Merge branch 'master' into for-next

Sync with Linus' branch in order to be able to apply fixups
of more recent patches.

166 files changed, 11117 insertions, 10226 deletions

diff --git a/fs/xfs/Makefile b/fs/xfs/Makefile
index aceca2f9a3db..04611a1068b4 100644
--- a/fs/xfs/Makefile
+++ b/fs/xfs/Makefile
@@ -7,8 +7,6 @@
 ccflags-y += -I $(srctree)/$(src)		# needed for trace events
 ccflags-y += -I $(srctree)/$(src)/libxfs
 
-ccflags-$(CONFIG_XFS_DEBUG) += -g
-
 obj-$(CONFIG_XFS_FS)		+= xfs.o
 
 # this one should be compiled first, as the tracing macros can easily blow up
@@ -26,6 +24,7 @@ xfs-y				+= $(addprefix libxfs/, \
 				   xfs_bmap.o \
 				   xfs_bmap_btree.o \
 				   xfs_btree.o \
+				   xfs_btree_staging.o \
 				   xfs_da_btree.o \
 				   xfs_defer.o \
 				   xfs_dir2.o \
@@ -100,9 +99,12 @@ xfs-y				+= xfs_log.o \
 				   xfs_log_cil.o \
 				   xfs_bmap_item.o \
 				   xfs_buf_item.o \
+				   xfs_buf_item_recover.o \
+				   xfs_dquot_item_recover.o \
 				   xfs_extfree_item.o \
 				   xfs_icreate_item.o \
 				   xfs_inode_item.o \
+				   xfs_inode_item_recover.o \
 				   xfs_refcount_item.o \
 				   xfs_rmap_item.o \
 				   xfs_log_recover.o \
diff --git a/fs/xfs/kmem.c b/fs/xfs/kmem.c
index 1da94237a8cf..e841ed781a25 100644
--- a/fs/xfs/kmem.c
+++ b/fs/xfs/kmem.c
@@ -48,7 +48,7 @@ __kmem_vmalloc(size_t size, xfs_km_flags_t flags)
 	if (flags & KM_NOFS)
 		nofs_flag = memalloc_nofs_save();
 
-	ptr = __vmalloc(size, lflags, PAGE_KERNEL);
+	ptr = __vmalloc(size, lflags);
 
 	if (flags & KM_NOFS)
 		memalloc_nofs_restore(nofs_flag);
@@ -115,24 +115,3 @@ kmem_realloc(const void *old, size_t newsize, xfs_km_flags_t flags)
 		congestion_wait(BLK_RW_ASYNC, HZ/50);
 	} while (1);
 }
-
-void *
-kmem_zone_alloc(kmem_zone_t *zone, xfs_km_flags_t flags)
-{
-	int	retries = 0;
-	gfp_t	lflags = kmem_flags_convert(flags);
-	void	*ptr;
-
-	trace_kmem_zone_alloc(kmem_cache_size(zone), flags, _RET_IP_);
-	do {
-		ptr = kmem_cache_alloc(zone, lflags);
-		if (ptr || (flags & KM_MAYFAIL))
-			return ptr;
-		if (!(++retries % 100))
-			xfs_err(NULL,
-		"%s(%u) possible memory allocation deadlock in %s (mode:0x%x)",
-				current->comm, current->pid,
-				__func__, lflags);
-		congestion_wait(BLK_RW_ASYNC, HZ/50);
-	} while (1);
-}
diff --git a/fs/xfs/kmem.h b/fs/xfs/kmem.h
index 6143117770e9..8e8555817e6d 100644
--- a/fs/xfs/kmem.h
+++ b/fs/xfs/kmem.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000-2005 Silicon Graphics, Inc.
  * All Rights Reserved.
@@ -19,6 +19,7 @@ typedef unsigned __bitwise xfs_km_flags_t;
 #define KM_NOFS		((__force xfs_km_flags_t)0x0004u)
 #define KM_MAYFAIL	((__force xfs_km_flags_t)0x0008u)
 #define KM_ZERO		((__force xfs_km_flags_t)0x0010u)
+#define KM_NOLOCKDEP	((__force xfs_km_flags_t)0x0020u)
 
 /*
  * We use a special process flag to avoid recursive callbacks into
@@ -30,7 +31,7 @@ kmem_flags_convert(xfs_km_flags_t flags)
 {
 	gfp_t	lflags;
 
-	BUG_ON(flags & ~(KM_NOFS|KM_MAYFAIL|KM_ZERO));
+	BUG_ON(flags & ~(KM_NOFS | KM_MAYFAIL | KM_ZERO | KM_NOLOCKDEP));
 
 	lflags = GFP_KERNEL | __GFP_NOWARN;
 	if (flags & KM_NOFS)
@@ -49,6 +50,9 @@ kmem_flags_convert(xfs_km_flags_t flags)
 	if (flags & KM_ZERO)
 		lflags |= __GFP_ZERO;
 
+	if (flags & KM_NOLOCKDEP)
+		lflags |= __GFP_NOLOCKDEP;
+
 	return lflags;
 }
 
@@ -81,14 +85,6 @@ kmem_zalloc_large(size_t size, xfs_km_flags_t flags)
 #define kmem_zone	kmem_cache
 #define kmem_zone_t	struct kmem_cache
 
-extern void *kmem_zone_alloc(kmem_zone_t *, xfs_km_flags_t);
-
-static inline void *
-kmem_zone_zalloc(kmem_zone_t *zone, xfs_km_flags_t flags)
-{
-	return kmem_zone_alloc(zone, flags | KM_ZERO);
-}
-
 static inline struct page *
 kmem_to_page(void *addr)
 {
diff --git a/fs/xfs/libxfs/xfs_ag.c b/fs/xfs/libxfs/xfs_ag.c
index 08d6beb54f8c..8cf73fe4338e 100644
--- a/fs/xfs/libxfs/xfs_ag.c
+++ b/fs/xfs/libxfs/xfs_ag.c
@@ -231,7 +231,7 @@ xfs_sbblock_init(
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	struct xfs_dsb		*dsb = XFS_BUF_TO_SBP(bp);
+	struct xfs_dsb		*dsb = bp->b_addr;
 
 	xfs_sb_to_disk(dsb, &mp->m_sb);
 	dsb->sb_inprogress = 1;
@@ -243,7 +243,7 @@ xfs_agfblock_init(
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(bp);
+	struct xfs_agf		*agf = bp->b_addr;
 	xfs_extlen_t		tmpsize;
 
 	agf->agf_magicnum = cpu_to_be32(XFS_AGF_MAGIC);
@@ -301,7 +301,7 @@ xfs_agflblock_init(
 		uuid_copy(&agfl->agfl_uuid, &mp->m_sb.sb_meta_uuid);
 	}
 
-	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, bp);
+	agfl_bno = xfs_buf_to_agfl_bno(bp);
 	for (bucket = 0; bucket < xfs_agfl_size(mp); bucket++)
 		agfl_bno[bucket] = cpu_to_be32(NULLAGBLOCK);
 }
@@ -312,7 +312,7 @@ xfs_agiblock_init(
 	struct xfs_buf		*bp,
 	struct aghdr_init_data	*id)
 {
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(bp);
+	struct xfs_agi		*agi = bp->b_addr;
 	int			bucket;
 
 	agi->agi_magicnum = cpu_to_be32(XFS_AGI_MAGIC);
@@ -502,7 +502,7 @@ xfs_ag_extend_space(
 	if (error)
 		return error;
 
