RFC parent inode pointers.

cover

#ifdef HERE
	if (error == ENOSPC) {
		spaceres = 0;
		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_rename, 0, 0);
	}
#endif 
	... 
       ----- kernel space bits -----
       ----- user space bits -----
---
 fs/xfs/xfs_da_btree.h   |    1 +
 fs/xfs/xfs_dir2.c       |    6 +++++-
 fs/xfs/xfs_dir2.h       |    3 ++-
 fs/xfs/xfs_dir2_block.c |    1 +
 fs/xfs/xfs_dir2_leaf.c  |    2 ++
 fs/xfs/xfs_dir2_node.c  |    2 ++
 fs/xfs/xfs_dir2_sf.c    |    2 ++
 fs/xfs/xfs_inode.c      |    7 ++++---
 fs/xfs/xfs_symlink.c    |    2 +-
 9 files changed, 20 insertions(+), 6 deletions(-)
Index: b/fs/xfs/xfs_da_btree.h
===================================================================
@@ -66,6 +66,7 @@ typedef struct xfs_da_args {
 	int		rmtblkcnt2;	/* remote attr value block count */
 	int		op_flags;	/* operation flags */
 	enum xfs_dacmp	cmpresult;	/* name compare result for lookups */
+	__uint32_t	offset;		/* OUT: offset in directory */
 } xfs_da_args_t;
 
 /*

Index: b/fs/xfs/xfs_dir2.c
===================================================================
@@ -203,7 +203,8 @@ xfs_dir_createname(
 	xfs_ino_t		inum,		/* new entry inode number */
 	xfs_fsblock_t		*first,		/* bmap's firstblock */
 	xfs_bmap_free_t		*flist,		/* bmap's freeblock list */
-	xfs_extlen_t		total)		/* bmap's total block count */
+	xfs_extlen_t		total,		/* bmap's total block count */
+	__uint32_t		*offset)	/* OUT entry's dir offset */
 {
 	xfs_da_args_t		args;
 	int			rval;

@@ -240,6 +241,9 @@ xfs_dir_createname(
 		rval = xfs_dir2_leaf_addname(&args);
 	else
 		rval = xfs_dir2_node_addname(&args);
+	/* return the location that this entry was place in the parent inode */
+	if (offset)
+		*offset = args.offset;
 	return rval;
 }
 

Index: b/fs/xfs/xfs_dir2.h
===================================================================
@@ -119,7 +119,8 @@ extern int xfs_dir_init(struct xfs_trans
 extern int xfs_dir_createname(struct xfs_trans *tp, struct xfs_inode *dp,
 				struct xfs_name *name, xfs_ino_t inum,
 				xfs_fsblock_t *first,
-				struct xfs_bmap_free *flist, xfs_extlen_t tot);
+				struct xfs_bmap_free *flist, xfs_extlen_t tot,
+				 __uint32_t *offset);
 extern int xfs_dir_lookup(struct xfs_trans *tp, struct xfs_inode *dp,
 				struct xfs_name *name, xfs_ino_t *inum,
 				struct xfs_name *ci_name);

Index: b/fs/xfs/xfs_dir2_block.c
===================================================================
@@ -554,6 +554,7 @@ xfs_dir2_block_addname(
 	dp->d_ops->data_put_ftype(dep, args->filetype);
 	tagp = dp->d_ops->data_entry_tag_p(dep);
 	*tagp = cpu_to_be16((char *)dep - (char *)hdr);
+	args->offset = xfs_dir2_byte_to_dataptr(mp, (char *)dep - (char *)hdr);
 	/*
 	 * Clean up the bestfree array and log the header, tail, and entry.
 	 */

Index: b/fs/xfs/xfs_dir2_leaf.c
===================================================================
@@ -860,6 +860,8 @@ xfs_dir2_leaf_addname(
 	dp->d_ops->data_put_ftype(dep, args->filetype);
 	tagp = dp->d_ops->data_entry_tag_p(dep);
 	*tagp = cpu_to_be16((char *)dep - (char *)hdr);
+	args->offset = xfs_dir2_db_off_to_dataptr(mp, use_block,
+						(char *)dep - (char *)hdr);
 	/*
 	 * Need to scan fix up the bestfree table.
 	 */

Index: b/fs/xfs/xfs_dir2_node.c
===================================================================
@@ -1973,6 +1973,8 @@ xfs_dir2_node_addname_int(
 	dp->d_ops->data_put_ftype(dep, args->filetype);
 	tagp = dp->d_ops->data_entry_tag_p(dep);
 	*tagp = cpu_to_be16((char *)dep - (char *)hdr);
+	args->offset = xfs_dir2_db_off_to_dataptr(mp, dbno,
+						  (char *)dep - (char *)hdr);
 	xfs_dir2_data_log_entry(tp, dp, dbp, dep);
 	/*
 	 * Rescan the block for bestfree if needed.

Index: b/fs/xfs/xfs_dir2_sf.c
===================================================================
@@ -425,6 +425,7 @@ xfs_dir2_sf_addname_easy(
 	memcpy(sfep->name, args->name, sfep->namelen);
 	dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
 	dp->d_ops->sf_put_ftype(sfep, args->filetype);
+	args->offset = xfs_dir2_byte_to_dataptr(dp->i_mount, offset);
 
 	/*
 	 * Update the header and inode.

@@ -520,6 +521,7 @@ xfs_dir2_sf_addname_hard(
 	memcpy(sfep->name, args->name, sfep->namelen);
 	dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
 	dp->d_ops->sf_put_ftype(sfep, args->filetype);
+	args->offset = xfs_dir2_byte_to_dataptr(dp->i_mount, offset);
 	sfp->count++;
 #if XFS_BIG_INUMS
 	if (args->inumber > XFS_DIR2_MAX_SHORT_INUM && !objchange)

Index: b/fs/xfs/xfs_inode.c
===================================================================
@@ -1264,7 +1264,8 @@ xfs_create(
 
 	error = xfs_dir_createname(tp, dp, name, ip->i_ino,
 					&first_block, &free_list, resblks ?
-					resblks - XFS_IALLOC_SPACE_RES(mp) : 0);
+					resblks - XFS_IALLOC_SPACE_RES(mp) : 0,
+					NULL);
 	if (error) {
 		ASSERT(error != ENOSPC);
 		goto out_trans_abort;

@@ -1402,7 +1403,7 @@ xfs_link(
 	xfs_bmap_init(&free_list, &first_block);
 
 	error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino,
-					&first_block, &free_list, resblks);
+				   &first_block, &free_list, resblks, NULL);
 	if (error)
 		goto abort_return;
 	xfs_trans_ichgtime(tp, tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);

@@ -2736,7 +2737,7 @@ xfs_rename(
 		 */
 		error = xfs_dir_createname(tp, target_dp, target_name,
 						src_ip->i_ino, &first_block,
-						&free_list, spaceres);
+						&free_list, spaceres, NULL);
 		if (error == ENOSPC)
 			goto error_return;
 		if (error)

Index: b/fs/xfs/xfs_symlink.c
===================================================================
@@ -375,7 +375,7 @@ xfs_symlink(
 	 * Create the directory entry for the symlink.
 	 */
 	error = xfs_dir_createname(tp, dp, link_name, ip->i_ino,
-					&first_block, &free_list, resblks);
+				   &first_block, &free_list, resblks, NULL);
 	if (error)
 		goto error2;
 	xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);

---
 fs/xfs/xfs_dir2.c       |    5 ++++-
 fs/xfs/xfs_dir2.h       |    3 ++-
 fs/xfs/xfs_dir2_block.c |    1 +
 fs/xfs/xfs_dir2_leaf.c  |    1 +
 fs/xfs/xfs_dir2_node.c  |    1 +
 fs/xfs/xfs_dir2_sf.c    |    2 ++
 fs/xfs/xfs_inode.c      |    4 ++--
 7 files changed, 13 insertions(+), 4 deletions(-)
Index: b/fs/xfs/xfs_dir2.c
===================================================================
@@ -340,7 +340,8 @@ xfs_dir_removename(
 	xfs_ino_t	ino,
 	xfs_fsblock_t	*first,		/* bmap's firstblock */
 	xfs_bmap_free_t	*flist,		/* bmap's freeblock list */
-	xfs_extlen_t	total)		/* bmap's total block count */
+	xfs_extlen_t	total,		/* bmap's total block count */
+	__uint32_t	*offset)	/* out: return entry's dir offset */
 {
 	xfs_da_args_t	args;
 	int		rval;

@@ -374,6 +375,8 @@ xfs_dir_removename(
 		rval = xfs_dir2_leaf_removename(&args);
 	else
 		rval = xfs_dir2_node_removename(&args);
+	if (offset)
+		*offset = args.offset;
 	return rval;
 }
 

Index: b/fs/xfs/xfs_dir2.h
===================================================================
@@ -127,7 +127,8 @@ extern int xfs_dir_lookup(struct xfs_tra
 extern int xfs_dir_removename(struct xfs_trans *tp, struct xfs_inode *dp,
 				struct xfs_name *name, xfs_ino_t ino,
 				xfs_fsblock_t *first,
-				struct xfs_bmap_free *flist, xfs_extlen_t tot);
+				struct xfs_bmap_free *flist, xfs_extlen_t tot,
+				__uint32_t *offset);
 extern int xfs_dir_replace(struct xfs_trans *tp, struct xfs_inode *dp,
 				struct xfs_name *name, xfs_ino_t inum,
 				xfs_fsblock_t *first,

Index: b/fs/xfs/xfs_dir2_block.c
===================================================================
@@ -796,6 +796,7 @@ xfs_dir2_block_removename(
 	/*
 	 * Point to the data entry using the leaf entry.
 	 */
+	args->offset = be32_to_cpu(blp[ent].address);
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)hdr + xfs_dir2_dataptr_to_off(mp, be32_to_cpu(blp[ent].address)));
 	/*

