X-Git-Url: https://git.whamcloud.com/?p=fs%2Flustre-release.git;a=blobdiff_plain;f=lustre%2Fosd-zfs%2Fosd_index.c;h=765099ab00917b35006ebb8a5090d3acad7c3afb;hp=b96e6422182e2d6c1082b4c4deb136b634c5fd51;hb=7a18fe124c215ad7ab1806abd62bea9c7b9592c6;hpb=cf753f6b4bdc267b5169e97293f69b5362b826de

diff --git a/lustre/osd-zfs/osd_index.c b/lustre/osd-zfs/osd_index.c
index b96e642..765099a 100644
--- a/lustre/osd-zfs/osd_index.c
+++ b/lustre/osd-zfs/osd_index.c
@@ -15,21 +15,15 @@
  *
  * You should have received a copy of the GNU General Public License
  * version 2 along with this program; If not, see
- * http://www.sun.com/software/products/lustre/docs/GPLv2.pdf
- *
- * Please contact Sun Microsystems, Inc., 4150 Network Circle, Santa Clara,
- * CA 95054 USA or visit www.sun.com if you need additional information or
- * have any questions.
+ * http://www.gnu.org/licenses/gpl-2.0.html
  *
  * GPL HEADER END
  */
 /*
  * Copyright (c) 2009, 2010, Oracle and/or its affiliates. All rights reserved.
  * Use is subject to license terms.
- */
-/*
- * Copyright (c) 2012, 2013, Intel Corporation.
- * Use is subject to license terms.
+ *
+ * Copyright (c) 2012, 2017, Intel Corporation.
  */
 /*
  * This file is part of Lustre, http://www.lustre.org/
@@ -43,7 +37,6 @@
 
 #define DEBUG_SUBSYSTEM S_OSD
 
-#include <lustre_ver.h>
 #include <libcfs/libcfs.h>
 #include <obd_support.h>
 #include <lustre_net.h>
@@ -55,7 +48,6 @@
 #include "osd_internal.h"
 
 #include <sys/dnode.h>
-#include <sys/dbuf.h>
 #include <sys/spa.h>
 #include <sys/stat.h>
 #include <sys/zap.h>
@@ -66,34 +58,106 @@
 #include <sys/dsl_prop.h>
 #include <sys/sa_impl.h>
 #include <sys/txg.h>
+#include <lustre_scrub.h>
+
+/* We don't actually have direct access to the zap_hashbits() function
+ * so just pretend like we do for now.  If this ever breaks we can look at
+ * it at that time. */
+#define zap_hashbits(zc) 48
+/*
+ * ZFS hash format:
+ * | cd (16 bits) | hash (48 bits) |
+ * we need it in other form:
+ * |0| hash (48 bit) | cd (15 bit) |
+ * to be a full 64-bit ordered hash so that Lustre readdir can use it to merge
+ * the readdir hashes from multiple directory stripes uniformly on the client.
+ * Another point is sign bit, the hash range should be in [0, 2^63-1] because
+ * loff_t (for llseek) needs to be a positive value.  This means the "cd" field
+ * should only be the low 15 bits.
+ */
+uint64_t osd_zap_cursor_serialize(zap_cursor_t *zc)
+{
+	uint64_t zfs_hash = zap_cursor_serialize(zc) & (~0ULL >> 1);
+
+	return (zfs_hash >> zap_hashbits(zc)) |
+		(zfs_hash << (63 - zap_hashbits(zc)));
+}
+
+void osd_zap_cursor_init_serialized(zap_cursor_t *zc, struct objset *os,
+				    uint64_t id, uint64_t dirhash)
+{
+	uint64_t zfs_hash = ((dirhash << zap_hashbits(zc)) & (~0ULL >> 1)) |
+		(dirhash >> (63 - zap_hashbits(zc)));
+
+	zap_cursor_init_serialized(zc, os, id, zfs_hash);
+}
+
+int osd_zap_cursor_init(zap_cursor_t **zc, struct objset *os,
+			uint64_t id, uint64_t dirhash)
+{
+	zap_cursor_t *t;
+
+	OBD_ALLOC_PTR(t);
+	if (unlikely(t == NULL))
+		return -ENOMEM;
+
+	osd_zap_cursor_init_serialized(t, os, id, dirhash);
+	*zc = t;
+
+	return 0;
+}
+
+void osd_zap_cursor_fini(zap_cursor_t *zc)
+{
+	zap_cursor_fini(zc);
+	OBD_FREE_PTR(zc);
+}
+
+static inline void osd_obj_cursor_init_serialized(zap_cursor_t *zc,
+						 struct osd_object *o,
+						 uint64_t dirhash)
+{
+	struct osd_device *d = osd_obj2dev(o);
+	osd_zap_cursor_init_serialized(zc, d->od_os,
+				       o->oo_dn->dn_object, dirhash);
+}
+
+static inline int osd_obj_cursor_init(zap_cursor_t **zc, struct osd_object *o,
+			uint64_t dirhash)
+{
+	struct osd_device *d = osd_obj2dev(o);
+	return osd_zap_cursor_init(zc, d->od_os, o->oo_dn->dn_object, dirhash);
+}
 
 static struct dt_it *osd_index_it_init(const struct lu_env *env,
 				       struct dt_object *dt,
-				       __u32 unused,
-				       struct lustre_capa *capa)
+				       __u32 unused)
 {
 	struct osd_thread_info  *info = osd_oti_get(env);
 	struct osd_zap_it       *it;
 	struct osd_object       *obj = osd_dt_obj(dt);
-	struct osd_device       *osd = osd_obj2dev(obj);
 	struct lu_object        *lo  = &dt->do_lu;
+	int			 rc;
 	ENTRY;
 
-	/* XXX: check capa ? */
+	if (obj->oo_destroyed)
+		RETURN(ERR_PTR(-ENOENT));
 
 	LASSERT(lu_object_exists(lo));
-	LASSERT(obj->oo_db);
-	LASSERT(udmu_object_is_zap(obj->oo_db));
+	LASSERT(obj->oo_dn);
 	LASSERT(info);
 
-	it = &info->oti_it_zap;
-
-	if (udmu_zap_cursor_init(&it->ozi_zc, &osd->od_objset,
-				 obj->oo_db->db_object, 0))
+	OBD_SLAB_ALLOC_PTR_GFP(it, osd_zapit_cachep, GFP_NOFS);
+	if (it == NULL)
 		RETURN(ERR_PTR(-ENOMEM));
 
+	rc = osd_obj_cursor_init(&it->ozi_zc, obj, 0);
+	if (rc != 0) {
+		OBD_SLAB_FREE_PTR(it, osd_zapit_cachep);
+		RETURN(ERR_PTR(rc));
+	}
+
 	it->ozi_obj   = obj;
-	it->ozi_capa  = capa;
 	it->ozi_reset = 1;
 	lu_object_get(lo);
 
@@ -102,8 +166,8 @@ static struct dt_it *osd_index_it_init(const struct lu_env *env,
 
 static void osd_index_it_fini(const struct lu_env *env, struct dt_it *di)
 {
-	struct osd_zap_it *it = (struct osd_zap_it *)di;
-	struct osd_object *obj;
+	struct osd_zap_it	*it	= (struct osd_zap_it *)di;
+	struct osd_object	*obj;
 	ENTRY;
 
 	LASSERT(it);
@@ -111,8 +175,9 @@ static void osd_index_it_fini(const struct lu_env *env, struct dt_it *di)
 
 	obj = it->ozi_obj;
 
-	udmu_zap_cursor_fini(it->ozi_zc);
-	lu_object_put(env, &obj->oo_dt.do_lu);
+	osd_zap_cursor_fini(it->ozi_zc);
+	osd_object_put(env, obj);
+	OBD_SLAB_FREE_PTR(it, osd_zapit_cachep);
 
 	EXIT;
 }
@@ -124,57 +189,6 @@ static void osd_index_it_put(const struct lu_env *env, struct dt_it *di)
 	 *      next/finish. */
 }
 
-int udmu_zap_cursor_retrieve_key(const struct lu_env *env,
-				 zap_cursor_t *zc, char *key, int max)
-{
-	zap_attribute_t *za = &osd_oti_get(env)->oti_za;
-	int		 err;
-
-	if ((err = zap_cursor_retrieve(zc, za)))
-		return err;
-
-	if (key)
-		strcpy(key, za->za_name);
-
-	return 0;
-}
-
-/*
- * zap_cursor_retrieve read from current record.
- * to read bytes we need to call zap_lookup explicitly.
- */
-int udmu_zap_cursor_retrieve_value(const struct lu_env *env,
-		zap_cursor_t *zc,  char *buf,
-		int buf_size, int *bytes_read)
-{
-	zap_attribute_t *za = &osd_oti_get(env)->oti_za;
-	int err, actual_size;
-
-	if ((err = zap_cursor_retrieve(zc, za)))
-		return err;
-
-	if (za->za_integer_length <= 0)
-		return (ERANGE);
-
-	actual_size = za->za_integer_length * za->za_num_integers;
-
-	if (actual_size > buf_size) {
-		actual_size = buf_size;
-		buf_size = actual_size / za->za_integer_length;
-	} else {
-		buf_size = za->za_num_integers;
-	}
-
-	err = -zap_lookup(zc->zc_objset, zc->zc_zapobj,
-			za->za_name, za->za_integer_length,
-			buf_size, buf);
-
-	if (!err)
-		*bytes_read = actual_size;
-
-	return err;
-}
-
 static inline void osd_it_append_attrs(struct lu_dirent *ent, __u32 attr,
 				       int len, __u16 type)
 {
@@ -193,95 +207,149 @@ static inline void osd_it_append_attrs(struct lu_dirent *ent, __u32 attr,
 	ent->lde_attrs = cpu_to_le32(ent->lde_attrs);
 }
 
-/*
- * as we don't know FID, we can't use LU object, so this function
- * partially duplicate __osd_xattr_get() which is built around
- * LU-object and uses it to cache data like regular EA dnode, etc
+int __osd_xattr_load_by_oid(struct osd_device *osd, uint64_t oid, nvlist_t **sa)
+{
+	sa_handle_t *hdl;
+	dmu_buf_t *db;
+	int rc;
+
+	rc = -dmu_bonus_hold(osd->od_os, oid, osd_obj_tag, &db);
+	if (rc < 0) {
+		CERROR("%s: can't get bonus, rc = %d\n", osd->od_svname, rc);
+		return rc;
+	}
+
+	rc = -sa_handle_get_from_db(osd->od_os, db, NULL, SA_HDL_PRIVATE, &hdl);
+	if (rc) {
+		dmu_buf_rele(db, osd_obj_tag);
+		return rc;
+	}
+
+	rc = __osd_xattr_load(osd, hdl, sa);
+
+	sa_handle_destroy(hdl);
+
+	return rc;
+}
+/**
+ * Get the object's FID from its LMA EA.
+ *
+ * \param[in] env	pointer to the thread context
+ * \param[in] osd	pointer to the OSD device
+ * \param[in] oid	the object's local identifier
+ * \param[out] fid	the buffer to hold the object's FID
+ *
+ * \retval		0 for success
+ * \retval		negative error number on failure
  */
-static int osd_find_parent_by_dnode(const struct lu_env *env,
-				    struct dt_object *o,
-				    struct lu_fid *fid)
+int osd_get_fid_by_oid(const struct lu_env *env, struct osd_device *osd,
+		       uint64_t oid, struct lu_fid *fid)
 {
-	struct lustre_mdt_attrs	*lma;
-	udmu_objset_t		*uos = &osd_obj2dev(osd_dt_obj(o))->od_objset;
+	struct objset		*os	  = osd->od_os;
+	struct osd_thread_info	*oti	  = osd_oti_get(env);
+	struct lustre_mdt_attrs	*lma	  =
+			(struct lustre_mdt_attrs *)oti->oti_buf;
 	struct lu_buf		 buf;
-	sa_handle_t		*sa_hdl;
-	nvlist_t		*nvbuf = NULL;
-	uchar_t			*value;
-	uint64_t		 dnode;
-	int			 rc, size;
+	nvlist_t		*sa_xattr = NULL;
+	sa_handle_t		*sa_hdl   = NULL;
+	uchar_t			*nv_value = NULL;
+	uint64_t		 xattr	  = ZFS_NO_OBJECT;
+	int			 size	  = 0;
+	int			 rc;
 	ENTRY;
 
