acl_ids->z_aclp, cr, tx); + ASSERT3P(err, ==, 0); + } + ZFS_OBJ_HOLD_EXIT(zfsvfs, obj); } +/* + * zfs_xvattr_set only updates the 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) +zfs_xvattr_set(znode_t *zp, xvattr_t *xvap, dmu_tx_t *tx) { xoptattr_t *xoap; @@ -858,60 +1018,86 @@ zfs_xvattr_set(znode_t *zp, xvattr_t *xvap) ASSERT(xoap); if (XVA_ISSET_REQ(xvap, XAT_CREATETIME)) { - ZFS_TIME_ENCODE(&xoap->xoa_createtime, zp->z_phys->zp_crtime); + 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); + 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); + 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); + 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); + 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); + 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); + 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); + 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); + 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); + 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); + 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); + 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)) { - (void) memcpy(zp->z_phys + 1, xoap->xoa_av_scanstamp, - sizeof (xoap->xoa_av_scanstamp)); - zp->z_phys->zp_flags |= ZFS_BONUS_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 @@ -921,35 +1107,42 @@ zfs_zget(zfsvfs_t *zfsvfs, uint64_t obj_num, znode_t **zpp) dmu_buf_t *db; znode_t *zp; int err; + sa_handle_t *hdl; *zpp = NULL; ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num); - err = dmu_bonus_hold(zfsvfs->z_os, obj_num, NULL, &db); + 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_ZNODE || - doi.doi_bonus_size < sizeof (znode_phys_t)) { - dmu_buf_rele(db, NULL); + 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 (EINVAL); } - zp = dmu_buf_get_user(db); - if (zp != NULL) { - mutex_enter(&zp->z_lock); + hdl = dmu_buf_get_user(db); + if (hdl != NULL) { + zp = sa_get_userdata(hdl); + /* - * Since we do immediate eviction of the z_dbuf, we - * should never find a dbuf with a znode that doesn't - * know about the dbuf. + * Since "SA" does immediate eviction we + * should never find a sa handle that doesn't + * know about the znode. */ - ASSERT3P(zp->z_dbuf, ==, db); + + ASSERT3P(zp, !=, NULL); + + mutex_enter(&zp->z_lock); ASSERT3U(zp->z_id, ==, obj_num); if (zp->z_unlinked) { err = ENOENT; @@ -958,7 +1151,7 @@ zfs_zget(zfsvfs_t *zfsvfs, uint64_t obj_num, znode_t **zpp) *zpp = zp; err = 0; } - dmu_buf_rele(db, NULL); + sa_buf_rele(db, NULL); mutex_exit(&zp->z_lock); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (err); @@ -966,11 +1159,24 @@ zfs_zget(zfsvfs_t *zfsvfs, uint64_t obj_num, znode_t **zpp) /* * 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); + zp = zfs_znode_alloc(zfsvfs, db, doi.doi_data_block_size, + doi.doi_bonus_type, NULL); + if (zp == NULL) { + err = ENOENT; + } else { + *zpp = zp; + } ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); - *zpp = zp; - return (0); + return (err); } int @@ -980,32 +1186,73 @@ zfs_rezget(znode_t *zp) dmu_object_info_t doi; dmu_buf_t *db; uint64_t obj_num = zp->z_id; + uint64_t mode; + sa_bulk_attr_t bulk[8]; int err; + int count = 0; + uint64_t gen; ZFS_OBJ_HOLD_ENTER(zfsvfs, obj_num); - err = dmu_bonus_hold(zfsvfs->z_os, obj_num, NULL, &db); + 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_ZNODE || - doi.doi_bonus_size < sizeof (znode_phys_t)) { - dmu_buf_rele(db, NULL); + 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 (EINVAL); } - if (((znode_phys_t *)db->db_data)->zp_gen != zp->z_gen) { - dmu_buf_rele(db, NULL); + zfs_znode_sa_init(zfsvfs, zp, db, doi.doi_bonus_type, NULL); + + /* 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 (EIO); + } + + zp->z_mode = mode; + + if (gen != zp->z_gen) { + zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); return (EIO); } - zfs_znode_dmu_init(zfsvfs, zp, db); - zp->z_unlinked = (zp->z_phys->zp_links == 0); + zp->z_unlinked = (zp->z_links == 0); zp->z_blksz = doi.doi_data_block_size; ZFS_OBJ_HOLD_EXIT(zfsvfs, obj_num); @@ -1019,11 +1266,13 @@ 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 = zp->z_phys->zp_acl.z_acl_extern_obj; + uint64_t acl_obj = zfs_external_acl(zp); ZFS_OBJ_HOLD_ENTER(zfsvfs, obj); - if (acl_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); @@ -1037,7 +1286,7 @@ zfs_zinactive(znode_t *zp) zfsvfs_t *zfsvfs = zp->z_zfsvfs; uint64_t z_id = zp->z_id; - ASSERT(zp->z_dbuf && zp->z_phys); + ASSERT(zp->z_sa_hdl); /* * Don't allow a zfs_zget() while were trying to release this znode @@ -1076,6 +1325,7 @@ zfs_zinactive(znode_t *zp) zfs_rmnode(zp); return; } + mutex_exit(&zp->z_lock); zfs_znode_dmu_fini(zp); ZFS_OBJ_HOLD_EXIT(zfsvfs, z_id); @@ -1096,65 +1346,51 @@ zfs_znode_free(znode_t *zp) 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); VFS_RELE(zfsvfs->z_vfs); } void -zfs_time_stamper_locked(znode_t *zp, uint_t flag, dmu_tx_t *tx) +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; - ASSERT(MUTEX_HELD(&zp->z_lock)); - gethrestime(&now); - if (tx) { - dmu_buf_will_dirty(zp->z_dbuf, tx); + 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_phys->zp_atime); + if (flag & AT_ATIME) { + ZFS_TIME_ENCODE(&now, zp->z_atime); + } if (flag & AT_MTIME) { - ZFS_TIME_ENCODE(&now, zp->z_phys->zp_mtime); - if (zp->z_zfsvfs->z_use_fuids) - zp->z_phys->zp_flags |= (ZFS_ARCHIVE | ZFS_AV_MODIFIED); + 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, zp->z_phys->zp_ctime); + ZFS_TIME_ENCODE(&now, ctime); if (zp->z_zfsvfs->z_use_fuids) - zp->z_phys->zp_flags |= ZFS_ARCHIVE; + zp->z_pflags |= ZFS_ARCHIVE; } } /* - * Update the requested znode timestamps with the current time. - * If we are in a transaction, then go ahead and mark the znode - * dirty in the transaction so the timestamps will go to disk. - * Otherwise, we will get pushed next time the znode is updated - * in a transaction, or when this znode eventually goes inactive. - * - * Why is this OK? - * 1 - Only the ACCESS time is ever updated outside of a transaction. - * 2 - Multiple consecutive updates will be collapsed into a single - * znode update by the transaction grouping semantics of the DMU. - */ -void -zfs_time_stamper(znode_t *zp, uint_t flag, dmu_tx_t *tx) -{ - mutex_enter(&zp->z_lock); - zfs_time_stamper_locked(zp, flag, tx); - mutex_exit(&zp->z_lock); -} - -/* * Grow the block size for a file. * * IN: zp - znode of file to free data in. @@ -1176,17 +1412,18 @@ zfs_grow_blocksize(znode_t *zp, uint64_t size, dmu_tx_t *tx) * we will not grow. If there is more than one block in a file, * the blocksize cannot change. */ - if (zp->z_blksz && zp->z_phys->zp_size > zp->z_blksz) + 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; ASSERT3U(error, ==, 0); /* What blocksize did we actually get? */ - dmu_object_size_from_db(zp->z_dbuf, &zp->z_blksz, &dummy); + dmu_object_size_from_db(sa_get_db(zp->z_sa_hdl), &zp->z_blksz, &dummy); } /* @@ -1229,13 +1466,14 @@ zfs_extend(znode_t *zp, uint64_t end) /* * Nothing to do if file already at desired length. */ - if (end <= zp->z_phys->zp_size) { + if (end <= zp->z_size) { zfs_range_unlock(rl); return (0); } top: tx = dmu_tx_create(zfsvfs->z_os); - dmu_tx_hold_bonus(tx, zp->z_id); + 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)) { /* @@ -1263,12 +1501,14 @@ top: zfs_range_unlock(rl); return (error); } - dmu_buf_will_dirty(zp->z_dbuf, tx); if (newblksz) zfs_grow_blocksize(zp, newblksz, tx); - zp->z_phys->zp_size = end; + 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)); zfs_range_unlock(rl); @@ -1302,13 +1542,13 @@ zfs_free_range(znode_t *zp, uint64_t off, uint64_t len) /* * Nothing to do if file already at desired length. */ - if (off >= zp->z_phys->zp_size) { + if (off >= zp->z_size) { zfs_range_unlock(rl); return (0); } - if (off + len > zp->z_phys->zp_size) - len = zp->z_phys->zp_size - off; + if (off + len > zp->z_size) + len = zp->z_size - off; error = dmu_free_long_range(zfsvfs->z_os, zp->z_id, off, len); @@ -1334,6 +1574,8 @@ zfs_trunc(znode_t *zp, uint64_t end) 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. @@ -1343,7 +1585,7 @@ zfs_trunc(znode_t *zp, uint64_t end) /* * Nothing to do if file already at desired length. */ - if (end >= zp->z_phys->zp_size) { + if (end >= zp->z_size) { zfs_range_unlock(rl); return (0); } @@ -1355,7 +1597,8 @@ zfs_trunc(znode_t *zp, uint64_t end) } top: tx = dmu_tx_create(zfsvfs->z_os); - dmu_tx_hold_bonus(tx, zp->z_id); + dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); + zfs_sa_upgrade_txholds(tx, zp); error = dmu_tx_assign(tx, TXG_NOWAIT); if (error) { if (error == ERESTART) { @@ -1367,9 +1610,17 @@ top: zfs_range_unlock(rl); return (error); } - dmu_buf_will_dirty(zp->z_dbuf, tx); - zp->z_phys->zp_size = end; + 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); @@ -1421,9 +1672,17 @@ zfs_freesp(znode_t *zp, uint64_t off, uint64_t len, int flag, boolean_t log) 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 (off > zp->z_phys->zp_size) { + 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; @@ -1434,8 +1693,9 @@ zfs_freesp(znode_t *zp, uint64_t off, uint64_t len, int flag, boolean_t log) /* * Check for any locks in the region to be freed. */ - if (MANDLOCK(vp, (mode_t)zp->z_phys->zp_mode)) { - uint64_t length = (len ? len : zp->z_phys->zp_size - off); + + 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); } @@ -1444,14 +1704,15 @@ zfs_freesp(znode_t *zp, uint64_t off, uint64_t len, int flag, boolean_t log) error = zfs_trunc(zp, off); } else { if ((error = zfs_free_range(zp, off, len)) == 0 && - off + len > zp->z_phys->zp_size) + 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_bonus(tx, zp->z_id); + dmu_tx_hold_sa(tx, zp->z_sa_hdl, B_FALSE); + zfs_sa_upgrade_txholds(tx, zp); error = dmu_tx_assign(tx, TXG_NOWAIT); if (error) { if (error == ERESTART) { @@ -1463,7 +1724,14 @@ log: return (error); } - zfs_time_stamper(zp, CONTENT_MODIFIED, tx); + 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); @@ -1474,15 +1742,17 @@ void zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx) { zfsvfs_t zfsvfs; - uint64_t moid, doid, version; + 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; vnode_t *vp; vattr_t vattr; znode_t *zp; + zfs_acl_ids_t acl_ids; /* * First attempt to create master node. @@ -1499,12 +1769,7 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx) /* * Set starting attributes. */ - if (spa_version(dmu_objset_spa(os)) >= SPA_VERSION_FUID) - version = ZPL_VERSION; - else - version = ZPL_VERSION_FUID - 1; - error = zap_update(os, moid, ZPL_VERSION_STR, - 8, 1, &version, tx); + 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 */ @@ -1515,9 +1780,8 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx) VERIFY(nvpair_value_uint64(elem, &val) == 0); name = nvpair_name(elem); if (strcmp(name, zfs_prop_to_name(ZFS_PROP_VERSION)) == 0) { - version = val; - error = zap_update(os, moid, ZPL_VERSION_STR, - 8, 1, &version, tx); + if (val < version) + version = val; } else { error = zap_update(os, moid, name, 8, 1, &val, tx); } @@ -1528,13 +1792,26 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx) 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. */ - doid = zap_create(os, DMU_OT_UNLINKED_SET, DMU_OT_NONE, 0, tx); + obj = zap_create(os, DMU_OT_UNLINKED_SET, DMU_OT_NONE, 