Index: b/fs/xfs/xfs_dir2_leaf.c
===================================================================
@@ -1381,6 +1381,7 @@ xfs_dir2_leaf_removename(
 	 * Point to the leaf entry, use that to point to the data entry.
 	 */
 	lep = &ents[index];
+	args->offset = be32_to_cpu(lep->address);
 	db = xfs_dir2_dataptr_to_db(mp, be32_to_cpu(lep->address));
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)hdr + xfs_dir2_dataptr_to_off(mp, be32_to_cpu(lep->address)));

Index: b/fs/xfs/xfs_dir2_node.c
===================================================================
@@ -1192,6 +1192,7 @@ xfs_dir2_leafn_remove(
 	/*
 	 * Extract the data block and offset from the entry.
 	 */
+	args->offset = be32_to_cpu(lep->address);
 	db = xfs_dir2_dataptr_to_db(mp, be32_to_cpu(lep->address));
 	ASSERT(dblk->blkno == db);
 	off = xfs_dir2_dataptr_to_off(mp, be32_to_cpu(lep->address));

Index: b/fs/xfs/xfs_dir2_sf.c
===================================================================
@@ -847,6 +847,8 @@ xfs_dir2_sf_removename(
 								XFS_CMP_EXACT) {
 			ASSERT(dp->d_ops->sf_get_ino(sfp, sfep) ==
 			       args->inumber);
+			args->offset = xfs_dir2_byte_to_dataptr(dp->i_mount,
+						xfs_dir2_sf_get_offset(sfep));
 			break;
 		}
 	}

Index: b/fs/xfs/xfs_inode.c
===================================================================
@@ -2542,7 +2542,7 @@ xfs_remove(
 
 	xfs_bmap_init(&free_list, &first_block);
 	error = xfs_dir_removename(tp, dp, name, ip->i_ino,
-					&first_block, &free_list, resblks);
+				     &first_block, &free_list, resblks, NULL);
 	if (error) {
 		ASSERT(error != ENOENT);
 		goto out_bmap_cancel;

@@ -2847,7 +2847,7 @@ xfs_rename(
 	}
 
 	error = xfs_dir_removename(tp, src_dp, src_name, src_ip->i_ino,
-					&first_block, &free_list, spaceres);
+				     &first_block, &free_list, spaceres, NULL);
 	if (error)
 		goto abort_return;
 

---
 fs/xfs/xfs_dir2.c       |    5 ++++-
 fs/xfs/xfs_dir2.h       |    3 ++-
 fs/xfs/xfs_dir2_block.c |    1 +
 fs/xfs/xfs_dir2_leaf.c  |    1 +
 fs/xfs/xfs_dir2_node.c  |    1 +
 fs/xfs/xfs_dir2_sf.c    |    3 +++
 fs/xfs/xfs_inode.c      |    8 ++++----
 7 files changed, 16 insertions(+), 6 deletions(-)
Index: b/fs/xfs/xfs_dir2.c
===================================================================
@@ -391,7 +391,8 @@ xfs_dir_replace(
 	xfs_ino_t	inum,		/* new inode number */
 	xfs_fsblock_t	*first,		/* bmap's firstblock */
 	xfs_bmap_free_t	*flist,		/* bmap's freeblock list */
-	xfs_extlen_t	total)		/* bmap's total block count */
+	xfs_extlen_t	total,		/* bmap's total block count */
+	__uint32_t	*offset)	/* out: return entry's dir offset */
 {
 	xfs_da_args_t	args;
 	int		rval;

@@ -427,6 +428,8 @@ xfs_dir_replace(
 		rval = xfs_dir2_leaf_replace(&args);
 	else
 		rval = xfs_dir2_node_replace(&args);
+	if (offset)
+		*offset = args.offset;
 	return rval;
 }
 

Index: b/fs/xfs/xfs_dir2.h
===================================================================
@@ -132,7 +132,8 @@ extern int xfs_dir_removename(struct xfs
 extern int xfs_dir_replace(struct xfs_trans *tp, struct xfs_inode *dp,
 				struct xfs_name *name, xfs_ino_t inum,
 				xfs_fsblock_t *first,
-				struct xfs_bmap_free *flist, xfs_extlen_t tot);
+				struct xfs_bmap_free *flist, xfs_extlen_t tot,
+				__uint32_t *offset);
 extern int xfs_dir_canenter(struct xfs_trans *tp, struct xfs_inode *dp,
 				struct xfs_name *name, uint resblks);
 

Index: b/fs/xfs/xfs_dir2_block.c
===================================================================
@@ -872,6 +872,7 @@ xfs_dir2_block_replace(
 	/*
 	 * Point to the data entry we need to change.
 	 */
+	args->offset = be32_to_cpu(blp[ent].address);
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)hdr + xfs_dir2_dataptr_to_off(mp, be32_to_cpu(blp[ent].address)));
 	ASSERT(be64_to_cpu(dep->inumber) != args->inumber);

Index: b/fs/xfs/xfs_dir2_leaf.c
===================================================================
@@ -1515,6 +1515,7 @@ xfs_dir2_leaf_replace(
 	/*
 	 * Point to the data entry.
 	 */
+	args->offset = be32_to_cpu(lep->address);
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)dbp->b_addr +
 	       xfs_dir2_dataptr_to_off(dp->i_mount, be32_to_cpu(lep->address)));

Index: b/fs/xfs/xfs_dir2_node.c
===================================================================
@@ -2186,6 +2186,7 @@ xfs_dir2_node_replace(
 		hdr = state->extrablk.bp->b_addr;
 		ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
 		       hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC));
+		args->offset = be32_to_cpu(lep->address);
 		dep = (xfs_dir2_data_entry_t *)
 		      ((char *)hdr +
 		       xfs_dir2_dataptr_to_off(state->mp, be32_to_cpu(lep->address)));

Index: b/fs/xfs/xfs_dir2_sf.c
===================================================================
@@ -987,6 +987,9 @@ xfs_dir2_sf_replace(
 #endif
 				dp->d_ops->sf_put_ino(sfp, sfep, args->inumber);
 				dp->d_ops->sf_put_ftype(sfep, args->filetype);
+				args->offset = xfs_dir2_byte_to_dataptr(
+						  dp->i_mount,
+						  xfs_dir2_sf_get_offset(sfep));
 				break;
 			}
 		}

Index: b/fs/xfs/xfs_inode.c
===================================================================
@@ -2778,8 +2778,8 @@ xfs_rename(
 		 * name at the destination directory, remove it first.
 		 */
 		error = xfs_dir_replace(tp, target_dp, target_name,
-					src_ip->i_ino,
-					&first_block, &free_list, spaceres);
+					src_ip->i_ino, &first_block,
+					&free_list, spaceres, NULL);
 		if (error)
 			goto abort_return;
 

@@ -2813,8 +2813,8 @@ xfs_rename(
 		 * directory.
 		 */
 		error = xfs_dir_replace(tp, src_ip, &xfs_name_dotdot,
-					target_dp->i_ino,
-					&first_block, &free_list, spaceres);
+					target_dp->i_ino, &first_block,
+					&free_list, spaceres, NULL);
 		ASSERT(error != EEXIST);
 		if (error)
 			goto abort_return;

---
 fs/xfs/xfs_sb.h |   11 ++++++++++-
 1 file changed, 10 insertions(+), 1 deletion(-)
Index: b/fs/xfs/xfs_sb.h
===================================================================
@@ -89,6 +89,7 @@ struct xfs_trans;
 #define	XFS_SB_VERSION2_OKREALFBITS	\
 	(XFS_SB_VERSION2_LAZYSBCOUNTBIT	| \
 	 XFS_SB_VERSION2_ATTR2BIT	| \
+	 XFS_SB_VERSION2_PARENTBIT	| \
 	 XFS_SB_VERSION2_PROJID32BIT	| \
 	 XFS_SB_VERSION2_FTYPE)
 #define	XFS_SB_VERSION2_OKSASHFBITS	\

@@ -596,8 +597,10 @@ xfs_sb_has_ro_compat_feature(
 }
 
 #define XFS_SB_FEAT_INCOMPAT_FTYPE	(1 << 0)	/* filetype in dirent */
+#define XFS_SB_FEAT_INCOMPAT_PARENT	(2 << 0)	/* parent inode ptr */
 #define XFS_SB_FEAT_INCOMPAT_ALL \
-		(XFS_SB_FEAT_INCOMPAT_FTYPE)
+		(XFS_SB_FEAT_INCOMPAT_FTYPE | \
+		 XFS_SB_FEAT_INCOMPAT_PARENT)
 
 #define XFS_SB_FEAT_INCOMPAT_UNKNOWN	~XFS_SB_FEAT_INCOMPAT_ALL
 static inline bool

@@ -639,6 +642,12 @@ static inline int xfs_sb_version_hasftyp
 		 (sbp->sb_features2 & XFS_SB_VERSION2_FTYPE));
 }
 
+static inline int xfs_sb_version_hasparent(struct xfs_sb *sbp)
+{
+	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 &&
+		xfs_sb_has_incompat_feature(sbp, XFS_SB_FEAT_INCOMPAT_PARENT));
+}
+
 /*
  * end of superblock version macros
  */

---
 fs/xfs/xfs_attr.c      |   20 ++++++++++++--------
 fs/xfs/xfs_attr.h      |    2 ++
 fs/xfs/xfs_da_format.h |   12 ++++++++----
 fs/xfs/xfs_fs.h        |    1 +
 fs/xfs/xfs_fsops.c     |    4 +++-
 6 files changed, 65 insertions(+), 13 deletions(-)
Index: b/fs/xfs/xfs_attr.c
===================================================================
@@ -234,7 +234,7 @@ xfs_attr_set_int(
 	int			error, err2, committed;
 	struct xfs_mount	*mp = dp->i_mount;
 	struct xfs_trans_res	tres;
-	int			rsvd = (flags & ATTR_ROOT) != 0;
+	int			rsvd = (flags & (ATTR_ROOT | ATTR_PARENT)) != 0;
 	int			local;
 