-	/* first of all, get parent dnode from own attributes */
-	LASSERT(osd_dt_obj(o)->oo_db);
-	rc = -sa_handle_get(uos->os, osd_dt_obj(o)->oo_db->db_object,
-			    NULL, SA_HDL_PRIVATE, &sa_hdl);
-	if (rc)
-		RETURN(rc);
+	rc = __osd_xattr_load_by_oid(osd, oid, &sa_xattr);
+	if (rc == -ENOENT)
+		goto regular;
 
-	dnode = ZFS_NO_OBJECT;
-	rc = -sa_lookup(sa_hdl, SA_ZPL_PARENT(uos), &dnode, 8);
-	sa_handle_destroy(sa_hdl);
-	if (rc)
-		RETURN(rc);
+	if (rc != 0)
+		GOTO(out, rc);
 
-	/* now get EA buffer */
-	rc = __osd_xattr_load(uos, dnode, &nvbuf);
-	if (rc)
-		GOTO(regular, rc);
+	rc = -nvlist_lookup_byte_array(sa_xattr, XATTR_NAME_LMA, &nv_value,
+				       &size);
+	if (rc == -ENOENT)
+		goto regular;
 
-	/* XXX: if we get that far.. should we cache the result? */
+	if (rc != 0)
+		GOTO(out, rc);
 
-	/* try to find LMA attribute */
-	LASSERT(nvbuf != NULL);
-	rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA, &value, &size);
-	if (rc == 0 && size >= sizeof(*lma)) {
-		lma = (struct lustre_mdt_attrs *)value;
-		lustre_lma_swab(lma);
-		*fid = lma->lma_self_fid;
-		GOTO(out, rc = 0);
-	}
+	if (unlikely(size > sizeof(oti->oti_buf)))
+		GOTO(out, rc = -ERANGE);
 
-regular:
-	/* no LMA attribute in SA, let's try regular EA */
+	memcpy(lma, nv_value, size);
 
-	/* first of all, get parent dnode storing regular EA */
-	rc = -sa_handle_get(uos->os, dnode, NULL, SA_HDL_PRIVATE, &sa_hdl);
-	if (rc)
+	goto found;
+
+regular:
+	rc = -sa_handle_get(os, oid, NULL, SA_HDL_PRIVATE, &sa_hdl);
+	if (rc != 0)
 		GOTO(out, rc);
 
-	dnode = ZFS_NO_OBJECT;
-	rc = -sa_lookup(sa_hdl, SA_ZPL_XATTR(uos), &dnode, 8);
+	rc = -sa_lookup(sa_hdl, SA_ZPL_XATTR(osd), &xattr, 8);
 	sa_handle_destroy(sa_hdl);
-	if (rc)
+	if (rc != 0)
 		GOTO(out, rc);
 
-	CLASSERT(sizeof(*lma) <= sizeof(osd_oti_get(env)->oti_buf));
-	buf.lb_buf = osd_oti_get(env)->oti_buf;
-	buf.lb_len = sizeof(osd_oti_get(env)->oti_buf);
-
-	/* now try to find LMA */
-	rc = __osd_xattr_get_large(env, uos, dnode, &buf,
+	buf.lb_buf = lma;
+	buf.lb_len = sizeof(oti->oti_buf);
+	rc = __osd_xattr_get_large(env, osd, xattr, &buf,
 				   XATTR_NAME_LMA, &size);
-	if (rc == 0 && size >= sizeof(*lma)) {
-		lma = buf.lb_buf;
-		lustre_lma_swab(lma);
-		*fid = lma->lma_self_fid;
-		GOTO(out, rc = 0);
-	} else if (rc < 0) {
+	if (rc != 0)
 		GOTO(out, rc);
-	} else {
+
+found:
+	if (size < sizeof(*lma))
 		GOTO(out, rc = -EIO);
+
+	lustre_lma_swab(lma);
+	if (unlikely((lma->lma_incompat & ~LMA_INCOMPAT_SUPP) ||
+		     CFS_FAIL_CHECK(OBD_FAIL_OSD_LMA_INCOMPAT))) {
+		CWARN("%s: unsupported incompat LMA feature(s) %#x for "
+		      "oid = %#llx\n", osd->od_svname,
+		      lma->lma_incompat & ~LMA_INCOMPAT_SUPP, oid);
+		GOTO(out, rc = -EOPNOTSUPP);
+	} else {
+		*fid = lma->lma_self_fid;
+		GOTO(out, rc = 0);
 	}
 
 out:
-	if (nvbuf != NULL)
-		nvlist_free(nvbuf);
+	if (sa_xattr != NULL)
+		nvlist_free(sa_xattr);
+	return rc;
+}
+
+/*
+ * As we don't know FID, we can't use LU object, so this function
+ * partially duplicate osd_xattr_get_internal() which is built around
+ * LU-object and uses it to cache data like regular EA dnode, etc
+ */
+static int osd_find_parent_by_dnode(const struct lu_env *env,
+				    struct dt_object *o,
+				    struct lu_fid *fid, uint64_t *oid)
+{
+	struct osd_object	*obj = osd_dt_obj(o);
+	struct osd_device	*osd = osd_obj2dev(obj);
+	uint64_t		 dnode = ZFS_NO_OBJECT;
+	int			 rc;
+	ENTRY;
+
+	/* first of all, get parent dnode from own attributes */
+	rc = osd_sa_handle_get(obj);
+	if (rc != 0)
+		RETURN(rc);
+	rc = -sa_lookup(obj->oo_sa_hdl, SA_ZPL_PARENT(osd), &dnode, 8);
+	if (!rc) {
+		if (oid)
+			*oid = dnode;
+		rc = osd_get_fid_by_oid(env, osd, dnode, fid);
+	}
+
 	RETURN(rc);
 }
 
 static int osd_find_parent_fid(const struct lu_env *env, struct dt_object *o,
-			       struct lu_fid *fid)
+			       struct lu_fid *fid, uint64_t *oid)
 {
 	struct link_ea_header  *leh;
 	struct link_ea_entry   *lee;
@@ -292,10 +360,9 @@ static int osd_find_parent_fid(const struct lu_env *env, struct dt_object *o,
 	buf.lb_buf = osd_oti_get(env)->oti_buf;
 	buf.lb_len = sizeof(osd_oti_get(env)->oti_buf);
 
-	rc = osd_xattr_get(env, o, &buf, XATTR_NAME_LINK, BYPASS_CAPA);
+	rc = osd_xattr_get(env, o, &buf, XATTR_NAME_LINK);
 	if (rc == -ERANGE) {
-		rc = osd_xattr_get(env, o, &LU_BUF_NULL,
-				   XATTR_NAME_LINK, BYPASS_CAPA);
+		rc = osd_xattr_get(env, o, &LU_BUF_NULL, XATTR_NAME_LINK);
 		if (rc < 0)
 			RETURN(rc);
 		LASSERT(rc > 0);
@@ -303,7 +370,7 @@ static int osd_find_parent_fid(const struct lu_env *env, struct dt_object *o,
 		if (buf.lb_buf == NULL)
 			RETURN(-ENOMEM);
 		buf.lb_len = rc;
-		rc = osd_xattr_get(env, o, &buf, XATTR_NAME_LINK, BYPASS_CAPA);
+		rc = osd_xattr_get(env, o, &buf, XATTR_NAME_LINK);
 	}
 	if (rc < 0)
 		GOTO(out, rc);
@@ -335,7 +402,7 @@ out:
 	if (rc == 0) {
 		struct lu_fid fid2;
 		int rc2;
-		rc2 = osd_find_parent_by_dnode(env, o, &fid2);
+		rc2 = osd_find_parent_by_dnode(env, o, &fid2, oid);
 		if (rc2 == 0)
 			if (lu_fid_eq(fid, &fid2) == 0)
 				CERROR("wrong parent: "DFID" != "DFID"\n",
@@ -345,163 +412,575 @@ out:
 
 	/* no LinkEA is found, let's try to find the fid in parent's LMA */
 	if (unlikely(rc != 0))
-		rc = osd_find_parent_by_dnode(env, o, fid);
+		rc = osd_find_parent_by_dnode(env, o, fid, oid);
 