-	agi = XFS_BUF_TO_AGI(bp);
+	agi = bp->b_addr;
 	be32_add_cpu(&agi->agi_length, len);
 	ASSERT(id->agno == mp->m_sb.sb_agcount - 1 ||
 	       be32_to_cpu(agi->agi_length) == mp->m_sb.sb_agblocks);
@@ -515,7 +515,7 @@ xfs_ag_extend_space(
 	if (error)
 		return error;
 
-	agf = XFS_BUF_TO_AGF(bp);
+	agf = bp->b_addr;
 	be32_add_cpu(&agf->agf_length, len);
 	ASSERT(agf->agf_length == agi->agi_length);
 	xfs_alloc_log_agf(tp, bp, XFS_AGF_LENGTH);
@@ -563,17 +563,18 @@ xfs_ag_get_geometry(
 	error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agf_bp);
 	if (error)
 		goto out_agi;
-	pag = xfs_perag_get(mp, agno);
+
+	pag = agi_bp->b_pag;
 
 	/* Fill out form. */
 	memset(ageo, 0, sizeof(*ageo));
 	ageo->ag_number = agno;
 
-	agi = XFS_BUF_TO_AGI(agi_bp);
+	agi = agi_bp->b_addr;
 	ageo->ag_icount = be32_to_cpu(agi->agi_count);
 	ageo->ag_ifree = be32_to_cpu(agi->agi_freecount);
 
-	agf = XFS_BUF_TO_AGF(agf_bp);
+	agf = agf_bp->b_addr;
 	ageo->ag_length = be32_to_cpu(agf->agf_length);
 	freeblks = pag->pagf_freeblks +
 		   pag->pagf_flcount +
@@ -583,7 +584,6 @@ xfs_ag_get_geometry(
 	xfs_ag_geom_health(pag, ageo);
 
 	/* Release resources. */
-	xfs_perag_put(pag);
 	xfs_buf_relse(agf_bp);
 out_agi:
 	xfs_buf_relse(agi_bp);
diff --git a/fs/xfs/libxfs/xfs_ag_resv.h b/fs/xfs/libxfs/xfs_ag_resv.h
index c0352edc8e41..8a8eb4bc48bb 100644
--- a/fs/xfs/libxfs/xfs_ag_resv.h
+++ b/fs/xfs/libxfs/xfs_ag_resv.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0+
+/* SPDX-License-Identifier: GPL-2.0+ */
 /*
  * Copyright (C) 2016 Oracle.  All Rights Reserved.
  * Author: Darrick J. Wong <darrick.wong@oracle.com>
@@ -37,16 +37,4 @@ xfs_ag_resv_rmapbt_alloc(
 	xfs_perag_put(pag);
 }
 
-static inline void
-xfs_ag_resv_rmapbt_free(
-	struct xfs_mount	*mp,
-	xfs_agnumber_t		agno)
-{
-	struct xfs_perag	*pag;
-
-	pag = xfs_perag_get(mp, agno);
-	xfs_ag_resv_free_extent(pag, XFS_AG_RESV_RMAPBT, NULL, 1);
-	xfs_perag_put(pag);
-}
-
 #endif	/* __XFS_AG_RESV_H__ */
diff --git a/fs/xfs/libxfs/xfs_alloc.c b/fs/xfs/libxfs/xfs_alloc.c
index d8053bc96c4d..852b536551b5 100644
--- a/fs/xfs/libxfs/xfs_alloc.c
+++ b/fs/xfs/libxfs/xfs_alloc.c
@@ -151,7 +151,7 @@ xfs_alloc_lookup_eq(
 	cur->bc_rec.a.ar_startblock = bno;
 	cur->bc_rec.a.ar_blockcount = len;
 	error = xfs_btree_lookup(cur, XFS_LOOKUP_EQ, stat);
-	cur->bc_private.a.priv.abt.active = (*stat == 1);
+	cur->bc_ag.abt.active = (*stat == 1);
 	return error;
 }
 
@@ -171,7 +171,7 @@ xfs_alloc_lookup_ge(
 	cur->bc_rec.a.ar_startblock = bno;
 	cur->bc_rec.a.ar_blockcount = len;
 	error = xfs_btree_lookup(cur, XFS_LOOKUP_GE, stat);
-	cur->bc_private.a.priv.abt.active = (*stat == 1);
+	cur->bc_ag.abt.active = (*stat == 1);
 	return error;
 }
 
@@ -190,7 +190,7 @@ xfs_alloc_lookup_le(
 	cur->bc_rec.a.ar_startblock = bno;
 	cur->bc_rec.a.ar_blockcount = len;
 	error = xfs_btree_lookup(cur, XFS_LOOKUP_LE, stat);
-	cur->bc_private.a.priv.abt.active = (*stat == 1);
+	cur->bc_ag.abt.active = (*stat == 1);
 	return error;
 }
 
@@ -198,7 +198,7 @@ static inline bool
 xfs_alloc_cur_active(
 	struct xfs_btree_cur	*cur)
 {
-	return cur && cur->bc_private.a.priv.abt.active;
+	return cur && cur->bc_ag.abt.active;
 }
 
 /*
@@ -230,7 +230,7 @@ xfs_alloc_get_rec(
 	int			*stat)	/* output: success/failure */
 {
 	struct xfs_mount	*mp = cur->bc_mp;
-	xfs_agnumber_t		agno = cur->bc_private.a.agno;
+	xfs_agnumber_t		agno = cur->bc_ag.agno;
 	union xfs_btree_rec	*rec;
 	int			error;
 
@@ -589,6 +589,7 @@ xfs_agfl_verify(
 {
 	struct xfs_mount *mp = bp->b_mount;
 	struct xfs_agfl	*agfl = XFS_BUF_TO_AGFL(bp);
+	__be32		*agfl_bno = xfs_buf_to_agfl_bno(bp);
 	int		i;
 
 	/*
@@ -614,8 +615,8 @@ xfs_agfl_verify(
 		return __this_address;
 
 	for (i = 0; i < xfs_agfl_size(mp); i++) {
-		if (be32_to_cpu(agfl->agfl_bno[i]) != NULLAGBLOCK &&
-		    be32_to_cpu(agfl->agfl_bno[i]) >= mp->m_sb.sb_agblocks)
+		if (be32_to_cpu(agfl_bno[i]) != NULLAGBLOCK &&
+		    be32_to_cpu(agfl_bno[i]) >= mp->m_sb.sb_agblocks)
 			return __this_address;
 	}
 
@@ -709,19 +710,18 @@ xfs_alloc_read_agfl(
 STATIC int
 xfs_alloc_update_counters(
 	struct xfs_trans	*tp,
-	struct xfs_perag	*pag,
 	struct xfs_buf		*agbp,
 	long			len)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_agf		*agf = agbp->b_addr;
 
-	pag->pagf_freeblks += len;
+	agbp->b_pag->pagf_freeblks += len;
 	be32_add_cpu(&agf->agf_freeblks, len);
 
 	xfs_trans_agblocks_delta(tp, len);
 	if (unlikely(be32_to_cpu(agf->agf_freeblks) >
 		     be32_to_cpu(agf->agf_length))) {
-		xfs_buf_corruption_error(agbp);
+		xfs_buf_mark_corrupt(agbp);
 		return -EFSCORRUPTED;
 	}
 
