Merge tag 'xfs-for-linus-3.16-rc1' of git://oss.sgi.com/xfs/xfs

Pull xfs updates from Dave Chinner:
 "This update contains:
   - cleanup removing unused function args
   - rework of the filestreams allocator to use dentry cache parent
     lookups
   - new on-disk free inode btree and optimised inode allocator
   - various bug fixes
   - rework of internal attribute API
   - cleanup of superblock feature bit support to remove historic cruft
   - more fixes and minor cleanups
   - added a new directory/attribute geometry abstraction
   - yet more fixes and minor cleanups"

* tag 'xfs-for-linus-3.16-rc1' of git://oss.sgi.com/xfs/xfs: (86 commits)
  xfs: fix xfs_da_args sparse warning in xfs_readdir
  xfs: Fix rounding in xfs_alloc_fix_len()
  xfs: tone down writepage/releasepage WARN_ONs
  xfs: small cleanup in xfs_lowbit64()
  xfs: kill xfs_buf_geterror()
  xfs: xfs_readsb needs to check for magic numbers
  xfs: block allocation work needs to be kswapd aware
  xfs: remove redundant geometry information from xfs_da_state
  xfs: replace attr LBSIZE with xfs_da_geometry
  xfs: pass xfs_da_args to xfs_attr_leaf_newentsize
  xfs: use xfs_da_geometry for block size in attr code
  xfs: remove mp->m_dir_geo from directory logging
  xfs: reduce direct usage of mp->m_dir_geo
  xfs: move node entry counts to xfs_da_geometry
  xfs: convert dir/attr btree threshold to xfs_da_geometry
  xfs: convert m_dirblksize to xfs_da_geometry
  xfs: convert m_dirblkfsbs to xfs_da_geometry
  xfs: convert directory segment limits to xfs_da_geometry
  xfs: convert directory db conversion to xfs_da_geometry
  xfs: convert directory dablk conversion to xfs_da_geometry
  ...

fs/quota/quota.c

@@ -278,6 +278,17 @@ static int quota_getxquota(struct super_block *sb, int type, qid_t id,
 	return ret;
 }
 
+static int quota_rmxquota(struct super_block *sb, void __user *addr)
+{
+	__u32 flags;
+
+	if (copy_from_user(&flags, addr, sizeof(flags)))
+		return -EFAULT;
+	if (!sb->s_qcop->rm_xquota)
+		return -ENOSYS;
+	return sb->s_qcop->rm_xquota(sb, flags);
+}
+
 /* Copy parameters and call proper function */
 static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id,
 		       void __user *addr, struct path *path)
@@ -316,8 +327,9 @@ static int do_quotactl(struct super_block *sb, int type, int cmd, qid_t id,
 		return sb->s_qcop->quota_sync(sb, type);
 	case Q_XQUOTAON:
 	case Q_XQUOTAOFF:
-	case Q_XQUOTARM:
 		return quota_setxstate(sb, cmd, addr);
+	case Q_XQUOTARM:
+		return quota_rmxquota(sb, addr);
 	case Q_XGETQSTAT:
 		return quota_getxstate(sb, addr);
 	case Q_XGETQSTATV:

fs/xfs/xfs_ag.h

@@ -160,30 +160,38 @@ typedef struct xfs_agi {
 	 * still being referenced.
 	 */
 	__be32		agi_unlinked[XFS_AGI_UNLINKED_BUCKETS];
-
+	/*
+	 * This marks the end of logging region 1 and start of logging region 2.
+	 */
 	uuid_t		agi_uuid;	/* uuid of filesystem */
 	__be32		agi_crc;	/* crc of agi sector */
 	__be32		agi_pad32;
 	__be64		agi_lsn;	/* last write sequence */
 
+	__be32		agi_free_root; /* root of the free inode btree */
+	__be32		agi_free_level;/* levels in free inode btree */
+
 	/* structure must be padded to 64 bit alignment */
 } xfs_agi_t;
 
 #define XFS_AGI_CRC_OFF		offsetof(struct xfs_agi, agi_crc)
 
-#define	XFS_AGI_MAGICNUM	0x00000001
-#define	XFS_AGI_VERSIONNUM	0x00000002
-#define	XFS_AGI_SEQNO		0x00000004
-#define	XFS_AGI_LENGTH		0x00000008
-#define	XFS_AGI_COUNT		0x00000010
-#define	XFS_AGI_ROOT		0x00000020
-#define	XFS_AGI_LEVEL		0x00000040
-#define	XFS_AGI_FREECOUNT	0x00000080
-#define	XFS_AGI_NEWINO		0x00000100
-#define	XFS_AGI_DIRINO		0x00000200
-#define	XFS_AGI_UNLINKED	0x00000400
-#define	XFS_AGI_NUM_BITS	11
-#define	XFS_AGI_ALL_BITS	((1 << XFS_AGI_NUM_BITS) - 1)
+#define	XFS_AGI_MAGICNUM	(1 << 0)
+#define	XFS_AGI_VERSIONNUM	(1 << 1)
+#define	XFS_AGI_SEQNO		(1 << 2)
+#define	XFS_AGI_LENGTH		(1 << 3)
+#define	XFS_AGI_COUNT		(1 << 4)
+#define	XFS_AGI_ROOT		(1 << 5)
+#define	XFS_AGI_LEVEL		(1 << 6)
+#define	XFS_AGI_FREECOUNT	(1 << 7)
+#define	XFS_AGI_NEWINO		(1 << 8)
+#define	XFS_AGI_DIRINO		(1 << 9)
+#define	XFS_AGI_UNLINKED	(1 << 10)
+#define	XFS_AGI_NUM_BITS_R1	11	/* end of the 1st agi logging region */
+#define	XFS_AGI_ALL_BITS_R1	((1 << XFS_AGI_NUM_BITS_R1) - 1)
+#define	XFS_AGI_FREE_ROOT	(1 << 11)
+#define	XFS_AGI_FREE_LEVEL	(1 << 12)
+#define	XFS_AGI_NUM_BITS_R2	13
 
 /* disk block (xfs_daddr_t) in the AG */
 #define XFS_AGI_DADDR(mp)	((xfs_daddr_t)(2 << (mp)->m_sectbb_log))

fs/xfs/xfs_alloc.c

@@ -257,16 +257,14 @@ xfs_alloc_fix_len(
 	k = rlen % args->prod;
 	if (k == args->mod)
 		return;
-	if (k > args->mod) {
-		if ((int)(rlen = rlen - k - args->mod) < (int)args->minlen)
-			return;
-	} else {
-		if ((int)(rlen = rlen - args->prod - (args->mod - k)) <
-		    (int)args->minlen)
-			return;
-	}
-	ASSERT(rlen >= args->minlen);
-	ASSERT(rlen <= args->maxlen);
+	if (k > args->mod)
+		rlen = rlen - (k - args->mod);
+	else
+		rlen = rlen - args->prod + (args->mod - k);
+	if ((int)rlen < (int)args->minlen)
+		return;
+	ASSERT(rlen >= args->minlen && rlen <= args->maxlen);
+	ASSERT(rlen % args->prod == args->mod);
 	args->len = rlen;
 }
 
@@ -541,7 +539,6 @@ xfs_alloc_read_agfl(
 			XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_agfl_buf_ops);
 	if (error)
 		return error;
-	ASSERT(!xfs_buf_geterror(bp));
 	xfs_buf_set_ref(bp, XFS_AGFL_REF);
 	*bpp = bp;
 	return 0;

fs/xfs/xfs_alloc_btree.c

@@ -70,7 +70,6 @@ xfs_allocbt_alloc_block(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*start,
 	union xfs_btree_ptr	*new,
-	int			length,
 	int			*stat)
 {
 	int			error;

fs/xfs/xfs_aops.c

@@ -975,14 +975,39 @@ xfs_vm_writepage(
 	 * Given that we do not allow direct reclaim to call us, we should
 	 * never be called while in a filesystem transaction.
 	 */
-	if (WARN_ON(current->flags & PF_FSTRANS))
+	if (WARN_ON_ONCE(current->flags & PF_FSTRANS))
 		goto redirty;
 
 	/* Is this page beyond the end of the file? */
 	offset = i_size_read(inode);
 	end_index = offset >> PAGE_CACHE_SHIFT;
 	last_index = (offset - 1) >> PAGE_CACHE_SHIFT;
-	if (page->index >= end_index) {
+
+	/*
+	 * The page index is less than the end_index, adjust the end_offset
+	 * to the highest offset that this page should represent.
+	 * -----------------------------------------------------
+	 * |			file mapping	       | <EOF> |
+	 * -----------------------------------------------------
+	 * | Page ... | Page N-2 | Page N-1 |  Page N  |       |
+	 * ^--------------------------------^----------|--------
+	 * |     desired writeback range    |      see else    |
+	 * ---------------------------------^------------------|
+	 */
+	if (page->index < end_index)
+		end_offset = (xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT;
+	else {
+		/*
+		 * Check whether the page to write out is beyond or straddles
+		 * i_size or not.
+		 * -------------------------------------------------------
+		 * |		file mapping		        | <EOF>  |
+		 * -------------------------------------------------------
+		 * | Page ... | Page N-2 | Page N-1 |  Page N   | Beyond |
+		 * ^--------------------------------^-----------|---------
+		 * |				    |      Straddles     |
+		 * ---------------------------------^-----------|--------|
+		 */
 		unsigned offset_into_page = offset & (PAGE_CACHE_SIZE - 1);
 
 		/*
@@ -990,24 +1015,36 @@ xfs_vm_writepage(
 		 * truncate operation that is in progress. We must redirty the
 		 * page so that reclaim stops reclaiming it. Otherwise
 		 * xfs_vm_releasepage() is called on it and gets confused.
+		 *
+		 * Note that the end_index is unsigned long, it would overflow
+		 * if the given offset is greater than 16TB on 32-bit system
+		 * and if we do check the page is fully outside i_size or not
+		 * via "if (page->index >= end_index + 1)" as "end_index + 1"
+		 * will be evaluated to 0.  Hence this page will be redirtied
+		 * and be written out repeatedly which would result in an
+		 * infinite loop, the user program that perform this operation
+		 * will hang.  Instead, we can verify this situation by checking
+		 * if the page to write is totally beyond the i_size or if it's
+		 * offset is just equal to the EOF.
 		 */
-		if (page->index >= end_index + 1 || offset_into_page == 0)
+		if (page->index > end_index ||
+		    (page->index == end_index && offset_into_page == 0))
 			goto redirty;
 
 		/*
 		 * The page straddles i_size.  It must be zeroed out on each
 		 * and every writepage invocation because it may be mmapped.
 		 * "A file is mapped in multiples of the page size.  For a file
-		 * that is not a multiple of the  page size, the remaining
+		 * that is not a multiple of the page size, the remaining
 		 * memory is zeroed when mapped, and writes to that region are
 		 * not written out to the file."
 		 */
 		zero_user_segment(page, offset_into_page, PAGE_CACHE_SIZE);
+
+		/* Adjust the end_offset to the end of file */
+		end_offset = offset;
 	}
 
-	end_offset = min_t(unsigned long long,
-			(xfs_off_t)(page->index + 1) << PAGE_CACHE_SHIFT,
-			offset);
 	len = 1 << inode->i_blkbits;
 
 	bh = head = page_buffers(page);
@@ -1188,9 +1225,9 @@ xfs_vm_releasepage(
 
 	xfs_count_page_state(page, &delalloc, &unwritten);
 
-	if (WARN_ON(delalloc))
+	if (WARN_ON_ONCE(delalloc))
 		return 0;
-	if (WARN_ON(unwritten))
+	if (WARN_ON_ONCE(unwritten))
 		return 0;
 
 	return try_to_free_buffers(page);

fs/xfs/xfs_attr.c

@@ -77,17 +77,27 @@ STATIC int xfs_attr_refillstate(xfs_da_state_t *state);
 
 
 STATIC int
-xfs_attr_name_to_xname(
-	struct xfs_name	*xname,
-	const unsigned char *aname)
+xfs_attr_args_init(
+	struct xfs_da_args	*args,
+	struct xfs_inode	*dp,
+	const unsigned char	*name,
+	int			flags)
 {
-	if (!aname)
+
+	if (!name)
 		return EINVAL;
-	xname->name = aname;
-	xname->len = strlen((char *)aname);
-	if (xname->len >= MAXNAMELEN)
+
+	memset(args, 0, sizeof(*args));
+	args->geo = dp->i_mount->m_attr_geo;
+	args->whichfork = XFS_ATTR_FORK;
+	args->dp = dp;
+	args->flags = flags;
+	args->name = name;
+	args->namelen = strlen((const char *)name);
+	if (args->namelen >= MAXNAMELEN)
 		return EFAULT;		/* match IRIX behaviour */
 
+	args->hashval = xfs_da_hashname(args->name, args->namelen);
 	return 0;
 }
 
@@ -106,79 +116,46 @@ xfs_inode_hasattr(
  * Overall external interface routines.
  *========================================================================*/
 
-STATIC int
-xfs_attr_get_int(
+int
+xfs_attr_get(
 	struct xfs_inode	*ip,
-	struct xfs_name		*name,
+	const unsigned char	*name,
 	unsigned char		*value,
 	int			*valuelenp,
 	int			flags)
 {
-	xfs_da_args_t   args;
-	int             error;
+	struct xfs_da_args	args;
+	uint			lock_mode;
+	int			error;
+
+	XFS_STATS_INC(xs_attr_get);
+
+	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+		return EIO;
 
 	if (!xfs_inode_hasattr(ip))
 		return ENOATTR;
 
-	/*
-	 * Fill in the arg structure for this request.
-	 */
-	memset((char *)&args, 0, sizeof(args));
-	args.name = name->name;
-	args.namelen = name->len;
+	error = xfs_attr_args_init(&args, ip, name, flags);
+	if (error)
+		return error;
+
 	args.value = value;
 	args.valuelen = *valuelenp;
-	args.flags = flags;
-	args.hashval = xfs_da_hashname(args.name, args.namelen);
-	args.dp = ip;
-	args.whichfork = XFS_ATTR_FORK;
 
-	/*
-	 * Decide on what work routines to call based on the inode size.
-	 */
-	if (ip->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
+	lock_mode = xfs_ilock_attr_map_shared(ip);
+	if (!xfs_inode_hasattr(ip))
+		error = ENOATTR;
+	else if (ip->i_d.di_aformat == XFS_DINODE_FMT_LOCAL)
 		error = xfs_attr_shortform_getvalue(&args);
-	} else if (xfs_bmap_one_block(ip, XFS_ATTR_FORK)) {
+	else if (xfs_bmap_one_block(ip, XFS_ATTR_FORK))
 		error = xfs_attr_leaf_get(&args);
-	} else {
+	else
 		error = xfs_attr_node_get(&args);
-	}
+	xfs_iunlock(ip, lock_mode);
 
-	/*
-	 * Return the number of bytes in the value to the caller.
-	 */
 	*valuelenp = args.valuelen;
-
-	if (error == EEXIST)
-		error = 0;
-	return(error);
-}
-
-int
-xfs_attr_get(
-	xfs_inode_t	*ip,
-	const unsigned char *name,
-	unsigned char	*value,
-	int		*valuelenp,
-	int		flags)
-{
-	int		error;
-	struct xfs_name	xname;
-	uint		lock_mode;
-
-	XFS_STATS_INC(xs_attr_get);
-
-	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
-		return(EIO);
-
-	error = xfs_attr_name_to_xname(&xname, name);
-	if (error)
-		return error;
-
-	lock_mode = xfs_ilock_attr_map_shared(ip);
-	error = xfs_attr_get_int(ip, &xname, value, valuelenp, flags);
-	xfs_iunlock(ip, lock_mode);
-	return(error);
+	return error == EEXIST ? 0 : error;
 }
 
 /*
@@ -186,12 +163,10 @@ xfs_attr_get(
  */
 STATIC int
 xfs_attr_calc_size(
-	struct xfs_inode 	*ip,
-	int			namelen,
-	int			valuelen,
+	struct xfs_da_args	*args,
 	int			*local)
 {
-	struct xfs_mount 	*mp = ip->i_mount;
+	struct xfs_mount	*mp = args->dp->i_mount;
 	int			size;
 	int			nblks;
 
@@ -199,12 +174,10 @@ xfs_attr_calc_size(
 	 * Determine space new attribute will use, and if it would be
 	 * "local" or "remote" (note: local != inline).
 	 */
-	size = xfs_attr_leaf_newentsize(namelen, valuelen,
-					mp->m_sb.sb_blocksize, local);
-
+	size = xfs_attr_leaf_newentsize(args, local);
 	nblks = XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK);
 	if (*local) {
-		if (size > (mp->m_sb.sb_blocksize >> 1)) {
+		if (size > (args->geo->blksize / 2)) {
 			/* Double split possible */
 			nblks *= 2;
 		}
@@ -213,7 +186,7 @@ xfs_attr_calc_size(
 		 * Out of line attribute, cannot double split, but
 		 * make room for the attribute value itself.
 		 */
-		uint	dblocks = xfs_attr3_rmt_blocks(mp, valuelen);
+		uint	dblocks = xfs_attr3_rmt_blocks(mp, args->valuelen);
 		nblks += dblocks;
 		nblks += XFS_NEXTENTADD_SPACE_RES(mp, dblocks, XFS_ATTR_FORK);
 	}
@@ -221,26 +194,38 @@ xfs_attr_calc_size(
 	return nblks;
 }
 
-STATIC int
-xfs_attr_set_int(
-	struct xfs_inode *dp,
-	struct xfs_name	*name,
-	unsigned char	*value,
-	int		valuelen,
-	int		flags)
+int
+xfs_attr_set(
+	struct xfs_inode	*dp,
+	const unsigned char	*name,
+	unsigned char		*value,
+	int			valuelen,
+	int			flags)
 {
-	xfs_da_args_t		args;
-	xfs_fsblock_t		firstblock;
-	xfs_bmap_free_t		flist;
-	int			error, err2, committed;
 	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_da_args	args;
+	struct xfs_bmap_free	flist;
 	struct xfs_trans_res	tres;
+	xfs_fsblock_t		firstblock;
 	int			rsvd = (flags & ATTR_ROOT) != 0;
-	int			local;
+	int			error, err2, committed, local;
+
+	XFS_STATS_INC(xs_attr_set);
+
+	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
+		return EIO;
+
+	error = xfs_attr_args_init(&args, dp, name, flags);
+	if (error)
+		return error;
+
+	args.value = value;
+	args.valuelen = valuelen;
+	args.firstblock = &firstblock;
+	args.flist = &flist;
+	args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT;
+	args.total = xfs_attr_calc_size(&args, &local);
 
-	/*
-	 * Attach the dquots to the inode.
-	 */
 	error = xfs_qm_dqattach(dp, 0);
 	if (error)
 		return error;
@@ -251,31 +236,13 @@ xfs_attr_set_int(
 	 */
 	if (XFS_IFORK_Q(dp) == 0) {
 		int sf_size = sizeof(xfs_attr_sf_hdr_t) +
-			      XFS_ATTR_SF_ENTSIZE_BYNAME(name->len, valuelen);
+			XFS_ATTR_SF_ENTSIZE_BYNAME(args.namelen, valuelen);
 
-		if ((error = xfs_bmap_add_attrfork(dp, sf_size, rsvd)))
-			return(error);
+		error = xfs_bmap_add_attrfork(dp, sf_size, rsvd);
+		if (error)
+			return error;
 	}
 
-	/*
-	 * Fill in the arg structure for this request.
-	 */
-	memset((char *)&args, 0, sizeof(args));
-	args.name = name->name;
-	args.namelen = name->len;
-	args.value = value;
-	args.valuelen = valuelen;
-	args.flags = flags;
-	args.hashval = xfs_da_hashname(args.name, args.namelen);
-	args.dp = dp;
-	args.firstblock = &firstblock;
-	args.flist = &flist;
-	args.whichfork = XFS_ATTR_FORK;
-	args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT;
-
-	/* Size is now blocks for attribute data */
-	args.total = xfs_attr_calc_size(dp, name->len, valuelen, &local);
-
 	/*
 	 * Start our first transaction of the day.
 	 *
@@ -303,7 +270,7 @@ xfs_attr_set_int(
 	error = xfs_trans_reserve(args.trans, &tres, args.total, 0);
 	if (error) {
 		xfs_trans_cancel(args.trans, 0);
-		return(error);
+		return error;
 	}
 	xfs_ilock(dp, XFS_ILOCK_EXCL);
 
@@ -313,7 +280,7 @@ xfs_attr_set_int(
 	if (error) {
 		xfs_iunlock(dp, XFS_ILOCK_EXCL);
 		xfs_trans_cancel(args.trans, XFS_TRANS_RELEASE_LOG_RES);
-		return (error);
+		return error;
 	}
 
 	xfs_trans_ijoin(args.trans, dp, 0);
@@ -322,9 +289,9 @@ xfs_attr_set_int(
 	 * If the attribute list is non-existent or a shortform list,
 	 * upgrade it to a single-leaf-block attribute list.
 	 */
-	if ((dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) ||
-	    ((dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS) &&
-	     (dp->i_d.di_anextents == 0))) {
+	if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL ||
+	    (dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
+	     dp->i_d.di_anextents == 0)) {
 
 		/*
 		 * Build initial attribute list (if required).
@@ -349,9 +316,8 @@ xfs_attr_set_int(
 			 * the transaction goes to disk before returning
 			 * to the user.
 			 */
-			if (mp->m_flags & XFS_MOUNT_WSYNC) {
+			if (mp->m_flags & XFS_MOUNT_WSYNC)
 				xfs_trans_set_sync(args.trans);
-			}
 
 			if (!error && (flags & ATTR_KERNOTIME) == 0) {
 				xfs_trans_ichgtime(args.trans, dp,
@@ -361,7 +327,7 @@ xfs_attr_set_int(
 						 XFS_TRANS_RELEASE_LOG_RES);
 			xfs_iunlock(dp, XFS_ILOCK_EXCL);
 
-			return(error == 0 ? err2 : error);
+			return error ? error : err2;
 		}
 
 		/*
@@ -399,22 +365,19 @@ xfs_attr_set_int(
 
 	}
 
-	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
+	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK))
 		error = xfs_attr_leaf_addname(&args);
-	} else {
+	else
 		error = xfs_attr_node_addname(&args);
-	}
-	if (error) {
+	if (error)
 		goto out;
-	}
 
 	/*
 	 * If this is a synchronous mount, make sure that the
 	 * transaction goes to disk before returning to the user.
 	 */
-	if (mp->m_flags & XFS_MOUNT_WSYNC) {
+	if (mp->m_flags & XFS_MOUNT_WSYNC)
 		xfs_trans_set_sync(args.trans);
-	}
 
 	if ((flags & ATTR_KERNOTIME) == 0)
 		xfs_trans_ichgtime(args.trans, dp, XFS_ICHGTIME_CHG);
@@ -426,65 +389,47 @@ xfs_attr_set_int(
 	error = xfs_trans_commit(args.trans, XFS_TRANS_RELEASE_LOG_RES);
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);
 
-	return(error);
+	return error;
 
 out:
-	if (args.trans)
+	if (args.trans) {
 		xfs_trans_cancel(args.trans,
 			XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
+	}
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);
-	return(error);
+	return error;
 }
 
+/*
+ * Generic handler routine to remove a name from an attribute list.
+ * Transitions attribute list from Btree to shortform as necessary.
+ */
 int
-xfs_attr_set(
-	xfs_inode_t	*dp,
-	const unsigned char *name,
-	unsigned char	*value,
-	int		valuelen,
-	int		flags)
+xfs_attr_remove(
+	struct xfs_inode	*dp,
+	const unsigned char	*name,
+	int			flags)
 {
-	int             error;
-	struct xfs_name	xname;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_da_args	args;
+	struct xfs_bmap_free	flist;
+	xfs_fsblock_t		firstblock;
+	int			error;
 
-	XFS_STATS_INC(xs_attr_set);
+	XFS_STATS_INC(xs_attr_remove);
 
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
-		return (EIO);
+		return EIO;
+
+	if (!xfs_inode_hasattr(dp))
+		return ENOATTR;
 
-	error = xfs_attr_name_to_xname(&xname, name);
+	error = xfs_attr_args_init(&args, dp, name, flags);
 	if (error)
 		return error;
 
-	return xfs_attr_set_int(dp, &xname, value, valuelen, flags);
-}
-
-/*
- * Generic handler routine to remove a name from an attribute list.
- * Transitions attribute list from Btree to shortform as necessary.
- */
-STATIC int
-xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
-{
-	xfs_da_args_t	args;
-	xfs_fsblock_t	firstblock;
-	xfs_bmap_free_t	flist;
-	int		error;
-	xfs_mount_t	*mp = dp->i_mount;
-
-	/*
-	 * Fill in the arg structure for this request.
-	 */
-	memset((char *)&args, 0, sizeof(args));
-	args.name = name->name;
-	args.namelen = name->len;
-	args.flags = flags;
-	args.hashval = xfs_da_hashname(args.name, args.namelen);
-	args.dp = dp;
 	args.firstblock = &firstblock;
 	args.flist = &flist;
-	args.total = 0;
-	args.whichfork = XFS_ATTR_FORK;
 
 	/*
 	 * we have no control over the attribute names that userspace passes us
@@ -493,9 +438,6 @@ xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
 	 */
 	args.op_flags = XFS_DA_OP_OKNOENT;
 
-	/*
-	 * Attach the dquots to the inode.
-	 */
 	error = xfs_qm_dqattach(dp, 0);
 	if (error)
 		return error;
@@ -524,7 +466,7 @@ xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
 				  XFS_ATTRRM_SPACE_RES(mp), 0);
 	if (error) {
 		xfs_trans_cancel(args.trans, 0);
-		return(error);
+		return error;
 	}
 
 	xfs_ilock(dp, XFS_ILOCK_EXCL);
@@ -534,35 +476,26 @@ xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
 	 */
 	xfs_trans_ijoin(args.trans, dp, 0);
 
-	/*
-	 * Decide on what work routines to call based on the inode size.
-	 */
 	if (!xfs_inode_hasattr(dp)) {
 		error = XFS_ERROR(ENOATTR);
-		goto out;
-	}
-	if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
+	} else if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
 		ASSERT(dp->i_afp->if_flags & XFS_IFINLINE);
 		error = xfs_attr_shortform_remove(&args);
-		if (error) {
-			goto out;
-		}
 	} else if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
 		error = xfs_attr_leaf_removename(&args);
 	} else {
 		error = xfs_attr_node_removename(&args);
 	}
-	if (error) {
+
+	if (error)
 		goto out;
-	}
 
 	/*
 	 * If this is a synchronous mount, make sure that the
 	 * transaction goes to disk before returning to the user.
 	 */
-	if (mp->m_flags & XFS_MOUNT_WSYNC) {
+	if (mp->m_flags & XFS_MOUNT_WSYNC)
 		xfs_trans_set_sync(args.trans);
-	}
 
 	if ((flags & ATTR_KERNOTIME) == 0)
 		xfs_trans_ichgtime(args.trans, dp, XFS_ICHGTIME_CHG);
@@ -574,45 +507,17 @@ xfs_attr_remove_int(xfs_inode_t *dp, struct xfs_name *name, int flags)
 	error = xfs_trans_commit(args.trans, XFS_TRANS_RELEASE_LOG_RES);
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);
 
-	return(error);
+	return error;
 
 out:
-	if (args.trans)
+	if (args.trans) {
 		xfs_trans_cancel(args.trans,
 			XFS_TRANS_RELEASE_LOG_RES|XFS_TRANS_ABORT);
-	xfs_iunlock(dp, XFS_ILOCK_EXCL);
-	return(error);
-}
-
-int
-xfs_attr_remove(
-	xfs_inode_t	*dp,
-	const unsigned char *name,
-	int		flags)
-{
-	int		error;
-	struct xfs_name	xname;
-
-	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;
-
-	xfs_ilock(dp, XFS_ILOCK_SHARED);
-	if (!xfs_inode_hasattr(dp)) {
-		xfs_iunlock(dp, XFS_ILOCK_SHARED);
-		return XFS_ERROR(ENOATTR);
 	}
-	xfs_iunlock(dp, XFS_ILOCK_SHARED);
-
-	return xfs_attr_remove_int(dp, &xname, flags);
+	xfs_iunlock(dp, XFS_ILOCK_EXCL);
+	return error;
 }
 
-
 /*========================================================================
  * External routines when attribute list is inside the inode
  *========================================================================*/
@@ -958,7 +863,7 @@ xfs_attr_leaf_get(xfs_da_args_t *args)
 }
 
 /*========================================================================
- * External routines when attribute list size > XFS_LBSIZE(mp).
+ * External routines when attribute list size > geo->blksize
  *========================================================================*/
 
 /*
@@ -991,8 +896,6 @@ restart:
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = mp;
-	state->blocksize = state->mp->m_sb.sb_blocksize;
-	state->node_ents = state->mp->m_attr_node_ents;
 
 	/*
 	 * Search to see if name already exists, and get back a pointer
@@ -1170,8 +1073,6 @@ restart:
 		state = xfs_da_state_alloc();
 		state->args = args;
 		state->mp = mp;
-		state->blocksize = state->mp->m_sb.sb_blocksize;
-		state->node_ents = state->mp->m_attr_node_ents;
 		state->inleaf = 0;
 		error = xfs_da3_node_lookup_int(state, &retval);
 		if (error)
@@ -1262,8 +1163,6 @@ xfs_attr_node_removename(xfs_da_args_t *args)
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = dp->i_mount;
-	state->blocksize = state->mp->m_sb.sb_blocksize;
-	state->node_ents = state->mp->m_attr_node_ents;
 
 	/*
 	 * Search to see if name exists, and get back a pointer to it.
@@ -1525,8 +1424,6 @@ xfs_attr_node_get(xfs_da_args_t *args)
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = args->dp->i_mount;
-	state->blocksize = state->mp->m_sb.sb_blocksize;
-	state->node_ents = state->mp->m_attr_node_ents;
 
 	/*
 	 * Search to see if name exists, and get back a pointer to it.

fs/xfs/xfs_attr_leaf.c

@@ -80,11 +80,12 @@ STATIC int xfs_attr3_leaf_figure_balance(xfs_da_state_t *state,
 /*
  * Utility routines.
  */
-STATIC void xfs_attr3_leaf_moveents(struct xfs_attr_leafblock *src_leaf,
+STATIC void xfs_attr3_leaf_moveents(struct xfs_da_args *args,
+			struct xfs_attr_leafblock *src_leaf,
 			struct xfs_attr3_icleaf_hdr *src_ichdr, int src_start,
 			struct xfs_attr_leafblock *dst_leaf,
 			struct xfs_attr3_icleaf_hdr *dst_ichdr, int dst_start,
-			int move_count, struct xfs_mount *mp);
+			int move_count);
 STATIC int xfs_attr_leaf_entsize(xfs_attr_leafblock_t *leaf, int index);
 
 void
@@ -711,6 +712,7 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 
 	memset((char *)&nargs, 0, sizeof(nargs));
 	nargs.dp = dp;
+	nargs.geo = args->geo;
 	nargs.firstblock = args->firstblock;
 	nargs.flist = args->flist;
 	nargs.total = args->total;
@@ -805,18 +807,18 @@ xfs_attr3_leaf_to_shortform(
 
 	trace_xfs_attr_leaf_to_sf(args);
 
-	tmpbuffer = kmem_alloc(XFS_LBSIZE(dp->i_mount), KM_SLEEP);
+	tmpbuffer = kmem_alloc(args->geo->blksize, KM_SLEEP);
 	if (!tmpbuffer)
 		return ENOMEM;
 
-	memcpy(tmpbuffer, bp->b_addr, XFS_LBSIZE(dp->i_mount));
+	memcpy(tmpbuffer, bp->b_addr, args->geo->blksize);
 
 	leaf = (xfs_attr_leafblock_t *)tmpbuffer;
 	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
 	entry = xfs_attr3_leaf_entryp(leaf);
 
 	/* XXX (dgc): buffer is about to be marked stale - why zero it? */
-	memset(bp->b_addr, 0, XFS_LBSIZE(dp->i_mount));
+	memset(bp->b_addr, 0, args->geo->blksize);
 
 	/*
 	 * Clean out the prior contents of the attribute list.
@@ -838,6 +840,7 @@ xfs_attr3_leaf_to_shortform(
 	 * Copy the attributes
 	 */
 	memset((char *)&nargs, 0, sizeof(nargs));
+	nargs.geo = args->geo;
 	nargs.dp = dp;
 	nargs.firstblock = args->firstblock;
 	nargs.flist = args->flist;
@@ -904,12 +907,12 @@ xfs_attr3_leaf_to_node(
 	/* copy leaf to new buffer, update identifiers */
 	xfs_trans_buf_set_type(args->trans, bp2, XFS_BLFT_ATTR_LEAF_BUF);
 	bp2->b_ops = bp1->b_ops;
-	memcpy(bp2->b_addr, bp1->b_addr, XFS_LBSIZE(mp));
+	memcpy(bp2->b_addr, bp1->b_addr, args->geo->blksize);
 	if (xfs_sb_version_hascrc(&mp->m_sb)) {
 		struct xfs_da3_blkinfo *hdr3 = bp2->b_addr;
 		hdr3->blkno = cpu_to_be64(bp2->b_bn);
 	}
-	xfs_trans_log_buf(args->trans, bp2, 0, XFS_LBSIZE(mp) - 1);
+	xfs_trans_log_buf(args->trans, bp2, 0, args->geo->blksize - 1);
 
 	/*
 	 * Set up the new root node.
@@ -930,7 +933,7 @@ xfs_attr3_leaf_to_node(
 	btree[0].before = cpu_to_be32(blkno);
 	icnodehdr.count = 1;
 	dp->d_ops->node_hdr_to_disk(node, &icnodehdr);
-	xfs_trans_log_buf(args->trans, bp1, 0, XFS_LBSIZE(mp) - 1);
+	xfs_trans_log_buf(args->trans, bp1, 0, args->geo->blksize - 1);
 	error = 0;
 out:
 	return error;
@@ -966,10 +969,10 @@ xfs_attr3_leaf_create(
 	bp->b_ops = &xfs_attr3_leaf_buf_ops;
 	xfs_trans_buf_set_type(args->trans, bp, XFS_BLFT_ATTR_LEAF_BUF);
 	leaf = bp->b_addr;
-	memset(leaf, 0, XFS_LBSIZE(mp));
+	memset(leaf, 0, args->geo->blksize);
 
 	memset(&ichdr, 0, sizeof(ichdr));
-	ichdr.firstused = XFS_LBSIZE(mp);
+	ichdr.firstused = args->geo->blksize;
 
 	if (xfs_sb_version_hascrc(&mp->m_sb)) {
 		struct xfs_da3_blkinfo *hdr3 = bp->b_addr;
@@ -988,7 +991,7 @@ xfs_attr3_leaf_create(
 	ichdr.freemap[0].size = ichdr.firstused - ichdr.freemap[0].base;
 
 	xfs_attr3_leaf_hdr_to_disk(leaf, &ichdr);
-	xfs_trans_log_buf(args->trans, bp, 0, XFS_LBSIZE(mp) - 1);
+	xfs_trans_log_buf(args->trans, bp, 0, args->geo->blksize - 1);
 
 	*bpp = bp;
 	return 0;
@@ -1074,8 +1077,7 @@ xfs_attr3_leaf_add(
 	leaf = bp->b_addr;
 	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
 	ASSERT(args->index >= 0 && args->index <= ichdr.count);
-	entsize = xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
-			   args->trans->t_mountp->m_sb.sb_blocksize, NULL);
+	entsize = xfs_attr_leaf_newentsize(args, NULL);
 
 	/*
 	 * Search through freemap for first-fit on new name length.
@@ -1174,17 +1176,14 @@ xfs_attr3_leaf_add_work(
 	 * Allocate space for the new string (at the end of the run).
 	 */
 	mp = args->trans->t_mountp;
-	ASSERT(ichdr->freemap[mapindex].base < XFS_LBSIZE(mp));
+	ASSERT(ichdr->freemap[mapindex].base < args->geo->blksize);
 	ASSERT((ichdr->freemap[mapindex].base & 0x3) == 0);
 	ASSERT(ichdr->freemap[mapindex].size >=
-		xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
-					 mp->m_sb.sb_blocksize, NULL));
-	ASSERT(ichdr->freemap[mapindex].size < XFS_LBSIZE(mp));
+		xfs_attr_leaf_newentsize(args, NULL));
+	ASSERT(ichdr->freemap[mapindex].size < args->geo->blksize);
 	ASSERT((ichdr->freemap[mapindex].size & 0x3) == 0);
 
-	ichdr->freemap[mapindex].size -=
-			xfs_attr_leaf_newentsize(args->namelen, args->valuelen,
-						 mp->m_sb.sb_blocksize, &tmp);
+	ichdr->freemap[mapindex].size -= xfs_attr_leaf_newentsize(args, &tmp);
 
 	entry->nameidx = cpu_to_be16(ichdr->freemap[mapindex].base +
 				     ichdr->freemap[mapindex].size);
@@ -1269,14 +1268,13 @@ xfs_attr3_leaf_compact(
 	struct xfs_attr_leafblock *leaf_dst;
 	struct xfs_attr3_icleaf_hdr ichdr_src;
 	struct xfs_trans	*trans = args->trans;
-	struct xfs_mount	*mp = trans->t_mountp;
 	char			*tmpbuffer;
 
 	trace_xfs_attr_leaf_compact(args);
 
-	tmpbuffer = kmem_alloc(XFS_LBSIZE(mp), KM_SLEEP);
-	memcpy(tmpbuffer, bp->b_addr, XFS_LBSIZE(mp));
-	memset(bp->b_addr, 0, XFS_LBSIZE(mp));
+	tmpbuffer = kmem_alloc(args->geo->blksize, KM_SLEEP);
+	memcpy(tmpbuffer, bp->b_addr, args->geo->blksize);
+	memset(bp->b_addr, 0, args->geo->blksize);
 	leaf_src = (xfs_attr_leafblock_t *)tmpbuffer;
 	leaf_dst = bp->b_addr;
 
@@ -1289,7 +1287,7 @@ xfs_attr3_leaf_compact(
 
 	/* Initialise the incore headers */
 	ichdr_src = *ichdr_dst;	/* struct copy */
-	ichdr_dst->firstused = XFS_LBSIZE(mp);
+	ichdr_dst->firstused = args->geo->blksize;
 	ichdr_dst->usedbytes = 0;
 	ichdr_dst->count = 0;
 	ichdr_dst->holes = 0;
@@ -1304,13 +1302,13 @@ xfs_attr3_leaf_compact(
 	 * Copy all entry's in the same (sorted) order,
 	 * but allocate name/value pairs packed and in sequence.
 	 */
-	xfs_attr3_leaf_moveents(leaf_src, &ichdr_src, 0, leaf_dst, ichdr_dst, 0,
-				ichdr_src.count, mp);
+	xfs_attr3_leaf_moveents(args, leaf_src, &ichdr_src, 0,
+				leaf_dst, ichdr_dst, 0, ichdr_src.count);
 	/*
 	 * this logs the entire buffer, but the caller must write the header
 	 * back to the buffer when it is finished modifying it.
 	 */
-	xfs_trans_log_buf(trans, bp, 0, XFS_LBSIZE(mp) - 1);
+	xfs_trans_log_buf(trans, bp, 0, args->geo->blksize - 1);
 
 	kmem_free(tmpbuffer);
 }
@@ -1461,8 +1459,8 @@ xfs_attr3_leaf_rebalance(
 		/*
 		 * Move high entries from leaf1 to low end of leaf2.
 		 */
-		xfs_attr3_leaf_moveents(leaf1, &ichdr1, ichdr1.count - count,
-				leaf2, &ichdr2, 0, count, state->mp);
+		xfs_attr3_leaf_moveents(args, leaf1, &ichdr1,
+				ichdr1.count - count, leaf2, &ichdr2, 0, count);
 
 	} else if (count > ichdr1.count) {
 		/*
@@ -1490,14 +1488,14 @@ xfs_attr3_leaf_rebalance(
 		/*
 		 * Move low entries from leaf2 to high end of leaf1.
 		 */
-		xfs_attr3_leaf_moveents(leaf2, &ichdr2, 0, leaf1, &ichdr1,
-					ichdr1.count, count, state->mp);
+		xfs_attr3_leaf_moveents(args, leaf2, &ichdr2, 0, leaf1, &ichdr1,
+					ichdr1.count, count);
 	}
 
 	xfs_attr3_leaf_hdr_to_disk(leaf1, &ichdr1);
 	xfs_attr3_leaf_hdr_to_disk(leaf2, &ichdr2);
-	xfs_trans_log_buf(args->trans, blk1->bp, 0, state->blocksize-1);
-	xfs_trans_log_buf(args->trans, blk2->bp, 0, state->blocksize-1);
+	xfs_trans_log_buf(args->trans, blk1->bp, 0, args->geo->blksize - 1);
+	xfs_trans_log_buf(args->trans, blk2->bp, 0, args->geo->blksize - 1);
 
 	/*
 	 * Copy out last hashval in each block for B-tree code.
@@ -1592,11 +1590,9 @@ xfs_attr3_leaf_figure_balance(
 	max = ichdr1->count + ichdr2->count;
 	half = (max + 1) * sizeof(*entry);
 	half += ichdr1->usedbytes + ichdr2->usedbytes +
-			xfs_attr_leaf_newentsize(state->args->namelen,
-						 state->args->valuelen,
-						 state->blocksize, NULL);
+			xfs_attr_leaf_newentsize(state->args, NULL);
 	half /= 2;
-	lastdelta = state->blocksize;
+	lastdelta = state->args->geo->blksize;
 	entry = xfs_attr3_leaf_entryp(leaf1);
 	for (count = index = 0; count < max; entry++, index++, count++) {
 
@@ -1606,10 +1602,7 @@ xfs_attr3_leaf_figure_balance(
 		 */
 		if (count == blk1->index) {
 			tmp = totallen + sizeof(*entry) +
-				xfs_attr_leaf_newentsize(
-						state->args->namelen,
-						state->args->valuelen,
-						state->blocksize, NULL);
+				xfs_attr_leaf_newentsize(state->args, NULL);
 			if (XFS_ATTR_ABS(half - tmp) > lastdelta)
 				break;
 			lastdelta = XFS_ATTR_ABS(half - tmp);
@@ -1645,10 +1638,7 @@ xfs_attr3_leaf_figure_balance(
 	totallen -= count * sizeof(*entry);
 	if (foundit) {
 		totallen -= sizeof(*entry) +
-				xfs_attr_leaf_newentsize(
-						state->args->namelen,
-						state->args->valuelen,
-						state->blocksize, NULL);
+				xfs_attr_leaf_newentsize(state->args, NULL);
 	}
 
 	*countarg = count;
@@ -1700,7 +1690,7 @@ xfs_attr3_leaf_toosmall(
 	bytes = xfs_attr3_leaf_hdr_size(leaf) +
 		ichdr.count * sizeof(xfs_attr_leaf_entry_t) +
 		ichdr.usedbytes;
-	if (bytes > (state->blocksize >> 1)) {
+	if (bytes > (state->args->geo->blksize >> 1)) {
 		*action = 0;	/* blk over 50%, don't try to join */
 		return(0);
 	}
@@ -1754,7 +1744,8 @@ xfs_attr3_leaf_toosmall(
 
 		xfs_attr3_leaf_hdr_from_disk(&ichdr2, bp->b_addr);
 
-		bytes = state->blocksize - (state->blocksize >> 2) -
+		bytes = state->args->geo->blksize -
+			(state->args->geo->blksize >> 2) -
 			ichdr.usedbytes - ichdr2.usedbytes -
 			((ichdr.count + ichdr2.count) *
 					sizeof(xfs_attr_leaf_entry_t)) -
@@ -1805,7 +1796,6 @@ xfs_attr3_leaf_remove(
 	struct xfs_attr_leafblock *leaf;
 	struct xfs_attr3_icleaf_hdr ichdr;
 	struct xfs_attr_leaf_entry *entry;
-	struct xfs_mount	*mp = args->trans->t_mountp;
 	int			before;
 	int			after;
 	int			smallest;
@@ -1819,7 +1809,7 @@ xfs_attr3_leaf_remove(
 	leaf = bp->b_addr;
 	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
 
-	ASSERT(ichdr.count > 0 && ichdr.count < XFS_LBSIZE(mp) / 8);
+	ASSERT(ichdr.count > 0 && ichdr.count < args->geo->blksize / 8);
 	ASSERT(args->index >= 0 && args->index < ichdr.count);
 	ASSERT(ichdr.firstused >= ichdr.count * sizeof(*entry) +
 					xfs_attr3_leaf_hdr_size(leaf));
@@ -1827,7 +1817,7 @@ xfs_attr3_leaf_remove(
 	entry = &xfs_attr3_leaf_entryp(leaf)[args->index];
 
 	ASSERT(be16_to_cpu(entry->nameidx) >= ichdr.firstused);
-	ASSERT(be16_to_cpu(entry->nameidx) < XFS_LBSIZE(mp));
+	ASSERT(be16_to_cpu(entry->nameidx) < args->geo->blksize);
 
 	/*
 	 * Scan through free region table:
@@ -1842,8 +1832,8 @@ xfs_attr3_leaf_remove(
 	smallest = XFS_ATTR_LEAF_MAPSIZE - 1;
 	entsize = xfs_attr_leaf_entsize(leaf, args->index);
 	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; i++) {
-		ASSERT(ichdr.freemap[i].base < XFS_LBSIZE(mp));
-		ASSERT(ichdr.freemap[i].size < XFS_LBSIZE(mp));
+		ASSERT(ichdr.freemap[i].base < args->geo->blksize);
+		ASSERT(ichdr.freemap[i].size < args->geo->blksize);
 		if (ichdr.freemap[i].base == tablesize) {
 			ichdr.freemap[i].base -= sizeof(xfs_attr_leaf_entry_t);
 			ichdr.freemap[i].size += sizeof(xfs_attr_leaf_entry_t);
@@ -1920,11 +1910,11 @@ xfs_attr3_leaf_remove(
 	 * removing the name.
 	 */
 	if (smallest) {
-		tmp = XFS_LBSIZE(mp);
+		tmp = args->geo->blksize;
 		entry = xfs_attr3_leaf_entryp(leaf);
 		for (i = ichdr.count - 1; i >= 0; entry++, i--) {
 			ASSERT(be16_to_cpu(entry->nameidx) >= ichdr.firstused);
-			ASSERT(be16_to_cpu(entry->nameidx) < XFS_LBSIZE(mp));
+			ASSERT(be16_to_cpu(entry->nameidx) < args->geo->blksize);
 
 			if (be16_to_cpu(entry->nameidx) < tmp)
 				tmp = be16_to_cpu(entry->nameidx);
@@ -1947,7 +1937,7 @@ xfs_attr3_leaf_remove(
 	tmp = ichdr.usedbytes + xfs_attr3_leaf_hdr_size(leaf) +
 	      ichdr.count * sizeof(xfs_attr_leaf_entry_t);
 
-	return tmp < mp->m_attr_magicpct; /* leaf is < 37% full */
+	return tmp < args->geo->magicpct; /* leaf is < 37% full */
 }
 
 /*
@@ -1964,7 +1954,6 @@ xfs_attr3_leaf_unbalance(
 	struct xfs_attr3_icleaf_hdr drophdr;
 	struct xfs_attr3_icleaf_hdr savehdr;
 	struct xfs_attr_leaf_entry *entry;
-	struct xfs_mount	*mp = state->mp;
 
 	trace_xfs_attr_leaf_unbalance(state->args);
 
@@ -1991,13 +1980,15 @@ xfs_attr3_leaf_unbalance(
 		 */
 		if (xfs_attr3_leaf_order(save_blk->bp, &savehdr,
 					 drop_blk->bp, &drophdr)) {
-			xfs_attr3_leaf_moveents(drop_leaf, &drophdr, 0,
+			xfs_attr3_leaf_moveents(state->args,
+						drop_leaf, &drophdr, 0,
 						save_leaf, &savehdr, 0,
-						drophdr.count, mp);
+						drophdr.count);
 		} else {
-			xfs_attr3_leaf_moveents(drop_leaf, &drophdr, 0,
+			xfs_attr3_leaf_moveents(state->args,
+						drop_leaf, &drophdr, 0,
 						save_leaf, &savehdr,
-						savehdr.count, drophdr.count, mp);
+						savehdr.count, drophdr.count);
 		}
 	} else {
 		/*
@@ -2007,7 +1998,7 @@ xfs_attr3_leaf_unbalance(
 		struct xfs_attr_leafblock *tmp_leaf;
 		struct xfs_attr3_icleaf_hdr tmphdr;
 
-		tmp_leaf = kmem_zalloc(state->blocksize, KM_SLEEP);
+		tmp_leaf = kmem_zalloc(state->args->geo->blksize, KM_SLEEP);
 
 		/*
 		 * Copy the header into the temp leaf so that all the stuff
@@ -2020,35 +2011,39 @@ xfs_attr3_leaf_unbalance(
 		tmphdr.magic = savehdr.magic;
 		tmphdr.forw = savehdr.forw;
 		tmphdr.back = savehdr.back;
-		tmphdr.firstused = state->blocksize;
+		tmphdr.firstused = state->args->geo->blksize;
 
 		/* write the header to the temp buffer to initialise it */
 		xfs_attr3_leaf_hdr_to_disk(tmp_leaf, &tmphdr);
 
 		if (xfs_attr3_leaf_order(save_blk->bp, &savehdr,
 					 drop_blk->bp, &drophdr)) {
-			xfs_attr3_leaf_moveents(drop_leaf, &drophdr, 0,
+			xfs_attr3_leaf_moveents(state->args,
+						drop_leaf, &drophdr, 0,
 						tmp_leaf, &tmphdr, 0,
-						drophdr.count, mp);
-			xfs_attr3_leaf_moveents(save_leaf, &savehdr, 0,
+						drophdr.count);
+			xfs_attr3_leaf_moveents(state->args,
+						save_leaf, &savehdr, 0,
 						tmp_leaf, &tmphdr, tmphdr.count,
-						savehdr.count, mp);
+						savehdr.count);
 		} else {
-			xfs_attr3_leaf_moveents(save_leaf, &savehdr, 0,
+			xfs_attr3_leaf_moveents(state->args,
+						save_leaf, &savehdr, 0,
 						tmp_leaf, &tmphdr, 0,
-						savehdr.count, mp);
-			xfs_attr3_leaf_moveents(drop_leaf, &drophdr, 0,
+						savehdr.count);
+			xfs_attr3_leaf_moveents(state->args,
+						drop_leaf, &drophdr, 0,
 						tmp_leaf, &tmphdr, tmphdr.count,
-						drophdr.count, mp);
+						drophdr.count);
 		}
-		memcpy(save_leaf, tmp_leaf, state->blocksize);
+		memcpy(save_leaf, tmp_leaf, state->args->geo->blksize);
 		savehdr = tmphdr; /* struct copy */
 		kmem_free(tmp_leaf);
 	}
 
 	xfs_attr3_leaf_hdr_to_disk(save_leaf, &savehdr);
 	xfs_trans_log_buf(state->args->trans, save_blk->bp, 0,
-					   state->blocksize - 1);
+					   state->args->geo->blksize - 1);
 
 	/*
 	 * Copy out last hashval in each block for B-tree code.
@@ -2094,7 +2089,7 @@ xfs_attr3_leaf_lookup_int(
 	leaf = bp->b_addr;
 	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
 	entries = xfs_attr3_leaf_entryp(leaf);
-	ASSERT(ichdr.count < XFS_LBSIZE(args->dp->i_mount) / 8);
+	ASSERT(ichdr.count < args->geo->blksize / 8);
 
 	/*
 	 * Binary search.  (note: small blocks will skip this loop)
@@ -2198,7 +2193,7 @@ xfs_attr3_leaf_getvalue(
 
 	leaf = bp->b_addr;
 	xfs_attr3_leaf_hdr_from_disk(&ichdr, leaf);
-	ASSERT(ichdr.count < XFS_LBSIZE(args->dp->i_mount) / 8);
+	ASSERT(ichdr.count < args->geo->blksize / 8);
 	ASSERT(args->index < ichdr.count);
 
 	entry = &xfs_attr3_leaf_entryp(leaf)[args->index];
@@ -2249,14 +2244,14 @@ xfs_attr3_leaf_getvalue(
 /*ARGSUSED*/
 STATIC void
 xfs_attr3_leaf_moveents(
+	struct xfs_da_args		*args,
 	struct xfs_attr_leafblock	*leaf_s,
 	struct xfs_attr3_icleaf_hdr	*ichdr_s,
 	int				start_s,
 	struct xfs_attr_leafblock	*leaf_d,
 	struct xfs_attr3_icleaf_hdr	*ichdr_d,
 	int				start_d,
-	int				count,
-	struct xfs_mount		*mp)
+	int				count)
 {
 	struct xfs_attr_leaf_entry	*entry_s;
 	struct xfs_attr_leaf_entry	*entry_d;
@@ -2276,10 +2271,10 @@ xfs_attr3_leaf_moveents(
 	ASSERT(ichdr_s->magic == XFS_ATTR_LEAF_MAGIC ||
 	       ichdr_s->magic == XFS_ATTR3_LEAF_MAGIC);
 	ASSERT(ichdr_s->magic == ichdr_d->magic);
-	ASSERT(ichdr_s->count > 0 && ichdr_s->count < XFS_LBSIZE(mp) / 8);
+	ASSERT(ichdr_s->count > 0 && ichdr_s->count < args->geo->blksize / 8);
 	ASSERT(ichdr_s->firstused >= (ichdr_s->count * sizeof(*entry_s))
 					+ xfs_attr3_leaf_hdr_size(leaf_s));
-	ASSERT(ichdr_d->count < XFS_LBSIZE(mp) / 8);
+	ASSERT(ichdr_d->count < args->geo->blksize / 8);
 	ASSERT(ichdr_d->firstused >= (ichdr_d->count * sizeof(*entry_d))
 					+ xfs_attr3_leaf_hdr_size(leaf_d));
 
@@ -2331,11 +2326,11 @@ xfs_attr3_leaf_moveents(
 			entry_d->nameidx = cpu_to_be16(ichdr_d->firstused);
 			entry_d->flags = entry_s->flags;
 			ASSERT(be16_to_cpu(entry_d->nameidx) + tmp
-							<= XFS_LBSIZE(mp));
+							<= args->geo->blksize);
 			memmove(xfs_attr3_leaf_name(leaf_d, desti),
 				xfs_attr3_leaf_name(leaf_s, start_s + i), tmp);
 			ASSERT(be16_to_cpu(entry_s->nameidx) + tmp
-							<= XFS_LBSIZE(mp));
+							<= args->geo->blksize);
 			memset(xfs_attr3_leaf_name(leaf_s, start_s + i), 0, tmp);
 			ichdr_s->usedbytes -= tmp;
 			ichdr_d->usedbytes += tmp;
@@ -2356,7 +2351,7 @@ xfs_attr3_leaf_moveents(
 		tmp = count * sizeof(xfs_attr_leaf_entry_t);
 		entry_s = &xfs_attr3_leaf_entryp(leaf_s)[start_s];
 		ASSERT(((char *)entry_s + tmp) <=
-		       ((char *)leaf_s + XFS_LBSIZE(mp)));
+		       ((char *)leaf_s + args->geo->blksize));
 		memset(entry_s, 0, tmp);
 	} else {
 		/*
@@ -2371,7 +2366,7 @@ xfs_attr3_leaf_moveents(
 		tmp = count * sizeof(xfs_attr_leaf_entry_t);
 		entry_s = &xfs_attr3_leaf_entryp(leaf_s)[ichdr_s->count];
 		ASSERT(((char *)entry_s + tmp) <=
-		       ((char *)leaf_s + XFS_LBSIZE(mp)));
+		       ((char *)leaf_s + args->geo->blksize));
 		memset(entry_s, 0, tmp);
 	}
 
@@ -2439,22 +2434,21 @@ xfs_attr_leaf_entsize(xfs_attr_leafblock_t *leaf, int index)
  * a "local" or a "remote" attribute.
  */
 int
-xfs_attr_leaf_newentsize(int namelen, int valuelen, int blocksize, int *local)
+xfs_attr_leaf_newentsize(
+	struct xfs_da_args	*args,
+	int			*local)
 {
-	int size;
+	int			size;
 
-	size = xfs_attr_leaf_entsize_local(namelen, valuelen);
-	if (size < xfs_attr_leaf_entsize_local_max(blocksize)) {
-		if (local) {
+	size = xfs_attr_leaf_entsize_local(args->namelen, args->valuelen);
+	if (size < xfs_attr_leaf_entsize_local_max(args->geo->blksize)) {
+		if (local)
 			*local = 1;
-		}
-	} else {
-		size = xfs_attr_leaf_entsize_remote(namelen);
-		if (local) {
-			*local = 0;
-		}
+		return size;
 	}
-	return size;
+	if (local)
+		*local = 0;
+	return xfs_attr_leaf_entsize_remote(args->namelen);
 }
 
 

fs/xfs/xfs_attr_leaf.h

@@ -96,8 +96,7 @@ int	xfs_attr3_root_inactive(struct xfs_trans **trans, struct xfs_inode *dp);
 xfs_dahash_t	xfs_attr_leaf_lasthash(struct xfs_buf *bp, int *count);
 int	xfs_attr_leaf_order(struct xfs_buf *leaf1_bp,
 				   struct xfs_buf *leaf2_bp);
-int	xfs_attr_leaf_newentsize(int namelen, int valuelen, int blocksize,
-					int *local);
+int	xfs_attr_leaf_newentsize(struct xfs_da_args *args, int *local);
 int	xfs_attr3_leaf_read(struct xfs_trans *tp, struct xfs_inode *dp,
 			xfs_dablk_t bno, xfs_daddr_t mappedbno,
 			struct xfs_buf **bpp);

fs/xfs/xfs_attr_list.c

@@ -444,6 +444,7 @@ xfs_attr3_leaf_list_int(
 				xfs_da_args_t args;
 
 				memset((char *)&args, 0, sizeof(args));
+				args.geo = context->dp->i_mount->m_attr_geo;
 				args.dp = context->dp;
 				args.whichfork = XFS_ATTR_FORK;
 				args.valuelen = valuelen;

fs/xfs/xfs_attr_remote.c

@@ -68,7 +68,6 @@ xfs_attr3_rmt_blocks(
  */
 static bool
 xfs_attr3_rmt_hdr_ok(
-	struct xfs_mount	*mp,
 	void			*ptr,
 	xfs_ino_t		ino,
 	uint32_t		offset,
@@ -126,6 +125,7 @@ xfs_attr3_rmt_read_verify(
 	char		*ptr;
 	int		len;
 	xfs_daddr_t	bno;
+	int		blksize = mp->m_attr_geo->blksize;
 