 	RETURN(rc);
 }
 
+/*
+ * When lookup item under striped directory, we need to locate the master
+ * MDT-object of the striped directory firstly, then the client will send
+ * lookup (getattr_by_name) RPC to the MDT with some slave MDT-object's FID
+ * and the item's name. If the system is restored from MDT file level backup,
+ * then before the OI scrub completely built the OI files, the OI mappings of
+ * the master MDT-object and slave MDT-object may be invalid. Usually, it is
+ * not a problem for the master MDT-object. Because when locate the master
+ * MDT-object, we will do name based lookup (for the striped directory itself)
+ * firstly, during such process we can setup the correct OI mapping for the
+ * master MDT-object. But it will be trouble for the slave MDT-object. Because
+ * the client will not trigger name based lookup on the MDT to locate the slave
+ * MDT-object before locating item under the striped directory, then when
+ * osd_fid_lookup(), it will find that the OI mapping for the slave MDT-object
+ * is invalid and does not know what the right OI mapping is, then the MDT has
+ * to return -EINPROGRESS to the client to notify that the OI scrub is rebuiding
+ * the OI file, related OI mapping is unknown yet, please try again later. And
+ * then client will re-try the RPC again and again until related OI mapping has
+ * been updated. That is quite inefficient.
+ *
+ * To resolve above trouble, we will handle it as the following two cases:
+ *
+ * 1) The slave MDT-object and the master MDT-object are on different MDTs.
+ *    It is relative easy. Be as one of remote MDT-objects, the slave MDT-object
+ *    is linked under /REMOTE_PARENT_DIR with the name of its FID string.
+ *    We can locate the slave MDT-object via lookup the /REMOTE_PARENT_DIR
+ *    directly. Please check osd_fid_lookup().
+ *
+ * 2) The slave MDT-object and the master MDT-object reside on the same MDT.
+ *    Under such case, during lookup the master MDT-object, we will lookup the
+ *    slave MDT-object via readdir against the master MDT-object, because the
+ *    slave MDT-objects information are stored as sub-directories with the name
+ *    "${FID}:${index}". Then when find the local slave MDT-object, its OI
+ *    mapping will be recorded. Then subsequent osd_fid_lookup() will know
+ *    the correct OI mapping for the slave MDT-object.
+ */
+static int osd_check_lmv(const struct lu_env *env, struct osd_device *osd,
+			 uint64_t oid, const struct lu_fid *fid)
+{
+	struct osd_thread_info *info = osd_oti_get(env);
+	struct luz_direntry *zde = &info->oti_zde;
+	zap_attribute_t *za = &info->oti_za;
+	zap_cursor_t *zc = &info->oti_zc;
+	struct lu_fid *tfid = &info->oti_fid;
+	nvlist_t *nvbuf = NULL;
+	struct lmv_mds_md_v1 *lmv = NULL;
+	int size;
+	int rc;
+	ENTRY;
+
+	rc = __osd_xattr_load_by_oid(osd, oid, &nvbuf);
+	if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
+		RETURN(0);
+
+	if (rc)
+		RETURN(rc);
+
+	rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMV,
+				       (uchar_t **)&lmv, &size);
+	if (rc == -ENOENT || rc == -EEXIST || rc == -ENODATA)
+		GOTO(out_nvbuf, rc = 0);
+
+	if (rc)
+		GOTO(out_nvbuf, rc);
+
+	if (le32_to_cpu(lmv->lmv_magic) != LMV_MAGIC_V1)
+		GOTO(out_nvbuf, rc = -EINVAL);
+
+	zap_cursor_init_serialized(zc, osd->od_os, oid, 0);
+	rc = -zap_cursor_retrieve(zc, za);
+	if (rc == -ENOENT) {
+		zap_cursor_advance(zc);
+	} else if (rc) {
+		CERROR("%s: fail to init for check LMV "DFID"(%llu): rc = %d\n",
+		       osd_name(osd), PFID(fid), oid, rc);
+		GOTO(out_zc, rc);
+	}
+
+	while (1) {
+		rc = -zap_cursor_retrieve(zc, za);
+		if (rc == -ENOENT)
+			GOTO(out_zc, rc = 0);
+
+		if (rc) {
+			CERROR("%s: fail to locate next for check LMV "
+			       DFID"(%llu): rc = %d\n",
+			       osd_name(osd), PFID(fid), oid, rc);
+			GOTO(out_zc, rc);
+		}
+
+		fid_zero(tfid);
+		sscanf(za->za_name + 1, SFID, RFID(tfid));
+		if (fid_is_sane(tfid) && !osd_remote_fid(env, osd, tfid)) {
+			rc = osd_zap_lookup(osd, oid, NULL, za->za_name,
+					za->za_integer_length,
+					sizeof(*zde) / za->za_integer_length,
+					(void *)zde);
+			if (rc) {
+				CERROR("%s: fail to lookup for check LMV "
+				       DFID"(%llu): rc = %d\n",
+				       osd_name(osd), PFID(fid), oid, rc);
+				GOTO(out_zc, rc);
+			}
+
+			rc = osd_oii_insert(env, osd, tfid,
+					    zde->lzd_reg.zde_dnode, false);
+			GOTO(out_zc, rc);
+		}
+
+		zap_cursor_advance(zc);
+	}
+
+out_zc:
+	zap_cursor_fini(zc);
+out_nvbuf:
+	nvlist_free(nvbuf);
+
+	return rc;
+}
+
+static int
+osd_consistency_check(const struct lu_env *env, struct osd_device *osd,
+		      struct osd_object *obj, const struct lu_fid *fid,
+		      uint64_t oid, bool is_dir)
+{
+	struct lustre_scrub *scrub = &osd->od_scrub;
+	dnode_t *dn = NULL;
+	uint64_t oid2;
+	int once = 0;
+	bool insert;
+	int rc;
+	ENTRY;
+
+	if (!fid_is_norm(fid) && !fid_is_igif(fid))
+		RETURN(0);
+
+	/* oid == ZFS_NO_OBJECT must be for lookup ".." case */
+	if (oid == ZFS_NO_OBJECT) {
+		rc = osd_sa_handle_get(obj);
+		if (rc)
+			RETURN(rc);
+
+		rc = -sa_lookup(obj->oo_sa_hdl, SA_ZPL_PARENT(osd), &oid, 8);
+		if (rc)
+			RETURN(rc);
+	}
+
+	if (scrub->os_running) {
+		if (scrub->os_pos_current > oid)
+			RETURN(0);
+	} else if (osd->od_auto_scrub_interval == AS_NEVER) {
+		RETURN(0);
+	} else {
+		if (ktime_get_real_seconds() <
+		    scrub->os_file.sf_time_last_complete +
+		    osd->od_auto_scrub_interval)
+			RETURN(0);
+	}
+
+again:
+	rc = osd_fid_lookup(env, osd, fid, &oid2);
+	if (rc == -ENOENT) {
+		insert = true;
+		if (dn)
+			goto trigger;
+
+		rc = __osd_obj2dnode(osd->od_os, oid, &dn);
+		/* The object has been removed (by race maybe). */
+		if (rc)
+			RETURN(rc = (rc == -EEXIST ? -ENOENT : rc));
+
+		goto trigger;
+	} else if (rc || oid == oid2) {
+		GOTO(out, rc);
+	}
+
+	insert = false;
+
+trigger:
+	if (scrub->os_running) {
+		if (!dn) {
+			rc = __osd_obj2dnode(osd->od_os, oid, &dn);
+			/* The object has been removed (by race maybe). */
+			if (rc)
+				RETURN(rc = (rc == -EEXIST ? -ENOENT : rc));
+		}
+
+		rc = osd_oii_insert(env, osd, fid, oid, insert);
+		/* There is race condition between osd_oi_lookup and OI scrub.
+		 * The OI scrub finished just after osd_oi_lookup() failure.
+		 * Under such case, it is unnecessary to trigger OI scrub again,
+		 * but try to call osd_oi_lookup() again. */
+		if (unlikely(rc == -EAGAIN))
+			goto again;
+
+		if (is_dir)
+			rc = osd_check_lmv(env, osd, oid, fid);
+		else
+			rc = 0;
+
+		GOTO(out, rc);
+	}
+
+	if (osd->od_auto_scrub_interval != AS_NEVER && ++once == 1) {
+		rc = osd_scrub_start(env, osd, SS_AUTO_FULL |
+				     SS_CLEAR_DRYRUN | SS_CLEAR_FAILOUT);
+		CDEBUG(D_LFSCK | D_CONSOLE | D_WARNING,
+		       "%s: trigger partial OI scrub for RPC inconsistency "
+		       "checking FID "DFID": rc = %d\n",
+		       osd_name(osd), PFID(fid), rc);
+		if (!rc)
+			goto again;
+	}
+
+	GOTO(out, rc);
+
+out:
+	if (dn)
+		osd_dnode_rele(dn);
+
+	return rc;
+}
+
 static int osd_dir_lookup(const struct lu_env *env, struct dt_object *dt,
-			  struct dt_rec *rec, const struct dt_key *key,
-			  struct lustre_capa *capa)
+			  struct dt_rec *rec, const struct dt_key *key)
 {
 	struct osd_thread_info *oti = osd_oti_get(env);
-	struct osd_object  *obj = osd_dt_obj(dt);
-	struct osd_device  *osd = osd_obj2dev(obj);
-	char		   *name = (char *)key;
-	int                 rc;
+	struct osd_object *obj = osd_dt_obj(dt);
+	struct osd_device *osd = osd_obj2dev(obj);
+	struct lu_fid *fid = (struct lu_fid *)rec;
+	char *name = (char *)key;
+	uint64_t oid = ZFS_NO_OBJECT;
+	int rc;
 	ENTRY;
 
-	LASSERT(udmu_object_is_zap(obj->oo_db));
-
 	if (name[0] == '.') {
 		if (name[1] == 0) {
 			const struct lu_fid *f = lu_object_fid(&dt->do_lu);
 			memcpy(rec, f, sizeof(*f));
 			RETURN(1);
 		} else if (name[1] == '.' && name[2] == 0) {
-			rc = osd_find_parent_fid(env, dt, (struct lu_fid *)rec);
-			RETURN(rc == 0 ? 1 : rc);
+			rc = osd_find_parent_fid(env, dt, fid, &oid);
+			GOTO(out, rc);
 		}
 	}
 
-	rc = -zap_lookup(osd->od_objset.os, obj->oo_db->db_object,
-			 (char *)key, 8, sizeof(oti->oti_zde) / 8,
-			 (void *)&oti->oti_zde);
-	memcpy(rec, &oti->oti_zde.lzd_fid, sizeof(struct lu_fid));
+	memset(&oti->oti_zde.lzd_fid, 0, sizeof(struct lu_fid));
+	rc = osd_zap_lookup(osd, obj->oo_dn->dn_object, obj->oo_dn,
+			    (char *)key, 8, sizeof(oti->oti_zde) / 8,
+			    (void *)&oti->oti_zde);
+	if (rc != 0)
+		RETURN(rc);
 
-	RETURN(rc == 0 ? 1 : rc);
+	oid = oti->oti_zde.lzd_reg.zde_dnode;
+	if (likely(fid_is_sane(&oti->oti_zde.lzd_fid))) {
+		memcpy(rec, &oti->oti_zde.lzd_fid, sizeof(struct lu_fid));
+		GOTO(out, rc = 0);
+	}
+
+	rc = osd_get_fid_by_oid(env, osd, oti->oti_zde.lzd_reg.zde_dnode, fid);
+
+	GOTO(out, rc);
+
+out:
+	if (!rc && !osd_remote_fid(env, osd, fid)) {
+		/*
+		 * this should ask the scrubber to check OI given
+		 * the mapping we just found in the dir entry.
+		 * but result of that check should not affect
+		 * result of the lookup in the directory.
+		 * otherwise such a direntry becomes hidden
+		 * from the layers above, including LFSCK which
+		 * is supposed to fix dangling entries.
+		 */
+		osd_consistency_check(env, osd, obj, fid, oid,
+				S_ISDIR(DTTOIF(oti->oti_zde.lzd_reg.zde_type)));
+	}
+
+	return rc == 0 ? 1 : (rc == -ENOENT ? -ENODATA : rc);
 }
 
-static int osd_declare_dir_insert(const struct lu_env *env,
-				  struct dt_object *dt,
-				  const struct dt_rec *rec,
-				  const struct dt_key *key,
-				  struct thandle *th)
+/*
+ * In DNE environment, the object and its name entry may reside on different
+ * MDTs. Under such case, we will create an agent object on the MDT where the
+ * name entry resides. The agent object is empty, and indicates that the real
+ * object for the name entry resides on another MDT. If without agent object,
+ * related name entry will be skipped when perform MDT side file level backup
+ * and restore via ZPL by userspace tool, such as 'tar'.
+ */
+static int osd_create_agent_object(const struct lu_env *env,
+				   struct osd_device *osd,
+				   struct luz_direntry *zde,
+				   uint64_t parent, dmu_tx_t *tx)
 {
-	struct osd_object  *obj = osd_dt_obj(dt);
-	struct osd_thandle *oh;
+	struct osd_thread_info *info = osd_oti_get(env);
+	struct lustre_mdt_attrs *lma = &info->oti_mdt_attrs;
+	struct lu_attr *la = &info->oti_la;
+	nvlist_t *nvbuf = NULL;
+	dnode_t *dn = NULL;
+	sa_handle_t *hdl;
+	int rc = 0;
 	ENTRY;
 
-	LASSERT(th != NULL);
-	oh = container_of0(th, struct osd_thandle, ot_super);
+	if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_NO_AGENTOBJ))
+		RETURN(0);
 
-	LASSERT(obj->oo_db);
-	LASSERT(udmu_object_is_zap(obj->oo_db));
+	rc = -nvlist_alloc(&nvbuf, NV_UNIQUE_NAME, KM_SLEEP);
+	if (rc)
+		RETURN(rc);
 
-	dmu_tx_hold_bonus(oh->ot_tx, obj->oo_db->db_object);
-	dmu_tx_hold_zap(oh->ot_tx, obj->oo_db->db_object, TRUE, (char *)key);
+	lustre_lma_init(lma, &zde->lzd_fid, 0, LMAI_AGENT);
+	lustre_lma_swab(lma);
+	rc = -nvlist_add_byte_array(nvbuf, XATTR_NAME_LMA, (uchar_t *)lma,
+				    sizeof(*lma));
+	if (rc)
+		GOTO(out, rc);
 
-	RETURN(0);
+	la->la_valid = LA_TYPE | LA_MODE;
+	la->la_mode = (DTTOIF(zde->lzd_reg.zde_type) & S_IFMT) |
+			S_IRUGO | S_IWUSR | S_IXUGO;
+
+	if (S_ISDIR(la->la_mode))
+		rc = __osd_zap_create(env, osd, &dn, tx, la,
+				osd_find_dnsize(osd, OSD_BASE_EA_IN_BONUS), 0);
+	else
+		rc = __osd_object_create(env, osd, NULL, &zde->lzd_fid,
+					 &dn, tx, la);
+	if (rc)
+		GOTO(out, rc);
+
+	zde->lzd_reg.zde_dnode = dn->dn_object;
+	rc = -sa_handle_get(osd->od_os, dn->dn_object, NULL,
+			    SA_HDL_PRIVATE, &hdl);
+	if (!rc) {
+		rc = __osd_attr_init(env, osd, NULL, hdl, tx,
+				     la, parent, nvbuf);
+		sa_handle_destroy(hdl);
+	}
+
+	GOTO(out, rc);
+
+out:
+	if (dn) {
+		if (rc)
+			dmu_object_free(osd->od_os, dn->dn_object, tx);
+		osd_dnode_rele(dn);
+	}
+
+	if (nvbuf)
+		nvlist_free(nvbuf);
+
+	return rc;
 }
 
-/**
- * Find the osd object for given fid.
- *
- * \param fid need to find the osd object having this fid
- *
- * \retval osd_object on success
- * \retval        -ve on error
- */
-struct osd_object *osd_object_find(const struct lu_env *env,
-				   struct dt_object *dt,
-				   const struct lu_fid *fid)
+int osd_add_to_remote_parent(const struct lu_env *env,
+			     struct osd_device *osd,
+			     struct osd_object *obj,
+			     struct osd_thandle *oh)
 {
-	struct lu_device         *ludev = dt->do_lu.lo_dev;
-	struct osd_object        *child = NULL;
-	struct lu_object         *luch;
-	struct lu_object         *lo;
+	struct osd_thread_info *info = osd_oti_get(env);
+	struct luz_direntry *zde = &info->oti_zde;
+	char *name = info->oti_str;
+	const struct lu_fid *fid = lu_object_fid(&obj->oo_dt.do_lu);
+	struct lustre_mdt_attrs *lma = (struct lustre_mdt_attrs *)info->oti_buf;
+	struct lu_buf buf = {
+		.lb_buf = lma,
+		.lb_len = sizeof(info->oti_buf),
+	};
+	int size = 0;
+	int rc;
+	ENTRY;
 