@@ -907,7 +907,7 @@ xfs_alloc_cur_check(
 		deactivate = true;
 out:
 	if (deactivate)
-		cur->bc_private.a.priv.abt.active = false;
+		cur->bc_ag.abt.active = false;
 	trace_xfs_alloc_cur_check(args->mp, cur->bc_btnum, bno, len, diff,
 				  *new);
 	return 0;
@@ -922,13 +922,13 @@ xfs_alloc_cur_finish(
 	struct xfs_alloc_arg	*args,
 	struct xfs_alloc_cur	*acur)
 {
+	struct xfs_agf __maybe_unused *agf = args->agbp->b_addr;
 	int			error;
 
 	ASSERT(acur->cnt && acur->bnolt);
 	ASSERT(acur->bno >= acur->rec_bno);
 	ASSERT(acur->bno + acur->len <= acur->rec_bno + acur->rec_len);
-	ASSERT(acur->rec_bno + acur->rec_len <=
-	       be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length));
+	ASSERT(acur->rec_bno + acur->rec_len <= be32_to_cpu(agf->agf_length));
 
 	error = xfs_alloc_fixup_trees(acur->cnt, acur->bnolt, acur->rec_bno,
 				      acur->rec_len, acur->bno, acur->len, 0);
@@ -1026,6 +1026,7 @@ xfs_alloc_ag_vextent_small(
 	xfs_extlen_t		*flenp,	/* result length */
 	int			*stat)	/* status: 0-freelist, 1-normal/none */
 {
+	struct xfs_agf		*agf = args->agbp->b_addr;
 	int			error = 0;
 	xfs_agblock_t		fbno = NULLAGBLOCK;
 	xfs_extlen_t		flen = 0;
@@ -1054,8 +1055,7 @@ xfs_alloc_ag_vextent_small(
 
 	if (args->minlen != 1 || args->alignment != 1 ||
 	    args->resv == XFS_AG_RESV_AGFL ||
-	    (be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_flcount) <=
-	     args->minleft))
+	    be32_to_cpu(agf->agf_flcount) <= args->minleft)
 		goto out;
 
 	error = xfs_alloc_get_freelist(args->tp, args->agbp, &fbno, 0);
@@ -1079,9 +1079,7 @@ xfs_alloc_ag_vextent_small(
 	}
 	*fbnop = args->agbno = fbno;
 	*flenp = args->len = 1;
-	if (XFS_IS_CORRUPT(args->mp,
-			   fbno >= be32_to_cpu(
-				   XFS_BUF_TO_AGF(args->agbp)->agf_length))) {
+	if (XFS_IS_CORRUPT(args->mp, fbno >= be32_to_cpu(agf->agf_length))) {
 		error = -EFSCORRUPTED;
 		goto error;
 	}
@@ -1176,8 +1174,7 @@ xfs_alloc_ag_vextent(
 	}
 
 	if (!args->wasfromfl) {
-		error = xfs_alloc_update_counters(args->tp, args->pag,
-						  args->agbp,
+		error = xfs_alloc_update_counters(args->tp, args->agbp,
 						  -((long)(args->len)));
 		if (error)
 			return error;
@@ -1203,6 +1200,7 @@ STATIC int			/* error */
 xfs_alloc_ag_vextent_exact(
 	xfs_alloc_arg_t	*args)	/* allocation argument structure */
 {
+	struct xfs_agf __maybe_unused *agf = args->agbp->b_addr;
 	xfs_btree_cur_t	*bno_cur;/* by block-number btree cursor */
 	xfs_btree_cur_t	*cnt_cur;/* by count btree cursor */
 	int		error;
@@ -1281,8 +1279,7 @@ xfs_alloc_ag_vextent_exact(
 	 */
 	cnt_cur = xfs_allocbt_init_cursor(args->mp, args->tp, args->agbp,
 		args->agno, XFS_BTNUM_CNT);
-	ASSERT(args->agbno + args->len <=
-		be32_to_cpu(XFS_BUF_TO_AGF(args->agbp)->agf_length));
+	ASSERT(args->agbno + args->len <= be32_to_cpu(agf->agf_length));
 	error = xfs_alloc_fixup_trees(cnt_cur, bno_cur, fbno, flen, args->agbno,
 				      args->len, XFSA_FIXUP_BNO_OK);
 	if (error) {
@@ -1353,7 +1350,7 @@ xfs_alloc_walk_iter(
 		if (error)
 			return error;
 		if (i == 0)
-			cur->bc_private.a.priv.abt.active = false;
+			cur->bc_ag.abt.active = false;
 
 		if (count > 0)
 			count--;
@@ -1468,7 +1465,7 @@ xfs_alloc_ag_vextent_locality(
 		if (error)
 			return error;
 		if (i) {
-			acur->cnt->bc_private.a.priv.abt.active = true;
+			acur->cnt->bc_ag.abt.active = true;
 			fbcur = acur->cnt;
 			fbinc = false;
 		}
@@ -1515,7 +1512,7 @@ xfs_alloc_ag_vextent_lastblock(
 	 * maxlen, go to the start of this block, and skip all those smaller
 	 * than minlen.
 	 */
-	if (len || args->alignment > 1) {
+	if (*len || args->alignment > 1) {
 		acur->cnt->bc_ptrs[0] = 1;
 		do {
 			error = xfs_alloc_get_rec(acur->cnt, bno, len, &i);
@@ -1661,6 +1658,7 @@ STATIC int				/* error */
 xfs_alloc_ag_vextent_size(
 	xfs_alloc_arg_t	*args)		/* allocation argument structure */
 {
+	struct xfs_agf	*agf = args->agbp->b_addr;
 	xfs_btree_cur_t	*bno_cur;	/* cursor for bno btree */
 	xfs_btree_cur_t	*cnt_cur;	/* cursor for cnt btree */
 	int		error;		/* error result */
@@ -1851,8 +1849,7 @@ restart:
 	args->agbno = rbno;
 	if (XFS_IS_CORRUPT(args->mp,
 			   args->agbno + args->len >
-			   be32_to_cpu(
-				   XFS_BUF_TO_AGF(args->agbp)->agf_length))) {
+			   be32_to_cpu(agf->agf_length))) {
 		error = -EFSCORRUPTED;
 		goto error0;
 	}
@@ -1888,7 +1885,6 @@ xfs_free_ag_extent(
 	enum xfs_ag_resv_type		type)
 {
 	struct xfs_mount		*mp;
-	struct xfs_perag		*pag;
 	struct xfs_btree_cur		*bno_cur;
 	struct xfs_btree_cur		*cnt_cur;
 	xfs_agblock_t			gtbno; /* start of right neighbor */
@@ -2168,10 +2164,8 @@ xfs_free_ag_extent(
 	/*
 	 * Update the freespace totals in the ag and superblock.
 	 */
-	pag = xfs_perag_get(mp, agno);
-	error = xfs_alloc_update_counters(tp, pag, agbp, len);
-	xfs_ag_resv_free_extent(pag, type, tp, len);
-	xfs_perag_put(pag);
+	error = xfs_alloc_update_counters(tp, agbp, len);
+	xfs_ag_resv_free_extent(agbp->b_pag, type, tp, len);
 	if (error)
 		goto error0;
 
@@ -2424,7 +2418,7 @@ xfs_agfl_reset(
 	struct xfs_perag	*pag)
 {
 	struct xfs_mount	*mp = tp->t_mountp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_agf		*agf = agbp->b_addr;
 