 	/* no verification of non-crc buffers */
 	if (!xfs_sb_version_hascrc(&mp->m_sb))
@@ -134,21 +134,20 @@ xfs_attr3_rmt_read_verify(
 	ptr = bp->b_addr;
 	bno = bp->b_bn;
 	len = BBTOB(bp->b_length);
-	ASSERT(len >= XFS_LBSIZE(mp));
+	ASSERT(len >= blksize);
 
 	while (len > 0) {
-		if (!xfs_verify_cksum(ptr, XFS_LBSIZE(mp),
-				      XFS_ATTR3_RMT_CRC_OFF)) {
+		if (!xfs_verify_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF)) {
 			xfs_buf_ioerror(bp, EFSBADCRC);
 			break;
 		}
-		if (!xfs_attr3_rmt_verify(mp, ptr, XFS_LBSIZE(mp), bno)) {
+		if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
 			xfs_buf_ioerror(bp, EFSCORRUPTED);
 			break;
 		}
-		len -= XFS_LBSIZE(mp);
-		ptr += XFS_LBSIZE(mp);
-		bno += mp->m_bsize;
+		len -= blksize;
+		ptr += blksize;
+		bno += BTOBB(blksize);
 	}
 
 	if (bp->b_error)
@@ -166,6 +165,7 @@ xfs_attr3_rmt_write_verify(
 	char		*ptr;
 	int		len;
 	xfs_daddr_t	bno;
+	int		blksize = mp->m_attr_geo->blksize;
 
 	/* no verification of non-crc buffers */
 	if (!xfs_sb_version_hascrc(&mp->m_sb))
@@ -174,10 +174,10 @@ xfs_attr3_rmt_write_verify(
 	ptr = bp->b_addr;
 	bno = bp->b_bn;
 	len = BBTOB(bp->b_length);
-	ASSERT(len >= XFS_LBSIZE(mp));
+	ASSERT(len >= blksize);
 
 	while (len > 0) {
-		if (!xfs_attr3_rmt_verify(mp, ptr, XFS_LBSIZE(mp), bno)) {
+		if (!xfs_attr3_rmt_verify(mp, ptr, blksize, bno)) {
 			xfs_buf_ioerror(bp, EFSCORRUPTED);
 			xfs_verifier_error(bp);
 			return;
@@ -188,11 +188,11 @@ xfs_attr3_rmt_write_verify(
 			rmt = (struct xfs_attr3_rmt_hdr *)ptr;
 			rmt->rm_lsn = cpu_to_be64(bip->bli_item.li_lsn);
 		}
-		xfs_update_cksum(ptr, XFS_LBSIZE(mp), XFS_ATTR3_RMT_CRC_OFF);
+		xfs_update_cksum(ptr, blksize, XFS_ATTR3_RMT_CRC_OFF);
 
-		len -= XFS_LBSIZE(mp);
-		ptr += XFS_LBSIZE(mp);
-		bno += mp->m_bsize;
+		len -= blksize;
+		ptr += blksize;
+		bno += BTOBB(blksize);
 	}
 	ASSERT(len == 0);
 }
@@ -241,17 +241,18 @@ xfs_attr_rmtval_copyout(
 	char		*src = bp->b_addr;
 	xfs_daddr_t	bno = bp->b_bn;
 	int		len = BBTOB(bp->b_length);
+	int		blksize = mp->m_attr_geo->blksize;
 
-	ASSERT(len >= XFS_LBSIZE(mp));
+	ASSERT(len >= blksize);
 
 	while (len > 0 && *valuelen > 0) {
 		int hdr_size = 0;
-		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, XFS_LBSIZE(mp));
+		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
 
 		byte_cnt = min(*valuelen, byte_cnt);
 
 		if (xfs_sb_version_hascrc(&mp->m_sb)) {
-			if (!xfs_attr3_rmt_hdr_ok(mp, src, ino, *offset,
+			if (!xfs_attr3_rmt_hdr_ok(src, ino, *offset,
 						  byte_cnt, bno)) {
 				xfs_alert(mp,
 "remote attribute header mismatch bno/off/len/owner (0x%llx/0x%x/Ox%x/0x%llx)",
@@ -264,9 +265,9 @@ xfs_attr_rmtval_copyout(
 		memcpy(*dst, src + hdr_size, byte_cnt);
 
 		/* roll buffer forwards */
-		len -= XFS_LBSIZE(mp);
-		src += XFS_LBSIZE(mp);
-		bno += mp->m_bsize;
+		len -= blksize;
+		src += blksize;
+		bno += BTOBB(blksize);
 
 		/* roll attribute data forwards */
 		*valuelen -= byte_cnt;
@@ -288,12 +289,13 @@ xfs_attr_rmtval_copyin(
 	char		*dst = bp->b_addr;
 	xfs_daddr_t	bno = bp->b_bn;
 	int		len = BBTOB(bp->b_length);
+	int		blksize = mp->m_attr_geo->blksize;
 
-	ASSERT(len >= XFS_LBSIZE(mp));
+	ASSERT(len >= blksize);
 
 	while (len > 0 && *valuelen > 0) {
 		int hdr_size;
-		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, XFS_LBSIZE(mp));
+		int byte_cnt = XFS_ATTR3_RMT_BUF_SPACE(mp, blksize);
 
 		byte_cnt = min(*valuelen, byte_cnt);
 		hdr_size = xfs_attr3_rmt_hdr_set(mp, dst, ino, *offset,
@@ -305,17 +307,17 @@ xfs_attr_rmtval_copyin(
 		 * If this is the last block, zero the remainder of it.
 		 * Check that we are actually the last block, too.
 		 */
-		if (byte_cnt + hdr_size < XFS_LBSIZE(mp)) {
+		if (byte_cnt + hdr_size < blksize) {
 			ASSERT(*valuelen - byte_cnt == 0);
-			ASSERT(len == XFS_LBSIZE(mp));
+			ASSERT(len == blksize);
 			memset(dst + hdr_size + byte_cnt, 0,
-					XFS_LBSIZE(mp) - hdr_size - byte_cnt);
+					blksize - hdr_size - byte_cnt);
 		}
 
 		/* roll buffer forwards */
-		len -= XFS_LBSIZE(mp);
-		dst += XFS_LBSIZE(mp);
-		bno += mp->m_bsize;
+		len -= blksize;
+		dst += blksize;
+		bno += BTOBB(blksize);
 
 		/* roll attribute data forwards */
 		*valuelen -= byte_cnt;

fs/xfs/xfs_bit.h

@@ -66,8 +66,11 @@ static inline int xfs_lowbit64(__uint64_t v)
 		n = ffs(w);
 	} else {	/* upper bits */
 		w = (__uint32_t)(v >> 32);
-		if (w && (n = ffs(w)))
-		n += 32;
+		if (w) {
+			n = ffs(w);
+			if (n)
+				n += 32;
+		}
 	}
 	return n - 1;
 }

fs/xfs/xfs_bmap.c

@@ -94,7 +94,7 @@ xfs_bmap_compute_maxlevels(
 		maxleafents = MAXAEXTNUM;
 		sz = XFS_BMDR_SPACE_CALC(MINABTPTRS);
 	}
-	maxrootrecs = xfs_bmdr_maxrecs(mp, sz, 0);
+	maxrootrecs = xfs_bmdr_maxrecs(sz, 0);
 	minleafrecs = mp->m_bmap_dmnr[0];
 	minnoderecs = mp->m_bmap_dmnr[1];
 	maxblocks = (maxleafents + minleafrecs - 1) / minleafrecs;
@@ -233,7 +233,6 @@ xfs_default_attroffset(
  */
 STATIC void
 xfs_bmap_forkoff_reset(
-	xfs_mount_t	*mp,
 	xfs_inode_t	*ip,
 	int		whichfork)
 {
@@ -905,7 +904,7 @@ xfs_bmap_local_to_extents_empty(
 	ASSERT(ifp->if_bytes == 0);
 	ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) == 0);
 
-	xfs_bmap_forkoff_reset(ip->i_mount, ip, whichfork);
+	xfs_bmap_forkoff_reset(ip, whichfork);
 	ifp->if_flags &= ~XFS_IFINLINE;
 	ifp->if_flags |= XFS_IFEXTENTS;
 	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
@@ -1099,10 +1098,11 @@ xfs_bmap_add_attrfork_local(
 
 	if (S_ISDIR(ip->i_d.di_mode)) {
 		memset(&dargs, 0, sizeof(dargs));
+		dargs.geo = ip->i_mount->m_dir_geo;
 		dargs.dp = ip;
 		dargs.firstblock = firstblock;
 		dargs.flist = flist;
-		dargs.total = ip->i_mount->m_dirblkfsbs;
+		dargs.total = dargs.geo->fsbcount;
 		dargs.whichfork = XFS_DATA_FORK;
 		dargs.trans = tp;
 		return xfs_dir2_sf_to_block(&dargs);
@@ -1675,7 +1675,6 @@ xfs_bmap_isaeof(
  */
 int
 xfs_bmap_last_offset(
-	struct xfs_trans	*tp,
 	struct xfs_inode	*ip,
 	xfs_fileoff_t		*last_block,
 	int			whichfork)
@@ -3517,6 +3516,67 @@ xfs_bmap_adjacent(
 #undef ISVALID
 }
 
+static int
+xfs_bmap_longest_free_extent(
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		ag,
+	xfs_extlen_t		*blen,
+	int			*notinit)
+{
+	struct xfs_mount	*mp = tp->t_mountp;
+	struct xfs_perag	*pag;
+	xfs_extlen_t		longest;
+	int			error = 0;
+
+	pag = xfs_perag_get(mp, ag);
+	if (!pag->pagf_init) {
+		error = xfs_alloc_pagf_init(mp, tp, ag, XFS_ALLOC_FLAG_TRYLOCK);
+		if (error)
+			goto out;
+
+		if (!pag->pagf_init) {
+			*notinit = 1;
+			goto out;
+		}
+	}
+
+	longest = xfs_alloc_longest_free_extent(mp, pag);
+	if (*blen < longest)
+		*blen = longest;
+
+out:
+	xfs_perag_put(pag);
+	return error;
+}
+
+static void
+xfs_bmap_select_minlen(
+	struct xfs_bmalloca	*ap,
+	struct xfs_alloc_arg	*args,
+	xfs_extlen_t		*blen,
+	int			notinit)
+{
+	if (notinit || *blen < ap->minlen) {
+		/*
+		 * Since we did a BUF_TRYLOCK above, it is possible that
+		 * there is space for this request.
+		 */
+		args->minlen = ap->minlen;
+	} else if (*blen < args->maxlen) {
+		/*
+		 * If the best seen length is less than the request length,
+		 * use the best as the minimum.
+		 */
+		args->minlen = *blen;
+	} else {
+		/*
+		 * Otherwise we've seen an extent as big as maxlen, use that
+		 * as the minimum.
+		 */
+		args->minlen = args->maxlen;
+	}
+}
+
 STATIC int
 xfs_bmap_btalloc_nullfb(
 	struct xfs_bmalloca	*ap,
@@ -3524,111 +3584,74 @@ xfs_bmap_btalloc_nullfb(
 	xfs_extlen_t		*blen)
 {
 	struct xfs_mount	*mp = ap->ip->i_mount;
-	struct xfs_perag	*pag;
 	xfs_agnumber_t		ag, startag;
 	int			notinit = 0;
 	int			error;
 
-	if (ap->userdata && xfs_inode_is_filestream(ap->ip))
-		args->type = XFS_ALLOCTYPE_NEAR_BNO;
-	else
-		args->type = XFS_ALLOCTYPE_START_BNO;
+	args->type = XFS_ALLOCTYPE_START_BNO;
 	args->total = ap->total;
 
-	/*
-	 * Search for an allocation group with a single extent large enough
-	 * for the request.  If one isn't found, then adjust the minimum
-	 * allocation size to the largest space found.
-	 */
 	startag = ag = XFS_FSB_TO_AGNO(mp, args->fsbno);
 	if (startag == NULLAGNUMBER)
 		startag = ag = 0;
 
-	pag = xfs_perag_get(mp, ag);
 	while (*blen < args->maxlen) {
-		if (!pag->pagf_init) {
-			error = xfs_alloc_pagf_init(mp, args->tp, ag,
-						    XFS_ALLOC_FLAG_TRYLOCK);
-			if (error) {
-				xfs_perag_put(pag);
-				return error;
-			}
-		}
-
-		/*
-		 * See xfs_alloc_fix_freelist...
-		 */
-		if (pag->pagf_init) {
-			xfs_extlen_t	longest;
-			longest = xfs_alloc_longest_free_extent(mp, pag);
-			if (*blen < longest)
-				*blen = longest;
-		} else
-			notinit = 1;
-
-		if (xfs_inode_is_filestream(ap->ip)) {
-			if (*blen >= args->maxlen)
-				break;
-
-			if (ap->userdata) {
-				/*
-				 * If startag is an invalid AG, we've
-				 * come here once before and
-				 * xfs_filestream_new_ag picked the
-				 * best currently available.
-				 *
-				 * Don't continue looping, since we
-				 * could loop forever.
-				 */
-				if (startag == NULLAGNUMBER)
-					break;
-
-				error = xfs_filestream_new_ag(ap, &ag);
-				xfs_perag_put(pag);
-				if (error)
-					return error;
+		error = xfs_bmap_longest_free_extent(args->tp, ag, blen,
+						     &notinit);
+		if (error)
+			return error;
 
-				/* loop again to set 'blen'*/
-				startag = NULLAGNUMBER;
-				pag = xfs_perag_get(mp, ag);
-				continue;
-			}
-		}
 		if (++ag == mp->m_sb.sb_agcount)
 			ag = 0;
 		if (ag == startag)
 			break;
-		xfs_perag_put(pag);
-		pag = xfs_perag_get(mp, ag);
 	}
-	xfs_perag_put(pag);
 
-	/*
-	 * Since the above loop did a BUF_TRYLOCK, it is
-	 * possible that there is space for this request.
-	 */
-	if (notinit || *blen < ap->minlen)
-		args->minlen = ap->minlen;
-	/*
-	 * If the best seen length is less than the request
-	 * length, use the best as the minimum.
-	 */
-	else if (*blen < args->maxlen)
-		args->minlen = *blen;
-	/*
-	 * Otherwise we've seen an extent as big as maxlen,
-	 * use that as the minimum.
-	 */
-	else
-		args->minlen = args->maxlen;
+	xfs_bmap_select_minlen(ap, args, blen, notinit);
+	return 0;
+}
+
+STATIC int
+xfs_bmap_btalloc_filestreams(
+	struct xfs_bmalloca	*ap,
+	struct xfs_alloc_arg	*args,
+	xfs_extlen_t		*blen)
+{
+	struct xfs_mount	*mp = ap->ip->i_mount;
+	xfs_agnumber_t		ag;
+	int			notinit = 0;
+	int			error;
+
+	args->type = XFS_ALLOCTYPE_NEAR_BNO;
+	args->total = ap->total;
+
+	ag = XFS_FSB_TO_AGNO(mp, args->fsbno);
+	if (ag == NULLAGNUMBER)
+		ag = 0;
+
+	error = xfs_bmap_longest_free_extent(args->tp, ag, blen, &notinit);
+	if (error)
+		return error;
+
+	if (*blen < args->maxlen) {
+		error = xfs_filestream_new_ag(ap, &ag);
+		if (error)
+			return error;
+
+		error = xfs_bmap_longest_free_extent(args->tp, ag, blen,
+						     &notinit);
+		if (error)
+			return error;
+
+	}
+
+	xfs_bmap_select_minlen(ap, args, blen, notinit);
 
 	/*
-	 * set the failure fallback case to look in the selected
-	 * AG as the stream may have moved.
+	 * Set the failure fallback case to look in the selected AG as stream
+	 * may have moved.
 	 */
-	if (xfs_inode_is_filestream(ap->ip))
-		ap->blkno = args->fsbno = XFS_AGB_TO_FSB(mp, ag, 0);
-
+	ap->blkno = args->fsbno = XFS_AGB_TO_FSB(mp, ag, 0);
 	return 0;
 }
 
@@ -3708,7 +3731,15 @@ xfs_bmap_btalloc(
 	args.firstblock = *ap->firstblock;
 	blen = 0;
 	if (nullfb) {
-		error = xfs_bmap_btalloc_nullfb(ap, &args, &blen);
+		/*
+		 * Search for an allocation group with a single extent large
+		 * enough for the request.  If one isn't found, then adjust
+		 * the minimum allocation size to the largest space found.
+		 */
+		if (ap->userdata && xfs_inode_is_filestream(ap->ip))
+			error = xfs_bmap_btalloc_filestreams(ap, &args, &blen);
+		else
+			error = xfs_bmap_btalloc_nullfb(ap, &args, &blen);
 		if (error)
 			return error;
 	} else if (ap->flist->xbf_low) {

fs/xfs/xfs_bmap.h

@@ -156,8 +156,8 @@ int	xfs_bmap_first_unused(struct xfs_trans *tp, struct xfs_inode *ip,
 		xfs_extlen_t len, xfs_fileoff_t *unused, int whichfork);
 int	xfs_bmap_last_before(struct xfs_trans *tp, struct xfs_inode *ip,
 		xfs_fileoff_t *last_block, int whichfork);
-int	xfs_bmap_last_offset(struct xfs_trans *tp, struct xfs_inode *ip,
-		xfs_fileoff_t *unused, int whichfork);
+int	xfs_bmap_last_offset(struct xfs_inode *ip, xfs_fileoff_t *unused,
+		int whichfork);
 int	xfs_bmap_one_block(struct xfs_inode *ip, int whichfork);
 int	xfs_bmap_read_extents(struct xfs_trans *tp, struct xfs_inode *ip,
 		int whichfork);

fs/xfs/xfs_bmap_btree.c

@@ -84,7 +84,7 @@ xfs_bmdr_to_bmbt(
 	rblock->bb_level = dblock->bb_level;
 	ASSERT(be16_to_cpu(rblock->bb_level) > 0);
 	rblock->bb_numrecs = dblock->bb_numrecs;
-	dmxr = xfs_bmdr_maxrecs(mp, dblocklen, 0);
+	dmxr = xfs_bmdr_maxrecs(dblocklen, 0);
 	fkp = XFS_BMDR_KEY_ADDR(dblock, 1);
 	tkp = XFS_BMBT_KEY_ADDR(mp, rblock, 1);
 	fpp = XFS_BMDR_PTR_ADDR(dblock, 1, dmxr);
@@ -443,7 +443,7 @@ xfs_bmbt_to_bmdr(
 	ASSERT(rblock->bb_level != 0);
 	dblock->bb_level = rblock->bb_level;
 	dblock->bb_numrecs = rblock->bb_numrecs;
-	dmxr = xfs_bmdr_maxrecs(mp, dblocklen, 0);
+	dmxr = xfs_bmdr_maxrecs(dblocklen, 0);
 	fkp = XFS_BMBT_KEY_ADDR(mp, rblock, 1);
 	tkp = XFS_BMDR_KEY_ADDR(dblock, 1);
 	fpp = XFS_BMAP_BROOT_PTR_ADDR(mp, rblock, 1, rblocklen);
@@ -519,7 +519,6 @@ xfs_bmbt_alloc_block(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*start,
 	union xfs_btree_ptr	*new,
-	int			length,
 	int			*stat)
 {
 	xfs_alloc_arg_t		args;		/* block allocation args */
@@ -672,8 +671,7 @@ xfs_bmbt_get_dmaxrecs(
 {
 	if (level != cur->bc_nlevels - 1)
 		return cur->bc_mp->m_bmap_dmxr[level != 0];
-	return xfs_bmdr_maxrecs(cur->bc_mp, cur->bc_private.b.forksize,
-				level == 0);
+	return xfs_bmdr_maxrecs(cur->bc_private.b.forksize, level == 0);
 }
 
 STATIC void
@@ -914,7 +912,6 @@ xfs_bmbt_maxrecs(
  */
 int
 xfs_bmdr_maxrecs(
-	struct xfs_mount	*mp,
 	int			blocklen,
 	int			leaf)
 {

fs/xfs/xfs_bmap_btree.h

@@ -130,7 +130,7 @@ extern void xfs_bmbt_to_bmdr(struct xfs_mount *, struct xfs_btree_block *, int,
 			xfs_bmdr_block_t *, int);
 
 extern int xfs_bmbt_get_maxrecs(struct xfs_btree_cur *, int level);
-extern int xfs_bmdr_maxrecs(struct xfs_mount *, int blocklen, int leaf);
+extern int xfs_bmdr_maxrecs(int blocklen, int leaf);
 extern int xfs_bmbt_maxrecs(struct xfs_mount *, int blocklen, int leaf);
 
 extern int xfs_bmbt_change_owner(struct xfs_trans *tp, struct xfs_inode *ip,

fs/xfs/xfs_bmap_util.c

@@ -258,14 +258,23 @@ xfs_bmapi_allocate_worker(
 	struct xfs_bmalloca	*args = container_of(work,
 						struct xfs_bmalloca, work);
 	unsigned long		pflags;
+	unsigned long		new_pflags = PF_FSTRANS;
 
-	/* we are in a transaction context here */
-	current_set_flags_nested(&pflags, PF_FSTRANS);
+	/*
+	 * we are in a transaction context here, but may also be doing work
+	 * in kswapd context, and hence we may need to inherit that state
+	 * temporarily to ensure that we don't block waiting for memory reclaim
+	 * in any way.
+	 */
+	if (args->kswapd)
+		new_pflags |= PF_MEMALLOC | PF_SWAPWRITE | PF_KSWAPD;
+
+	current_set_flags_nested(&pflags, new_pflags);
 
 	args->result = __xfs_bmapi_allocate(args);
 	complete(args->done);
 
-	current_restore_flags_nested(&pflags, PF_FSTRANS);
+	current_restore_flags_nested(&pflags, new_pflags);
 }
 
 /*
@@ -284,6 +293,7 @@ xfs_bmapi_allocate(
 
 
 	args->done = &done;
+	args->kswapd = current_is_kswapd();
 	INIT_WORK_ONSTACK(&args->work, xfs_bmapi_allocate_worker);
 	queue_work(xfs_alloc_wq, &args->work);
 	wait_for_completion(&done);
@@ -1519,7 +1529,6 @@ xfs_collapse_file_space(
 
 	while (!error && !done) {
 		tp = xfs_trans_alloc(mp, XFS_TRANS_DIOSTRAT);
-		tp->t_flags |= XFS_TRANS_RESERVE;
 		/*
 		 * We would need to reserve permanent block for transaction.
 		 * This will come into picture when after shifting extent into
@@ -1529,7 +1538,6 @@ xfs_collapse_file_space(
 		error = xfs_trans_reserve(tp, &M_RES(mp)->tr_write,
 				XFS_DIOSTRAT_SPACE_RES(mp, 0), 0);
 		if (error) {
-			ASSERT(error == ENOSPC || XFS_FORCED_SHUTDOWN(mp));
 			xfs_trans_cancel(tp, 0);
 			break;
 		}

fs/xfs/xfs_bmap_util.h

@@ -50,12 +50,13 @@ struct xfs_bmalloca {
 	xfs_extlen_t		total;	/* total blocks needed for xaction */
 	xfs_extlen_t		minlen;	/* minimum allocation size (blocks) */
 	xfs_extlen_t		minleft; /* amount must be left after alloc */
-	char			eof;	/* set if allocating past last extent */
-	char			wasdel;	/* replacing a delayed allocation */
-	char			userdata;/* set if is user data */
-	char			aeof;	/* allocated space at eof */
-	char			conv;	/* overwriting unwritten extents */
-	char			stack_switch;
+	bool			eof;	/* set if allocating past last extent */
+	bool			wasdel;	/* replacing a delayed allocation */
+	bool			userdata;/* set if is user data */
+	bool			aeof;	/* allocated space at eof */
+	bool			conv;	/* overwriting unwritten extents */
+	bool			stack_switch;
+	bool			kswapd;	/* allocation in kswapd context */
 	int			flags;
 	struct completion	*done;
 	struct work_struct	work;

fs/xfs/xfs_btree.c

@@ -43,9 +43,10 @@ kmem_zone_t	*xfs_btree_cur_zone;
  * Btree magic numbers.
  */
 static const __uint32_t xfs_magics[2][XFS_BTNUM_MAX] = {
-	{ XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, XFS_BMAP_MAGIC, XFS_IBT_MAGIC },
+	{ XFS_ABTB_MAGIC, XFS_ABTC_MAGIC, XFS_BMAP_MAGIC, XFS_IBT_MAGIC,
+	  XFS_FIBT_MAGIC },
 	{ XFS_ABTB_CRC_MAGIC, XFS_ABTC_CRC_MAGIC,
-	  XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC }
+	  XFS_BMAP_CRC_MAGIC, XFS_IBT_CRC_MAGIC, XFS_FIBT_CRC_MAGIC }
 };
 #define xfs_btree_magic(cur) \
 	xfs_magics[!!((cur)->bc_flags & XFS_BTREE_CRC_BLOCKS)][cur->bc_btnum]
@@ -552,14 +553,11 @@ xfs_btree_get_bufl(
 	xfs_fsblock_t	fsbno,		/* file system block number */
 	uint		lock)		/* lock flags for get_buf */
 {
-	xfs_buf_t	*bp;		/* buffer pointer (return value) */
 	xfs_daddr_t		d;		/* real disk block address */
 
 	ASSERT(fsbno != NULLFSBLOCK);
 	d = XFS_FSB_TO_DADDR(mp, fsbno);
-	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, lock);
-	ASSERT(!xfs_buf_geterror(bp));
-	return bp;
+	return xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, lock);
 }
 
 /*
@@ -574,15 +572,12 @@ xfs_btree_get_bufs(
 	xfs_agblock_t	agbno,		/* allocation group block number */
 	uint		lock)		/* lock flags for get_buf */
 {
-	xfs_buf_t	*bp;		/* buffer pointer (return value) */
 	xfs_daddr_t		d;		/* real disk block address */
 
 	ASSERT(agno != NULLAGNUMBER);
 	ASSERT(agbno != NULLAGBLOCK);
 	d = XFS_AGB_TO_DADDR(mp, agno, agbno);
-	bp = xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, lock);
-	ASSERT(!xfs_buf_geterror(bp));
-	return bp;
+	return xfs_trans_get_buf(tp, mp->m_ddev_targp, d, mp->m_bsize, lock);
 }
 
 /*
@@ -722,7 +717,6 @@ xfs_btree_read_bufl(
 				   mp->m_bsize, lock, &bp, ops);
 	if (error)
 		return error;
-	ASSERT(!xfs_buf_geterror(bp));
 	if (bp)
 		xfs_buf_set_ref(bp, refval);
 	*bpp = bp;
@@ -1115,6 +1109,7 @@ xfs_btree_set_refs(
 		xfs_buf_set_ref(bp, XFS_ALLOC_BTREE_REF);
 		break;
 	case XFS_BTNUM_INO:
+	case XFS_BTNUM_FINO:
 		xfs_buf_set_ref(bp, XFS_INO_BTREE_REF);
 		break;
 	case XFS_BTNUM_BMAP:
@@ -1159,7 +1154,6 @@ STATIC int
 xfs_btree_read_buf_block(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*ptr,
-	int			level,
 	int			flags,
 	struct xfs_btree_block	**block,
 	struct xfs_buf		**bpp)
@@ -1178,7 +1172,6 @@ xfs_btree_read_buf_block(
 	if (error)
 		return error;
 
-	ASSERT(!xfs_buf_geterror(*bpp));
 	xfs_btree_set_refs(cur, *bpp);
 	*block = XFS_BUF_TO_BLOCK(*bpp);
 	return 0;
@@ -1517,8 +1510,8 @@ xfs_btree_increment(
 		union xfs_btree_ptr	*ptrp;
 
 		ptrp = xfs_btree_ptr_addr(cur, cur->bc_ptrs[lev], block);
-		error = xfs_btree_read_buf_block(cur, ptrp, --lev,
-							0, &block, &bp);
+		--lev;
+		error = xfs_btree_read_buf_block(cur, ptrp, 0, &block, &bp);
 		if (error)
 			goto error0;
 
@@ -1616,8 +1609,8 @@ xfs_btree_decrement(
 		union xfs_btree_ptr	*ptrp;
 
 		ptrp = xfs_btree_ptr_addr(cur, cur->bc_ptrs[lev], block);
-		error = xfs_btree_read_buf_block(cur, ptrp, --lev,
-							0, &block, &bp);
+		--lev;
+		error = xfs_btree_read_buf_block(cur, ptrp, 0, &block, &bp);
 		if (error)
 			goto error0;
 		xfs_btree_setbuf(cur, lev, bp);
@@ -1667,7 +1660,7 @@ xfs_btree_lookup_get_block(
 		return 0;
 	}
 
-	error = xfs_btree_read_buf_block(cur, pp, level, 0, blkp, &bp);
+	error = xfs_btree_read_buf_block(cur, pp, 0, blkp, &bp);
 	if (error)
 		return error;
 
@@ -2018,7 +2011,7 @@ xfs_btree_lshift(
 		goto out0;
 
 	/* Set up the left neighbor as "left". */
-	error = xfs_btree_read_buf_block(cur, &lptr, level, 0, &left, &lbp);
+	error = xfs_btree_read_buf_block(cur, &lptr, 0, &left, &lbp);
 	if (error)
 		goto error0;
 
@@ -2202,7 +2195,7 @@ xfs_btree_rshift(
 		goto out0;
 
 	/* Set up the right neighbor as "right". */
-	error = xfs_btree_read_buf_block(cur, &rptr, level, 0, &right, &rbp);
+	error = xfs_btree_read_buf_block(cur, &rptr, 0, &right, &rbp);
 	if (error)
 		goto error0;
 
@@ -2372,7 +2365,7 @@ xfs_btree_split(
 	xfs_btree_buf_to_ptr(cur, lbp, &lptr);
 
 	/* Allocate the new block. If we can't do it, we're toast. Give up. */
-	error = cur->bc_ops->alloc_block(cur, &lptr, &rptr, 1, stat);
+	error = cur->bc_ops->alloc_block(cur, &lptr, &rptr, stat);
 	if (error)
 		goto error0;
 	if (*stat == 0)
@@ -2470,7 +2463,7 @@ xfs_btree_split(
 	 * point back to right instead of to left.
 	 */
 	if (!xfs_btree_ptr_is_null(cur, &rrptr)) {
-		error = xfs_btree_read_buf_block(cur, &rrptr, level,
+		error = xfs_btree_read_buf_block(cur, &rrptr,
 							0, &rrblock, &rrbp);
 		if (error)
 			goto error0;
@@ -2545,7 +2538,7 @@ xfs_btree_new_iroot(
 	pp = xfs_btree_ptr_addr(cur, 1, block);
 
 	/* Allocate the new block. If we can't do it, we're toast. Give up. */
-	error = cur->bc_ops->alloc_block(cur, pp, &nptr, 1, stat);
+	error = cur->bc_ops->alloc_block(cur, pp, &nptr, stat);
 	if (error)
 		goto error0;
 	if (*stat == 0) {
@@ -2649,7 +2642,7 @@ xfs_btree_new_root(
 	cur->bc_ops->init_ptr_from_cur(cur, &rptr);
 
 	/* Allocate the new block. If we can't do it, we're toast. Give up. */
-	error = cur->bc_ops->alloc_block(cur, &rptr, &lptr, 1, stat);
+	error = cur->bc_ops->alloc_block(cur, &rptr, &lptr, stat);
 	if (error)
 		goto error0;
 	if (*stat == 0)
@@ -2684,8 +2677,7 @@ xfs_btree_new_root(
 		lbp = bp;
 		xfs_btree_buf_to_ptr(cur, lbp, &lptr);
 		left = block;
-		error = xfs_btree_read_buf_block(cur, &rptr,
-					cur->bc_nlevels - 1, 0, &right, &rbp);
+		error = xfs_btree_read_buf_block(cur, &rptr, 0, &right, &rbp);
 		if (error)
 			goto error0;
 		bp = rbp;
@@ -2696,8 +2688,7 @@ xfs_btree_new_root(
 		xfs_btree_buf_to_ptr(cur, rbp, &rptr);
 		right = block;
 		xfs_btree_get_sibling(cur, right, &lptr, XFS_BB_LEFTSIB);
-		error = xfs_btree_read_buf_block(cur, &lptr,
-					cur->bc_nlevels - 1, 0, &left, &lbp);
+		error = xfs_btree_read_buf_block(cur, &lptr, 0, &left, &lbp);
 		if (error)
 			goto error0;
 		bp = lbp;
@@ -3649,8 +3640,7 @@ xfs_btree_delrec(
 		rptr = cptr;
 		right = block;
 		rbp = bp;
-		error = xfs_btree_read_buf_block(cur, &lptr, level,
-							0, &left, &lbp);
+		error = xfs_btree_read_buf_block(cur, &lptr, 0, &left, &lbp);
 		if (error)
 			goto error0;
 
@@ -3667,8 +3657,7 @@ xfs_btree_delrec(
 		lptr = cptr;
 		left = block;
 		lbp = bp;
-		error = xfs_btree_read_buf_block(cur, &rptr, level,
-							0, &right, &rbp);
+		error = xfs_btree_read_buf_block(cur, &rptr, 0, &right, &rbp);
 		if (error)
 			goto error0;
 
@@ -3740,8 +3729,7 @@ xfs_btree_delrec(
 	/* If there is a right sibling, point it to the remaining block. */
 	xfs_btree_get_sibling(cur, left, &cptr, XFS_BB_RIGHTSIB);
 	if (!xfs_btree_ptr_is_null(cur, &cptr)) {
-		error = xfs_btree_read_buf_block(cur, &cptr, level,
-							0, &rrblock, &rrbp);
+		error = xfs_btree_read_buf_block(cur, &cptr, 0, &rrblock, &rrbp);
 		if (error)
 			goto error0;
 		xfs_btree_set_sibling(cur, rrblock, &lptr, XFS_BB_LEFTSIB);

fs/xfs/xfs_btree.h

@@ -62,6 +62,7 @@ union xfs_btree_rec {
 #define	XFS_BTNUM_CNT	((xfs_btnum_t)XFS_BTNUM_CNTi)
 #define	XFS_BTNUM_BMAP	((xfs_btnum_t)XFS_BTNUM_BMAPi)
 #define	XFS_BTNUM_INO	((xfs_btnum_t)XFS_BTNUM_INOi)
+#define	XFS_BTNUM_FINO	((xfs_btnum_t)XFS_BTNUM_FINOi)
 
 /*
  * For logging record fields.
@@ -92,6 +93,7 @@ do {    \
 	case XFS_BTNUM_CNT: __XFS_BTREE_STATS_INC(abtc, stat); break;	\
 	case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_INC(bmbt, stat); break;	\
 	case XFS_BTNUM_INO: __XFS_BTREE_STATS_INC(ibt, stat); break;	\
+	case XFS_BTNUM_FINO: __XFS_BTREE_STATS_INC(fibt, stat); break;	\
 	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break;	\
 	}       \
 } while (0)
@@ -105,6 +107,7 @@ do {    \
 	case XFS_BTNUM_CNT: __XFS_BTREE_STATS_ADD(abtc, stat, val); break; \
 	case XFS_BTNUM_BMAP: __XFS_BTREE_STATS_ADD(bmbt, stat, val); break; \
 	case XFS_BTNUM_INO: __XFS_BTREE_STATS_ADD(ibt, stat, val); break; \
+	case XFS_BTNUM_FINO: __XFS_BTREE_STATS_ADD(fibt, stat, val); break; \
 	case XFS_BTNUM_MAX: ASSERT(0); /* fucking gcc */ ; break;	\
 	}       \
 } while (0)
@@ -129,7 +132,7 @@ struct xfs_btree_ops {
 	int	(*alloc_block)(struct xfs_btree_cur *cur,
 			       union xfs_btree_ptr *start_bno,
 			       union xfs_btree_ptr *new_bno,
-			       int length, int *stat);
+			       int *stat);
 	int	(*free_block)(struct xfs_btree_cur *cur, struct xfs_buf *bp);
 
 	/* update last record information */

fs/xfs/xfs_buf.c

@@ -216,8 +216,7 @@ _xfs_buf_alloc(
 STATIC int
 _xfs_buf_get_pages(
 	xfs_buf_t		*bp,
-	int			page_count,
-	xfs_buf_flags_t		flags)
+	int			page_count)
 {
 	/* Make sure that we have a page list */
 	if (bp->b_pages == NULL) {
@@ -330,7 +329,7 @@ use_alloc_page:
 	end = (BBTOB(bp->b_maps[0].bm_bn + bp->b_length) + PAGE_SIZE - 1)
 								>> PAGE_SHIFT;
 	page_count = end - start;
-	error = _xfs_buf_get_pages(bp, page_count, flags);
+	error = _xfs_buf_get_pages(bp, page_count);
 	if (unlikely(error))
 		return error;
 
@@ -778,7 +777,7 @@ xfs_buf_associate_memory(
 	bp->b_pages = NULL;
 	bp->b_addr = mem;
 
-	rval = _xfs_buf_get_pages(bp, page_count, 0);
+	rval = _xfs_buf_get_pages(bp, page_count);
 	if (rval)
 		return rval;
 
@@ -811,7 +810,7 @@ xfs_buf_get_uncached(
 		goto fail;
 
 	page_count = PAGE_ALIGN(numblks << BBSHIFT) >> PAGE_SHIFT;
-	error = _xfs_buf_get_pages(bp, page_count, 0);
+	error = _xfs_buf_get_pages(bp, page_count);
 	if (error)
 		goto fail_free_buf;
 
@@ -1615,7 +1614,6 @@ xfs_free_buftarg(
 int
 xfs_setsize_buftarg(
 	xfs_buftarg_t		*btp,
-	unsigned int		blocksize,
 	unsigned int		sectorsize)
 {
 	/* Set up metadata sector size info */
@@ -1650,16 +1648,13 @@ xfs_setsize_buftarg_early(
 	xfs_buftarg_t		*btp,
 	struct block_device	*bdev)
 {
-	return xfs_setsize_buftarg(btp, PAGE_SIZE,
-				   bdev_logical_block_size(bdev));
+	return xfs_setsize_buftarg(btp, bdev_logical_block_size(bdev));
 }
 
 xfs_buftarg_t *
 xfs_alloc_buftarg(
 	struct xfs_mount	*mp,
-	struct block_device	*bdev,
-	int			external,
-	const char		*fsname)
+	struct block_device	*bdev)
 {
 	xfs_buftarg_t		*btp;
 

fs/xfs/xfs_buf.h

@@ -298,11 +298,6 @@ extern void xfs_buf_iomove(xfs_buf_t *, size_t, size_t, void *,
 
 extern int xfs_bioerror_relse(struct xfs_buf *);
 
-static inline int xfs_buf_geterror(xfs_buf_t *bp)
-{
-	return bp ? bp->b_error : ENOMEM;
-}
-
 /* Buffer Utility Routines */
 extern xfs_caddr_t xfs_buf_offset(xfs_buf_t *, size_t);
 
@@ -387,10 +382,10 @@ xfs_buf_update_cksum(struct xfs_buf *bp, unsigned long cksum_offset)
  *	Handling of buftargs.
  */
 extern xfs_buftarg_t *xfs_alloc_buftarg(struct xfs_mount *,
-			struct block_device *, int, const char *);
+			struct block_device *);
 extern void xfs_free_buftarg(struct xfs_mount *, struct xfs_buftarg *);
 extern void xfs_wait_buftarg(xfs_buftarg_t *);
-extern int xfs_setsize_buftarg(xfs_buftarg_t *, unsigned int, unsigned int);
+extern int xfs_setsize_buftarg(xfs_buftarg_t *, unsigned int);
 
 #define xfs_getsize_buftarg(buftarg)	block_size((buftarg)->bt_bdev)
 #define xfs_readonly_buftarg(buftarg)	bdev_read_only((buftarg)->bt_bdev)

fs/xfs/xfs_buf_item.c

@@ -812,7 +812,6 @@ xfs_buf_item_init(
  */
 static void
 xfs_buf_item_log_segment(
-	struct xfs_buf_log_item	*bip,
 	uint			first,
 	uint			last,
 	uint			*map)
@@ -920,7 +919,7 @@ xfs_buf_item_log(
 		if (end > last)
 			end = last;
 
-		xfs_buf_item_log_segment(bip, first, end,
+		xfs_buf_item_log_segment(first, end,
 					 &bip->bli_formats[i].blf_data_map[0]);
 
 		start += bp->b_maps[i].bm_len;
@@ -1053,7 +1052,7 @@ xfs_buf_iodone_callbacks(
 	static ulong		lasttime;
 	static xfs_buftarg_t	*lasttarg;
 
-	if (likely(!xfs_buf_geterror(bp)))
+	if (likely(!bp->b_error))
 		goto do_callbacks;
 
 	/*

fs/xfs/xfs_da_btree.c

@@ -167,8 +167,8 @@ xfs_da3_node_verify(
 	 * we don't know if the node is for and attribute or directory tree,
 	 * so only fail if the count is outside both bounds
 	 */
-	if (ichdr.count > mp->m_dir_node_ents &&
-	    ichdr.count > mp->m_attr_node_ents)
+	if (ichdr.count > mp->m_dir_geo->node_ents &&
+	    ichdr.count > mp->m_attr_geo->node_ents)
 		return false;
 
 	/* XXX: hash order check? */
@@ -598,7 +598,7 @@ xfs_da3_root_split(
 	 * Set up the new root node.
 	 */
 	error = xfs_da3_node_create(args,
-		(args->whichfork == XFS_DATA_FORK) ? mp->m_dirleafblk : 0,
+		(args->whichfork == XFS_DATA_FORK) ? args->geo->leafblk : 0,
 		level + 1, &bp, args->whichfork);
 	if (error)
 		return error;
@@ -616,10 +616,10 @@ xfs_da3_root_split(
 #ifdef DEBUG
 	if (oldroot->hdr.info.magic == cpu_to_be16(XFS_DIR2_LEAFN_MAGIC) ||
 	    oldroot->hdr.info.magic == cpu_to_be16(XFS_DIR3_LEAFN_MAGIC)) {
-		ASSERT(blk1->blkno >= mp->m_dirleafblk &&
-		       blk1->blkno < mp->m_dirfreeblk);
-		ASSERT(blk2->blkno >= mp->m_dirleafblk &&
-		       blk2->blkno < mp->m_dirfreeblk);
+		ASSERT(blk1->blkno >= args->geo->leafblk &&
+		       blk1->blkno < args->geo->freeblk);
+		ASSERT(blk2->blkno >= args->geo->leafblk &&
+		       blk2->blkno < args->geo->freeblk);
 	}
 #endif
 
@@ -663,7 +663,7 @@ xfs_da3_node_split(
 	/*
 	 * Do we have to split the node?
 	 */
-	if (nodehdr.count + newcount > state->node_ents) {
+	if (nodehdr.count + newcount > state->args->geo->node_ents) {
 		/*
 		 * Allocate a new node, add to the doubly linked chain of
 		 * nodes, then move some of our excess entries into it.
@@ -894,8 +894,8 @@ xfs_da3_node_add(
 	ASSERT(oldblk->index >= 0 && oldblk->index <= nodehdr.count);
 	ASSERT(newblk->blkno != 0);
 	if (state->args->whichfork == XFS_DATA_FORK)
-		ASSERT(newblk->blkno >= state->mp->m_dirleafblk &&
-		       newblk->blkno < state->mp->m_dirfreeblk);
+		ASSERT(newblk->blkno >= state->args->geo->leafblk &&
+		       newblk->blkno < state->args->geo->freeblk);
 
 	/*
 	 * We may need to make some room before we insert the new node.
@@ -1089,14 +1089,15 @@ xfs_da3_root_join(
 	 * that could occur. For dir3 blocks we also need to update the block
 	 * number in the buffer header.
 	 */
-	memcpy(root_blk->bp->b_addr, bp->b_addr, state->blocksize);
+	memcpy(root_blk->bp->b_addr, bp->b_addr, args->geo->blksize);
 	root_blk->bp->b_ops = bp->b_ops;
 	xfs_trans_buf_copy_type(root_blk->bp, bp);
 	if (oldroothdr.magic == XFS_DA3_NODE_MAGIC) {
 		struct xfs_da3_blkinfo *da3 = root_blk->bp->b_addr;
 		da3->blkno = cpu_to_be64(root_blk->bp->b_bn);
 	}
-	xfs_trans_log_buf(args->trans, root_blk->bp, 0, state->blocksize - 1);
+	xfs_trans_log_buf(args->trans, root_blk->bp, 0,
+			  args->geo->blksize - 1);
 	error = xfs_da_shrink_inode(args, child, bp);
 	return(error);
 }
@@ -1139,7 +1140,7 @@ xfs_da3_node_toosmall(
 	info = blk->bp->b_addr;
 	node = (xfs_da_intnode_t *)info;
 	dp->d_ops->node_hdr_from_disk(&nodehdr, node);
-	if (nodehdr.count > (state->node_ents >> 1)) {
+	if (nodehdr.count > (state->args->geo->node_ents >> 1)) {
 		*action = 0;	/* blk over 50%, don't try to join */
 		return(0);	/* blk over 50%, don't try to join */
 	}
@@ -1176,8 +1177,8 @@ xfs_da3_node_toosmall(
 	 * We prefer coalescing with the lower numbered sibling so as
 	 * to shrink a directory over time.
 	 */
-	count  = state->node_ents;
-	count -= state->node_ents >> 2;
+	count  = state->args->geo->node_ents;
+	count -= state->args->geo->node_ents >> 2;
 	count -= nodehdr.count;
 
 	/* start with smaller blk num */
@@ -1472,7 +1473,7 @@ xfs_da3_node_lookup_int(
 	 * Descend thru the B-tree searching each level for the right
 	 * node to use, until the right hashval is found.
 	 */
-	blkno = (args->whichfork == XFS_DATA_FORK)? state->mp->m_dirleafblk : 0;
+	blkno = (args->whichfork == XFS_DATA_FORK)? args->geo->leafblk : 0;
 	for (blk = &state->path.blk[0], state->path.active = 1;
 			 state->path.active <= XFS_DA_NODE_MAXDEPTH;
 			 blk++, state->path.active++) {
@@ -2090,20 +2091,12 @@ xfs_da_grow_inode(
 	xfs_dablk_t		*new_blkno)
 {
 	xfs_fileoff_t		bno;
-	int			count;
 	int			error;
 
 	trace_xfs_da_grow_inode(args);
 
-	if (args->whichfork == XFS_DATA_FORK) {
-		bno = args->dp->i_mount->m_dirleafblk;
-		count = args->dp->i_mount->m_dirblkfsbs;
-	} else {
-		bno = 0;
-		count = 1;
-	}
-
-	error = xfs_da_grow_inode_int(args, &bno, count);
+	bno = args->geo->leafblk;
+	error = xfs_da_grow_inode_int(args, &bno, args->geo->fsbcount);
 	if (!error)
 		*new_blkno = (xfs_dablk_t)bno;
 	return error;
@@ -2158,7 +2151,7 @@ xfs_da3_swap_lastblock(
 	w = args->whichfork;
 	ASSERT(w == XFS_DATA_FORK);
 	mp = dp->i_mount;
-	lastoff = mp->m_dirfreeblk;
+	lastoff = args->geo->freeblk;
 	error = xfs_bmap_last_before(tp, dp, &lastoff, w);
 	if (error)
 		return error;
@@ -2170,15 +2163,15 @@ xfs_da3_swap_lastblock(
 	/*
 	 * Read the last block in the btree space.
 	 */
-	last_blkno = (xfs_dablk_t)lastoff - mp->m_dirblkfsbs;
+	last_blkno = (xfs_dablk_t)lastoff - args->geo->fsbcount;
 	error = xfs_da3_node_read(tp, dp, last_blkno, -1, &last_buf, w);
 	if (error)
 		return error;
 	/*
 	 * Copy the last block into the dead buffer and log it.
 	 */
-	memcpy(dead_buf->b_addr, last_buf->b_addr, mp->m_dirblksize);
-	xfs_trans_log_buf(tp, dead_buf, 0, mp->m_dirblksize - 1);
+	memcpy(dead_buf->b_addr, last_buf->b_addr, args->geo->blksize);
+	xfs_trans_log_buf(tp, dead_buf, 0, args->geo->blksize - 1);
 	dead_info = dead_buf->b_addr;
 	/*
 	 * Get values from the moved block.
@@ -2247,7 +2240,7 @@ xfs_da3_swap_lastblock(
 					sizeof(sib_info->back)));
 		sib_buf = NULL;
 	}
-	par_blkno = mp->m_dirleafblk;
+	par_blkno = args->geo->leafblk;
 	level = -1;
 	/*
 	 * Walk down the tree looking for the parent of the moved block.
@@ -2357,10 +2350,7 @@ xfs_da_shrink_inode(
 	w = args->whichfork;
 	tp = args->trans;
 	mp = dp->i_mount;
-	if (w == XFS_DATA_FORK)
-		count = mp->m_dirblkfsbs;
-	else
-		count = 1;
+	count = args->geo->fsbcount;
 	for (;;) {
 		/*
 		 * Remove extents.  If we get ENOSPC for a dir we have to move
@@ -2462,7 +2452,6 @@ xfs_buf_map_from_irec(
  */
 static int
 xfs_dabuf_map(
-	struct xfs_trans	*trans,
 	struct xfs_inode	*dp,
 	xfs_dablk_t		bno,
 	xfs_daddr_t		mappedbno,
@@ -2480,7 +2469,10 @@ xfs_dabuf_map(
 	ASSERT(map && *map);
 	ASSERT(*nmaps == 1);
 
-	nfsb = (whichfork == XFS_DATA_FORK) ? mp->m_dirblkfsbs : 1;
+	if (whichfork == XFS_DATA_FORK)
+		nfsb = mp->m_dir_geo->fsbcount;
+	else
+		nfsb = mp->m_attr_geo->fsbcount;
 
 	/*
 	 * Caller doesn't have a mapping.  -2 means don't complain
@@ -2558,7 +2550,7 @@ xfs_da_get_buf(
 	*bpp = NULL;
 	mapp = &map;
 	nmap = 1;
-	error = xfs_dabuf_map(trans, dp, bno, mappedbno, whichfork,
+	error = xfs_dabuf_map(dp, bno, mappedbno, whichfork,
 				&mapp, &nmap);
 	if (error) {
 		/* mapping a hole is not an error, but we don't continue */
@@ -2606,7 +2598,7 @@ xfs_da_read_buf(
 	*bpp = NULL;
 	mapp = &map;
 	nmap = 1;
-	error = xfs_dabuf_map(trans, dp, bno, mappedbno, whichfork,
+	error = xfs_dabuf_map(dp, bno, mappedbno, whichfork,
 				&mapp, &nmap);
 	if (error) {
 		/* mapping a hole is not an error, but we don't continue */
@@ -2625,47 +2617,6 @@ xfs_da_read_buf(
 		xfs_buf_set_ref(bp, XFS_ATTR_BTREE_REF);
 	else
 		xfs_buf_set_ref(bp, XFS_DIR_BTREE_REF);
-
-	/*
-	 * This verification code will be moved to a CRC verification callback
-	 * function so just leave it here unchanged until then.
-	 */
-	{
-		xfs_dir2_data_hdr_t	*hdr = bp->b_addr;
-		xfs_dir2_free_t		*free = bp->b_addr;
-		xfs_da_blkinfo_t	*info = bp->b_addr;
-		uint			magic, magic1;
-		struct xfs_mount	*mp = dp->i_mount;
-
-		magic = be16_to_cpu(info->magic);
-		magic1 = be32_to_cpu(hdr->magic);
-		if (unlikely(
-		    XFS_TEST_ERROR((magic != XFS_DA_NODE_MAGIC) &&
-				   (magic != XFS_DA3_NODE_MAGIC) &&
-				   (magic != XFS_ATTR_LEAF_MAGIC) &&
-				   (magic != XFS_ATTR3_LEAF_MAGIC) &&
-				   (magic != XFS_DIR2_LEAF1_MAGIC) &&
-				   (magic != XFS_DIR3_LEAF1_MAGIC) &&
-				   (magic != XFS_DIR2_LEAFN_MAGIC) &&
-				   (magic != XFS_DIR3_LEAFN_MAGIC) &&
-				   (magic1 != XFS_DIR2_BLOCK_MAGIC) &&
-				   (magic1 != XFS_DIR3_BLOCK_MAGIC) &&
-				   (magic1 != XFS_DIR2_DATA_MAGIC) &&
-				   (magic1 != XFS_DIR3_DATA_MAGIC) &&
-				   (free->hdr.magic !=
-					cpu_to_be32(XFS_DIR2_FREE_MAGIC)) &&
-				   (free->hdr.magic !=
-					cpu_to_be32(XFS_DIR3_FREE_MAGIC)),
-				mp, XFS_ERRTAG_DA_READ_BUF,
-				XFS_RANDOM_DA_READ_BUF))) {
-			trace_xfs_da_btree_corrupt(bp, _RET_IP_);
-			XFS_CORRUPTION_ERROR("xfs_da_do_buf(2)",
-					     XFS_ERRLEVEL_LOW, mp, info);
-			error = XFS_ERROR(EFSCORRUPTED);
-			xfs_trans_brelse(trans, bp);
-			goto out_free;
-		}
-	}
 	*bpp = bp;
 out_free:
 	if (mapp != &map)
@@ -2679,7 +2630,6 @@ out_free:
  */
 xfs_daddr_t
 xfs_da_reada_buf(
-	struct xfs_trans	*trans,
 	struct xfs_inode	*dp,
 	xfs_dablk_t		bno,
 	xfs_daddr_t		mappedbno,
@@ -2693,7 +2643,7 @@ xfs_da_reada_buf(
 
 	mapp = &map;
 	nmap = 1;
-	error = xfs_dabuf_map(trans, dp, bno, mappedbno, whichfork,
+	error = xfs_dabuf_map(dp, bno, mappedbno, whichfork,
 				&mapp, &nmap);
 	if (error) {
 		/* mapping a hole is not an error, but we don't continue */

fs/xfs/xfs_da_btree.h

@@ -25,6 +25,23 @@ struct xfs_trans;
 struct zone;
 struct xfs_dir_ops;
 
+/*
+ * Directory/attribute geometry information. There will be one of these for each
+ * data fork type, and it will be passed around via the xfs_da_args. Global
+ * structures will be attached to the xfs_mount.
+ */
+struct xfs_da_geometry {
+	int		blksize;	/* da block size in bytes */
+	int		fsbcount;	/* da block size in filesystem blocks */
+	uint8_t		fsblog;		/* log2 of _filesystem_ block size */
+	uint8_t		blklog;		/* log2 of da block size */
+	uint		node_ents;	/* # of entries in a danode */
+	int		magicpct;	/* 37% of block size in bytes */
+	xfs_dablk_t	datablk;	/* blockno of dir data v2 */
+	xfs_dablk_t	leafblk;	/* blockno of leaf data v2 */
+	xfs_dablk_t	freeblk;	/* blockno of free data v2 */
+};
+
 /*========================================================================
  * Btree searching and modification structure definitions.
  *========================================================================*/
@@ -42,6 +59,7 @@ enum xfs_dacmp {
  * Structure to ease passing around component names.
  */
 typedef struct xfs_da_args {
+	struct xfs_da_geometry *geo;	/* da block geometry */
 	const __uint8_t	*name;		/* string (maybe not NULL terminated) */
 	int		namelen;	/* length of string (maybe no NULL) */
 	__uint8_t	filetype;	/* filetype of inode for directories */
@@ -110,8 +128,6 @@ typedef struct xfs_da_state_path {
 typedef struct xfs_da_state {
 	xfs_da_args_t		*args;		/* filename arguments */
 	struct xfs_mount	*mp;		/* filesystem mount point */
-	unsigned int		blocksize;	/* logical block size */
-	unsigned int		node_ents;	/* how many entries in danode */
 	xfs_da_state_path_t	path;		/* search/split paths */
 	xfs_da_state_path_t	altpath;	/* alternate path for join */
 	unsigned char		inleaf;		/* insert into 1->lf, 0->splf */
@@ -185,9 +201,9 @@ int	xfs_da_read_buf(struct xfs_trans *trans, struct xfs_inode *dp,
 			       xfs_dablk_t bno, xfs_daddr_t mappedbno,
 			       struct xfs_buf **bpp, int whichfork,
 			       const struct xfs_buf_ops *ops);
-xfs_daddr_t	xfs_da_reada_buf(struct xfs_trans *trans, struct xfs_inode *dp,
-				xfs_dablk_t bno, xfs_daddr_t mapped_bno,
-				int whichfork, const struct xfs_buf_ops *ops);
+xfs_daddr_t	xfs_da_reada_buf(struct xfs_inode *dp, xfs_dablk_t bno,
+				xfs_daddr_t mapped_bno, int whichfork,
+				const struct xfs_buf_ops *ops);
 int	xfs_da_shrink_inode(xfs_da_args_t *args, xfs_dablk_t dead_blkno,
 					  struct xfs_buf *dead_buf);
 

fs/xfs/xfs_da_format.c

@@ -26,8 +26,10 @@
 #include "xfs_ag.h"
 #include "xfs_mount.h"
 #include "xfs_da_format.h"
+#include "xfs_da_btree.h"
 #include "xfs_inode.h"
 #include "xfs_dir2.h"
+#include "xfs_dir2_priv.h"
 
 /*
  * Shortform directory ops
@@ -425,9 +427,9 @@ xfs_dir3_data_unused_p(struct xfs_dir2_data_hdr *hdr)
  * Directory Leaf block operations
  */
 static int
-xfs_dir2_max_leaf_ents(struct xfs_mount *mp)
+xfs_dir2_max_leaf_ents(struct xfs_da_geometry *geo)
 {
-	return (mp->m_dirblksize - sizeof(struct xfs_dir2_leaf_hdr)) /
+	return (geo->blksize - sizeof(struct xfs_dir2_leaf_hdr)) /
 		(uint)sizeof(struct xfs_dir2_leaf_entry);
 }
 
@@ -438,9 +440,9 @@ xfs_dir2_leaf_ents_p(struct xfs_dir2_leaf *lp)
 }
 
 static int
-xfs_dir3_max_leaf_ents(struct xfs_mount *mp)
+xfs_dir3_max_leaf_ents(struct xfs_da_geometry *geo)
 {
-	return (mp->m_dirblksize - sizeof(struct xfs_dir3_leaf_hdr)) /
+	return (geo->blksize - sizeof(struct xfs_dir3_leaf_hdr)) /
 		(uint)sizeof(struct xfs_dir2_leaf_entry);
 }
 