-	/*
-	 * at this point topdev might not exist yet
-	 * (i.e. MGS is preparing profiles). so we can
-	 * not rely on topdev and instead lookup with
-	 * our device passed as topdev. this can't work
-	 * if the object isn't cached yet (as osd doesn't
-	 * allocate lu_header). IOW, the object must be
-	 * in the cache, otherwise lu_object_alloc() crashes
-	 * -bzzz
-	 */
-	luch = lu_object_find_at(env, ludev, fid, NULL);
-	if (IS_ERR(luch))
-		return (void *)luch;
-
-	if (lu_object_exists(luch)) {
-		lo = lu_object_locate(luch->lo_header, ludev->ld_type);
-		if (lo != NULL)
-			child = osd_obj(lo);
-		else
-			LU_OBJECT_DEBUG(D_ERROR, env, luch,
-					"%s: object can't be located "DFID"\n",
-					osd_dev(ludev)->od_svname, PFID(fid));
-
-		if (child == NULL) {
-			lu_object_put(env, luch);
-			CERROR("%s: Unable to get osd_object "DFID"\n",
-			       osd_dev(ludev)->od_svname, PFID(fid));
-			child = ERR_PTR(-ENOENT);
-		}
-	} else {
-		LU_OBJECT_DEBUG(D_ERROR, env, luch,
-				"%s: lu_object does not exists "DFID"\n",
-				osd_dev(ludev)->od_svname, PFID(fid));
-		lu_object_put(env, luch);
-		child = ERR_PTR(-ENOENT);
+	if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_NO_AGENTENT))
+		RETURN(0);
+
+	rc = osd_xattr_get_internal(env, obj, &buf, XATTR_NAME_LMA, &size);
+	if (rc) {
+		CWARN("%s: fail to load LMA for adding "
+		      DFID" to remote parent: rc = %d\n",
+		      osd_name(osd), PFID(fid), rc);
+		RETURN(rc);
 	}
 
-	return child;
+	lustre_lma_swab(lma);
+	lma->lma_incompat |= LMAI_REMOTE_PARENT;
+	lustre_lma_swab(lma);
+	buf.lb_len = size;
+	rc = osd_xattr_set_internal(env, obj, &buf, XATTR_NAME_LMA,
+				    LU_XATTR_REPLACE, oh);
+	if (rc) {
+		CWARN("%s: fail to update LMA for adding "
+		      DFID" to remote parent: rc = %d\n",
+		      osd_name(osd), PFID(fid), rc);
+		RETURN(rc);
+	}
+
+	osd_fid2str(name, fid, sizeof(info->oti_str));
+	zde->lzd_reg.zde_dnode = obj->oo_dn->dn_object;
+	zde->lzd_reg.zde_type = IFTODT(S_IFDIR);
+	zde->lzd_fid = *fid;
+
+	rc = osd_zap_add(osd, osd->od_remote_parent_dir, NULL,
+			 name, 8, sizeof(*zde) / 8, zde, oh->ot_tx);
+	if (unlikely(rc == -EEXIST))
+		rc = 0;
+	if (rc)
+		CWARN("%s: fail to add name entry for "
+		      DFID" to remote parent: rc = %d\n",
+		      osd_name(osd), PFID(fid), rc);
+	else
+		lu_object_set_agent_entry(&obj->oo_dt.do_lu);
+
+	RETURN(rc);
 }
 
-/**
- * Put the osd object once done with it.
- *
- * \param obj osd object that needs to be put
- */
-static inline void osd_object_put(const struct lu_env *env,
-				  struct osd_object *obj)
+int osd_delete_from_remote_parent(const struct lu_env *env,
+				  struct osd_device *osd,
+				  struct osd_object *obj,
+				  struct osd_thandle *oh, bool destroy)
+{
+	struct osd_thread_info *info = osd_oti_get(env);
+	char *name = info->oti_str;
+	const struct lu_fid *fid = lu_object_fid(&obj->oo_dt.do_lu);
+	struct lustre_mdt_attrs *lma = (struct lustre_mdt_attrs *)info->oti_buf;
+	struct lu_buf buf = {
+		.lb_buf = lma,
+		.lb_len = sizeof(info->oti_buf),
+	};
+	int size = 0;
+	int rc;
+	ENTRY;
+
+	osd_fid2str(name, fid, sizeof(info->oti_str));
+	rc = osd_zap_remove(osd, osd->od_remote_parent_dir, NULL,
+			    name, oh->ot_tx);
+	if (unlikely(rc == -ENOENT))
+		rc = 0;
+	if (rc)
+		CERROR("%s: fail to remove entry under remote "
+		       "parent for "DFID": rc = %d\n",
+		       osd_name(osd), PFID(fid), rc);
+
+	if (destroy || rc)
+		RETURN(rc);
+
+	rc = osd_xattr_get_internal(env, obj, &buf, XATTR_NAME_LMA, &size);
+	if (rc) {
+		CERROR("%s: fail to load LMA for removing "
+		       DFID" from remote parent: rc = %d\n",
+		       osd_name(osd), PFID(fid), rc);
+		RETURN(rc);
+	}
+
+	lustre_lma_swab(lma);
+	lma->lma_incompat &= ~LMAI_REMOTE_PARENT;
+	lustre_lma_swab(lma);
+	buf.lb_len = size;
+	rc = osd_xattr_set_internal(env, obj, &buf, XATTR_NAME_LMA,
+				    LU_XATTR_REPLACE, oh);
+	if (rc)
+		CERROR("%s: fail to update LMA for removing "
+		       DFID" from remote parent: rc = %d\n",
+		       osd_name(osd), PFID(fid), rc);
+	else
+		lu_object_clear_agent_entry(&obj->oo_dt.do_lu);
+
+	RETURN(rc);
+}
+
+static int osd_declare_dir_insert(const struct lu_env *env,
+				  struct dt_object *dt,
+				  const struct dt_rec *rec,
+				  const struct dt_key *key,
+				  struct thandle *th)
 {
-	lu_object_put(env, &obj->oo_dt.do_lu);
+	struct osd_object	*obj = osd_dt_obj(dt);
+	struct osd_device	*osd = osd_obj2dev(obj);
+	const struct dt_insert_rec *rec1;
+	const struct lu_fid	*fid;
+	struct osd_thandle	*oh;
+	uint64_t		 object;
+	struct osd_idmap_cache *idc;
+	ENTRY;
+
+	rec1 = (struct dt_insert_rec *)rec;
+	fid = rec1->rec_fid;
+	LASSERT(fid != NULL);
+	LASSERT(rec1->rec_type != 0);
+
+	LASSERT(th != NULL);
+	oh = container_of(th, struct osd_thandle, ot_super);
+
+	idc = osd_idc_find_or_init(env, osd, fid);
+	if (IS_ERR(idc))
+		RETURN(PTR_ERR(idc));
+
+	if (idc->oic_remote) {
+		const char *name = (const char *)key;
+
+		if (name[0] != '.' || name[1] != '.' || name[2] != 0) {
+			/* Prepare agent object for remote entry that will
+			 * be used for operations via ZPL, such as MDT side
+			 * file-level backup and restore. */
+			dmu_tx_hold_sa_create(oh->ot_tx,
+				osd_find_dnsize(osd, OSD_BASE_EA_IN_BONUS));
+			if (S_ISDIR(rec1->rec_type))
+				dmu_tx_hold_zap(oh->ot_tx, DMU_NEW_OBJECT,
+						FALSE, NULL);
+		}
+	}
+
+	/* This is for inserting dot/dotdot for new created dir. */
+	if (obj->oo_dn == NULL)
+		object = DMU_NEW_OBJECT;
+	else
+		object = obj->oo_dn->dn_object;
+
+	/* do not specify the key as then DMU is trying to look it up
+	 * which is very expensive. usually the layers above lookup
+	 * before insertion */
+	osd_tx_hold_zap(oh->ot_tx, object, obj->oo_dn, TRUE, NULL);
+
+	RETURN(0);
 }
 
 static int osd_seq_exists(const struct lu_env *env, struct osd_device *osd,
-			  obd_seq seq)
+			  u64 seq)
 {
 	struct lu_seq_range	*range = &osd_oti_get(env)->oti_seq_range;
 	struct seq_server_site	*ss = osd_seq_site(osd);
 	int			rc;
 	ENTRY;
 
-	if (ss == NULL)
-		RETURN(1);
+	LASSERT(ss != NULL);
+	LASSERT(ss->ss_server_fld != NULL);
 
 	rc = osd_fld_lookup(env, osd, seq, range);
 	if (rc != 0) {
-		CERROR("%s: Can not lookup fld for "LPX64"\n",
-		       osd_name(osd), seq);
+		if (rc != -ENOENT)
+			CERROR("%s: Can not lookup fld for %#llx\n",
+			       osd_name(osd), seq);
 		RETURN(0);
 	}
 
 	RETURN(ss->ss_node_id == range->lsr_index);
 }
 
-static int osd_remote_fid(const struct lu_env *env, struct osd_device *osd,
-			  struct lu_fid *fid)
+int osd_remote_fid(const struct lu_env *env, struct osd_device *osd,
+		   const struct lu_fid *fid)
 {
+	struct seq_server_site	*ss = osd_seq_site(osd);
 	ENTRY;
 
 	/* FID seqs not in FLDB, must be local seq */
 	if (unlikely(!fid_seq_in_fldb(fid_seq(fid))))
 		RETURN(0);
 
+	/* If FLD is not being initialized yet, it only happens during the
+	 * initialization, likely during mgs initialization, and we assume
+	 * this is local FID. */
+	if (ss == NULL || ss->ss_server_fld == NULL)
+		RETURN(0);
+
+	/* Only check the local FLDB here */
 	if (osd_seq_exists(env, osd, fid_seq(fid)))
 		RETURN(0);
 
@@ -515,91 +994,120 @@ static int osd_remote_fid(const struct lu_env *env, struct osd_device *osd,
  *      \param  key     key for index
  *      \param  rec     record reference
  *      \param  th      transaction handler
- *      \param  capa    capability descriptor
- *      \param  ignore_quota update should not affect quota
  *
  *      \retval  0  success
  *      \retval -ve failure
  */
 static int osd_dir_insert(const struct lu_env *env, struct dt_object *dt,
 			  const struct dt_rec *rec, const struct dt_key *key,
-			  struct thandle *th, struct lustre_capa *capa,
-			  int ignore_quota)
+			  struct thandle *th)
 {
 	struct osd_thread_info *oti = osd_oti_get(env);
 	struct osd_object   *parent = osd_dt_obj(dt);
 	struct osd_device   *osd = osd_obj2dev(parent);
-	struct lu_fid       *fid = (struct lu_fid *)rec;
-	struct osd_thandle  *oh;
-	struct osd_object   *child = NULL;
-	__u32                attr;
-	char		    *name = (char *)key;
-	int                  rc;
+	struct dt_insert_rec *rec1 = (struct dt_insert_rec *)rec;
+	const struct lu_fid *fid = rec1->rec_fid;
+	struct osd_thandle *oh;
+	struct osd_idmap_cache *idc;
+	const char *name = (const char *)key;
+	struct luz_direntry *zde = &oti->oti_zde;
+	int num = sizeof(*zde) / 8;
+	int rc;
 	ENTRY;
 
-	LASSERT(parent->oo_db);
-	LASSERT(udmu_object_is_zap(parent->oo_db));
+	LASSERT(parent->oo_dn);
 
 	LASSERT(dt_object_exists(dt));
 	LASSERT(osd_invariant(parent));
 
 	LASSERT(th != NULL);
-	oh = container_of0(th, struct osd_thandle, ot_super);
-
-	rc = osd_remote_fid(env, osd, fid);
-	if (rc < 0) {
-		CERROR("%s: Can not find object "DFID": rc = %d\n",
-		       osd->od_svname, PFID(fid), rc);
-		RETURN(rc);
+	oh = container_of(th, struct osd_thandle, ot_super);
+
+	idc = osd_idc_find(env, osd, fid);
+	if (unlikely(idc == NULL)) {
+		/* this dt_insert() wasn't declared properly, so
+		 * FID is missing in OI cache. we better do not
+		 * lookup FID in FLDB/OI and don't risk to deadlock,
+		 * but in some special cases (lfsck testing, etc)
+		 * it's much simpler than fixing a caller */
+		idc = osd_idc_find_or_init(env, osd, fid);
+		if (IS_ERR(idc)) {
+			CERROR("%s: "DFID" wasn't declared for insert\n",
+			       osd_name(osd), PFID(fid));
+			RETURN(PTR_ERR(idc));
+		}
 	}
 