 	ASSERT(pag->pagf_agflreset);
 	trace_xfs_agfl_reset(mp, agf, 0, _RET_IP_);
@@ -2468,7 +2462,8 @@ xfs_defer_agfl_block(
 	ASSERT(xfs_bmap_free_item_zone != NULL);
 	ASSERT(oinfo != NULL);
 
-	new = kmem_zone_alloc(xfs_bmap_free_item_zone, 0);
+	new = kmem_cache_alloc(xfs_bmap_free_item_zone,
+			       GFP_KERNEL | __GFP_NOFAIL);
 	new->xefi_startblock = XFS_AGB_TO_FSB(mp, agno, agbno);
 	new->xefi_blockcount = 1;
 	new->xefi_oinfo = *oinfo;
@@ -2655,7 +2650,7 @@ xfs_alloc_get_freelist(
 	xfs_agblock_t	*bnop,	/* block address retrieved from freelist */
 	int		btreeblk) /* destination is a AGF btree */
 {
-	xfs_agf_t	*agf;	/* a.g. freespace structure */
+	struct xfs_agf	*agf = agbp->b_addr;
 	xfs_buf_t	*agflbp;/* buffer for a.g. freelist structure */
 	xfs_agblock_t	bno;	/* block number returned */
 	__be32		*agfl_bno;
@@ -2667,7 +2662,6 @@ xfs_alloc_get_freelist(
 	/*
 	 * Freelist is empty, give up.
 	 */
-	agf = XFS_BUF_TO_AGF(agbp);
 	if (!agf->agf_flcount) {
 		*bnop = NULLAGBLOCK;
 		return 0;
@@ -2684,14 +2678,14 @@ xfs_alloc_get_freelist(
 	/*
 	 * Get the block number and update the data structures.
 	 */
-	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp);
+	agfl_bno = xfs_buf_to_agfl_bno(agflbp);
 	bno = be32_to_cpu(agfl_bno[be32_to_cpu(agf->agf_flfirst)]);
 	be32_add_cpu(&agf->agf_flfirst, 1);
 	xfs_trans_brelse(tp, agflbp);
 	if (be32_to_cpu(agf->agf_flfirst) == xfs_agfl_size(mp))
 		agf->agf_flfirst = 0;
 
-	pag = xfs_perag_get(mp, be32_to_cpu(agf->agf_seqno));
+	pag = agbp->b_pag;
 	ASSERT(!pag->pagf_agflreset);
 	be32_add_cpu(&agf->agf_flcount, -1);
 	xfs_trans_agflist_delta(tp, -1);
@@ -2703,7 +2697,6 @@ xfs_alloc_get_freelist(
 		pag->pagf_btreeblks++;
 		logflags |= XFS_AGF_BTREEBLKS;
 	}
-	xfs_perag_put(pag);
 
 	xfs_alloc_log_agf(tp, agbp, logflags);
 	*bnop = bno;
@@ -2745,7 +2738,7 @@ xfs_alloc_log_agf(
 		sizeof(xfs_agf_t)
 	};
 
-	trace_xfs_agf(tp->t_mountp, XFS_BUF_TO_AGF(bp), fields, _RET_IP_);
+	trace_xfs_agf(tp->t_mountp, bp->b_addr, fields, _RET_IP_);
 
 	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_AGF_BUF);
 
@@ -2783,18 +2776,15 @@ xfs_alloc_put_freelist(
 	xfs_agblock_t		bno,	/* block being freed */
 	int			btreeblk) /* block came from a AGF btree */
 {
-	xfs_agf_t		*agf;	/* a.g. freespace structure */
+	struct xfs_mount	*mp = tp->t_mountp;
+	struct xfs_agf		*agf = agbp->b_addr;
 	__be32			*blockp;/* pointer to array entry */
 	int			error;
 	int			logflags;
-	xfs_mount_t		*mp;	/* mount structure */
 	xfs_perag_t		*pag;	/* per allocation group data */
 	__be32			*agfl_bno;
 	int			startoff;
 
-	agf = XFS_BUF_TO_AGF(agbp);
-	mp = tp->t_mountp;
-
 	if (!agflbp && (error = xfs_alloc_read_agfl(mp, tp,
 			be32_to_cpu(agf->agf_seqno), &agflbp)))
 		return error;
@@ -2802,7 +2792,7 @@ xfs_alloc_put_freelist(
 	if (be32_to_cpu(agf->agf_fllast) == xfs_agfl_size(mp))
 		agf->agf_fllast = 0;
 
-	pag = xfs_perag_get(mp, be32_to_cpu(agf->agf_seqno));
+	pag = agbp->b_pag;
 	ASSERT(!pag->pagf_agflreset);
 	be32_add_cpu(&agf->agf_flcount, 1);
 	xfs_trans_agflist_delta(tp, 1);
@@ -2814,13 +2804,12 @@ xfs_alloc_put_freelist(
 		pag->pagf_btreeblks--;
 		logflags |= XFS_AGF_BTREEBLKS;
 	}
-	xfs_perag_put(pag);
 
 	xfs_alloc_log_agf(tp, agbp, logflags);
 
 	ASSERT(be32_to_cpu(agf->agf_flcount) <= xfs_agfl_size(mp));
 
-	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp);
+	agfl_bno = xfs_buf_to_agfl_bno(agflbp);
 	blockp = &agfl_bno[be32_to_cpu(agf->agf_fllast)];
 	*blockp = cpu_to_be32(bno);
 	startoff = (char *)blockp - (char *)agflbp->b_addr;
@@ -2838,13 +2827,12 @@ xfs_agf_verify(
 	struct xfs_buf		*bp)
 {
 	struct xfs_mount	*mp = bp->b_mount;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(bp);
+	struct xfs_agf		*agf = bp->b_addr;
 
 	if (xfs_sb_version_hascrc(&mp->m_sb)) {
 		if (!uuid_equal(&agf->agf_uuid, &mp->m_sb.sb_meta_uuid))
 			return __this_address;
-		if (!xfs_log_check_lsn(mp,
-				be64_to_cpu(XFS_BUF_TO_AGF(bp)->agf_lsn)))
+		if (!xfs_log_check_lsn(mp, be64_to_cpu(agf->agf_lsn)))
 			return __this_address;
 	}
 
@@ -2858,6 +2846,13 @@ xfs_agf_verify(
 	      be32_to_cpu(agf->agf_flcount) <= xfs_agfl_size(mp)))
 		return __this_address;
 
+	if (be32_to_cpu(agf->agf_length) > mp->m_sb.sb_dblocks)
+		return __this_address;
+
+	if (be32_to_cpu(agf->agf_freeblks) < be32_to_cpu(agf->agf_longest) ||
+	    be32_to_cpu(agf->agf_freeblks) > be32_to_cpu(agf->agf_length))
+		return __this_address;
+
 	if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) < 1 ||
 	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]) < 1 ||
 	    be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) > XFS_BTREE_MAXLEVELS ||
@@ -2869,6 +2864,10 @@ xfs_agf_verify(
 	     be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]) > XFS_BTREE_MAXLEVELS))
 		return __this_address;
 