@@ -591,9 +593,9 @@ xfs_da3_node_hdr_to_disk(
  * Directory free space block operations
  */
 static int
-xfs_dir2_free_max_bests(struct xfs_mount *mp)
+xfs_dir2_free_max_bests(struct xfs_da_geometry *geo)
 {
-	return (mp->m_dirblksize - sizeof(struct xfs_dir2_free_hdr)) /
+	return (geo->blksize - sizeof(struct xfs_dir2_free_hdr)) /
 		sizeof(xfs_dir2_data_off_t);
 }
 
@@ -607,24 +609,25 @@ xfs_dir2_free_bests_p(struct xfs_dir2_free *free)
  * Convert data space db to the corresponding free db.
  */
 static xfs_dir2_db_t
-xfs_dir2_db_to_fdb(struct xfs_mount *mp, xfs_dir2_db_t db)
+xfs_dir2_db_to_fdb(struct xfs_da_geometry *geo, xfs_dir2_db_t db)
 {
-	return XFS_DIR2_FREE_FIRSTDB(mp) + db / xfs_dir2_free_max_bests(mp);
+	return xfs_dir2_byte_to_db(geo, XFS_DIR2_FREE_OFFSET) +
+			(db / xfs_dir2_free_max_bests(geo));
 }
 
 /*
  * Convert data space db to the corresponding index in a free db.
  */
 static int
-xfs_dir2_db_to_fdindex(struct xfs_mount *mp, xfs_dir2_db_t db)
+xfs_dir2_db_to_fdindex(struct xfs_da_geometry *geo, xfs_dir2_db_t db)
 {
-	return db % xfs_dir2_free_max_bests(mp);
+	return db % xfs_dir2_free_max_bests(geo);
 }
 
 static int
-xfs_dir3_free_max_bests(struct xfs_mount *mp)
+xfs_dir3_free_max_bests(struct xfs_da_geometry *geo)
 {
-	return (mp->m_dirblksize - sizeof(struct xfs_dir3_free_hdr)) /
+	return (geo->blksize - sizeof(struct xfs_dir3_free_hdr)) /
 		sizeof(xfs_dir2_data_off_t);
 }
 
@@ -638,18 +641,19 @@ xfs_dir3_free_bests_p(struct xfs_dir2_free *free)
  * Convert data space db to the corresponding free db.
  */
 static xfs_dir2_db_t
-xfs_dir3_db_to_fdb(struct xfs_mount *mp, xfs_dir2_db_t db)
+xfs_dir3_db_to_fdb(struct xfs_da_geometry *geo, xfs_dir2_db_t db)
 {
-	return XFS_DIR2_FREE_FIRSTDB(mp) + db / xfs_dir3_free_max_bests(mp);
+	return xfs_dir2_byte_to_db(geo, XFS_DIR2_FREE_OFFSET) +
+			(db / xfs_dir3_free_max_bests(geo));
 }
 
 /*
  * Convert data space db to the corresponding index in a free db.
  */
 static int
-xfs_dir3_db_to_fdindex(struct xfs_mount *mp, xfs_dir2_db_t db)
+xfs_dir3_db_to_fdindex(struct xfs_da_geometry *geo, xfs_dir2_db_t db)
 {
-	return db % xfs_dir3_free_max_bests(mp);
+	return db % xfs_dir3_free_max_bests(geo);
 }
 
 static void

fs/xfs/xfs_da_format.h

@@ -19,10 +19,6 @@
 #ifndef __XFS_DA_FORMAT_H__
 #define __XFS_DA_FORMAT_H__
 
-/*========================================================================
- * Directory Structure when greater than XFS_LBSIZE(mp) bytes.
- *========================================================================*/
-
 /*
  * This structure is common to both leaf nodes and non-leaf nodes in the Btree.
  *
@@ -122,8 +118,6 @@ struct xfs_da3_icnode_hdr {
 	__uint16_t	level;
 };
 
-#define	XFS_LBSIZE(mp)	(mp)->m_sb.sb_blocksize
-
 /*
  * Directory version 2.
  *
@@ -330,8 +324,6 @@ xfs_dir2_sf_firstentry(struct xfs_dir2_sf_hdr *hdr)
 #define	XFS_DIR2_SPACE_SIZE	(1ULL << (32 + XFS_DIR2_DATA_ALIGN_LOG))
 #define	XFS_DIR2_DATA_SPACE	0
 #define	XFS_DIR2_DATA_OFFSET	(XFS_DIR2_DATA_SPACE * XFS_DIR2_SPACE_SIZE)
-#define	XFS_DIR2_DATA_FIRSTDB(mp)	\
-	xfs_dir2_byte_to_db(mp, XFS_DIR2_DATA_OFFSET)
 
 /*
  * Describe a free area in the data block.
@@ -456,8 +448,6 @@ xfs_dir2_data_unused_tag_p(struct xfs_dir2_data_unused *dup)
  */
 #define	XFS_DIR2_LEAF_SPACE	1
 #define	XFS_DIR2_LEAF_OFFSET	(XFS_DIR2_LEAF_SPACE * XFS_DIR2_SPACE_SIZE)
-#define	XFS_DIR2_LEAF_FIRSTDB(mp)	\
-	xfs_dir2_byte_to_db(mp, XFS_DIR2_LEAF_OFFSET)
 
 /*
  * Leaf block header.
@@ -513,17 +503,6 @@ struct xfs_dir3_leaf {
 
 #define XFS_DIR3_LEAF_CRC_OFF  offsetof(struct xfs_dir3_leaf_hdr, info.crc)
 
-/*
- * Get address of the bestcount field in the single-leaf block.
- */
-static inline struct xfs_dir2_leaf_tail *
-xfs_dir2_leaf_tail_p(struct xfs_mount *mp, struct xfs_dir2_leaf *lp)
-{
-	return (struct xfs_dir2_leaf_tail *)
-		((char *)lp + mp->m_dirblksize -
-		  sizeof(struct xfs_dir2_leaf_tail));
-}
-
 /*
  * Get address of the bests array in the single-leaf block.
  */
@@ -533,123 +512,6 @@ xfs_dir2_leaf_bests_p(struct xfs_dir2_leaf_tail *ltp)
 	return (__be16 *)ltp - be32_to_cpu(ltp->bestcount);
 }
 
-/*
- * DB blocks here are logical directory block numbers, not filesystem blocks.
- */
-
-/*
- * Convert dataptr to byte in file space
- */
-static inline xfs_dir2_off_t
-xfs_dir2_dataptr_to_byte(struct xfs_mount *mp, xfs_dir2_dataptr_t dp)
-{
-	return (xfs_dir2_off_t)dp << XFS_DIR2_DATA_ALIGN_LOG;
-}
-
-/*
- * Convert byte in file space to dataptr.  It had better be aligned.
- */
-static inline xfs_dir2_dataptr_t
-xfs_dir2_byte_to_dataptr(struct xfs_mount *mp, xfs_dir2_off_t by)
-{
-	return (xfs_dir2_dataptr_t)(by >> XFS_DIR2_DATA_ALIGN_LOG);
-}
-
-/*
- * Convert byte in space to (DB) block
- */
-static inline xfs_dir2_db_t
-xfs_dir2_byte_to_db(struct xfs_mount *mp, xfs_dir2_off_t by)
-{
-	return (xfs_dir2_db_t)
-		(by >> (mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog));
-}
-
-/*
- * Convert dataptr to a block number
- */
-static inline xfs_dir2_db_t
-xfs_dir2_dataptr_to_db(struct xfs_mount *mp, xfs_dir2_dataptr_t dp)
-{
-	return xfs_dir2_byte_to_db(mp, xfs_dir2_dataptr_to_byte(mp, dp));
-}
-
-/*
- * Convert byte in space to offset in a block
- */
-static inline xfs_dir2_data_aoff_t
-xfs_dir2_byte_to_off(struct xfs_mount *mp, xfs_dir2_off_t by)
-{
-	return (xfs_dir2_data_aoff_t)(by &
-		((1 << (mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog)) - 1));
-}
-
-/*
- * Convert dataptr to a byte offset in a block
- */
-static inline xfs_dir2_data_aoff_t
-xfs_dir2_dataptr_to_off(struct xfs_mount *mp, xfs_dir2_dataptr_t dp)
-{
-	return xfs_dir2_byte_to_off(mp, xfs_dir2_dataptr_to_byte(mp, dp));
-}
-
-/*
- * Convert block and offset to byte in space
- */
-static inline xfs_dir2_off_t
-xfs_dir2_db_off_to_byte(struct xfs_mount *mp, xfs_dir2_db_t db,
-			xfs_dir2_data_aoff_t o)
-{
-	return ((xfs_dir2_off_t)db <<
-		(mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog)) + o;
-}
-
-/*
- * Convert block (DB) to block (dablk)
- */
-static inline xfs_dablk_t
-xfs_dir2_db_to_da(struct xfs_mount *mp, xfs_dir2_db_t db)
-{
-	return (xfs_dablk_t)(db << mp->m_sb.sb_dirblklog);
-}
-
-/*
- * Convert byte in space to (DA) block
- */
-static inline xfs_dablk_t
-xfs_dir2_byte_to_da(struct xfs_mount *mp, xfs_dir2_off_t by)
-{
-	return xfs_dir2_db_to_da(mp, xfs_dir2_byte_to_db(mp, by));
-}
-
-/*
- * Convert block and offset to dataptr
- */
-static inline xfs_dir2_dataptr_t
-xfs_dir2_db_off_to_dataptr(struct xfs_mount *mp, xfs_dir2_db_t db,
-			   xfs_dir2_data_aoff_t o)
-{
-	return xfs_dir2_byte_to_dataptr(mp, xfs_dir2_db_off_to_byte(mp, db, o));
-}
-
-/*
- * Convert block (dablk) to block (DB)
- */
-static inline xfs_dir2_db_t
-xfs_dir2_da_to_db(struct xfs_mount *mp, xfs_dablk_t da)
-{
-	return (xfs_dir2_db_t)(da >> mp->m_sb.sb_dirblklog);
-}
-
-/*
- * Convert block (dablk) to byte offset in space
- */
-static inline xfs_dir2_off_t
-xfs_dir2_da_to_byte(struct xfs_mount *mp, xfs_dablk_t da)
-{
-	return xfs_dir2_db_off_to_byte(mp, xfs_dir2_da_to_db(mp, da), 0);
-}
-
 /*
  * Free space block defintions for the node format.
  */
@@ -659,8 +521,6 @@ xfs_dir2_da_to_byte(struct xfs_mount *mp, xfs_dablk_t da)
  */
 #define	XFS_DIR2_FREE_SPACE	2
 #define	XFS_DIR2_FREE_OFFSET	(XFS_DIR2_FREE_SPACE * XFS_DIR2_SPACE_SIZE)
-#define	XFS_DIR2_FREE_FIRSTDB(mp)	\
-	xfs_dir2_byte_to_db(mp, XFS_DIR2_FREE_OFFSET)
 
 typedef	struct xfs_dir2_free_hdr {
 	__be32			magic;		/* XFS_DIR2_FREE_MAGIC */
@@ -735,16 +595,6 @@ typedef struct xfs_dir2_block_tail {
 	__be32		stale;			/* count of stale lf entries */
 } xfs_dir2_block_tail_t;
 
-/*
- * Pointer to the leaf header embedded in a data block (1-block format)
- */
-static inline struct xfs_dir2_block_tail *
-xfs_dir2_block_tail_p(struct xfs_mount *mp, struct xfs_dir2_data_hdr *hdr)
-{
-	return ((struct xfs_dir2_block_tail *)
-		((char *)hdr + mp->m_dirblksize)) - 1;
-}
-
 /*
  * Pointer to the leaf entries embedded in a data block (1-block format)
  */
@@ -764,10 +614,6 @@ xfs_dir2_block_leaf_p(struct xfs_dir2_block_tail *btp)
  * of an attribute name may not be unique, we may have duplicate keys.  The
  * internal links in the Btree are logical block offsets into the file.
  *
- *========================================================================
- * Attribute structure when equal to XFS_LBSIZE(mp) bytes.
- *========================================================================
- *
  * Struct leaf_entry's are packed from the top.  Name/values grow from the
  * bottom but are not packed.  The freemap contains run-length-encoded entries
  * for the free bytes after the leaf_entry's, but only the N largest such,

fs/xfs/xfs_dir2.c

@@ -85,38 +85,74 @@ static struct xfs_nameops xfs_ascii_ci_nameops = {
 	.compname	= xfs_ascii_ci_compname,
 };
 
-void
-xfs_dir_mount(
-	xfs_mount_t	*mp)
+int
+xfs_da_mount(
+	struct xfs_mount	*mp)
 {
-	int	nodehdr_size;
+	struct xfs_da_geometry	*dageo;
+	int			nodehdr_size;
 
 
-	ASSERT(xfs_sb_version_hasdirv2(&mp->m_sb));
+	ASSERT(mp->m_sb.sb_versionnum & XFS_SB_VERSION_DIRV2BIT);
 	ASSERT((1 << (mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog)) <=
 	       XFS_MAX_BLOCKSIZE);
 
 	mp->m_dir_inode_ops = xfs_dir_get_ops(mp, NULL);
 	mp->m_nondir_inode_ops = xfs_nondir_get_ops(mp, NULL);
 
-	mp->m_dirblksize = 1 << (mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog);
-	mp->m_dirblkfsbs = 1 << mp->m_sb.sb_dirblklog;
-	mp->m_dirdatablk = xfs_dir2_db_to_da(mp, XFS_DIR2_DATA_FIRSTDB(mp));
-	mp->m_dirleafblk = xfs_dir2_db_to_da(mp, XFS_DIR2_LEAF_FIRSTDB(mp));
-	mp->m_dirfreeblk = xfs_dir2_db_to_da(mp, XFS_DIR2_FREE_FIRSTDB(mp));
-
 	nodehdr_size = mp->m_dir_inode_ops->node_hdr_size;
-	mp->m_attr_node_ents = (mp->m_sb.sb_blocksize - nodehdr_size) /
+	mp->m_dir_geo = kmem_zalloc(sizeof(struct xfs_da_geometry),
+				    KM_SLEEP | KM_MAYFAIL);
+	mp->m_attr_geo = kmem_zalloc(sizeof(struct xfs_da_geometry),
+				     KM_SLEEP | KM_MAYFAIL);
+	if (!mp->m_dir_geo || !mp->m_attr_geo) {
+		kmem_free(mp->m_dir_geo);
+		kmem_free(mp->m_attr_geo);
+		return ENOMEM;
+	}
+
+	/* set up directory geometry */
+	dageo = mp->m_dir_geo;
+	dageo->blklog = mp->m_sb.sb_blocklog + mp->m_sb.sb_dirblklog;
+	dageo->fsblog = mp->m_sb.sb_blocklog;
+	dageo->blksize = 1 << dageo->blklog;
+	dageo->fsbcount = 1 << mp->m_sb.sb_dirblklog;
+
+	/*
+	 * Now we've set up the block conversion variables, we can calculate the
+	 * segment block constants using the geometry structure.
+	 */
+	dageo->datablk = xfs_dir2_byte_to_da(dageo, XFS_DIR2_DATA_OFFSET);
+	dageo->leafblk = xfs_dir2_byte_to_da(dageo, XFS_DIR2_LEAF_OFFSET);
+	dageo->freeblk = xfs_dir2_byte_to_da(dageo, XFS_DIR2_FREE_OFFSET);
+	dageo->node_ents = (dageo->blksize - nodehdr_size) /
 				(uint)sizeof(xfs_da_node_entry_t);
-	mp->m_dir_node_ents = (mp->m_dirblksize - nodehdr_size) /
+	dageo->magicpct = (dageo->blksize * 37) / 100;
+
+	/* set up attribute geometry - single fsb only */
+	dageo = mp->m_attr_geo;
+	dageo->blklog = mp->m_sb.sb_blocklog;
+	dageo->fsblog = mp->m_sb.sb_blocklog;
+	dageo->blksize = 1 << dageo->blklog;
+	dageo->fsbcount = 1;
+	dageo->node_ents = (dageo->blksize - nodehdr_size) /
 				(uint)sizeof(xfs_da_node_entry_t);
+	dageo->magicpct = (dageo->blksize * 37) / 100;
 
-	mp->m_dir_magicpct = (mp->m_dirblksize * 37) / 100;
 	if (xfs_sb_version_hasasciici(&mp->m_sb))
 		mp->m_dirnameops = &xfs_ascii_ci_nameops;
 	else
 		mp->m_dirnameops = &xfs_default_nameops;
 
+	return 0;
+}
+
+void
+xfs_da_unmount(
+	struct xfs_mount	*mp)
+{
+	kmem_free(mp->m_dir_geo);
+	kmem_free(mp->m_attr_geo);
 }
 
 /*
@@ -192,6 +228,7 @@ xfs_dir_init(
 	if (!args)
 		return ENOMEM;
 
+	args->geo = dp->i_mount->m_dir_geo;
 	args->dp = dp;
 	args->trans = tp;
 	error = xfs_dir2_sf_create(args, pdp->i_ino);
@@ -226,6 +263,7 @@ xfs_dir_createname(
 	if (!args)
 		return ENOMEM;
 
+	args->geo = dp->i_mount->m_dir_geo;
 	args->name = name->name;
 	args->namelen = name->len;
 	args->filetype = name->type;
@@ -244,7 +282,7 @@ xfs_dir_createname(
 		goto out_free;
 	}
 
-	rval = xfs_dir2_isblock(tp, dp, &v);
+	rval = xfs_dir2_isblock(args, &v);
 	if (rval)
 		goto out_free;
 	if (v) {
@@ -252,7 +290,7 @@ xfs_dir_createname(
 		goto out_free;
 	}
 
-	rval = xfs_dir2_isleaf(tp, dp, &v);
+	rval = xfs_dir2_isleaf(args, &v);
 	if (rval)
 		goto out_free;
 	if (v)
@@ -320,6 +358,7 @@ xfs_dir_lookup(
 	 * annotations into the reclaim path for the ilock.
 	 */
 	args = kmem_zalloc(sizeof(*args), KM_SLEEP | KM_NOFS);
+	args->geo = dp->i_mount->m_dir_geo;
 	args->name = name->name;
 	args->namelen = name->len;
 	args->filetype = name->type;
@@ -336,7 +375,7 @@ xfs_dir_lookup(
 		goto out_check_rval;
 	}
 
-	rval = xfs_dir2_isblock(tp, dp, &v);
+	rval = xfs_dir2_isblock(args, &v);
 	if (rval)
 		goto out_free;
 	if (v) {
@@ -344,7 +383,7 @@ xfs_dir_lookup(
 		goto out_check_rval;
 	}
 
-	rval = xfs_dir2_isleaf(tp, dp, &v);
+	rval = xfs_dir2_isleaf(args, &v);
 	if (rval)
 		goto out_free;
 	if (v)
@@ -391,6 +430,7 @@ xfs_dir_removename(
 	if (!args)
 		return ENOMEM;
 
+	args->geo = dp->i_mount->m_dir_geo;
 	args->name = name->name;
 	args->namelen = name->len;
 	args->filetype = name->type;
@@ -408,7 +448,7 @@ xfs_dir_removename(
 		goto out_free;
 	}
 
-	rval = xfs_dir2_isblock(tp, dp, &v);
+	rval = xfs_dir2_isblock(args, &v);
 	if (rval)
 		goto out_free;
 	if (v) {
@@ -416,7 +456,7 @@ xfs_dir_removename(
 		goto out_free;
 	}
 
-	rval = xfs_dir2_isleaf(tp, dp, &v);
+	rval = xfs_dir2_isleaf(args, &v);
 	if (rval)
 		goto out_free;
 	if (v)
@@ -455,6 +495,7 @@ xfs_dir_replace(
 	if (!args)
 		return ENOMEM;
 
+	args->geo = dp->i_mount->m_dir_geo;
 	args->name = name->name;
 	args->namelen = name->len;
 	args->filetype = name->type;
@@ -472,7 +513,7 @@ xfs_dir_replace(
 		goto out_free;
 	}
 
-	rval = xfs_dir2_isblock(tp, dp, &v);
+	rval = xfs_dir2_isblock(args, &v);
 	if (rval)
 		goto out_free;
 	if (v) {
@@ -480,7 +521,7 @@ xfs_dir_replace(
 		goto out_free;
 	}
 
-	rval = xfs_dir2_isleaf(tp, dp, &v);
+	rval = xfs_dir2_isleaf(args, &v);
 	if (rval)
 		goto out_free;
 	if (v)
@@ -516,6 +557,7 @@ xfs_dir_canenter(
 	if (!args)
 		return ENOMEM;
 
+	args->geo = dp->i_mount->m_dir_geo;
 	args->name = name->name;
 	args->namelen = name->len;
 	args->filetype = name->type;
@@ -531,7 +573,7 @@ xfs_dir_canenter(
 		goto out_free;
 	}
 
-	rval = xfs_dir2_isblock(tp, dp, &v);
+	rval = xfs_dir2_isblock(args, &v);
 	if (rval)
 		goto out_free;
 	if (v) {
@@ -539,7 +581,7 @@ xfs_dir_canenter(
 		goto out_free;
 	}
 
-	rval = xfs_dir2_isleaf(tp, dp, &v);
+	rval = xfs_dir2_isleaf(args, &v);
 	if (rval)
 		goto out_free;
 	if (v)
@@ -579,13 +621,13 @@ xfs_dir2_grow_inode(
 	 * Set lowest possible block in the space requested.
 	 */
 	bno = XFS_B_TO_FSBT(mp, space * XFS_DIR2_SPACE_SIZE);
-	count = mp->m_dirblkfsbs;
+	count = args->geo->fsbcount;
 
 	error = xfs_da_grow_inode_int(args, &bno, count);
 	if (error)
 		return error;
 
-	*dbp = xfs_dir2_da_to_db(mp, (xfs_dablk_t)bno);
+	*dbp = xfs_dir2_da_to_db(args->geo, (xfs_dablk_t)bno);
 
 	/*
 	 * Update file's size if this is the data space and it grew.
@@ -607,19 +649,16 @@ xfs_dir2_grow_inode(
  */
 int
 xfs_dir2_isblock(
-	xfs_trans_t	*tp,
-	xfs_inode_t	*dp,
-	int		*vp)		/* out: 1 is block, 0 is not block */
+	struct xfs_da_args	*args,
+	int			*vp)	/* out: 1 is block, 0 is not block */
 {
-	xfs_fileoff_t	last;		/* last file offset */
-	xfs_mount_t	*mp;
-	int		rval;
+	xfs_fileoff_t		last;	/* last file offset */
+	int			rval;
 
-	mp = dp->i_mount;
-	if ((rval = xfs_bmap_last_offset(tp, dp, &last, XFS_DATA_FORK)))
+	if ((rval = xfs_bmap_last_offset(args->dp, &last, XFS_DATA_FORK)))
 		return rval;
-	rval = XFS_FSB_TO_B(mp, last) == mp->m_dirblksize;
-	ASSERT(rval == 0 || dp->i_d.di_size == mp->m_dirblksize);
+	rval = XFS_FSB_TO_B(args->dp->i_mount, last) == args->geo->blksize;
+	ASSERT(rval == 0 || args->dp->i_d.di_size == args->geo->blksize);
 	*vp = rval;
 	return 0;
 }
@@ -629,18 +668,15 @@ xfs_dir2_isblock(
  */
 int
 xfs_dir2_isleaf(
-	xfs_trans_t	*tp,
-	xfs_inode_t	*dp,
-	int		*vp)		/* out: 1 is leaf, 0 is not leaf */
+	struct xfs_da_args	*args,
+	int			*vp)	/* out: 1 is block, 0 is not block */
 {
-	xfs_fileoff_t	last;		/* last file offset */
-	xfs_mount_t	*mp;
-	int		rval;
+	xfs_fileoff_t		last;	/* last file offset */
+	int			rval;
 
-	mp = dp->i_mount;
-	if ((rval = xfs_bmap_last_offset(tp, dp, &last, XFS_DATA_FORK)))
+	if ((rval = xfs_bmap_last_offset(args->dp, &last, XFS_DATA_FORK)))
 		return rval;
-	*vp = last == mp->m_dirleafblk + (1 << mp->m_sb.sb_dirblklog);
+	*vp = last == args->geo->leafblk + args->geo->fsbcount;
 	return 0;
 }
 
@@ -668,11 +704,11 @@ xfs_dir2_shrink_inode(
 	dp = args->dp;
 	mp = dp->i_mount;
 	tp = args->trans;
-	da = xfs_dir2_db_to_da(mp, db);
+	da = xfs_dir2_db_to_da(args->geo, db);
 	/*
 	 * Unmap the fsblock(s).
 	 */
-	if ((error = xfs_bunmapi(tp, dp, da, mp->m_dirblkfsbs,
+	if ((error = xfs_bunmapi(tp, dp, da, args->geo->fsbcount,
 			XFS_BMAPI_METADATA, 0, args->firstblock, args->flist,
 			&done))) {
 		/*
@@ -699,12 +735,12 @@ xfs_dir2_shrink_inode(
 	/*
 	 * If it's not a data block, we're done.
 	 */
-	if (db >= XFS_DIR2_LEAF_FIRSTDB(mp))
+	if (db >= xfs_dir2_byte_to_db(args->geo, XFS_DIR2_LEAF_OFFSET))
 		return 0;
 	/*
 	 * If the block isn't the last one in the directory, we're done.
 	 */
-	if (dp->i_d.di_size > xfs_dir2_db_off_to_byte(mp, db + 1, 0))
+	if (dp->i_d.di_size > xfs_dir2_db_off_to_byte(args->geo, db + 1, 0))
 		return 0;
 	bno = da;
 	if ((error = xfs_bmap_last_before(tp, dp, &bno, XFS_DATA_FORK))) {
@@ -713,7 +749,7 @@ xfs_dir2_shrink_inode(
 		 */
 		return error;
 	}
-	if (db == mp->m_dirdatablk)
+	if (db == args->geo->datablk)
 		ASSERT(bno == 0);
 	else
 		ASSERT(bno > 0);

fs/xfs/xfs_dir2.h

@@ -80,7 +80,7 @@ struct xfs_dir_ops {
 				    struct xfs_dir3_icleaf_hdr *from);
 	void	(*leaf_hdr_from_disk)(struct xfs_dir3_icleaf_hdr *to,
 				      struct xfs_dir2_leaf *from);
-	int	(*leaf_max_ents)(struct xfs_mount *mp);
+	int	(*leaf_max_ents)(struct xfs_da_geometry *geo);
 	struct xfs_dir2_leaf_entry *
 		(*leaf_ents_p)(struct xfs_dir2_leaf *lp);
 
@@ -97,10 +97,12 @@ struct xfs_dir_ops {
 				    struct xfs_dir3_icfree_hdr *from);
 	void	(*free_hdr_from_disk)(struct xfs_dir3_icfree_hdr *to,
 				      struct xfs_dir2_free *from);
-	int	(*free_max_bests)(struct xfs_mount *mp);
+	int	(*free_max_bests)(struct xfs_da_geometry *geo);
 	__be16 * (*free_bests_p)(struct xfs_dir2_free *free);
-	xfs_dir2_db_t (*db_to_fdb)(struct xfs_mount *mp, xfs_dir2_db_t db);
-	int	(*db_to_fdindex)(struct xfs_mount *mp, xfs_dir2_db_t db);
+	xfs_dir2_db_t (*db_to_fdb)(struct xfs_da_geometry *geo,
+				   xfs_dir2_db_t db);
+	int	(*db_to_fdindex)(struct xfs_da_geometry *geo,
+				 xfs_dir2_db_t db);
 };
 
 extern const struct xfs_dir_ops *
@@ -112,7 +114,9 @@ extern const struct xfs_dir_ops *
  * Generic directory interface routines
  */
 extern void xfs_dir_startup(void);
-extern void xfs_dir_mount(struct xfs_mount *mp);
+extern int xfs_da_mount(struct xfs_mount *mp);
+extern void xfs_da_unmount(struct xfs_mount *mp);
+
 extern int xfs_dir_isempty(struct xfs_inode *dp);
 extern int xfs_dir_init(struct xfs_trans *tp, struct xfs_inode *dp,
 				struct xfs_inode *pdp);
@@ -142,23 +146,23 @@ extern int xfs_dir2_sf_to_block(struct xfs_da_args *args);
 /*
  * Interface routines used by userspace utilities
  */
-extern int xfs_dir2_isblock(struct xfs_trans *tp, struct xfs_inode *dp, int *r);
-extern int xfs_dir2_isleaf(struct xfs_trans *tp, struct xfs_inode *dp, int *r);
+extern int xfs_dir2_isblock(struct xfs_da_args *args, int *r);
+extern int xfs_dir2_isleaf(struct xfs_da_args *args, int *r);
 extern int xfs_dir2_shrink_inode(struct xfs_da_args *args, xfs_dir2_db_t db,
 				struct xfs_buf *bp);
 
 extern void xfs_dir2_data_freescan(struct xfs_inode *dp,
 		struct xfs_dir2_data_hdr *hdr, int *loghead);
-extern void xfs_dir2_data_log_entry(struct xfs_trans *tp, struct xfs_inode *dp,
+extern void xfs_dir2_data_log_entry(struct xfs_da_args *args,
 		struct xfs_buf *bp, struct xfs_dir2_data_entry *dep);
-extern void xfs_dir2_data_log_header(struct xfs_trans *tp, struct xfs_inode *dp,
+extern void xfs_dir2_data_log_header(struct xfs_da_args *args,
 		struct xfs_buf *bp);
-extern void xfs_dir2_data_log_unused(struct xfs_trans *tp, struct xfs_buf *bp,
-		struct xfs_dir2_data_unused *dup);
-extern void xfs_dir2_data_make_free(struct xfs_trans *tp, struct xfs_inode *dp,
+extern void xfs_dir2_data_log_unused(struct xfs_da_args *args,
+		struct xfs_buf *bp, struct xfs_dir2_data_unused *dup);
+extern void xfs_dir2_data_make_free(struct xfs_da_args *args,
 		struct xfs_buf *bp, xfs_dir2_data_aoff_t offset,
 		xfs_dir2_data_aoff_t len, int *needlogp, int *needscanp);
-extern void xfs_dir2_data_use_free(struct xfs_trans *tp, struct xfs_inode *dp,
+extern void xfs_dir2_data_use_free(struct xfs_da_args *args,
 		struct xfs_buf *bp, struct xfs_dir2_data_unused *dup,
 		xfs_dir2_data_aoff_t offset, xfs_dir2_data_aoff_t len,
 		int *needlogp, int *needscanp);

fs/xfs/xfs_dir2_block.c

@@ -136,7 +136,7 @@ xfs_dir3_block_read(
 	struct xfs_mount	*mp = dp->i_mount;
 	int			err;
 
-	err = xfs_da_read_buf(tp, dp, mp->m_dirdatablk, -1, bpp,
+	err = xfs_da_read_buf(tp, dp, mp->m_dir_geo->datablk, -1, bpp,
 				XFS_DATA_FORK, &xfs_dir3_block_buf_ops);
 	if (!err && tp)
 		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_BLOCK_BUF);
@@ -281,8 +281,7 @@ out:
  */
 static void
 xfs_dir2_block_compact(
-	struct xfs_trans		*tp,
-	struct xfs_inode		*dp,
+	struct xfs_da_args		*args,
 	struct xfs_buf			*bp,
 	struct xfs_dir2_data_hdr	*hdr,
 	struct xfs_dir2_block_tail	*btp,
@@ -315,18 +314,17 @@ xfs_dir2_block_compact(
 	*lfloglow = toidx + 1 - (be32_to_cpu(btp->stale) - 1);
 	*lfloghigh -= be32_to_cpu(btp->stale) - 1;
 	be32_add_cpu(&btp->count, -(be32_to_cpu(btp->stale) - 1));
-	xfs_dir2_data_make_free(tp, dp, bp,
+	xfs_dir2_data_make_free(args, bp,
 		(xfs_dir2_data_aoff_t)((char *)blp - (char *)hdr),
 		(xfs_dir2_data_aoff_t)((be32_to_cpu(btp->stale) - 1) * sizeof(*blp)),
 		needlog, &needscan);
-	blp += be32_to_cpu(btp->stale) - 1;
 	btp->stale = cpu_to_be32(1);
 	/*
 	 * If we now need to rebuild the bestfree map, do so.
 	 * This needs to happen before the next call to use_free.
 	 */
 	if (needscan)
-		xfs_dir2_data_freescan(dp, hdr, needlog);
+		xfs_dir2_data_freescan(args->dp, hdr, needlog);
 }
 
 /*
@@ -378,7 +376,7 @@ xfs_dir2_block_addname(
 	 * Set up pointers to parts of the block.
 	 */
 	hdr = bp->b_addr;
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(args->geo, hdr);
 	blp = xfs_dir2_block_leaf_p(btp);
 
 	/*
@@ -421,7 +419,7 @@ xfs_dir2_block_addname(
 	 * If need to compact the leaf entries, do it now.
 	 */
 	if (compact) {
-		xfs_dir2_block_compact(tp, dp, bp, hdr, btp, blp, &needlog,
+		xfs_dir2_block_compact(args, bp, hdr, btp, blp, &needlog,
 				      &lfloghigh, &lfloglow);
 		/* recalculate blp post-compaction */
 		blp = xfs_dir2_block_leaf_p(btp);
@@ -456,7 +454,7 @@ xfs_dir2_block_addname(
 		/*
 		 * Mark the space needed for the new leaf entry, now in use.
 		 */
-		xfs_dir2_data_use_free(tp, dp, bp, enddup,
+		xfs_dir2_data_use_free(args, bp, enddup,
 			(xfs_dir2_data_aoff_t)
 			((char *)enddup - (char *)hdr + be16_to_cpu(enddup->length) -
 			 sizeof(*blp)),
@@ -537,13 +535,13 @@ xfs_dir2_block_addname(
 	 * Fill in the leaf entry.
 	 */
 	blp[mid].hashval = cpu_to_be32(args->hashval);
-	blp[mid].address = cpu_to_be32(xfs_dir2_byte_to_dataptr(mp,
+	blp[mid].address = cpu_to_be32(xfs_dir2_byte_to_dataptr(
 				(char *)dep - (char *)hdr));
 	xfs_dir2_block_log_leaf(tp, bp, lfloglow, lfloghigh);
 	/*
 	 * Mark space for the data entry used.
 	 */
-	xfs_dir2_data_use_free(tp, dp, bp, dup,
+	xfs_dir2_data_use_free(args, bp, dup,
 		(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr),
 		(xfs_dir2_data_aoff_t)len, &needlog, &needscan);
 	/*
@@ -561,9 +559,9 @@ xfs_dir2_block_addname(
 	if (needscan)
 		xfs_dir2_data_freescan(dp, hdr, &needlog);
 	if (needlog)
-		xfs_dir2_data_log_header(tp, dp, bp);
+		xfs_dir2_data_log_header(args, bp);
 	xfs_dir2_block_log_tail(tp, bp);
-	xfs_dir2_data_log_entry(tp, dp, bp, dep);
+	xfs_dir2_data_log_entry(args, bp, dep);
 	xfs_dir3_data_check(dp, bp);
 	return 0;
 }
@@ -582,7 +580,7 @@ xfs_dir2_block_log_leaf(
 	xfs_dir2_leaf_entry_t	*blp;
 	xfs_dir2_block_tail_t	*btp;
 
-	btp = xfs_dir2_block_tail_p(tp->t_mountp, hdr);
+	btp = xfs_dir2_block_tail_p(tp->t_mountp->m_dir_geo, hdr);
 	blp = xfs_dir2_block_leaf_p(btp);
 	xfs_trans_log_buf(tp, bp, (uint)((char *)&blp[first] - (char *)hdr),
 		(uint)((char *)&blp[last + 1] - (char *)hdr - 1));
@@ -599,7 +597,7 @@ xfs_dir2_block_log_tail(
 	xfs_dir2_data_hdr_t	*hdr = bp->b_addr;
 	xfs_dir2_block_tail_t	*btp;
 
-	btp = xfs_dir2_block_tail_p(tp->t_mountp, hdr);
+	btp = xfs_dir2_block_tail_p(tp->t_mountp->m_dir_geo, hdr);
 	xfs_trans_log_buf(tp, bp, (uint)((char *)btp - (char *)hdr),
 		(uint)((char *)(btp + 1) - (char *)hdr - 1));
 }
@@ -634,13 +632,14 @@ xfs_dir2_block_lookup(
 	mp = dp->i_mount;
 	hdr = bp->b_addr;
 	xfs_dir3_data_check(dp, bp);
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(args->geo, hdr);
 	blp = xfs_dir2_block_leaf_p(btp);
 	/*
 	 * Get the offset from the leaf entry, to point to the data.
 	 */
 	dep = (xfs_dir2_data_entry_t *)((char *)hdr +
-		xfs_dir2_dataptr_to_off(mp, be32_to_cpu(blp[ent].address)));
+			xfs_dir2_dataptr_to_off(args->geo,
+						be32_to_cpu(blp[ent].address)));
 	/*
 	 * Fill in inode number, CI name if appropriate, release the block.
 	 */
@@ -686,7 +685,7 @@ xfs_dir2_block_lookup_int(
 
 	hdr = bp->b_addr;
 	xfs_dir3_data_check(dp, bp);
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(args->geo, hdr);
 	blp = xfs_dir2_block_leaf_p(btp);
 	/*
 	 * Loop doing a binary search for our hash value.
@@ -724,7 +723,7 @@ xfs_dir2_block_lookup_int(
 		 * Get pointer to the entry from the leaf.
 		 */
 		dep = (xfs_dir2_data_entry_t *)
-			((char *)hdr + xfs_dir2_dataptr_to_off(mp, addr));
+			((char *)hdr + xfs_dir2_dataptr_to_off(args->geo, addr));
 		/*
 		 * Compare name and if it's an exact match, return the index
 		 * and buffer. If it's the first case-insensitive match, store
@@ -791,18 +790,19 @@ xfs_dir2_block_removename(
 	tp = args->trans;
 	mp = dp->i_mount;
 	hdr = bp->b_addr;
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(args->geo, hdr);
 	blp = xfs_dir2_block_leaf_p(btp);
 	/*
 	 * Point to the data entry using the leaf entry.
 	 */
-	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)hdr + xfs_dir2_dataptr_to_off(mp, be32_to_cpu(blp[ent].address)));
+	dep = (xfs_dir2_data_entry_t *)((char *)hdr +
+			xfs_dir2_dataptr_to_off(args->geo,
+						be32_to_cpu(blp[ent].address)));
 	/*
 	 * Mark the data entry's space free.
 	 */
 	needlog = needscan = 0;
-	xfs_dir2_data_make_free(tp, dp, bp,
+	xfs_dir2_data_make_free(args, bp,
 		(xfs_dir2_data_aoff_t)((char *)dep - (char *)hdr),
 		dp->d_ops->data_entsize(dep->namelen), &needlog, &needscan);
 	/*
@@ -821,7 +821,7 @@ xfs_dir2_block_removename(
 	if (needscan)
 		xfs_dir2_data_freescan(dp, hdr, &needlog);
 	if (needlog)
-		xfs_dir2_data_log_header(tp, dp, bp);
+		xfs_dir2_data_log_header(args, bp);
 	xfs_dir3_data_check(dp, bp);
 	/*
 	 * See if the size as a shortform is good enough.
@@ -866,20 +866,21 @@ xfs_dir2_block_replace(
 	dp = args->dp;
 	mp = dp->i_mount;
 	hdr = bp->b_addr;
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(args->geo, hdr);
 	blp = xfs_dir2_block_leaf_p(btp);
 	/*
 	 * Point to the data entry we need to change.
 	 */
-	dep = (xfs_dir2_data_entry_t *)
-	      ((char *)hdr + xfs_dir2_dataptr_to_off(mp, be32_to_cpu(blp[ent].address)));
+	dep = (xfs_dir2_data_entry_t *)((char *)hdr +
+			xfs_dir2_dataptr_to_off(args->geo,
+						be32_to_cpu(blp[ent].address)));
 	ASSERT(be64_to_cpu(dep->inumber) != args->inumber);
 	/*
 	 * Change the inode number to the new value.
 	 */
 	dep->inumber = cpu_to_be64(args->inumber);
 	dp->d_ops->data_put_ftype(dep, args->filetype);
-	xfs_dir2_data_log_entry(args->trans, dp, bp, dep);
+	xfs_dir2_data_log_entry(args, bp, dep);
 	xfs_dir3_data_check(dp, bp);
 	return 0;
 }
@@ -939,7 +940,7 @@ xfs_dir2_leaf_to_block(
 	leaf = lbp->b_addr;
 	dp->d_ops->leaf_hdr_from_disk(&leafhdr, leaf);
 	ents = dp->d_ops->leaf_ents_p(leaf);
-	ltp = xfs_dir2_leaf_tail_p(mp, leaf);
+	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 
 	ASSERT(leafhdr.magic == XFS_DIR2_LEAF1_MAGIC ||
 	       leafhdr.magic == XFS_DIR3_LEAF1_MAGIC);
@@ -949,13 +950,13 @@ xfs_dir2_leaf_to_block(
 	 * been left behind during no-space-reservation operations.
 	 * These will show up in the leaf bests table.
 	 */
-	while (dp->i_d.di_size > mp->m_dirblksize) {
+	while (dp->i_d.di_size > args->geo->blksize) {
 		int hdrsz;
 
 		hdrsz = dp->d_ops->data_entry_offset;
 		bestsp = xfs_dir2_leaf_bests_p(ltp);
 		if (be16_to_cpu(bestsp[be32_to_cpu(ltp->bestcount) - 1]) ==
-					    mp->m_dirblksize - hdrsz) {
+					    args->geo->blksize - hdrsz) {
 			if ((error =
 			    xfs_dir2_leaf_trim_data(args, lbp,
 				    (xfs_dir2_db_t)(be32_to_cpu(ltp->bestcount) - 1))))
@@ -967,7 +968,7 @@ xfs_dir2_leaf_to_block(
 	 * Read the data block if we don't already have it, give up if it fails.
 	 */
 	if (!dbp) {
-		error = xfs_dir3_data_read(tp, dp, mp->m_dirdatablk, -1, &dbp);
+		error = xfs_dir3_data_read(tp, dp, args->geo->datablk, -1, &dbp);
 		if (error)
 			return error;
 	}
@@ -983,7 +984,7 @@ xfs_dir2_leaf_to_block(
 	/*
 	 * Look at the last data entry.
 	 */
-	tagp = (__be16 *)((char *)hdr + mp->m_dirblksize) - 1;
+	tagp = (__be16 *)((char *)hdr + args->geo->blksize) - 1;
 	dup = (xfs_dir2_data_unused_t *)((char *)hdr + be16_to_cpu(*tagp));
 	/*
 	 * If it's not free or is too short we can't do it.
@@ -1002,12 +1003,12 @@ xfs_dir2_leaf_to_block(
 	/*
 	 * Use up the space at the end of the block (blp/btp).
 	 */
-	xfs_dir2_data_use_free(tp, dp, dbp, dup, mp->m_dirblksize - size, size,
+	xfs_dir2_data_use_free(args, dbp, dup, args->geo->blksize - size, size,
 		&needlog, &needscan);
 	/*
 	 * Initialize the block tail.
 	 */
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(args->geo, hdr);
 	btp->count = cpu_to_be32(leafhdr.count - leafhdr.stale);
 	btp->stale = 0;
 	xfs_dir2_block_log_tail(tp, dbp);
@@ -1028,11 +1029,11 @@ xfs_dir2_leaf_to_block(
 	if (needscan)
 		xfs_dir2_data_freescan(dp, hdr, &needlog);
 	if (needlog)
-		xfs_dir2_data_log_header(tp, dp, dbp);
+		xfs_dir2_data_log_header(args, dbp);
 	/*
 	 * Pitch the old leaf block.
 	 */
-	error = xfs_da_shrink_inode(args, mp->m_dirleafblk, lbp);
+	error = xfs_da_shrink_inode(args, args->geo->leafblk, lbp);
 	if (error)
 		return error;
 
@@ -1141,13 +1142,13 @@ xfs_dir2_sf_to_block(
 	 */
 	dup = dp->d_ops->data_unused_p(hdr);
 	needlog = needscan = 0;
-	xfs_dir2_data_use_free(tp, dp, bp, dup, mp->m_dirblksize - i, i, &needlog,
-		&needscan);
+	xfs_dir2_data_use_free(args, bp, dup, args->geo->blksize - i,
+			       i, &needlog, &needscan);
 	ASSERT(needscan == 0);
 	/*
 	 * Fill in the tail.
 	 */
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(args->geo, hdr);
 	btp->count = cpu_to_be32(sfp->count + 2);	/* ., .. */
 	btp->stale = 0;
 	blp = xfs_dir2_block_leaf_p(btp);
@@ -1155,7 +1156,7 @@ xfs_dir2_sf_to_block(
 	/*
 	 * Remove the freespace, we'll manage it.
 	 */
-	xfs_dir2_data_use_free(tp, dp, bp, dup,
+	xfs_dir2_data_use_free(args, bp, dup,
 		(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr),
 		be16_to_cpu(dup->length), &needlog, &needscan);
 	/*
@@ -1168,9 +1169,9 @@ xfs_dir2_sf_to_block(
 	dp->d_ops->data_put_ftype(dep, XFS_DIR3_FT_DIR);
 	tagp = dp->d_ops->data_entry_tag_p(dep);
 	*tagp = cpu_to_be16((char *)dep - (char *)hdr);
-	xfs_dir2_data_log_entry(tp, dp, bp, dep);
+	xfs_dir2_data_log_entry(args, bp, dep);
 	blp[0].hashval = cpu_to_be32(xfs_dir_hash_dot);
-	blp[0].address = cpu_to_be32(xfs_dir2_byte_to_dataptr(mp,
+	blp[0].address = cpu_to_be32(xfs_dir2_byte_to_dataptr(
 				(char *)dep - (char *)hdr));
 	/*
 	 * Create entry for ..
@@ -1182,9 +1183,9 @@ xfs_dir2_sf_to_block(
 	dp->d_ops->data_put_ftype(dep, XFS_DIR3_FT_DIR);
 	tagp = dp->d_ops->data_entry_tag_p(dep);
 	*tagp = cpu_to_be16((char *)dep - (char *)hdr);
-	xfs_dir2_data_log_entry(tp, dp, bp, dep);
+	xfs_dir2_data_log_entry(args, bp, dep);
 	blp[1].hashval = cpu_to_be32(xfs_dir_hash_dotdot);
-	blp[1].address = cpu_to_be32(xfs_dir2_byte_to_dataptr(mp,
+	blp[1].address = cpu_to_be32(xfs_dir2_byte_to_dataptr(
 				(char *)dep - (char *)hdr));
 	offset = dp->d_ops->data_first_offset;
 	/*
@@ -1216,7 +1217,7 @@ xfs_dir2_sf_to_block(
 			dup->length = cpu_to_be16(newoffset - offset);
 			*xfs_dir2_data_unused_tag_p(dup) = cpu_to_be16(
 				((char *)dup - (char *)hdr));
-			xfs_dir2_data_log_unused(tp, bp, dup);
+			xfs_dir2_data_log_unused(args, bp, dup);
 			xfs_dir2_data_freeinsert(hdr,
 						 dp->d_ops->data_bestfree_p(hdr),
 						 dup, &dummy);
@@ -1233,12 +1234,12 @@ xfs_dir2_sf_to_block(
 		memcpy(dep->name, sfep->name, dep->namelen);
 		tagp = dp->d_ops->data_entry_tag_p(dep);
 		*tagp = cpu_to_be16((char *)dep - (char *)hdr);
-		xfs_dir2_data_log_entry(tp, dp, bp, dep);
+		xfs_dir2_data_log_entry(args, bp, dep);
 		name.name = sfep->name;
 		name.len = sfep->namelen;
 		blp[2 + i].hashval = cpu_to_be32(mp->m_dirnameops->
 							hashname(&name));
-		blp[2 + i].address = cpu_to_be32(xfs_dir2_byte_to_dataptr(mp,
+		blp[2 + i].address = cpu_to_be32(xfs_dir2_byte_to_dataptr(
 						 (char *)dep - (char *)hdr));
 		offset = (int)((char *)(tagp + 1) - (char *)hdr);
 		if (++i == sfp->count)

fs/xfs/xfs_dir2_data.c

@@ -63,8 +63,10 @@ __xfs_dir3_data_check(
 	int			stale;		/* count of stale leaves */
 	struct xfs_name		name;
 	const struct xfs_dir_ops *ops;
+	struct xfs_da_geometry	*geo;
 
 	mp = bp->b_target->bt_mount;
+	geo = mp->m_dir_geo;
 
 	/*
 	 * We can be passed a null dp here from a verifier, so we need to go the
@@ -78,7 +80,7 @@ __xfs_dir3_data_check(
 	switch (hdr->magic) {
 	case cpu_to_be32(XFS_DIR3_BLOCK_MAGIC):
 	case cpu_to_be32(XFS_DIR2_BLOCK_MAGIC):
-		btp = xfs_dir2_block_tail_p(mp, hdr);
+		btp = xfs_dir2_block_tail_p(geo, hdr);
 		lep = xfs_dir2_block_leaf_p(btp);
 		endp = (char *)lep;
 
@@ -94,7 +96,7 @@ __xfs_dir3_data_check(
 		break;
 	case cpu_to_be32(XFS_DIR3_DATA_MAGIC):
 	case cpu_to_be32(XFS_DIR2_DATA_MAGIC):
-		endp = (char *)hdr + mp->m_dirblksize;
+		endp = (char *)hdr + geo->blksize;
 		break;
 	default:
 		XFS_ERROR_REPORT("Bad Magic", XFS_ERRLEVEL_LOW, mp);
@@ -172,9 +174,9 @@ __xfs_dir3_data_check(
 		lastfree = 0;
 		if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
 		    hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) {
-			addr = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
-				(xfs_dir2_data_aoff_t)
-				((char *)dep - (char *)hdr));
+			addr = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
+						(xfs_dir2_data_aoff_t)
+						((char *)dep - (char *)hdr));
 			name.name = dep->name;
 			name.len = dep->namelen;
 			hash = mp->m_dirnameops->hashname(&name);
@@ -329,12 +331,11 @@ xfs_dir3_data_read(
 
 int
 xfs_dir3_data_readahead(
-	struct xfs_trans	*tp,
 	struct xfs_inode	*dp,
 	xfs_dablk_t		bno,
 	xfs_daddr_t		mapped_bno)
 {
-	return xfs_da_reada_buf(tp, dp, bno, mapped_bno,
+	return xfs_da_reada_buf(dp, bno, mapped_bno,
 				XFS_DATA_FORK, &xfs_dir3_data_reada_buf_ops);
 }
 
@@ -510,6 +511,7 @@ xfs_dir2_data_freescan(
 	struct xfs_dir2_data_free *bf;
 	char			*endp;		/* end of block's data */
 	char			*p;		/* current entry pointer */
+	struct xfs_da_geometry	*geo = dp->i_mount->m_dir_geo;
 
 	ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
 	       hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC) ||
@@ -528,10 +530,10 @@ xfs_dir2_data_freescan(
 	p = (char *)dp->d_ops->data_entry_p(hdr);
 	if (hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
 	    hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC)) {
-		btp = xfs_dir2_block_tail_p(dp->i_mount, hdr);
+		btp = xfs_dir2_block_tail_p(geo, hdr);
 		endp = (char *)xfs_dir2_block_leaf_p(btp);
 	} else
-		endp = (char *)hdr + dp->i_mount->m_dirblksize;
+		endp = (char *)hdr + geo->blksize;
 	/*
 	 * Loop over the block's entries.
 	 */
@@ -585,8 +587,8 @@ xfs_dir3_data_init(
 	/*
 	 * Get the buffer set up for the block.
 	 */
-	error = xfs_da_get_buf(tp, dp, xfs_dir2_db_to_da(mp, blkno), -1, &bp,
-		XFS_DATA_FORK);
+	error = xfs_da_get_buf(tp, dp, xfs_dir2_db_to_da(args->geo, blkno),
+			       -1, &bp, XFS_DATA_FORK);
 	if (error)
 		return error;
 	bp->b_ops = &xfs_dir3_data_buf_ops;
@@ -621,15 +623,15 @@ xfs_dir3_data_init(
 	dup = dp->d_ops->data_unused_p(hdr);
 	dup->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
 
-	t = mp->m_dirblksize - (uint)dp->d_ops->data_entry_offset;
+	t = args->geo->blksize - (uint)dp->d_ops->data_entry_offset;
 	bf[0].length = cpu_to_be16(t);
 	dup->length = cpu_to_be16(t);
 	*xfs_dir2_data_unused_tag_p(dup) = cpu_to_be16((char *)dup - (char *)hdr);
 	/*
 	 * Log it and return it.
 	 */
-	xfs_dir2_data_log_header(tp, dp, bp);
-	xfs_dir2_data_log_unused(tp, bp, dup);
+	xfs_dir2_data_log_header(args, bp);
+	xfs_dir2_data_log_unused(args, bp, dup);
 	*bpp = bp;
 	return 0;
 }
@@ -639,8 +641,7 @@ xfs_dir3_data_init(
  */
 void
 xfs_dir2_data_log_entry(
-	struct xfs_trans	*tp,
-	struct xfs_inode	*dp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp,
 	xfs_dir2_data_entry_t	*dep)		/* data entry pointer */
 {
@@ -651,8 +652,8 @@ xfs_dir2_data_log_entry(
 	       hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
 	       hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
 
-	xfs_trans_log_buf(tp, bp, (uint)((char *)dep - (char *)hdr),
-		(uint)((char *)(dp->d_ops->data_entry_tag_p(dep) + 1) -
+	xfs_trans_log_buf(args->trans, bp, (uint)((char *)dep - (char *)hdr),
+		(uint)((char *)(args->dp->d_ops->data_entry_tag_p(dep) + 1) -
 		       (char *)hdr - 1));
 }
 
@@ -661,8 +662,7 @@ xfs_dir2_data_log_entry(
  */
 void
 xfs_dir2_data_log_header(
-	struct xfs_trans	*tp,
-	struct xfs_inode	*dp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp)
 {
 #ifdef DEBUG
@@ -674,7 +674,8 @@ xfs_dir2_data_log_header(
 	       hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
 #endif
 
-	xfs_trans_log_buf(tp, bp, 0, dp->d_ops->data_entry_offset - 1);
+	xfs_trans_log_buf(args->trans, bp, 0,
+			  args->dp->d_ops->data_entry_offset - 1);
 }
 
 /*
@@ -682,7 +683,7 @@ xfs_dir2_data_log_header(
  */
 void
 xfs_dir2_data_log_unused(
-	struct xfs_trans	*tp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp,
 	xfs_dir2_data_unused_t	*dup)		/* data unused pointer */
 {
@@ -696,13 +697,13 @@ xfs_dir2_data_log_unused(
 	/*
 	 * Log the first part of the unused entry.
 	 */
-	xfs_trans_log_buf(tp, bp, (uint)((char *)dup - (char *)hdr),
+	xfs_trans_log_buf(args->trans, bp, (uint)((char *)dup - (char *)hdr),
 		(uint)((char *)&dup->length + sizeof(dup->length) -
 		       1 - (char *)hdr));
 	/*
 	 * Log the end (tag) of the unused entry.
 	 */
-	xfs_trans_log_buf(tp, bp,
+	xfs_trans_log_buf(args->trans, bp,
 		(uint)((char *)xfs_dir2_data_unused_tag_p(dup) - (char *)hdr),
 		(uint)((char *)xfs_dir2_data_unused_tag_p(dup) - (char *)hdr +
 		       sizeof(xfs_dir2_data_off_t) - 1));
@@ -714,8 +715,7 @@ xfs_dir2_data_log_unused(
  */
 void
 xfs_dir2_data_make_free(
-	struct xfs_trans	*tp,
-	struct xfs_inode	*dp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp,
 	xfs_dir2_data_aoff_t	offset,		/* starting byte offset */
 	xfs_dir2_data_aoff_t	len,		/* length in bytes */
@@ -725,14 +725,12 @@ xfs_dir2_data_make_free(
 	xfs_dir2_data_hdr_t	*hdr;		/* data block pointer */
 	xfs_dir2_data_free_t	*dfp;		/* bestfree pointer */
 	char			*endptr;	/* end of data area */
-	xfs_mount_t		*mp;		/* filesystem mount point */
 	int			needscan;	/* need to regen bestfree */
 	xfs_dir2_data_unused_t	*newdup;	/* new unused entry */
 	xfs_dir2_data_unused_t	*postdup;	/* unused entry after us */
 	xfs_dir2_data_unused_t	*prevdup;	/* unused entry before us */
 	struct xfs_dir2_data_free *bf;
 
-	mp = tp->t_mountp;
 	hdr = bp->b_addr;
 
 	/*
@@ -740,20 +738,20 @@ xfs_dir2_data_make_free(
 	 */
 	if (hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
 	    hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC))
-		endptr = (char *)hdr + mp->m_dirblksize;
+		endptr = (char *)hdr + args->geo->blksize;
 	else {
 		xfs_dir2_block_tail_t	*btp;	/* block tail */
 
 		ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
 			hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
-		btp = xfs_dir2_block_tail_p(mp, hdr);
+		btp = xfs_dir2_block_tail_p(args->geo, hdr);
 		endptr = (char *)xfs_dir2_block_leaf_p(btp);
 	}
 	/*
 	 * If this isn't the start of the block, then back up to
 	 * the previous entry and see if it's free.
 	 */
-	if (offset > dp->d_ops->data_entry_offset) {
+	if (offset > args->dp->d_ops->data_entry_offset) {
 		__be16			*tagp;	/* tag just before us */
 
 		tagp = (__be16 *)((char *)hdr + offset) - 1;
@@ -779,7 +777,7 @@ xfs_dir2_data_make_free(
 	 * Previous and following entries are both free,
 	 * merge everything into a single free entry.
 	 */
-	bf = dp->d_ops->data_bestfree_p(hdr);
+	bf = args->dp->d_ops->data_bestfree_p(hdr);
 	if (prevdup && postdup) {
 		xfs_dir2_data_free_t	*dfp2;	/* another bestfree pointer */
 