 	/*

@@ -444,18 +444,22 @@ xfs_attr_set(
 	int		flags)
 {
 	int             error;
-	struct xfs_name	xname;
+	struct xfs_name	xname, *xnamep;
 
 	XFS_STATS_INC(xs_attr_set);
 
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return (EIO);
 
-	error = xfs_attr_name_to_xname(&xname, name);
-	if (error)
-		return error;
-
-	return xfs_attr_set_int(dp, &xname, value, valuelen, flags);
+	if ((flags & ATTR_PARENT) == 0) {
+		error = xfs_attr_name_to_xname(&xname, name);
+		if (error)
+			return error;
+		xnamep = &xname;
+	} else
+		xnamep = (struct xfs_name *) name;
+	return xfs_attr_set_int(dp, xnamep, value, valuelen, flags);
+					
 }
 
 /*

@@ -516,7 +520,7 @@ xfs_attr_remove_int(xfs_inode_t *dp, str
 	 * operation if necessary
 	 */
 
-	if (flags & ATTR_ROOT)
+	if (flags & (ATTR_ROOT | ATTR_PARENT))
 		args.trans->t_flags |= XFS_TRANS_RESERVE;
 
 	error = xfs_trans_reserve(args.trans, &M_RES(mp)->tr_attrrm,

Index: b/fs/xfs/xfs_attr.h
===================================================================
@@ -44,6 +44,7 @@ struct xfs_attr_list_context;
 #define ATTR_SECURE	0x0008	/* use attrs in security namespace */
 #define ATTR_CREATE	0x0010	/* pure create: fail if attr already exists */
 #define ATTR_REPLACE	0x0020	/* pure set: fail if attr does not exist */
+#define ATTR_PARENT	0x0040	/*  use attrs in parent namespace */
 
 #define ATTR_KERNOTIME	0x1000	/* [kernel] don't update inode timestamps */
 #define ATTR_KERNOVAL	0x2000	/* [kernel] get attr size only, not value */

@@ -55,6 +56,7 @@ struct xfs_attr_list_context;
 	{ ATTR_SECURE,		"SECURE" }, \
 	{ ATTR_CREATE,		"CREATE" }, \
 	{ ATTR_REPLACE,		"REPLACE" }, \
+	{ ATTR_PARENT,		"PARENT" }, \
 	{ ATTR_KERNOTIME,	"KERNOTIME" }, \
 	{ ATTR_KERNOVAL,	"KERNOVAL" }
 

Index: b/fs/xfs/xfs_da_format.h
===================================================================
@@ -895,24 +895,28 @@ struct xfs_attr3_icleaf_hdr {
 #define	XFS_ATTR_LOCAL_BIT	0	/* attr is stored locally */
 #define	XFS_ATTR_ROOT_BIT	1	/* limit access to trusted attrs */
 #define	XFS_ATTR_SECURE_BIT	2	/* limit access to secure attrs */
+#define XFS_ATTR_PARENT_BIT	3	/* parent pointer secure attrs */
 #define	XFS_ATTR_INCOMPLETE_BIT	7	/* attr in middle of create/delete */
 #define XFS_ATTR_LOCAL		(1 << XFS_ATTR_LOCAL_BIT)
 #define XFS_ATTR_ROOT		(1 << XFS_ATTR_ROOT_BIT)
 #define XFS_ATTR_SECURE		(1 << XFS_ATTR_SECURE_BIT)
+#define XFS_ATTR_PARENT		(1 << XFS_ATTR_PARENT_BIT)
 #define XFS_ATTR_INCOMPLETE	(1 << XFS_ATTR_INCOMPLETE_BIT)
 
 /*
  * Conversion macros for converting namespace bits from argument flags
  * to ondisk flags.
  */
-#define XFS_ATTR_NSP_ARGS_MASK		(ATTR_ROOT | ATTR_SECURE)
-#define XFS_ATTR_NSP_ONDISK_MASK	(XFS_ATTR_ROOT | XFS_ATTR_SECURE)
+#define XFS_ATTR_NSP_ARGS_MASK		(ATTR_ROOT | ATTR_SECURE | XFS_ATTR_PARENT)
+#define XFS_ATTR_NSP_ONDISK_MASK	(XFS_ATTR_ROOT | XFS_ATTR_SECURE | XFS_ATTR_PARENT)
 #define XFS_ATTR_NSP_ONDISK(flags)	((flags) & XFS_ATTR_NSP_ONDISK_MASK)
 #define XFS_ATTR_NSP_ARGS(flags)	((flags) & XFS_ATTR_NSP_ARGS_MASK)
 #define XFS_ATTR_NSP_ARGS_TO_ONDISK(x)	(((x) & ATTR_ROOT ? XFS_ATTR_ROOT : 0) |\
-					 ((x) & ATTR_SECURE ? XFS_ATTR_SECURE : 0))
+					 ((x) & ATTR_SECURE ? XFS_ATTR_SECURE : 0) | \
+					 ((x) & ATTR_PARENT ? XFS_ATTR_PARENT : 0))
 #define XFS_ATTR_NSP_ONDISK_TO_ARGS(x)	(((x) & XFS_ATTR_ROOT ? ATTR_ROOT : 0) |\
-					 ((x) & XFS_ATTR_SECURE ? ATTR_SECURE : 0))
+					 ((x) & XFS_ATTR_SECURE ? ATTR_SECURE : 0) | \
+					 ((x) & XFS_ATTR_PARENT ? ATTR_PARENT : 0))
 
 /*
  * Alignment for namelist and valuelist entries (since they are mixed

Index: b/fs/xfs/xfs_fs.h
===================================================================
@@ -238,6 +238,7 @@ typedef struct xfs_fsop_resblks {
 #define XFS_FSOP_GEOM_FLAGS_LAZYSB	0x4000	/* lazy superblock counters */
 #define XFS_FSOP_GEOM_FLAGS_V5SB	0x8000	/* version 5 superblock */
 #define XFS_FSOP_GEOM_FLAGS_FTYPE	0x10000	/* inode directory types */
+#define XFS_FSOP_GEOM_FLAGS_PARENT	0x20000	/* parent inode support */
 
 /*
  * Minimum and maximum sizes need for growth checks.

Index: b/fs/xfs/xfs_fsops.c
===================================================================
@@ -104,7 +104,9 @@ xfs_fs_geometry(
 			(xfs_sb_version_hascrc(&mp->m_sb) ?
 				XFS_FSOP_GEOM_FLAGS_V5SB : 0) |
 			(xfs_sb_version_hasftype(&mp->m_sb) ?
-				XFS_FSOP_GEOM_FLAGS_FTYPE : 0);
+				XFS_FSOP_GEOM_FLAGS_FTYPE : 0) |
+			(xfs_sb_version_hasparent(&mp->m_sb) ?
+				XFS_FSOP_GEOM_FLAGS_PARENT : 0);
 		geo->logsectsize = xfs_sb_version_hassector(&mp->m_sb) ?
 				mp->m_sb.sb_logsectsize : BBSIZE;
 		geo->rtsectsize = mp->m_sb.sb_blocksize;

---
 fs/xfs/xfs_dinode.h     |    4 +++-
 fs/xfs/xfs_inode_buf.c  |    6 ++++++
 fs/xfs/xfs_log_format.h |    4 +++-
 3 files changed, 12 insertions(+), 2 deletions(-)
Index: b/fs/xfs/xfs_dinode.h
===================================================================
@@ -79,7 +79,9 @@ typedef struct xfs_dinode {
 	__be64		di_changecount;	/* number of attribute changes */
 	__be64		di_lsn;		/* flush sequence */
 	__be64		di_flags2;	/* more random flags */
-	__u8		di_pad2[16];	/* more padding for future expansion */
+	__be64		di_parent;	/* inode of parent directory */
+	__be32		di_poffset;	/* offset into parent directory */
+	__u8		di_pad2[4];	/* more padding for future expansion */
 
 	/* fields only written to during inode creation */
 	xfs_timestamp_t	di_crtime;	/* time created */

Index: b/fs/xfs/xfs_inode_buf.c
===================================================================
@@ -238,6 +238,9 @@ xfs_dinode_from_disk(
 		to->di_lsn = be64_to_cpu(from->di_lsn);
 		memcpy(to->di_pad2, from->di_pad2, sizeof(to->di_pad2));
 		uuid_copy(&to->di_uuid, &from->di_uuid);
+ 		/* parent pointer information is left in disk order */
+		to->di_parent = from->di_parent;
+		to->di_poffset = from->di_poffset;
 	}
 }
 

@@ -285,6 +288,9 @@ xfs_dinode_to_disk(
 		memcpy(to->di_pad2, from->di_pad2, sizeof(to->di_pad2));
 		uuid_copy(&to->di_uuid, &from->di_uuid);
 		to->di_flushiter = 0;
+ 		/* parent pointer information is in disk order */
+		to->di_parent = from->di_parent;
+		to->di_poffset = from->di_poffset;
 	} else {
 		to->di_flushiter = cpu_to_be16(from->di_flushiter);
 	}

Index: b/fs/xfs/xfs_log_format.h
===================================================================
@@ -399,7 +399,9 @@ typedef struct xfs_icdinode {
 	__uint64_t	di_changecount;	/* number of attribute changes */
 	xfs_lsn_t	di_lsn;		/* flush sequence */
 	__uint64_t	di_flags2;	/* more random flags */
-	__uint8_t	di_pad2[16];	/* more padding for future expansion */
+	__uint64_t	di_parent;	/* inode of parent directory */
+	__uint32_t	di_poffset;	/* offset into parent directory */
+	__uint8_t	di_pad2[4];	/* more padding for future expansion */
 