-	if (unlikely(rc == 1)) {
-		/* Insert remote entry */
-		memset(&oti->oti_zde.lzd_reg, 0, sizeof(oti->oti_zde.lzd_reg));
-		oti->oti_zde.lzd_reg.zde_type = IFTODT(S_IFDIR & S_IFMT);
+	BUILD_BUG_ON(sizeof(zde->lzd_reg) != 8);
+	BUILD_BUG_ON(sizeof(*zde) % 8 != 0);
+
+	memset(&zde->lzd_reg, 0, sizeof(zde->lzd_reg));
+	zde->lzd_reg.zde_type = IFTODT(rec1->rec_type & S_IFMT);
+	zde->lzd_fid = *fid;
+
+	if (idc->oic_remote) {
+		if (name[0] != '.' || name[1] != '.' || name[2] != 0) {
+			/* Create agent inode for remote object that will
+			 * be used for MDT file-level backup and restore. */
+			rc = osd_create_agent_object(env, osd, zde,
+					parent->oo_dn->dn_object, oh->ot_tx);
+			if (rc) {
+				CWARN("%s: Fail to create agent object for "
+				      DFID": rc = %d\n",
+				      osd_name(osd), PFID(fid), rc);
+				/* Ignore the failure since the system can go
+				 * ahead if we do not care about the MDT side
+				 * file-level backup and restore. */
+				rc = 0;
+			}
+		}
 	} else {
-		/*
-		 * To simulate old Orion setups with ./..  stored in the
-		 * directories
-		 */
-		/* Insert local entry */
-		child = osd_object_find(env, dt, fid);
-		if (IS_ERR(child))
-			RETURN(PTR_ERR(child));
-
-		LASSERT(child->oo_db);
+		if (unlikely(idc->oic_dnode == 0)) {
+			/* for a reason OI cache wasn't filled properly */
+			CERROR("%s: OIC for "DFID" isn't filled\n",
+			       osd_name(osd), PFID(fid));
+			RETURN(-EINVAL);
+		}
 		if (name[0] == '.') {
 			if (name[1] == 0) {
 				/* do not store ".", instead generate it
 				 * during iteration */
 				GOTO(out, rc = 0);
 			} else if (name[1] == '.' && name[2] == 0) {
+				uint64_t dnode = idc->oic_dnode;
+				if (OBD_FAIL_CHECK(OBD_FAIL_LFSCK_BAD_PARENT))
+					dnode--;
+
 				/* update parent dnode in the child.
 				 * later it will be used to generate ".." */
-				udmu_objset_t *uos = &osd->od_objset;
 				rc = osd_object_sa_update(parent,
-						 SA_ZPL_PARENT(uos),
-						 &child->oo_db->db_object,
-						 8, oh);
+						 SA_ZPL_PARENT(osd),
+						 &dnode, 8, oh);
+
 				GOTO(out, rc);
 			}
 		}
-		CLASSERT(sizeof(oti->oti_zde.lzd_reg) == 8);
-		CLASSERT(sizeof(oti->oti_zde) % 8 == 0);
-		attr = child->oo_dt.do_lu.lo_header ->loh_attr;
-		oti->oti_zde.lzd_reg.zde_type = IFTODT(attr & S_IFMT);
-		oti->oti_zde.lzd_reg.zde_dnode = child->oo_db->db_object;
+		zde->lzd_reg.zde_dnode = idc->oic_dnode;
 	}
 
-	oti->oti_zde.lzd_fid = *fid;
+	if (OBD_FAIL_CHECK(OBD_FAIL_FID_INDIR))
+		zde->lzd_fid.f_ver = ~0;
+
+	/* The logic is not related with IGIF, just re-use the fail_loc value
+	 * to be consistent with ldiskfs case, then share the same test logic */
+	if (OBD_FAIL_CHECK(OBD_FAIL_FID_IGIF))
+		num = 1;
+
 	/* Insert (key,oid) into ZAP */
-	rc = -zap_add(osd->od_objset.os, parent->oo_db->db_object,
-		      (char *)key, 8, sizeof(oti->oti_zde) / 8,
-		      (void *)&oti->oti_zde, oh->ot_tx);
+	rc = osd_zap_add(osd, parent->oo_dn->dn_object, parent->oo_dn,
+			 name, 8, num, (void *)zde, oh->ot_tx);
+	if (unlikely(rc == -EEXIST &&
+		     name[0] == '.' && name[1] == '.' && name[2] == 0))
+		/* Update (key,oid) in ZAP */
+		rc = -zap_update(osd->od_os, parent->oo_dn->dn_object, name, 8,
+				 sizeof(*zde) / 8, (void *)zde, oh->ot_tx);
 
 out:
-	if (child != NULL)
-		osd_object_put(env, child);
 
 	RETURN(rc);
 }
@@ -610,40 +1118,56 @@ static int osd_declare_dir_delete(const struct lu_env *env,
 				  struct thandle *th)
 {
 	struct osd_object *obj = osd_dt_obj(dt);
+	dnode_t *zap_dn = obj->oo_dn;
 	struct osd_thandle *oh;
+	const char *name = (const char *)key;
 	ENTRY;
 
 	LASSERT(dt_object_exists(dt));
 	LASSERT(osd_invariant(obj));
+	LASSERT(zap_dn != NULL);
 
 	LASSERT(th != NULL);
-	oh = container_of0(th, struct osd_thandle, ot_super);
+	oh = container_of(th, struct osd_thandle, ot_super);
+
+	/*
+	 * In Orion . and .. were stored in the directory (not generated upon
+	 * request as now). We preserve them for backward compatibility.
+	 */
+	if (name[0] == '.') {
+		if (name[1] == 0)
+			RETURN(0);
+		else if (name[1] == '.' && name[2] == 0)
+			RETURN(0);
+	}
 
-	LASSERT(obj->oo_db);
-	LASSERT(udmu_object_is_zap(obj->oo_db));
+	/* do not specify the key as then DMU is trying to look it up
+	 * which is very expensive. usually the layers above lookup
+	 * before deletion */
+	osd_tx_hold_zap(oh->ot_tx, zap_dn->dn_object, zap_dn, FALSE, NULL);
 
-	dmu_tx_hold_zap(oh->ot_tx, obj->oo_db->db_object, TRUE, (char *)key);
+	/* For destroying agent object if have. */
+	dmu_tx_hold_bonus(oh->ot_tx, DMU_NEW_OBJECT);
 
 	RETURN(0);
 }
 
 static int osd_dir_delete(const struct lu_env *env, struct dt_object *dt,
-			  const struct dt_key *key, struct thandle *th,
-			  struct lustre_capa *capa)
+			  const struct dt_key *key, struct thandle *th)
 {
+	struct luz_direntry *zde = &osd_oti_get(env)->oti_zde;
 	struct osd_object *obj = osd_dt_obj(dt);
 	struct osd_device *osd = osd_obj2dev(obj);
 	struct osd_thandle *oh;
-	dmu_buf_t *zap_db = obj->oo_db;
+	dnode_t *zap_dn = obj->oo_dn;
 	char	  *name = (char *)key;
 	int rc;
 	ENTRY;
 
-	LASSERT(obj->oo_db);
-	LASSERT(udmu_object_is_zap(obj->oo_db));
+	LASSERT(zap_dn);
 
 	LASSERT(th != NULL);
-	oh = container_of0(th, struct osd_thandle, ot_super);
+	oh = container_of(th, struct osd_thandle, ot_super);
 
 	/*
 	 * In Orion . and .. were stored in the directory (not generated upon
@@ -657,31 +1181,54 @@ static int osd_dir_delete(const struct lu_env *env, struct dt_object *dt,
 		}
 	}
 
-	/* Remove key from the ZAP */
-	rc = -zap_remove(osd->od_objset.os, zap_db->db_object,
-			 (char *) key, oh->ot_tx);
+	/* XXX: We have to say that lookup during delete_declare will affect
+	 *	performance, but we have to check whether the name entry (to
+	 *	be deleted) has agent object or not to avoid orphans.
+	 *
+	 *	We will improve that in the future, some possible solutions,
+	 *	for example:
+	 *	1) Some hint from the caller via transaction handle to make
+	 *	   the lookup conditionally.
+	 *	2) Enhance the ZFS logic to recognize the OSD lookup result
+	 *	   and delete the given entry directly without lookup again
+	 *	   internally. LU-10190 */
+	memset(&zde->lzd_fid, 0, sizeof(zde->lzd_fid));
+	rc = osd_zap_lookup(osd, zap_dn->dn_object, zap_dn, name, 8, 3, zde);
+	if (unlikely(rc)) {
+		if (rc != -ENOENT)
+			CERROR("%s: failed to locate entry  %s: rc = %d\n",
+			       osd->od_svname, name, rc);
+		RETURN(rc);
+	}
 
-#if LUSTRE_VERSION_CODE <= OBD_OCD_VERSION(2, 4, 53, 0)
-	if (unlikely(rc == -ENOENT && name[0] == '.' &&
-	    (name[1] == 0 || (name[1] == '.' && name[2] == 0))))
-		rc = 0;
-#endif
-	if (unlikely(rc && rc != -ENOENT))
-		CERROR("%s: zap_remove failed: rc = %d\n", osd->od_svname, rc);
+	if (unlikely(osd_remote_fid(env, osd, &zde->lzd_fid) > 0)) {
+		rc = -dmu_object_free(osd->od_os, zde->lzd_reg.zde_dnode,
+				      oh->ot_tx);
+		if (rc)
+			CERROR("%s: failed to destroy agent object (%llu) "
+			       "for the entry %s: rc = %d\n", osd->od_svname,
+			       (__u64)zde->lzd_reg.zde_dnode, name, rc);
+	}
+
+	/* Remove key from the ZAP */
+	rc = osd_zap_remove(osd, zap_dn->dn_object, zap_dn,
+			    (char *)key, oh->ot_tx);
+	if (unlikely(rc))
+		CERROR("%s: zap_remove %s failed: rc = %d\n",
+		       osd->od_svname, name, rc);
 
 	RETURN(rc);
 }
 
 static struct dt_it *osd_dir_it_init(const struct lu_env *env,
 				     struct dt_object *dt,
-				     __u32 unused,
-				     struct lustre_capa *capa)
+				     __u32 unused)
 {
 	struct osd_zap_it *it;
 
-	it = (struct osd_zap_it *)osd_index_it_init(env, dt, unused, capa);
+	it = (struct osd_zap_it *)osd_index_it_init(env, dt, unused);
 	if (!IS_ERR(it))
-		it->ozi_pos = 0;
+		it->ozi_pos = OZI_POS_INIT;
 
 	RETURN((struct dt_it *)it);
 }
@@ -701,7 +1248,6 @@ static int osd_dir_it_get(const struct lu_env *env,
 {
 	struct osd_zap_it *it = (struct osd_zap_it *)di;
 	struct osd_object *obj = it->ozi_obj;
-	struct osd_device *osd = osd_obj2dev(obj);
 	char		  *name = (char *)key;
 	int		   rc;
 	ENTRY;
@@ -709,32 +1255,30 @@ static int osd_dir_it_get(const struct lu_env *env,
 	LASSERT(it);
 	LASSERT(it->ozi_zc);
 
-	udmu_zap_cursor_fini(it->ozi_zc);
-
-	if (udmu_zap_cursor_init(&it->ozi_zc, &osd->od_objset,
-				 obj->oo_db->db_object, 0))
-		RETURN(-ENOMEM);
+	/* reset the cursor */
+	zap_cursor_fini(it->ozi_zc);
+	osd_obj_cursor_init_serialized(it->ozi_zc, obj, 0);
 
 	/* XXX: implementation of the API is broken at the moment */
 	LASSERT(((const char *)key)[0] == 0);
 
 	if (name[0] == 0) {
-		it->ozi_pos = 0;
+		it->ozi_pos = OZI_POS_INIT;
 		RETURN(1);
 	}
 
 	if (name[0] == '.') {
 		if (name[1] == 0) {
-			it->ozi_pos = 1;
+			it->ozi_pos = OZI_POS_DOT;
 			GOTO(out, rc = 1);
 		} else if (name[1] == '.' && name[2] == 0) {
-			it->ozi_pos = 2;
+			it->ozi_pos = OZI_POS_DOTDOT;
 			GOTO(out, rc = 1);
 		}
 	}
 
 	/* neither . nor .. - some real record */
-	it->ozi_pos = 3;
+	it->ozi_pos = OZI_POS_REAL;
 	rc = +1;
 
 out:
@@ -792,21 +1336,24 @@ static int osd_dir_it_next(const struct lu_env *env, struct dt_it *di)
 	zap_attribute_t	  *za = &osd_oti_get(env)->oti_za;
 	int		   rc;
 