@@ -801,7 +799,7 @@ xfs_dir2_data_make_free(
 		be16_add_cpu(&prevdup->length, len + be16_to_cpu(postdup->length));
 		*xfs_dir2_data_unused_tag_p(prevdup) =
 			cpu_to_be16((char *)prevdup - (char *)hdr);
-		xfs_dir2_data_log_unused(tp, bp, prevdup);
+		xfs_dir2_data_log_unused(args, bp, prevdup);
 		if (!needscan) {
 			/*
 			 * Has to be the case that entries 0 and 1 are
@@ -836,7 +834,7 @@ xfs_dir2_data_make_free(
 		be16_add_cpu(&prevdup->length, len);
 		*xfs_dir2_data_unused_tag_p(prevdup) =
 			cpu_to_be16((char *)prevdup - (char *)hdr);
-		xfs_dir2_data_log_unused(tp, bp, prevdup);
+		xfs_dir2_data_log_unused(args, bp, prevdup);
 		/*
 		 * If the previous entry was in the table, the new entry
 		 * is longer, so it will be in the table too.  Remove
@@ -864,7 +862,7 @@ xfs_dir2_data_make_free(
 		newdup->length = cpu_to_be16(len + be16_to_cpu(postdup->length));
 		*xfs_dir2_data_unused_tag_p(newdup) =
 			cpu_to_be16((char *)newdup - (char *)hdr);
-		xfs_dir2_data_log_unused(tp, bp, newdup);
+		xfs_dir2_data_log_unused(args, bp, newdup);
 		/*
 		 * If the following entry was in the table, the new entry
 		 * is longer, so it will be in the table too.  Remove
@@ -891,7 +889,7 @@ xfs_dir2_data_make_free(
 		newdup->length = cpu_to_be16(len);
 		*xfs_dir2_data_unused_tag_p(newdup) =
 			cpu_to_be16((char *)newdup - (char *)hdr);
-		xfs_dir2_data_log_unused(tp, bp, newdup);
+		xfs_dir2_data_log_unused(args, bp, newdup);
 		xfs_dir2_data_freeinsert(hdr, bf, newdup, needlogp);
 	}
 	*needscanp = needscan;
@@ -902,8 +900,7 @@ xfs_dir2_data_make_free(
  */
 void
 xfs_dir2_data_use_free(
-	struct xfs_trans	*tp,
-	struct xfs_inode	*dp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp,
 	xfs_dir2_data_unused_t	*dup,		/* unused entry */
 	xfs_dir2_data_aoff_t	offset,		/* starting offset to use */
@@ -934,7 +931,7 @@ xfs_dir2_data_use_free(
 	 * Look up the entry in the bestfree table.
 	 */
 	oldlen = be16_to_cpu(dup->length);
-	bf = dp->d_ops->data_bestfree_p(hdr);
+	bf = args->dp->d_ops->data_bestfree_p(hdr);
 	dfp = xfs_dir2_data_freefind(hdr, bf, dup);
 	ASSERT(dfp || oldlen <= be16_to_cpu(bf[2].length));
 	/*
@@ -966,7 +963,7 @@ xfs_dir2_data_use_free(
 		newdup->length = cpu_to_be16(oldlen - len);
 		*xfs_dir2_data_unused_tag_p(newdup) =
 			cpu_to_be16((char *)newdup - (char *)hdr);
-		xfs_dir2_data_log_unused(tp, bp, newdup);
+		xfs_dir2_data_log_unused(args, bp, newdup);
 		/*
 		 * If it was in the table, remove it and add the new one.
 		 */
@@ -994,7 +991,7 @@ xfs_dir2_data_use_free(
 		newdup->length = cpu_to_be16(((char *)hdr + offset) - (char *)newdup);
 		*xfs_dir2_data_unused_tag_p(newdup) =
 			cpu_to_be16((char *)newdup - (char *)hdr);
-		xfs_dir2_data_log_unused(tp, bp, newdup);
+		xfs_dir2_data_log_unused(args, bp, newdup);
 		/*
 		 * If it was in the table, remove it and add the new one.
 		 */
@@ -1022,13 +1019,13 @@ xfs_dir2_data_use_free(
 		newdup->length = cpu_to_be16(((char *)hdr + offset) - (char *)newdup);
 		*xfs_dir2_data_unused_tag_p(newdup) =
 			cpu_to_be16((char *)newdup - (char *)hdr);
-		xfs_dir2_data_log_unused(tp, bp, newdup);
+		xfs_dir2_data_log_unused(args, bp, newdup);
 		newdup2 = (xfs_dir2_data_unused_t *)((char *)hdr + offset + len);
 		newdup2->freetag = cpu_to_be16(XFS_DIR2_DATA_FREE_TAG);
 		newdup2->length = cpu_to_be16(oldlen - len - be16_to_cpu(newdup->length));
 		*xfs_dir2_data_unused_tag_p(newdup2) =
 			cpu_to_be16((char *)newdup2 - (char *)hdr);
-		xfs_dir2_data_log_unused(tp, bp, newdup2);
+		xfs_dir2_data_log_unused(args, bp, newdup2);
 		/*
 		 * If the old entry was in the table, we need to scan
 		 * if the 3rd entry was valid, since these entries

fs/xfs/xfs_dir2_leaf.c

@@ -41,9 +41,10 @@
  */
 static int xfs_dir2_leaf_lookup_int(xfs_da_args_t *args, struct xfs_buf **lbpp,
 				    int *indexp, struct xfs_buf **dbpp);
-static void xfs_dir3_leaf_log_bests(struct xfs_trans *tp, struct xfs_buf *bp,
-				    int first, int last);
-static void xfs_dir3_leaf_log_tail(struct xfs_trans *tp, struct xfs_buf *bp);
+static void xfs_dir3_leaf_log_bests(struct xfs_da_args *args,
+				    struct xfs_buf *bp, int first, int last);
+static void xfs_dir3_leaf_log_tail(struct xfs_da_args *args,
+				   struct xfs_buf *bp);
 
 /*
  * Check the internal consistency of a leaf1 block.
@@ -92,6 +93,7 @@ xfs_dir3_leaf_check_int(
 	int			i;
 	const struct xfs_dir_ops *ops;
 	struct xfs_dir3_icleaf_hdr leafhdr;
+	struct xfs_da_geometry	*geo = mp->m_dir_geo;
 
 	/*
 	 * we can be passed a null dp here from a verifier, so we need to go the
@@ -105,14 +107,14 @@ xfs_dir3_leaf_check_int(
 	}
 
 	ents = ops->leaf_ents_p(leaf);
-	ltp = xfs_dir2_leaf_tail_p(mp, leaf);
+	ltp = xfs_dir2_leaf_tail_p(geo, leaf);
 
 	/*
 	 * XXX (dgc): This value is not restrictive enough.
 	 * Should factor in the size of the bests table as well.
 	 * We can deduce a value for that from di_size.
 	 */
-	if (hdr->count > ops->leaf_max_ents(mp))
+	if (hdr->count > ops->leaf_max_ents(geo))
 		return false;
 
 	/* Leaves and bests don't overlap in leaf format. */
@@ -323,7 +325,7 @@ xfs_dir3_leaf_init(
 	if (type == XFS_DIR2_LEAF1_MAGIC) {
 		struct xfs_dir2_leaf_tail *ltp;
 
-		ltp = xfs_dir2_leaf_tail_p(mp, leaf);
+		ltp = xfs_dir2_leaf_tail_p(mp->m_dir_geo, leaf);
 		ltp->bestcount = 0;
 		bp->b_ops = &xfs_dir3_leaf1_buf_ops;
 		xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DIR_LEAF1_BUF);
@@ -347,18 +349,18 @@ xfs_dir3_leaf_get_buf(
 	int			error;
 
 	ASSERT(magic == XFS_DIR2_LEAF1_MAGIC || magic == XFS_DIR2_LEAFN_MAGIC);
-	ASSERT(bno >= XFS_DIR2_LEAF_FIRSTDB(mp) &&
-	       bno < XFS_DIR2_FREE_FIRSTDB(mp));
+	ASSERT(bno >= xfs_dir2_byte_to_db(args->geo, XFS_DIR2_LEAF_OFFSET) &&
+	       bno < xfs_dir2_byte_to_db(args->geo, XFS_DIR2_FREE_OFFSET));
 
-	error = xfs_da_get_buf(tp, dp, xfs_dir2_db_to_da(mp, bno), -1, &bp,
-			       XFS_DATA_FORK);
+	error = xfs_da_get_buf(tp, dp, xfs_dir2_db_to_da(args->geo, bno),
+			       -1, &bp, XFS_DATA_FORK);
 	if (error)
 		return error;
 
 	xfs_dir3_leaf_init(mp, tp, bp, dp->i_ino, magic);
-	xfs_dir3_leaf_log_header(tp, dp, bp);
+	xfs_dir3_leaf_log_header(args, bp);
 	if (magic == XFS_DIR2_LEAF1_MAGIC)
-		xfs_dir3_leaf_log_tail(tp, bp);
+		xfs_dir3_leaf_log_tail(args, bp);
 	*bpp = bp;
 	return 0;
 }
@@ -403,8 +405,8 @@ xfs_dir2_block_to_leaf(
 	if ((error = xfs_da_grow_inode(args, &blkno))) {
 		return error;
 	}
-	ldb = xfs_dir2_da_to_db(mp, blkno);
-	ASSERT(ldb == XFS_DIR2_LEAF_FIRSTDB(mp));
+	ldb = xfs_dir2_da_to_db(args->geo, blkno);
+	ASSERT(ldb == xfs_dir2_byte_to_db(args->geo, XFS_DIR2_LEAF_OFFSET));
 	/*
 	 * Initialize the leaf block, get a buffer for it.
 	 */
@@ -415,7 +417,7 @@ xfs_dir2_block_to_leaf(
 	leaf = lbp->b_addr;
 	hdr = dbp->b_addr;
 	xfs_dir3_data_check(dp, dbp);
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(args->geo, hdr);
 	blp = xfs_dir2_block_leaf_p(btp);
 	bf = dp->d_ops->data_bestfree_p(hdr);
 	ents = dp->d_ops->leaf_ents_p(leaf);
@@ -427,23 +429,23 @@ xfs_dir2_block_to_leaf(
 	leafhdr.count = be32_to_cpu(btp->count);
 	leafhdr.stale = be32_to_cpu(btp->stale);
 	dp->d_ops->leaf_hdr_to_disk(leaf, &leafhdr);
-	xfs_dir3_leaf_log_header(tp, dp, lbp);
+	xfs_dir3_leaf_log_header(args, lbp);
 
 	/*
 	 * Could compact these but I think we always do the conversion
 	 * after squeezing out stale entries.
 	 */
 	memcpy(ents, blp, be32_to_cpu(btp->count) * sizeof(xfs_dir2_leaf_entry_t));
-	xfs_dir3_leaf_log_ents(tp, dp, lbp, 0, leafhdr.count - 1);
+	xfs_dir3_leaf_log_ents(args, lbp, 0, leafhdr.count - 1);
 	needscan = 0;
 	needlog = 1;
 	/*
 	 * Make the space formerly occupied by the leaf entries and block
 	 * tail be free.
 	 */
-	xfs_dir2_data_make_free(tp, dp, dbp,
+	xfs_dir2_data_make_free(args, dbp,
 		(xfs_dir2_data_aoff_t)((char *)blp - (char *)hdr),
-		(xfs_dir2_data_aoff_t)((char *)hdr + mp->m_dirblksize -
+		(xfs_dir2_data_aoff_t)((char *)hdr + args->geo->blksize -
 				       (char *)blp),
 		&needlog, &needscan);
 	/*
@@ -461,7 +463,7 @@ xfs_dir2_block_to_leaf(
 	/*
 	 * Set up leaf tail and bests table.
 	 */
-	ltp = xfs_dir2_leaf_tail_p(mp, leaf);
+	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 	ltp->bestcount = cpu_to_be32(1);
 	bestsp = xfs_dir2_leaf_bests_p(ltp);
 	bestsp[0] =  bf[0].length;
@@ -469,10 +471,10 @@ xfs_dir2_block_to_leaf(
 	 * Log the data header and leaf bests table.
 	 */
 	if (needlog)
-		xfs_dir2_data_log_header(tp, dp, dbp);
+		xfs_dir2_data_log_header(args, dbp);
 	xfs_dir3_leaf_check(dp, lbp);
 	xfs_dir3_data_check(dp, dbp);
-	xfs_dir3_leaf_log_bests(tp, lbp, 0, 0);
+	xfs_dir3_leaf_log_bests(args, lbp, 0, 0);
 	return 0;
 }
 
@@ -641,7 +643,7 @@ xfs_dir2_leaf_addname(
 	tp = args->trans;
 	mp = dp->i_mount;
 
-	error = xfs_dir3_leaf_read(tp, dp, mp->m_dirleafblk, -1, &lbp);
+	error = xfs_dir3_leaf_read(tp, dp, args->geo->leafblk, -1, &lbp);
 	if (error)
 		return error;
 
@@ -653,7 +655,7 @@ xfs_dir2_leaf_addname(
 	 */
 	index = xfs_dir2_leaf_search_hash(args, lbp);
 	leaf = lbp->b_addr;
-	ltp = xfs_dir2_leaf_tail_p(mp, leaf);
+	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 	ents = dp->d_ops->leaf_ents_p(leaf);
 	dp->d_ops->leaf_hdr_from_disk(&leafhdr, leaf);
 	bestsp = xfs_dir2_leaf_bests_p(ltp);
@@ -670,7 +672,7 @@ xfs_dir2_leaf_addname(
 	     index++, lep++) {
 		if (be32_to_cpu(lep->address) == XFS_DIR2_NULL_DATAPTR)
 			continue;
-		i = xfs_dir2_dataptr_to_db(mp, be32_to_cpu(lep->address));
+		i = xfs_dir2_dataptr_to_db(args->geo, be32_to_cpu(lep->address));
 		ASSERT(i < be32_to_cpu(ltp->bestcount));
 		ASSERT(bestsp[i] != cpu_to_be16(NULLDATAOFF));
 		if (be16_to_cpu(bestsp[i]) >= length) {
@@ -810,14 +812,15 @@ xfs_dir2_leaf_addname(
 			memmove(&bestsp[0], &bestsp[1],
 				be32_to_cpu(ltp->bestcount) * sizeof(bestsp[0]));
 			be32_add_cpu(&ltp->bestcount, 1);
-			xfs_dir3_leaf_log_tail(tp, lbp);
-			xfs_dir3_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
+			xfs_dir3_leaf_log_tail(args, lbp);
+			xfs_dir3_leaf_log_bests(args, lbp, 0,
+						be32_to_cpu(ltp->bestcount) - 1);
 		}
 		/*
 		 * If we're filling in a previously empty block just log it.
 		 */
 		else
-			xfs_dir3_leaf_log_bests(tp, lbp, use_block, use_block);
+			xfs_dir3_leaf_log_bests(args, lbp, use_block, use_block);
 		hdr = dbp->b_addr;
 		bf = dp->d_ops->data_bestfree_p(hdr);
 		bestsp[use_block] = bf[0].length;
@@ -828,8 +831,8 @@ xfs_dir2_leaf_addname(
 		 * Just read that one in.
 		 */
 		error = xfs_dir3_data_read(tp, dp,
-					   xfs_dir2_db_to_da(mp, use_block),
-					   -1, &dbp);
+				   xfs_dir2_db_to_da(args->geo, use_block),
+				   -1, &dbp);
 		if (error) {
 			xfs_trans_brelse(tp, lbp);
 			return error;
@@ -848,7 +851,7 @@ xfs_dir2_leaf_addname(
 	/*
 	 * Mark the initial part of our freespace in use for the new entry.
 	 */
-	xfs_dir2_data_use_free(tp, dp, dbp, dup,
+	xfs_dir2_data_use_free(args, dbp, dup,
 		(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr), length,
 		&needlog, &needscan);
 	/*
@@ -870,8 +873,8 @@ xfs_dir2_leaf_addname(
 	 * Need to log the data block's header.
 	 */
 	if (needlog)
-		xfs_dir2_data_log_header(tp, dp, dbp);
-	xfs_dir2_data_log_entry(tp, dp, dbp, dep);
+		xfs_dir2_data_log_header(args, dbp);
+	xfs_dir2_data_log_entry(args, dbp, dep);
 	/*
 	 * If the bests table needs to be changed, do it.
 	 * Log the change unless we've already done that.
@@ -879,7 +882,7 @@ xfs_dir2_leaf_addname(
 	if (be16_to_cpu(bestsp[use_block]) != be16_to_cpu(bf[0].length)) {
 		bestsp[use_block] = bf[0].length;
 		if (!grown)
-			xfs_dir3_leaf_log_bests(tp, lbp, use_block, use_block);
+			xfs_dir3_leaf_log_bests(args, lbp, use_block, use_block);
 	}
 
 	lep = xfs_dir3_leaf_find_entry(&leafhdr, ents, index, compact, lowstale,
@@ -889,14 +892,15 @@ xfs_dir2_leaf_addname(
 	 * Fill in the new leaf entry.
 	 */
 	lep->hashval = cpu_to_be32(args->hashval);
-	lep->address = cpu_to_be32(xfs_dir2_db_off_to_dataptr(mp, use_block,
+	lep->address = cpu_to_be32(
+				xfs_dir2_db_off_to_dataptr(args->geo, use_block,
 				be16_to_cpu(*tagp)));
 	/*
 	 * Log the leaf fields and give up the buffers.
 	 */
 	dp->d_ops->leaf_hdr_to_disk(leaf, &leafhdr);
-	xfs_dir3_leaf_log_header(tp, dp, lbp);
-	xfs_dir3_leaf_log_ents(tp, dp, lbp, lfloglow, lfloghigh);
+	xfs_dir3_leaf_log_header(args, lbp);
+	xfs_dir3_leaf_log_ents(args, lbp, lfloglow, lfloghigh);
 	xfs_dir3_leaf_check(dp, lbp);
 	xfs_dir3_data_check(dp, dbp);
 	return 0;
@@ -948,9 +952,9 @@ xfs_dir3_leaf_compact(
 	leafhdr->stale = 0;
 
 	dp->d_ops->leaf_hdr_to_disk(leaf, leafhdr);
-	xfs_dir3_leaf_log_header(args->trans, dp, bp);
+	xfs_dir3_leaf_log_header(args, bp);
 	if (loglow != -1)
-		xfs_dir3_leaf_log_ents(args->trans, dp, bp, loglow, to - 1);
+		xfs_dir3_leaf_log_ents(args, bp, loglow, to - 1);
 }
 
 /*
@@ -1052,7 +1056,7 @@ xfs_dir3_leaf_compact_x1(
  */
 static void
 xfs_dir3_leaf_log_bests(
-	xfs_trans_t		*tp,		/* transaction pointer */
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp,		/* leaf buffer */
 	int			first,		/* first entry to log */
 	int			last)		/* last entry to log */
@@ -1065,10 +1069,11 @@ xfs_dir3_leaf_log_bests(
 	ASSERT(leaf->hdr.info.magic == cpu_to_be16(XFS_DIR2_LEAF1_MAGIC) ||
 	       leaf->hdr.info.magic == cpu_to_be16(XFS_DIR3_LEAF1_MAGIC));
 
-	ltp = xfs_dir2_leaf_tail_p(tp->t_mountp, leaf);
+	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 	firstb = xfs_dir2_leaf_bests_p(ltp) + first;
 	lastb = xfs_dir2_leaf_bests_p(ltp) + last;
-	xfs_trans_log_buf(tp, bp, (uint)((char *)firstb - (char *)leaf),
+	xfs_trans_log_buf(args->trans, bp,
+		(uint)((char *)firstb - (char *)leaf),
 		(uint)((char *)lastb - (char *)leaf + sizeof(*lastb) - 1));
 }
 
@@ -1077,8 +1082,7 @@ xfs_dir3_leaf_log_bests(
  */
 void
 xfs_dir3_leaf_log_ents(
-	struct xfs_trans	*tp,
-	struct xfs_inode	*dp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp,
 	int			first,
 	int			last)
@@ -1093,10 +1097,11 @@ xfs_dir3_leaf_log_ents(
 	       leaf->hdr.info.magic == cpu_to_be16(XFS_DIR2_LEAFN_MAGIC) ||
 	       leaf->hdr.info.magic == cpu_to_be16(XFS_DIR3_LEAFN_MAGIC));
 
-	ents = dp->d_ops->leaf_ents_p(leaf);
+	ents = args->dp->d_ops->leaf_ents_p(leaf);
 	firstlep = &ents[first];
 	lastlep = &ents[last];
-	xfs_trans_log_buf(tp, bp, (uint)((char *)firstlep - (char *)leaf),
+	xfs_trans_log_buf(args->trans, bp,
+		(uint)((char *)firstlep - (char *)leaf),
 		(uint)((char *)lastlep - (char *)leaf + sizeof(*lastlep) - 1));
 }
 
@@ -1105,8 +1110,7 @@ xfs_dir3_leaf_log_ents(
  */
 void
 xfs_dir3_leaf_log_header(
-	struct xfs_trans	*tp,
-	struct xfs_inode	*dp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp)
 {
 	struct xfs_dir2_leaf	*leaf = bp->b_addr;
@@ -1116,8 +1120,9 @@ xfs_dir3_leaf_log_header(
 	       leaf->hdr.info.magic == cpu_to_be16(XFS_DIR2_LEAFN_MAGIC) ||
 	       leaf->hdr.info.magic == cpu_to_be16(XFS_DIR3_LEAFN_MAGIC));
 
-	xfs_trans_log_buf(tp, bp, (uint)((char *)&leaf->hdr - (char *)leaf),
-			  dp->d_ops->leaf_hdr_size - 1);
+	xfs_trans_log_buf(args->trans, bp,
+			  (uint)((char *)&leaf->hdr - (char *)leaf),
+			  args->dp->d_ops->leaf_hdr_size - 1);
 }
 
 /*
@@ -1125,21 +1130,20 @@ xfs_dir3_leaf_log_header(
  */
 STATIC void
 xfs_dir3_leaf_log_tail(
-	struct xfs_trans	*tp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp)
 {
 	struct xfs_dir2_leaf	*leaf = bp->b_addr;
 	xfs_dir2_leaf_tail_t	*ltp;		/* leaf tail structure */
-	struct xfs_mount	*mp = tp->t_mountp;
 
 	ASSERT(leaf->hdr.info.magic == cpu_to_be16(XFS_DIR2_LEAF1_MAGIC) ||
 	       leaf->hdr.info.magic == cpu_to_be16(XFS_DIR3_LEAF1_MAGIC) ||
 	       leaf->hdr.info.magic == cpu_to_be16(XFS_DIR2_LEAFN_MAGIC) ||
 	       leaf->hdr.info.magic == cpu_to_be16(XFS_DIR3_LEAFN_MAGIC));
 
-	ltp = xfs_dir2_leaf_tail_p(mp, leaf);
-	xfs_trans_log_buf(tp, bp, (uint)((char *)ltp - (char *)leaf),
-		(uint)(mp->m_dirblksize - 1));
+	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
+	xfs_trans_log_buf(args->trans, bp, (uint)((char *)ltp - (char *)leaf),
+		(uint)(args->geo->blksize - 1));
 }
 
 /*
@@ -1185,7 +1189,7 @@ xfs_dir2_leaf_lookup(
 	 */
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)dbp->b_addr +
-	       xfs_dir2_dataptr_to_off(dp->i_mount, be32_to_cpu(lep->address)));
+	       xfs_dir2_dataptr_to_off(args->geo, be32_to_cpu(lep->address)));
 	/*
 	 * Return the found inode number & CI name if appropriate
 	 */
@@ -1231,7 +1235,7 @@ xfs_dir2_leaf_lookup_int(
 	tp = args->trans;
 	mp = dp->i_mount;
 
-	error = xfs_dir3_leaf_read(tp, dp, mp->m_dirleafblk, -1, &lbp);
+	error = xfs_dir3_leaf_read(tp, dp, args->geo->leafblk, -1, &lbp);
 	if (error)
 		return error;
 
@@ -1260,7 +1264,8 @@ xfs_dir2_leaf_lookup_int(
 		/*
 		 * Get the new data block number.
 		 */
-		newdb = xfs_dir2_dataptr_to_db(mp, be32_to_cpu(lep->address));
+		newdb = xfs_dir2_dataptr_to_db(args->geo,
+					       be32_to_cpu(lep->address));
 		/*
 		 * If it's not the same as the old data block number,
 		 * need to pitch the old one and read the new one.
@@ -1269,8 +1274,8 @@ xfs_dir2_leaf_lookup_int(
 			if (dbp)
 				xfs_trans_brelse(tp, dbp);
 			error = xfs_dir3_data_read(tp, dp,
-						   xfs_dir2_db_to_da(mp, newdb),
-						   -1, &dbp);
+					   xfs_dir2_db_to_da(args->geo, newdb),
+					   -1, &dbp);
 			if (error) {
 				xfs_trans_brelse(tp, lbp);
 				return error;
@@ -1281,7 +1286,8 @@ xfs_dir2_leaf_lookup_int(
 		 * Point to the data entry.
 		 */
 		dep = (xfs_dir2_data_entry_t *)((char *)dbp->b_addr +
-			xfs_dir2_dataptr_to_off(mp, be32_to_cpu(lep->address)));
+			xfs_dir2_dataptr_to_off(args->geo,
+						be32_to_cpu(lep->address)));
 		/*
 		 * Compare name and if it's an exact match, return the index
 		 * and buffer. If it's the first case-insensitive match, store
@@ -1310,8 +1316,8 @@ xfs_dir2_leaf_lookup_int(
 		if (cidb != curdb) {
 			xfs_trans_brelse(tp, dbp);
 			error = xfs_dir3_data_read(tp, dp,
-						   xfs_dir2_db_to_da(mp, cidb),
-						   -1, &dbp);
+					   xfs_dir2_db_to_da(args->geo, cidb),
+					   -1, &dbp);
 			if (error) {
 				xfs_trans_brelse(tp, lbp);
 				return error;
@@ -1380,18 +1386,18 @@ xfs_dir2_leaf_removename(
 	 * Point to the leaf entry, use that to point to the data entry.
 	 */
 	lep = &ents[index];
-	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)));
+	db = xfs_dir2_dataptr_to_db(args->geo, be32_to_cpu(lep->address));
+	dep = (xfs_dir2_data_entry_t *)((char *)hdr +
+		xfs_dir2_dataptr_to_off(args->geo, be32_to_cpu(lep->address)));
 	needscan = needlog = 0;
 	oldbest = be16_to_cpu(bf[0].length);
-	ltp = xfs_dir2_leaf_tail_p(mp, leaf);
+	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 	bestsp = xfs_dir2_leaf_bests_p(ltp);
 	ASSERT(be16_to_cpu(bestsp[db]) == oldbest);
 	/*
 	 * Mark the former data entry unused.
 	 */
-	xfs_dir2_data_make_free(tp, dp, dbp,
+	xfs_dir2_data_make_free(args, dbp,
 		(xfs_dir2_data_aoff_t)((char *)dep - (char *)hdr),
 		dp->d_ops->data_entsize(dep->namelen), &needlog, &needscan);
 	/*
@@ -1399,10 +1405,10 @@ xfs_dir2_leaf_removename(
 	 */
 	leafhdr.stale++;
 	dp->d_ops->leaf_hdr_to_disk(leaf, &leafhdr);
-	xfs_dir3_leaf_log_header(tp, dp, lbp);
+	xfs_dir3_leaf_log_header(args, lbp);
 
 	lep->address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
-	xfs_dir3_leaf_log_ents(tp, dp, lbp, index, index);
+	xfs_dir3_leaf_log_ents(args, lbp, index, index);
 
 	/*
 	 * Scan the freespace in the data block again if necessary,
@@ -1411,22 +1417,22 @@ xfs_dir2_leaf_removename(
 	if (needscan)
 		xfs_dir2_data_freescan(dp, hdr, &needlog);
 	if (needlog)
-		xfs_dir2_data_log_header(tp, dp, dbp);
+		xfs_dir2_data_log_header(args, dbp);
 	/*
 	 * If the longest freespace in the data block has changed,
 	 * put the new value in the bests table and log that.
 	 */
 	if (be16_to_cpu(bf[0].length) != oldbest) {
 		bestsp[db] = bf[0].length;
-		xfs_dir3_leaf_log_bests(tp, lbp, db, db);
+		xfs_dir3_leaf_log_bests(args, lbp, db, db);
 	}
 	xfs_dir3_data_check(dp, dbp);
 	/*
 	 * If the data block is now empty then get rid of the data block.
 	 */
 	if (be16_to_cpu(bf[0].length) ==
-			mp->m_dirblksize - dp->d_ops->data_entry_offset) {
-		ASSERT(db != mp->m_dirdatablk);
+			args->geo->blksize - dp->d_ops->data_entry_offset) {
+		ASSERT(db != args->geo->datablk);
 		if ((error = xfs_dir2_shrink_inode(args, db, dbp))) {
 			/*
 			 * Nope, can't get rid of it because it caused
@@ -1459,15 +1465,16 @@ xfs_dir2_leaf_removename(
 			memmove(&bestsp[db - i], bestsp,
 				(be32_to_cpu(ltp->bestcount) - (db - i)) * sizeof(*bestsp));
 			be32_add_cpu(&ltp->bestcount, -(db - i));
-			xfs_dir3_leaf_log_tail(tp, lbp);
-			xfs_dir3_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
+			xfs_dir3_leaf_log_tail(args, lbp);
+			xfs_dir3_leaf_log_bests(args, lbp, 0,
+						be32_to_cpu(ltp->bestcount) - 1);
 		} else
 			bestsp[db] = cpu_to_be16(NULLDATAOFF);
 	}
 	/*
 	 * If the data block was not the first one, drop it.
 	 */
-	else if (db != mp->m_dirdatablk)
+	else if (db != args->geo->datablk)
 		dbp = NULL;
 
 	xfs_dir3_leaf_check(dp, lbp);
@@ -1515,7 +1522,7 @@ xfs_dir2_leaf_replace(
 	 */
 	dep = (xfs_dir2_data_entry_t *)
 	      ((char *)dbp->b_addr +
-	       xfs_dir2_dataptr_to_off(dp->i_mount, be32_to_cpu(lep->address)));
+	       xfs_dir2_dataptr_to_off(args->geo, be32_to_cpu(lep->address)));
 	ASSERT(args->inumber != be64_to_cpu(dep->inumber));
 	/*
 	 * Put the new inode number in, log it.
@@ -1523,7 +1530,7 @@ xfs_dir2_leaf_replace(
 	dep->inumber = cpu_to_be64(args->inumber);
 	dp->d_ops->data_put_ftype(dep, args->filetype);
 	tp = args->trans;
-	xfs_dir2_data_log_entry(tp, dp, dbp, dep);
+	xfs_dir2_data_log_entry(args, dbp, dep);
 	xfs_dir3_leaf_check(dp, lbp);
 	xfs_trans_brelse(tp, lbp);
 	return 0;
@@ -1609,12 +1616,13 @@ xfs_dir2_leaf_trim_data(
 	/*
 	 * Read the offending data block.  We need its buffer.
 	 */
-	error = xfs_dir3_data_read(tp, dp, xfs_dir2_db_to_da(mp, db), -1, &dbp);
+	error = xfs_dir3_data_read(tp, dp, xfs_dir2_db_to_da(args->geo, db),
+				   -1, &dbp);
 	if (error)
 		return error;
 
 	leaf = lbp->b_addr;
-	ltp = xfs_dir2_leaf_tail_p(mp, leaf);
+	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 
 #ifdef DEBUG
 {
@@ -1624,7 +1632,7 @@ xfs_dir2_leaf_trim_data(
 	ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
 	       hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC));
 	ASSERT(be16_to_cpu(bf[0].length) ==
-	       mp->m_dirblksize - dp->d_ops->data_entry_offset);
+	       args->geo->blksize - dp->d_ops->data_entry_offset);
 	ASSERT(db == be32_to_cpu(ltp->bestcount) - 1);
 }
 #endif
@@ -1643,8 +1651,8 @@ xfs_dir2_leaf_trim_data(
 	bestsp = xfs_dir2_leaf_bests_p(ltp);
 	be32_add_cpu(&ltp->bestcount, -1);
 	memmove(&bestsp[1], &bestsp[0], be32_to_cpu(ltp->bestcount) * sizeof(*bestsp));
-	xfs_dir3_leaf_log_tail(tp, lbp);
-	xfs_dir3_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
+	xfs_dir3_leaf_log_tail(args, lbp);
+	xfs_dir3_leaf_log_bests(args, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
 	return 0;
 }
 
@@ -1708,22 +1716,22 @@ xfs_dir2_node_to_leaf(
 	/*
 	 * Get the last offset in the file.
 	 */
-	if ((error = xfs_bmap_last_offset(tp, dp, &fo, XFS_DATA_FORK))) {
+	if ((error = xfs_bmap_last_offset(dp, &fo, XFS_DATA_FORK))) {
 		return error;
 	}
-	fo -= mp->m_dirblkfsbs;
+	fo -= args->geo->fsbcount;
 	/*
 	 * If there are freespace blocks other than the first one,
 	 * take this opportunity to remove trailing empty freespace blocks
 	 * that may have been left behind during no-space-reservation
 	 * operations.
 	 */
-	while (fo > mp->m_dirfreeblk) {
+	while (fo > args->geo->freeblk) {
 		if ((error = xfs_dir2_node_trim_free(args, fo, &rval))) {
 			return error;
 		}
 		if (rval)
-			fo -= mp->m_dirblkfsbs;
+			fo -= args->geo->fsbcount;
 		else
 			return 0;
 	}
@@ -1736,7 +1744,7 @@ xfs_dir2_node_to_leaf(
 	/*
 	 * If it's not the single leaf block, give up.
 	 */
-	if (XFS_FSB_TO_B(mp, fo) > XFS_DIR2_LEAF_OFFSET + mp->m_dirblksize)
+	if (XFS_FSB_TO_B(mp, fo) > XFS_DIR2_LEAF_OFFSET + args->geo->blksize)
 		return 0;
 	lbp = state->path.blk[0].bp;
 	leaf = lbp->b_addr;
@@ -1748,7 +1756,7 @@ xfs_dir2_node_to_leaf(
 	/*
 	 * Read the freespace block.
 	 */
-	error = xfs_dir2_free_read(tp, dp,  mp->m_dirfreeblk, &fbp);
+	error = xfs_dir2_free_read(tp, dp,  args->geo->freeblk, &fbp);
 	if (error)
 		return error;
 	free = fbp->b_addr;
@@ -1760,7 +1768,7 @@ xfs_dir2_node_to_leaf(
 	 * Now see if the leafn and free data will fit in a leaf1.
 	 * If not, release the buffer and give up.
 	 */
-	if (xfs_dir3_leaf_size(&leafhdr, freehdr.nvalid) > mp->m_dirblksize) {
+	if (xfs_dir3_leaf_size(&leafhdr, freehdr.nvalid) > args->geo->blksize) {
 		xfs_trans_brelse(tp, fbp);
 		return 0;
 	}
@@ -1780,7 +1788,7 @@ xfs_dir2_node_to_leaf(
 	/*
 	 * Set up the leaf tail from the freespace block.
 	 */
-	ltp = xfs_dir2_leaf_tail_p(mp, leaf);
+	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 	ltp->bestcount = cpu_to_be32(freehdr.nvalid);
 
 	/*
@@ -1790,15 +1798,17 @@ xfs_dir2_node_to_leaf(
 		freehdr.nvalid * sizeof(xfs_dir2_data_off_t));
 
 	dp->d_ops->leaf_hdr_to_disk(leaf, &leafhdr);
-	xfs_dir3_leaf_log_header(tp, dp, lbp);
-	xfs_dir3_leaf_log_bests(tp, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
-	xfs_dir3_leaf_log_tail(tp, lbp);
+	xfs_dir3_leaf_log_header(args, lbp);
+	xfs_dir3_leaf_log_bests(args, lbp, 0, be32_to_cpu(ltp->bestcount) - 1);
+	xfs_dir3_leaf_log_tail(args, lbp);
 	xfs_dir3_leaf_check(dp, lbp);
 
 	/*
 	 * Get rid of the freespace block.
 	 */
-	error = xfs_dir2_shrink_inode(args, XFS_DIR2_FREE_FIRSTDB(mp), fbp);
+	error = xfs_dir2_shrink_inode(args,
+			xfs_dir2_byte_to_db(args->geo, XFS_DIR2_FREE_OFFSET),
+			fbp);
 	if (error) {
 		/*
 		 * This can't fail here because it can only happen when

fs/xfs/xfs_dir2_node.c

@@ -195,17 +195,18 @@ xfs_dir2_free_try_read(
 
 static int
 xfs_dir3_free_get_buf(
-	struct xfs_trans	*tp,
-	struct xfs_inode	*dp,
+	xfs_da_args_t		*args,
 	xfs_dir2_db_t		fbno,
 	struct xfs_buf		**bpp)
 {
+	struct xfs_trans	*tp = args->trans;
+	struct xfs_inode	*dp = args->dp;
 	struct xfs_mount	*mp = dp->i_mount;
 	struct xfs_buf		*bp;
 	int			error;
 	struct xfs_dir3_icfree_hdr hdr;
 
-	error = xfs_da_get_buf(tp, dp, xfs_dir2_db_to_da(mp, fbno),
+	error = xfs_da_get_buf(tp, dp, xfs_dir2_db_to_da(args->geo, fbno),
 				   -1, &bp, XFS_DATA_FORK);
 	if (error)
 		return error;
@@ -240,8 +241,7 @@ xfs_dir3_free_get_buf(
  */
 STATIC void
 xfs_dir2_free_log_bests(
-	struct xfs_trans	*tp,
-	struct xfs_inode	*dp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp,
 	int			first,		/* first entry to log */
 	int			last)		/* last entry to log */
@@ -250,10 +250,10 @@ xfs_dir2_free_log_bests(
 	__be16			*bests;
 
 	free = bp->b_addr;
-	bests = dp->d_ops->free_bests_p(free);
+	bests = args->dp->d_ops->free_bests_p(free);
 	ASSERT(free->hdr.magic == cpu_to_be32(XFS_DIR2_FREE_MAGIC) ||
 	       free->hdr.magic == cpu_to_be32(XFS_DIR3_FREE_MAGIC));
-	xfs_trans_log_buf(tp, bp,
+	xfs_trans_log_buf(args->trans, bp,
 		(uint)((char *)&bests[first] - (char *)free),
 		(uint)((char *)&bests[last] - (char *)free +
 		       sizeof(bests[0]) - 1));
@@ -264,8 +264,7 @@ xfs_dir2_free_log_bests(
  */
 static void
 xfs_dir2_free_log_header(
-	struct xfs_trans	*tp,
-	struct xfs_inode	*dp,
+	struct xfs_da_args	*args,
 	struct xfs_buf		*bp)
 {
 #ifdef DEBUG
@@ -275,7 +274,8 @@ xfs_dir2_free_log_header(
 	ASSERT(free->hdr.magic == cpu_to_be32(XFS_DIR2_FREE_MAGIC) ||
 	       free->hdr.magic == cpu_to_be32(XFS_DIR3_FREE_MAGIC));
 #endif
-	xfs_trans_log_buf(tp, bp, 0, dp->d_ops->free_hdr_size - 1);
+	xfs_trans_log_buf(args->trans, bp, 0,
+			  args->dp->d_ops->free_hdr_size - 1);
 }
 
 /*
@@ -315,20 +315,20 @@ xfs_dir2_leaf_to_node(
 	if ((error = xfs_dir2_grow_inode(args, XFS_DIR2_FREE_SPACE, &fdb))) {
 		return error;
 	}
-	ASSERT(fdb == XFS_DIR2_FREE_FIRSTDB(mp));
+	ASSERT(fdb == xfs_dir2_byte_to_db(args->geo, XFS_DIR2_FREE_OFFSET));
 	/*
 	 * Get the buffer for the new freespace block.
 	 */
-	error = xfs_dir3_free_get_buf(tp, dp, fdb, &fbp);
+	error = xfs_dir3_free_get_buf(args, fdb, &fbp);
 	if (error)
 		return error;
 
 	free = fbp->b_addr;
 	dp->d_ops->free_hdr_from_disk(&freehdr, free);
 	leaf = lbp->b_addr;
-	ltp = xfs_dir2_leaf_tail_p(mp, leaf);
+	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 	ASSERT(be32_to_cpu(ltp->bestcount) <=
-				(uint)dp->i_d.di_size / mp->m_dirblksize);
+				(uint)dp->i_d.di_size / args->geo->blksize);
 
 	/*
 	 * Copy freespace entries from the leaf block to the new block.
@@ -349,8 +349,8 @@ xfs_dir2_leaf_to_node(
 	freehdr.nvalid = be32_to_cpu(ltp->bestcount);
 
 	dp->d_ops->free_hdr_to_disk(fbp->b_addr, &freehdr);
-	xfs_dir2_free_log_bests(tp, dp, fbp, 0, freehdr.nvalid - 1);
-	xfs_dir2_free_log_header(tp, dp, fbp);
+	xfs_dir2_free_log_bests(args, fbp, 0, freehdr.nvalid - 1);
+	xfs_dir2_free_log_header(args, fbp);
 
 	/*
 	 * Converting the leaf to a leafnode is just a matter of changing the
@@ -364,7 +364,7 @@ xfs_dir2_leaf_to_node(
 		leaf->hdr.info.magic = cpu_to_be16(XFS_DIR3_LEAFN_MAGIC);
 	lbp->b_ops = &xfs_dir3_leafn_buf_ops;
 	xfs_trans_buf_set_type(tp, lbp, XFS_BLFT_DIR_LEAFN_BUF);
-	xfs_dir3_leaf_log_header(tp, dp, lbp);
+	xfs_dir3_leaf_log_header(args, lbp);
 	xfs_dir3_leaf_check(dp, lbp);
 	return 0;
 }
@@ -415,7 +415,7 @@ xfs_dir2_leafn_add(
 	 * a compact.
 	 */
 
-	if (leafhdr.count == dp->d_ops->leaf_max_ents(mp)) {
+	if (leafhdr.count == dp->d_ops->leaf_max_ents(args->geo)) {
 		if (!leafhdr.stale)
 			return XFS_ERROR(ENOSPC);
 		compact = leafhdr.stale > 1;
@@ -450,12 +450,12 @@ xfs_dir2_leafn_add(
 				       highstale, &lfloglow, &lfloghigh);
 
 	lep->hashval = cpu_to_be32(args->hashval);
-	lep->address = cpu_to_be32(xfs_dir2_db_off_to_dataptr(mp,
+	lep->address = cpu_to_be32(xfs_dir2_db_off_to_dataptr(args->geo,
 				args->blkno, args->index));
 
 	dp->d_ops->leaf_hdr_to_disk(leaf, &leafhdr);
-	xfs_dir3_leaf_log_header(tp, dp, bp);
-	xfs_dir3_leaf_log_ents(tp, dp, bp, lfloglow, lfloghigh);
+	xfs_dir3_leaf_log_header(args, bp);
+	xfs_dir3_leaf_log_ents(args, bp, lfloglow, lfloghigh);
 	xfs_dir3_leaf_check(dp, bp);
 	return 0;
 }
@@ -471,7 +471,8 @@ xfs_dir2_free_hdr_check(
 
 	dp->d_ops->free_hdr_from_disk(&hdr, bp->b_addr);
 
-	ASSERT((hdr.firstdb % dp->d_ops->free_max_bests(dp->i_mount)) == 0);
+	ASSERT((hdr.firstdb %
+		dp->d_ops->free_max_bests(dp->i_mount->m_dir_geo)) == 0);
 	ASSERT(hdr.firstdb <= db);
 	ASSERT(db < hdr.firstdb + hdr.nvalid);
 }
@@ -576,7 +577,8 @@ xfs_dir2_leafn_lookup_for_addname(
 		/*
 		 * Pull the data block number from the entry.
 		 */
-		newdb = xfs_dir2_dataptr_to_db(mp, be32_to_cpu(lep->address));
+		newdb = xfs_dir2_dataptr_to_db(args->geo,
+					       be32_to_cpu(lep->address));
 		/*
 		 * For addname, we're looking for a place to put the new entry.
 		 * We want to use a data block with an entry of equal
@@ -593,7 +595,7 @@ xfs_dir2_leafn_lookup_for_addname(
 			 * Convert the data block to the free block
 			 * holding its freespace information.
 			 */
-			newfdb = dp->d_ops->db_to_fdb(mp, newdb);
+			newfdb = dp->d_ops->db_to_fdb(args->geo, newdb);
 			/*
 			 * If it's not the one we have in hand, read it in.
 			 */
@@ -605,7 +607,8 @@ xfs_dir2_leafn_lookup_for_addname(
 					xfs_trans_brelse(tp, curbp);
 
 				error = xfs_dir2_free_read(tp, dp,
-						xfs_dir2_db_to_da(mp, newfdb),
+						xfs_dir2_db_to_da(args->geo,
+								  newfdb),
 						&curbp);
 				if (error)
 					return error;
@@ -616,7 +619,7 @@ xfs_dir2_leafn_lookup_for_addname(
 			/*
 			 * Get the index for our entry.
 			 */
-			fi = dp->d_ops->db_to_fdindex(mp, curdb);
+			fi = dp->d_ops->db_to_fdindex(args->geo, curdb);
 			/*
 			 * If it has room, return it.
 			 */
@@ -721,7 +724,8 @@ xfs_dir2_leafn_lookup_for_entry(
 		/*
 		 * Pull the data block number from the entry.
 		 */
-		newdb = xfs_dir2_dataptr_to_db(mp, be32_to_cpu(lep->address));
+		newdb = xfs_dir2_dataptr_to_db(args->geo,
+					       be32_to_cpu(lep->address));
 		/*
 		 * Not adding a new entry, so we really want to find
 		 * the name given to us.
@@ -746,7 +750,8 @@ xfs_dir2_leafn_lookup_for_entry(
 				curbp = state->extrablk.bp;
 			} else {
 				error = xfs_dir3_data_read(tp, dp,
-						xfs_dir2_db_to_da(mp, newdb),
+						xfs_dir2_db_to_da(args->geo,
+								  newdb),
 						-1, &curbp);
 				if (error)
 					return error;
@@ -758,7 +763,8 @@ xfs_dir2_leafn_lookup_for_entry(
 		 * Point to the data entry.
 		 */
 		dep = (xfs_dir2_data_entry_t *)((char *)curbp->b_addr +
-			xfs_dir2_dataptr_to_off(mp, be32_to_cpu(lep->address)));
+			xfs_dir2_dataptr_to_off(args->geo,
+						be32_to_cpu(lep->address)));
 		/*
 		 * Compare the entry and if it's an exact match, return
 		 * EEXIST immediately. If it's the first case-insensitive
@@ -844,7 +850,6 @@ xfs_dir3_leafn_moveents(
 	int				start_d,/* destination leaf index */
 	int				count)	/* count of leaves to copy */
 {
-	struct xfs_trans		*tp = args->trans;
 	int				stale;	/* count stale leaves copied */
 
 	trace_xfs_dir2_leafn_moveents(args, start_s, start_d, count);
@@ -863,7 +868,7 @@ xfs_dir3_leafn_moveents(
 	if (start_d < dhdr->count) {
 		memmove(&dents[start_d + count], &dents[start_d],
 			(dhdr->count - start_d) * sizeof(xfs_dir2_leaf_entry_t));
-		xfs_dir3_leaf_log_ents(tp, args->dp, bp_d, start_d + count,
+		xfs_dir3_leaf_log_ents(args, bp_d, start_d + count,
 				       count + dhdr->count - 1);
 	}
 	/*
@@ -885,8 +890,7 @@ xfs_dir3_leafn_moveents(
 	 */
 	memcpy(&dents[start_d], &sents[start_s],
 		count * sizeof(xfs_dir2_leaf_entry_t));
-	xfs_dir3_leaf_log_ents(tp, args->dp, bp_d,
-			       start_d, start_d + count - 1);
+	xfs_dir3_leaf_log_ents(args, bp_d, start_d, start_d + count - 1);
 
 	/*
 	 * If there are source entries after the ones we copied,
@@ -895,8 +899,7 @@ xfs_dir3_leafn_moveents(
 	if (start_s + count < shdr->count) {
 		memmove(&sents[start_s], &sents[start_s + count],
 			count * sizeof(xfs_dir2_leaf_entry_t));
-		xfs_dir3_leaf_log_ents(tp, args->dp, bp_s,
-				       start_s, start_s + count - 1);
+		xfs_dir3_leaf_log_ents(args, bp_s, start_s, start_s + count - 1);
 	}
 
 	/*
@@ -1032,8 +1035,8 @@ xfs_dir2_leafn_rebalance(
 	/* log the changes made when moving the entries */
 	dp->d_ops->leaf_hdr_to_disk(leaf1, &hdr1);
 	dp->d_ops->leaf_hdr_to_disk(leaf2, &hdr2);
-	xfs_dir3_leaf_log_header(args->trans, dp, blk1->bp);
-	xfs_dir3_leaf_log_header(args->trans, dp, blk2->bp);
+	xfs_dir3_leaf_log_header(args, blk1->bp);
+	xfs_dir3_leaf_log_header(args, blk2->bp);
 
 	xfs_dir3_leaf_check(dp, blk1->bp);
 	xfs_dir3_leaf_check(dp, blk2->bp);
@@ -1076,7 +1079,6 @@ xfs_dir3_data_block_free(
 	struct xfs_buf		*fbp,
 	int			longest)
 {
-	struct xfs_trans	*tp = args->trans;
 	int			logfree = 0;
 	__be16			*bests;
 	struct xfs_dir3_icfree_hdr freehdr;
@@ -1090,7 +1092,7 @@ xfs_dir3_data_block_free(
 		 * value.
 		 */
 		bests[findex] = cpu_to_be16(longest);
-		xfs_dir2_free_log_bests(tp, dp, fbp, findex, findex);
+		xfs_dir2_free_log_bests(args, fbp, findex, findex);
 		return 0;
 	}
 
@@ -1118,7 +1120,7 @@ xfs_dir3_data_block_free(
 	}
 
 	dp->d_ops->free_hdr_to_disk(free, &freehdr);
-	xfs_dir2_free_log_header(tp, dp, fbp);
+	xfs_dir2_free_log_header(args, fbp);
 
 	/*
 	 * If there are no useful entries left in the block, get rid of the
@@ -1142,7 +1144,7 @@ xfs_dir3_data_block_free(
 
 	/* Log the free entry that changed, unless we got rid of it.  */
 	if (logfree)
-		xfs_dir2_free_log_bests(tp, dp, fbp, findex, findex);
+		xfs_dir2_free_log_bests(args, fbp, findex, findex);
 	return 0;
 }
 
@@ -1193,9 +1195,9 @@ xfs_dir2_leafn_remove(
 	/*
 	 * Extract the data block and offset from the entry.
 	 */
-	db = xfs_dir2_dataptr_to_db(mp, be32_to_cpu(lep->address));
+	db = xfs_dir2_dataptr_to_db(args->geo, be32_to_cpu(lep->address));
 	ASSERT(dblk->blkno == db);
-	off = xfs_dir2_dataptr_to_off(mp, be32_to_cpu(lep->address));
+	off = xfs_dir2_dataptr_to_off(args->geo, be32_to_cpu(lep->address));
 	ASSERT(dblk->index == off);
 
 	/*
@@ -1204,10 +1206,10 @@ xfs_dir2_leafn_remove(
 	 */
 	leafhdr.stale++;
 	dp->d_ops->leaf_hdr_to_disk(leaf, &leafhdr);
-	xfs_dir3_leaf_log_header(tp, dp, bp);
+	xfs_dir3_leaf_log_header(args, bp);
 
 	lep->address = cpu_to_be32(XFS_DIR2_NULL_DATAPTR);
-	xfs_dir3_leaf_log_ents(tp, dp, bp, index, index);
+	xfs_dir3_leaf_log_ents(args, bp, index, index);
 
 	/*
 	 * Make the data entry free.  Keep track of the longest freespace
@@ -1219,7 +1221,7 @@ xfs_dir2_leafn_remove(
 	bf = dp->d_ops->data_bestfree_p(hdr);
 	longest = be16_to_cpu(bf[0].length);
 	needlog = needscan = 0;
-	xfs_dir2_data_make_free(tp, dp, dbp, off,
+	xfs_dir2_data_make_free(args, dbp, off,
 		dp->d_ops->data_entsize(dep->namelen), &needlog, &needscan);
 	/*
 	 * Rescan the data block freespaces for bestfree.
@@ -1228,7 +1230,7 @@ xfs_dir2_leafn_remove(
 	if (needscan)
 		xfs_dir2_data_freescan(dp, hdr, &needlog);
 	if (needlog)
-		xfs_dir2_data_log_header(tp, dp, dbp);
+		xfs_dir2_data_log_header(args, dbp);
 	xfs_dir3_data_check(dp, dbp);
 	/*
 	 * If the longest data block freespace changes, need to update
@@ -1245,8 +1247,9 @@ xfs_dir2_leafn_remove(
 		 * Convert the data block number to a free block,
 		 * read in the free block.
 		 */
-		fdb = dp->d_ops->db_to_fdb(mp, db);
-		error = xfs_dir2_free_read(tp, dp, xfs_dir2_db_to_da(mp, fdb),
+		fdb = dp->d_ops->db_to_fdb(args->geo, db);
+		error = xfs_dir2_free_read(tp, dp,
+					   xfs_dir2_db_to_da(args->geo, fdb),
 					   &fbp);
 		if (error)
 			return error;
@@ -1255,20 +1258,21 @@ xfs_dir2_leafn_remove(
 	{
 		struct xfs_dir3_icfree_hdr freehdr;
 		dp->d_ops->free_hdr_from_disk(&freehdr, free);
-		ASSERT(freehdr.firstdb == dp->d_ops->free_max_bests(mp) *
-					  (fdb - XFS_DIR2_FREE_FIRSTDB(mp)));
+		ASSERT(freehdr.firstdb == dp->d_ops->free_max_bests(args->geo) *
+			(fdb - xfs_dir2_byte_to_db(args->geo,
+						   XFS_DIR2_FREE_OFFSET)));
 	}
 #endif
 		/*
 		 * Calculate which entry we need to fix.
 		 */
-		findex = dp->d_ops->db_to_fdindex(mp, db);
+		findex = dp->d_ops->db_to_fdindex(args->geo, db);
 		longest = be16_to_cpu(bf[0].length);
 		/*
 		 * If the data block is now empty we can get rid of it
 		 * (usually).
 		 */
-		if (longest == mp->m_dirblksize -
+		if (longest == args->geo->blksize -
 			       dp->d_ops->data_entry_offset) {
 			/*
 			 * Try to punch out the data block.
@@ -1303,7 +1307,7 @@ xfs_dir2_leafn_remove(
 	 */
 	*rval = (dp->d_ops->leaf_hdr_size +
 		 (uint)sizeof(ents[0]) * (leafhdr.count - leafhdr.stale)) <
-		mp->m_dir_magicpct;
+		args->geo->magicpct;
 	return 0;
 }
 
@@ -1336,7 +1340,7 @@ xfs_dir2_leafn_split(
 	/*
 	 * Initialize the new leaf block.
 	 */
-	error = xfs_dir3_leaf_get_buf(args, xfs_dir2_da_to_db(mp, blkno),
+	error = xfs_dir3_leaf_get_buf(args, xfs_dir2_da_to_db(args->geo, blkno),
 				      &newblk->bp, XFS_DIR2_LEAFN_MAGIC);
 	if (error)
 		return error;
@@ -1410,7 +1414,7 @@ xfs_dir2_leafn_toosmall(
 
 	count = leafhdr.count - leafhdr.stale;
 	bytes = dp->d_ops->leaf_hdr_size + count * sizeof(ents[0]);
-	if (bytes > (state->blocksize >> 1)) {
+	if (bytes > (state->args->geo->blksize >> 1)) {
 		/*
 		 * Blk over 50%, don't try to join.
 		 */
@@ -1463,7 +1467,8 @@ xfs_dir2_leafn_toosmall(
 		 * Count bytes in the two blocks combined.
 		 */
 		count = leafhdr.count - leafhdr.stale;
-		bytes = state->blocksize - (state->blocksize >> 2);
+		bytes = state->args->geo->blksize -
+			(state->args->geo->blksize >> 2);
 
 		leaf = bp->b_addr;
 		dp->d_ops->leaf_hdr_from_disk(&hdr2, leaf);
@@ -1560,8 +1565,8 @@ xfs_dir2_leafn_unbalance(
 	/* log the changes made when moving the entries */
 	dp->d_ops->leaf_hdr_to_disk(save_leaf, &savehdr);
 	dp->d_ops->leaf_hdr_to_disk(drop_leaf, &drophdr);
-	xfs_dir3_leaf_log_header(args->trans, dp, save_blk->bp);
-	xfs_dir3_leaf_log_header(args->trans, dp, drop_blk->bp);
+	xfs_dir3_leaf_log_header(args, save_blk->bp);
+	xfs_dir3_leaf_log_header(args, drop_blk->bp);
 
 	xfs_dir3_leaf_check(dp, save_blk->bp);
 	xfs_dir3_leaf_check(dp, drop_blk->bp);
@@ -1587,8 +1592,6 @@ xfs_dir2_node_addname(
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = args->dp->i_mount;
-	state->blocksize = state->mp->m_dirblksize;
-	state->node_ents = state->mp->m_dir_node_ents;
 	/*
 	 * Look up the name.  We're not supposed to find it, but
 	 * this gives us the insertion point.
@@ -1727,9 +1730,9 @@ xfs_dir2_node_addname_int(
 	if (dbno == -1) {
 		xfs_fileoff_t	fo;		/* freespace block number */
 
-		if ((error = xfs_bmap_last_offset(tp, dp, &fo, XFS_DATA_FORK)))
+		if ((error = xfs_bmap_last_offset(dp, &fo, XFS_DATA_FORK)))
 			return error;
-		lastfbno = xfs_dir2_da_to_db(mp, (xfs_dablk_t)fo);
+		lastfbno = xfs_dir2_da_to_db(args->geo, (xfs_dablk_t)fo);
 		fbno = ifbno;
 	}
 	/*
@@ -1747,7 +1750,8 @@ xfs_dir2_node_addname_int(
 			 * us a freespace block to start with.
 			 */
 			if (++fbno == 0)
-				fbno = XFS_DIR2_FREE_FIRSTDB(mp);
+				fbno = xfs_dir2_byte_to_db(args->geo,
+							XFS_DIR2_FREE_OFFSET);
 			/*
 			 * If it's ifbno we already looked at it.
 			 */
@@ -1765,8 +1769,8 @@ xfs_dir2_node_addname_int(
 			 * to avoid it.
 			 */
 			error = xfs_dir2_free_try_read(tp, dp,
-						xfs_dir2_db_to_da(mp, fbno),
-						&fbp);
+					xfs_dir2_db_to_da(args->geo, fbno),
+					&fbp);
 			if (error)
 				return error;
 			if (!fbp)
@@ -1834,10 +1838,10 @@ xfs_dir2_node_addname_int(
 		 * Get the freespace block corresponding to the data block
 		 * that was just allocated.
 		 */
-		fbno = dp->d_ops->db_to_fdb(mp, dbno);
+		fbno = dp->d_ops->db_to_fdb(args->geo, dbno);
 		error = xfs_dir2_free_try_read(tp, dp,
-					       xfs_dir2_db_to_da(mp, fbno),
-					       &fbp);
+				       xfs_dir2_db_to_da(args->geo, fbno),
+				       &fbp);
 		if (error)
 			return error;
 