 	/* fields only written to during inode creation */
 	xfs_ictimestamp_t di_crtime;	/* time created */

---
 fs/xfs/xfs_inode.c |    9 ++++++++-
 1 file changed, 8 insertions(+), 1 deletion(-)
Index: b/fs/xfs/xfs_inode.c
===================================================================
@@ -1163,6 +1163,7 @@ xfs_create(
 	struct xfs_dquot	*pdqp = NULL;
 	struct xfs_trans_res	tres;
 	uint			resblks;
+	uint			offset;
 
 	trace_xfs_create(dp, name);
 

@@ -1265,11 +1266,17 @@ xfs_create(
 	error = xfs_dir_createname(tp, dp, name, ip->i_ino,
 					&first_block, &free_list, resblks ?
 					resblks - XFS_IALLOC_SPACE_RES(mp) : 0,
-					NULL);
+					&offset);
 	if (error) {
 		ASSERT(error != ENOSPC);
 		goto out_trans_abort;
 	}
+
+	if (xfs_sb_version_hasparent(&mp->m_sb)) {
+		/* set the parent pointer and offset to the inode core fields */
+		ip->i_d.di_parent = cpu_to_be64(dp->i_ino);
+		ip->i_d.di_poffset = cpu_to_be32(offset);
+	}
 	xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 	xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
 

---
 fs/xfs/xfs_symlink.c |   10 +++++++++-
 1 file changed, 9 insertions(+), 1 deletion(-)
Index: b/fs/xfs/xfs_symlink.c
===================================================================
@@ -36,6 +36,7 @@
 #include "xfs_bmap_util.h"
 #include "xfs_error.h"
 #include "xfs_quota.h"
+#include "xfs_attr.h"
 #include "xfs_trans_space.h"
 #include "xfs_trace.h"
 #include "xfs_symlink.h"

@@ -192,6 +193,7 @@ xfs_symlink(
 	struct xfs_dquot	*gdqp = NULL;
 	struct xfs_dquot	*pdqp = NULL;
 	uint			resblks;
+	uint			offset;
 
 	*ipp = NULL;
 

@@ -375,9 +377,15 @@ xfs_symlink(
 	 * Create the directory entry for the symlink.
 	 */
 	error = xfs_dir_createname(tp, dp, link_name, ip->i_ino,
-				   &first_block, &free_list, resblks, NULL);
+				   &first_block, &free_list, resblks, &offset);
 	if (error)
 		goto error2;
+
+	if (xfs_sb_version_hasparent(&mp->m_sb)) {
+		/* set the parent pointer and offset to the inode core fields */
+		ip->i_d.di_parent = cpu_to_be64(dp->i_ino);
+		ip->i_d.di_poffset = cpu_to_be32(offset);
+	}
 	xfs_trans_ichgtime(tp, dp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 	xfs_trans_log_inode(tp, dp, XFS_ILOG_CORE);
 

---
 fs/xfs/xfs_inode.c  |   43 ++++++++++++++++++++++++++++++++++++++++---
 fs/xfs/xfs_parent.h |   39 +++++++++++++++++++++++++++++++++++++++
 2 files changed, 79 insertions(+), 3 deletions(-)
Index: b/fs/xfs/xfs_inode.c
===================================================================
@@ -51,6 +51,7 @@
 #include "xfs_trans_priv.h"
 #include "xfs_log.h"
 #include "xfs_bmap_btree.h"
+#include "xfs_parent.h"
 
 kmem_zone_t *xfs_inode_zone;
 

@@ -1358,6 +1359,7 @@ xfs_link(
 	int			cancel_flags;
 	int			committed;
 	int			resblks;
+	uint			offset;
 
 	trace_xfs_link(tdp, target_name);
 

@@ -1407,12 +1409,26 @@ xfs_link(
 	if (error)
 		goto error_return;
 
-	xfs_bmap_init(&free_list, &first_block);
+ 	xfs_bmap_init(&free_list, &first_block);
 
 	error = xfs_dir_createname(tp, tdp, target_name, sip->i_ino,
-				   &first_block, &free_list, resblks, NULL);
+				   &first_block, &free_list, resblks, &offset);
 	if (error)
 		goto abort_return;
+
+	if (xfs_sb_version_hasparent(&mp->m_sb) &&
+	    sip->i_d.di_parent == NULLFSINO) {
+		/*
+		 * store the parent information in the inode if the fields
+		 * are empty. zero the offset as a flag to not add the
+		 * information in an extended attribute.
+		 */
+		sip->i_d.di_parent = cpu_to_be64(tdp->i_ino);
+		sip->i_d.di_poffset = cpu_to_be32(offset);
+		xfs_trans_log_inode(tp, sip, XFS_ILOG_CORE);
+		offset = 0;
+	}
+
 	xfs_trans_ichgtime(tp, tdp, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 	xfs_trans_log_inode(tp, tdp, XFS_ILOG_CORE);
 

@@ -1435,7 +1451,28 @@ xfs_link(
 		goto abort_return;
 	}
 
-	return xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+	error = xfs_trans_commit(tp, XFS_TRANS_RELEASE_LOG_RES);
+	if (error)
+		return error;
+
+	if (xfs_sb_version_hasparent(&mp->m_sb) && offset) {
+		struct xfs_pattr	p_entry;
+		struct xfs_name		p_name;
+
+		/*
+		 * The parent pointer information could not be stored
+		 * inode core, and must be stored in an extended attribute.
+		 */
+		xfs_parent_pname(tdp->i_ino, offset, &p_entry, &p_name);
+		error = xfs_attr_set(sip, (char *) &p_name, NULL, 0,
+				     ATTR_PARENT|ATTR_CREATE);
+		if (error)
+			xfs_notice(mp,
+			   "%s: attr %llx/%x failed inode %llx %d\n",
+			   __func__, tdp->i_ino, offset, sip->i_ino,
+			   error);
+	}
+	return error;
 
  abort_return:
 	cancel_flags |= XFS_TRANS_ABORT;

Index: b/fs/xfs/xfs_parent.h
===================================================================
@@ -0,0 +1,39 @@
+/*
+ * Copyright (c) 2014 SGI
+ * All Rights Reserved.
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License as
+ * published by the Free Software Foundation.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301  USA
+ */
+
+#ifndef	__XFS_PARENT_H__
+#define	__XFS_PARENT_H__
+
+#define XFSREMOVESRC	(1 << 0)
+#define XFSREMOVETAR	(1 << 1)
+#define XFSADD		(1 << 2)
+
+struct xfs_pattr {
+	__be64  p_ino;
+	__be32  p_offset;
+} __attribute__((packed));
+
+static inline void xfs_parent_pname(xfs_ino_t ino, __uint32_t off, struct xfs_pattr *pe, struct xfs_name *pn)
+{
+	pe->p_ino = cpu_to_be64(ino);
+	pe->p_offset = cpu_to_be32(off);
+	pn->name = (char *)pe;
+	pn->len = sizeof(struct xfs_pattr);
+}
+#endif /* __XFS_PARENT_H__ */
+

---
 fs/xfs/xfs_attr.c  |   16 ++++++++++------
 fs/xfs/xfs_inode.c |   36 ++++++++++++++++++++++++++++++++----
 2 files changed, 42 insertions(+), 10 deletions(-)
Index: b/fs/xfs/xfs_attr.c
===================================================================
@@ -466,7 +466,7 @@ xfs_attr_set(
  * Generic handler routine to remove a name from an attribute list.
  * Transitions attribute list from Btree to shortform as necessary.
  */
-STATIC int
+int
 xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
 {
 	xfs_da_args_t	args;

@@ -594,16 +594,20 @@ xfs_attr_remove(
 	int		flags)
 {
 	int		error;
-	struct xfs_name	xname;
+	struct xfs_name	xname, *xnamep;
 
 	XFS_STATS_INC(xs_attr_remove);
 
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return (EIO);
 
-	error = xfs_attr_name_to_xname(&xname, name);
-	if (error)
-		return error;
+	if ((flags & ATTR_PARENT) == 0) {
+		xnamep = &xname;
+		error = xfs_attr_name_to_xname(xnamep, name);
+		if (error)
+			return error;
+	} else
+		xnamep = (struct xfs_name *) name;
 
 	xfs_ilock(dp, XFS_ILOCK_SHARED);
 	if (!xfs_inode_hasattr(dp)) {

@@ -612,7 +616,7 @@ xfs_attr_remove(
 	}
 	xfs_iunlock(dp, XFS_ILOCK_SHARED);
 
-	return xfs_attr_remove_int(dp, &xname, flags);
+	return xfs_attr_remove_int(dp, xnamep, flags);
 }
 
 

Index: b/fs/xfs/xfs_inode.c
===================================================================
@@ -2493,6 +2493,7 @@ xfs_remove(
 	int			link_zero;
 	uint			resblks;
 	uint			log_count;
+	uint			offset;
 
 	trace_xfs_remove(dp, name);
 

@@ -2585,13 +2586,24 @@ xfs_remove(
 	link_zero = (ip->i_d.di_nlink == 0);
 
 	xfs_bmap_init(&free_list, &first_block);
-	error = xfs_dir_removename(tp, dp, name, ip->i_ino,
-				     &first_block, &free_list, resblks, NULL);
+	error = xfs_dir_removename(tp, dp, name, ip->i_ino, &first_block,
+				   &free_list, resblks, &offset);
 	if (error) {
 		ASSERT(error != ENOENT);
 		goto out_bmap_cancel;
 	}
 
+	if (xfs_sb_version_hasparent(&mp->m_sb) &&
+	    atomic_read(&VFS_I(ip)->i_count) > 1 && !link_zero &&
+	    ip->i_d.di_parent == cpu_to_be64(dp->i_ino) &&
+	    ip->i_d.di_poffset == cpu_to_be32(offset)) {
+		/* Remove the incore parent information */
+		ip->i_d.di_parent = NULLFSINO;
+		ip->i_d.di_poffset = 0;
+		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
+		offset = 0;
+	}
+
 	/*
 	 * If this is a synchronous mount, make sure that the
 	 * remove transaction goes to disk before returning to