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb) &&
+	    be32_to_cpu(agf->agf_rmap_blocks) > be32_to_cpu(agf->agf_length))
+		return __this_address;
+
 	/*
 	 * during growfs operations, the perag is not fully initialised,
 	 * so we can't use it for any useful checking. growfs ensures we can't
@@ -2883,6 +2882,11 @@ xfs_agf_verify(
 		return __this_address;
 
 	if (xfs_sb_version_hasreflink(&mp->m_sb) &&
+	    be32_to_cpu(agf->agf_refcount_blocks) >
+	    be32_to_cpu(agf->agf_length))
+		return __this_address;
+
+	if (xfs_sb_version_hasreflink(&mp->m_sb) &&
 	    (be32_to_cpu(agf->agf_refcount_level) < 1 ||
 	     be32_to_cpu(agf->agf_refcount_level) > XFS_BTREE_MAXLEVELS))
 		return __this_address;
@@ -2914,6 +2918,7 @@ xfs_agf_write_verify(
 {
 	struct xfs_mount	*mp = bp->b_mount;
 	struct xfs_buf_log_item	*bip = bp->b_log_item;
+	struct xfs_agf		*agf = bp->b_addr;
 	xfs_failaddr_t		fa;
 
 	fa = xfs_agf_verify(bp);
@@ -2926,7 +2931,7 @@ xfs_agf_write_verify(
 		return;
 
 	if (bip)
-		XFS_BUF_TO_AGF(bp)->agf_lsn = cpu_to_be64(bip->bli_item.li_lsn);
+		agf->agf_lsn = cpu_to_be64(bip->bli_item.li_lsn);
 
 	xfs_buf_update_cksum(bp, XFS_AGF_CRC_OFF);
 }
@@ -2994,8 +2999,8 @@ xfs_alloc_read_agf(
 		return error;
 	ASSERT(!(*bpp)->b_error);
 
-	agf = XFS_BUF_TO_AGF(*bpp);
-	pag = xfs_perag_get(mp, agno);
+	agf = (*bpp)->b_addr;
+	pag = (*bpp)->b_pag;
 	if (!pag->pagf_init) {
 		pag->pagf_freeblks = be32_to_cpu(agf->agf_freeblks);
 		pag->pagf_btreeblks = be32_to_cpu(agf->agf_btreeblks);
@@ -3023,7 +3028,6 @@ xfs_alloc_read_agf(
 		       be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNTi]));
 	}
 #endif
-	xfs_perag_put(pag);
 	return 0;
 }
 
@@ -3275,6 +3279,7 @@ __xfs_free_extent(
 	struct xfs_buf			*agbp;
 	xfs_agnumber_t			agno = XFS_FSB_TO_AGNO(mp, bno);
 	xfs_agblock_t			agbno = XFS_FSB_TO_AGBNO(mp, bno);
+	struct xfs_agf			*agf;
 	int				error;
 	unsigned int			busy_flags = 0;
 
@@ -3288,6 +3293,7 @@ __xfs_free_extent(
 	error = xfs_free_extent_fix_freelist(tp, agno, &agbp);
 	if (error)
 		return error;
+	agf = agbp->b_addr;
 
 	if (XFS_IS_CORRUPT(mp, agbno >= mp->m_sb.sb_agblocks)) {
 		error = -EFSCORRUPTED;
@@ -3295,9 +3301,7 @@ __xfs_free_extent(
 	}
 
 	/* validate the extent size is legal now we have the agf locked */
-	if (XFS_IS_CORRUPT(mp,
-			   agbno + len >
-			   be32_to_cpu(XFS_BUF_TO_AGF(agbp)->agf_length))) {
+	if (XFS_IS_CORRUPT(mp, agbno + len > be32_to_cpu(agf->agf_length))) {
 		error = -EFSCORRUPTED;
 		goto err;
 	}
@@ -3408,7 +3412,7 @@ xfs_agfl_walk(
 	unsigned int		i;
 	int			error;
 
-	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp);
+	agfl_bno = xfs_buf_to_agfl_bno(agflbp);
 	i = be32_to_cpu(agf->agf_flfirst);
 
 	/* Nothing to walk in an empty AGFL. */
diff --git a/fs/xfs/libxfs/xfs_alloc.h b/fs/xfs/libxfs/xfs_alloc.h
index 7380fbe4a3ff..6c22b12176b8 100644
--- a/fs/xfs/libxfs/xfs_alloc.h
+++ b/fs/xfs/libxfs/xfs_alloc.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
  * All Rights Reserved.
@@ -236,4 +236,13 @@ typedef int (*xfs_agfl_walk_fn)(struct xfs_mount *mp, xfs_agblock_t bno,
 int xfs_agfl_walk(struct xfs_mount *mp, struct xfs_agf *agf,
 		struct xfs_buf *agflbp, xfs_agfl_walk_fn walk_fn, void *priv);
 
+static inline __be32 *
+xfs_buf_to_agfl_bno(
+	struct xfs_buf		*bp)
+{
+	if (xfs_sb_version_hascrc(&bp->b_mount->m_sb))
+		return bp->b_addr + sizeof(struct xfs_agfl);
+	return bp->b_addr;
+}
+
 #endif	/* __XFS_ALLOC_H__ */
diff --git a/fs/xfs/libxfs/xfs_alloc_btree.c b/fs/xfs/libxfs/xfs_alloc_btree.c
index 279694d73e4e..8e01231b308e 100644
--- a/fs/xfs/libxfs/xfs_alloc_btree.c
+++ b/fs/xfs/libxfs/xfs_alloc_btree.c
@@ -12,6 +12,7 @@
 #include "xfs_sb.h"
 #include "xfs_mount.h"
 #include "xfs_btree.h"
+#include "xfs_btree_staging.h"
 #include "xfs_alloc_btree.h"
 #include "xfs_alloc.h"
 #include "xfs_extent_busy.h"
@@ -25,7 +26,7 @@ xfs_allocbt_dup_cursor(
 	struct xfs_btree_cur	*cur)
 {
 	return xfs_allocbt_init_cursor(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agbp, cur->bc_private.a.agno,
+			cur->bc_ag.agbp, cur->bc_ag.agno,
 			cur->bc_btnum);
 }
 
@@ -35,18 +36,16 @@ xfs_allocbt_set_root(
 	union xfs_btree_ptr	*ptr,
 	int			inc)
 {
-	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
-	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agf		*agf = agbp->b_addr;
 	int			btnum = cur->bc_btnum;
-	struct xfs_perag	*pag = xfs_perag_get(cur->bc_mp, seqno);
+	struct xfs_perag	*pag = agbp->b_pag;
 
 	ASSERT(ptr->s != 0);
 
 	agf->agf_roots[btnum] = ptr->s;
 	be32_add_cpu(&agf->agf_levels[btnum], inc);
 	pag->pagf_levels[btnum] += inc;
-	xfs_perag_put(pag);
 
 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
 }
@@ -62,7 +61,7 @@ xfs_allocbt_alloc_block(
 	xfs_agblock_t		bno;
 
 	/* Allocate the new block from the freelist. If we can't, give up.  */
-	error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_private.a.agbp,
+	error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_ag.agbp,
 				       &bno, 1);
 	if (error)
 		return error;
@@ -72,7 +71,7 @@ xfs_allocbt_alloc_block(
 		return 0;
 	}
 