+	ENTRY;
+
 	/* temp. storage should be enough for any key supported by ZFS */
-	CLASSERT(sizeof(za->za_name) <= sizeof(it->ozi_name));
+	BUILD_BUG_ON(sizeof(za->za_name) > sizeof(it->ozi_name));
 
 	/*
 	 * the first ->next() moves the cursor to .
 	 * the second ->next() moves the cursor to ..
 	 * then we get to the real records and have to verify any exist
 	 */
-	if (it->ozi_pos <= 2) {
+	if (it->ozi_pos <= OZI_POS_DOTDOT) {
 		it->ozi_pos++;
-		if (it->ozi_pos <=2)
+		if (it->ozi_pos <= OZI_POS_DOTDOT)
 			RETURN(0);
-	}
 
-	zap_cursor_advance(it->ozi_zc);
+	} else {
+		zap_cursor_advance(it->ozi_zc);
+	}
 
 	/*
 	 * According to current API we need to return error if its last entry.
@@ -830,10 +1377,10 @@ static struct dt_key *osd_dir_it_key(const struct lu_env *env,
 	int		   rc = 0;
 	ENTRY;
 
-	if (it->ozi_pos <= 1) {
-		it->ozi_pos = 1;
+	if (it->ozi_pos <= OZI_POS_DOT) {
+		it->ozi_pos = OZI_POS_DOT;
 		RETURN((struct dt_key *)".");
-	} else if (it->ozi_pos == 2) {
+	} else if (it->ozi_pos == OZI_POS_DOTDOT) {
 		RETURN((struct dt_key *)"..");
 	}
 
@@ -842,19 +1389,6 @@ static struct dt_key *osd_dir_it_key(const struct lu_env *env,
 
 	strcpy(it->ozi_name, za->za_name);
 
-#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 3, 90, 0)
-	if (za->za_name[0] == '.') {
-		if (za->za_name[1] == 0 || (za->za_name[1] == '.' &&
-		    za->za_name[2] == 0)) {
-			/* we should not get onto . and ..
-			 * stored in the directory. ->next() and
-			 * other methods should prevent this
-			 */
-			LBUG();
-		}
-	}
-#endif
-
 	RETURN((struct dt_key *)it->ozi_name);
 }
 
@@ -865,107 +1399,255 @@ static int osd_dir_it_key_size(const struct lu_env *env, const struct dt_it *di)
 	int		   rc;
 	ENTRY;
 
-	if (it->ozi_pos <= 1) {
-		it->ozi_pos = 1;
+	if (it->ozi_pos <= OZI_POS_DOT) {
+		it->ozi_pos = OZI_POS_DOT;
 		RETURN(2);
-	} else if (it->ozi_pos == 2) {
+	} else if (it->ozi_pos == OZI_POS_DOTDOT) {
 		RETURN(3);
 	}
 
 	if ((rc = -zap_cursor_retrieve(it->ozi_zc, za)) == 0)
 		rc = strlen(za->za_name);
 
-#if LUSTRE_VERSION_CODE < OBD_OCD_VERSION(2, 3, 90, 0)
-	if (rc == 0 && za->za_name[0] == '.') {
-		if (za->za_name[1] == 0 || (za->za_name[1] == '.' &&
-		    za->za_name[2] == 0)) {
-			/* we should not get onto . and ..
-			 * stored in the directory. ->next() and
-			 * other methods should prevent this
-			 */
-			LBUG();
-		}
-	}
-#endif
 	RETURN(rc);
 }
 
+static int
+osd_dirent_update(const struct lu_env *env, struct osd_device *dev,
+		  uint64_t zap, const char *key, struct luz_direntry *zde)
+{
+	dmu_tx_t *tx;
+	int rc;
+	ENTRY;
+
+	tx = dmu_tx_create(dev->od_os);
+	if (!tx)
+		RETURN(-ENOMEM);
+
+	dmu_tx_hold_zap(tx, zap, TRUE, NULL);
+	rc = -dmu_tx_assign(tx, TXG_WAIT);
+	if (!rc)
+		rc = -zap_update(dev->od_os, zap, key, 8, sizeof(*zde) / 8,
+				 (const void *)zde, tx);
+	if (rc)
+		dmu_tx_abort(tx);
+	else
+		dmu_tx_commit(tx);
+
+	RETURN(rc);
+}
+
+static int osd_update_entry_for_agent(const struct lu_env *env,
+				      struct osd_device *osd,
+				      uint64_t zap, const char *name,
+				      struct luz_direntry *zde, __u32 attr)
+{
+	dmu_tx_t *tx = NULL;
+	int rc = 0;
+	ENTRY;
+
+	if (attr & LUDA_VERIFY_DRYRUN)
+		GOTO(out, rc = 0);
+
+	tx = dmu_tx_create(osd->od_os);
+	if (!tx)
+		GOTO(out, rc = -ENOMEM);
+
+	dmu_tx_hold_sa_create(tx, osd_find_dnsize(osd, OSD_BASE_EA_IN_BONUS));
+	dmu_tx_hold_zap(tx, zap, FALSE, NULL);
+	rc = -dmu_tx_assign(tx, TXG_WAIT);
+	if (rc) {
+		dmu_tx_abort(tx);
+		GOTO(out, rc);
+	}
+
+	rc = osd_create_agent_object(env, osd, zde, zap, tx);
+	if (!rc)
+		rc = -zap_update(osd->od_os, zap, name, 8, sizeof(*zde) / 8,
+				 (const void *)zde, tx);
+	dmu_tx_commit(tx);
+
+	GOTO(out, rc);
+
+out:
+	CDEBUG(D_LFSCK, "%s: Updated (%s) remote entry for "DFID": rc = %d\n",
+	       osd_name(osd), (attr & LUDA_VERIFY_DRYRUN) ? "(ro)" : "(rw)",
+	       PFID(&zde->lzd_fid), rc);
+	return rc;
+}
+
 static int osd_dir_it_rec(const struct lu_env *env, const struct dt_it *di,
 			  struct dt_rec *dtrec, __u32 attr)
 {
-	struct osd_zap_it   *it = (struct osd_zap_it *)di;
-	struct lu_dirent    *lde = (struct lu_dirent *)dtrec;
-	struct luz_direntry *zde = &osd_oti_get(env)->oti_zde;
-	zap_attribute_t     *za = &osd_oti_get(env)->oti_za;
-	int		     rc, namelen;
+	struct osd_zap_it *it = (struct osd_zap_it *)di;
+	struct lu_dirent *lde = (struct lu_dirent *)dtrec;
+	struct osd_thread_info *info = osd_oti_get(env);
+	struct luz_direntry *zde = &info->oti_zde;
+	zap_attribute_t *za = &info->oti_za;
+	struct lu_fid *fid = &info->oti_fid;
+	struct osd_device *osd = osd_obj2dev(it->ozi_obj);
+	int rc, namelen;
 	ENTRY;
 
-	if (it->ozi_pos <= 1) {
-		lde->lde_hash = cpu_to_le64(1);
+	lde->lde_attrs = 0;
+	if (it->ozi_pos <= OZI_POS_DOT) {
+		/* notice hash=0 here, this is needed to avoid
+		 * case when some real entry (after ./..) may
+		 * have hash=0. in this case the client would
+		 * be confused having records out of hash order. */
+		lde->lde_hash = cpu_to_le64(0);
 		strcpy(lde->lde_name, ".");
 		lde->lde_namelen = cpu_to_le16(1);
-		lde->lde_fid = *lu_object_fid(&it->ozi_obj->oo_dt.do_lu);
+		fid_cpu_to_le(&lde->lde_fid,
+			      lu_object_fid(&it->ozi_obj->oo_dt.do_lu));
 		lde->lde_attrs = LUDA_FID;
 		/* append lustre attributes */
 		osd_it_append_attrs(lde, attr, 1, IFTODT(S_IFDIR));
 		lde->lde_reclen = cpu_to_le16(lu_dirent_calc_size(1, attr));
-		it->ozi_pos = 1;
-		GOTO(out, rc = 0);
-
-	} else if (it->ozi_pos == 2) {
-		lde->lde_hash = cpu_to_le64(2);
+		it->ozi_pos = OZI_POS_DOT;
+		RETURN(0);
+	} else if (it->ozi_pos == OZI_POS_DOTDOT) {
+		/* same as for . above */
+		lde->lde_hash = cpu_to_le64(0);
 		strcpy(lde->lde_name, "..");
 		lde->lde_namelen = cpu_to_le16(2);
-		lde->lde_attrs = LUDA_FID;
+		rc = osd_find_parent_fid(env, &it->ozi_obj->oo_dt, fid, NULL);
+		if (!rc) {
+			fid_cpu_to_le(&lde->lde_fid, fid);
+			lde->lde_attrs = LUDA_FID;
+		} else if (rc != -ENOENT) {
+			/* ENOENT happens at the root of filesystem, ignore */
+			RETURN(rc);
+		}
+
 		/* append lustre attributes */
 		osd_it_append_attrs(lde, attr, 2, IFTODT(S_IFDIR));
 		lde->lde_reclen = cpu_to_le16(lu_dirent_calc_size(2, attr));
-		rc = osd_find_parent_fid(env, &it->ozi_obj->oo_dt, &lde->lde_fid);
-		/*
-		 * early Orion code was not setting LinkEA, so it's possible
-		 * some setups still have objects with no LinkEA set.
-		 * but at that time .. was a real record in the directory
-		 * so we should try to lookup .. in ZAP
-		 */
-		if (rc != -ENOENT)
-			GOTO(out, rc);
+		RETURN(0);
 	}
 
 	LASSERT(lde);
 
 	rc = -zap_cursor_retrieve(it->ozi_zc, za);
-	if (unlikely(rc != 0))
-		GOTO(out, rc);
+	if (unlikely(rc))
+		RETURN(rc);
 
-	lde->lde_hash = cpu_to_le64(udmu_zap_cursor_serialize(it->ozi_zc));
+	lde->lde_hash = cpu_to_le64(osd_zap_cursor_serialize(it->ozi_zc));
 	namelen = strlen(za->za_name);
 	if (namelen > NAME_MAX)
-		GOTO(out, rc = -EOVERFLOW);
+		RETURN(-EOVERFLOW);
 	strcpy(lde->lde_name, za->za_name);
 	lde->lde_namelen = cpu_to_le16(namelen);
 
-	if (za->za_integer_length != 8 || za->za_num_integers < 3) {
+	if (za->za_integer_length != 8) {
 		CERROR("%s: unsupported direntry format: %d %d\n",
-		       osd_obj2dev(it->ozi_obj)->od_svname,
+		       osd->od_svname,
 		       za->za_integer_length, (int)za->za_num_integers);
-
-		GOTO(out, rc = -EIO);
+		RETURN(-EIO);
 	}
 
-	rc = -zap_lookup(it->ozi_zc->zc_objset, it->ozi_zc->zc_zapobj,
-			 za->za_name, za->za_integer_length, 3, zde);
+	rc = osd_zap_lookup(osd, it->ozi_zc->zc_zapobj, it->ozi_obj->oo_dn,
+			    za->za_name, za->za_integer_length, 3, zde);
 	if (rc)
-		GOTO(out, rc);
+		RETURN(rc);
+
+	if (za->za_num_integers >= 3 && fid_is_sane(&zde->lzd_fid)) {
+		lde->lde_attrs = LUDA_FID;
+		fid_cpu_to_le(&lde->lde_fid, &zde->lzd_fid);
+		if (unlikely(zde->lzd_reg.zde_dnode == ZFS_NO_OBJECT &&
+			     osd_remote_fid(env, osd, &zde->lzd_fid) > 0 &&
+			     attr & LUDA_VERIFY)) {
+			/* It is mainly used for handling the MDT
+			 * upgraded from old ZFS based backend. */
+			rc = osd_update_entry_for_agent(env, osd,
+					it->ozi_obj->oo_dn->dn_object,
+					za->za_name, zde, attr);
+			if (!rc)
+				lde->lde_attrs |= LUDA_REPAIR;
+			else
+				lde->lde_attrs |= LUDA_UNKNOWN;
+		}
+
+		GOTO(pack_attr, rc = 0);
+	}
+
+	if (OBD_FAIL_CHECK(OBD_FAIL_FID_LOOKUP))
+		RETURN(-ENOENT);
+
+	rc = osd_get_fid_by_oid(env, osd, zde->lzd_reg.zde_dnode, fid);
+	if (rc) {
+		lde->lde_attrs = LUDA_UNKNOWN;
+		GOTO(pack_attr, rc = 0);
+	}
+
+	if (!(attr & LUDA_VERIFY)) {
+		fid_cpu_to_le(&lde->lde_fid, fid);
+		lde->lde_attrs = LUDA_FID;
+		GOTO(pack_attr, rc = 0);
+	}
+
+	if (attr & LUDA_VERIFY_DRYRUN) {
+		fid_cpu_to_le(&lde->lde_fid, fid);
+		lde->lde_attrs = LUDA_FID | LUDA_REPAIR;
+		GOTO(pack_attr, rc = 0);
+	}
 