@@ -2608,6 +2620,22 @@ xfs_remove(
 	if (error)
 		goto std_return;
 
+	if (xfs_sb_version_hasparent(&mp->m_sb) &&
+	    atomic_read(&VFS_I(ip)->i_count) > 1 && !link_zero && offset) {
+		struct xfs_pattr	p_entry;
+		struct xfs_name		p_name;
+
+		xfs_parent_pname(dp->i_ino, offset, &p_entry, &p_name);
+		error = xfs_attr_remove(ip, (char *) &p_name, ATTR_PARENT);
+		if (error) {
+			xfs_notice(mp,
+				"%s: attr %llx/%x failed inode %llx %d\n",
+				__func__, dp->i_ino, offset,
+				 ip->i_ino, error);
+			if (error == ENOATTR)
+				error = 0;
+		}
+	}
 	/*
 	 * If we are using filestreams, kill the stream association.
 	 * If the file is still open it may get a new one but that

@@ -2617,7 +2645,7 @@ xfs_remove(
 	if (!is_dir && link_zero && xfs_inode_is_filestream(ip))
 		xfs_filestream_deassociate(ip);
 
-	return 0;
+	return error;
 
  out_bmap_cancel:
 	xfs_bmap_cancel(&free_list);

---
 fs/xfs/Makefile     |    1 
 fs/xfs/xfs_inode.c  |   69 +++++++++++++++++++++-
 fs/xfs/xfs_parent.c |  158 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/xfs_parent.h |    5 +
 4 files changed, 228 insertions(+), 5 deletions(-)
Index: b/fs/xfs/Makefile
===================================================================
@@ -49,6 +49,7 @@ xfs-y				+= xfs_aops.o \
 				   xfs_message.o \
 				   xfs_mount.o \
 				   xfs_mru_cache.o \
+				   xfs_parent.o \
 				   xfs_super.o \
 				   xfs_symlink.o \
 				   xfs_trans.o \

Index: b/fs/xfs/xfs_parent.h
===================================================================
@@ -35,5 +35,10 @@ static inline void xfs_parent_pname(xfs_
 	pn->name = (char *)pe;
 	pn->len = sizeof(struct xfs_pattr);
 }
+
+int xfs_pptr_rename(struct xfs_mount *mp, struct xfs_trans *tp,
+	struct xfs_inode *src_dp, struct xfs_inode *target_dp,
+	struct xfs_inode *src_ip, struct xfs_inode *target_ip,
+	uint src_offset, uint tar_offset);
 #endif /* __XFS_PARENT_H__ */
 

---
 fs/xfs/xfs_attr.h   |    2 
 fs/xfs/xfs_fs.h     |   10 +
 fs/xfs/xfs_ioctl.c  |   72 ++++++++
 fs/xfs/xfs_parent.c |  435 ++++++++++++++++++++++++++++++++++++++++++++++++++++
 fs/xfs/xfs_parent.h |   30 +++
 5 files changed, 548 insertions(+), 1 deletion(-)
Index: b/fs/xfs/xfs_attr.h
===================================================================
@@ -131,6 +131,8 @@ typedef struct xfs_attr_list_context {
 	int				put_value;	/* T/F: need value for listent */
 	put_listent_func_t		put_listent;	/* list output fmt function */
 	int				index;		/* index into output buffer */
+	char				*scratch;	/* parent private buf */
+	int				srtchlen;	/* parent priv buf len */
 } xfs_attr_list_context_t;
 
 

Index: b/fs/xfs/xfs_fs.h
===================================================================
@@ -438,6 +438,14 @@ typedef struct xfs_fsop_attrmulti_handle
 	struct xfs_attr_multiop		__user *ops; /* attr_multi data */
 } xfs_fsop_attrmulti_handlereq_t;
 
+typedef struct xfs_fsop_parentlist_handlereq {
+	struct xfs_fsop_handlereq	hreq;	/* handle interface structure */
+	u32				flags;	/* which namespace to use */
+	u32				buflen;	/* length of supplied buffer */
+	__s32				__user *ocount; /* output count ptr */
+	void				__user *buffer; /* returned names */
+} xfs_fsop_parentlist_handlereq_t;
+
 /*
  * per machine unique filesystem identifier types.
  */

@@ -553,6 +561,8 @@ typedef struct xfs_swapext
 #define XFS_IOC_ATTRMULTI_BY_HANDLE  _IOW ('X', 123, struct xfs_fsop_attrmulti_handlereq)
 #define XFS_IOC_FSGEOMETRY	     _IOR ('X', 124, struct xfs_fsop_geom)
 #define XFS_IOC_GOINGDOWN	     _IOR ('X', 125, __uint32_t)
+#define XFS_IOC_GETPARENTS_BY_HANDLE _IOWR('X', 126, struct xfs_fsop_parentlist_handlereq)
+#define XFS_IOC_GETPARENTPATHS_BY_HANDLE _IOWR('X', 127, struct xfs_fsop_parentlist_handlereq)
 /*	XFS_IOC_GETFSUUID ---------- deprecated 140	 */
 
 

Index: b/fs/xfs/xfs_parent.h
===================================================================
@@ -28,6 +28,22 @@ struct xfs_pattr {
 	__be32  p_offset;
 } __attribute__((packed));
 
+/* xfs_readdir fill program private data */
+struct xfs_pfillinfo {
+	struct dir_context	ctx;		/* must be first entry */
+	int			gp_len;
+	int			gp_off;
+	char			*gp_path;
+};
+
+typedef struct xfs_parent {
+	uint64_t	p_ino;		/* parent inode number */
+	uint32_t	p_offset;	/* entry offset in parent inode */
+	uint16_t	p_reclen;	/* name length */
+	uint16_t	p_pad;		/* padding for future */
+	char		p_name[1];	/* variable length name */
+} xfs_parent_t;
+
 static inline void xfs_parent_pname(xfs_ino_t ino, __uint32_t off, struct xfs_pattr *pe, struct xfs_name *pn)
 {
 	pe->p_ino = cpu_to_be64(ino);

@@ -36,9 +52,21 @@ static inline void xfs_parent_pname(xfs_
 	pn->len = sizeof(struct xfs_pattr);
 }
 
+/* actual bytes used by parent entry */
+#define PARENT_ENTSIZE(namelen)					\
+	((offsetof(struct xfs_parent, p_name) + (namelen) + 1 +	\
+	sizeof(uint64_t)-1) & ~(sizeof(uint64_t)-1))
+
+int xfs_getpath(struct xfs_mount *mp, xfs_ino_t pino, xfs_off_t offset,
+		char *path, int *plen);
+int xfs_getpname(struct xfs_mount *mp, xfs_ino_t pino, xfs_off_t offset,
+		char *path, int *plen);
+int xfs_parents_attr_list(struct xfs_inode *dp, char *buffer, int bufsize,
+		int *count);
+int xfs_parentpaths_attr_list(struct xfs_inode *dp, char *buffer, int bufsize,
+		int *count);
 int xfs_pptr_rename(struct xfs_mount *mp, struct xfs_trans *tp,
 	struct xfs_inode *src_dp, struct xfs_inode *target_dp,
 	struct xfs_inode *src_ip, struct xfs_inode *target_ip,
 	uint src_offset, uint tar_offset);
 #endif /* __XFS_PARENT_H__ */
-

---
 include/xfs_dinode.h     |    4 +++-
 include/xfs_log_format.h |    4 +++-
 include/xfs_sb.h         |   14 ++++++++++++--
 libxfs/xfs_ialloc.c      |   12 ++++++++++++
 man/man8/mkfs.xfs.8      |    6 ++++++
 mkfs/xfs_mkfs.c          |   24 +++++++++++++++++++-----
 6 files changed, 55 insertions(+), 9 deletions(-)
Index: b/include/xfs_dinode.h
===================================================================
@@ -79,7 +79,9 @@ typedef struct xfs_dinode {
 	__be64		di_changecount;	/* number of attribute changes */
 	__be64		di_lsn;		/* flush sequence */
 	__be64		di_flags2;	/* more random flags */
-	__u8		di_pad2[16];	/* more padding for future expansion */
+	__be64		di_parent;	/* parent directory inode */
+	__be32		di_poffset;	/* offset into parent directory */
+	__u8		di_pad2[4];	/* more padding for future expansion */
 
 	/* fields only written to during inode creation */
 	xfs_timestamp_t	di_crtime;	/* time created */

Index: b/include/xfs_log_format.h
===================================================================
@@ -567,7 +567,9 @@ typedef struct xfs_icdinode {
 	__uint64_t	di_changecount;	/* number of attribute changes */
 	xfs_lsn_t	di_lsn;		/* flush sequence */
 	__uint64_t	di_flags2;	/* more random flags */
-	__uint8_t	di_pad2[16];	/* more padding for future expansion */
+	__uint64_t	di_parent;	/* parent directory inode */
+	__uint32_t	di_poffset;	/* offset into parent directory */
+	__uint8_t	di_pad2[4];	/* more padding for future expansion */
 
 	/* fields only written to during inode creation */
 	xfs_ictimestamp_t di_crtime;	/* time created */

Index: b/include/xfs_sb.h
===================================================================
@@ -90,6 +90,7 @@ struct xfs_trans;
 	(XFS_SB_VERSION2_LAZYSBCOUNTBIT	| \
 	 XFS_SB_VERSION2_ATTR2BIT	| \
 	 XFS_SB_VERSION2_PROJID32BIT	| \
+	 XFS_SB_VERSION2_PARENTBIT	| \
 	 XFS_SB_VERSION2_FTYPE)
 #define	XFS_SB_VERSION2_OKSASHFBITS	\
 	(0)

@@ -328,7 +329,6 @@ typedef enum {
 	 XFS_SB_FEATURES_RO_COMPAT | XFS_SB_FEATURES_INCOMPAT | \
 	 XFS_SB_FEATURES_LOG_INCOMPAT | XFS_SB_PQUOTINO)
 
-
 /*
  * Misc. Flags - warning - these will be cleared by xfs_repair unless
  * a feature bit is set when the flag is used.