@@ -1851,12 +1855,13 @@ xfs_dir2_node_addname_int(
 			if (error)
 				return error;
 
-			if (unlikely(dp->d_ops->db_to_fdb(mp, dbno) != fbno)) {
+			if (dp->d_ops->db_to_fdb(args->geo, dbno) != fbno) {
 				xfs_alert(mp,
 			"%s: dir ino %llu needed freesp block %lld for\n"
 			"  data block %lld, got %lld ifbno %llu lastfbno %d",
 					__func__, (unsigned long long)dp->i_ino,
-					(long long)dp->d_ops->db_to_fdb(mp, dbno),
+					(long long)dp->d_ops->db_to_fdb(
+								args->geo, dbno),
 					(long long)dbno, (long long)fbno,
 					(unsigned long long)ifbno, lastfbno);
 				if (fblk) {
@@ -1877,7 +1882,7 @@ xfs_dir2_node_addname_int(
 			/*
 			 * Get a buffer for the new block.
 			 */
-			error = xfs_dir3_free_get_buf(tp, dp, fbno, &fbp);
+			error = xfs_dir3_free_get_buf(args, fbno, &fbp);
 			if (error)
 				return error;
 			free = fbp->b_addr;
@@ -1887,8 +1892,10 @@ xfs_dir2_node_addname_int(
 			/*
 			 * Remember the first slot as our empty slot.
 			 */
-			freehdr.firstdb = (fbno - XFS_DIR2_FREE_FIRSTDB(mp)) *
-					dp->d_ops->free_max_bests(mp);
+			freehdr.firstdb =
+				(fbno - xfs_dir2_byte_to_db(args->geo,
+							XFS_DIR2_FREE_OFFSET)) *
+					dp->d_ops->free_max_bests(args->geo);
 		} else {
 			free = fbp->b_addr;
 			bests = dp->d_ops->free_bests_p(free);
@@ -1898,13 +1905,13 @@ xfs_dir2_node_addname_int(
 		/*
 		 * Set the freespace block index from the data block number.
 		 */
-		findex = dp->d_ops->db_to_fdindex(mp, dbno);
+		findex = dp->d_ops->db_to_fdindex(args->geo, dbno);
 		/*
 		 * If it's after the end of the current entries in the
 		 * freespace block, extend that table.
 		 */
 		if (findex >= freehdr.nvalid) {
-			ASSERT(findex < dp->d_ops->free_max_bests(mp));
+			ASSERT(findex < dp->d_ops->free_max_bests(args->geo));
 			freehdr.nvalid = findex + 1;
 			/*
 			 * Tag new entry so nused will go up.
@@ -1918,7 +1925,7 @@ xfs_dir2_node_addname_int(
 		if (bests[findex] == cpu_to_be16(NULLDATAOFF)) {
 			freehdr.nused++;
 			dp->d_ops->free_hdr_to_disk(fbp->b_addr, &freehdr);
-			xfs_dir2_free_log_header(tp, dp, fbp);
+			xfs_dir2_free_log_header(args, fbp);
 		}
 		/*
 		 * Update the real value in the table.
@@ -1943,7 +1950,8 @@ xfs_dir2_node_addname_int(
 		/*
 		 * Read the data block in.
 		 */
-		error = xfs_dir3_data_read(tp, dp, xfs_dir2_db_to_da(mp, dbno),
+		error = xfs_dir3_data_read(tp, dp,
+					   xfs_dir2_db_to_da(args->geo, dbno),
 					   -1, &dbp);
 		if (error)
 			return error;
@@ -1961,7 +1969,7 @@ xfs_dir2_node_addname_int(
 	/*
 	 * Mark the first part of the unused space, inuse for us.
 	 */
-	xfs_dir2_data_use_free(tp, dp, dbp, dup,
+	xfs_dir2_data_use_free(args, dbp, dup,
 		(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr), length,
 		&needlog, &needscan);
 	/*
@@ -1974,7 +1982,7 @@ 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);
-	xfs_dir2_data_log_entry(tp, dp, dbp, dep);
+	xfs_dir2_data_log_entry(args, dbp, dep);
 	/*
 	 * Rescan the block for bestfree if needed.
 	 */
@@ -1984,7 +1992,7 @@ xfs_dir2_node_addname_int(
 	 * Log the data block header if needed.
 	 */
 	if (needlog)
-		xfs_dir2_data_log_header(tp, dp, dbp);
+		xfs_dir2_data_log_header(args, dbp);
 	/*
 	 * If the freespace entry is now wrong, update it.
 	 */
@@ -1997,7 +2005,7 @@ xfs_dir2_node_addname_int(
 	 * Log the freespace entry if needed.
 	 */
 	if (logfree)
-		xfs_dir2_free_log_bests(tp, dp, fbp, findex, findex);
+		xfs_dir2_free_log_bests(args, fbp, findex, findex);
 	/*
 	 * Return the data block and offset in args, then drop the data block.
 	 */
@@ -2028,8 +2036,6 @@ xfs_dir2_node_lookup(
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = args->dp->i_mount;
-	state->blocksize = state->mp->m_dirblksize;
-	state->node_ents = state->mp->m_dir_node_ents;
 	/*
 	 * Fill in the path to the entry in the cursor.
 	 */
@@ -2083,8 +2089,6 @@ xfs_dir2_node_removename(
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = args->dp->i_mount;
-	state->blocksize = state->mp->m_dirblksize;
-	state->node_ents = state->mp->m_dir_node_ents;
 
 	/* Look up the entry we're deleting, set up the cursor. */
 	error = xfs_da3_node_lookup_int(state, &rval);
@@ -2153,8 +2157,6 @@ xfs_dir2_node_replace(
 	state = xfs_da_state_alloc();
 	state->args = args;
 	state->mp = args->dp->i_mount;
-	state->blocksize = state->mp->m_dirblksize;
-	state->node_ents = state->mp->m_dir_node_ents;
 	inum = args->inumber;
 	/*
 	 * Lookup the entry to change in the btree.
@@ -2186,15 +2188,15 @@ xfs_dir2_node_replace(
 		       hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC));
 		dep = (xfs_dir2_data_entry_t *)
 		      ((char *)hdr +
-		       xfs_dir2_dataptr_to_off(state->mp, be32_to_cpu(lep->address)));
+		       xfs_dir2_dataptr_to_off(args->geo,
+					       be32_to_cpu(lep->address)));
 		ASSERT(inum != be64_to_cpu(dep->inumber));
 		/*
 		 * Fill in the new inode number and log the entry.
 		 */
 		dep->inumber = cpu_to_be64(inum);
 		args->dp->d_ops->data_put_ftype(dep, args->filetype);
-		xfs_dir2_data_log_entry(args->trans, args->dp,
-					state->extrablk.bp, dep);
+		xfs_dir2_data_log_entry(args, state->extrablk.bp, dep);
 		rval = 0;
 	}
 	/*
@@ -2262,9 +2264,9 @@ xfs_dir2_node_trim_free(
 	/*
 	 * Blow the block away.
 	 */
-	if ((error =
-	    xfs_dir2_shrink_inode(args, xfs_dir2_da_to_db(mp, (xfs_dablk_t)fo),
-		    bp))) {
+	error = xfs_dir2_shrink_inode(args,
+			xfs_dir2_da_to_db(args->geo, (xfs_dablk_t)fo), bp);
+	if (error) {
 		/*
 		 * Can't fail with ENOSPC since that only happens with no
 		 * space reservation, when breaking up an extent into two

fs/xfs/xfs_dir2_priv.h

@@ -20,6 +20,140 @@
 
 struct dir_context;
 
+/*
+ * Directory offset/block conversion functions.
+ *
+ * DB blocks here are logical directory block numbers, not filesystem blocks.
+ */
+
+/*
+ * Convert dataptr to byte in file space
+ */
+static inline xfs_dir2_off_t
+xfs_dir2_dataptr_to_byte(xfs_dir2_dataptr_t dp)
+{
+	return (xfs_dir2_off_t)dp << XFS_DIR2_DATA_ALIGN_LOG;
+}
+
+/*
+ * Convert byte in file space to dataptr.  It had better be aligned.
+ */
+static inline xfs_dir2_dataptr_t
+xfs_dir2_byte_to_dataptr(xfs_dir2_off_t by)
+{
+	return (xfs_dir2_dataptr_t)(by >> XFS_DIR2_DATA_ALIGN_LOG);
+}
+
+/*
+ * Convert byte in space to (DB) block
+ */
+static inline xfs_dir2_db_t
+xfs_dir2_byte_to_db(struct xfs_da_geometry *geo, xfs_dir2_off_t by)
+{
+	return (xfs_dir2_db_t)(by >> geo->blklog);
+}
+
+/*
+ * Convert dataptr to a block number
+ */
+static inline xfs_dir2_db_t
+xfs_dir2_dataptr_to_db(struct xfs_da_geometry *geo, xfs_dir2_dataptr_t dp)
+{
+	return xfs_dir2_byte_to_db(geo, xfs_dir2_dataptr_to_byte(dp));
+}
+
+/*
+ * Convert byte in space to offset in a block
+ */
+static inline xfs_dir2_data_aoff_t
+xfs_dir2_byte_to_off(struct xfs_da_geometry *geo, xfs_dir2_off_t by)
+{
+	return (xfs_dir2_data_aoff_t)(by & (geo->blksize - 1));
+}
+
+/*
+ * Convert dataptr to a byte offset in a block
+ */
+static inline xfs_dir2_data_aoff_t
+xfs_dir2_dataptr_to_off(struct xfs_da_geometry *geo, xfs_dir2_dataptr_t dp)
+{
+	return xfs_dir2_byte_to_off(geo, xfs_dir2_dataptr_to_byte(dp));
+}
+
+/*
+ * Convert block and offset to byte in space
+ */
+static inline xfs_dir2_off_t
+xfs_dir2_db_off_to_byte(struct xfs_da_geometry *geo, xfs_dir2_db_t db,
+			xfs_dir2_data_aoff_t o)
+{
+	return ((xfs_dir2_off_t)db << geo->blklog) + o;
+}
+
+/*
+ * Convert block (DB) to block (dablk)
+ */
+static inline xfs_dablk_t
+xfs_dir2_db_to_da(struct xfs_da_geometry *geo, xfs_dir2_db_t db)
+{
+	return (xfs_dablk_t)(db << (geo->blklog - geo->fsblog));
+}
+
+/*
+ * Convert byte in space to (DA) block
+ */
+static inline xfs_dablk_t
+xfs_dir2_byte_to_da(struct xfs_da_geometry *geo, xfs_dir2_off_t by)
+{
+	return xfs_dir2_db_to_da(geo, xfs_dir2_byte_to_db(geo, by));
+}
+
+/*
+ * Convert block and offset to dataptr
+ */
+static inline xfs_dir2_dataptr_t
+xfs_dir2_db_off_to_dataptr(struct xfs_da_geometry *geo, xfs_dir2_db_t db,
+			   xfs_dir2_data_aoff_t o)
+{
+	return xfs_dir2_byte_to_dataptr(xfs_dir2_db_off_to_byte(geo, db, o));
+}
+
+/*
+ * Convert block (dablk) to block (DB)
+ */
+static inline xfs_dir2_db_t
+xfs_dir2_da_to_db(struct xfs_da_geometry *geo, xfs_dablk_t da)
+{
+	return (xfs_dir2_db_t)(da >> (geo->blklog - geo->fsblog));
+}
+
+/*
+ * Convert block (dablk) to byte offset in space
+ */
+static inline xfs_dir2_off_t
+xfs_dir2_da_to_byte(struct xfs_da_geometry *geo, xfs_dablk_t da)
+{
+	return xfs_dir2_db_off_to_byte(geo, xfs_dir2_da_to_db(geo, da), 0);
+}
+
+/*
+ * Directory tail pointer accessor functions. Based on block geometry.
+ */
+static inline struct xfs_dir2_block_tail *
+xfs_dir2_block_tail_p(struct xfs_da_geometry *geo, struct xfs_dir2_data_hdr *hdr)
+{
+	return ((struct xfs_dir2_block_tail *)
+		((char *)hdr + geo->blksize)) - 1;
+}
+
+static inline struct xfs_dir2_leaf_tail *
+xfs_dir2_leaf_tail_p(struct xfs_da_geometry *geo, struct xfs_dir2_leaf *lp)
+{
+	return (struct xfs_dir2_leaf_tail *)
+		((char *)lp + geo->blksize -
+		  sizeof(struct xfs_dir2_leaf_tail));
+}
+
 /* xfs_dir2.c */
 extern int xfs_dir_ino_validate(struct xfs_mount *mp, xfs_ino_t ino);
 extern int xfs_dir2_grow_inode(struct xfs_da_args *args, int space,
@@ -54,8 +188,8 @@ extern int xfs_dir2_leaf_to_block(struct xfs_da_args *args,
 extern int __xfs_dir3_data_check(struct xfs_inode *dp, struct xfs_buf *bp);
 extern int xfs_dir3_data_read(struct xfs_trans *tp, struct xfs_inode *dp,
 		xfs_dablk_t bno, xfs_daddr_t mapped_bno, struct xfs_buf **bpp);
-extern int xfs_dir3_data_readahead(struct xfs_trans *tp, struct xfs_inode *dp,
-		xfs_dablk_t bno, xfs_daddr_t mapped_bno);
+extern int xfs_dir3_data_readahead(struct xfs_inode *dp, xfs_dablk_t bno,
+		xfs_daddr_t mapped_bno);
 
 extern struct xfs_dir2_data_free *
 xfs_dir2_data_freeinsert(struct xfs_dir2_data_hdr *hdr,
@@ -77,9 +211,9 @@ extern void xfs_dir3_leaf_compact_x1(struct xfs_dir3_icleaf_hdr *leafhdr,
 		int *lowstalep, int *highstalep, int *lowlogp, int *highlogp);
 extern int xfs_dir3_leaf_get_buf(struct xfs_da_args *args, xfs_dir2_db_t bno,
 		struct xfs_buf **bpp, __uint16_t magic);
-extern void xfs_dir3_leaf_log_ents(struct xfs_trans *tp, struct xfs_inode *dp,
+extern void xfs_dir3_leaf_log_ents(struct xfs_da_args *args,
 		struct xfs_buf *bp, int first, int last);
-extern void xfs_dir3_leaf_log_header(struct xfs_trans *tp, struct xfs_inode *dp,
+extern void xfs_dir3_leaf_log_header(struct xfs_da_args *args,
 		struct xfs_buf *bp);
 extern int xfs_dir2_leaf_lookup(struct xfs_da_args *args);
 extern int xfs_dir2_leaf_removename(struct xfs_da_args *args);

fs/xfs/xfs_dir2_readdir.c

@@ -76,26 +76,25 @@ const unsigned char xfs_mode_to_ftype[S_IFMT >> S_SHIFT] = {
 
 STATIC int
 xfs_dir2_sf_getdents(
-	xfs_inode_t		*dp,		/* incore directory inode */
+	struct xfs_da_args	*args,
 	struct dir_context	*ctx)
 {
 	int			i;		/* shortform entry number */
-	xfs_mount_t		*mp;		/* filesystem mount point */
+	struct xfs_inode	*dp = args->dp;	/* incore directory inode */
 	xfs_dir2_dataptr_t	off;		/* current entry's offset */
 	xfs_dir2_sf_entry_t	*sfep;		/* shortform directory entry */
 	xfs_dir2_sf_hdr_t	*sfp;		/* shortform structure */
 	xfs_dir2_dataptr_t	dot_offset;
 	xfs_dir2_dataptr_t	dotdot_offset;
 	xfs_ino_t		ino;
-
-	mp = dp->i_mount;
+	struct xfs_da_geometry	*geo = args->geo;
 
 	ASSERT(dp->i_df.if_flags & XFS_IFINLINE);
 	/*
 	 * Give up if the directory is way too short.
 	 */
 	if (dp->i_d.di_size < offsetof(xfs_dir2_sf_hdr_t, parent)) {
-		ASSERT(XFS_FORCED_SHUTDOWN(mp));
+		ASSERT(XFS_FORCED_SHUTDOWN(dp->i_mount));
 		return XFS_ERROR(EIO);
 	}
 
@@ -109,18 +108,18 @@ xfs_dir2_sf_getdents(
 	/*
 	 * If the block number in the offset is out of range, we're done.
 	 */
-	if (xfs_dir2_dataptr_to_db(mp, ctx->pos) > mp->m_dirdatablk)
+	if (xfs_dir2_dataptr_to_db(geo, ctx->pos) > geo->datablk)
 		return 0;
 
 	/*
 	 * Precalculate offsets for . and .. as we will always need them.
 	 *
 	 * XXX(hch): the second argument is sometimes 0 and sometimes
-	 * mp->m_dirdatablk.
+	 * geo->datablk
 	 */
-	dot_offset = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
+	dot_offset = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
 						dp->d_ops->data_dot_offset);
-	dotdot_offset = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
+	dotdot_offset = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
 						dp->d_ops->data_dotdot_offset);
 
 	/*
@@ -149,7 +148,7 @@ xfs_dir2_sf_getdents(
 	for (i = 0; i < sfp->count; i++) {
 		__uint8_t filetype;
 
-		off = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
+		off = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
 				xfs_dir2_sf_get_offset(sfep));
 
 		if (ctx->pos > off) {
@@ -161,13 +160,13 @@ xfs_dir2_sf_getdents(
 		filetype = dp->d_ops->sf_get_ftype(sfep);
 		ctx->pos = off & 0x7fffffff;
 		if (!dir_emit(ctx, (char *)sfep->name, sfep->namelen, ino,
-			    xfs_dir3_get_dtype(mp, filetype)))
+			    xfs_dir3_get_dtype(dp->i_mount, filetype)))
 			return 0;
 		sfep = dp->d_ops->sf_nextentry(sfp, sfep);
 	}
 
-	ctx->pos = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk + 1, 0) &
-			0x7fffffff;
+	ctx->pos = xfs_dir2_db_off_to_dataptr(geo, geo->datablk + 1, 0) &
+								0x7fffffff;
 	return 0;
 }
 
@@ -176,9 +175,10 @@ xfs_dir2_sf_getdents(
  */
 STATIC int
 xfs_dir2_block_getdents(
-	xfs_inode_t		*dp,		/* incore inode */
+	struct xfs_da_args	*args,
 	struct dir_context	*ctx)
 {
+	struct xfs_inode	*dp = args->dp;	/* incore directory inode */
 	xfs_dir2_data_hdr_t	*hdr;		/* block header */
 	struct xfs_buf		*bp;		/* buffer for block */
 	xfs_dir2_block_tail_t	*btp;		/* block tail */
@@ -186,16 +186,15 @@ xfs_dir2_block_getdents(
 	xfs_dir2_data_unused_t	*dup;		/* block unused entry */
 	char			*endptr;	/* end of the data entries */
 	int			error;		/* error return value */
-	xfs_mount_t		*mp;		/* filesystem mount point */
 	char			*ptr;		/* current data entry */
 	int			wantoff;	/* starting block offset */
 	xfs_off_t		cook;
+	struct xfs_da_geometry	*geo = args->geo;
 
-	mp = dp->i_mount;
 	/*
 	 * If the block number in the offset is out of range, we're done.
 	 */
-	if (xfs_dir2_dataptr_to_db(mp, ctx->pos) > mp->m_dirdatablk)
+	if (xfs_dir2_dataptr_to_db(geo, ctx->pos) > geo->datablk)
 		return 0;
 
 	error = xfs_dir3_block_read(NULL, dp, &bp);
@@ -206,13 +205,13 @@ xfs_dir2_block_getdents(
 	 * Extract the byte offset we start at from the seek pointer.
 	 * We'll skip entries before this.
 	 */
-	wantoff = xfs_dir2_dataptr_to_off(mp, ctx->pos);
+	wantoff = xfs_dir2_dataptr_to_off(geo, ctx->pos);
 	hdr = bp->b_addr;
 	xfs_dir3_data_check(dp, bp);
 	/*
 	 * Set up values for the loop.
 	 */
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(geo, hdr);
 	ptr = (char *)dp->d_ops->data_entry_p(hdr);
 	endptr = (char *)xfs_dir2_block_leaf_p(btp);
 
@@ -244,7 +243,7 @@ xfs_dir2_block_getdents(
 		if ((char *)dep - (char *)hdr < wantoff)
 			continue;
 
-		cook = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk,
+		cook = xfs_dir2_db_off_to_dataptr(geo, geo->datablk,
 					    (char *)dep - (char *)hdr);
 
 		ctx->pos = cook & 0x7fffffff;
@@ -254,7 +253,7 @@ xfs_dir2_block_getdents(
 		 */
 		if (!dir_emit(ctx, (char *)dep->name, dep->namelen,
 			    be64_to_cpu(dep->inumber),
-			    xfs_dir3_get_dtype(mp, filetype))) {
+			    xfs_dir3_get_dtype(dp->i_mount, filetype))) {
 			xfs_trans_brelse(NULL, bp);
 			return 0;
 		}
@@ -264,8 +263,8 @@ xfs_dir2_block_getdents(
 	 * Reached the end of the block.
 	 * Set the offset to a non-existent block 1 and return.
 	 */
-	ctx->pos = xfs_dir2_db_off_to_dataptr(mp, mp->m_dirdatablk + 1, 0) &
-			0x7fffffff;
+	ctx->pos = xfs_dir2_db_off_to_dataptr(geo, geo->datablk + 1, 0) &
+								0x7fffffff;
 	xfs_trans_brelse(NULL, bp);
 	return 0;
 }
@@ -286,13 +285,13 @@ struct xfs_dir2_leaf_map_info {
 
 STATIC int
 xfs_dir2_leaf_readbuf(
-	struct xfs_inode	*dp,
+	struct xfs_da_args	*args,
 	size_t			bufsize,
 	struct xfs_dir2_leaf_map_info *mip,
 	xfs_dir2_off_t		*curoff,
 	struct xfs_buf		**bpp)
 {
-	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_inode	*dp = args->dp;
 	struct xfs_buf		*bp = *bpp;
 	struct xfs_bmbt_irec	*map = mip->map;
 	struct blk_plug		plug;
@@ -300,6 +299,7 @@ xfs_dir2_leaf_readbuf(
 	int			length;
 	int			i;
 	int			j;
+	struct xfs_da_geometry	*geo = args->geo;
 
 	/*
 	 * If we have a buffer, we need to release it and
@@ -309,12 +309,12 @@ xfs_dir2_leaf_readbuf(
 	if (bp) {
 		xfs_trans_brelse(NULL, bp);
 		bp = NULL;
-		mip->map_blocks -= mp->m_dirblkfsbs;
+		mip->map_blocks -= geo->fsbcount;
 		/*
 		 * Loop to get rid of the extents for the
 		 * directory block.
 		 */
-		for (i = mp->m_dirblkfsbs; i > 0; ) {
+		for (i = geo->fsbcount; i > 0; ) {
 			j = min_t(int, map->br_blockcount, i);
 			map->br_blockcount -= j;
 			map->br_startblock += j;
@@ -333,8 +333,7 @@ xfs_dir2_leaf_readbuf(
 	/*
 	 * Recalculate the readahead blocks wanted.
 	 */
-	mip->ra_want = howmany(bufsize + mp->m_dirblksize,
-			       mp->m_sb.sb_blocksize) - 1;
+	mip->ra_want = howmany(bufsize + geo->blksize, (1 << geo->fsblog)) - 1;
 	ASSERT(mip->ra_want >= 0);
 
 	/*
@@ -342,14 +341,14 @@ xfs_dir2_leaf_readbuf(
 	 * run out of data blocks, get some more mappings.
 	 */
 	if (1 + mip->ra_want > mip->map_blocks &&
-	    mip->map_off < xfs_dir2_byte_to_da(mp, XFS_DIR2_LEAF_OFFSET)) {
+	    mip->map_off < xfs_dir2_byte_to_da(geo, XFS_DIR2_LEAF_OFFSET)) {
 		/*
 		 * Get more bmaps, fill in after the ones
 		 * we already have in the table.
 		 */
 		mip->nmap = mip->map_size - mip->map_valid;
 		error = xfs_bmapi_read(dp, mip->map_off,
-				xfs_dir2_byte_to_da(mp, XFS_DIR2_LEAF_OFFSET) -
+				xfs_dir2_byte_to_da(geo, XFS_DIR2_LEAF_OFFSET) -
 								mip->map_off,
 				&map[mip->map_valid], &mip->nmap, 0);
 
@@ -370,7 +369,7 @@ xfs_dir2_leaf_readbuf(
 			i = mip->map_valid + mip->nmap - 1;
 			mip->map_off = map[i].br_startoff + map[i].br_blockcount;
 		} else
-			mip->map_off = xfs_dir2_byte_to_da(mp,
+			mip->map_off = xfs_dir2_byte_to_da(geo,
 							XFS_DIR2_LEAF_OFFSET);
 
 		/*
@@ -396,18 +395,18 @@ xfs_dir2_leaf_readbuf(
 	 * No valid mappings, so no more data blocks.
 	 */
 	if (!mip->map_valid) {
-		*curoff = xfs_dir2_da_to_byte(mp, mip->map_off);
+		*curoff = xfs_dir2_da_to_byte(geo, mip->map_off);
 		goto out;
 	}
 
 	/*
 	 * Read the directory block starting at the first mapping.
 	 */
-	mip->curdb = xfs_dir2_da_to_db(mp, map->br_startoff);
+	mip->curdb = xfs_dir2_da_to_db(geo, map->br_startoff);
 	error = xfs_dir3_data_read(NULL, dp, map->br_startoff,
-			map->br_blockcount >= mp->m_dirblkfsbs ?
-			    XFS_FSB_TO_DADDR(mp, map->br_startblock) : -1, &bp);
-
+			map->br_blockcount >= geo->fsbcount ?
+			    XFS_FSB_TO_DADDR(dp->i_mount, map->br_startblock) :
+			    -1, &bp);
 	/*
 	 * Should just skip over the data block instead of giving up.
 	 */
@@ -419,7 +418,7 @@ xfs_dir2_leaf_readbuf(
 	 * was previously ra.
 	 */
 	if (mip->ra_current)
-		mip->ra_current -= mp->m_dirblkfsbs;
+		mip->ra_current -= geo->fsbcount;
 
 	/*
 	 * Do we need more readahead?
@@ -427,16 +426,16 @@ xfs_dir2_leaf_readbuf(
 	blk_start_plug(&plug);
 	for (mip->ra_index = mip->ra_offset = i = 0;
 	     mip->ra_want > mip->ra_current && i < mip->map_blocks;
-	     i += mp->m_dirblkfsbs) {
+	     i += geo->fsbcount) {
 		ASSERT(mip->ra_index < mip->map_valid);
 		/*
 		 * Read-ahead a contiguous directory block.
 		 */
 		if (i > mip->ra_current &&
-		    map[mip->ra_index].br_blockcount >= mp->m_dirblkfsbs) {
-			xfs_dir3_data_readahead(NULL, dp,
+		    map[mip->ra_index].br_blockcount >= geo->fsbcount) {
+			xfs_dir3_data_readahead(dp,
 				map[mip->ra_index].br_startoff + mip->ra_offset,
-				XFS_FSB_TO_DADDR(mp,
+				XFS_FSB_TO_DADDR(dp->i_mount,
 					map[mip->ra_index].br_startblock +
 							mip->ra_offset));
 			mip->ra_current = i;
@@ -447,7 +446,7 @@ xfs_dir2_leaf_readbuf(
 		 * use our mapping, but this is a very rare case.
 		 */
 		else if (i > mip->ra_current) {
-			xfs_dir3_data_readahead(NULL, dp,
+			xfs_dir3_data_readahead(dp,
 					map[mip->ra_index].br_startoff +
 							mip->ra_offset, -1);
 			mip->ra_current = i;
@@ -456,15 +455,14 @@ xfs_dir2_leaf_readbuf(
 		/*
 		 * Advance offset through the mapping table.
 		 */
-		for (j = 0; j < mp->m_dirblkfsbs; j++) {
+		for (j = 0; j < geo->fsbcount; j += length ) {
 			/*
 			 * The rest of this extent but not more than a dir
 			 * block.
 			 */
-			length = min_t(int, mp->m_dirblkfsbs,
+			length = min_t(int, geo->fsbcount,
 					map[mip->ra_index].br_blockcount -
 							mip->ra_offset);
-			j += length;
 			mip->ra_offset += length;
 
 			/*
@@ -489,22 +487,23 @@ out:
  */
 STATIC int
 xfs_dir2_leaf_getdents(
-	xfs_inode_t		*dp,		/* incore directory inode */
+	struct xfs_da_args	*args,
 	struct dir_context	*ctx,
 	size_t			bufsize)
 {
+	struct xfs_inode	*dp = args->dp;
 	struct xfs_buf		*bp = NULL;	/* data block buffer */
 	xfs_dir2_data_hdr_t	*hdr;		/* data block header */
 	xfs_dir2_data_entry_t	*dep;		/* data entry */
 	xfs_dir2_data_unused_t	*dup;		/* unused entry */
 	int			error = 0;	/* error return value */
 	int			length;		/* temporary length value */
-	xfs_mount_t		*mp;		/* filesystem mount point */
 	int			byteoff;	/* offset in current block */
 	xfs_dir2_off_t		curoff;		/* current overall offset */
 	xfs_dir2_off_t		newoff;		/* new curoff after new blk */
 	char			*ptr = NULL;	/* pointer to current data */
 	struct xfs_dir2_leaf_map_info *map_info;
+	struct xfs_da_geometry	*geo = args->geo;
 
 	/*
 	 * If the offset is at or past the largest allowed value,
@@ -513,15 +512,12 @@ xfs_dir2_leaf_getdents(
 	if (ctx->pos >= XFS_DIR2_MAX_DATAPTR)
 		return 0;
 
-	mp = dp->i_mount;
-
 	/*
 	 * Set up to bmap a number of blocks based on the caller's
 	 * buffer size, the directory block size, and the filesystem
 	 * block size.
 	 */
-	length = howmany(bufsize + mp->m_dirblksize,
-				     mp->m_sb.sb_blocksize);
+	length = howmany(bufsize + geo->blksize, (1 << geo->fsblog));
 	map_info = kmem_zalloc(offsetof(struct xfs_dir2_leaf_map_info, map) +
 				(length * sizeof(struct xfs_bmbt_irec)),
 			       KM_SLEEP | KM_NOFS);
@@ -531,14 +527,14 @@ xfs_dir2_leaf_getdents(
 	 * Inside the loop we keep the main offset value as a byte offset
 	 * in the directory file.
 	 */
-	curoff = xfs_dir2_dataptr_to_byte(mp, ctx->pos);
+	curoff = xfs_dir2_dataptr_to_byte(ctx->pos);
 
 	/*
 	 * Force this conversion through db so we truncate the offset
 	 * down to get the start of the data block.
 	 */
-	map_info->map_off = xfs_dir2_db_to_da(mp,
-					      xfs_dir2_byte_to_db(mp, curoff));
+	map_info->map_off = xfs_dir2_db_to_da(geo,
+					      xfs_dir2_byte_to_db(geo, curoff));
 
 	/*
 	 * Loop over directory entries until we reach the end offset.
@@ -551,9 +547,9 @@ xfs_dir2_leaf_getdents(
 		 * If we have no buffer, or we're off the end of the
 		 * current buffer, need to get another one.
 		 */
-		if (!bp || ptr >= (char *)bp->b_addr + mp->m_dirblksize) {
+		if (!bp || ptr >= (char *)bp->b_addr + geo->blksize) {
 
-			error = xfs_dir2_leaf_readbuf(dp, bufsize, map_info,
+			error = xfs_dir2_leaf_readbuf(args, bufsize, map_info,
 						      &curoff, &bp);
 			if (error || !map_info->map_valid)
 				break;
@@ -561,7 +557,8 @@ xfs_dir2_leaf_getdents(
 			/*
 			 * Having done a read, we need to set a new offset.
 			 */
-			newoff = xfs_dir2_db_off_to_byte(mp, map_info->curdb, 0);
+			newoff = xfs_dir2_db_off_to_byte(geo,
+							 map_info->curdb, 0);
 			/*
 			 * Start of the current block.
 			 */
@@ -571,7 +568,7 @@ xfs_dir2_leaf_getdents(
 			 * Make sure we're in the right block.
 			 */
 			else if (curoff > newoff)
-				ASSERT(xfs_dir2_byte_to_db(mp, curoff) ==
+				ASSERT(xfs_dir2_byte_to_db(geo, curoff) ==
 				       map_info->curdb);
 			hdr = bp->b_addr;
 			xfs_dir3_data_check(dp, bp);
@@ -579,7 +576,7 @@ xfs_dir2_leaf_getdents(
 			 * Find our position in the block.
 			 */
 			ptr = (char *)dp->d_ops->data_entry_p(hdr);
-			byteoff = xfs_dir2_byte_to_off(mp, curoff);
+			byteoff = xfs_dir2_byte_to_off(geo, curoff);
 			/*
 			 * Skip past the header.
 			 */
@@ -608,10 +605,10 @@ xfs_dir2_leaf_getdents(
 				 * Now set our real offset.
 				 */
 				curoff =
-					xfs_dir2_db_off_to_byte(mp,
-					    xfs_dir2_byte_to_db(mp, curoff),
+					xfs_dir2_db_off_to_byte(geo,
+					    xfs_dir2_byte_to_db(geo, curoff),
 					    (char *)ptr - (char *)hdr);
-				if (ptr >= (char *)hdr + mp->m_dirblksize) {
+				if (ptr >= (char *)hdr + geo->blksize) {
 					continue;
 				}
 			}
@@ -635,10 +632,10 @@ xfs_dir2_leaf_getdents(
 		length = dp->d_ops->data_entsize(dep->namelen);
 		filetype = dp->d_ops->data_get_ftype(dep);
 
-		ctx->pos = xfs_dir2_byte_to_dataptr(mp, curoff) & 0x7fffffff;
+		ctx->pos = xfs_dir2_byte_to_dataptr(curoff) & 0x7fffffff;
 		if (!dir_emit(ctx, (char *)dep->name, dep->namelen,
 			    be64_to_cpu(dep->inumber),
-			    xfs_dir3_get_dtype(mp, filetype)))
+			    xfs_dir3_get_dtype(dp->i_mount, filetype)))
 			break;
 
 		/*
@@ -653,10 +650,10 @@ xfs_dir2_leaf_getdents(
 	/*
 	 * All done.  Set output offset value to current offset.
 	 */
-	if (curoff > xfs_dir2_dataptr_to_byte(mp, XFS_DIR2_MAX_DATAPTR))
+	if (curoff > xfs_dir2_dataptr_to_byte(XFS_DIR2_MAX_DATAPTR))
 		ctx->pos = XFS_DIR2_MAX_DATAPTR & 0x7fffffff;
 	else
-		ctx->pos = xfs_dir2_byte_to_dataptr(mp, curoff) & 0x7fffffff;
+		ctx->pos = xfs_dir2_byte_to_dataptr(curoff) & 0x7fffffff;
 	kmem_free(map_info);
 	if (bp)
 		xfs_trans_brelse(NULL, bp);
@@ -668,13 +665,14 @@ xfs_dir2_leaf_getdents(
  */
 int
 xfs_readdir(
-	xfs_inode_t	*dp,
-	struct dir_context *ctx,
-	size_t		bufsize)
+	struct xfs_inode	*dp,
+	struct dir_context	*ctx,
+	size_t			bufsize)
 {
-	int		rval;		/* return value */
-	int		v;		/* type-checking value */
-	uint		lock_mode;
+	struct xfs_da_args	args = { NULL };
+	int			rval;
+	int			v;
+	uint			lock_mode;
 
 	trace_xfs_readdir(dp);
 
@@ -684,15 +682,18 @@ xfs_readdir(
 	ASSERT(S_ISDIR(dp->i_d.di_mode));
 	XFS_STATS_INC(xs_dir_getdents);
 
+	args.dp = dp;
+	args.geo = dp->i_mount->m_dir_geo;
+
 	lock_mode = xfs_ilock_data_map_shared(dp);
 	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL)
-		rval = xfs_dir2_sf_getdents(dp, ctx);
-	else if ((rval = xfs_dir2_isblock(NULL, dp, &v)))
+		rval = xfs_dir2_sf_getdents(&args, ctx);
+	else if ((rval = xfs_dir2_isblock(&args, &v)))
 		;
 	else if (v)
-		rval = xfs_dir2_block_getdents(dp, ctx);
+		rval = xfs_dir2_block_getdents(&args, ctx);
 	else
-		rval = xfs_dir2_leaf_getdents(dp, ctx, bufsize);
+		rval = xfs_dir2_leaf_getdents(&args, ctx, bufsize);
 	xfs_iunlock(dp, lock_mode);
 
 	return rval;

fs/xfs/xfs_dir2_sf.c

@@ -82,8 +82,10 @@ xfs_dir2_block_sfsize(
 	xfs_ino_t		parent = 0;	/* parent inode number */
 	int			size=0;		/* total computed size */
 	int			has_ftype;
+	struct xfs_da_geometry	*geo;
 
 	mp = dp->i_mount;
+	geo = mp->m_dir_geo;
 
 	/*
 	 * if there is a filetype field, add the extra byte to the namelen
@@ -92,7 +94,7 @@ xfs_dir2_block_sfsize(
 	has_ftype = xfs_sb_version_hasftype(&mp->m_sb) ? 1 : 0;
 
 	count = i8count = namelen = 0;
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(geo, hdr);
 	blp = xfs_dir2_block_leaf_p(btp);
 
 	/*
@@ -104,8 +106,8 @@ xfs_dir2_block_sfsize(
 		/*
 		 * Calculate the pointer to the entry at hand.
 		 */
-		dep = (xfs_dir2_data_entry_t *)
-		      ((char *)hdr + xfs_dir2_dataptr_to_off(mp, addr));
+		dep = (xfs_dir2_data_entry_t *)((char *)hdr +
+				xfs_dir2_dataptr_to_off(geo, addr));
 		/*
 		 * Detect . and .., so we can special-case them.
 		 * . is not included in sf directories.
@@ -195,7 +197,7 @@ xfs_dir2_block_to_sf(
 	/*
 	 * Set up to loop over the block's entries.
 	 */
-	btp = xfs_dir2_block_tail_p(mp, hdr);
+	btp = xfs_dir2_block_tail_p(args->geo, hdr);
 	ptr = (char *)dp->d_ops->data_entry_p(hdr);
 	endptr = (char *)xfs_dir2_block_leaf_p(btp);
 	sfep = xfs_dir2_sf_firstentry(sfp);
@@ -247,7 +249,7 @@ xfs_dir2_block_to_sf(
 
 	/* now we are done with the block, we can shrink the inode */
 	logflags = XFS_ILOG_CORE;
-	error = xfs_dir2_shrink_inode(args, mp->m_dirdatablk, bp);
+	error = xfs_dir2_shrink_inode(args, args->geo->datablk, bp);
 	if (error) {
 		ASSERT(error != ENOSPC);
 		goto out;
@@ -285,14 +287,12 @@ int						/* error */
 xfs_dir2_sf_addname(
 	xfs_da_args_t		*args)		/* operation arguments */
 {
-	int			add_entsize;	/* size of the new entry */
 	xfs_inode_t		*dp;		/* incore directory inode */
 	int			error;		/* error return value */
 	int			incr_isize;	/* total change in size */
 	int			new_isize;	/* di_size after adding name */
 	int			objchange;	/* changing to 8-byte inodes */
 	xfs_dir2_data_aoff_t	offset = 0;	/* offset for new entry */
-	int			old_isize;	/* di_size before adding name */
 	int			pick;		/* which algorithm to use */
 	xfs_dir2_sf_hdr_t	*sfp;		/* shortform structure */
 	xfs_dir2_sf_entry_t	*sfep = NULL;	/* shortform entry */
@@ -316,8 +316,7 @@ xfs_dir2_sf_addname(
 	/*
 	 * Compute entry (and change in) size.
 	 */
-	add_entsize = dp->d_ops->sf_entsize(sfp, args->namelen);
-	incr_isize = add_entsize;
+	incr_isize = dp->d_ops->sf_entsize(sfp, args->namelen);
 	objchange = 0;
 #if XFS_BIG_INUMS
 	/*
@@ -325,11 +324,8 @@ xfs_dir2_sf_addname(
 	 */
 	if (args->inumber > XFS_DIR2_MAX_SHORT_INUM && sfp->i8count == 0) {
 		/*
-		 * Yes, adjust the entry size and the total size.
+		 * Yes, adjust the inode size.  old count + (parent + new)
 		 */
-		add_entsize +=
-			(uint)sizeof(xfs_dir2_ino8_t) -
-			(uint)sizeof(xfs_dir2_ino4_t);
 		incr_isize +=
 			(sfp->count + 2) *
 			((uint)sizeof(xfs_dir2_ino8_t) -
@@ -337,8 +333,7 @@ xfs_dir2_sf_addname(
 		objchange = 1;
 	}
 #endif
-	old_isize = (int)dp->i_d.di_size;
-	new_isize = old_isize + incr_isize;
+	new_isize = (int)dp->i_d.di_size + incr_isize;
 	/*
 	 * Won't fit as shortform any more (due to size),
 	 * or the pick routine says it won't (due to offset values).
@@ -593,7 +588,7 @@ xfs_dir2_sf_addname_pick(
 	 * we'll go back, convert to block, then try the insert and convert
 	 * to leaf.
 	 */
-	if (used + (holefit ? 0 : size) > mp->m_dirblksize)
+	if (used + (holefit ? 0 : size) > args->geo->blksize)
 		return 0;
 	/*
 	 * If changing the inode number size, do it the hard way.
@@ -608,7 +603,7 @@ xfs_dir2_sf_addname_pick(
 	/*
 	 * If it won't fit at the end then do it the hard way (use the hole).
 	 */
-	if (used + size > mp->m_dirblksize)
+	if (used + size > args->geo->blksize)
 		return 2;
 	/*
 	 * Do it the easy way.
@@ -659,7 +654,7 @@ xfs_dir2_sf_check(
 	ASSERT((char *)sfep - (char *)sfp == dp->i_d.di_size);
 	ASSERT(offset +
 	       (sfp->count + 2) * (uint)sizeof(xfs_dir2_leaf_entry_t) +
-	       (uint)sizeof(xfs_dir2_block_tail_t) <= mp->m_dirblksize);
+	       (uint)sizeof(xfs_dir2_block_tail_t) <= args->geo->blksize);
 }
 #endif	/* DEBUG */
 
@@ -1110,9 +1105,9 @@ xfs_dir2_sf_toino4(
 }
 
 /*
- * Convert from 4-byte inode numbers to 8-byte inode numbers.
- * The new 8-byte inode number is not there yet, we leave with the
- * count 1 but no corresponding entry.
+ * Convert existing entries from 4-byte inode numbers to 8-byte inode numbers.
+ * The new entry w/ an 8-byte inode number is not there yet; we leave with
+ * i8count set to 1, but no corresponding 8-byte entry.
  */
 static void
 xfs_dir2_sf_toino8(
@@ -1145,7 +1140,7 @@ xfs_dir2_sf_toino8(
 	ASSERT(oldsfp->i8count == 0);
 	memcpy(buf, oldsfp, oldsize);
 	/*
-	 * Compute the new inode size.
+	 * Compute the new inode size (nb: entry count + 1 for parent)
 	 */
 	newsize =
 		oldsize +

fs/xfs/xfs_dquot.c

@@ -353,10 +353,10 @@ xfs_qm_dqalloc(
 			       dqp->q_blkno,
 			       mp->m_quotainfo->qi_dqchunklen,
 			       0);
-
-	error = xfs_buf_geterror(bp);
-	if (error)
+	if (!bp) {
+		error = ENOMEM;
 		goto error1;
+	}
 	bp->b_ops = &xfs_dquot_buf_ops;
 
 	/*
@@ -832,47 +832,6 @@ restart:
 	return (0);
 }
 
-
-STATIC void
-xfs_qm_dqput_final(
-	struct xfs_dquot	*dqp)
-{
-	struct xfs_quotainfo	*qi = dqp->q_mount->m_quotainfo;
-	struct xfs_dquot	*gdqp;
-	struct xfs_dquot	*pdqp;
-
-	trace_xfs_dqput_free(dqp);
-
-	if (list_lru_add(&qi->qi_lru, &dqp->q_lru))
-		XFS_STATS_INC(xs_qm_dquot_unused);
-
-	/*
-	 * If we just added a udquot to the freelist, then we want to release
-	 * the gdquot/pdquot reference that it (probably) has. Otherwise it'll
-	 * keep the gdquot/pdquot from getting reclaimed.
-	 */
-	gdqp = dqp->q_gdquot;
-	if (gdqp) {
-		xfs_dqlock(gdqp);
-		dqp->q_gdquot = NULL;
-	}
-
-	pdqp = dqp->q_pdquot;
-	if (pdqp) {
-		xfs_dqlock(pdqp);
-		dqp->q_pdquot = NULL;
-	}
-	xfs_dqunlock(dqp);
-
-	/*
-	 * If we had a group/project quota hint, release it now.
-	 */
-	if (gdqp)
-		xfs_qm_dqput(gdqp);
-	if (pdqp)
-		xfs_qm_dqput(pdqp);
-}
-
 /*
  * Release a reference to the dquot (decrement ref-count) and unlock it.
  *
@@ -888,10 +847,14 @@ xfs_qm_dqput(
 
 	trace_xfs_dqput(dqp);
 
-	if (--dqp->q_nrefs > 0)
-		xfs_dqunlock(dqp);
-	else
-		xfs_qm_dqput_final(dqp);
+	if (--dqp->q_nrefs == 0) {
+		struct xfs_quotainfo	*qi = dqp->q_mount->m_quotainfo;
+		trace_xfs_dqput_free(dqp);
+
+		if (list_lru_add(&qi->qi_lru, &dqp->q_lru))
+			XFS_STATS_INC(xs_qm_dquot_unused);
+	}
+	xfs_dqunlock(dqp);
 }
 
 /*

fs/xfs/xfs_dquot.h

@@ -52,8 +52,6 @@ typedef struct xfs_dquot {
 	int		 q_bufoffset;	/* off of dq in buffer (# dquots) */
 	xfs_fileoff_t	 q_fileoffset;	/* offset in quotas file */
 
-	struct xfs_dquot*q_gdquot;	/* group dquot, hint only */
-	struct xfs_dquot*q_pdquot;	/* project dquot, hint only */
 	xfs_disk_dquot_t q_core;	/* actual usage & quotas */
 	xfs_dq_logitem_t q_logitem;	/* dquot log item */
 	xfs_qcnt_t	 q_res_bcount;	/* total regular nblks used+reserved */

fs/xfs/xfs_dquot_buf.c

@@ -35,7 +35,6 @@
 
 int
 xfs_calc_dquots_per_chunk(
-	struct xfs_mount	*mp,
 	unsigned int		nbblks)	/* basic block units */
 {
 	unsigned int	ndquots;
@@ -194,7 +193,7 @@ xfs_dquot_buf_verify_crc(
 	if (mp->m_quotainfo)
 		ndquots = mp->m_quotainfo->qi_dqperchunk;
 	else
-		ndquots = xfs_calc_dquots_per_chunk(mp,
+		ndquots = xfs_calc_dquots_per_chunk(
 					XFS_BB_TO_FSB(mp, bp->b_length));
 
 	for (i = 0; i < ndquots; i++, d++) {
@@ -225,7 +224,7 @@ xfs_dquot_buf_verify(
 	if (mp->m_quotainfo)
 		ndquots = mp->m_quotainfo->qi_dqperchunk;
 	else
-		ndquots = xfs_calc_dquots_per_chunk(mp, bp->b_length);
+		ndquots = xfs_calc_dquots_per_chunk(bp->b_length);
 
 	/*
 	 * On the first read of the buffer, verify that each dquot is valid.

fs/xfs/xfs_file.c

@@ -944,7 +944,7 @@ xfs_dir_open(
 	 */
 	mode = xfs_ilock_data_map_shared(ip);
 	if (ip->i_d.di_nextents > 0)
-		xfs_dir3_data_readahead(NULL, ip, 0, -1);
+		xfs_dir3_data_readahead(ip, 0, -1);
 	xfs_iunlock(ip, mode);
 	return 0;
 }

fs/xfs/xfs_filestream.c

@@ -1,5 +1,6 @@
 /*
  * Copyright (c) 2006-2007 Silicon Graphics, Inc.
+ * Copyright (c) 2014 Christoph Hellwig.
  * All Rights Reserved.
  *
  * This program is free software; you can redistribute it and/or
@@ -32,100 +33,20 @@
 #include "xfs_filestream.h"
 #include "xfs_trace.h"
 
-#ifdef XFS_FILESTREAMS_TRACE
-
-ktrace_t *xfs_filestreams_trace_buf;
-
-STATIC void
-xfs_filestreams_trace(
-	xfs_mount_t	*mp,	/* mount point */
-	int		type,	/* type of trace */
-	const char	*func,	/* source function */
-	int		line,	/* source line number */
-	__psunsigned_t	arg0,
-	__psunsigned_t	arg1,
-	__psunsigned_t	arg2,
-	__psunsigned_t	arg3,
-	__psunsigned_t	arg4,
-	__psunsigned_t	arg5)
-{
-	ktrace_enter(xfs_filestreams_trace_buf,
-		(void *)(__psint_t)(type | (line << 16)),
-		(void *)func,
-		(void *)(__psunsigned_t)current_pid(),
-		(void *)mp,
-		(void *)(__psunsigned_t)arg0,
-		(void *)(__psunsigned_t)arg1,
-		(void *)(__psunsigned_t)arg2,
-		(void *)(__psunsigned_t)arg3,
-		(void *)(__psunsigned_t)arg4,
-		(void *)(__psunsigned_t)arg5,
-		NULL, NULL, NULL, NULL, NULL, NULL);
-}
-
-#define TRACE0(mp,t)			TRACE6(mp,t,0,0,0,0,0,0)
-#define TRACE1(mp,t,a0)			TRACE6(mp,t,a0,0,0,0,0,0)
-#define TRACE2(mp,t,a0,a1)		TRACE6(mp,t,a0,a1,0,0,0,0)
-#define TRACE3(mp,t,a0,a1,a2)		TRACE6(mp,t,a0,a1,a2,0,0,0)
-#define TRACE4(mp,t,a0,a1,a2,a3)	TRACE6(mp,t,a0,a1,a2,a3,0,0)
-#define TRACE5(mp,t,a0,a1,a2,a3,a4)	TRACE6(mp,t,a0,a1,a2,a3,a4,0)
-#define TRACE6(mp,t,a0,a1,a2,a3,a4,a5) \
-	xfs_filestreams_trace(mp, t, __func__, __LINE__, \
-				(__psunsigned_t)a0, (__psunsigned_t)a1, \
-				(__psunsigned_t)a2, (__psunsigned_t)a3, \
-				(__psunsigned_t)a4, (__psunsigned_t)a5)
-
-#define TRACE_AG_SCAN(mp, ag, ag2) \
-		TRACE2(mp, XFS_FSTRM_KTRACE_AGSCAN, ag, ag2);
-#define TRACE_AG_PICK1(mp, max_ag, maxfree) \
-		TRACE2(mp, XFS_FSTRM_KTRACE_AGPICK1, max_ag, maxfree);
-#define TRACE_AG_PICK2(mp, ag, ag2, cnt, free, scan, flag) \
-		TRACE6(mp, XFS_FSTRM_KTRACE_AGPICK2, ag, ag2, \
-			 cnt, free, scan, flag)
-#define TRACE_UPDATE(mp, ip, ag, cnt, ag2, cnt2) \
-		TRACE5(mp, XFS_FSTRM_KTRACE_UPDATE, ip, ag, cnt, ag2, cnt2)
-#define TRACE_FREE(mp, ip, pip, ag, cnt) \
-		TRACE4(mp, XFS_FSTRM_KTRACE_FREE, ip, pip, ag, cnt)
-#define TRACE_LOOKUP(mp, ip, pip, ag, cnt) \
-		TRACE4(mp, XFS_FSTRM_KTRACE_ITEM_LOOKUP, ip, pip, ag, cnt)
-#define TRACE_ASSOCIATE(mp, ip, pip, ag, cnt) \
-		TRACE4(mp, XFS_FSTRM_KTRACE_ASSOCIATE, ip, pip, ag, cnt)
-#define TRACE_MOVEAG(mp, ip, pip, oag, ocnt, nag, ncnt) \
-		TRACE6(mp, XFS_FSTRM_KTRACE_MOVEAG, ip, pip, oag, ocnt, nag, ncnt)
-#define TRACE_ORPHAN(mp, ip, ag) \
-		TRACE2(mp, XFS_FSTRM_KTRACE_ORPHAN, ip, ag);
-
-
-#else
-#define TRACE_AG_SCAN(mp, ag, ag2)
-#define TRACE_AG_PICK1(mp, max_ag, maxfree)
-#define TRACE_AG_PICK2(mp, ag, ag2, cnt, free, scan, flag)
-#define TRACE_UPDATE(mp, ip, ag, cnt, ag2, cnt2)
-#define TRACE_FREE(mp, ip, pip, ag, cnt)
-#define TRACE_LOOKUP(mp, ip, pip, ag, cnt)
-#define TRACE_ASSOCIATE(mp, ip, pip, ag, cnt)
-#define TRACE_MOVEAG(mp, ip, pip, oag, ocnt, nag, ncnt)
-#define TRACE_ORPHAN(mp, ip, ag)
-#endif
-
-static kmem_zone_t *item_zone;
+struct xfs_fstrm_item {
+	struct xfs_mru_cache_elem	mru;
+	struct xfs_inode		*ip;
+	xfs_agnumber_t			ag; /* AG in use for this directory */
+};
 
-/*
- * Structure for associating a file or a directory with an allocation group.
- * The parent directory pointer is only needed for files, but since there will
- * generally be vastly more files than directories in the cache, using the same
- * data structure simplifies the code with very little memory overhead.
- */
-typedef struct fstrm_item
-{
-	xfs_agnumber_t	ag;	/* AG currently in use for the file/directory. */
-	xfs_inode_t	*ip;	/* inode self-pointer. */
-	xfs_inode_t	*pip;	/* Parent directory inode pointer. */
-} fstrm_item_t;
+enum xfs_fstrm_alloc {
+	XFS_PICK_USERDATA = 1,
+	XFS_PICK_LOWSPACE = 2,
+};
 
 /*
  * Allocation group filestream associations are tracked with per-ag atomic
- * counters.  These counters allow _xfs_filestream_pick_ag() to tell whether a
+ * counters.  These counters allow xfs_filestream_pick_ag() to tell whether a
  * particular AG already has active filestreams associated with it. The mount
  * point's m_peraglock is used to protect these counters from per-ag array
  * re-allocation during a growfs operation.  When xfs_growfs_data_private() is
@@ -160,7 +81,7 @@ typedef struct fstrm_item
  * the cache that reference per-ag array elements that have since been
  * reallocated.
  */
-static int
+int
 xfs_filestream_peek_ag(
 	xfs_mount_t	*mp,
 	xfs_agnumber_t	agno)
@@ -200,23 +121,40 @@ xfs_filestream_put_ag(
 	xfs_perag_put(pag);
 }
 
+static void
+xfs_fstrm_free_func(
+	struct xfs_mru_cache_elem *mru)
+{
+	struct xfs_fstrm_item	*item =
+		container_of(mru, struct xfs_fstrm_item, mru);
+
+	xfs_filestream_put_ag(item->ip->i_mount, item->ag);
+
+	trace_xfs_filestream_free(item->ip, item->ag);
+
+	kmem_free(item);
+}
+
 /*
  * Scan the AGs starting at startag looking for an AG that isn't in use and has
  * at least minlen blocks free.
  */
 static int
-_xfs_filestream_pick_ag(
-	xfs_mount_t	*mp,
-	xfs_agnumber_t	startag,
-	xfs_agnumber_t	*agp,
-	int		flags,
-	xfs_extlen_t	minlen)
+xfs_filestream_pick_ag(
+	struct xfs_inode	*ip,
+	xfs_agnumber_t		startag,
+	xfs_agnumber_t		*agp,
+	int			flags,
+	xfs_extlen_t		minlen)
 {
-	int		streams, max_streams;
-	int		err, trylock, nscan;
-	xfs_extlen_t	longest, free, minfree, maxfree = 0;
-	xfs_agnumber_t	ag, max_ag = NULLAGNUMBER;
-	struct xfs_perag *pag;
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_fstrm_item	*item;
+	struct xfs_perag	*pag;
+	xfs_extlen_t		longest, free = 0, minfree, maxfree = 0;
+	xfs_agnumber_t		ag, max_ag = NULLAGNUMBER;
+	int			err, trylock, nscan;
+
+	ASSERT(S_ISDIR(ip->i_d.di_mode));
 
 	/* 2% of an AG's blocks must be free for it to be chosen. */
 	minfree = mp->m_sb.sb_agblocks / 50;
@@ -228,8 +166,9 @@ _xfs_filestream_pick_ag(
 	trylock = XFS_ALLOC_FLAG_TRYLOCK;
 
 	for (nscan = 0; 1; nscan++) {
+		trace_xfs_filestream_scan(ip, ag);
+
 		pag = xfs_perag_get(mp, ag);
-		TRACE_AG_SCAN(mp, ag, atomic_read(&pag->pagf_fstrms));
 
 		if (!pag->pagf_init) {
 			err = xfs_alloc_pagf_init(mp, NULL, ag, trylock);
@@ -246,7 +185,6 @@ _xfs_filestream_pick_ag(
 		/* Keep track of the AG with the most free blocks. */
 		if (pag->pagf_freeblks > maxfree) {
 			maxfree = pag->pagf_freeblks;
-			max_streams = atomic_read(&pag->pagf_fstrms);
 			max_ag = ag;
 		}
 
@@ -269,7 +207,6 @@ _xfs_filestream_pick_ag(
 
 			/* Break out, retaining the reference on the AG. */
 			free = pag->pagf_freeblks;
-			streams = atomic_read(&pag->pagf_fstrms);
 			xfs_perag_put(pag);
 			*agp = ag;
 			break;
@@ -305,317 +242,98 @@ next_ag:
 		 */
 		if (max_ag != NULLAGNUMBER) {
 			xfs_filestream_get_ag(mp, max_ag);
-			TRACE_AG_PICK1(mp, max_ag, maxfree);
-			streams = max_streams;
 			free = maxfree;
 			*agp = max_ag;
 			break;
 		}
 