-	xfs_extent_busy_reuse(cur->bc_mp, cur->bc_private.a.agno, bno, 1, false);
+	xfs_extent_busy_reuse(cur->bc_mp, cur->bc_ag.agno, bno, 1, false);
 
 	xfs_trans_agbtree_delta(cur->bc_tp, 1);
 	new->s = cpu_to_be32(bno);
@@ -86,8 +85,8 @@ xfs_allocbt_free_block(
 	struct xfs_btree_cur	*cur,
 	struct xfs_buf		*bp)
 {
-	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agf		*agf = agbp->b_addr;
 	xfs_agblock_t		bno;
 	int			error;
 
@@ -113,8 +112,7 @@ xfs_allocbt_update_lastrec(
 	int			ptr,
 	int			reason)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
-	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
+	struct xfs_agf		*agf = cur->bc_ag.agbp->b_addr;
 	struct xfs_perag	*pag;
 	__be32			len;
 	int			numrecs;
@@ -159,10 +157,9 @@ xfs_allocbt_update_lastrec(
 	}
 
 	agf->agf_longest = len;
-	pag = xfs_perag_get(cur->bc_mp, seqno);
+	pag = cur->bc_ag.agbp->b_pag;
 	pag->pagf_longest = be32_to_cpu(len);
-	xfs_perag_put(pag);
-	xfs_alloc_log_agf(cur->bc_tp, cur->bc_private.a.agbp, XFS_AGF_LONGEST);
+	xfs_alloc_log_agf(cur->bc_tp, cur->bc_ag.agbp, XFS_AGF_LONGEST);
 }
 
 STATIC int
@@ -226,9 +223,9 @@ xfs_allocbt_init_ptr_from_cur(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*ptr)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
+	struct xfs_agf		*agf = cur->bc_ag.agbp->b_addr;
 
-	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno));
+	ASSERT(cur->bc_ag.agno == be32_to_cpu(agf->agf_seqno));
 
 	ptr->s = agf->agf_roots[cur->bc_btnum];
 }
@@ -471,23 +468,19 @@ static const struct xfs_btree_ops xfs_cntbt_ops = {
 	.recs_inorder		= xfs_cntbt_recs_inorder,
 };
 
-/*
- * Allocate a new allocation btree cursor.
- */
-struct xfs_btree_cur *			/* new alloc btree cursor */
-xfs_allocbt_init_cursor(
-	struct xfs_mount	*mp,		/* file system mount point */
-	struct xfs_trans	*tp,		/* transaction pointer */
-	struct xfs_buf		*agbp,		/* buffer for agf structure */
-	xfs_agnumber_t		agno,		/* allocation group number */
-	xfs_btnum_t		btnum)		/* btree identifier */
+/* Allocate most of a new allocation btree cursor. */
+STATIC struct xfs_btree_cur *
+xfs_allocbt_init_common(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	xfs_agnumber_t		agno,
+	xfs_btnum_t		btnum)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
 	struct xfs_btree_cur	*cur;
 
 	ASSERT(btnum == XFS_BTNUM_BNO || btnum == XFS_BTNUM_CNT);
 
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
+	cur = kmem_cache_zalloc(xfs_btree_cur_zone, GFP_NOFS | __GFP_NOFAIL);
 
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
@@ -495,19 +488,16 @@ xfs_allocbt_init_cursor(
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
 
 	if (btnum == XFS_BTNUM_CNT) {
-		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtc_2);
 		cur->bc_ops = &xfs_cntbt_ops;
-		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtc_2);
 		cur->bc_flags = XFS_BTREE_LASTREC_UPDATE;
 	} else {
-		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtb_2);
 		cur->bc_ops = &xfs_bnobt_ops;
-		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]);
+		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_abtb_2);
 	}
 
-	cur->bc_private.a.agbp = agbp;
-	cur->bc_private.a.agno = agno;
-	cur->bc_private.a.priv.abt.active = false;
+	cur->bc_ag.agno = agno;
+	cur->bc_ag.abt.active = false;
 
 	if (xfs_sb_version_hascrc(&mp->m_sb))
 		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
@@ -516,6 +506,73 @@ xfs_allocbt_init_cursor(
 }
 
 /*
+ * Allocate a new allocation btree cursor.
+ */
+struct xfs_btree_cur *			/* new alloc btree cursor */
+xfs_allocbt_init_cursor(
+	struct xfs_mount	*mp,		/* file system mount point */
+	struct xfs_trans	*tp,		/* transaction pointer */
+	struct xfs_buf		*agbp,		/* buffer for agf structure */
+	xfs_agnumber_t		agno,		/* allocation group number */
+	xfs_btnum_t		btnum)		/* btree identifier */
+{
+	struct xfs_agf		*agf = agbp->b_addr;
+	struct xfs_btree_cur	*cur;
+
+	cur = xfs_allocbt_init_common(mp, tp, agno, btnum);
+	if (btnum == XFS_BTNUM_CNT)
+		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);
+	else
+		cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]);
+
+	cur->bc_ag.agbp = agbp;
+
+	return cur;
+}
+
+/* Create a free space btree cursor with a fake root for staging. */
+struct xfs_btree_cur *
+xfs_allocbt_stage_cursor(
+	struct xfs_mount	*mp,
+	struct xbtree_afakeroot	*afake,
+	xfs_agnumber_t		agno,
+	xfs_btnum_t		btnum)
+{
+	struct xfs_btree_cur	*cur;
+
+	cur = xfs_allocbt_init_common(mp, NULL, agno, btnum);
+	xfs_btree_stage_afakeroot(cur, afake);
+	return cur;
+}
+
+/*
+ * Install a new free space btree root.  Caller is responsible for invalidating
+ * and freeing the old btree blocks.
+ */
+void
+xfs_allocbt_commit_staged_btree(
+	struct xfs_btree_cur	*cur,
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp)
+{
+	struct xfs_agf		*agf = agbp->b_addr;
+	struct xbtree_afakeroot	*afake = cur->bc_ag.afake;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+
+	agf->agf_roots[cur->bc_btnum] = cpu_to_be32(afake->af_root);
+	agf->agf_levels[cur->bc_btnum] = cpu_to_be32(afake->af_levels);
+	xfs_alloc_log_agf(tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
+
+	if (cur->bc_btnum == XFS_BTNUM_BNO) {
+		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_bnobt_ops);
+	} else {
+		cur->bc_flags |= XFS_BTREE_LASTREC_UPDATE;
+		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_cntbt_ops);
+	}
+}
+
+/*
  * Calculate number of records in an alloc btree block.
  */
 int
diff --git a/fs/xfs/libxfs/xfs_alloc_btree.h b/fs/xfs/libxfs/xfs_alloc_btree.h
index c9305ebb69f6..a5b998e950fe 100644
--- a/fs/xfs/libxfs/xfs_alloc_btree.h
+++ b/fs/xfs/libxfs/xfs_alloc_btree.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000,2005 Silicon Graphics, Inc.
  * All Rights Reserved.
@@ -13,6 +13,7 @@
 struct xfs_buf;
 struct xfs_btree_cur;
 struct xfs_mount;
+struct xbtree_afakeroot;
 
 /*
  * Btree block header size depends on a superblock flag.
@@ -48,8 +49,14 @@ struct xfs_mount;
 extern struct xfs_btree_cur *xfs_allocbt_init_cursor(struct xfs_mount *,
 		struct xfs_trans *, struct xfs_buf *,
 		xfs_agnumber_t, xfs_btnum_t);
+struct xfs_btree_cur *xfs_allocbt_stage_cursor(struct xfs_mount *mp,
+		struct xbtree_afakeroot *afake, xfs_agnumber_t agno,
+		xfs_btnum_t btnum);
 extern int xfs_allocbt_maxrecs(struct xfs_mount *, int, int);
 extern xfs_extlen_t xfs_allocbt_calc_size(struct xfs_mount *mp,
 		unsigned long long len);
 