@@ -596,8 +596,10 @@ xfs_sb_has_ro_compat_feature(
 }
 
 #define XFS_SB_FEAT_INCOMPAT_FTYPE	(1 << 0)	/* filetype in dirent */
+#define XFS_SB_FEAT_INCOMPAT_PARENT	(2 << 0)	/* parent inode ptrs */
 #define XFS_SB_FEAT_INCOMPAT_ALL \
-		(XFS_SB_FEAT_INCOMPAT_FTYPE)
+		(XFS_SB_FEAT_INCOMPAT_FTYPE | \
+		 XFS_SB_FEAT_INCOMPAT_PARENT)
 
 #define XFS_SB_FEAT_INCOMPAT_UNKNOWN	~XFS_SB_FEAT_INCOMPAT_ALL
 static inline bool

@@ -639,6 +641,14 @@ static inline int xfs_sb_version_hasftyp
 		 (sbp->sb_features2 & XFS_SB_VERSION2_FTYPE));
 }
 
+static inline int xfs_sb_version_hasparent(struct xfs_sb *sbp)
+{
+	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 &&
+		xfs_sb_has_incompat_feature(sbp, XFS_SB_FEAT_INCOMPAT_PARENT)) ||
+		(xfs_sb_version_hasmorebits(sbp) &&
+		(sbp->sb_features2 & XFS_SB_VERSION2_PARENTBIT));
+}
+
 /*
  * end of superblock version macros
  */

Index: b/libxfs/xfs_ialloc.c
===================================================================
@@ -233,6 +233,18 @@ xfs_ialloc_inode_init(
 			free->di_next_unlinked = cpu_to_be32(NULLAGINO);
 
 			if (version == 3) {
+				if (xfs_sb_version_hasparent(&mp->m_sb) &&
+					   (mp->m_sb.sb_rootino == NULLFSINO ||
+					    mp->m_sb.sb_rootino == ino)) {
+					/*
+					 * set the parent pointer information
+					 * in the root inode.
+					 */
+					free->di_parent = cpu_to_be64(ino);
+					free->di_poffset =
+					    cpu_to_be32(xfs_dir3_data_dotdot_offset(mp)
+						    >> XFS_DIR2_DATA_ALIGN_LOG);
+				}
 				free->di_ino = cpu_to_be64(ino);
 				ino++;
 				uuid_copy(&free->di_uuid, &mp->m_sb.sb_uuid);

Index: b/man/man8/mkfs.xfs.8
===================================================================
@@ -364,6 +364,12 @@ This is used to enable 32bit quota proje
 is either 0 or 1, with 1 signifying that 32bit projid are to be enabled.
 If the value is omitted, 1 is assumed.  (This default changed
 in release version 3.2.0.)
+.TP
+.BI parent[= value ]
+This is used to enable parent inode pointers in CRC enabled filesystems. The
+.I value
+is either 0 or 1, with 1 signifying that parent inode pointers are to be enabled.
+If the value is omitted, 0 is assumed.
 .RE
 .TP
 .BI \-l " log_section_options"

Index: b/mkfs/xfs_mkfs.c
===================================================================
@@ -109,6 +109,8 @@ char	*iopts[] = {
 	"attr",
 #define	I_PROJID32BIT	6
 	"projid32bit",
+#define	I_PARENT	7
+	"parent",
 	NULL
 };
 

@@ -149,7 +151,7 @@ char	*nopts[] = {
 	"version",
 #define	N_FTYPE		3
 	"ftype",
-	NULL,
+	NULL
 };
 
 char	*ropts[] = {

@@ -962,6 +964,7 @@ main(
 	struct fs_topology	ft;
 	int			lazy_sb_counters;
 	int			crcs_enabled;
+	int			parent;
 
 	progname = basename(argv[0]);
 	setlocale(LC_ALL, "");

@@ -995,6 +998,7 @@ main(
 	worst_freelist = 0;
 	lazy_sb_counters = 1;
 	crcs_enabled = 0;
+	parent = 0;
 	memset(&fsx, 0, sizeof(fsx));
 
 	memset(&xi, 0, sizeof(xi));

@@ -1316,6 +1320,14 @@ main(
 						illegal(value, "i projid32bit");
 					projid16bit = c ? 0 : 1;
 					break;
+				case I_PARENT:
+					if (!value || *value == '\0')
+						value = "0";
+					c = atoi(value);
+					if (c < 0 || c > 1)
+						illegal(value, "i parent");
+					parent = c;
+					break;
 				default:
 					unknown('i', value);
 				}

@@ -2463,12 +2475,14 @@ _("size %s specified for log subvolume i
 	if (crcs_enabled) {
 		sbp->sb_features_incompat = XFS_SB_FEAT_INCOMPAT_FTYPE;
 		dirftype = 1;
-	}
+		if (parent)
+			sbp->sb_features_incompat |= XFS_SB_FEAT_INCOMPAT_PARENT;
+	 }
 
 	if (!qflag || Nflag) {
 		printf(_(
 		   "meta-data=%-22s isize=%-6d agcount=%lld, agsize=%lld blks\n"
-		   "         =%-22s sectsz=%-5u attr=%u, projid32bit=%u\n"
+		   "         =%-22s sectsz=%-5u attr=%u, projid32bit=%u, parent=%d\n"
 		   "         =%-22s crc=%u\n"
 		   "data     =%-22s bsize=%-6u blocks=%llu, imaxpct=%u\n"
 		   "         =%-22s sunit=%-6u swidth=%u blks\n"

@@ -2477,7 +2491,7 @@ _("size %s specified for log subvolume i
 		   "         =%-22s sectsz=%-5u sunit=%d blks, lazy-count=%d\n"
 		   "realtime =%-22s extsz=%-6d blocks=%lld, rtextents=%lld\n"),
 			dfile, isize, (long long)agcount, (long long)agsize,
-			"", sectorsize, attrversion, !projid16bit,
+			"", sectorsize, attrversion, !projid16bit, parent,
 			"", crcs_enabled,
 			"", blocksize, (long long)dblocks, imaxpct,
 			"", dsunit, dswidth,

@@ -3085,7 +3099,7 @@ usage( void )
 			    sectlog=n|sectsize=num\n\
 /* force overwrite */	[-f]\n\
 /* inode size */	[-i log=n|perblock=n|size=num,maxpct=n,attr=0|1|2,\n\
-			    projid32bit=0|1]\n\
+			    projid32bit=0|1,parent==0|1]\n\
 /* no discard */	[-K]\n\
 /* log subvol */	[-l agnum=n,internal,size=num,logdev=xxx,version=n\n\
 			    sunit=value|su=num,sectlog=n|sectsize=num,\n\

---
 include/libxfs.h        |    1 +
 include/xfs_da_format.h |   12 ++++++++----
 libxfs/xfs.h            |    1 +
 repair/attr_repair.c    |   18 +++++++++++-------
 4 files changed, 21 insertions(+), 11 deletions(-)
Index: b/include/libxfs.h
===================================================================
@@ -603,6 +603,7 @@ typedef struct xfs_inode {
 #define LIBXFS_ATTR_SECURE	0x0008	/* use attrs in security namespace */
 #define LIBXFS_ATTR_CREATE	0x0010	/* create, but fail if attr exists */
 #define LIBXFS_ATTR_REPLACE	0x0020	/* set, but fail if attr not exists */
+#define LIBXFS_ATTR_PARENT	0x0040	/* parent pointer entry */
 
 /*
  * Project quota id helpers (previously projid was 16bit only and using two

Index: b/include/xfs_da_format.h
===================================================================
@@ -1242,24 +1242,28 @@ struct xfs_attr3_icleaf_hdr {
 #define	XFS_ATTR_LOCAL_BIT	0	/* attr is stored locally */
 #define	XFS_ATTR_ROOT_BIT	1	/* limit access to trusted attrs */
 #define	XFS_ATTR_SECURE_BIT	2	/* limit access to secure attrs */
+#define	XFS_ATTR_PARENT_BIT	3	/* parent pointer entry */
 #define	XFS_ATTR_INCOMPLETE_BIT	7	/* attr in middle of create/delete */
 #define XFS_ATTR_LOCAL		(1 << XFS_ATTR_LOCAL_BIT)
 #define XFS_ATTR_ROOT		(1 << XFS_ATTR_ROOT_BIT)
 #define XFS_ATTR_SECURE		(1 << XFS_ATTR_SECURE_BIT)
+#define XFS_ATTR_PARENT		(1 << XFS_ATTR_PARENT_BIT)
 #define XFS_ATTR_INCOMPLETE	(1 << XFS_ATTR_INCOMPLETE_BIT)
 