-	lde->lde_fid = zde->lzd_fid;
+	fid_cpu_to_le(&lde->lde_fid, fid);
 	lde->lde_attrs = LUDA_FID;
+	zde->lzd_fid = *fid;
+	rc = osd_dirent_update(env, osd, it->ozi_zc->zc_zapobj,
+			       za->za_name, zde);
+	if (rc) {
+		lde->lde_attrs |= LUDA_UNKNOWN;
+		GOTO(pack_attr, rc = 0);
+	}
 
-	/* append lustre attributes */
-	osd_it_append_attrs(lde, attr, namelen, zde->lzd_reg.zde_type);
+	lde->lde_attrs |= LUDA_REPAIR;
+
+	GOTO(pack_attr, rc = 0);
 
+pack_attr:
+	osd_it_append_attrs(lde, attr, namelen, zde->lzd_reg.zde_type);
 	lde->lde_reclen = cpu_to_le16(lu_dirent_calc_size(namelen, attr));
+	return rc;
+}
+
+static int osd_dir_it_rec_size(const struct lu_env *env, const struct dt_it *di,
+			       __u32 attr)
+{
+	struct osd_zap_it   *it = (struct osd_zap_it *)di;
+	zap_attribute_t     *za = &osd_oti_get(env)->oti_za;
+	size_t		     namelen = 0;
+	int		     rc;
+	ENTRY;
+
+	if (it->ozi_pos <= OZI_POS_DOT)
+		namelen = 1;
+	else if (it->ozi_pos == OZI_POS_DOTDOT)
+		namelen = 2;
+
+	if (namelen > 0) {
+		rc = lu_dirent_calc_size(namelen, attr);
+		RETURN(rc);
+	}
+
+	rc = -zap_cursor_retrieve(it->ozi_zc, za);
+	if (unlikely(rc != 0))
+		RETURN(rc);
+
+	if (za->za_integer_length != 8 || za->za_num_integers < 3) {
+		CERROR("%s: unsupported direntry format: %d %d\n",
+		       osd_obj2dev(it->ozi_obj)->od_svname,
+		       za->za_integer_length, (int)za->za_num_integers);
+		RETURN(-EIO);
+	}
+
+	namelen = strlen(za->za_name);
+	if (namelen > NAME_MAX)
+		RETURN(-EOVERFLOW);
+
+	rc = lu_dirent_calc_size(namelen, attr);
 
-out:
 	RETURN(rc);
 }
 
@@ -975,10 +1657,10 @@ static __u64 osd_dir_it_store(const struct lu_env *env, const struct dt_it *di)
 	__u64		   pos;
 	ENTRY;
 
-	if (it->ozi_pos <= 2)
-		pos = it->ozi_pos;
+	if (it->ozi_pos <= OZI_POS_DOTDOT)
+		pos = 0;
 	else
-		pos = udmu_zap_cursor_serialize(it->ozi_zc);
+		pos = osd_zap_cursor_serialize(it->ozi_zc);
 
 	RETURN(pos);
 }
@@ -994,21 +1676,19 @@ static int osd_dir_it_load(const struct lu_env *env,
 {
 	struct osd_zap_it *it = (struct osd_zap_it *)di;
 	struct osd_object *obj = it->ozi_obj;
-	struct osd_device *osd = osd_obj2dev(obj);
 	zap_attribute_t   *za = &osd_oti_get(env)->oti_za;
 	int		   rc;
 	ENTRY;
 
-	udmu_zap_cursor_fini(it->ozi_zc);
-	if (udmu_zap_cursor_init(&it->ozi_zc, &osd->od_objset,
-				 obj->oo_db->db_object, hash))
-		RETURN(-ENOMEM);
+	/* reset the cursor */
+	zap_cursor_fini(it->ozi_zc);
+	osd_obj_cursor_init_serialized(it->ozi_zc, obj, hash);
 
-	if (hash <= 2) {
-		it->ozi_pos = hash;
-		rc = +1;
+	if (hash == 0) {
+		it->ozi_pos = OZI_POS_INIT;
+		rc = +1; /* there will be ./.. at least */
 	} else {
-		it->ozi_pos = 3;
+		it->ozi_pos = OZI_POS_REAL;
 		/* to return whether the end has been reached */
 		rc = osd_index_retrieve_skip_dots(it, za);
 		if (rc == 0)
@@ -1020,7 +1700,7 @@ static int osd_dir_it_load(const struct lu_env *env,
 	RETURN(rc);
 }
 
-static struct dt_index_operations osd_dir_ops = {
+struct dt_index_operations osd_dir_ops = {
 	.dio_lookup         = osd_dir_lookup,
 	.dio_declare_insert = osd_declare_dir_insert,
 	.dio_insert         = osd_dir_insert,
@@ -1035,6 +1715,7 @@ static struct dt_index_operations osd_dir_ops = {
 		.key      = osd_dir_it_key,
 		.key_size = osd_dir_it_key_size,
 		.rec      = osd_dir_it_rec,
+		.rec_size = osd_dir_it_rec_size,
 		.store    = osd_dir_it_store,
 		.load     = osd_dir_it_load
 	}
@@ -1067,8 +1748,7 @@ static int osd_prepare_key_uint64(struct osd_object *o, __u64 *dst,
 }
 
 static int osd_index_lookup(const struct lu_env *env, struct dt_object *dt,
-			struct dt_rec *rec, const struct dt_key *key,
-			struct lustre_capa *capa)
+			struct dt_rec *rec, const struct dt_key *key)
 {
 	struct osd_object *obj = osd_dt_obj(dt);
 	struct osd_device *osd = osd_obj2dev(obj);
@@ -1078,7 +1758,7 @@ static int osd_index_lookup(const struct lu_env *env, struct dt_object *dt,
 
 	rc = osd_prepare_key_uint64(obj, k, key);
 
-	rc = -zap_lookup_uint64(osd->od_objset.os, obj->oo_db->db_object,
+	rc = -zap_lookup_uint64(osd->od_os, obj->oo_dn->dn_object,
 				k, rc, obj->oo_recusize, obj->oo_recsize,
 				(void *)rec);
 	RETURN(rc == 0 ? 1 : rc);
@@ -1095,24 +1775,22 @@ static int osd_declare_index_insert(const struct lu_env *env,
 	ENTRY;
 
 	LASSERT(th != NULL);
-	oh = container_of0(th, struct osd_thandle, ot_super);
+	oh = container_of(th, struct osd_thandle, ot_super);
 
-	LASSERT(obj->oo_db);
+	LASSERT(obj->oo_dn);
 
-	dmu_tx_hold_bonus(oh->ot_tx, obj->oo_db->db_object);
-
-	/* It is not clear what API should be used for binary keys, so we pass
-	 * a null name which has the side effect of over-reserving space,
-	 * accounting for the worst case. See zap_count_write() */
-	dmu_tx_hold_zap(oh->ot_tx, obj->oo_db->db_object, TRUE, NULL);
+	/* do not specify the key as then DMU is trying to look it up
+	 * which is very expensive. usually the layers above lookup
+	 * before insertion */
+	osd_tx_hold_zap(oh->ot_tx, obj->oo_dn->dn_object, obj->oo_dn,
+			TRUE, NULL);
 
 	RETURN(0);
 }
 
 static int osd_index_insert(const struct lu_env *env, struct dt_object *dt,
 			    const struct dt_rec *rec, const struct dt_key *key,
-			    struct thandle *th, struct lustre_capa *capa,
-			    int ignore_quota)
+			    struct thandle *th)
 {
 	struct osd_object  *obj = osd_dt_obj(dt);
 	struct osd_device  *osd = osd_obj2dev(obj);
@@ -1121,17 +1799,17 @@ static int osd_index_insert(const struct lu_env *env, struct dt_object *dt,
 	int                 rc;
 	ENTRY;
 
-	LASSERT(obj->oo_db);
+	LASSERT(obj->oo_dn);
 	LASSERT(dt_object_exists(dt));
 	LASSERT(osd_invariant(obj));
 	LASSERT(th != NULL);
 
-	oh = container_of0(th, struct osd_thandle, ot_super);
+	oh = container_of(th, struct osd_thandle, ot_super);
 
 	rc = osd_prepare_key_uint64(obj, k, key);
 
 	/* Insert (key,oid) into ZAP */
-	rc = -zap_add_uint64(osd->od_objset.os, obj->oo_db->db_object,
+	rc = -zap_add_uint64(osd->od_os, obj->oo_dn->dn_object,
 			     k, rc, obj->oo_recusize, obj->oo_recsize,
 			     (void *)rec, oh->ot_tx);
 	RETURN(rc);
@@ -1149,17 +1827,21 @@ static int osd_declare_index_delete(const struct lu_env *env,
 	LASSERT(dt_object_exists(dt));
 	LASSERT(osd_invariant(obj));
 	LASSERT(th != NULL);
-	LASSERT(obj->oo_db);
+	LASSERT(obj->oo_dn);
+
+	oh = container_of(th, struct osd_thandle, ot_super);
 
-	oh = container_of0(th, struct osd_thandle, ot_super);
-	dmu_tx_hold_zap(oh->ot_tx, obj->oo_db->db_object, TRUE, NULL);
+	/* do not specify the key as then DMU is trying to look it up
+	 * which is very expensive. usually the layers above lookup
+	 * before deletion */
+	osd_tx_hold_zap(oh->ot_tx, obj->oo_dn->dn_object, obj->oo_dn,
+			FALSE, NULL);
 
 	RETURN(0);
 }
 
 static int osd_index_delete(const struct lu_env *env, struct dt_object *dt,
-			    const struct dt_key *key, struct thandle *th,
-			    struct lustre_capa *capa)
+			    const struct dt_key *key, struct thandle *th)
 {
 	struct osd_object  *obj = osd_dt_obj(dt);
 	struct osd_device  *osd = osd_obj2dev(obj);
@@ -1168,14 +1850,14 @@ static int osd_index_delete(const struct lu_env *env, struct dt_object *dt,
 	int                 rc;
 	ENTRY;
 
-	LASSERT(obj->oo_db);
+	LASSERT(obj->oo_dn);
 	LASSERT(th != NULL);
-	oh = container_of0(th, struct osd_thandle, ot_super);
+	oh = container_of(th, struct osd_thandle, ot_super);
 
 	rc = osd_prepare_key_uint64(obj, k, key);
 
 	/* Remove binary key from the ZAP */
-	rc = -zap_remove_uint64(osd->od_objset.os, obj->oo_db->db_object,
+	rc = -zap_remove_uint64(osd->od_os, obj->oo_dn->dn_object,
 				k, rc, oh->ot_tx);
 	RETURN(rc);
 }
@@ -1195,11 +1877,11 @@ static int osd_index_it_get(const struct lu_env *env, struct dt_it *di,
 	 * XXX: we need a binary version of zap_cursor_move_to_key()
 	 *	to implement this API */
 	if (*((const __u64 *)key) != 0)
-		CERROR("NOT IMPLEMETED YET (move to %Lx)\n", *((__u64 *)key));
+		CERROR("NOT IMPLEMETED YET (move to %#llx)\n",
+		       *((__u64 *)key));
 
 	zap_cursor_fini(it->ozi_zc);
-	memset(it->ozi_zc, 0, sizeof(*it->ozi_zc));
-	zap_cursor_init(it->ozi_zc, osd->od_objset.os, obj->oo_db->db_object);
+	zap_cursor_init(it->ozi_zc, osd->od_os, obj->oo_dn->dn_object);
 	it->ozi_reset = 1;
 
 	RETURN(+1);
@@ -1275,7 +1957,7 @@ static int osd_index_it_rec(const struct lu_env *env, const struct dt_it *di,
 
 	rc = osd_prepare_key_uint64(obj, k, (const struct dt_key *)za->za_name);
 