0, tx); - error = zap_add(os, moid, ZFS_UNLINKED_SET, 8, 1, &doid, tx); + error = zap_add(os, moid, ZFS_UNLINKED_SET, 8, 1, &obj, tx); ASSERT(error == 0); /* @@ -1548,8 +1825,11 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx) vattr.va_gid = crgetgid(cr); 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); vp = ZTOV(rootzp); vn_reinit(vp); @@ -1561,7 +1841,14 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx) 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. @@ -1573,60 +1860,153 @@ zfs_create_fs(objset_t *os, cred_t *cr, nvlist_t *zplprops, dmu_tx_t *tx) list_create(&zfsvfs.z_all_znodes, sizeof (znode_t), offsetof(znode_t, z_link_node)); - ASSERT(!POINTER_IS_VALID(rootzp->z_zfsvfs)); + for (i = 0; i != ZFS_OBJ_MTX_SZ; i++) + mutex_init(&zfsvfs.z_hold_mtx[i], NULL, MUTEX_DEFAULT, NULL); + rootzp->z_zfsvfs = &zfsvfs; - zfs_mknode(rootzp, &vattr, tx, cr, IS_ROOT_NODE, &zp, 0, NULL, NULL); + 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); ASSERT(!vn_in_dnlc(ZTOV(rootzp))); /* not valid to move */ 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); ZTOV(rootzp)->v_count = 0; - dmu_buf_rele(rootzp->z_dbuf, NULL); - rootzp->z_dbuf = NULL; + 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]); } #endif /* _KERNEL */ -/* - * 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, uint64_t obj, uint64_t *pobjp, int *is_xattrdir) +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) { - dmu_buf_t *db; dmu_object_info_t doi; - znode_phys_t *zp; int error; - if ((error = dmu_bonus_hold(osp, obj, FTAG, &db)) != 0) + if ((error = sa_buf_hold(osp, obj, FTAG, db)) != 0) return (error); - dmu_object_info_from_db(db, &doi); - if (doi.doi_bonus_type != DMU_OT_ZNODE || + 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)) { - dmu_buf_rele(db, FTAG); - return (EINVAL); + sa_buf_rele(*db, FTAG); + return (ENOTSUP); } - zp = db->db_data; - *pobjp = zp->zp_parent; - *is_xattrdir = ((zp->zp_flags & ZFS_XATTR) != 0) && - S_ISDIR(zp->zp_mode); - dmu_buf_rele(db, FTAG); + error = sa_handle_get(osp, obj, NULL, SA_HDL_PRIVATE, hdlp); + if (error != 0) { + sa_buf_rele(*db, FTAG); + return (error); + } return (0); } -int -zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len) +void +zfs_release_sa_handle(sa_handle_t *hdl, dmu_buf_t *db) +{ + sa_handle_destroy(hdl); + sa_buf_rele(db, FTAG); +} + +/* + * 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(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; + sa_bulk_attr_t bulk[3]; + 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); + + *pobjp = parent; + *is_xattrdir = ((pflags & ZFS_XATTR) != 0) && S_ISDIR(mode); + + 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; for (;;) { uint64_t pobj; @@ -1634,7 +2014,10 @@ zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len) size_t complen; int is_xattrdir; - if ((error = zfs_obj_to_pobj(osp, obj, &pobj, + if (prevdb) + zfs_release_sa_handle(prevhdl, prevdb); + + if ((error = zfs_obj_to_pobj(sa_hdl, sa_table, &pobj, &is_xattrdir)) != 0) break; @@ -1659,9 +2042,80 @@ zfs_obj_to_path(objset_t *osp, uint64_t obj, char *buf, int len) 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); + 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); } 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); + if (error != 0) + return (error); + + error = zfs_obj_to_path_impl(osp, obj, hdl, sa_table, buf, len); + + zfs_release_sa_handle(hdl, db); + 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); + if (error != 0) + return (error); + + error = zfs_obj_to_stats_impl(hdl, sa_table, sb); + if (error != 0) { + zfs_release_sa_handle(hdl, db); + return (error); + } + + error = zfs_obj_to_path_impl(osp, obj, hdl, sa_table, buf, len); + + zfs_release_sa_handle(hdl, db); return (error); }