 /*
  * Conversion macros for converting namespace bits from argument flags
  * to ondisk flags.
  */
-#define XFS_ATTR_NSP_ARGS_MASK		(ATTR_ROOT | ATTR_SECURE)
-#define XFS_ATTR_NSP_ONDISK_MASK	(XFS_ATTR_ROOT | XFS_ATTR_SECURE)
+#define XFS_ATTR_NSP_ARGS_MASK		(ATTR_ROOT | ATTR_SECURE | ATTR_PARENT)
+#define XFS_ATTR_NSP_ONDISK_MASK	(XFS_ATTR_ROOT | XFS_ATTR_SECURE | XFS_ATTR_PARENT)
 #define XFS_ATTR_NSP_ONDISK(flags)	((flags) & XFS_ATTR_NSP_ONDISK_MASK)
 #define XFS_ATTR_NSP_ARGS(flags)	((flags) & XFS_ATTR_NSP_ARGS_MASK)
 #define XFS_ATTR_NSP_ARGS_TO_ONDISK(x)	(((x) & ATTR_ROOT ? XFS_ATTR_ROOT : 0) |\
-					 ((x) & ATTR_SECURE ? XFS_ATTR_SECURE : 0))
+					 ((x) & ATTR_SECURE ? XFS_ATTR_SECURE : 0) | \
+					 ((x) & ATTR_PARENT ? XFS_ATTR_PARENT : 0))
 #define XFS_ATTR_NSP_ONDISK_TO_ARGS(x)	(((x) & XFS_ATTR_ROOT ? ATTR_ROOT : 0) |\
-					 ((x) & XFS_ATTR_SECURE ? ATTR_SECURE : 0))
+					 ((x) & XFS_ATTR_SECURE ? ATTR_SECURE : 0) | \
+					 ((x) & XFS_ATTR_PARENT ? ATTR_PARENT : 0))
 
 /*
  * Alignment for namelist and valuelist entries (since they are mixed

Index: b/libxfs/xfs.h
===================================================================
@@ -97,6 +97,7 @@ typedef struct xfs_bmalloca {
 #define ATTR_SECURE			LIBXFS_ATTR_SECURE
 #define ATTR_CREATE			LIBXFS_ATTR_CREATE
 #define ATTR_REPLACE			LIBXFS_ATTR_REPLACE
+#define ATTR_PARENT			LIBXFS_ATTR_PARENT
 #define ATTR_KERNOTIME			0
 #define ATTR_KERNOVAL			0
 

Index: b/repair/attr_repair.c
===================================================================
@@ -17,6 +17,7 @@
  */
 
 #include <libxfs.h>
+#include <xfs_attr_leaf.h>
 #include "globals.h"
 #include "err_protos.h"
 #include "attr_repair.h"

@@ -880,7 +881,8 @@ process_shortform_attr(
 		/* namecheck checks for / and null terminated for file names.
 		 * attributes names currently follow the same rules.
 		*/
-		if (namecheck((char *)&currententry->nameval[0],
+		if (!(currententry->flags & XFS_ATTR_PARENT) &&
+		     namecheck((char *)&currententry->nameval[0],
 						currententry->namelen))  {
 			do_warn(
 	_("entry contains illegal character in shortform attribute name\n"));

@@ -1026,8 +1028,9 @@ process_leaf_attr_local(
 	xfs_attr_leaf_name_local_t *local;
 
 	local = xfs_attr3_leaf_name_local(leaf, i);
-	if (local->namelen == 0 || namecheck((char *)&local->nameval[0], 
-							local->namelen)) {
+	if (local->namelen == 0 ||
+	     (!(entry->flags & XFS_ATTR_PARENT) &&
+	      namecheck((char *)&local->nameval[0], local->namelen))) {
 		do_warn(
 	_("attribute entry %d in attr block %u, inode %" PRIu64 " has bad name (namelen = %d)\n"),
 			i, da_bno, ino, local->namelen);

@@ -1081,10 +1084,11 @@ process_leaf_attr_remote(
 
 	remotep = xfs_attr3_leaf_name_remote(leaf, i);
 
-	if (remotep->namelen == 0 || namecheck((char *)&remotep->name[0], 
-						remotep->namelen) || 
-			be32_to_cpu(entry->hashval) != 
-				libxfs_da_hashname((uchar_t *)&remotep->name[0], 
+	if (remotep->namelen == 0 ||
+	    (!(entry->flags & XFS_ATTR_PARENT) &&
+	     namecheck((char *)&remotep->name[0],remotep->namelen)) ||
+	     be32_to_cpu(entry->hashval) != 
+				libxfs_da_hashname((uchar_t *)&remotep->name[0],
 						remotep->namelen) ||
 			be32_to_cpu(entry->hashval) < last_hashval ||
 			be32_to_cpu(remotep->valueblk) == 0) {

---
 db/inode.c              |    2 ++
---
 db/inode.c |    2 ++
 1 file changed, 2 insertions(+)
Index: b/db/inode.c
===================================================================
@@ -175,6 +175,8 @@ const field_t	inode_v3_flds[] = {
 	{ "crtime", FLDT_TIMESTAMP, OI(COFF(crtime)), C1, 0, TYP_NONE },
 	{ "inumber", FLDT_INO, OI(COFF(ino)), C1, 0, TYP_NONE },
 	{ "uuid", FLDT_UUID, OI(COFF(uuid)), C1, 0, TYP_NONE },
+	{ "parent", FLDT_UINT64D, OI(COFF(parent)), C1, 0, TYP_NONE },
+	{ "poffset", FLDT_UINT32X, OI(OFF(poffset)), C1, 0, TYP_NONE },
 	{ NULL }
 };
 

---
 include/handle.h   |    6 ++--
 include/jdm.h      |   14 +++------
 include/parent.h   |   13 ++++----
 include/xfs_fs.h   |   11 +++++++
 io/parent.c        |   56 +++++++++++++++++++++++---------------
 libhandle/handle.c |   78 +++++++++++++++++++++++++++++++++++++++++++----------
 libhandle/jdm.c    |   48 ++++++++++++++++++++++----------
 7 files changed, 158 insertions(+), 68 deletions(-)
Index: b/include/handle.h
===================================================================
@@ -24,7 +24,7 @@ extern "C" {
 
 struct fsdmidata;
 struct attrlist_cursor;
-struct parent;
+struct xfs_parent;
 
 extern int  path_to_handle (char *__path, void **__hanp, size_t *__hlen);
 extern int  path_to_fshandle (char *__path, void **__fshanp, size_t *__fshlen);

@@ -41,10 +41,10 @@ extern int  attr_list_by_handle (void *_
 				 size_t __bufsize, int __flags,
 				 struct attrlist_cursor *__cursor);
 extern int  parents_by_handle(void *__hanp, size_t __hlen,
-			      struct parent *__buf, size_t __bufsize,
+			      struct xfs_parent *__buf, size_t __bufsize,
 			      unsigned int *__count);
 extern int  parentpaths_by_handle(void *__hanp, size_t __hlen,
-				  struct parent *__buf, size_t __bufsize,
+				  struct xfs_parent *__buf, size_t __bufsize,
 				  unsigned int *__count);
 extern int  fssetdm_by_handle (void *__hanp, size_t __hlen,
 			       struct fsdmidata *__fsdmi);

Index: b/include/jdm.h
===================================================================
@@ -24,7 +24,7 @@ typedef void	jdm_filehandle_t;	/* fileha
 
 struct xfs_bstat;
 struct attrlist_cursor;
-struct parent;
+struct xfs_parent;
 
 extern jdm_fshandle_t *
 jdm_getfshandle( char *mntpnt);

@@ -62,16 +62,12 @@ jdm_attr_list(	jdm_fshandle_t *fshp,
 		struct attrlist_cursor *cursor);
 
 extern int
-jdm_parents( jdm_fshandle_t *fshp,
-		xfs_bstat_t *statp,
-		struct parent *bufp, size_t bufsz,
-		unsigned int *count);
+jdm_parents( jdm_fshandle_t *fshp, struct xfs_bstat *statp,
+	     struct xfs_parent *bufp, size_t bufsz, unsigned int *count);
 
 extern int
-jdm_parentpaths( jdm_fshandle_t *fshp,
-		xfs_bstat_t *statp,
-		struct parent *bufp, size_t bufsz,
-		unsigned int *count);
+jdm_parentpaths( jdm_fshandle_t *fshp, struct xfs_bstat *statp,
+	         struct xfs_parent *bufp, size_t bufsz, unsigned int *count);
 
 /* macro for determining the size of a structure member */
 #define sizeofmember( t, m )	sizeof( ( ( t * )0 )->m )

Index: b/include/parent.h
===================================================================
@@ -18,12 +18,13 @@
 #ifndef __PARENT_H__
 #define	__PARENT_H__
 
-typedef struct parent {
-	__u64	p_ino;
-	__u32	p_gen;
-	__u16	p_reclen;
-	char	p_name[1];
-} parent_t;
+struct xfs_parent {
+	uint64_t	p_ino;          /*parent inode number */
+	uint32_t	p_offset;       /* entry offset in parent inode */
+	uint16_t	p_reclen;       /* name length */
+	uint16_t	p_spare;        /* padding for future */
+	char		p_name[1];      /* name */
+};
 
 typedef struct parent_cursor {
 	__u32	opaque[4];      /* an opaque cookie */

Index: b/include/xfs_fs.h
===================================================================
@@ -438,6 +438,15 @@ typedef struct xfs_fsop_attrmulti_handle
 	struct xfs_attr_multiop		__user *ops; /* attr_multi data */
 } xfs_fsop_attrmulti_handlereq_t;
 
+
+typedef struct xfs_fsop_parentlist_handlereq {
+	struct xfs_fsop_handlereq	hreq;	/* handle interface structure */
+	__u32				flags;	/* which namespace to use */
+	__u32				buflen;	/* length of buffer supplied */
+	__u32				__user *ocount;	/* output count ptr */
+	void				__user *buffer;	/* returned names */
+} xfs_fsop_parentlist_handlereq_t;
+
 /*
  * per machine unique filesystem identifier types.
  */

@@ -553,6 +562,8 @@ typedef struct xfs_swapext
 #define XFS_IOC_ATTRMULTI_BY_HANDLE  _IOW ('X', 123, struct xfs_fsop_attrmulti_handlereq)
 #define XFS_IOC_FSGEOMETRY	     _IOR ('X', 124, struct xfs_fsop_geom)
 #define XFS_IOC_GOINGDOWN	     _IOR ('X', 125, __uint32_t)
+#define XFS_IOC_GETPARENTS_BY_HANDLE	 _IOWR('X', 126, struct xfs_fsop_parentlist_handlereq)
+#define XFS_IOC_GETPARENTPATHS_BY_HANDLE _IOWR('X', 127, struct xfs_fsop_parentlist_handlereq)
 /*	XFS_IOC_GETFSUUID ---------- deprecated 140	 */
 