-	rc = -zap_lookup_uint64(osd->od_objset.os, obj->oo_db->db_object,
+	rc = -zap_lookup_uint64(osd->od_os, obj->oo_dn->dn_object,
 				k, rc, obj->oo_recusize, obj->oo_recsize,
 				(void *)rec);
 	RETURN(rc);
@@ -1300,13 +1982,10 @@ static int osd_index_it_load(const struct lu_env *env, const struct dt_it *di,
 	int                rc;
 	ENTRY;
 
-	/* close the current cursor */
+	/* reset the cursor */
 	zap_cursor_fini(it->ozi_zc);
-
-	/* create a new one starting at hash */
-	memset(it->ozi_zc, 0, sizeof(*it->ozi_zc));
-	zap_cursor_init_serialized(it->ozi_zc, osd->od_objset.os,
-				   obj->oo_db->db_object, hash);
+	zap_cursor_init_serialized(it->ozi_zc, osd->od_os,
+				   obj->oo_dn->dn_object, hash);
 	it->ozi_reset = 0;
 
 	rc = -zap_cursor_retrieve(it->ozi_zc, za);
@@ -1338,260 +2017,56 @@ static struct dt_index_operations osd_index_ops = {
 	}
 };
 
-struct osd_metadnode_it {
-	struct osd_device       *mit_dev;
-	__u64			 mit_pos;
-	struct lu_fid		 mit_fid;
-	int			 mit_prefetched;
-	__u64			 mit_prefetched_dnode;
-};
-
-static struct dt_it *osd_zfs_otable_it_init(const struct lu_env *env,
-					    struct dt_object *dt, __u32 attr,
-					    struct lustre_capa *capa)
-{
-	struct osd_device	*dev   = osd_dev(dt->do_lu.lo_dev);
-	struct osd_metadnode_it *it;
-	ENTRY;
-
-	OBD_ALLOC_PTR(it);
-	if (unlikely(it == NULL))
-		RETURN(ERR_PTR(-ENOMEM));
-
-	it->mit_dev = dev;
-
-	RETURN((struct dt_it *)it);
-}
-
-static void osd_zfs_otable_it_fini(const struct lu_env *env, struct dt_it *di)
-{
-	struct osd_metadnode_it *it  = (struct osd_metadnode_it *)di;
-
-	OBD_FREE_PTR(it);
-}
-
-static int osd_zfs_otable_it_get(const struct lu_env *env,
-				 struct dt_it *di, const struct dt_key *key)
-{
-	return 0;
-}
-
-static void osd_zfs_otable_it_put(const struct lu_env *env, struct dt_it *di)
-{
-}
-
-#define OTABLE_PREFETCH		256
-
-static void osd_zfs_otable_prefetch(const struct lu_env *env,
-				    struct osd_metadnode_it *it)
-{
-	struct osd_device	*dev = it->mit_dev;
-	udmu_objset_t		*uos = &dev->od_objset;
-	int			 rc;
-
-	/* can go negative on the very first access to the iterator
-	 * or if some non-Lustre objects were found */
-	if (unlikely(it->mit_prefetched < 0))
-		it->mit_prefetched = 0;
-
-	if (it->mit_prefetched >= (OTABLE_PREFETCH >> 1))
-		return;
-
-	if (it->mit_prefetched_dnode == 0)
-		it->mit_prefetched_dnode = it->mit_pos;
-
-	while (it->mit_prefetched < OTABLE_PREFETCH) {
-		rc = -dmu_object_next(uos->os, &it->mit_prefetched_dnode,
-				      B_FALSE, 0);
-		if (unlikely(rc != 0))
-			break;
-
-		/* dmu_prefetch() was exported in 0.6.2, if you use with
-		 * an older release, just comment it out - this is an
-		 * optimization */
-		dmu_prefetch(uos->os, it->mit_prefetched_dnode, 0, 0);
-
-		it->mit_prefetched++;
-	}
-}
-
-static int osd_zfs_otable_it_next(const struct lu_env *env, struct dt_it *di)
-{
-	struct osd_metadnode_it *it  = (struct osd_metadnode_it *)di;
-	struct lustre_mdt_attrs *lma;
-	struct osd_device	*dev = it->mit_dev;
-	udmu_objset_t		*uos = &dev->od_objset;
-	nvlist_t		*nvbuf = NULL;
-	uchar_t			*v;
-	__u64			 dnode;
-	int			 rc, s;
-	ENTRY;
-
-	memset(&it->mit_fid, 0, sizeof(it->mit_fid));
-
-	dnode = it->mit_pos;
-	do {
-		rc = -dmu_object_next(uos->os, &it->mit_pos, B_FALSE, 0);
-		if (unlikely(rc != 0))
-			GOTO(out, rc = 1);
-		it->mit_prefetched--;
-
-		/* LMA is required for this to be a Lustre object.
-		 * If there is no xattr skip it. */
-		rc = __osd_xattr_load(uos, it->mit_pos, &nvbuf);
-		if (unlikely(rc != 0))
-			continue;
-
-		LASSERT(nvbuf != NULL);
-		rc = -nvlist_lookup_byte_array(nvbuf, XATTR_NAME_LMA, &v, &s);
-		if (likely(rc == 0)) {
-			/* Lustre object */
-			lma = (struct lustre_mdt_attrs *)v;
-			lustre_lma_swab(lma);
-			it->mit_fid = lma->lma_self_fid;
-			nvlist_free(nvbuf);
-			break;
-		} else {
-			/* not a Lustre object, try next one */
-			nvlist_free(nvbuf);
-		}
-
-	} while (1);
-
-
-	/* we aren't prefetching in the above loop because the number of
-	 * non-Lustre objects is very small and we will be repeating very
-	 * rare. in case we want to use this to iterate over non-Lustre
-	 * objects (i.e. when we convert regular ZFS in Lustre) it makes
-	 * sense to initiate prefetching in the loop */
-
-	/* 0 - there are more items, +1 - the end */
-	if (likely(rc == 0))
-		osd_zfs_otable_prefetch(env, it);
-
-	CDEBUG(D_OTHER, "advance: %llu -> %llu "DFID": %d\n", dnode,
-	       it->mit_pos, PFID(&it->mit_fid), rc);
-
-out:
-	RETURN(rc);
-}
-
-static struct dt_key *osd_zfs_otable_it_key(const struct lu_env *env,
-					    const struct dt_it *di)
-{
-	return NULL;
-}
-
-static int osd_zfs_otable_it_key_size(const struct lu_env *env,
-				      const struct dt_it *di)
-{
-	return sizeof(__u64);
-}
-
-static int osd_zfs_otable_it_rec(const struct lu_env *env,
-				 const struct dt_it *di,
-				 struct dt_rec *rec, __u32 attr)
-{
-	struct osd_metadnode_it *it  = (struct osd_metadnode_it *)di;
-	struct lu_fid *fid = (struct lu_fid *)rec;
-	ENTRY;
-
-	*fid = it->mit_fid;
-
-	RETURN(0);
-}
-
-
-static __u64 osd_zfs_otable_it_store(const struct lu_env *env,
-				     const struct dt_it *di)
-{
-	struct osd_metadnode_it *it  = (struct osd_metadnode_it *)di;
-
-	return it->mit_pos;
-}
-
-static int osd_zfs_otable_it_load(const struct lu_env *env,
-				  const struct dt_it *di, __u64 hash)
-{
-	struct osd_metadnode_it *it  = (struct osd_metadnode_it *)di;
-
-	it->mit_pos = hash;
-	it->mit_prefetched = 0;
-	it->mit_prefetched_dnode = 0;
-
-	return osd_zfs_otable_it_next(env, (struct dt_it *)di);
-}
-
-static int osd_zfs_otable_it_key_rec(const struct lu_env *env,
-				     const struct dt_it *di, void *key_rec)
-{
-	return 0;
-}
-
-const struct dt_index_operations osd_zfs_otable_ops = {
-	.dio_it = {
-		.init     = osd_zfs_otable_it_init,
-		.fini     = osd_zfs_otable_it_fini,
-		.get      = osd_zfs_otable_it_get,
-		.put	  = osd_zfs_otable_it_put,
-		.next     = osd_zfs_otable_it_next,
-		.key	  = osd_zfs_otable_it_key,
-		.key_size = osd_zfs_otable_it_key_size,
-		.rec      = osd_zfs_otable_it_rec,
-		.store    = osd_zfs_otable_it_store,
-		.load     = osd_zfs_otable_it_load,
-		.key_rec  = osd_zfs_otable_it_key_rec,
-	}
-};
-
 int osd_index_try(const struct lu_env *env, struct dt_object *dt,
 		const struct dt_index_features *feat)
 {
 	struct osd_object *obj = osd_dt_obj(dt);
+	struct osd_device *osd = osd_obj2dev(obj);
+	const struct lu_fid *fid = lu_object_fid(&dt->do_lu);
+	int rc = 0;
 	ENTRY;
 
-	LASSERT(dt_object_exists(dt));
+	down_read(&obj->oo_guard);
 
 	/*
 	 * XXX: implement support for fixed-size keys sorted with natural
 	 *      numerical way (not using internal hash value)
 	 */
 	if (feat->dif_flags & DT_IND_RANGE)
-		RETURN(-ERANGE);
+		GOTO(out, rc = -ERANGE);
 
 	if (unlikely(feat == &dt_otable_features)) {
-		dt->do_index_ops = &osd_zfs_otable_ops;
-		RETURN(0);
+		dt->do_index_ops = &osd_otable_ops;
+		GOTO(out, rc = 0);
 	}
 
-	LASSERT(obj->oo_db != NULL);
+	LASSERT(!dt_object_exists(dt) || obj->oo_dn != NULL);
 	if (likely(feat == &dt_directory_features)) {
-		if (udmu_object_is_zap(obj->oo_db))
+		if (!dt_object_exists(dt) || osd_object_is_zap(obj->oo_dn))
 			dt->do_index_ops = &osd_dir_ops;
 		else
-			RETURN(-ENOTDIR);
+			GOTO(out, rc = -ENOTDIR);
 	} else if (unlikely(feat == &dt_acct_features)) {
-		LASSERT(fid_is_acct(lu_object_fid(&dt->do_lu)));
+		LASSERT(fid_is_acct(fid));
 		dt->do_index_ops = &osd_acct_index_ops;
-	} else if (udmu_object_is_zap(obj->oo_db) &&
-		   dt->do_index_ops == NULL) {
+	} else if (dt->do_index_ops == NULL) {
 		/* For index file, we don't support variable key & record sizes
 		 * and the key has to be unique */
 		if ((feat->dif_flags & ~DT_IND_UPDATE) != 0)
-			RETURN(-EINVAL);
+			GOTO(out, rc = -EINVAL);
 
 		if (feat->dif_keysize_max > ZAP_MAXNAMELEN)
-			RETURN(-E2BIG);
+			GOTO(out, rc = -E2BIG);
 		if (feat->dif_keysize_max != feat->dif_keysize_min)
-			RETURN(-EINVAL);
+			GOTO(out, rc = -EINVAL);
 
 		/* As for the record size, it should be a multiple of 8 bytes
 		 * and smaller than the maximum value length supported by ZAP.
 		 */
 		if (feat->dif_recsize_max > ZAP_MAXVALUELEN)
-			RETURN(-E2BIG);
+			GOTO(out, rc = -E2BIG);
 		if (feat->dif_recsize_max != feat->dif_recsize_min)
-			RETURN(-EINVAL);
+			GOTO(out, rc = -EINVAL);
 
 		obj->oo_keysize = feat->dif_keysize_max;
 		obj->oo_recsize = feat->dif_recsize_max;
@@ -1603,7 +2078,28 @@ int osd_index_try(const struct lu_env *env, struct dt_object *dt,
 			obj->oo_recusize = 8;
 		}
 		dt->do_index_ops = &osd_index_ops;
+
+		if (feat == &dt_lfsck_layout_orphan_features ||
+		    feat == &dt_lfsck_layout_dangling_features ||
+		    feat == &dt_lfsck_namespace_features)
+			GOTO(out, rc = 0);
+
+		rc = osd_index_register(osd, fid, obj->oo_keysize,
+					obj->oo_recusize * obj->oo_recsize);
+		if (rc < 0)
+			CWARN("%s: failed to register index "DFID": rc = %d\n",
+			      osd_name(osd), PFID(fid), rc);
+		else if (rc > 0)
+			rc = 0;
+		else
+			CDEBUG(D_LFSCK, "%s: index object "DFID
+			       " (%u/%u/%u) registered\n",
+			       osd_name(osd), PFID(fid), obj->oo_keysize,
+			       obj->oo_recusize, obj->oo_recsize);
 	}
 
-	RETURN(0);
+out:
+	up_read(&obj->oo_guard);
+
+	RETURN(rc);
 }