+void xfs_allocbt_commit_staged_btree(struct xfs_btree_cur *cur,
+		struct xfs_trans *tp, struct xfs_buf *agbp);
+
 #endif	/* __XFS_ALLOC_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_attr.c b/fs/xfs/libxfs/xfs_attr.c
index e6149720ce02..2e055c079f39 100644
--- a/fs/xfs/libxfs/xfs_attr.c
+++ b/fs/xfs/libxfs/xfs_attr.c
@@ -46,6 +46,7 @@ STATIC int xfs_attr_shortform_addname(xfs_da_args_t *args);
 STATIC int xfs_attr_leaf_get(xfs_da_args_t *args);
 STATIC int xfs_attr_leaf_addname(xfs_da_args_t *args);
 STATIC int xfs_attr_leaf_removename(xfs_da_args_t *args);
+STATIC int xfs_attr_leaf_hasname(struct xfs_da_args *args, struct xfs_buf **bp);
 
 /*
  * Internal routines when attribute list is more than one block.
@@ -53,43 +54,18 @@ STATIC int xfs_attr_leaf_removename(xfs_da_args_t *args);
 STATIC int xfs_attr_node_get(xfs_da_args_t *args);
 STATIC int xfs_attr_node_addname(xfs_da_args_t *args);
 STATIC int xfs_attr_node_removename(xfs_da_args_t *args);
+STATIC int xfs_attr_node_hasname(xfs_da_args_t *args,
+				 struct xfs_da_state **state);
 STATIC int xfs_attr_fillstate(xfs_da_state_t *state);
 STATIC int xfs_attr_refillstate(xfs_da_state_t *state);
 
-
-STATIC int
-xfs_attr_args_init(
-	struct xfs_da_args	*args,
-	struct xfs_inode	*dp,
-	const unsigned char	*name,
-	size_t			namelen,
-	int			flags)
-{
-
-	if (!name)
-		return -EINVAL;
-
-	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 = namelen;
-	if (args->namelen >= MAXNAMELEN)
-		return -EFAULT;		/* match IRIX behaviour */
-
-	args->hashval = xfs_da_hashname(args->name, args->namelen);
-	return 0;
-}
-
 int
 xfs_inode_hasattr(
 	struct xfs_inode	*ip)
 {
 	if (!XFS_IFORK_Q(ip) ||
-	    (ip->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS &&
-	     ip->i_d.di_anextents == 0))
+	    (ip->i_afp->if_format == XFS_DINODE_FMT_EXTENTS &&
+	     ip->i_afp->if_nextents == 0))
 		return 0;
 	return 1;
 }
@@ -104,85 +80,60 @@ xfs_inode_hasattr(
  */
 int
 xfs_attr_get_ilocked(
-	struct xfs_inode	*ip,
 	struct xfs_da_args	*args)
 {
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
+	ASSERT(xfs_isilocked(args->dp, XFS_ILOCK_SHARED | XFS_ILOCK_EXCL));
 
-	if (!xfs_inode_hasattr(ip))
+	if (!xfs_inode_hasattr(args->dp))
 		return -ENOATTR;
-	else if (ip->i_d.di_aformat == XFS_DINODE_FMT_LOCAL)
+
+	if (args->dp->i_afp->if_format == XFS_DINODE_FMT_LOCAL)
 		return xfs_attr_shortform_getvalue(args);
-	else if (xfs_bmap_one_block(ip, XFS_ATTR_FORK))
+	if (xfs_bmap_one_block(args->dp, XFS_ATTR_FORK))
 		return xfs_attr_leaf_get(args);
-	else
-		return xfs_attr_node_get(args);
+	return xfs_attr_node_get(args);
 }
 
 /*
  * Retrieve an extended attribute by name, and its value if requested.
  *
- * If ATTR_KERNOVAL is set in @flags, then the caller does not want the value,
- * just an indication whether the attribute exists and the size of the value if
- * it exists. The size is returned in @valuelenp,
+ * If args->valuelen is zero, then the caller does not want the value, just an
+ * indication whether the attribute exists and the size of the value if it
+ * exists. The size is returned in args.valuelen.
  *
- * If the attribute is found, but exceeds the size limit set by the caller in
- * @valuelenp, return -ERANGE with the size of the attribute that was found in
- * @valuelenp.
+ * If args->value is NULL but args->valuelen is non-zero, allocate the buffer
+ * for the value after existence of the attribute has been determined. The
+ * caller always has to free args->value if it is set, no matter if this
+ * function was successful or not.
  *
- * If ATTR_ALLOC is set in @flags, allocate the buffer for the value after
- * existence of the attribute has been determined. On success, return that
- * buffer to the caller and leave them to free it. On failure, free any
- * allocated buffer and ensure the buffer pointer returned to the caller is
- * null.
+ * If the attribute is found, but exceeds the size limit set by the caller in
+ * args->valuelen, return -ERANGE with the size of the attribute that was found
+ * in args->valuelen.
  */
 int
 xfs_attr_get(
-	struct xfs_inode	*ip,
-	const unsigned char	*name,
-	size_t			namelen,
-	unsigned char		**value,
-	int			*valuelenp,
-	int			flags)
+	struct xfs_da_args	*args)
 {
-	struct xfs_da_args	args;
 	uint			lock_mode;
 	int			error;
 
-	ASSERT((flags & (ATTR_ALLOC | ATTR_KERNOVAL)) || *value);
-
-	XFS_STATS_INC(ip->i_mount, xs_attr_get);
+	XFS_STATS_INC(args->dp->i_mount, xs_attr_get);
 
-	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
+	if (XFS_FORCED_SHUTDOWN(args->dp->i_mount))
 		return -EIO;
 
-	error = xfs_attr_args_init(&args, ip, name, namelen, flags);
-	if (error)
-		return error;
+	args->geo = args->dp->i_mount->m_attr_geo;
+	args->whichfork = XFS_ATTR_FORK;
+	args->hashval = xfs_da_hashname(args->name, args->namelen);
 
 	/* Entirely possible to look up a name which doesn't exist */
-	args.op_flags = XFS_DA_OP_OKNOENT;
-	if (flags & ATTR_ALLOC)
-		args.op_flags |= XFS_DA_OP_ALLOCVAL;
-	else
-		args.value = *value;
-	args.valuelen = *valuelenp;
+	args->op_flags = XFS_DA_OP_OKNOENT;
 
-	lock_mode = xfs_ilock_attr_map_shared(ip);
-	error = xfs_attr_get_ilocked(ip, &args);
-	xfs_iunlock(ip, lock_mode);
-	*valuelenp = args.valuelen;
+	lock_mode = xfs_ilock_attr_map_shared(args->dp);
+	error = xfs_attr_get_ilocked(args);
+	xfs_iunlock(args->dp, lock_mode);
 
-	/* on error, we have to clean up allocated value buffers */
-	if (error) {
-		if (flags & ATTR_ALLOC) {
-			kmem_free(args.value);
-			*value = NULL;
-		}
-		return error;
-	}
-	*value = args.value;
-	return 0;
+	return error;
 }
 