 

---
 growfs/xfs_growfs.c |   14 +++++++++-----
 include/xfs_fs.h    |    2 +-
 2 files changed, 10 insertions(+), 6 deletions(-)
Index: b/growfs/xfs_growfs.c
===================================================================
@@ -56,7 +56,8 @@ report_info(
 	int		projid32bit,
 	int		crcs_enabled,
 	int		cimode,
-	int		ftype_enabled)
+	int		ftype_enabled,
+	int		parentptr)
 {
 	printf(_(
 	    "meta-data=%-22s isize=%-6u agcount=%u, agsize=%u blks\n"

@@ -64,7 +65,7 @@ report_info(
 	    "         =%-22s crc=%u\n"
 	    "data     =%-22s bsize=%-6u blocks=%llu, imaxpct=%u\n"
 	    "         =%-22s sunit=%-6u swidth=%u blks\n"
-	    "naming   =version %-14u bsize=%-6u ascii-ci=%d ftype=%d\n"
+	    "naming   =version %-14u bsize=%-6u ascii-ci=%d ftype=%d parent=%d\n"
 	    "log      =%-22s bsize=%-6u blocks=%u, version=%u\n"
 	    "         =%-22s sectsz=%-5u sunit=%u blks, lazy-count=%u\n"
 	    "realtime =%-22s extsz=%-6u blocks=%llu, rtextents=%llu\n"),

@@ -75,7 +76,7 @@ report_info(
 		"", geo.blocksize, (unsigned long long)geo.datablocks,
 			geo.imaxpct,
 		"", geo.sunit, geo.swidth,
-  		dirversion, geo.dirblocksize, cimode, ftype_enabled,
+  		dirversion, geo.dirblocksize, cimode, ftype_enabled, parentptr,
 		isint ? _("internal") : logname ? logname : _("external"),
 			geo.blocksize, geo.logblocks, logversion,
 		"", geo.logsectsize, geo.logsunit / geo.blocksize, lazycount,

@@ -123,6 +124,7 @@ main(int argc, char **argv)
 	int			projid32bit;
 	int			crcs_enabled;
 	int			ftype_enabled = 0;
+	int			parent_enabled = 0;
 
 	progname = basename(argv[0]);
 	setlocale(LC_ALL, "");

@@ -245,11 +247,12 @@ main(int argc, char **argv)
 	projid32bit = geo.flags & XFS_FSOP_GEOM_FLAGS_PROJID32 ? 1 : 0;
 	crcs_enabled = geo.flags & XFS_FSOP_GEOM_FLAGS_V5SB ? 1 : 0;
 	ftype_enabled = geo.flags & XFS_FSOP_GEOM_FLAGS_FTYPE ? 1 : 0;
+	parent_enabled = geo.flags & XFS_FSOP_GEOM_FLAGS_PARENT ? 1 : 0;
 	if (nflag) {
 		report_info(geo, datadev, isint, logdev, rtdev,
 				lazycount, dirversion, logversion,
 				attrversion, projid32bit, crcs_enabled, ci,
-				ftype_enabled);
+				ftype_enabled, parent_enabled);
 		exit(0);
 	}
 

@@ -286,7 +289,8 @@ main(int argc, char **argv)
 
 	report_info(geo, datadev, isint, logdev, rtdev,
 			lazycount, dirversion, logversion,
-			attrversion, projid32bit, crcs_enabled, ci, ftype_enabled);
+			attrversion, projid32bit, crcs_enabled, ci,
+			ftype_enabled, parent_enabled);
 
 	ddsize = xi.dsize;
 	dlsize = ( xi.logBBsize? xi.logBBsize :

Index: b/include/xfs_fs.h
===================================================================
@@ -238,7 +238,7 @@ typedef struct xfs_fsop_resblks {
 #define XFS_FSOP_GEOM_FLAGS_LAZYSB	0x4000	/* lazy superblock counters */
 #define XFS_FSOP_GEOM_FLAGS_V5SB	0x8000	/* version 5 superblock */
 #define XFS_FSOP_GEOM_FLAGS_FTYPE	0x10000	/* inode directory types */
-
+#define XFS_FSOP_GEOM_FLAGS_PARENT	0x20000	/* parent inode support */
 
 /*
  * Minimum and maximum sizes need for growth checks.

	http://thread.gmane.org/gmane.comp.file-systems.xfs.general/27772
Dave Chinner
[email protected]

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
	http://thread.gmane.org/gmane.comp.file-systems.xfs.general/27772
	http://thread.gmane.org/gmane.comp.file-systems.xfs.general/27772
Dave Chinner
[email protected]

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
Dave Chinner
[email protected]

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
| On Mon, Jan 27, 2014 at 01:41:47PM -0600, Mark Tinguely wrote:
| > 2) Add the filename to EA. Not a fan, but I will ask but if DMF needs it
| >    for performance then it has to be done. My point was this assumes
| >    that we can keep all the links' EA entries inline in the inode. A
| >    couple 255 character files or several links of modest sized filenames
| >    would negate that assumption. I tried to minimize the EA entries to
| >    keep them inline in the inode. I will talk to the DMF group.
| 
| Actually, I made the point about DMF needing them inline performance
| because that's an argument SGI might find persuasive. What I didn't
| say just then is that *I* need them inline, too, as online directory
| tree scrubbing needs to be able to do bulks scans, as does
| xfs_repair. However, I have idefinitely said this before in previous
| parent poitner discussions, so it should be no surprise here...
Geoffrey Wehrman

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
| On Mon, Jan 27, 2014 at 01:41:47PM -0600, Mark Tinguely wrote:
| > 2) Add the filename to EA. Not a fan, but I will ask but if DMF needs it
| >    for performance then it has to be done. My point was this assumes
| >    that we can keep all the links' EA entries inline in the inode. A
| >    couple 255 character files or several links of modest sized filenames
| >    would negate that assumption. I tried to minimize the EA entries to
| >    keep them inline in the inode. I will talk to the DMF group.
| 
| Actually, I made the point about DMF needing them inline performance
| because that's an argument SGI might find persuasive. What I didn't
| say just then is that *I* need them inline, too, as online directory
| tree scrubbing needs to be able to do bulks scans, as does
| xfs_repair. However, I have idefinitely said this before in previous
| parent poitner discussions, so it should be no surprise here...
Dave Chinner
[email protected]

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs
| On Tue, Jan 28, 2014 at 04:02:30PM -0600, Geoffrey Wehrman wrote:
| > On Tue, Jan 28, 2014 at 02:00:40PM +1100, Dave Chinner wrote:
| > | On Mon, Jan 27, 2014 at 01:41:47PM -0600, Mark Tinguely wrote:
| > | > 2) Add the filename to EA. Not a fan, but I will ask but if DMF needs it
| > | >    for performance then it has to be done. My point was this assumes
| > | >    that we can keep all the links' EA entries inline in the inode. A
| > | >    couple 255 character files or several links of modest sized filenames
| > | >    would negate that assumption. I tried to minimize the EA entries to
| > | >    keep them inline in the inode. I will talk to the DMF group.
| > | 
| > | Actually, I made the point about DMF needing them inline performance
| > | because that's an argument SGI might find persuasive. What I didn't
| > | say just then is that *I* need them inline, too, as online directory
| > | tree scrubbing needs to be able to do bulks scans, as does
| > | xfs_repair. However, I have idefinitely said this before in previous
| > | parent poitner discussions, so it should be no surprise here...
| > 
| > I appologize in advance for my ignorance.  What is "online directory
| > tree scrubbing" and how does it benefit from parent inode pointers?
| 
| It will be a kernel thread that walks the directory tree
| periodically verifying that all inodes are reachable and that the
| directory tree is intact. It's the same principle as RAID array
| media scrubbing - proactive detection of errors that could cause
| failures and repairing them before an error is delivered to the
| application. The reason for wanting parent pointers is that it
| enables immediate repair of most corruptions as the scrubbing
| detects them without requiring duplicate copies of the metadata.
| 
| IOWs, parent pointers are a fundamental part of bringing on-line
| repair functionality to XFS. While parent pointers are not directly
| mentioned in this document:
| 
| http://xfs.org/index.php/Reliable_Detection_and_Repair_of_Metadata_Corruption
| 
| There is a section on reverse mapping for used blocks and how to use
| that for online repair of detected damage. Parent pointers provide
| reverse mappings for the directory structure, and hence provide the
| same functionality. Parent pointers also means that if we get an IO
| error in a block in a file, we can identify the owner of the block
| by full path and offset into the file where the error has
| occurred...
| 
| The v5 metadata format has an owner field in all metadata that has a
| single parent object for keeping such parent information, and it is
| intended for improving scrub-based validation and for identifying
| the owner of lost metadata blocks to improve recover from errors.
| i.e. reverse mapping is a one of the fundamental architectural
| requirements underlying the v5 metadata changes.
| 
| The two missing pieces are the parent pointers for directory
| structure reverse mapping, and an AGF reverse mapping btree to
| enable arbitrary block->owner lookups. I have prototype code for the
| AGF rmap btree, and the parent pointers provide that functionality
| for the directory structure. With both of these implemented, we will
| be able to do fully-online filesystem metadata validity checking
| equivalent to 'xfs_repair -n' or better.....
| 
| Reverse block mapping and directory structure parent pointers have
| been considered necessary for robust exception handling and online
| repair since we first started thinking about the CRC metadata format
| changes....
Geoffrey Wehrman  651-683-5496  [email protected]

_______________________________________________
xfs mailing list
[email protected]
http://oss.sgi.com/mailman/listinfo/xfs