 		/* take AG 0 if none matched */
-		TRACE_AG_PICK1(mp, max_ag, maxfree);
+		trace_xfs_filestream_pick(ip, *agp, free, nscan);
 		*agp = 0;
 		return 0;
 	}
 
-	TRACE_AG_PICK2(mp, startag, *agp, streams, free, nscan, flags);
-
-	return 0;
-}
+	trace_xfs_filestream_pick(ip, *agp, free, nscan);
 
-/*
- * Set the allocation group number for a file or a directory, updating inode
- * references and per-AG references as appropriate.
- */
-static int
-_xfs_filestream_update_ag(
-	xfs_inode_t	*ip,
-	xfs_inode_t	*pip,
-	xfs_agnumber_t	ag)
-{
-	int		err = 0;
-	xfs_mount_t	*mp;
-	xfs_mru_cache_t	*cache;
-	fstrm_item_t	*item;
-	xfs_agnumber_t	old_ag;
-	xfs_inode_t	*old_pip;
-
-	/*
-	 * Either ip is a regular file and pip is a directory, or ip is a
-	 * directory and pip is NULL.
-	 */
-	ASSERT(ip && ((S_ISREG(ip->i_d.di_mode) && pip &&
-	               S_ISDIR(pip->i_d.di_mode)) ||
-	              (S_ISDIR(ip->i_d.di_mode) && !pip)));
-
-	mp = ip->i_mount;
-	cache = mp->m_filestream;
-
-	item = xfs_mru_cache_lookup(cache, ip->i_ino);
-	if (item) {
-		ASSERT(item->ip == ip);
-		old_ag = item->ag;
-		item->ag = ag;
-		old_pip = item->pip;
-		item->pip = pip;
-		xfs_mru_cache_done(cache);
-
-		/*
-		 * If the AG has changed, drop the old ref and take a new one,
-		 * effectively transferring the reference from old to new AG.
-		 */
-		if (ag != old_ag) {
-			xfs_filestream_put_ag(mp, old_ag);
-			xfs_filestream_get_ag(mp, ag);
-		}
-
-		/*
-		 * If ip is a file and its pip has changed, drop the old ref and
-		 * take a new one.
-		 */
-		if (pip && pip != old_pip) {
-			IRELE(old_pip);
-			IHOLD(pip);
-		}
-
-		TRACE_UPDATE(mp, ip, old_ag, xfs_filestream_peek_ag(mp, old_ag),
-				ag, xfs_filestream_peek_ag(mp, ag));
+	if (*agp == NULLAGNUMBER)
 		return 0;
-	}
 
-	item = kmem_zone_zalloc(item_zone, KM_MAYFAIL);
+	err = ENOMEM;
+	item = kmem_alloc(sizeof(*item), KM_MAYFAIL);
 	if (!item)
-		return ENOMEM;
+		goto out_put_ag;
 
-	item->ag = ag;
+	item->ag = *agp;
 	item->ip = ip;
-	item->pip = pip;
 
-	err = xfs_mru_cache_insert(cache, ip->i_ino, item);
+	err = xfs_mru_cache_insert(mp->m_filestream, ip->i_ino, &item->mru);
 	if (err) {
-		kmem_zone_free(item_zone, item);
-		return err;
+		if (err == EEXIST)
+			err = 0;
+		goto out_free_item;
 	}
 
-	/* Take a reference on the AG. */
-	xfs_filestream_get_ag(mp, ag);
-
-	/*
-	 * Take a reference on the inode itself regardless of whether it's a
-	 * regular file or a directory.
-	 */
-	IHOLD(ip);
-
-	/*
-	 * In the case of a regular file, take a reference on the parent inode
-	 * as well to ensure it remains in-core.
-	 */
-	if (pip)
-		IHOLD(pip);
-
-	TRACE_UPDATE(mp, ip, ag, xfs_filestream_peek_ag(mp, ag),
-			ag, xfs_filestream_peek_ag(mp, ag));
-
 	return 0;
-}
-
-/* xfs_fstrm_free_func(): callback for freeing cached stream items. */
-STATIC void
-xfs_fstrm_free_func(
-	unsigned long	ino,
-	void		*data)
-{
-	fstrm_item_t	*item  = (fstrm_item_t *)data;
-	xfs_inode_t	*ip = item->ip;
-
-	ASSERT(ip->i_ino == ino);
-
-	xfs_iflags_clear(ip, XFS_IFILESTREAM);
-
-	/* Drop the reference taken on the AG when the item was added. */
-	xfs_filestream_put_ag(ip->i_mount, item->ag);
-
-	TRACE_FREE(ip->i_mount, ip, item->pip, item->ag,
-		xfs_filestream_peek_ag(ip->i_mount, item->ag));
-
-	/*
-	 * _xfs_filestream_update_ag() always takes a reference on the inode
-	 * itself, whether it's a file or a directory.  Release it here.
-	 * This can result in the inode being freed and so we must
-	 * not hold any inode locks when freeing filesstreams objects
-	 * otherwise we can deadlock here.
-	 */
-	IRELE(ip);
-
-	/*
-	 * In the case of a regular file, _xfs_filestream_update_ag() also
-	 * takes a ref on the parent inode to keep it in-core.  Release that
-	 * too.
-	 */
-	if (item->pip)
-		IRELE(item->pip);
-
-	/* Finally, free the memory allocated for the item. */
-	kmem_zone_free(item_zone, item);
-}
-
-/*
- * xfs_filestream_init() is called at xfs initialisation time to set up the
- * memory zone that will be used for filestream data structure allocation.
- */
-int
-xfs_filestream_init(void)
-{
-	item_zone = kmem_zone_init(sizeof(fstrm_item_t), "fstrm_item");
-	if (!item_zone)
-		return -ENOMEM;
-
-	return 0;
-}
-
-/*
- * xfs_filestream_uninit() is called at xfs termination time to destroy the
- * memory zone that was used for filestream data structure allocation.
- */
-void
-xfs_filestream_uninit(void)
-{
-	kmem_zone_destroy(item_zone);
-}
-
-/*
- * xfs_filestream_mount() is called when a file system is mounted with the
- * filestream option.  It is responsible for allocating the data structures
- * needed to track the new file system's file streams.
- */
-int
-xfs_filestream_mount(
-	xfs_mount_t	*mp)
-{
-	int		err;
-	unsigned int	lifetime, grp_count;
-
-	/*
-	 * The filestream timer tunable is currently fixed within the range of
-	 * one second to four minutes, with five seconds being the default.  The
-	 * group count is somewhat arbitrary, but it'd be nice to adhere to the
-	 * timer tunable to within about 10 percent.  This requires at least 10
-	 * groups.
-	 */
-	lifetime  = xfs_fstrm_centisecs * 10;
-	grp_count = 10;
-
-	err = xfs_mru_cache_create(&mp->m_filestream, lifetime, grp_count,
-	                     xfs_fstrm_free_func);
 
+out_free_item:
+	kmem_free(item);
+out_put_ag:
+	xfs_filestream_put_ag(mp, *agp);
 	return err;
 }
 
-/*
- * xfs_filestream_unmount() is called when a file system that was mounted with
- * the filestream option is unmounted.  It drains the data structures created
- * to track the file system's file streams and frees all the memory that was
- * allocated.
- */
-void
-xfs_filestream_unmount(
-	xfs_mount_t	*mp)
+static struct xfs_inode *
+xfs_filestream_get_parent(
+	struct xfs_inode	*ip)
 {
-	xfs_mru_cache_destroy(mp->m_filestream);
-}
+	struct inode		*inode = VFS_I(ip), *dir = NULL;
+	struct dentry		*dentry, *parent;
 
-/*
- * Return the AG of the filestream the file or directory belongs to, or
- * NULLAGNUMBER otherwise.
- */
-xfs_agnumber_t
-xfs_filestream_lookup_ag(
-	xfs_inode_t	*ip)
-{
-	xfs_mru_cache_t	*cache;
-	fstrm_item_t	*item;
-	xfs_agnumber_t	ag;
-	int		ref;
-
-	if (!S_ISREG(ip->i_d.di_mode) && !S_ISDIR(ip->i_d.di_mode)) {
-		ASSERT(0);
-		return NULLAGNUMBER;
-	}
+	dentry = d_find_alias(inode);
+	if (!dentry)
+		goto out;
 
-	cache = ip->i_mount->m_filestream;
-	item = xfs_mru_cache_lookup(cache, ip->i_ino);
-	if (!item) {
-		TRACE_LOOKUP(ip->i_mount, ip, NULL, NULLAGNUMBER, 0);
-		return NULLAGNUMBER;
-	}
+	parent = dget_parent(dentry);
+	if (!parent)
+		goto out_dput;
 
-	ASSERT(ip == item->ip);
-	ag = item->ag;
-	ref = xfs_filestream_peek_ag(ip->i_mount, ag);
-	xfs_mru_cache_done(cache);
+	dir = igrab(parent->d_inode);
+	dput(parent);
 
-	TRACE_LOOKUP(ip->i_mount, ip, item->pip, ag, ref);
-	return ag;
+out_dput:
+	dput(dentry);
+out:
+	return dir ? XFS_I(dir) : NULL;
 }
 
 /*
- * xfs_filestream_associate() should only be called to associate a regular file
- * with its parent directory.  Calling it with a child directory isn't
- * appropriate because filestreams don't apply to entire directory hierarchies.
- * Creating a file in a child directory of an existing filestream directory
- * starts a new filestream with its own allocation group association.
+ * Find the right allocation group for a file, either by finding an
+ * existing file stream or creating a new one.
  *
- * Returns < 0 on error, 0 if successful association occurred, > 0 if
- * we failed to get an association because of locking issues.
+ * Returns NULLAGNUMBER in case of an error.
  */
-int
-xfs_filestream_associate(
-	xfs_inode_t	*pip,
-	xfs_inode_t	*ip)
+xfs_agnumber_t
+xfs_filestream_lookup_ag(
+	struct xfs_inode	*ip)
 {
-	xfs_mount_t	*mp;
-	xfs_mru_cache_t	*cache;
-	fstrm_item_t	*item;
-	xfs_agnumber_t	ag, rotorstep, startag;
-	int		err = 0;
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_inode	*pip = NULL;
+	xfs_agnumber_t		startag, ag = NULLAGNUMBER;
+	struct xfs_mru_cache_elem *mru;
 
-	ASSERT(S_ISDIR(pip->i_d.di_mode));
 	ASSERT(S_ISREG(ip->i_d.di_mode));
-	if (!S_ISDIR(pip->i_d.di_mode) || !S_ISREG(ip->i_d.di_mode))
-		return -EINVAL;
 
-	mp = pip->i_mount;
-	cache = mp->m_filestream;
+	pip = xfs_filestream_get_parent(ip);
+	if (!pip)
+		goto out;
 
-	/*
-	 * We have a problem, Houston.
-	 *
-	 * Taking the iolock here violates inode locking order - we already
-	 * hold the ilock. Hence if we block getting this lock we may never
-	 * wake. Unfortunately, that means if we can't get the lock, we're
-	 * screwed in terms of getting a stream association - we can't spin
-	 * waiting for the lock because someone else is waiting on the lock we
-	 * hold and we cannot drop that as we are in a transaction here.
-	 *
-	 * Lucky for us, this inversion is not a problem because it's a
-	 * directory inode that we are trying to lock here.
-	 *
-	 * So, if we can't get the iolock without sleeping then just give up
-	 */
-	if (!xfs_ilock_nowait(pip, XFS_IOLOCK_EXCL))
-		return 1;
-
-	/* If the parent directory is already in the cache, use its AG. */
-	item = xfs_mru_cache_lookup(cache, pip->i_ino);
-	if (item) {
-		ASSERT(item->ip == pip);
-		ag = item->ag;
-		xfs_mru_cache_done(cache);
-
-		TRACE_LOOKUP(mp, pip, pip, ag, xfs_filestream_peek_ag(mp, ag));
-		err = _xfs_filestream_update_ag(ip, pip, ag);
+	mru = xfs_mru_cache_lookup(mp->m_filestream, pip->i_ino);
+	if (mru) {
+		ag = container_of(mru, struct xfs_fstrm_item, mru)->ag;
+		xfs_mru_cache_done(mp->m_filestream);
 
-		goto exit;
+		trace_xfs_filestream_lookup(ip, ag);
+		goto out;
 	}
 
 	/*
@@ -623,202 +341,94 @@ xfs_filestream_associate(
 	 * use the directory inode's AG.
 	 */
 	if (mp->m_flags & XFS_MOUNT_32BITINODES) {
-		rotorstep = xfs_rotorstep;
+		xfs_agnumber_t	 rotorstep = xfs_rotorstep;
 		startag = (mp->m_agfrotor / rotorstep) % mp->m_sb.sb_agcount;
 		mp->m_agfrotor = (mp->m_agfrotor + 1) %
 		                 (mp->m_sb.sb_agcount * rotorstep);
 	} else
 		startag = XFS_INO_TO_AGNO(mp, pip->i_ino);
 
-	/* Pick a new AG for the parent inode starting at startag. */
-	err = _xfs_filestream_pick_ag(mp, startag, &ag, 0, 0);
-	if (err || ag == NULLAGNUMBER)
-		goto exit_did_pick;
-
-	/* Associate the parent inode with the AG. */
-	err = _xfs_filestream_update_ag(pip, NULL, ag);
-	if (err)
-		goto exit_did_pick;
-
-	/* Associate the file inode with the AG. */
-	err = _xfs_filestream_update_ag(ip, pip, ag);
-	if (err)
-		goto exit_did_pick;
-
-	TRACE_ASSOCIATE(mp, ip, pip, ag, xfs_filestream_peek_ag(mp, ag));
-
-exit_did_pick:
-	/*
-	 * If _xfs_filestream_pick_ag() returned a valid AG, remove the
-	 * reference it took on it, since the file and directory will have taken
-	 * their own now if they were successfully cached.
-	 */
-	if (ag != NULLAGNUMBER)
-		xfs_filestream_put_ag(mp, ag);
-
-exit:
-	xfs_iunlock(pip, XFS_IOLOCK_EXCL);
-	return -err;
+	if (xfs_filestream_pick_ag(pip, startag, &ag, 0, 0))
+		ag = NULLAGNUMBER;
+out:
+	IRELE(pip);
+	return ag;
 }
 
 /*
- * Pick a new allocation group for the current file and its file stream.  This
- * function is called by xfs_bmap_filestreams() with the mount point's per-ag
- * lock held.
+ * Pick a new allocation group for the current file and its file stream.
+ *
+ * This is called when the allocator can't find a suitable extent in the
+ * current AG, and we have to move the stream into a new AG with more space.
  */
 int
 xfs_filestream_new_ag(
 	struct xfs_bmalloca	*ap,
 	xfs_agnumber_t		*agp)
 {
-	int		flags, err;
-	xfs_inode_t	*ip, *pip = NULL;
-	xfs_mount_t	*mp;
-	xfs_mru_cache_t	*cache;
-	xfs_extlen_t	minlen;
-	fstrm_item_t	*dir, *file;
-	xfs_agnumber_t	ag = NULLAGNUMBER;
-
-	ip = ap->ip;
-	mp = ip->i_mount;
-	cache = mp->m_filestream;
-	minlen = ap->length;
-	*agp = NULLAGNUMBER;
+	struct xfs_inode	*ip = ap->ip, *pip;
+	struct xfs_mount	*mp = ip->i_mount;
+	xfs_extlen_t		minlen = ap->length;
+	xfs_agnumber_t		startag = 0;
+	int			flags, err = 0;
+	struct xfs_mru_cache_elem *mru;
 
-	/*
-	 * Look for the file in the cache, removing it if it's found.  Doing
-	 * this allows it to be held across the dir lookup that follows.
-	 */
-	file = xfs_mru_cache_remove(cache, ip->i_ino);
-	if (file) {
-		ASSERT(ip == file->ip);
-
-		/* Save the file's parent inode and old AG number for later. */
-		pip = file->pip;
-		ag = file->ag;
-
-		/* Look for the file's directory in the cache. */
-		dir = xfs_mru_cache_lookup(cache, pip->i_ino);
-		if (dir) {
-			ASSERT(pip == dir->ip);
-
-			/*
-			 * If the directory has already moved on to a new AG,
-			 * use that AG as the new AG for the file. Don't
-			 * forget to twiddle the AG refcounts to match the
-			 * movement.
-			 */
-			if (dir->ag != file->ag) {
-				xfs_filestream_put_ag(mp, file->ag);
-				xfs_filestream_get_ag(mp, dir->ag);
-				*agp = file->ag = dir->ag;
-			}
-
-			xfs_mru_cache_done(cache);
-		}
+	*agp = NULLAGNUMBER;
 
-		/*
-		 * Put the file back in the cache.  If this fails, the free
-		 * function needs to be called to tidy up in the same way as if
-		 * the item had simply expired from the cache.
-		 */
-		err = xfs_mru_cache_insert(cache, ip->i_ino, file);
-		if (err) {
-			xfs_fstrm_free_func(ip->i_ino, file);
-			return err;
-		}
+	pip = xfs_filestream_get_parent(ip);
+	if (!pip)
+		goto exit;
 
-		/*
-		 * If the file's AG was moved to the directory's new AG, there's
-		 * nothing more to be done.
-		 */
-		if (*agp != NULLAGNUMBER) {
-			TRACE_MOVEAG(mp, ip, pip,
-					ag, xfs_filestream_peek_ag(mp, ag),
-					*agp, xfs_filestream_peek_ag(mp, *agp));
-			return 0;
-		}
+	mru = xfs_mru_cache_remove(mp->m_filestream, pip->i_ino);
+	if (mru) {
+		struct xfs_fstrm_item *item =
+			container_of(mru, struct xfs_fstrm_item, mru);
+		startag = (item->ag + 1) % mp->m_sb.sb_agcount;
 	}
 
-	/*
-	 * If the file's parent directory is known, take its iolock in exclusive
-	 * mode to prevent two sibling files from racing each other to migrate
-	 * themselves and their parent to different AGs.
-	 *
-	 * Note that we lock the parent directory iolock inside the child
-	 * iolock here.  That's fine as we never hold both parent and child
-	 * iolock in any other place.  This is different from the ilock,
-	 * which requires locking of the child after the parent for namespace
-	 * operations.
-	 */
-	if (pip)
-		xfs_ilock(pip, XFS_IOLOCK_EXCL | XFS_IOLOCK_PARENT);
-
-	/*
-	 * A new AG needs to be found for the file.  If the file's parent
-	 * directory is also known, it will be moved to the new AG as well to
-	 * ensure that files created inside it in future use the new AG.
-	 */
-	ag = (ag == NULLAGNUMBER) ? 0 : (ag + 1) % mp->m_sb.sb_agcount;
 	flags = (ap->userdata ? XFS_PICK_USERDATA : 0) |
 	        (ap->flist->xbf_low ? XFS_PICK_LOWSPACE : 0);
 
-	err = _xfs_filestream_pick_ag(mp, ag, agp, flags, minlen);
-	if (err || *agp == NULLAGNUMBER)
-		goto exit;
+	err = xfs_filestream_pick_ag(pip, startag, agp, flags, minlen);
 
 	/*
-	 * If the file wasn't found in the file cache, then its parent directory
-	 * inode isn't known.  For this to have happened, the file must either
-	 * be pre-existing, or it was created long enough ago that its cache
-	 * entry has expired.  This isn't the sort of usage that the filestreams
-	 * allocator is trying to optimise, so there's no point trying to track
-	 * its new AG somehow in the filestream data structures.
+	 * Only free the item here so we skip over the old AG earlier.
 	 */
-	if (!pip) {
-		TRACE_ORPHAN(mp, ip, *agp);
-		goto exit;
-	}
-
-	/* Associate the parent inode with the AG. */
-	err = _xfs_filestream_update_ag(pip, NULL, *agp);
-	if (err)
-		goto exit;
-
-	/* Associate the file inode with the AG. */
-	err = _xfs_filestream_update_ag(ip, pip, *agp);
-	if (err)
-		goto exit;
-
-	TRACE_MOVEAG(mp, ip, pip, NULLAGNUMBER, 0,
-			*agp, xfs_filestream_peek_ag(mp, *agp));
+	if (mru)
+		xfs_fstrm_free_func(mru);
 
+	IRELE(pip);
 exit:
-	/*
-	 * If _xfs_filestream_pick_ag() returned a valid AG, remove the
-	 * reference it took on it, since the file and directory will have taken
-	 * their own now if they were successfully cached.
-	 */
-	if (*agp != NULLAGNUMBER)
-		xfs_filestream_put_ag(mp, *agp);
-	else
+	if (*agp == NULLAGNUMBER)
 		*agp = 0;
-
-	if (pip)
-		xfs_iunlock(pip, XFS_IOLOCK_EXCL);
-
 	return err;
 }
 
-/*
- * Remove an association between an inode and a filestream object.
- * Typically this is done on last close of an unlinked file.
- */
 void
 xfs_filestream_deassociate(
-	xfs_inode_t	*ip)
+	struct xfs_inode	*ip)
 {
-	xfs_mru_cache_t	*cache = ip->i_mount->m_filestream;
+	xfs_mru_cache_delete(ip->i_mount->m_filestream, ip->i_ino);
+}
+
+int
+xfs_filestream_mount(
+	xfs_mount_t	*mp)
+{
+	/*
+	 * The filestream timer tunable is currently fixed within the range of
+	 * one second to four minutes, with five seconds being the default.  The
+	 * group count is somewhat arbitrary, but it'd be nice to adhere to the
+	 * timer tunable to within about 10 percent.  This requires at least 10
+	 * groups.
+	 */
+	return xfs_mru_cache_create(&mp->m_filestream, xfs_fstrm_centisecs * 10,
+				    10, xfs_fstrm_free_func);
+}
 
-	xfs_mru_cache_delete(cache, ip->i_ino);
+void
+xfs_filestream_unmount(
+	xfs_mount_t	*mp)
+{
+	xfs_mru_cache_destroy(mp->m_filestream);
 }

fs/xfs/xfs_filestream.h

@@ -20,50 +20,20 @@
 
 struct xfs_mount;
 struct xfs_inode;
-struct xfs_perag;
 struct xfs_bmalloca;
 
-#ifdef XFS_FILESTREAMS_TRACE
-#define XFS_FSTRM_KTRACE_INFO		1
-#define XFS_FSTRM_KTRACE_AGSCAN		2
-#define XFS_FSTRM_KTRACE_AGPICK1	3
-#define XFS_FSTRM_KTRACE_AGPICK2	4
-#define XFS_FSTRM_KTRACE_UPDATE		5
-#define XFS_FSTRM_KTRACE_FREE		6
-#define	XFS_FSTRM_KTRACE_ITEM_LOOKUP	7
-#define	XFS_FSTRM_KTRACE_ASSOCIATE	8
-#define	XFS_FSTRM_KTRACE_MOVEAG		9
-#define	XFS_FSTRM_KTRACE_ORPHAN		10
-
-#define XFS_FSTRM_KTRACE_SIZE	16384
-extern ktrace_t *xfs_filestreams_trace_buf;
-
-#endif
-
-/* allocation selection flags */
-typedef enum xfs_fstrm_alloc {
-	XFS_PICK_USERDATA = 1,
-	XFS_PICK_LOWSPACE = 2,
-} xfs_fstrm_alloc_t;
-
-/* prototypes for filestream.c */
-int xfs_filestream_init(void);
-void xfs_filestream_uninit(void);
 int xfs_filestream_mount(struct xfs_mount *mp);
 void xfs_filestream_unmount(struct xfs_mount *mp);
-xfs_agnumber_t xfs_filestream_lookup_ag(struct xfs_inode *ip);
-int xfs_filestream_associate(struct xfs_inode *dip, struct xfs_inode *ip);
 void xfs_filestream_deassociate(struct xfs_inode *ip);
+xfs_agnumber_t xfs_filestream_lookup_ag(struct xfs_inode *ip);
 int xfs_filestream_new_ag(struct xfs_bmalloca *ap, xfs_agnumber_t *agp);
+int xfs_filestream_peek_ag(struct xfs_mount *mp, xfs_agnumber_t agno);
 
-
-/* filestreams for the inode? */
 static inline int
 xfs_inode_is_filestream(
 	struct xfs_inode	*ip)
 {
 	return (ip->i_mount->m_flags & XFS_MOUNT_FILESTREAMS) ||
-		xfs_iflags_test(ip, XFS_IFILESTREAM) ||
 		(ip->i_d.di_flags & XFS_DIFLAG_FILESTREAM);
 }
 

fs/xfs/xfs_format.h

@@ -202,6 +202,8 @@ typedef __be32 xfs_alloc_ptr_t;
  */
 #define	XFS_IBT_MAGIC		0x49414254	/* 'IABT' */
 #define	XFS_IBT_CRC_MAGIC	0x49414233	/* 'IAB3' */
+#define	XFS_FIBT_MAGIC		0x46494254	/* 'FIBT' */
+#define	XFS_FIBT_CRC_MAGIC	0x46494233	/* 'FIB3' */
 
 typedef	__uint64_t	xfs_inofree_t;
 #define	XFS_INODES_PER_CHUNK		(NBBY * sizeof(xfs_inofree_t))
@@ -244,7 +246,17 @@ typedef __be32 xfs_inobt_ptr_t;
  * block numbers in the AG.
  */
 #define	XFS_IBT_BLOCK(mp)		((xfs_agblock_t)(XFS_CNT_BLOCK(mp) + 1))
-#define	XFS_PREALLOC_BLOCKS(mp)		((xfs_agblock_t)(XFS_IBT_BLOCK(mp) + 1))
+#define	XFS_FIBT_BLOCK(mp)		((xfs_agblock_t)(XFS_IBT_BLOCK(mp) + 1))
+
+/*
+ * The first data block of an AG depends on whether the filesystem was formatted
+ * with the finobt feature. If so, account for the finobt reserved root btree
+ * block.
+ */
+#define XFS_PREALLOC_BLOCKS(mp) \
+	(xfs_sb_version_hasfinobt(&((mp)->m_sb)) ? \
+	 XFS_FIBT_BLOCK(mp) + 1 : \
+	 XFS_IBT_BLOCK(mp) + 1)
 
 
 

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_FINOBT	0x20000	/* free inode btree */
 
 /*
  * Minimum and maximum sizes need for growth checks.

fs/xfs/xfs_fsops.c

@@ -24,6 +24,8 @@
 #include "xfs_sb.h"
 #include "xfs_ag.h"
 #include "xfs_mount.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
 #include "xfs_inode.h"
 #include "xfs_trans.h"
 #include "xfs_inode_item.h"
@@ -74,23 +76,18 @@ xfs_fs_geometry(
 	}
 	if (new_version >= 3) {
 		geo->version = XFS_FSOP_GEOM_VERSION;
-		geo->flags =
+		geo->flags = XFS_FSOP_GEOM_FLAGS_NLINK |
+			     XFS_FSOP_GEOM_FLAGS_DIRV2 |
 			(xfs_sb_version_hasattr(&mp->m_sb) ?
 				XFS_FSOP_GEOM_FLAGS_ATTR : 0) |
-			(xfs_sb_version_hasnlink(&mp->m_sb) ?
-				XFS_FSOP_GEOM_FLAGS_NLINK : 0) |
 			(xfs_sb_version_hasquota(&mp->m_sb) ?
 				XFS_FSOP_GEOM_FLAGS_QUOTA : 0) |
 			(xfs_sb_version_hasalign(&mp->m_sb) ?
 				XFS_FSOP_GEOM_FLAGS_IALIGN : 0) |
 			(xfs_sb_version_hasdalign(&mp->m_sb) ?
 				XFS_FSOP_GEOM_FLAGS_DALIGN : 0) |
-			(xfs_sb_version_hasshared(&mp->m_sb) ?
-				XFS_FSOP_GEOM_FLAGS_SHARED : 0) |
 			(xfs_sb_version_hasextflgbit(&mp->m_sb) ?
 				XFS_FSOP_GEOM_FLAGS_EXTFLG : 0) |
-			(xfs_sb_version_hasdirv2(&mp->m_sb) ?
-				XFS_FSOP_GEOM_FLAGS_DIRV2 : 0) |
 			(xfs_sb_version_hassector(&mp->m_sb) ?
 				XFS_FSOP_GEOM_FLAGS_SECTOR : 0) |
 			(xfs_sb_version_hasasciici(&mp->m_sb) ?
@@ -104,11 +101,13 @@ 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_hasfinobt(&mp->m_sb) ?
+				XFS_FSOP_GEOM_FLAGS_FINOBT : 0);
 		geo->logsectsize = xfs_sb_version_hassector(&mp->m_sb) ?
 				mp->m_sb.sb_logsectsize : BBSIZE;
 		geo->rtsectsize = mp->m_sb.sb_blocksize;
-		geo->dirblocksize = mp->m_dirblksize;
+		geo->dirblocksize = mp->m_dir_geo->blksize;
 	}
 	if (new_version >= 4) {
 		geo->flags |=
@@ -316,6 +315,10 @@ xfs_growfs_data_private(
 		agi->agi_dirino = cpu_to_be32(NULLAGINO);
 		if (xfs_sb_version_hascrc(&mp->m_sb))
 			uuid_copy(&agi->agi_uuid, &mp->m_sb.sb_uuid);
+		if (xfs_sb_version_hasfinobt(&mp->m_sb)) {
+			agi->agi_free_root = cpu_to_be32(XFS_FIBT_BLOCK(mp));
+			agi->agi_free_level = cpu_to_be32(1);
+		}
 		for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++)
 			agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO);
 
@@ -407,6 +410,34 @@ xfs_growfs_data_private(
 		xfs_buf_relse(bp);
 		if (error)
 			goto error0;
+
+		/*
+		 * FINO btree root block
+		 */
+		if (xfs_sb_version_hasfinobt(&mp->m_sb)) {
+			bp = xfs_growfs_get_hdr_buf(mp,
+				XFS_AGB_TO_DADDR(mp, agno, XFS_FIBT_BLOCK(mp)),
+				BTOBB(mp->m_sb.sb_blocksize), 0,
+				&xfs_inobt_buf_ops);
+			if (!bp) {
+				error = ENOMEM;
+				goto error0;
+			}
+
+			if (xfs_sb_version_hascrc(&mp->m_sb))
+				xfs_btree_init_block(mp, bp, XFS_FIBT_CRC_MAGIC,
+						     0, 0, agno,
+						     XFS_BTREE_CRC_BLOCKS);
+			else
+				xfs_btree_init_block(mp, bp, XFS_FIBT_MAGIC, 0,
+						     0, agno, 0);
+
+			error = xfs_bwrite(bp);
+			xfs_buf_relse(bp);
+			if (error)
+				goto error0;
+		}
+
 	}
 	xfs_trans_agblocks_delta(tp, nfree);
 	/*

fs/xfs/xfs_ialloc.c

@@ -111,6 +111,66 @@ xfs_inobt_get_rec(
 	return error;
 }
 
+/*
+ * Insert a single inobt record. Cursor must already point to desired location.
+ */
+STATIC int
+xfs_inobt_insert_rec(
+	struct xfs_btree_cur	*cur,
+	__int32_t		freecount,
+	xfs_inofree_t		free,
+	int			*stat)
+{
+	cur->bc_rec.i.ir_freecount = freecount;
+	cur->bc_rec.i.ir_free = free;
+	return xfs_btree_insert(cur, stat);
+}
+
+/*
+ * Insert records describing a newly allocated inode chunk into the inobt.
+ */
+STATIC int
+xfs_inobt_insert(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
+	xfs_agino_t		newino,
+	xfs_agino_t		newlen,
+	xfs_btnum_t		btnum)
+{
+	struct xfs_btree_cur	*cur;
+	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agbp);
+	xfs_agnumber_t		agno = be32_to_cpu(agi->agi_seqno);
+	xfs_agino_t		thisino;
+	int			i;
+	int			error;
+
+	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, btnum);
+
+	for (thisino = newino;
+	     thisino < newino + newlen;
+	     thisino += XFS_INODES_PER_CHUNK) {
+		error = xfs_inobt_lookup(cur, thisino, XFS_LOOKUP_EQ, &i);
+		if (error) {
+			xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
+			return error;
+		}
+		ASSERT(i == 0);
+
+		error = xfs_inobt_insert_rec(cur, XFS_INODES_PER_CHUNK,
+					     XFS_INOBT_ALL_FREE, &i);
+		if (error) {
+			xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
+			return error;
+		}
+		ASSERT(i == 1);
+	}
+
+	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
+
+	return 0;
+}
+
 /*
  * Verify that the number of free inodes in the AGI is correct.
  */
@@ -220,10 +280,8 @@ xfs_ialloc_inode_init(
 		if (tp)
 			xfs_icreate_log(tp, agno, agbno, mp->m_ialloc_inos,
 					mp->m_sb.sb_inodesize, length, gen);
-	} else if (xfs_sb_version_hasnlink(&mp->m_sb))
+	} else
 		version = 2;
-	else
-		version = 1;
 
 	for (j = 0; j < nbufs; j++) {
 		/*
@@ -303,13 +361,10 @@ xfs_ialloc_ag_alloc(
 {
 	xfs_agi_t	*agi;		/* allocation group header */
 	xfs_alloc_arg_t	args;		/* allocation argument structure */
-	xfs_btree_cur_t	*cur;		/* inode btree cursor */
 	xfs_agnumber_t	agno;
 	int		error;
-	int		i;
 	xfs_agino_t	newino;		/* new first inode's number */
 	xfs_agino_t	newlen;		/* new number of inodes */
-	xfs_agino_t	thisino;	/* current inode number, for loop */
 	int		isaligned = 0;	/* inode allocation at stripe unit */
 					/* boundary */
 	struct xfs_perag *pag;
@@ -459,29 +514,19 @@ xfs_ialloc_ag_alloc(
 	agi->agi_newino = cpu_to_be32(newino);
 
 	/*
-	 * Insert records describing the new inode chunk into the btree.
+	 * Insert records describing the new inode chunk into the btrees.
 	 */
-	cur = xfs_inobt_init_cursor(args.mp, tp, agbp, agno);
-	for (thisino = newino;
-	     thisino < newino + newlen;
-	     thisino += XFS_INODES_PER_CHUNK) {
-		cur->bc_rec.i.ir_startino = thisino;
-		cur->bc_rec.i.ir_freecount = XFS_INODES_PER_CHUNK;
-		cur->bc_rec.i.ir_free = XFS_INOBT_ALL_FREE;
-		error = xfs_btree_lookup(cur, XFS_LOOKUP_EQ, &i);
-		if (error) {
-			xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
-			return error;
-		}
-		ASSERT(i == 0);
-		error = xfs_btree_insert(cur, &i);
-		if (error) {
-			xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
+	error = xfs_inobt_insert(args.mp, tp, agbp, newino, newlen,
+				 XFS_BTNUM_INO);
+	if (error)
+		return error;
+
+	if (xfs_sb_version_hasfinobt(&args.mp->m_sb)) {
+		error = xfs_inobt_insert(args.mp, tp, agbp, newino, newlen,
+					 XFS_BTNUM_FINO);
+		if (error)
 			return error;
-		}
-		ASSERT(i == 1);
 	}
-	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 	/*
 	 * Log allocation group header fields
 	 */
@@ -675,13 +720,10 @@ xfs_ialloc_get_rec(
 }
 
 /*
- * Allocate an inode.
- *
- * The caller selected an AG for us, and made sure that free inodes are
- * available.
+ * Allocate an inode using the inobt-only algorithm.
  */
 STATIC int
-xfs_dialloc_ag(
+xfs_dialloc_ag_inobt(
 	struct xfs_trans	*tp,
 	struct xfs_buf		*agbp,
 	xfs_ino_t		parent,
@@ -707,7 +749,7 @@ xfs_dialloc_ag(
 	ASSERT(pag->pagi_freecount > 0);
 
  restart_pagno:
-	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
+	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO);
 	/*
 	 * If pagino is 0 (this is the root inode allocation) use newino.
 	 * This must work because we've just allocated some.
@@ -939,6 +981,294 @@ error0:
 	return error;
 }
 
+/*
+ * Use the free inode btree to allocate an inode based on distance from the
+ * parent. Note that the provided cursor may be deleted and replaced.
+ */
+STATIC int
+xfs_dialloc_ag_finobt_near(
+	xfs_agino_t			pagino,
+	struct xfs_btree_cur		**ocur,
+	struct xfs_inobt_rec_incore	*rec)
+{
+	struct xfs_btree_cur		*lcur = *ocur;	/* left search cursor */
+	struct xfs_btree_cur		*rcur;	/* right search cursor */
+	struct xfs_inobt_rec_incore	rrec;
+	int				error;
+	int				i, j;
+
+	error = xfs_inobt_lookup(lcur, pagino, XFS_LOOKUP_LE, &i);
+	if (error)
+		return error;
+
+	if (i == 1) {
+		error = xfs_inobt_get_rec(lcur, rec, &i);
+		if (error)
+			return error;
+		XFS_WANT_CORRUPTED_RETURN(i == 1);
+
+		/*
+		 * See if we've landed in the parent inode record. The finobt
+		 * only tracks chunks with at least one free inode, so record
+		 * existence is enough.
+		 */
+		if (pagino >= rec->ir_startino &&
+		    pagino < (rec->ir_startino + XFS_INODES_PER_CHUNK))
+			return 0;
+	}
+
+	error = xfs_btree_dup_cursor(lcur, &rcur);
+	if (error)
+		return error;
+
+	error = xfs_inobt_lookup(rcur, pagino, XFS_LOOKUP_GE, &j);
+	if (error)
+		goto error_rcur;
+	if (j == 1) {
+		error = xfs_inobt_get_rec(rcur, &rrec, &j);
+		if (error)
+			goto error_rcur;
+		XFS_WANT_CORRUPTED_GOTO(j == 1, error_rcur);
+	}
+
+	XFS_WANT_CORRUPTED_GOTO(i == 1 || j == 1, error_rcur);
+	if (i == 1 && j == 1) {
+		/*
+		 * Both the left and right records are valid. Choose the closer
+		 * inode chunk to the target.
+		 */
+		if ((pagino - rec->ir_startino + XFS_INODES_PER_CHUNK - 1) >
+		    (rrec.ir_startino - pagino)) {
+			*rec = rrec;
+			xfs_btree_del_cursor(lcur, XFS_BTREE_NOERROR);
+			*ocur = rcur;
+		} else {
+			xfs_btree_del_cursor(rcur, XFS_BTREE_NOERROR);
+		}
+	} else if (j == 1) {
+		/* only the right record is valid */
+		*rec = rrec;
+		xfs_btree_del_cursor(lcur, XFS_BTREE_NOERROR);
+		*ocur = rcur;
+	} else if (i == 1) {
+		/* only the left record is valid */
+		xfs_btree_del_cursor(rcur, XFS_BTREE_NOERROR);
+	}
+
+	return 0;
+
+error_rcur:
+	xfs_btree_del_cursor(rcur, XFS_BTREE_ERROR);
+	return error;
+}
+
+/*
+ * Use the free inode btree to find a free inode based on a newino hint. If
+ * the hint is NULL, find the first free inode in the AG.
+ */
+STATIC int
+xfs_dialloc_ag_finobt_newino(
+	struct xfs_agi			*agi,
+	struct xfs_btree_cur		*cur,
+	struct xfs_inobt_rec_incore	*rec)
+{
+	int error;
+	int i;
+
+	if (agi->agi_newino != cpu_to_be32(NULLAGINO)) {
+		error = xfs_inobt_lookup(cur, agi->agi_newino, XFS_LOOKUP_EQ,
+					 &i);
+		if (error)
+			return error;
+		if (i == 1) {
+			error = xfs_inobt_get_rec(cur, rec, &i);
+			if (error)
+				return error;
+			XFS_WANT_CORRUPTED_RETURN(i == 1);
+
+			return 0;
+		}
+	}
+
+	/*
+	 * Find the first inode available in the AG.
+	 */
+	error = xfs_inobt_lookup(cur, 0, XFS_LOOKUP_GE, &i);
+	if (error)
+		return error;
+	XFS_WANT_CORRUPTED_RETURN(i == 1);
+
+	error = xfs_inobt_get_rec(cur, rec, &i);
+	if (error)
+		return error;
+	XFS_WANT_CORRUPTED_RETURN(i == 1);
+
+	return 0;
+}
+
+/*
+ * Update the inobt based on a modification made to the finobt. Also ensure that
+ * the records from both trees are equivalent post-modification.
+ */
+STATIC int
+xfs_dialloc_ag_update_inobt(
+	struct xfs_btree_cur		*cur,	/* inobt cursor */
+	struct xfs_inobt_rec_incore	*frec,	/* finobt record */
+	int				offset) /* inode offset */
+{
+	struct xfs_inobt_rec_incore	rec;
+	int				error;
+	int				i;
+
+	error = xfs_inobt_lookup(cur, frec->ir_startino, XFS_LOOKUP_EQ, &i);
+	if (error)
+		return error;
+	XFS_WANT_CORRUPTED_RETURN(i == 1);
+
+	error = xfs_inobt_get_rec(cur, &rec, &i);
+	if (error)
+		return error;
+	XFS_WANT_CORRUPTED_RETURN(i == 1);
+	ASSERT((XFS_AGINO_TO_OFFSET(cur->bc_mp, rec.ir_startino) %
+				   XFS_INODES_PER_CHUNK) == 0);
+
+	rec.ir_free &= ~XFS_INOBT_MASK(offset);
+	rec.ir_freecount--;
+
+	XFS_WANT_CORRUPTED_RETURN((rec.ir_free == frec->ir_free) &&
+				  (rec.ir_freecount == frec->ir_freecount));
+
+	error = xfs_inobt_update(cur, &rec);
+	if (error)
+		return error;
+
+	return 0;
+}
+
+/*
+ * Allocate an inode using the free inode btree, if available. Otherwise, fall
+ * back to the inobt search algorithm.
+ *
+ * The caller selected an AG for us, and made sure that free inodes are
+ * available.
+ */
+STATIC int
+xfs_dialloc_ag(
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
+	xfs_ino_t		parent,
+	xfs_ino_t		*inop)
+{
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_agi			*agi = XFS_BUF_TO_AGI(agbp);
+	xfs_agnumber_t			agno = be32_to_cpu(agi->agi_seqno);
+	xfs_agnumber_t			pagno = XFS_INO_TO_AGNO(mp, parent);
+	xfs_agino_t			pagino = XFS_INO_TO_AGINO(mp, parent);
+	struct xfs_perag		*pag;
+	struct xfs_btree_cur		*cur;	/* finobt cursor */
+	struct xfs_btree_cur		*icur;	/* inobt cursor */
+	struct xfs_inobt_rec_incore	rec;
+	xfs_ino_t			ino;
+	int				error;
+	int				offset;
+	int				i;
+
+	if (!xfs_sb_version_hasfinobt(&mp->m_sb))
+		return xfs_dialloc_ag_inobt(tp, agbp, parent, inop);
+
+	pag = xfs_perag_get(mp, agno);
+
+	/*
+	 * If pagino is 0 (this is the root inode allocation) use newino.
+	 * This must work because we've just allocated some.
+	 */
+	if (!pagino)
+		pagino = be32_to_cpu(agi->agi_newino);
+
+	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_FINO);
+
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error_cur;
+
+	/*
+	 * The search algorithm depends on whether we're in the same AG as the
+	 * parent. If so, find the closest available inode to the parent. If
+	 * not, consider the agi hint or find the first free inode in the AG.
+	 */
+	if (agno == pagno)
+		error = xfs_dialloc_ag_finobt_near(pagino, &cur, &rec);
+	else
+		error = xfs_dialloc_ag_finobt_newino(agi, cur, &rec);
+	if (error)
+		goto error_cur;
+
+	offset = xfs_lowbit64(rec.ir_free);
+	ASSERT(offset >= 0);
+	ASSERT(offset < XFS_INODES_PER_CHUNK);
+	ASSERT((XFS_AGINO_TO_OFFSET(mp, rec.ir_startino) %
+				   XFS_INODES_PER_CHUNK) == 0);
+	ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino + offset);
+
+	/*
+	 * Modify or remove the finobt record.
+	 */
+	rec.ir_free &= ~XFS_INOBT_MASK(offset);
+	rec.ir_freecount--;
+	if (rec.ir_freecount)
+		error = xfs_inobt_update(cur, &rec);
+	else
+		error = xfs_btree_delete(cur, &i);
+	if (error)
+		goto error_cur;
+
+	/*
+	 * The finobt has now been updated appropriately. We haven't updated the
+	 * agi and superblock yet, so we can create an inobt cursor and validate
+	 * the original freecount. If all is well, make the equivalent update to
+	 * the inobt using the finobt record and offset information.
+	 */
+	icur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO);
+
+	error = xfs_check_agi_freecount(icur, agi);
+	if (error)
+		goto error_icur;
+
+	error = xfs_dialloc_ag_update_inobt(icur, &rec, offset);
+	if (error)
+		goto error_icur;
+
+	/*
+	 * Both trees have now been updated. We must update the perag and
+	 * superblock before we can check the freecount for each btree.
+	 */
+	be32_add_cpu(&agi->agi_freecount, -1);
+	xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
+	pag->pagi_freecount--;
+
+	xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
+
+	error = xfs_check_agi_freecount(icur, agi);
+	if (error)
+		goto error_icur;
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error_icur;
+
+	xfs_btree_del_cursor(icur, XFS_BTREE_NOERROR);
+	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
+	xfs_perag_put(pag);
+	*inop = ino;
+	return 0;
+
+error_icur:
+	xfs_btree_del_cursor(icur, XFS_BTREE_ERROR);
+error_cur:
+	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
+	xfs_perag_put(pag);
+	return error;
+}
+
 /*
  * Allocate an inode on disk.
  *
@@ -1098,78 +1428,34 @@ out_error:
 	return XFS_ERROR(error);
 }
 
-/*
- * Free disk inode.  Carefully avoids touching the incore inode, all
- * manipulations incore are the caller's responsibility.
- * The on-disk inode is not changed by this operation, only the
- * btree (free inode mask) is changed.
- */
-int
-xfs_difree(
-	xfs_trans_t	*tp,		/* transaction pointer */
-	xfs_ino_t	inode,		/* inode to be freed */
-	xfs_bmap_free_t	*flist,		/* extents to free */
-	int		*delete,	/* set if inode cluster was deleted */
-	xfs_ino_t	*first_ino)	/* first inode in deleted cluster */
+STATIC int
+xfs_difree_inobt(
+	struct xfs_mount		*mp,
+	struct xfs_trans		*tp,
+	struct xfs_buf			*agbp,
+	xfs_agino_t			agino,
+	struct xfs_bmap_free		*flist,
+	int				*deleted,
+	xfs_ino_t			*first_ino,
+	struct xfs_inobt_rec_incore	*orec)
 {
-	/* REFERENCED */
-	xfs_agblock_t	agbno;	/* block number containing inode */
-	xfs_buf_t	*agbp;	/* buffer containing allocation group header */
-	xfs_agino_t	agino;	/* inode number relative to allocation group */
-	xfs_agnumber_t	agno;	/* allocation group number */
-	xfs_agi_t	*agi;	/* allocation group header */
-	xfs_btree_cur_t	*cur;	/* inode btree cursor */
-	int		error;	/* error return value */
-	int		i;	/* result code */
-	int		ilen;	/* inodes in an inode cluster */
-	xfs_mount_t	*mp;	/* mount structure for filesystem */
-	int		off;	/* offset of inode in inode chunk */
-	xfs_inobt_rec_incore_t rec;	/* btree record */
-	struct xfs_perag *pag;
+	struct xfs_agi			*agi = XFS_BUF_TO_AGI(agbp);
+	xfs_agnumber_t			agno = be32_to_cpu(agi->agi_seqno);
+	struct xfs_perag		*pag;
+	struct xfs_btree_cur		*cur;
+	struct xfs_inobt_rec_incore	rec;
+	int				ilen;
+	int				error;
+	int				i;
+	int				off;
 
-	mp = tp->t_mountp;
-
-	/*
-	 * Break up inode number into its components.
-	 */
-	agno = XFS_INO_TO_AGNO(mp, inode);
-	if (agno >= mp->m_sb.sb_agcount)  {
-		xfs_warn(mp, "%s: agno >= mp->m_sb.sb_agcount (%d >= %d).",
-			__func__, agno, mp->m_sb.sb_agcount);
-		ASSERT(0);
-		return XFS_ERROR(EINVAL);
-	}
-	agino = XFS_INO_TO_AGINO(mp, inode);
-	if (inode != XFS_AGINO_TO_INO(mp, agno, agino))  {
-		xfs_warn(mp, "%s: inode != XFS_AGINO_TO_INO() (%llu != %llu).",
-			__func__, (unsigned long long)inode,
-			(unsigned long long)XFS_AGINO_TO_INO(mp, agno, agino));
-		ASSERT(0);
-		return XFS_ERROR(EINVAL);
-	}
-	agbno = XFS_AGINO_TO_AGBNO(mp, agino);
-	if (agbno >= mp->m_sb.sb_agblocks)  {
-		xfs_warn(mp, "%s: agbno >= mp->m_sb.sb_agblocks (%d >= %d).",
-			__func__, agbno, mp->m_sb.sb_agblocks);
-		ASSERT(0);
-		return XFS_ERROR(EINVAL);
-	}
-	/*
-	 * Get the allocation group header.
-	 */
-	error = xfs_ialloc_read_agi(mp, tp, agno, &agbp);
-	if (error) {
-		xfs_warn(mp, "%s: xfs_ialloc_read_agi() returned error %d.",
-			__func__, error);
-		return error;
-	}
-	agi = XFS_BUF_TO_AGI(agbp);
 	ASSERT(agi->agi_magicnum == cpu_to_be32(XFS_AGI_MAGIC));
-	ASSERT(agbno < be32_to_cpu(agi->agi_length));
+	ASSERT(XFS_AGINO_TO_AGBNO(mp, agino) < be32_to_cpu(agi->agi_length));
+
 	/*
 	 * Initialize the cursor.
 	 */
-	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
+	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO);
 
 	error = xfs_check_agi_freecount(cur, agi);
 	if (error)
@@ -1209,7 +1495,7 @@ xfs_difree(
 	if (!(mp->m_flags & XFS_MOUNT_IKEEP) &&
 	    (rec.ir_freecount == mp->m_ialloc_inos)) {
 
-		*delete = 1;
+		*deleted = 1;
 		*first_ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino);
 
 		/*
@@ -1237,7 +1523,7 @@ xfs_difree(
 				  XFS_AGINO_TO_AGBNO(mp, rec.ir_startino)),
 				  mp->m_ialloc_blks, flist, mp);
 	} else {
-		*delete = 0;
+		*deleted = 0;
 
 		error = xfs_inobt_update(cur, &rec);
 		if (error) {
@@ -1261,6 +1547,7 @@ xfs_difree(
 	if (error)
 		goto error0;
 
+	*orec = rec;
 	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 	return 0;
 
@@ -1269,6 +1556,182 @@ error0:
 	return error;
 }
 
+/*
+ * Free an inode in the free inode btree.
+ */
+STATIC int
+xfs_difree_finobt(
+	struct xfs_mount		*mp,
+	struct xfs_trans		*tp,
+	struct xfs_buf			*agbp,
+	xfs_agino_t			agino,
+	struct xfs_inobt_rec_incore	*ibtrec) /* inobt record */
+{
+	struct xfs_agi			*agi = XFS_BUF_TO_AGI(agbp);
+	xfs_agnumber_t			agno = be32_to_cpu(agi->agi_seqno);
+	struct xfs_btree_cur		*cur;
+	struct xfs_inobt_rec_incore	rec;
+	int				offset = agino - ibtrec->ir_startino;
+	int				error;
+	int				i;
+
+	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_FINO);
+
+	error = xfs_inobt_lookup(cur, ibtrec->ir_startino, XFS_LOOKUP_EQ, &i);
+	if (error)
+		goto error;
+	if (i == 0) {
+		/*
+		 * If the record does not exist in the finobt, we must have just
+		 * freed an inode in a previously fully allocated chunk. If not,
+		 * something is out of sync.
+		 */
+		XFS_WANT_CORRUPTED_GOTO(ibtrec->ir_freecount == 1, error);
+
+		error = xfs_inobt_insert_rec(cur, ibtrec->ir_freecount,
+					     ibtrec->ir_free, &i);
+		if (error)
+			goto error;
+		ASSERT(i == 1);
+
+		goto out;
+	}
+
+	/*
+	 * Read and update the existing record. We could just copy the ibtrec
+	 * across here, but that would defeat the purpose of having redundant
+	 * metadata. By making the modifications independently, we can catch
+	 * corruptions that we wouldn't see if we just copied from one record
+	 * to another.
+	 */
+	error = xfs_inobt_get_rec(cur, &rec, &i);
+	if (error)
+		goto error;
+	XFS_WANT_CORRUPTED_GOTO(i == 1, error);
+
+	rec.ir_free |= XFS_INOBT_MASK(offset);
+	rec.ir_freecount++;
+
+	XFS_WANT_CORRUPTED_GOTO((rec.ir_free == ibtrec->ir_free) &&
+				(rec.ir_freecount == ibtrec->ir_freecount),
+				error);
+
+	/*
+	 * The content of inobt records should always match between the inobt
+	 * and finobt. The lifecycle of records in the finobt is different from
+	 * the inobt in that the finobt only tracks records with at least one
+	 * free inode. Hence, if all of the inodes are free and we aren't
+	 * keeping inode chunks permanently on disk, remove the record.
+	 * Otherwise, update the record with the new information.
+	 */
+	if (rec.ir_freecount == mp->m_ialloc_inos &&
+	    !(mp->m_flags & XFS_MOUNT_IKEEP)) {
+		error = xfs_btree_delete(cur, &i);
+		if (error)
+			goto error;
+		ASSERT(i == 1);
+	} else {
+		error = xfs_inobt_update(cur, &rec);
+		if (error)
+			goto error;
+	}
+
+out:
+	error = xfs_check_agi_freecount(cur, agi);
+	if (error)
+		goto error;
+
+	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
+	return 0;
+
+error:
+	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
+	return error;
+}
+
+/*
+ * Free disk inode.  Carefully avoids touching the incore inode, all
+ * manipulations incore are the caller's responsibility.
+ * The on-disk inode is not changed by this operation, only the
+ * btree (free inode mask) is changed.
+ */
+int
+xfs_difree(
+	struct xfs_trans	*tp,		/* transaction pointer */
+	xfs_ino_t		inode,		/* inode to be freed */
+	struct xfs_bmap_free	*flist,		/* extents to free */
+	int			*deleted,/* set if inode cluster was deleted */
+	xfs_ino_t		*first_ino)/* first inode in deleted cluster */
+{
+	/* REFERENCED */
+	xfs_agblock_t		agbno;	/* block number containing inode */
+	struct xfs_buf		*agbp;	/* buffer for allocation group header */
+	xfs_agino_t		agino;	/* allocation group inode number */
+	xfs_agnumber_t		agno;	/* allocation group number */
+	int			error;	/* error return value */
+	struct xfs_mount	*mp;	/* mount structure for filesystem */
+	struct xfs_inobt_rec_incore rec;/* btree record */
+
+	mp = tp->t_mountp;
+
+	/*
+	 * Break up inode number into its components.
+	 */
+	agno = XFS_INO_TO_AGNO(mp, inode);
+	if (agno >= mp->m_sb.sb_agcount)  {
+		xfs_warn(mp, "%s: agno >= mp->m_sb.sb_agcount (%d >= %d).",
+			__func__, agno, mp->m_sb.sb_agcount);
+		ASSERT(0);
+		return XFS_ERROR(EINVAL);
+	}
+	agino = XFS_INO_TO_AGINO(mp, inode);
+	if (inode != XFS_AGINO_TO_INO(mp, agno, agino))  {
+		xfs_warn(mp, "%s: inode != XFS_AGINO_TO_INO() (%llu != %llu).",
+			__func__, (unsigned long long)inode,
+			(unsigned long long)XFS_AGINO_TO_INO(mp, agno, agino));
+		ASSERT(0);
+		return XFS_ERROR(EINVAL);
+	}
+	agbno = XFS_AGINO_TO_AGBNO(mp, agino);
+	if (agbno >= mp->m_sb.sb_agblocks)  {
+		xfs_warn(mp, "%s: agbno >= mp->m_sb.sb_agblocks (%d >= %d).",
+			__func__, agbno, mp->m_sb.sb_agblocks);
+		ASSERT(0);
+		return XFS_ERROR(EINVAL);
+	}
+	/*
+	 * Get the allocation group header.
+	 */
+	error = xfs_ialloc_read_agi(mp, tp, agno, &agbp);
+	if (error) {
+		xfs_warn(mp, "%s: xfs_ialloc_read_agi() returned error %d.",
+			__func__, error);
+		return error;
+	}
+
+	/*
+	 * Fix up the inode allocation btree.
+	 */
+	error = xfs_difree_inobt(mp, tp, agbp, agino, flist, deleted, first_ino,
+				 &rec);
+	if (error)
+		goto error0;
+
+	/*
+	 * Fix up the free inode btree.
+	 */
+	if (xfs_sb_version_hasfinobt(&mp->m_sb)) {
+		error = xfs_difree_finobt(mp, tp, agbp, agino, &rec);
+		if (error)
+			goto error0;
+	}
+
+	return 0;
+
+error0:
+	return error;
+}
+
 STATIC int
 xfs_imap_lookup(
 	struct xfs_mount	*mp,
@@ -1300,7 +1763,7 @@ xfs_imap_lookup(
 	 * we have a record, we need to ensure it contains the inode number
 	 * we are looking up.
 	 */
-	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno);
+	cur = xfs_inobt_init_cursor(mp, tp, agbp, agno, XFS_BTNUM_INO);
 	error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_LE, &i);
 	if (!error) {
 		if (i)
@@ -1488,7 +1951,16 @@ xfs_ialloc_compute_maxlevels(
 }
 
 /*
- * Log specified fields for the ag hdr (inode section)
+ * Log specified fields for the ag hdr (inode section). The growth of the agi
+ * structure over time requires that we interpret the buffer as two logical
+ * regions delineated by the end of the unlinked list. This is due to the size
+ * of the hash table and its location in the middle of the agi.
+ *
+ * For example, a request to log a field before agi_unlinked and a field after
+ * agi_unlinked could cause us to log the entire hash table and use an excessive
+ * amount of log space. To avoid this behavior, log the region up through
+ * agi_unlinked in one call and the region after agi_unlinked through the end of
+ * the structure in another.
  */
 void
 xfs_ialloc_log_agi(
@@ -1511,6 +1983,8 @@ xfs_ialloc_log_agi(
 		offsetof(xfs_agi_t, agi_newino),
 		offsetof(xfs_agi_t, agi_dirino),
 		offsetof(xfs_agi_t, agi_unlinked),
+		offsetof(xfs_agi_t, agi_free_root),
+		offsetof(xfs_agi_t, agi_free_level),
 		sizeof(xfs_agi_t)
 	};
 #ifdef DEBUG
@@ -1519,15 +1993,30 @@ xfs_ialloc_log_agi(
 	agi = XFS_BUF_TO_AGI(bp);
 	ASSERT(agi->agi_magicnum == cpu_to_be32(XFS_AGI_MAGIC));
 #endif
+
+	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_AGI_BUF);
+
 	/*
-	 * Compute byte offsets for the first and last fields.
+	 * Compute byte offsets for the first and last fields in the first
+	 * region and log the agi buffer. This only logs up through
+	 * agi_unlinked.
 	 */
-	xfs_btree_offsets(fields, offsets, XFS_AGI_NUM_BITS, &first, &last);
+	if (fields & XFS_AGI_ALL_BITS_R1) {
+		xfs_btree_offsets(fields, offsets, XFS_AGI_NUM_BITS_R1,
+				  &first, &last);
+		xfs_trans_log_buf(tp, bp, first, last);
+	}
+
 	/*
-	 * Log the allocation group inode header buffer.
+	 * Mask off the bits in the first region and calculate the first and
+	 * last field offsets for any bits in the second region.
 	 */
-	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_AGI_BUF);
-	xfs_trans_log_buf(tp, bp, first, last);
+	fields &= ~XFS_AGI_ALL_BITS_R1;
+	if (fields) {
+		xfs_btree_offsets(fields, offsets, XFS_AGI_NUM_BITS_R2,
+				  &first, &last);
+		xfs_trans_log_buf(tp, bp, first, last);
+	}
 }
 
 #ifdef DEBUG
@@ -1640,7 +2129,6 @@ xfs_read_agi(
 	if (error)
 		return error;
 
-	ASSERT(!xfs_buf_geterror(*bpp));
 	xfs_buf_set_ref(*bpp, XFS_AGI_REF);
 	return 0;
 }

fs/xfs/xfs_ialloc.h

@@ -90,7 +90,7 @@ xfs_difree(
 	struct xfs_trans *tp,		/* transaction pointer */
 	xfs_ino_t	inode,		/* inode to be freed */
 	struct xfs_bmap_free *flist,	/* extents to free */
-	int		*delete,	/* set if inode cluster was deleted */
+	int		*deleted,	/* set if inode cluster was deleted */
 	xfs_ino_t	*first_ino);	/* first inode in deleted cluster */
 
 /*

fs/xfs/xfs_ialloc_btree.c

@@ -49,7 +49,8 @@ xfs_inobt_dup_cursor(
 	struct xfs_btree_cur	*cur)
 {
 	return xfs_inobt_init_cursor(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agbp, cur->bc_private.a.agno);
+			cur->bc_private.a.agbp, cur->bc_private.a.agno,
+			cur->bc_btnum);
 }
 
 STATIC void
@@ -66,12 +67,26 @@ xfs_inobt_set_root(
 	xfs_ialloc_log_agi(cur->bc_tp, agbp, XFS_AGI_ROOT | XFS_AGI_LEVEL);
 }
 
+STATIC void
+xfs_finobt_set_root(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*nptr,
+	int			inc)	/* level change */
+{
+	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
+	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agbp);
+
+	agi->agi_free_root = nptr->s;
+	be32_add_cpu(&agi->agi_free_level, inc);
+	xfs_ialloc_log_agi(cur->bc_tp, agbp,
+			   XFS_AGI_FREE_ROOT | XFS_AGI_FREE_LEVEL);
+}
+
 STATIC int
 xfs_inobt_alloc_block(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*start,
 	union xfs_btree_ptr	*new,
-	int			length,
 	int			*stat)
 {
 	xfs_alloc_arg_t		args;		/* block allocation args */
@@ -173,6 +188,17 @@ xfs_inobt_init_ptr_from_cur(
 	ptr->s = agi->agi_root;
 }
 
+STATIC void
+xfs_finobt_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	struct xfs_agi		*agi = XFS_BUF_TO_AGI(cur->bc_private.a.agbp);
+
+	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agi->agi_seqno));
+	ptr->s = agi->agi_free_root;
+}
+
 STATIC __int64_t
 xfs_inobt_key_diff(
 	struct xfs_btree_cur	*cur,
@@ -203,6 +229,7 @@ xfs_inobt_verify(
 	 */
 	switch (block->bb_magic) {
 	case cpu_to_be32(XFS_IBT_CRC_MAGIC):
+	case cpu_to_be32(XFS_FIBT_CRC_MAGIC):
 		if (!xfs_sb_version_hascrc(&mp->m_sb))
 			return false;
 		if (!uuid_equal(&block->bb_u.s.bb_uuid, &mp->m_sb.sb_uuid))
@@ -214,6 +241,7 @@ xfs_inobt_verify(
 			return false;
 		/* fall through */
 	case cpu_to_be32(XFS_IBT_MAGIC):
+	case cpu_to_be32(XFS_FIBT_MAGIC):
 		break;
 	default:
 		return 0;
@@ -317,6 +345,28 @@ static const struct xfs_btree_ops xfs_inobt_ops = {
 #endif
 };
 
+static const struct xfs_btree_ops xfs_finobt_ops = {
+	.rec_len		= sizeof(xfs_inobt_rec_t),
+	.key_len		= sizeof(xfs_inobt_key_t),
+
+	.dup_cursor		= xfs_inobt_dup_cursor,
+	.set_root		= xfs_finobt_set_root,
+	.alloc_block		= xfs_inobt_alloc_block,
+	.free_block		= xfs_inobt_free_block,
+	.get_minrecs		= xfs_inobt_get_minrecs,
+	.get_maxrecs		= xfs_inobt_get_maxrecs,
+	.init_key_from_rec	= xfs_inobt_init_key_from_rec,
+	.init_rec_from_key	= xfs_inobt_init_rec_from_key,
+	.init_rec_from_cur	= xfs_inobt_init_rec_from_cur,
+	.init_ptr_from_cur	= xfs_finobt_init_ptr_from_cur,
+	.key_diff		= xfs_inobt_key_diff,
+	.buf_ops		= &xfs_inobt_buf_ops,
+#if defined(DEBUG) || defined(XFS_WARN)
+	.keys_inorder		= xfs_inobt_keys_inorder,
+	.recs_inorder		= xfs_inobt_recs_inorder,
+#endif
+};
+
 /*
  * Allocate a new inode btree cursor.
  */
@@ -325,7 +375,8 @@ xfs_inobt_init_cursor(
 	struct xfs_mount	*mp,		/* file system mount point */
 	struct xfs_trans	*tp,		/* transaction pointer */
 	struct xfs_buf		*agbp,		/* buffer for agi structure */
-	xfs_agnumber_t		agno)		/* allocation group number */
+	xfs_agnumber_t		agno,		/* allocation group number */
+	xfs_btnum_t		btnum)		/* ialloc or free ino btree */
 {
 	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agbp);
 	struct xfs_btree_cur	*cur;
@@ -334,11 +385,17 @@ xfs_inobt_init_cursor(
 
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
-	cur->bc_nlevels = be32_to_cpu(agi->agi_level);
-	cur->bc_btnum = XFS_BTNUM_INO;
+	cur->bc_btnum = btnum;
+	if (btnum == XFS_BTNUM_INO) {
+		cur->bc_nlevels = be32_to_cpu(agi->agi_level);
+		cur->bc_ops = &xfs_inobt_ops;
+	} else {
+		cur->bc_nlevels = be32_to_cpu(agi->agi_free_level);
+		cur->bc_ops = &xfs_finobt_ops;
+	}
+
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 
-	cur->bc_ops = &xfs_inobt_ops;
 	if (xfs_sb_version_hascrc(&mp->m_sb))
 		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
 

fs/xfs/xfs_ialloc_btree.h

@@ -58,7 +58,8 @@ struct xfs_mount;
 		 ((index) - 1) * sizeof(xfs_inobt_ptr_t)))
 
 extern struct xfs_btree_cur *xfs_inobt_init_cursor(struct xfs_mount *,
-		struct xfs_trans *, struct xfs_buf *, xfs_agnumber_t);
+		struct xfs_trans *, struct xfs_buf *, xfs_agnumber_t,
+		xfs_btnum_t);
 extern int xfs_inobt_maxrecs(struct xfs_mount *, int, int);
 
 #endif	/* __XFS_IALLOC_BTREE_H__ */

fs/xfs/xfs_icache.c

@@ -507,8 +507,7 @@ STATIC int
 xfs_inode_ag_walk(
 	struct xfs_mount	*mp,
 	struct xfs_perag	*pag,
-	int			(*execute)(struct xfs_inode *ip,
-					   struct xfs_perag *pag, int flags,
+	int			(*execute)(struct xfs_inode *ip, int flags,
 					   void *args),
 	int			flags,
 	void			*args,
@@ -582,7 +581,7 @@ restart:
 		for (i = 0; i < nr_found; i++) {
 			if (!batch[i])
 				continue;
-			error = execute(batch[i], pag, flags, args);
+			error = execute(batch[i], flags, args);
 			IRELE(batch[i]);
 			if (error == EAGAIN) {
 				skipped++;
@@ -636,8 +635,7 @@ xfs_eofblocks_worker(
 int
 xfs_inode_ag_iterator(
 	struct xfs_mount	*mp,
-	int			(*execute)(struct xfs_inode *ip,
-					   struct xfs_perag *pag, int flags,
+	int			(*execute)(struct xfs_inode *ip, int flags,
 					   void *args),
 	int			flags,
 	void			*args)
@@ -664,8 +662,7 @@ xfs_inode_ag_iterator(
 int
 xfs_inode_ag_iterator_tag(
 	struct xfs_mount	*mp,
-	int			(*execute)(struct xfs_inode *ip,
-					   struct xfs_perag *pag, int flags,
+	int			(*execute)(struct xfs_inode *ip, int flags,
 					   void *args),
 	int			flags,
 	void			*args,
@@ -1209,7 +1206,6 @@ xfs_inode_match_id(
 STATIC int
 xfs_inode_free_eofblocks(
 	struct xfs_inode	*ip,
-	struct xfs_perag	*pag,
 	int			flags,
 	void			*args)
 {

fs/xfs/xfs_icache.h

@@ -60,12 +60,10 @@ int xfs_icache_free_eofblocks(struct xfs_mount *, struct xfs_eofblocks *);
 void xfs_eofblocks_worker(struct work_struct *);
 
 int xfs_inode_ag_iterator(struct xfs_mount *mp,
-	int (*execute)(struct xfs_inode *ip, struct xfs_perag *pag,
-		int flags, void *args),
+	int (*execute)(struct xfs_inode *ip, int flags, void *args),
 	int flags, void *args);
 int xfs_inode_ag_iterator_tag(struct xfs_mount *mp,
-	int (*execute)(struct xfs_inode *ip, struct xfs_perag *pag,
-		int flags, void *args),
+	int (*execute)(struct xfs_inode *ip, int flags, void *args),
 	int flags, void *args, int tag);
 
 static inline int

fs/xfs/xfs_inode.c

@@ -655,7 +655,6 @@ xfs_ialloc(
 	uint		flags;
 	int		error;
 	timespec_t	tv;
-	int		filestreams = 0;
 
 	/*
 	 * Call the space management code to pick
@@ -682,6 +681,14 @@ xfs_ialloc(
 		return error;
 	ASSERT(ip != NULL);
 
+	/*
+	 * We always convert v1 inodes to v2 now - we only support filesystems
+	 * with >= v2 inode capability, so there is no reason for ever leaving
+	 * an inode in v1 format.
+	 */
+	if (ip->i_d.di_version == 1)
+		ip->i_d.di_version = 2;
+
 	ip->i_d.di_mode = mode;
 	ip->i_d.di_onlink = 0;
 	ip->i_d.di_nlink = nlink;
@@ -691,27 +698,6 @@ xfs_ialloc(
 	xfs_set_projid(ip, prid);
 	memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
 
-	/*
-	 * If the superblock version is up to where we support new format
-	 * inodes and this is currently an old format inode, then change
-	 * the inode version number now.  This way we only do the conversion
-	 * here rather than here and in the flush/logging code.
-	 */
-	if (xfs_sb_version_hasnlink(&mp->m_sb) &&
-	    ip->i_d.di_version == 1) {
-		ip->i_d.di_version = 2;
-		/*
-		 * We've already zeroed the old link count, the projid field,
-		 * and the pad field.
-		 */
-	}
-
-	/*
-	 * Project ids won't be stored on disk if we are using a version 1 inode.
-	 */
-	if ((prid != 0) && (ip->i_d.di_version == 1))
-		xfs_bump_ino_vers2(tp, ip);
-
 	if (pip && XFS_INHERIT_GID(pip)) {
 		ip->i_d.di_gid = pip->i_d.di_gid;
 		if ((pip->i_d.di_mode & S_ISGID) && S_ISDIR(mode)) {
@@ -772,13 +758,6 @@ xfs_ialloc(
 		flags |= XFS_ILOG_DEV;
 		break;
 	case S_IFREG:
-		/*
-		 * we can't set up filestreams until after the VFS inode
-		 * is set up properly.
-		 */
-		if (pip && xfs_inode_is_filestream(pip))
-			filestreams = 1;
-		/* fall through */
 	case S_IFDIR:
 		if (pip && (pip->i_d.di_flags & XFS_DIFLAG_ANY)) {
 			uint	di_flags = 0;
@@ -844,15 +823,6 @@ xfs_ialloc(
 	/* now that we have an i_mode we can setup inode ops and unlock */
 	xfs_setup_inode(ip);
 
-	/* now we have set up the vfs inode we can associate the filestream */
-	if (filestreams) {
-		error = xfs_filestream_associate(pip, ip);
-		if (error < 0)
-			return -error;
-		if (!error)
-			xfs_iflags_set(ip, XFS_IFILESTREAM);
-	}
-
 	*ipp = ip;
 	return 0;
 }
@@ -1072,40 +1042,6 @@ xfs_droplink(
 	return error;
 }
 
-/*
- * This gets called when the inode's version needs to be changed from 1 to 2.
- * Currently this happens when the nlink field overflows the old 16-bit value
- * or when chproj is called to change the project for the first time.
- * As a side effect the superblock version will also get rev'd
- * to contain the NLINK bit.
- */
-void
-xfs_bump_ino_vers2(
-	xfs_trans_t	*tp,
-	xfs_inode_t	*ip)
-{
-	xfs_mount_t	*mp;
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	ASSERT(ip->i_d.di_version == 1);
-
-	ip->i_d.di_version = 2;
-	ip->i_d.di_onlink = 0;
-	memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
-	mp = tp->t_mountp;
-	if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
-		spin_lock(&mp->m_sb_lock);
-		if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
-			xfs_sb_version_addnlink(&mp->m_sb);
-			spin_unlock(&mp->m_sb_lock);
-			xfs_mod_sb(tp, XFS_SB_VERSIONNUM);
-		} else {
-			spin_unlock(&mp->m_sb_lock);
-		}
-	}
-	/* Caller must log the inode */
-}
-
 /*
  * Increment the link count on an inode & log the change.
  */
@@ -1116,22 +1052,10 @@ xfs_bumplink(
 {
 	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
 
+	ASSERT(ip->i_d.di_version > 1);
 	ASSERT(ip->i_d.di_nlink > 0 || (VFS_I(ip)->i_state & I_LINKABLE));
 	ip->i_d.di_nlink++;
 	inc_nlink(VFS_I(ip));
-	if ((ip->i_d.di_version == 1) &&
-	    (ip->i_d.di_nlink > XFS_MAXLINK_1)) {
-		/*
-		 * The inode has increased its number of links beyond
-		 * what can fit in an old format inode.  It now needs
-		 * to be converted to a version 2 inode with a 32 bit
-		 * link count.  If this is the first inode in the file
-		 * system to do this, then we need to bump the superblock
-		 * version number as well.
-		 */
-		xfs_bump_ino_vers2(tp, ip);
-	}
-
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 	return 0;
 }
@@ -1698,16 +1622,6 @@ xfs_release(
 	if (!XFS_FORCED_SHUTDOWN(mp)) {
 		int truncated;
 
-		/*
-		 * If we are using filestreams, and we have an unlinked
-		 * file that we are processing the last close on, then nothing
-		 * will be able to reopen and write to this file. Purge this
-		 * inode from the filestreams cache so that it doesn't delay
-		 * teardown of the inode.
-		 */
-		if ((ip->i_d.di_nlink == 0) && xfs_inode_is_filestream(ip))
-			xfs_filestream_deassociate(ip);
-
 		/*
 		 * If we previously truncated this file and removed old data
 		 * in the process, we want to initiate "early" writeout on
@@ -1838,9 +1752,33 @@ xfs_inactive_ifree(
 	int			error;
 
 	tp = xfs_trans_alloc(mp, XFS_TRANS_INACTIVE);
-	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ifree, 0, 0);
+
+	/*
+	 * The ifree transaction might need to allocate blocks for record
+	 * insertion to the finobt. We don't want to fail here at ENOSPC, so
+	 * allow ifree to dip into the reserved block pool if necessary.
+	 *
+	 * Freeing large sets of inodes generally means freeing inode chunks,
+	 * directory and file data blocks, so this should be relatively safe.
+	 * Only under severe circumstances should it be possible to free enough
+	 * inodes to exhaust the reserve block pool via finobt expansion while
+	 * at the same time not creating free space in the filesystem.
+	 *
+	 * Send a warning if the reservation does happen to fail, as the inode
+	 * now remains allocated and sits on the unlinked list until the fs is
+	 * repaired.
+	 */
+	tp->t_flags |= XFS_TRANS_RESERVE;
+	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_ifree,
+				  XFS_IFREE_SPACE_RES(mp), 0);
 	if (error) {
-		ASSERT(XFS_FORCED_SHUTDOWN(mp));
+		if (error == ENOSPC) {
+			xfs_warn_ratelimited(mp,
+			"Failed to remove inode(s) from unlinked list. "
+			"Please free space, unmount and run xfs_repair.");
+		} else {
+			ASSERT(XFS_FORCED_SHUTDOWN(mp));
+		}
 		xfs_trans_cancel(tp, XFS_TRANS_RELEASE_LOG_RES);
 		return error;
 	}
@@ -2664,13 +2602,7 @@ xfs_remove(
 	if (error)
 		goto std_return;
 
-	/*
-	 * If we are using filestreams, kill the stream association.
-	 * If the file is still open it may get a new one but that
-	 * will get killed on last close in xfs_close() so we don't
-	 * have to worry about that.
-	 */
-	if (!is_dir && link_zero && xfs_inode_is_filestream(ip))
+	if (is_dir && xfs_inode_is_filestream(ip))
 		xfs_filestream_deassociate(ip);
 
 	return 0;
@@ -3258,6 +3190,7 @@ xfs_iflush_int(
 	ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
 	       ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK));
 	ASSERT(iip != NULL && iip->ili_fields != 0);
+	ASSERT(ip->i_d.di_version > 1);
 
 	/* set *dip = inode's place in the buffer */
 	dip = (xfs_dinode_t *)xfs_buf_offset(bp, ip->i_imap.im_boffset);
@@ -3318,7 +3251,7 @@ xfs_iflush_int(
 	}
 
 	/*
-	 * Inode item log recovery for v1/v2 inodes are dependent on the
+	 * Inode item log recovery for v2 inodes are dependent on the
 	 * di_flushiter count for correct sequencing. We bump the flush
 	 * iteration count so we can detect flushes which postdate a log record
 	 * during recovery. This is redundant as we now log every change and
@@ -3341,40 +3274,9 @@ xfs_iflush_int(
 	if (ip->i_d.di_flushiter == DI_MAX_FLUSH)
 		ip->i_d.di_flushiter = 0;
 
-	/*
-	 * If this is really an old format inode and the superblock version
-	 * has not been updated to support only new format inodes, then
-	 * convert back to the old inode format.  If the superblock version
-	 * has been updated, then make the conversion permanent.
-	 */
-	ASSERT(ip->i_d.di_version == 1 || xfs_sb_version_hasnlink(&mp->m_sb));
-	if (ip->i_d.di_version == 1) {
-		if (!xfs_sb_version_hasnlink(&mp->m_sb)) {
-			/*
-			 * Convert it back.
-			 */
-			ASSERT(ip->i_d.di_nlink <= XFS_MAXLINK_1);
-			dip->di_onlink = cpu_to_be16(ip->i_d.di_nlink);
-		} else {
-			/*
-			 * The superblock version has already been bumped,
-			 * so just make the conversion to the new inode
-			 * format permanent.
-			 */
-			ip->i_d.di_version = 2;
-			dip->di_version = 2;
-			ip->i_d.di_onlink = 0;
-			dip->di_onlink = 0;
-			memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
-			memset(&(dip->di_pad[0]), 0,
-			      sizeof(dip->di_pad));
-			ASSERT(xfs_get_projid(ip) == 0);
-		}
-	}
-
-	xfs_iflush_fork(ip, dip, iip, XFS_DATA_FORK, bp);
+	xfs_iflush_fork(ip, dip, iip, XFS_DATA_FORK);
 	if (XFS_IFORK_Q(ip))
-		xfs_iflush_fork(ip, dip, iip, XFS_ATTR_FORK, bp);
+		xfs_iflush_fork(ip, dip, iip, XFS_ATTR_FORK);
 	xfs_inobp_check(mp, bp);
 
 	/*

fs/xfs/xfs_inode.h

@@ -209,7 +209,6 @@ xfs_get_initial_prid(struct xfs_inode *dp)
 #define XFS_ISTALE		(1 << 1) /* inode has been staled */
 #define XFS_IRECLAIMABLE	(1 << 2) /* inode can be reclaimed */
 #define XFS_INEW		(1 << 3) /* inode has just been allocated */
-#define XFS_IFILESTREAM		(1 << 4) /* inode is in a filestream dir. */
 #define XFS_ITRUNCATED		(1 << 5) /* truncated down so flush-on-close */
 #define XFS_IDIRTY_RELEASE	(1 << 6) /* dirty release already seen */
 #define __XFS_IFLOCK_BIT	7	 /* inode is being flushed right now */
@@ -225,8 +224,7 @@ xfs_get_initial_prid(struct xfs_inode *dp)
  */
 #define XFS_IRECLAIM_RESET_FLAGS	\
 	(XFS_IRECLAIMABLE | XFS_IRECLAIM | \
-	 XFS_IDIRTY_RELEASE | XFS_ITRUNCATED | \
-	 XFS_IFILESTREAM);
+	 XFS_IDIRTY_RELEASE | XFS_ITRUNCATED)
 
 /*
  * Synchronize processes attempting to flush the in-core inode back to disk.
@@ -379,7 +377,6 @@ int		xfs_dir_ialloc(struct xfs_trans **, struct xfs_inode *, umode_t,
 			       struct xfs_inode **, int *);
 int		xfs_droplink(struct xfs_trans *, struct xfs_inode *);
 int		xfs_bumplink(struct xfs_trans *, struct xfs_inode *);
-void		xfs_bump_ino_vers2(struct xfs_trans *, struct xfs_inode *);
 
 /* from xfs_file.c */
 int		xfs_zero_eof(struct xfs_inode *, xfs_off_t, xfs_fsize_t);

fs/xfs/xfs_inode_buf.c

@@ -437,17 +437,16 @@ xfs_iread(
 	}
 
 	/*
-	 * The inode format changed when we moved the link count and
-	 * made it 32 bits long.  If this is an old format inode,
-	 * convert it in memory to look like a new one.  If it gets
-	 * flushed to disk we will convert back before flushing or
-	 * logging it.  We zero out the new projid field and the old link
-	 * count field.  We'll handle clearing the pad field (the remains
-	 * of the old uuid field) when we actually convert the inode to
-	 * the new format. We don't change the version number so that we
-	 * can distinguish this from a real new format inode.
+	 * Automatically convert version 1 inode formats in memory to version 2
+	 * inode format. If the inode is modified, it will get logged and
+	 * rewritten as a version 2 inode. We can do this because we set the
+	 * superblock feature bit for v2 inodes unconditionally during mount
+	 * and it means the reast of the code can assume the inode version is 2
+	 * or higher.
 	 */
 	if (ip->i_d.di_version == 1) {
+		ip->i_d.di_version = 2;
+		memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
 		ip->i_d.di_nlink = ip->i_d.di_onlink;
 		ip->i_d.di_onlink = 0;
 		xfs_set_projid(ip, 0);

fs/xfs/xfs_inode_fork.c

@@ -798,8 +798,7 @@ xfs_iflush_fork(
 	xfs_inode_t		*ip,
 	xfs_dinode_t		*dip,
 	xfs_inode_log_item_t	*iip,
-	int			whichfork,
-	xfs_buf_t		*bp)
+	int			whichfork)
 {
 	char			*cp;
 	xfs_ifork_t		*ifp;

fs/xfs/xfs_inode_fork.h

@@ -127,8 +127,7 @@ typedef struct xfs_ifork {
 
 int		xfs_iformat_fork(struct xfs_inode *, struct xfs_dinode *);
 void		xfs_iflush_fork(struct xfs_inode *, struct xfs_dinode *,
-				struct xfs_inode_log_item *, int,
-				struct xfs_buf *);
+				struct xfs_inode_log_item *, int);
 void		xfs_idestroy_fork(struct xfs_inode *, int);
 void		xfs_idata_realloc(struct xfs_inode *, int, int);
 void		xfs_iroot_realloc(struct xfs_inode *, int, int);

fs/xfs/xfs_inode_item.c

@@ -145,34 +145,6 @@ xfs_inode_item_size(
 		xfs_inode_item_attr_fork_size(iip, nvecs, nbytes);
 }
 
-/*
- * If this is a v1 format inode, then we need to log it as such.  This means
- * that we have to copy the link count from the new field to the old.  We
- * don't have to worry about the new fields, because nothing trusts them as
- * long as the old inode version number is there.
- */
-STATIC void
-xfs_inode_item_format_v1_inode(
-	struct xfs_inode	*ip)
-{
-	if (!xfs_sb_version_hasnlink(&ip->i_mount->m_sb)) {
-		/*
-		 * Convert it back.
-		 */
-		ASSERT(ip->i_d.di_nlink <= XFS_MAXLINK_1);
-		ip->i_d.di_onlink = ip->i_d.di_nlink;
-	} else {
-		/*
-		 * The superblock version has already been bumped,
-		 * so just make the conversion to the new inode
-		 * format permanent.
-		 */
-		ip->i_d.di_version = 2;
-		ip->i_d.di_onlink = 0;
-		memset(&(ip->i_d.di_pad[0]), 0, sizeof(ip->i_d.di_pad));
-	}
-}
-
 STATIC void
 xfs_inode_item_format_data_fork(
 	struct xfs_inode_log_item *iip,
@@ -370,6 +342,8 @@ xfs_inode_item_format(
 	struct xfs_inode_log_format *ilf;
 	struct xfs_log_iovec	*vecp = NULL;
 
+	ASSERT(ip->i_d.di_version > 1);
+
 	ilf = xlog_prepare_iovec(lv, &vecp, XLOG_REG_TYPE_IFORMAT);
 	ilf->ilf_type = XFS_LI_INODE;
 	ilf->ilf_ino = ip->i_ino;
@@ -380,8 +354,6 @@ xfs_inode_item_format(
 	ilf->ilf_size = 2; /* format + core */
 	xlog_finish_iovec(lv, vecp, sizeof(struct xfs_inode_log_format));
 
-	if (ip->i_d.di_version == 1)
-		xfs_inode_item_format_v1_inode(ip);
 	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_ICORE,
 			&ip->i_d,
 			xfs_icdinode_size(ip->i_d.di_version));

fs/xfs/xfs_ioctl.c

@@ -543,10 +543,11 @@ xfs_attrmulti_by_handle(
 
 	ops = memdup_user(am_hreq.ops, size);
 	if (IS_ERR(ops)) {
-		error = PTR_ERR(ops);
+		error = -PTR_ERR(ops);
 		goto out_dput;
 	}
 
+	error = ENOMEM;
 	attr_name = kmalloc(MAXNAMELEN, GFP_KERNEL);
 	if (!attr_name)
 		goto out_kfree_ops;
@@ -556,7 +557,7 @@ xfs_attrmulti_by_handle(
 		ops[i].am_error = strncpy_from_user((char *)attr_name,
 				ops[i].am_attrname, MAXNAMELEN);
 		if (ops[i].am_error == 0 || ops[i].am_error == MAXNAMELEN)
-			error = -ERANGE;
+			error = ERANGE;
 		if (ops[i].am_error < 0)
 			break;
 
@@ -1227,15 +1228,8 @@ xfs_ioctl_setattr(
 				olddquot = xfs_qm_vop_chown(tp, ip,
 							&ip->i_pdquot, pdqp);
 			}
+			ASSERT(ip->i_d.di_version > 1);
 			xfs_set_projid(ip, fa->fsx_projid);
-
-			/*
-			 * We may have to rev the inode as well as
-			 * the superblock version number since projids didn't
-			 * exist before DINODE_VERSION_2 and SB_VERSION_NLINK.
-			 */
-			if (ip->i_d.di_version == 1)
-				xfs_bump_ino_vers2(tp, ip);
 		}
 
 	}

fs/xfs/xfs_ioctl32.c

@@ -424,10 +424,11 @@ xfs_compat_attrmulti_by_handle(
 
 	ops = memdup_user(compat_ptr(am_hreq.ops), size);
 	if (IS_ERR(ops)) {
-		error = PTR_ERR(ops);
+		error = -PTR_ERR(ops);
 		goto out_dput;
 	}
 
+	error = ENOMEM;
 	attr_name = kmalloc(MAXNAMELEN, GFP_KERNEL);
 	if (!attr_name)
 		goto out_kfree_ops;
@@ -438,7 +439,7 @@ xfs_compat_attrmulti_by_handle(
 				compat_ptr(ops[i].am_attrname),
 				MAXNAMELEN);
 		if (ops[i].am_error == 0 || ops[i].am_error == MAXNAMELEN)
-			error = -ERANGE;
+			error = ERANGE;
 		if (ops[i].am_error < 0)
 			break;
 

fs/xfs/xfs_iomap.c

@@ -730,7 +730,7 @@ xfs_iomap_write_allocate(
 			 */
 			nimaps = 1;
 			end_fsb = XFS_B_TO_FSB(mp, XFS_ISIZE(ip));
-			error = xfs_bmap_last_offset(NULL, ip, &last_block,
+			error = xfs_bmap_last_offset(ip, &last_block,
 							XFS_DATA_FORK);
 			if (error)
 				goto trans_cancel;

fs/xfs/xfs_iops.c

@@ -829,22 +829,34 @@ xfs_setattr_size(
 	 */
 	inode_dio_wait(inode);
 
+	/*
+	 * Do all the page cache truncate work outside the transaction context
+	 * as the "lock" order is page lock->log space reservation.  i.e.
+	 * locking pages inside the transaction can ABBA deadlock with
+	 * writeback. We have to do the VFS inode size update before we truncate
+	 * the pagecache, however, to avoid racing with page faults beyond the
+	 * new EOF they are not serialised against truncate operations except by
+	 * page locks and size updates.
+	 *
+	 * Hence we are in a situation where a truncate can fail with ENOMEM
+	 * from xfs_trans_reserve(), but having already truncated the in-memory
+	 * version of the file (i.e. made user visible changes). There's not
+	 * much we can do about this, except to hope that the caller sees ENOMEM
+	 * and retries the truncate operation.
+	 */
 	error = -block_truncate_page(inode->i_mapping, newsize, xfs_get_blocks);
 	if (error)
 		return error;
+	truncate_setsize(inode, newsize);
 
 	tp = xfs_trans_alloc(mp, XFS_TRANS_SETATTR_SIZE);
 	error = xfs_trans_reserve(tp, &M_RES(mp)->tr_itruncate, 0, 0);
 	if (error)
 		goto out_trans_cancel;
 
-	truncate_setsize(inode, newsize);
-
 	commit_flags = XFS_TRANS_RELEASE_LOG_RES;
 	lock_flags |= XFS_ILOCK_EXCL;
-
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
-
 	xfs_trans_ijoin(tp, ip, 0);
 
 	/*

fs/xfs/xfs_itable.c

@@ -270,7 +270,8 @@ xfs_bulkstat(
 		/*
 		 * Allocate and initialize a btree cursor for ialloc btree.
 		 */
-		cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno);
+		cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno,
+					    XFS_BTNUM_INO);
 		irbp = irbuf;
 		irbufend = irbuf + nirbuf;
 		end_of_ag = 0;
@@ -621,7 +622,8 @@ xfs_inumbers(
 				agino = 0;
 				continue;
 			}
-			cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno);
+			cur = xfs_inobt_init_cursor(mp, NULL, agbp, agno,
+						    XFS_BTNUM_INO);
 			error = xfs_inobt_lookup(cur, agino, XFS_LOOKUP_GE,
 						 &tmp);
 			if (error) {

fs/xfs/xfs_log.c

@@ -1165,7 +1165,7 @@ xlog_iodone(xfs_buf_t *bp)
 	/*
 	 * Race to shutdown the filesystem if we see an error.
 	 */
-	if (XFS_TEST_ERROR((xfs_buf_geterror(bp)), l->l_mp,
+	if (XFS_TEST_ERROR(bp->b_error, l->l_mp,
 			XFS_ERRTAG_IODONE_IOERR, XFS_RANDOM_IODONE_IOERR)) {
 		xfs_buf_ioerror_alert(bp, __func__);
 		xfs_buf_stale(bp);
@@ -3952,11 +3952,14 @@ xfs_log_force_umount(
 		retval = xlog_state_ioerror(log);
 		spin_unlock(&log->l_icloglock);
 	}
+
 	/*
-	 * Wake up everybody waiting on xfs_log_force.
-	 * Callback all log item committed functions as if the
-	 * log writes were completed.
+	 * Wake up everybody waiting on xfs_log_force. Wake the CIL push first
+	 * as if the log writes were completed. The abort handling in the log
+	 * item committed callback functions will do this again under lock to
+	 * avoid races.
 	 */
+	wake_up_all(&log->l_cilp->xc_commit_wait);
 	xlog_state_do_callback(log, XFS_LI_ABORTED, NULL);
 
 #ifdef XFSERRORDEBUG

fs/xfs/xfs_log.h

@@ -24,7 +24,8 @@ struct xfs_log_vec {
 	struct xfs_log_iovec	*lv_iovecp;	/* iovec array */
 	struct xfs_log_item	*lv_item;	/* owner */
 	char			*lv_buf;	/* formatted buffer */
-	int			lv_buf_len;	/* size of formatted buffer */
+	int			lv_bytes;	/* accounted space in buffer */
+	int			lv_buf_len;	/* aligned size of buffer */
 	int			lv_size;	/* size of allocated lv */
 };
 
@@ -52,15 +53,21 @@ xlog_prepare_iovec(struct xfs_log_vec *lv, struct xfs_log_iovec **vecp,
 	return vec->i_addr;
 }
 
+/*
+ * We need to make sure the next buffer is naturally aligned for the biggest
+ * basic data type we put into it.  We already accounted for this padding when
+ * sizing the buffer.
+ *
+ * However, this padding does not get written into the log, and hence we have to
+ * track the space used by the log vectors separately to prevent log space hangs
+ * due to inaccurate accounting (i.e. a leak) of the used log space through the
+ * CIL context ticket.
+ */
 static inline void
 xlog_finish_iovec(struct xfs_log_vec *lv, struct xfs_log_iovec *vec, int len)
 {
-	/*
-	 * We need to make sure the next buffer is naturally aligned for the
-	 * biggest basic data type we put into it.  We already accounted for
-	 * this when sizing the buffer.
-	 */
 	lv->lv_buf_len += round_up(len, sizeof(uint64_t));
+	lv->lv_bytes += len;
 	vec->i_len = len;
 }
 

fs/xfs/xfs_log_cil.c

@@ -97,7 +97,7 @@ xfs_cil_prepare_item(
 {
 	/* Account for the new LV being passed in */
 	if (lv->lv_buf_len != XFS_LOG_VEC_ORDERED) {
-		*diff_len += lv->lv_buf_len;
+		*diff_len += lv->lv_bytes;
 		*diff_iovecs += lv->lv_niovecs;
 	}
 
@@ -111,7 +111,7 @@ xfs_cil_prepare_item(
 	else if (old_lv != lv) {
 		ASSERT(lv->lv_buf_len != XFS_LOG_VEC_ORDERED);
 
-		*diff_len -= old_lv->lv_buf_len;
+		*diff_len -= old_lv->lv_bytes;
 		*diff_iovecs -= old_lv->lv_niovecs;
 		kmem_free(old_lv);
 	}
@@ -239,7 +239,7 @@ xlog_cil_insert_format_items(
 			 * that the space reservation accounting is correct.
 			 */
 			*diff_iovecs -= lv->lv_niovecs;
-			*diff_len -= lv->lv_buf_len;
+			*diff_len -= lv->lv_bytes;
 		} else {
 			/* allocate new data chunk */
 			lv = kmem_zalloc(buf_size, KM_SLEEP|KM_NOFS);
@@ -259,6 +259,7 @@ xlog_cil_insert_format_items(
 
 		/* The allocated data region lies beyond the iovec region */
 		lv->lv_buf_len = 0;
+		lv->lv_bytes = 0;
 		lv->lv_buf = (char *)lv + buf_size - nbytes;
 		ASSERT(IS_ALIGNED((unsigned long)lv->lv_buf, sizeof(uint64_t)));
 
@@ -385,7 +386,15 @@ xlog_cil_committed(
 	xfs_extent_busy_clear(mp, &ctx->busy_extents,
 			     (mp->m_flags & XFS_MOUNT_DISCARD) && !abort);
 
+	/*
+	 * If we are aborting the commit, wake up anyone waiting on the
+	 * committing list.  If we don't, then a shutdown we can leave processes
+	 * waiting in xlog_cil_force_lsn() waiting on a sequence commit that
+	 * will never happen because we aborted it.
+	 */
 	spin_lock(&ctx->cil->xc_push_lock);
+	if (abort)
+		wake_up_all(&ctx->cil->xc_commit_wait);
 	list_del(&ctx->committing);
 	spin_unlock(&ctx->cil->xc_push_lock);
 
@@ -563,9 +572,19 @@ xlog_cil_push(
 restart:
 	spin_lock(&cil->xc_push_lock);
 	list_for_each_entry(new_ctx, &cil->xc_committing, committing) {
+		/*
+		 * Avoid getting stuck in this loop because we were woken by the
+		 * shutdown, but then went back to sleep once already in the
+		 * shutdown state.
+		 */
+		if (XLOG_FORCED_SHUTDOWN(log)) {
+			spin_unlock(&cil->xc_push_lock);
+			goto out_abort_free_ticket;
+		}
+
 		/*
 		 * Higher sequences will wait for this one so skip them.
-		 * Don't wait for own own sequence, either.
+		 * Don't wait for our own sequence, either.
 		 */
 		if (new_ctx->sequence >= ctx->sequence)
 			continue;
@@ -810,6 +829,13 @@ restart:
 	 */
 	spin_lock(&cil->xc_push_lock);
 	list_for_each_entry(ctx, &cil->xc_committing, committing) {
+		/*
+		 * Avoid getting stuck in this loop because we were woken by the
+		 * shutdown, but then went back to sleep once already in the
+		 * shutdown state.
+		 */
+		if (XLOG_FORCED_SHUTDOWN(log))
+			goto out_shutdown;
 		if (ctx->sequence > sequence)
 			continue;
 		if (!ctx->commit_lsn) {
@@ -833,14 +859,12 @@ restart:
 	 * push sequence after the above wait loop and the CIL still contains
 	 * dirty objects.
 	 *
-	 * When the push occurs, it will empty the CIL and
-	 * atomically increment the currect sequence past the push sequence and
-	 * move it into the committing list. Of course, if the CIL is clean at
-	 * the time of the push, it won't have pushed the CIL at all, so in that
-	 * case we should try the push for this sequence again from the start
-	 * just in case.
+	 * When the push occurs, it will empty the CIL and atomically increment
+	 * the currect sequence past the push sequence and move it into the
+	 * committing list. Of course, if the CIL is clean at the time of the
+	 * push, it won't have pushed the CIL at all, so in that case we should
+	 * try the push for this sequence again from the start just in case.
 	 */
-
 	if (sequence == cil->xc_current_sequence &&
 	    !list_empty(&cil->xc_cil)) {
 		spin_unlock(&cil->xc_push_lock);
@@ -849,6 +873,17 @@ restart:
 
 	spin_unlock(&cil->xc_push_lock);
 	return commit_lsn;
+
+	/*
+	 * We detected a shutdown in progress. We need to trigger the log force
+	 * to pass through it's iclog state machine error handling, even though
+	 * we are already in a shutdown state. Hence we can't return
+	 * NULLCOMMITLSN here as that has special meaning to log forces (i.e.
+	 * LSN is already stable), so we return a zero LSN instead.
+	 */
+out_shutdown:
+	spin_unlock(&cil->xc_push_lock);
+	return 0;
 }
 
 /*

fs/xfs/xfs_log_recover.c

@@ -2138,7 +2138,9 @@ xlog_recover_validate_buf_type(
 			bp->b_ops = &xfs_allocbt_buf_ops;
 			break;
 		case XFS_IBT_CRC_MAGIC:
+		case XFS_FIBT_CRC_MAGIC:
 		case XFS_IBT_MAGIC:
+		case XFS_FIBT_MAGIC:
 			bp->b_ops = &xfs_inobt_buf_ops;
 			break;
 		case XFS_BMAP_CRC_MAGIC:
@@ -3145,7 +3147,7 @@ xlog_recover_efd_pass2(
 		}
 		lip = xfs_trans_ail_cursor_next(ailp, &cur);
 	}
-	xfs_trans_ail_cursor_done(ailp, &cur);
+	xfs_trans_ail_cursor_done(&cur);
 	spin_unlock(&ailp->xa_lock);
 
 	return 0;
@@ -3520,8 +3522,7 @@ out:
 
 STATIC int
 xlog_recover_unmount_trans(
-	struct xlog		*log,
-	struct xlog_recover	*trans)
+	struct xlog		*log)
 {
 	/* Do nothing now */
 	xfs_warn(log->l_mp, "%s: Unmount LR", __func__);
@@ -3595,7 +3596,7 @@ xlog_recover_process_data(
 								trans, pass);
 				break;
 			case XLOG_UNMOUNT_TRANS:
-				error = xlog_recover_unmount_trans(log, trans);
+				error = xlog_recover_unmount_trans(log);
 				break;
 			case XLOG_WAS_CONT_TRANS:
 				error = xlog_recover_add_to_cont_trans(log,
@@ -3757,7 +3758,7 @@ xlog_recover_process_efis(
 		lip = xfs_trans_ail_cursor_next(ailp, &cur);
 	}
 out:
-	xfs_trans_ail_cursor_done(ailp, &cur);
+	xfs_trans_ail_cursor_done(&cur);
 	spin_unlock(&ailp->xa_lock);
 	return error;
 }

fs/xfs/xfs_log_rlimit.c

@@ -42,7 +42,7 @@ xfs_log_calc_max_attrsetm_res(
 	int			size;
 	int			nblks;
 
-	size = xfs_attr_leaf_entsize_local_max(mp->m_sb.sb_blocksize) -
+	size = xfs_attr_leaf_entsize_local_max(mp->m_attr_geo->blksize) -
 	       MAXNAMELEN - 1;
 	nblks = XFS_DAENTER_SPACE_RES(mp, XFS_ATTR_FORK);
 	nblks += XFS_B_TO_FSB(mp, size);

fs/xfs/xfs_mount.c

@@ -323,8 +323,19 @@ reread:
 	/*
 	 * Initialize the mount structure from the superblock.
 	 */
-	xfs_sb_from_disk(&mp->m_sb, XFS_BUF_TO_SBP(bp));
-	xfs_sb_quota_from_disk(&mp->m_sb);
+	xfs_sb_from_disk(sbp, XFS_BUF_TO_SBP(bp));
+	xfs_sb_quota_from_disk(sbp);
+
+	/*
+	 * If we haven't validated the superblock, do so now before we try
+	 * to check the sector size and reread the superblock appropriately.
+	 */
+	if (sbp->sb_magicnum != XFS_SB_MAGIC) {
+		if (loud)
+			xfs_warn(mp, "Invalid superblock magic number");
+		error = EINVAL;
+		goto release_buf;
+	}
 
 	/*
 	 * We must be able to do sector-sized and sector-aligned IO.
@@ -337,11 +348,11 @@ reread:
 		goto release_buf;
 	}
 
-	/*
-	 * Re-read the superblock so the buffer is correctly sized,
-	 * and properly verified.
-	 */
 	if (buf_ops == NULL) {
+		/*
+		 * Re-read the superblock so the buffer is correctly sized,
+		 * and properly verified.
+		 */
 		xfs_buf_relse(bp);
 		sector_size = sbp->sb_sectsize;
 		buf_ops = loud ? &xfs_sb_buf_ops : &xfs_sb_quiet_buf_ops;
@@ -697,6 +708,12 @@ xfs_mountfs(
 			mp->m_update_flags |= XFS_SB_VERSIONNUM;
 	}
 
+	/* always use v2 inodes by default now */
+	if (!(mp->m_sb.sb_versionnum & XFS_SB_VERSION_NLINKBIT)) {
+		mp->m_sb.sb_versionnum |= XFS_SB_VERSION_NLINKBIT;
+		mp->m_update_flags |= XFS_SB_VERSIONNUM;
+	}
+
 	/*
 	 * Check if sb_agblocks is aligned at stripe boundary
 	 * If sb_agblocks is NOT aligned turn off m_dalign since
@@ -774,12 +791,11 @@ xfs_mountfs(
 
 	mp->m_dmevmask = 0;	/* not persistent; set after each mount */
 
-	xfs_dir_mount(mp);
-
-	/*
-	 * Initialize the attribute manager's entries.
-	 */
-	mp->m_attr_magicpct = (mp->m_sb.sb_blocksize * 37) / 100;
+	error = xfs_da_mount(mp);
+	if (error) {
+		xfs_warn(mp, "Failed dir/attr init: %d", error);
+		goto out_remove_uuid;
+	}
 
 	/*
 	 * Initialize the precomputed transaction reservations values.
@@ -794,7 +810,7 @@ xfs_mountfs(
 	error = xfs_initialize_perag(mp, sbp->sb_agcount, &mp->m_maxagi);
 	if (error) {
 		xfs_warn(mp, "Failed per-ag init: %d", error);
-		goto out_remove_uuid;
+		goto out_free_dir;
 	}
 
 	if (!sbp->sb_logblocks) {
@@ -969,6 +985,8 @@ xfs_mountfs(
 	xfs_wait_buftarg(mp->m_ddev_targp);
  out_free_perag:
 	xfs_free_perag(mp);
+ out_free_dir:
+	xfs_da_unmount(mp);
  out_remove_uuid:
 	xfs_uuid_unmount(mp);
  out:
@@ -1046,6 +1064,7 @@ xfs_unmountfs(
 				"Freespace may not be correct on next mount.");
 
 	xfs_log_unmount(mp);
+	xfs_da_unmount(mp);
 	xfs_uuid_unmount(mp);
 
 #if defined(DEBUG)

fs/xfs/xfs_mount.h

@@ -27,6 +27,7 @@ struct xfs_nameops;
 struct xfs_ail;
 struct xfs_quotainfo;
 struct xfs_dir_ops;
+struct xfs_da_geometry;
 
 #ifdef HAVE_PERCPU_SB
 
@@ -96,6 +97,8 @@ typedef struct xfs_mount {
 	uint			m_readio_blocks; /* min read size blocks */
 	uint			m_writeio_log;	/* min write size log bytes */
 	uint			m_writeio_blocks; /* min write size blocks */
+	struct xfs_da_geometry	*m_dir_geo;	/* directory block geometry */
+	struct xfs_da_geometry	*m_attr_geo;	/* attribute block geometry */
 	struct xlog		*m_log;		/* log specific stuff */
 	int			m_logbufs;	/* number of log buffers */
 	int			m_logbsize;	/* size of each log buffer */
@@ -131,8 +134,6 @@ typedef struct xfs_mount {
 	int			m_fixedfsid[2];	/* unchanged for life of FS */
 	uint			m_dmevmask;	/* DMI events for this FS */
 	__uint64_t		m_flags;	/* global mount flags */
-	uint			m_dir_node_ents; /* #entries in a dir danode */
-	uint			m_attr_node_ents; /* #entries in attr danode */
 	int			m_ialloc_inos;	/* inodes in inode allocation */
 	int			m_ialloc_blks;	/* blocks in inode allocation */
 	int			m_inoalign_mask;/* mask sb_inoalignmt if used */
@@ -145,17 +146,10 @@ typedef struct xfs_mount {
 	int			m_dalign;	/* stripe unit */
 	int			m_swidth;	/* stripe width */
 	int			m_sinoalign;	/* stripe unit inode alignment */
-	int			m_attr_magicpct;/* 37% of the blocksize */
-	int			m_dir_magicpct;	/* 37% of the dir blocksize */
 	__uint8_t		m_sectbb_log;	/* sectlog - BBSHIFT */
 	const struct xfs_nameops *m_dirnameops;	/* vector of dir name ops */
 	const struct xfs_dir_ops *m_dir_inode_ops; /* vector of dir inode ops */
 	const struct xfs_dir_ops *m_nondir_inode_ops; /* !dir inode ops */
-	int			m_dirblksize;	/* directory block sz--bytes */
-	int			m_dirblkfsbs;	/* directory block sz--fsbs */
-	xfs_dablk_t		m_dirdatablk;	/* blockno of dir data v2 */
-	xfs_dablk_t		m_dirleafblk;	/* blockno of dir non-data v2 */
-	xfs_dablk_t		m_dirfreeblk;	/* blockno of dirfreeindex v2 */
 	uint			m_chsize;	/* size of next field */
 	atomic_t		m_active_trans;	/* number trans frozen */
 #ifdef HAVE_PERCPU_SB

fs/xfs/xfs_mru_cache.c

@@ -100,14 +100,20 @@
  * likely result in a loop in one of the lists.  That's a sure-fire recipe for
  * an infinite loop in the code.
  */
-typedef struct xfs_mru_cache_elem
-{
-	struct list_head list_node;
-	unsigned long	key;
-	void		*value;
-} xfs_mru_cache_elem_t;
+struct xfs_mru_cache {
+	struct radix_tree_root	store;     /* Core storage data structure.  */
+	struct list_head	*lists;    /* Array of lists, one per grp.  */
+	struct list_head	reap_list; /* Elements overdue for reaping. */
+	spinlock_t		lock;      /* Lock to protect this struct.  */
+	unsigned int		grp_count; /* Number of discrete groups.    */
+	unsigned int		grp_time;  /* Time period spanned by grps.  */
+	unsigned int		lru_grp;   /* Group containing time zero.   */
+	unsigned long		time_zero; /* Time first element was added. */
+	xfs_mru_cache_free_func_t free_func; /* Function pointer for freeing. */
+	struct delayed_work	work;      /* Workqueue data for reaping.   */
+	unsigned int		queued;	   /* work has been queued */
+};
 
-static kmem_zone_t		*xfs_mru_elem_zone;
 static struct workqueue_struct	*xfs_mru_reap_wq;
 
 /*
@@ -129,12 +135,12 @@ static struct workqueue_struct	*xfs_mru_reap_wq;
  */
 STATIC unsigned long
 _xfs_mru_cache_migrate(
-	xfs_mru_cache_t	*mru,
-	unsigned long	now)
+	struct xfs_mru_cache	*mru,
+	unsigned long		now)
 {
-	unsigned int	grp;
-	unsigned int	migrated = 0;
-	struct list_head *lru_list;
+	unsigned int		grp;
+	unsigned int		migrated = 0;
+	struct list_head	*lru_list;
 
 	/* Nothing to do if the data store is empty. */
 	if (!mru->time_zero)
@@ -193,11 +199,11 @@ _xfs_mru_cache_migrate(
  */
 STATIC void
 _xfs_mru_cache_list_insert(
-	xfs_mru_cache_t		*mru,
-	xfs_mru_cache_elem_t	*elem)
+	struct xfs_mru_cache	*mru,
+	struct xfs_mru_cache_elem *elem)
 {
-	unsigned int	grp = 0;
-	unsigned long	now = jiffies;
+	unsigned int		grp = 0;
+	unsigned long		now = jiffies;
 
 	/*
 	 * If the data store is empty, initialise time zero, leave grp set to
@@ -231,10 +237,10 @@ _xfs_mru_cache_list_insert(
  */
 STATIC void
 _xfs_mru_cache_clear_reap_list(
-	xfs_mru_cache_t		*mru) __releases(mru->lock) __acquires(mru->lock)
-
+	struct xfs_mru_cache	*mru)
+		__releases(mru->lock) __acquires(mru->lock)
 {
-	xfs_mru_cache_elem_t	*elem, *next;
+	struct xfs_mru_cache_elem *elem, *next;
 	struct list_head	tmp;
 
 	INIT_LIST_HEAD(&tmp);
@@ -252,15 +258,8 @@ _xfs_mru_cache_clear_reap_list(
 	spin_unlock(&mru->lock);
 
 	list_for_each_entry_safe(elem, next, &tmp, list_node) {
-
-		/* Remove the element from the reap list. */
 		list_del_init(&elem->list_node);
-
-		/* Call the client's free function with the key and value pointer. */
-		mru->free_func(elem->key, elem->value);
-
-		/* Free the element structure. */
-		kmem_zone_free(xfs_mru_elem_zone, elem);
+		mru->free_func(elem);
 	}
 
 	spin_lock(&mru->lock);
@@ -277,7 +276,8 @@ STATIC void
 _xfs_mru_cache_reap(
 	struct work_struct	*work)
 {
-	xfs_mru_cache_t		*mru = container_of(work, xfs_mru_cache_t, work.work);
+	struct xfs_mru_cache	*mru =
+		container_of(work, struct xfs_mru_cache, work.work);
 	unsigned long		now, next;
 
 	ASSERT(mru && mru->lists);
@@ -304,28 +304,16 @@ _xfs_mru_cache_reap(
 int
 xfs_mru_cache_init(void)
 {
-	xfs_mru_elem_zone = kmem_zone_init(sizeof(xfs_mru_cache_elem_t),
-	                                 "xfs_mru_cache_elem");
-	if (!xfs_mru_elem_zone)
-		goto out;
-
 	xfs_mru_reap_wq = alloc_workqueue("xfs_mru_cache", WQ_MEM_RECLAIM, 1);
 	if (!xfs_mru_reap_wq)
-		goto out_destroy_mru_elem_zone;
-
+		return -ENOMEM;
 	return 0;
-
- out_destroy_mru_elem_zone:
-	kmem_zone_destroy(xfs_mru_elem_zone);
- out:
-	return -ENOMEM;
 }
 
 void
 xfs_mru_cache_uninit(void)
 {
 	destroy_workqueue(xfs_mru_reap_wq);
-	kmem_zone_destroy(xfs_mru_elem_zone);
 }
 
 /*
@@ -336,14 +324,14 @@ xfs_mru_cache_uninit(void)
  */
 int
 xfs_mru_cache_create(
-	xfs_mru_cache_t		**mrup,
+	struct xfs_mru_cache	**mrup,
 	unsigned int		lifetime_ms,
 	unsigned int		grp_count,
 	xfs_mru_cache_free_func_t free_func)
 {
-	xfs_mru_cache_t	*mru = NULL;
-	int		err = 0, grp;
-	unsigned int	grp_time;
+	struct xfs_mru_cache	*mru = NULL;
+	int			err = 0, grp;
+	unsigned int		grp_time;
 
 	if (mrup)
 		*mrup = NULL;
@@ -400,7 +388,7 @@ exit:
  */
 static void
 xfs_mru_cache_flush(
-	xfs_mru_cache_t		*mru)
+	struct xfs_mru_cache	*mru)
 {
 	if (!mru || !mru->lists)
 		return;
@@ -420,7 +408,7 @@ xfs_mru_cache_flush(
 
 void
 xfs_mru_cache_destroy(
-	xfs_mru_cache_t		*mru)
+	struct xfs_mru_cache	*mru)
 {
 	if (!mru || !mru->lists)
 		return;
@@ -438,38 +426,30 @@ xfs_mru_cache_destroy(
  */
 int
 xfs_mru_cache_insert(
-	xfs_mru_cache_t	*mru,
-	unsigned long	key,
-	void		*value)
+	struct xfs_mru_cache	*mru,
+	unsigned long		key,
+	struct xfs_mru_cache_elem *elem)
 {
-	xfs_mru_cache_elem_t *elem;
+	int			error;
 
 	ASSERT(mru && mru->lists);
 	if (!mru || !mru->lists)
 		return EINVAL;
 
-	elem = kmem_zone_zalloc(xfs_mru_elem_zone, KM_SLEEP);
-	if (!elem)
+	if (radix_tree_preload(GFP_KERNEL))
 		return ENOMEM;
 
-	if (radix_tree_preload(GFP_KERNEL)) {
-		kmem_zone_free(xfs_mru_elem_zone, elem);
-		return ENOMEM;
-	}
-
 	INIT_LIST_HEAD(&elem->list_node);
 	elem->key = key;
-	elem->value = value;
 
 	spin_lock(&mru->lock);
-
-	radix_tree_insert(&mru->store, key, elem);
+	error = -radix_tree_insert(&mru->store, key, elem);
 	radix_tree_preload_end();
-	_xfs_mru_cache_list_insert(mru, elem);
-
+	if (!error)
+		_xfs_mru_cache_list_insert(mru, elem);
 	spin_unlock(&mru->lock);
 
-	return 0;
+	return error;
 }
 
 /*
@@ -478,13 +458,12 @@ xfs_mru_cache_insert(
  * the client data pointer for the removed element is returned, otherwise this
  * function will return a NULL pointer.
  */
-void *
+struct xfs_mru_cache_elem *
 xfs_mru_cache_remove(
-	xfs_mru_cache_t	*mru,
-	unsigned long	key)
+	struct xfs_mru_cache	*mru,
+	unsigned long		key)
 {
-	xfs_mru_cache_elem_t *elem;
-	void		*value = NULL;
+	struct xfs_mru_cache_elem *elem;
 
 	ASSERT(mru && mru->lists);
 	if (!mru || !mru->lists)
@@ -492,17 +471,11 @@ xfs_mru_cache_remove(
 
 	spin_lock(&mru->lock);
 	elem = radix_tree_delete(&mru->store, key);
-	if (elem) {
-		value = elem->value;
+	if (elem)
 		list_del(&elem->list_node);
-	}
-
 	spin_unlock(&mru->lock);
 
-	if (elem)
-		kmem_zone_free(xfs_mru_elem_zone, elem);
-
-	return value;
+	return elem;
 }
 
 /*
@@ -511,13 +484,14 @@ xfs_mru_cache_remove(
  */
 void
 xfs_mru_cache_delete(
-	xfs_mru_cache_t	*mru,
-	unsigned long	key)
+	struct xfs_mru_cache	*mru,
+	unsigned long		key)
 {
-	void		*value = xfs_mru_cache_remove(mru, key);
+	struct xfs_mru_cache_elem *elem;
 
-	if (value)
-		mru->free_func(key, value);
+	elem = xfs_mru_cache_remove(mru, key);
+	if (elem)
+		mru->free_func(elem);
 }
 
 /*
@@ -540,12 +514,12 @@ xfs_mru_cache_delete(
  * status, we need to help it get it right by annotating the path that does
  * not release the lock.
  */
-void *
+struct xfs_mru_cache_elem *
 xfs_mru_cache_lookup(
-	xfs_mru_cache_t	*mru,
-	unsigned long	key)
+	struct xfs_mru_cache	*mru,
+	unsigned long		key)
 {
-	xfs_mru_cache_elem_t *elem;
+	struct xfs_mru_cache_elem *elem;
 
 	ASSERT(mru && mru->lists);
 	if (!mru || !mru->lists)
@@ -560,7 +534,7 @@ xfs_mru_cache_lookup(
 	} else
 		spin_unlock(&mru->lock);
 
-	return elem ? elem->value : NULL;
+	return elem;
 }
 
 /*
@@ -570,7 +544,8 @@ xfs_mru_cache_lookup(
  */
 void
 xfs_mru_cache_done(
-	xfs_mru_cache_t	*mru) __releases(mru->lock)
+	struct xfs_mru_cache	*mru)
+		__releases(mru->lock)
 {
 	spin_unlock(&mru->lock);
 }

fs/xfs/xfs_mru_cache.h

@@ -18,24 +18,15 @@
 #ifndef __XFS_MRU_CACHE_H__
 #define __XFS_MRU_CACHE_H__
 
+struct xfs_mru_cache;
 
-/* Function pointer type for callback to free a client's data pointer. */
-typedef void (*xfs_mru_cache_free_func_t)(unsigned long, void*);
+struct xfs_mru_cache_elem {
+	struct list_head list_node;
+	unsigned long	key;
+};
 
-typedef struct xfs_mru_cache
-{
-	struct radix_tree_root	store;     /* Core storage data structure.  */
-	struct list_head	*lists;    /* Array of lists, one per grp.  */
-	struct list_head	reap_list; /* Elements overdue for reaping. */
-	spinlock_t		lock;      /* Lock to protect this struct.  */
-	unsigned int		grp_count; /* Number of discrete groups.    */
-	unsigned int		grp_time;  /* Time period spanned by grps.  */
-	unsigned int		lru_grp;   /* Group containing time zero.   */
-	unsigned long		time_zero; /* Time first element was added. */
-	xfs_mru_cache_free_func_t free_func; /* Function pointer for freeing. */
-	struct delayed_work	work;      /* Workqueue data for reaping.   */
-	unsigned int		queued;	   /* work has been queued */
-} xfs_mru_cache_t;
+/* Function pointer type for callback to free a client's data pointer. */
+typedef void (*xfs_mru_cache_free_func_t)(struct xfs_mru_cache_elem *elem);
 
 int xfs_mru_cache_init(void);
 void xfs_mru_cache_uninit(void);
@@ -44,10 +35,12 @@ int xfs_mru_cache_create(struct xfs_mru_cache **mrup, unsigned int lifetime_ms,
 			     xfs_mru_cache_free_func_t free_func);
 void xfs_mru_cache_destroy(struct xfs_mru_cache *mru);
 int xfs_mru_cache_insert(struct xfs_mru_cache *mru, unsigned long key,
-				void *value);
-void * xfs_mru_cache_remove(struct xfs_mru_cache *mru, unsigned long key);
+		struct xfs_mru_cache_elem *elem);
+struct xfs_mru_cache_elem *
+xfs_mru_cache_remove(struct xfs_mru_cache *mru, unsigned long key);
 void xfs_mru_cache_delete(struct xfs_mru_cache *mru, unsigned long key);
-void *xfs_mru_cache_lookup(struct xfs_mru_cache *mru, unsigned long key);
+struct xfs_mru_cache_elem *
+xfs_mru_cache_lookup(struct xfs_mru_cache *mru, unsigned long key);
 void xfs_mru_cache_done(struct xfs_mru_cache *mru);
 
 #endif /* __XFS_MRU_CACHE_H__ */

fs/xfs/xfs_qm.c

@@ -192,47 +192,6 @@ xfs_qm_dqpurge(
 	return 0;
 }
 
-/*
- * Release the group or project dquot pointers the user dquots maybe carrying
- * around as a hint, and proceed to purge the user dquot cache if requested.
-*/
-STATIC int
-xfs_qm_dqpurge_hints(
-	struct xfs_dquot	*dqp,
-	void			*data)
-{
-	struct xfs_dquot	*gdqp = NULL;
-	struct xfs_dquot	*pdqp = NULL;
-	uint			flags = *((uint *)data);
-
-	xfs_dqlock(dqp);
-	if (dqp->dq_flags & XFS_DQ_FREEING) {
-		xfs_dqunlock(dqp);
-		return EAGAIN;
-	}
-
-	/* If this quota has a hint attached, prepare for releasing it now */
-	gdqp = dqp->q_gdquot;
-	if (gdqp)
-		dqp->q_gdquot = NULL;
-
-	pdqp = dqp->q_pdquot;
-	if (pdqp)
-		dqp->q_pdquot = NULL;
-
-	xfs_dqunlock(dqp);
-
-	if (gdqp)
-		xfs_qm_dqrele(gdqp);
-	if (pdqp)
-		xfs_qm_dqrele(pdqp);
-
-	if (flags & XFS_QMOPT_UQUOTA)
-		return xfs_qm_dqpurge(dqp, NULL);
-
-	return 0;
-}
-
 /*
  * Purge the dquot cache.
  */
@@ -241,18 +200,8 @@ xfs_qm_dqpurge_all(
 	struct xfs_mount	*mp,
 	uint			flags)
 {
-	/*
-	 * We have to release group/project dquot hint(s) from the user dquot
-	 * at first if they are there, otherwise we would run into an infinite
-	 * loop while walking through radix tree to purge other type of dquots
-	 * since their refcount is not zero if the user dquot refers to them
-	 * as hint.
-	 *
-	 * Call the special xfs_qm_dqpurge_hints() will end up go through the
-	 * general xfs_qm_dqpurge() against user dquot cache if requested.
-	 */
-	xfs_qm_dquot_walk(mp, XFS_DQ_USER, xfs_qm_dqpurge_hints, &flags);
-
+	if (flags & XFS_QMOPT_UQUOTA)
+		xfs_qm_dquot_walk(mp, XFS_DQ_USER, xfs_qm_dqpurge, NULL);
 	if (flags & XFS_QMOPT_GQUOTA)
 		xfs_qm_dquot_walk(mp, XFS_DQ_GROUP, xfs_qm_dqpurge, NULL);
 	if (flags & XFS_QMOPT_PQUOTA)
@@ -409,7 +358,6 @@ xfs_qm_dqattach_one(
 	xfs_dqid_t	id,
 	uint		type,
 	uint		doalloc,
-	xfs_dquot_t	*udqhint, /* hint */
 	xfs_dquot_t	**IO_idqpp)
 {
 	xfs_dquot_t	*dqp;
@@ -419,9 +367,9 @@ xfs_qm_dqattach_one(
 	error = 0;
 
 	/*
-	 * See if we already have it in the inode itself. IO_idqpp is
-	 * &i_udquot or &i_gdquot. This made the code look weird, but
-	 * made the logic a lot simpler.
+	 * See if we already have it in the inode itself. IO_idqpp is &i_udquot
+	 * or &i_gdquot. This made the code look weird, but made the logic a lot
+	 * simpler.
 	 */
 	dqp = *IO_idqpp;
 	if (dqp) {
@@ -430,49 +378,10 @@ xfs_qm_dqattach_one(
 	}
 
 	/*
-	 * udqhint is the i_udquot field in inode, and is non-NULL only
-	 * when the type arg is group/project. Its purpose is to save a
-	 * lookup by dqid (xfs_qm_dqget) by caching a group dquot inside
-	 * the user dquot.
-	 */
-	if (udqhint) {
-		ASSERT(type == XFS_DQ_GROUP || type == XFS_DQ_PROJ);
-		xfs_dqlock(udqhint);
-
-		/*
-		 * No need to take dqlock to look at the id.
-		 *
-		 * The ID can't change until it gets reclaimed, and it won't
-		 * be reclaimed as long as we have a ref from inode and we
-		 * hold the ilock.
-		 */
-		if (type == XFS_DQ_GROUP)
-			dqp = udqhint->q_gdquot;
-		else
-			dqp = udqhint->q_pdquot;
-		if (dqp && be32_to_cpu(dqp->q_core.d_id) == id) {
-			ASSERT(*IO_idqpp == NULL);
-
-			*IO_idqpp = xfs_qm_dqhold(dqp);
-			xfs_dqunlock(udqhint);
-			return 0;
-		}
-
-		/*
-		 * We can't hold a dquot lock when we call the dqget code.
-		 * We'll deadlock in no time, because of (not conforming to)
-		 * lock ordering - the inodelock comes before any dquot lock,
-		 * and we may drop and reacquire the ilock in xfs_qm_dqget().
-		 */
-		xfs_dqunlock(udqhint);
-	}
-
-	/*
-	 * Find the dquot from somewhere. This bumps the
-	 * reference count of dquot and returns it locked.
-	 * This can return ENOENT if dquot didn't exist on
-	 * disk and we didn't ask it to allocate;
-	 * ESRCH if quotas got turned off suddenly.
+	 * Find the dquot from somewhere. This bumps the reference count of
+	 * dquot and returns it locked.  This can return ENOENT if dquot didn't
+	 * exist on disk and we didn't ask it to allocate; ESRCH if quotas got
+	 * turned off suddenly.
 	 */
 	error = xfs_qm_dqget(ip->i_mount, ip, id, type,
 			     doalloc | XFS_QMOPT_DOWARN, &dqp);
@@ -490,48 +399,6 @@ xfs_qm_dqattach_one(
 	return 0;
 }
 
-
-/*
- * Given a udquot and group/project type, attach the group/project
- * dquot pointer to the udquot as a hint for future lookups.
- */
-STATIC void
-xfs_qm_dqattach_hint(
-	struct xfs_inode	*ip,
-	int			type)
-{
-	struct xfs_dquot **dqhintp;
-	struct xfs_dquot *dqp;
-	struct xfs_dquot *udq = ip->i_udquot;
-
-	ASSERT(type == XFS_DQ_GROUP || type == XFS_DQ_PROJ);
-
-	xfs_dqlock(udq);
-
-	if (type == XFS_DQ_GROUP) {
-		dqp = ip->i_gdquot;
-		dqhintp = &udq->q_gdquot;
-	} else {
-		dqp = ip->i_pdquot;
-		dqhintp = &udq->q_pdquot;
-	}
-
-	if (*dqhintp) {
-		struct xfs_dquot *tmp;
-
-		if (*dqhintp == dqp)
-			goto done;
-
-		tmp = *dqhintp;
-		*dqhintp = NULL;
-		xfs_qm_dqrele(tmp);
-	}
-
-	*dqhintp = xfs_qm_dqhold(dqp);
-done:
-	xfs_dqunlock(udq);
-}
-
 static bool
 xfs_qm_need_dqattach(
 	struct xfs_inode	*ip)
@@ -562,7 +429,6 @@ xfs_qm_dqattach_locked(
 	uint		flags)
 {
 	xfs_mount_t	*mp = ip->i_mount;
-	uint		nquotas = 0;
 	int		error = 0;
 
 	if (!xfs_qm_need_dqattach(ip))
@@ -570,77 +436,39 @@ xfs_qm_dqattach_locked(
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 
-	if (XFS_IS_UQUOTA_ON(mp)) {
+	if (XFS_IS_UQUOTA_ON(mp) && !ip->i_udquot) {
 		error = xfs_qm_dqattach_one(ip, ip->i_d.di_uid, XFS_DQ_USER,
 						flags & XFS_QMOPT_DQALLOC,
-						NULL, &ip->i_udquot);
+						&ip->i_udquot);
 		if (error)
 			goto done;
-		nquotas++;
+		ASSERT(ip->i_udquot);
 	}
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	if (XFS_IS_GQUOTA_ON(mp)) {
+	if (XFS_IS_GQUOTA_ON(mp) && !ip->i_gdquot) {
 		error = xfs_qm_dqattach_one(ip, ip->i_d.di_gid, XFS_DQ_GROUP,
 						flags & XFS_QMOPT_DQALLOC,
-						ip->i_udquot, &ip->i_gdquot);
-		/*
-		 * Don't worry about the udquot that we may have
-		 * attached above. It'll get detached, if not already.
-		 */
+						&ip->i_gdquot);
 		if (error)
 			goto done;
-		nquotas++;
+		ASSERT(ip->i_gdquot);
 	}
 
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	if (XFS_IS_PQUOTA_ON(mp)) {
+	if (XFS_IS_PQUOTA_ON(mp) && !ip->i_pdquot) {
 		error = xfs_qm_dqattach_one(ip, xfs_get_projid(ip), XFS_DQ_PROJ,
 						flags & XFS_QMOPT_DQALLOC,
-						ip->i_udquot, &ip->i_pdquot);
-		/*
-		 * Don't worry about the udquot that we may have
-		 * attached above. It'll get detached, if not already.
-		 */
+						&ip->i_pdquot);
 		if (error)
 			goto done;
-		nquotas++;
+		ASSERT(ip->i_pdquot);
 	}
 
+done:
 	/*
-	 * Attach this group/project quota to the user quota as a hint.
-	 * This WON'T, in general, result in a thrash.
+	 * Don't worry about the dquots that we may have attached before any
+	 * error - they'll get detached later if it has not already been done.
 	 */
-	if (nquotas > 1 && ip->i_udquot) {
-		ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-		ASSERT(ip->i_gdquot || !XFS_IS_GQUOTA_ON(mp));
-		ASSERT(ip->i_pdquot || !XFS_IS_PQUOTA_ON(mp));
-
-		/*
-		 * We do not have i_udquot locked at this point, but this check
-		 * is OK since we don't depend on the i_gdquot to be accurate
-		 * 100% all the time. It is just a hint, and this will
-		 * succeed in general.
-		 */
-		if (ip->i_udquot->q_gdquot != ip->i_gdquot)
-			xfs_qm_dqattach_hint(ip, XFS_DQ_GROUP);
-
-		if (ip->i_udquot->q_pdquot != ip->i_pdquot)
-			xfs_qm_dqattach_hint(ip, XFS_DQ_PROJ);
-	}
-
- done:
-#ifdef DEBUG
-	if (!error) {
-		if (XFS_IS_UQUOTA_ON(mp))
-			ASSERT(ip->i_udquot);
-		if (XFS_IS_GQUOTA_ON(mp))
-			ASSERT(ip->i_gdquot);
-		if (XFS_IS_PQUOTA_ON(mp))
-			ASSERT(ip->i_pdquot);
-	}
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-#endif
 	return error;
 }
 
@@ -865,8 +693,7 @@ xfs_qm_init_quotainfo(
 
 	/* Precalc some constants */
 	qinf->qi_dqchunklen = XFS_FSB_TO_BB(mp, XFS_DQUOT_CLUSTER_SIZE_FSB);
-	qinf->qi_dqperchunk = xfs_calc_dquots_per_chunk(mp,
-							qinf->qi_dqchunklen);
+	qinf->qi_dqperchunk = xfs_calc_dquots_per_chunk(qinf->qi_dqchunklen);
 
 	mp->m_qflags |= (mp->m_sb.sb_qflags & XFS_ALL_QUOTA_CHKD);
 

fs/xfs/xfs_qm_syscalls.c

@@ -278,9 +278,10 @@ xfs_qm_scall_trunc_qfiles(
 	xfs_mount_t	*mp,
 	uint		flags)
 {
-	int		error;
+	int		error = EINVAL;
 
-	if (!xfs_sb_version_hasquota(&mp->m_sb) || flags == 0) {
+	if (!xfs_sb_version_hasquota(&mp->m_sb) || flags == 0 ||
+	    (flags & ~XFS_DQ_ALLTYPES)) {
 		xfs_debug(mp, "%s: flags=%x m_qflags=%x",
 			__func__, flags, mp->m_qflags);
 		return XFS_ERROR(EINVAL);
@@ -959,7 +960,6 @@ xfs_qm_export_flags(
 STATIC int
 xfs_dqrele_inode(
 	struct xfs_inode	*ip,
-	struct xfs_perag	*pag,
 	int			flags,
 	void			*args)
 {

fs/xfs/xfs_quota_defs.h

@@ -156,6 +156,6 @@ typedef __uint16_t	xfs_qwarncnt_t;
 
 extern int xfs_dqcheck(struct xfs_mount *mp, xfs_disk_dquot_t *ddq,
 		       xfs_dqid_t id, uint type, uint flags, char *str);
-extern int xfs_calc_dquots_per_chunk(struct xfs_mount *mp, unsigned int nbblks);
+extern int xfs_calc_dquots_per_chunk(unsigned int nbblks);
 
 #endif	/* __XFS_QUOTA_H__ */

fs/xfs/xfs_quotaops.c

@@ -100,15 +100,35 @@ xfs_fs_set_xstate(
 		if (!XFS_IS_QUOTA_ON(mp))
 			return -EINVAL;
 		return -xfs_qm_scall_quotaoff(mp, flags);
-	case Q_XQUOTARM:
-		if (XFS_IS_QUOTA_ON(mp))
-			return -EINVAL;
-		return -xfs_qm_scall_trunc_qfiles(mp, flags);
 	}
 
 	return -EINVAL;
 }
 
+STATIC int
+xfs_fs_rm_xquota(
+	struct super_block	*sb,
+	unsigned int		uflags)
+{
+	struct xfs_mount	*mp = XFS_M(sb);
+	unsigned int		flags = 0;
+	
+	if (sb->s_flags & MS_RDONLY)
+		return -EROFS;
+
+	if (XFS_IS_QUOTA_ON(mp))
+		return -EINVAL;
+
+	if (uflags & FS_USER_QUOTA)
+		flags |= XFS_DQ_USER;
+	if (uflags & FS_GROUP_QUOTA)
+		flags |= XFS_DQ_GROUP;
+	if (uflags & FS_USER_QUOTA)
+		flags |= XFS_DQ_PROJ;
+
+	return -xfs_qm_scall_trunc_qfiles(mp, flags);
+}	
+
 STATIC int
 xfs_fs_get_dqblk(
 	struct super_block	*sb,
@@ -149,6 +169,7 @@ const struct quotactl_ops xfs_quotactl_operations = {
 	.get_xstatev		= xfs_fs_get_xstatev,
 	.get_xstate		= xfs_fs_get_xstate,
 	.set_xstate		= xfs_fs_set_xstate,
+	.rm_xquota		= xfs_fs_rm_xquota,
 	.get_dqblk		= xfs_fs_get_dqblk,
 	.set_dqblk		= xfs_fs_set_dqblk,
 };

fs/xfs/xfs_rtbitmap.c

@@ -74,7 +74,6 @@ xfs_rtbuf_get(
 				   mp->m_bsize, 0, &bp, NULL);
 	if (error)
 		return error;
-	ASSERT(!xfs_buf_geterror(bp));
 	*bpp = bp;
 	return 0;
 }

fs/xfs/xfs_sb.c

@@ -291,7 +291,8 @@ xfs_mount_validate_sb(
 	    (sbp->sb_imax_pct > 100 /* zero sb_imax_pct is valid */)	||
 	    sbp->sb_dblocks == 0					||
 	    sbp->sb_dblocks > XFS_MAX_DBLOCKS(sbp)			||
-	    sbp->sb_dblocks < XFS_MIN_DBLOCKS(sbp))) {
+	    sbp->sb_dblocks < XFS_MIN_DBLOCKS(sbp)			||
+	    sbp->sb_shared_vn != 0)) {
 		xfs_notice(mp, "SB sanity check failed");
 		return XFS_ERROR(EFSCORRUPTED);
 	}
@@ -333,15 +334,6 @@ xfs_mount_validate_sb(
 		xfs_warn(mp, "Offline file system operation in progress!");
 		return XFS_ERROR(EFSCORRUPTED);
 	}
-
-	/*
-	 * Version 1 directory format has never worked on Linux.
-	 */
-	if (unlikely(!xfs_sb_version_hasdirv2(sbp))) {
-		xfs_warn(mp, "file system using version 1 directory format");
-		return XFS_ERROR(ENOSYS);
-	}
-
 	return 0;
 }
 

fs/xfs/xfs_sb.h

@@ -36,8 +36,6 @@ struct xfs_trans;
 #define	XFS_SB_VERSION_5	5		/* CRC enabled filesystem */
 #define	XFS_SB_VERSION_NUMBITS		0x000f
 #define	XFS_SB_VERSION_ALLFBITS		0xfff0
-#define	XFS_SB_VERSION_SASHFBITS	0xf000
-#define	XFS_SB_VERSION_REALFBITS	0x0ff0
 #define	XFS_SB_VERSION_ATTRBIT		0x0010
 #define	XFS_SB_VERSION_NLINKBIT		0x0020
 #define	XFS_SB_VERSION_QUOTABIT		0x0040
@@ -50,24 +48,15 @@ struct xfs_trans;
 #define	XFS_SB_VERSION_DIRV2BIT		0x2000
 #define	XFS_SB_VERSION_BORGBIT		0x4000	/* ASCII only case-insens. */
 #define	XFS_SB_VERSION_MOREBITSBIT	0x8000
-#define	XFS_SB_VERSION_OKSASHFBITS	\
-	(XFS_SB_VERSION_EXTFLGBIT | \
-	 XFS_SB_VERSION_DIRV2BIT | \
-	 XFS_SB_VERSION_BORGBIT)
-#define	XFS_SB_VERSION_OKREALFBITS	\
-	(XFS_SB_VERSION_ATTRBIT | \
-	 XFS_SB_VERSION_NLINKBIT | \
-	 XFS_SB_VERSION_QUOTABIT | \
-	 XFS_SB_VERSION_ALIGNBIT | \
-	 XFS_SB_VERSION_DALIGNBIT | \
-	 XFS_SB_VERSION_SHAREDBIT | \
-	 XFS_SB_VERSION_LOGV2BIT | \
-	 XFS_SB_VERSION_SECTORBIT | \
-	 XFS_SB_VERSION_MOREBITSBIT)
-#define	XFS_SB_VERSION_OKREALBITS	\
-	(XFS_SB_VERSION_NUMBITS | \
-	 XFS_SB_VERSION_OKREALFBITS | \
-	 XFS_SB_VERSION_OKSASHFBITS)
+
+/*
+ * Supported feature bit list is just all bits in the versionnum field because
+ * we've used them all up and understand them all. Except, of course, for the
+ * shared superblock bit, which nobody knows what it does and so is unsupported.
+ */
+#define	XFS_SB_VERSION_OKBITS		\
+	((XFS_SB_VERSION_NUMBITS | XFS_SB_VERSION_ALLFBITS) & \
+		~XFS_SB_VERSION_SHAREDBIT)
 
 /*
  * There are two words to hold XFS "feature" bits: the original
@@ -76,7 +65,6 @@ struct xfs_trans;
  *
  * These defines represent bits in sb_features2.
  */
-#define XFS_SB_VERSION2_REALFBITS	0x00ffffff	/* Mask: features */
 #define XFS_SB_VERSION2_RESERVED1BIT	0x00000001
 #define XFS_SB_VERSION2_LAZYSBCOUNTBIT	0x00000002	/* Superblk counters */
 #define XFS_SB_VERSION2_RESERVED4BIT	0x00000004
@@ -86,16 +74,11 @@ struct xfs_trans;
 #define XFS_SB_VERSION2_CRCBIT		0x00000100	/* metadata CRCs */
 #define XFS_SB_VERSION2_FTYPE		0x00000200	/* inode type in dir */
 
-#define	XFS_SB_VERSION2_OKREALFBITS	\
+#define	XFS_SB_VERSION2_OKBITS		\
 	(XFS_SB_VERSION2_LAZYSBCOUNTBIT	| \
 	 XFS_SB_VERSION2_ATTR2BIT	| \
 	 XFS_SB_VERSION2_PROJID32BIT	| \
 	 XFS_SB_VERSION2_FTYPE)
-#define	XFS_SB_VERSION2_OKSASHFBITS	\
-	(0)
-#define XFS_SB_VERSION2_OKREALBITS	\
-	(XFS_SB_VERSION2_OKREALFBITS |	\
-	 XFS_SB_VERSION2_OKSASHFBITS )
 
 /*
  * Superblock - in core version.  Must match the ondisk version below.
@@ -345,214 +328,140 @@ typedef enum {
 
 #define	XFS_SB_VERSION_NUM(sbp)	((sbp)->sb_versionnum & XFS_SB_VERSION_NUMBITS)
 
-static inline int xfs_sb_good_version(xfs_sb_t *sbp)
-{
-	/* We always support version 1-3 */
-	if (sbp->sb_versionnum >= XFS_SB_VERSION_1 &&
-	    sbp->sb_versionnum <= XFS_SB_VERSION_3)
-		return 1;
-
-	/* We support version 4 if all feature bits are supported */
-	if (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4) {
-		if ((sbp->sb_versionnum & ~XFS_SB_VERSION_OKREALBITS) ||
-		    ((sbp->sb_versionnum & XFS_SB_VERSION_MOREBITSBIT) &&
-		     (sbp->sb_features2 & ~XFS_SB_VERSION2_OKREALBITS)))
-			return 0;
-
-		if (sbp->sb_shared_vn > XFS_SB_MAX_SHARED_VN)
-			return 0;
-		return 1;
-	}
-	if (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5)
-		return 1;
-
-	return 0;
-}
-
 /*
- * Detect a mismatched features2 field.  Older kernels read/wrote
- * this into the wrong slot, so to be safe we keep them in sync.
+ * The first XFS version we support is a v4 superblock with V2 directories.
  */
-static inline int xfs_sb_has_mismatched_features2(xfs_sb_t *sbp)
+static inline bool xfs_sb_good_v4_features(struct xfs_sb *sbp)
 {
-	return (sbp->sb_bad_features2 != sbp->sb_features2);
-}
-
-static inline unsigned xfs_sb_version_tonew(unsigned v)
-{
-	if (v == XFS_SB_VERSION_1)
-		return XFS_SB_VERSION_4;
+	if (!(sbp->sb_versionnum & XFS_SB_VERSION_DIRV2BIT))
+		return false;
 
-	if (v == XFS_SB_VERSION_2)
-		return XFS_SB_VERSION_4 | XFS_SB_VERSION_ATTRBIT;
+	/* check for unknown features in the fs */
+	if ((sbp->sb_versionnum & ~XFS_SB_VERSION_OKBITS) ||
+	    ((sbp->sb_versionnum & XFS_SB_VERSION_MOREBITSBIT) &&
+	     (sbp->sb_features2 & ~XFS_SB_VERSION2_OKBITS)))
+		return false;
 
-	return XFS_SB_VERSION_4 | XFS_SB_VERSION_ATTRBIT |
-		XFS_SB_VERSION_NLINKBIT;
+	return true;
 }
 
-static inline unsigned xfs_sb_version_toold(unsigned v)
+static inline bool xfs_sb_good_version(struct xfs_sb *sbp)
 {
-	if (v & (XFS_SB_VERSION_QUOTABIT | XFS_SB_VERSION_ALIGNBIT))
-		return 0;
-	if (v & XFS_SB_VERSION_NLINKBIT)
-		return XFS_SB_VERSION_3;
-	if (v & XFS_SB_VERSION_ATTRBIT)
-		return XFS_SB_VERSION_2;
-	return XFS_SB_VERSION_1;
-}
-
-static inline int xfs_sb_version_hasattr(xfs_sb_t *sbp)
-{
-	return sbp->sb_versionnum == XFS_SB_VERSION_2 ||
-		sbp->sb_versionnum == XFS_SB_VERSION_3 ||
-		(XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4 &&
-		 (sbp->sb_versionnum & XFS_SB_VERSION_ATTRBIT));
+	if (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5)
+		return true;
+	if (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4)
+		return xfs_sb_good_v4_features(sbp);
+	return false;
 }
 
-static inline void xfs_sb_version_addattr(xfs_sb_t *sbp)
+/*
+ * Detect a mismatched features2 field.  Older kernels read/wrote
+ * this into the wrong slot, so to be safe we keep them in sync.
+ */
+static inline bool xfs_sb_has_mismatched_features2(struct xfs_sb *sbp)
 {
-	if (sbp->sb_versionnum == XFS_SB_VERSION_1)
-		sbp->sb_versionnum = XFS_SB_VERSION_2;
-	else if (XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4)
-		sbp->sb_versionnum |= XFS_SB_VERSION_ATTRBIT;
-	else
-		sbp->sb_versionnum = XFS_SB_VERSION_4 | XFS_SB_VERSION_ATTRBIT;
+	return sbp->sb_bad_features2 != sbp->sb_features2;
 }
 
-static inline int xfs_sb_version_hasnlink(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hasattr(struct xfs_sb *sbp)
 {
-	return sbp->sb_versionnum == XFS_SB_VERSION_3 ||
-		 (XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4 &&
-		  (sbp->sb_versionnum & XFS_SB_VERSION_NLINKBIT));
+	return (sbp->sb_versionnum & XFS_SB_VERSION_ATTRBIT);
 }
 
-static inline void xfs_sb_version_addnlink(xfs_sb_t *sbp)
+static inline void xfs_sb_version_addattr(struct xfs_sb *sbp)
 {
-	if (sbp->sb_versionnum <= XFS_SB_VERSION_2)
-		sbp->sb_versionnum = XFS_SB_VERSION_3;
-	else
-		sbp->sb_versionnum |= XFS_SB_VERSION_NLINKBIT;
+	sbp->sb_versionnum |= XFS_SB_VERSION_ATTRBIT;
 }
 
-static inline int xfs_sb_version_hasquota(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hasquota(struct xfs_sb *sbp)
 {
-	return XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4 &&
-		(sbp->sb_versionnum & XFS_SB_VERSION_QUOTABIT);
+	return (sbp->sb_versionnum & XFS_SB_VERSION_QUOTABIT);
 }
 
-static inline void xfs_sb_version_addquota(xfs_sb_t *sbp)
+static inline void xfs_sb_version_addquota(struct xfs_sb *sbp)
 {
-	if (XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4)
-		sbp->sb_versionnum |= XFS_SB_VERSION_QUOTABIT;
-	else
-		sbp->sb_versionnum = xfs_sb_version_tonew(sbp->sb_versionnum) |
-					XFS_SB_VERSION_QUOTABIT;
+	sbp->sb_versionnum |= XFS_SB_VERSION_QUOTABIT;
 }
 
-static inline int xfs_sb_version_hasalign(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hasalign(struct xfs_sb *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) ||
-	       (XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4 &&
+	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 ||
 		(sbp->sb_versionnum & XFS_SB_VERSION_ALIGNBIT));
 }
 
-static inline int xfs_sb_version_hasdalign(xfs_sb_t *sbp)
-{
-	return XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4 &&
-		(sbp->sb_versionnum & XFS_SB_VERSION_DALIGNBIT);
-}
-
-static inline int xfs_sb_version_hasshared(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hasdalign(struct xfs_sb *sbp)
 {
-	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
-		(sbp->sb_versionnum & XFS_SB_VERSION_SHAREDBIT);
+	return (sbp->sb_versionnum & XFS_SB_VERSION_DALIGNBIT);
 }
 
-static inline int xfs_sb_version_hasdirv2(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_haslogv2(struct xfs_sb *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) ||
-	       (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
-		(sbp->sb_versionnum & XFS_SB_VERSION_DIRV2BIT));
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 ||
+	       (sbp->sb_versionnum & XFS_SB_VERSION_LOGV2BIT);
 }
 
-static inline int xfs_sb_version_haslogv2(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hasextflgbit(struct xfs_sb *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) ||
-	       (XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4 &&
-		(sbp->sb_versionnum & XFS_SB_VERSION_LOGV2BIT));
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 ||
+	       (sbp->sb_versionnum & XFS_SB_VERSION_EXTFLGBIT);
 }
 
-static inline int xfs_sb_version_hasextflgbit(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hassector(struct xfs_sb *sbp)
 {
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) ||
-	       (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
-		(sbp->sb_versionnum & XFS_SB_VERSION_EXTFLGBIT));
+	return (sbp->sb_versionnum & XFS_SB_VERSION_SECTORBIT);
 }
 
-static inline int xfs_sb_version_hassector(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hasasciici(struct xfs_sb *sbp)
 {
-	return XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4 &&
-		(sbp->sb_versionnum & XFS_SB_VERSION_SECTORBIT);
+	return (sbp->sb_versionnum & XFS_SB_VERSION_BORGBIT);
 }
 
-static inline int xfs_sb_version_hasasciici(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hasmorebits(struct xfs_sb *sbp)
 {
-	return XFS_SB_VERSION_NUM(sbp) >= XFS_SB_VERSION_4 &&
-		(sbp->sb_versionnum & XFS_SB_VERSION_BORGBIT);
-}
-
-static inline int xfs_sb_version_hasmorebits(xfs_sb_t *sbp)
-{
-	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) ||
-	       (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_4 &&
-		(sbp->sb_versionnum & XFS_SB_VERSION_MOREBITSBIT));
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5 ||
+	       (sbp->sb_versionnum & XFS_SB_VERSION_MOREBITSBIT);
 }
 
 /*
  * sb_features2 bit version macros.
- *
- * For example, for a bit defined as XFS_SB_VERSION2_FUNBIT, has a macro:
- *
- * SB_VERSION_HASFUNBIT(xfs_sb_t *sbp)
- *	((xfs_sb_version_hasmorebits(sbp) &&
- *	 ((sbp)->sb_features2 & XFS_SB_VERSION2_FUNBIT)
  */
-
-static inline int xfs_sb_version_haslazysbcount(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_haslazysbcount(struct xfs_sb *sbp)
 {
 	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) ||
 	       (xfs_sb_version_hasmorebits(sbp) &&
 		(sbp->sb_features2 & XFS_SB_VERSION2_LAZYSBCOUNTBIT));
 }
 
-static inline int xfs_sb_version_hasattr2(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hasattr2(struct xfs_sb *sbp)
 {
 	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) ||
 	       (xfs_sb_version_hasmorebits(sbp) &&
 		(sbp->sb_features2 & XFS_SB_VERSION2_ATTR2BIT));
 }
 
-static inline void xfs_sb_version_addattr2(xfs_sb_t *sbp)
+static inline void xfs_sb_version_addattr2(struct xfs_sb *sbp)
 {
 	sbp->sb_versionnum |= XFS_SB_VERSION_MOREBITSBIT;
 	sbp->sb_features2 |= XFS_SB_VERSION2_ATTR2BIT;
+	sbp->sb_bad_features2 |= XFS_SB_VERSION2_ATTR2BIT;
 }
 
-static inline void xfs_sb_version_removeattr2(xfs_sb_t *sbp)
+static inline void xfs_sb_version_removeattr2(struct xfs_sb *sbp)
 {
 	sbp->sb_features2 &= ~XFS_SB_VERSION2_ATTR2BIT;
+	sbp->sb_bad_features2 &= ~XFS_SB_VERSION2_ATTR2BIT;
 	if (!sbp->sb_features2)
 		sbp->sb_versionnum &= ~XFS_SB_VERSION_MOREBITSBIT;
 }
 
-static inline int xfs_sb_version_hasprojid32bit(xfs_sb_t *sbp)
+static inline bool xfs_sb_version_hasprojid32bit(struct xfs_sb *sbp)
 {
 	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) ||
 	       (xfs_sb_version_hasmorebits(sbp) &&
 		(sbp->sb_features2 & XFS_SB_VERSION2_PROJID32BIT));
 }
 
-static inline void xfs_sb_version_addprojid32bit(xfs_sb_t *sbp)
+static inline void xfs_sb_version_addprojid32bit(struct xfs_sb *sbp)
 {
 	sbp->sb_versionnum |= XFS_SB_VERSION_MOREBITSBIT;
 	sbp->sb_features2 |= XFS_SB_VERSION2_PROJID32BIT;
@@ -587,7 +496,9 @@ xfs_sb_has_compat_feature(
 	return (sbp->sb_features_compat & feature) != 0;
 }
 
-#define XFS_SB_FEAT_RO_COMPAT_ALL 0
+#define XFS_SB_FEAT_RO_COMPAT_FINOBT   (1 << 0)		/* free inode btree */
+#define XFS_SB_FEAT_RO_COMPAT_ALL \
+		(XFS_SB_FEAT_RO_COMPAT_FINOBT)
 #define XFS_SB_FEAT_RO_COMPAT_UNKNOWN	~XFS_SB_FEAT_RO_COMPAT_ALL
 static inline bool
 xfs_sb_has_ro_compat_feature(
@@ -623,12 +534,12 @@ xfs_sb_has_incompat_log_feature(
 /*
  * V5 superblock specific feature checks
  */
-static inline int xfs_sb_version_hascrc(xfs_sb_t *sbp)
+static inline int xfs_sb_version_hascrc(struct xfs_sb *sbp)
 {
 	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5;
 }
 
-static inline int xfs_sb_version_has_pquotino(xfs_sb_t *sbp)
+static inline int xfs_sb_version_has_pquotino(struct xfs_sb *sbp)
 {
 	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5;
 }
@@ -641,6 +552,12 @@ static inline int xfs_sb_version_hasftype(struct xfs_sb *sbp)
 		 (sbp->sb_features2 & XFS_SB_VERSION2_FTYPE));
 }
 
+static inline int xfs_sb_version_hasfinobt(xfs_sb_t *sbp)
+{
+	return (XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5) &&
+		(sbp->sb_features_ro_compat & XFS_SB_FEAT_RO_COMPAT_FINOBT);
+}
+
 /*
  * end of superblock version macros
  */

fs/xfs/xfs_shared.h

@@ -238,7 +238,7 @@ int	xfs_log_calc_minimum_size(struct xfs_mount *);
 int xfs_symlink_blocks(struct xfs_mount *mp, int pathlen);
 int xfs_symlink_hdr_set(struct xfs_mount *mp, xfs_ino_t ino, uint32_t offset,
 			uint32_t size, struct xfs_buf *bp);
-bool xfs_symlink_hdr_ok(struct xfs_mount *mp, xfs_ino_t ino, uint32_t offset,
+bool xfs_symlink_hdr_ok(xfs_ino_t ino, uint32_t offset,
 			uint32_t size, struct xfs_buf *bp);
 void xfs_symlink_local_to_remote(struct xfs_trans *tp, struct xfs_buf *bp,
 				 struct xfs_inode *ip, struct xfs_ifork *ifp);

fs/xfs/xfs_stats.c

@@ -59,6 +59,7 @@ static int xfs_stat_proc_show(struct seq_file *m, void *v)
 		{ "abtc2",		XFSSTAT_END_ABTC_V2		},
 		{ "bmbt2",		XFSSTAT_END_BMBT_V2		},
 		{ "ibt2",		XFSSTAT_END_IBT_V2		},
+		{ "fibt2",		XFSSTAT_END_FIBT_V2		},
 		/* we print both series of quota information together */
 		{ "qm",			XFSSTAT_END_QM			},
 	};

fs/xfs/xfs_stats.h

@@ -183,7 +183,23 @@ struct xfsstats {
 	__uint32_t		xs_ibt_2_alloc;
 	__uint32_t		xs_ibt_2_free;
 	__uint32_t		xs_ibt_2_moves;
-#define XFSSTAT_END_XQMSTAT		(XFSSTAT_END_IBT_V2+6)
+#define XFSSTAT_END_FIBT_V2		(XFSSTAT_END_IBT_V2+15)
+	__uint32_t		xs_fibt_2_lookup;
+	__uint32_t		xs_fibt_2_compare;
+	__uint32_t		xs_fibt_2_insrec;
+	__uint32_t		xs_fibt_2_delrec;
+	__uint32_t		xs_fibt_2_newroot;
+	__uint32_t		xs_fibt_2_killroot;
+	__uint32_t		xs_fibt_2_increment;
+	__uint32_t		xs_fibt_2_decrement;
+	__uint32_t		xs_fibt_2_lshift;
+	__uint32_t		xs_fibt_2_rshift;
+	__uint32_t		xs_fibt_2_split;
+	__uint32_t		xs_fibt_2_join;
+	__uint32_t		xs_fibt_2_alloc;
+	__uint32_t		xs_fibt_2_free;
+	__uint32_t		xs_fibt_2_moves;
+#define XFSSTAT_END_XQMSTAT		(XFSSTAT_END_FIBT_V2+6)
 	__uint32_t		xs_qm_dqreclaims;
 	__uint32_t		xs_qm_dqreclaim_misses;
 	__uint32_t		xs_qm_dquot_dups;

fs/xfs/xfs_super.c

@@ -765,20 +765,18 @@ xfs_open_devices(
 	 * Setup xfs_mount buffer target pointers
 	 */
 	error = ENOMEM;
-	mp->m_ddev_targp = xfs_alloc_buftarg(mp, ddev, 0, mp->m_fsname);
+	mp->m_ddev_targp = xfs_alloc_buftarg(mp, ddev);
 	if (!mp->m_ddev_targp)
 		goto out_close_rtdev;
 
 	if (rtdev) {
-		mp->m_rtdev_targp = xfs_alloc_buftarg(mp, rtdev, 1,
-							mp->m_fsname);
+		mp->m_rtdev_targp = xfs_alloc_buftarg(mp, rtdev);
 		if (!mp->m_rtdev_targp)
 			goto out_free_ddev_targ;
 	}
 
 	if (logdev && logdev != ddev) {
-		mp->m_logdev_targp = xfs_alloc_buftarg(mp, logdev, 1,
-							mp->m_fsname);
+		mp->m_logdev_targp = xfs_alloc_buftarg(mp, logdev);
 		if (!mp->m_logdev_targp)
 			goto out_free_rtdev_targ;
 	} else {
@@ -811,8 +809,7 @@ xfs_setup_devices(
 {
 	int			error;
 
-	error = xfs_setsize_buftarg(mp->m_ddev_targp, mp->m_sb.sb_blocksize,
-				    mp->m_sb.sb_sectsize);
+	error = xfs_setsize_buftarg(mp->m_ddev_targp, mp->m_sb.sb_sectsize);
 	if (error)
 		return error;
 
@@ -822,14 +819,12 @@ xfs_setup_devices(
 		if (xfs_sb_version_hassector(&mp->m_sb))
 			log_sector_size = mp->m_sb.sb_logsectsize;
 		error = xfs_setsize_buftarg(mp->m_logdev_targp,
-					    mp->m_sb.sb_blocksize,
 					    log_sector_size);
 		if (error)
 			return error;
 	}
 	if (mp->m_rtdev_targp) {
 		error = xfs_setsize_buftarg(mp->m_rtdev_targp,
-					    mp->m_sb.sb_blocksize,
 					    mp->m_sb.sb_sectsize);
 		if (error)
 			return error;
@@ -1754,13 +1749,9 @@ init_xfs_fs(void)
 	if (error)
 		goto out_destroy_wq;
 
-	error = xfs_filestream_init();
-	if (error)
-		goto out_mru_cache_uninit;
-
 	error = xfs_buf_init();
 	if (error)
-		goto out_filestream_uninit;
+		goto out_mru_cache_uninit;
 
 	error = xfs_init_procfs();
 	if (error)
@@ -1787,8 +1778,6 @@ init_xfs_fs(void)
 	xfs_cleanup_procfs();
  out_buf_terminate:
 	xfs_buf_terminate();
- out_filestream_uninit:
-	xfs_filestream_uninit();
  out_mru_cache_uninit:
 	xfs_mru_cache_uninit();
  out_destroy_wq:
@@ -1807,7 +1796,6 @@ exit_xfs_fs(void)
 	xfs_sysctl_unregister();
 	xfs_cleanup_procfs();
 	xfs_buf_terminate();
-	xfs_filestream_uninit();
 	xfs_mru_cache_uninit();
 	xfs_destroy_workqueues();
 	xfs_destroy_zones();

fs/xfs/xfs_symlink.c

@@ -27,6 +27,7 @@
 #include "xfs_ag.h"
 #include "xfs_mount.h"
 #include "xfs_da_format.h"
+#include "xfs_da_btree.h"
 #include "xfs_dir2.h"
 #include "xfs_inode.h"
 #include "xfs_ialloc.h"
@@ -92,7 +93,7 @@ xfs_readlink_bmap(
 
 		cur_chunk = bp->b_addr;
 		if (xfs_sb_version_hascrc(&mp->m_sb)) {
-			if (!xfs_symlink_hdr_ok(mp, ip->i_ino, offset,
+			if (!xfs_symlink_hdr_ok(ip->i_ino, offset,
 							byte_cnt, bp)) {
 				error = EFSCORRUPTED;
 				xfs_alert(mp,

fs/xfs/xfs_symlink_remote.c

@@ -80,7 +80,6 @@ xfs_symlink_hdr_set(
  */
 bool
 xfs_symlink_hdr_ok(
-	struct xfs_mount	*mp,
 	xfs_ino_t		ino,
 	uint32_t		offset,
 	uint32_t		size,

fs/xfs/xfs_trace.c

@@ -46,6 +46,7 @@
 #include "xfs_log_recover.h"
 #include "xfs_inode_item.h"
 #include "xfs_bmap_btree.h"
+#include "xfs_filestream.h"
 
 /*
  * We include this last to have the helpers above available for the trace

fs/xfs/xfs_trace.h

@@ -538,6 +538,64 @@ DEFINE_BUF_ITEM_EVENT(xfs_trans_bhold_release);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_binval);
 DEFINE_BUF_ITEM_EVENT(xfs_trans_buf_ordered);
 
+DECLARE_EVENT_CLASS(xfs_filestream_class,
+	TP_PROTO(struct xfs_inode *ip, xfs_agnumber_t agno),
+	TP_ARGS(ip, agno),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_agnumber_t, agno)
+		__field(int, streams)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->agno = agno;
+		__entry->streams = xfs_filestream_peek_ag(ip->i_mount, agno);
+	),
+	TP_printk("dev %d:%d ino 0x%llx agno %u streams %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->agno,
+		  __entry->streams)
+)
+#define DEFINE_FILESTREAM_EVENT(name) \
+DEFINE_EVENT(xfs_filestream_class, name, \
+	TP_PROTO(struct xfs_inode *ip, xfs_agnumber_t agno), \
+	TP_ARGS(ip, agno))
+DEFINE_FILESTREAM_EVENT(xfs_filestream_free);
+DEFINE_FILESTREAM_EVENT(xfs_filestream_lookup);
+DEFINE_FILESTREAM_EVENT(xfs_filestream_scan);
+
+TRACE_EVENT(xfs_filestream_pick,
+	TP_PROTO(struct xfs_inode *ip, xfs_agnumber_t agno,
+		 xfs_extlen_t free, int nscan),
+	TP_ARGS(ip, agno, free, nscan),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(xfs_agnumber_t, agno)
+		__field(int, streams)
+		__field(xfs_extlen_t, free)
+		__field(int, nscan)
+	),
+	TP_fast_assign(
+		__entry->dev = VFS_I(ip)->i_sb->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->agno = agno;
+		__entry->streams = xfs_filestream_peek_ag(ip->i_mount, agno);
+		__entry->free = free;
+		__entry->nscan = nscan;
+	),
+	TP_printk("dev %d:%d ino 0x%llx agno %u streams %d free %d nscan %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->agno,
+		  __entry->streams,
+		  __entry->free,
+		  __entry->nscan)
+);
+
 DECLARE_EVENT_CLASS(xfs_lock_class,
 	TP_PROTO(struct xfs_inode *ip, unsigned lock_flags,
 		 unsigned long caller_ip),

fs/xfs/xfs_trans.c

@@ -827,7 +827,7 @@ xfs_trans_committed_bulk(
 		xfs_log_item_batch_insert(ailp, &cur, log_items, i, commit_lsn);
 
 	spin_lock(&ailp->xa_lock);
-	xfs_trans_ail_cursor_done(ailp, &cur);
+	xfs_trans_ail_cursor_done(&cur);
 	spin_unlock(&ailp->xa_lock);
 }
 

fs/xfs/xfs_trans_ail.c

@@ -173,7 +173,6 @@ xfs_trans_ail_cursor_next(
  */
 void
 xfs_trans_ail_cursor_done(
-	struct xfs_ail		*ailp,
 	struct xfs_ail_cursor	*cur)
 {
 	cur->item = NULL;
@@ -368,7 +367,7 @@ xfsaild_push(
 		 * If the AIL is empty or our push has reached the end we are
 		 * done now.
 		 */
-		xfs_trans_ail_cursor_done(ailp, &cur);
+		xfs_trans_ail_cursor_done(&cur);
 		spin_unlock(&ailp->xa_lock);
 		goto out_done;
 	}
@@ -453,7 +452,7 @@ xfsaild_push(
 			break;
 		lsn = lip->li_lsn;
 	}
-	xfs_trans_ail_cursor_done(ailp, &cur);
+	xfs_trans_ail_cursor_done(&cur);
 	spin_unlock(&ailp->xa_lock);
 
 	if (xfs_buf_delwri_submit_nowait(&ailp->xa_buf_list))

fs/xfs/xfs_trans_priv.h

@@ -133,8 +133,7 @@ struct xfs_log_item *	xfs_trans_ail_cursor_last(struct xfs_ail *ailp,
 					xfs_lsn_t lsn);
 struct xfs_log_item *	xfs_trans_ail_cursor_next(struct xfs_ail *ailp,
 					struct xfs_ail_cursor *cur);
-void			xfs_trans_ail_cursor_done(struct xfs_ail *ailp,
-					struct xfs_ail_cursor *cur);
+void			xfs_trans_ail_cursor_done(struct xfs_ail_cursor *cur);
 
 #if BITS_PER_LONG != 64
 static inline void

fs/xfs/xfs_trans_resv.c

@@ -26,6 +26,7 @@
 #include "xfs_ag.h"
 #include "xfs_mount.h"
 #include "xfs_da_format.h"
+#include "xfs_da_btree.h"
 #include "xfs_inode.h"
 #include "xfs_bmap_btree.h"
 #include "xfs_ialloc.h"
@@ -105,6 +106,47 @@ xfs_calc_inode_res(
 		 2 * XFS_BMBT_BLOCK_LEN(mp));
 }
 
+/*
+ * The free inode btree is a conditional feature and the log reservation
+ * requirements differ slightly from that of the traditional inode allocation
+ * btree. The finobt tracks records for inode chunks with at least one free
+ * inode. A record can be removed from the tree for an inode allocation
+ * or free and thus the finobt reservation is unconditional across:
+ *
+ * 	- inode allocation
+ * 	- inode free
+ * 	- inode chunk allocation
+ *
+ * The 'modify' param indicates to include the record modification scenario. The
+ * 'alloc' param indicates to include the reservation for free space btree
+ * modifications on behalf of finobt modifications. This is required only for
+ * transactions that do not already account for free space btree modifications.
+ *
+ * the free inode btree: max depth * block size
+ * the allocation btrees: 2 trees * (max depth - 1) * block size
+ * the free inode btree entry: block size
+ */
+STATIC uint
+xfs_calc_finobt_res(
+	struct xfs_mount 	*mp,
+	int			alloc,
+	int			modify)
+{
+	uint res;
+
+	if (!xfs_sb_version_hasfinobt(&mp->m_sb))
+		return 0;
+
+	res = xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1));
+	if (alloc)
+		res += xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1), 
+					XFS_FSB_TO_B(mp, 1));
+	if (modify)
+		res += (uint)XFS_FSB_TO_B(mp, 1);
+
+	return res;
+}
+
 /*
  * Various log reservation values.
  *
@@ -302,6 +344,7 @@ xfs_calc_remove_reservation(
  *    the superblock for the nlink flag: sector size
  *    the directory btree: (max depth + v2) * dir block size
  *    the directory inode's bmap btree: (max depth + v2) * block size
+ *    the finobt (record modification and allocation btrees)
  */
 STATIC uint
 xfs_calc_create_resv_modify(
@@ -310,7 +353,8 @@ xfs_calc_create_resv_modify(
 	return xfs_calc_inode_res(mp, 2) +
 		xfs_calc_buf_res(1, mp->m_sb.sb_sectsize) +
 		(uint)XFS_FSB_TO_B(mp, 1) +
-		xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1));
+		xfs_calc_buf_res(XFS_DIROP_LOG_COUNT(mp), XFS_FSB_TO_B(mp, 1)) +
+		xfs_calc_finobt_res(mp, 1, 1);
 }
 
 /*
@@ -348,6 +392,7 @@ __xfs_calc_create_reservation(
  *    the superblock for the nlink flag: sector size
  *    the inode btree: max depth * blocksize
  *    the allocation btrees: 2 trees * (max depth - 1) * block size
+ *    the finobt (record insertion)
  */
 STATIC uint
 xfs_calc_icreate_resv_alloc(
@@ -357,7 +402,8 @@ xfs_calc_icreate_resv_alloc(
 		mp->m_sb.sb_sectsize +
 		xfs_calc_buf_res(mp->m_in_maxlevels, XFS_FSB_TO_B(mp, 1)) +
 		xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
-				 XFS_FSB_TO_B(mp, 1));
+				 XFS_FSB_TO_B(mp, 1)) +
+		xfs_calc_finobt_res(mp, 0, 0);
 }
 
 STATIC uint
@@ -425,6 +471,7 @@ xfs_calc_symlink_reservation(
  *    the on disk inode before ours in the agi hash list: inode cluster size
  *    the inode btree: max depth * blocksize
  *    the allocation btrees: 2 trees * (max depth - 1) * block size
+ *    the finobt (record insertion, removal or modification)
  */
 STATIC uint
 xfs_calc_ifree_reservation(
@@ -439,7 +486,8 @@ xfs_calc_ifree_reservation(
 		xfs_calc_buf_res(2 + mp->m_ialloc_blks +
 				 mp->m_in_maxlevels, 0) +
 		xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),
-				 XFS_FSB_TO_B(mp, 1));
+				 XFS_FSB_TO_B(mp, 1)) +
+		xfs_calc_finobt_res(mp, 0, 1);
 }
 
 /*
@@ -562,7 +610,7 @@ xfs_calc_addafork_reservation(
 	return XFS_DQUOT_LOGRES(mp) +
 		xfs_calc_inode_res(mp, 1) +
 		xfs_calc_buf_res(2, mp->m_sb.sb_sectsize) +
-		xfs_calc_buf_res(1, mp->m_dirblksize) +
+		xfs_calc_buf_res(1, mp->m_dir_geo->blksize) +
 		xfs_calc_buf_res(XFS_DAENTER_BMAP1B(mp, XFS_DATA_FORK) + 1,
 				 XFS_FSB_TO_B(mp, 1)) +
 		xfs_calc_buf_res(XFS_ALLOCFREE_LOG_COUNT(mp, 1),

fs/xfs/xfs_trans_space.h

@@ -28,7 +28,8 @@
 	(((b + XFS_MAX_CONTIG_EXTENTS_PER_BLOCK(mp) - 1) / \
 	  XFS_MAX_CONTIG_EXTENTS_PER_BLOCK(mp)) * \
 	  XFS_EXTENTADD_SPACE_RES(mp,w))
-#define	XFS_DAENTER_1B(mp,w)	((w) == XFS_DATA_FORK ? (mp)->m_dirblkfsbs : 1)
+#define	XFS_DAENTER_1B(mp,w)	\
+	((w) == XFS_DATA_FORK ? (mp)->m_dir_geo->fsbcount : 1)
 #define	XFS_DAENTER_DBS(mp,w)	\
 	(XFS_DA_NODE_MAXDEPTH + (((w) == XFS_DATA_FORK) ? 2 : 0))
 #define	XFS_DAENTER_BLOCKS(mp,w)	\
@@ -47,13 +48,15 @@
 #define	XFS_DIRREMOVE_SPACE_RES(mp)	\
 	XFS_DAREMOVE_SPACE_RES(mp, XFS_DATA_FORK)
 #define	XFS_IALLOC_SPACE_RES(mp)	\
-	((mp)->m_ialloc_blks + (mp)->m_in_maxlevels - 1)
+	((mp)->m_ialloc_blks + \
+	 (xfs_sb_version_hasfinobt(&mp->m_sb) ? 2 : 1 * \
+	  ((mp)->m_in_maxlevels - 1)))
 
 /*
  * Space reservation values for various transactions.
  */
 #define	XFS_ADDAFORK_SPACE_RES(mp)	\
-	((mp)->m_dirblkfsbs + XFS_DAENTER_BMAP1B(mp, XFS_DATA_FORK))
+	((mp)->m_dir_geo->fsbcount + XFS_DAENTER_BMAP1B(mp, XFS_DATA_FORK))
 #define	XFS_ATTRRM_SPACE_RES(mp)	\
 	XFS_DAREMOVE_SPACE_RES(mp, XFS_ATTR_FORK)
 /* This macro is not used - see inline code in xfs_attr_set */
@@ -82,5 +85,8 @@
 	(XFS_DIRREMOVE_SPACE_RES(mp) + XFS_DIRENTER_SPACE_RES(mp,nl))
 #define	XFS_SYMLINK_SPACE_RES(mp,nl,b)	\
 	(XFS_IALLOC_SPACE_RES(mp) + XFS_DIRENTER_SPACE_RES(mp,nl) + (b))
+#define XFS_IFREE_SPACE_RES(mp)		\
+	(xfs_sb_version_hasfinobt(&mp->m_sb) ? (mp)->m_in_maxlevels : 0)
+
 
 #endif	/* __XFS_TRANS_SPACE_H__ */

fs/xfs/xfs_types.h

@@ -134,7 +134,7 @@ typedef enum {
 
 typedef enum {
 	XFS_BTNUM_BNOi, XFS_BTNUM_CNTi, XFS_BTNUM_BMAPi, XFS_BTNUM_INOi,
-	XFS_BTNUM_MAX
+	XFS_BTNUM_FINOi, XFS_BTNUM_MAX
 } xfs_btnum_t;
 
 struct xfs_name {

include/linux/quota.h

@@ -329,6 +329,7 @@ struct quotactl_ops {
 	int (*get_xstate)(struct super_block *, struct fs_quota_stat *);
 	int (*set_xstate)(struct super_block *, unsigned int, int);
 	int (*get_xstatev)(struct super_block *, struct fs_quota_statv *);
+	int (*rm_xquota)(struct super_block *, unsigned int);
 };
 
 struct quota_format_type {