 /*
@@ -227,8 +178,13 @@ xfs_attr_try_sf_addname(
 	struct xfs_da_args	*args)
 {
 
-	struct xfs_mount	*mp = dp->i_mount;
-	int			error, error2;
+	int			error;
+
+	/*
+	 * Build initial attribute list (if required).
+	 */
+	if (dp->i_afp->if_format == XFS_DINODE_FMT_EXTENTS)
+		xfs_attr_shortform_create(args);
 
 	error = xfs_attr_shortform_addname(args);
 	if (error == -ENOSPC)
@@ -238,15 +194,73 @@ xfs_attr_try_sf_addname(
 	 * Commit the shortform mods, and we're done.
 	 * NOTE: this is also the error path (EEXIST, etc).
 	 */
-	if (!error && (args->flags & ATTR_KERNOTIME) == 0)
+	if (!error && !(args->op_flags & XFS_DA_OP_NOTIME))
 		xfs_trans_ichgtime(args->trans, dp, XFS_ICHGTIME_CHG);
 
-	if (mp->m_flags & XFS_MOUNT_WSYNC)
+	if (dp->i_mount->m_flags & XFS_MOUNT_WSYNC)
 		xfs_trans_set_sync(args->trans);
 
-	error2 = xfs_trans_commit(args->trans);
-	args->trans = NULL;
-	return error ? error : error2;
+	return error;
+}
+
+/*
+ * Check to see if the attr should be upgraded from non-existent or shortform to
+ * single-leaf-block attribute list.
+ */
+static inline bool
+xfs_attr_is_shortform(
+	struct xfs_inode    *ip)
+{
+	return ip->i_afp->if_format == XFS_DINODE_FMT_LOCAL ||
+	       (ip->i_afp->if_format == XFS_DINODE_FMT_EXTENTS &&
+		ip->i_afp->if_nextents == 0);
+}
+
+/*
+ * Attempts to set an attr in shortform, or converts short form to leaf form if
+ * there is not enough room.  If the attr is set, the transaction is committed
+ * and set to NULL.
+ */
+STATIC int
+xfs_attr_set_shortform(
+	struct xfs_da_args	*args,
+	struct xfs_buf		**leaf_bp)
+{
+	struct xfs_inode	*dp = args->dp;
+	int			error, error2 = 0;
+
+	/*
+	 * Try to add the attr to the attribute list in the inode.
+	 */
+	error = xfs_attr_try_sf_addname(dp, args);
+	if (error != -ENOSPC) {
+		error2 = xfs_trans_commit(args->trans);
+		args->trans = NULL;
+		return error ? error : error2;
+	}
+	/*
+	 * It won't fit in the shortform, transform to a leaf block.  GROT:
+	 * another possible req'mt for a double-split btree op.
+	 */
+	error = xfs_attr_shortform_to_leaf(args, leaf_bp);
+	if (error)
+		return error;
+
+	/*
+	 * Prevent the leaf buffer from being unlocked so that a concurrent AIL
+	 * push cannot grab the half-baked leaf buffer and run into problems
+	 * with the write verifier. Once we're done rolling the transaction we
+	 * can release the hold and add the attr to the leaf.
+	 */
+	xfs_trans_bhold(args->trans, *leaf_bp);
+	error = xfs_defer_finish(&args->trans);
+	xfs_trans_bhold_release(args->trans, *leaf_bp);
+	if (error) {
+		xfs_trans_brelse(args->trans, *leaf_bp);
+		return error;
+	}
+
+	return 0;
 }
 
 /*
@@ -258,61 +272,94 @@ xfs_attr_set_args(
 {
 	struct xfs_inode	*dp = args->dp;
 	struct xfs_buf          *leaf_bp = NULL;
-	int			error;
+	int			error = 0;
 
 	/*
-	 * If the attribute list is non-existent or a shortform list,
-	 * upgrade it to a single-leaf-block attribute list.
+	 * If the attribute list is already in leaf format, jump straight to
+	 * leaf handling.  Otherwise, try to add the attribute to the shortform
+	 * list; if there's no room then convert the list to leaf format and try
+	 * again.
 	 */
-	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 (xfs_attr_is_shortform(dp)) {
 
 		/*
-		 * Build initial attribute list (if required).
+		 * If the attr was successfully set in shortform, the
+		 * transaction is committed and set to NULL.  Otherwise, is it
+		 * converted from shortform to leaf, and the transaction is
+		 * retained.
 		 */
-		if (dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS)
-			xfs_attr_shortform_create(args);
+		error = xfs_attr_set_shortform(args, &leaf_bp);
+		if (error || !args->trans)
+			return error;
+	}
 
-		/*
-		 * Try to add the attr to the attribute list in the inode.
-		 */
-		error = xfs_attr_try_sf_addname(dp, args);
+	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
+		error = xfs_attr_leaf_addname(args);
 		if (error != -ENOSPC)
 			return error;
 
 		/*
-		 * It won't fit in the shortform, transform to a leaf block.
-		 * GROT: another possible req'mt for a double-split btree op.
+		 * Promote the attribute list to the Btree format.
 		 */
-		error = xfs_attr_shortform_to_leaf(args, &leaf_bp);
+		error = xfs_attr3_leaf_to_node(args);
 		if (error)
 			return error;
 
 		/*
-		 * Prevent the leaf buffer from being unlocked so that a
-		 * concurrent AIL push cannot grab the half-baked leaf
-		 * buffer and run into problems with the write verifier.
-		 * Once we're done rolling the transaction we can release
-		 * the hold and add the attr to the leaf.
+		 * Finish any deferred work items and roll the transaction once
+		 * more.  The goal here is to call node_addname with the inode
+		 * and transaction in the same state (inode locked and joined,
+		 * transaction clean) no matter how we got to this step.
 		 */
-		xfs_trans_bhold(args->trans, leaf_bp);
 		error = xfs_defer_finish(&args->trans);
-		xfs_trans_bhold_release(args->trans, leaf_bp);
-		if (error) {
-			xfs_trans_brelse(args->trans, leaf_bp);
+		if (error)
+			return error;
+
+		/*
+		 * Commit the current trans (including the inode) and
+		 * start a new one.
+		 */
+		error = xfs_trans_roll_inode(&args->trans, dp);
+		if (error)
 			return error;
-		}
 	}
 
-	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK))
-		error = xfs_attr_leaf_addname(args);
-	else
-		error = xfs_attr_node_addname(args);
+	error = xfs_attr_node_addname(args);
 	return error;
 }
 
 /*
+ * Return EEXIST if attr is found, or ENOATTR if not
+ */
+int
+xfs_has_attr(
+	struct xfs_da_args	*args)
+{
+	struct xfs_inode	*dp = args->dp;
+	struct xfs_buf		*bp = NULL;
+	int			error;
+
+	if (!xfs_inode_hasattr(dp))
+		return -ENOATTR;
+
+	if (dp->i_afp->if_format == XFS_DINODE_FMT_LOCAL) {
+		ASSERT(dp->i_afp->if_flags & XFS_IFINLINE);
+		return xfs_attr_sf_findname(args, NULL, NULL);
+	}
+
+	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
+		error = xfs_attr_leaf_hasname(args, &bp);
+
+		if (bp)
+			xfs_trans_brelse(args->trans, bp);
+
+		return error;
+	}
+
+	return xfs_attr_node_hasname(args, NULL);
+}
+
+/*
  * Remove the attribute specified in @args.
  */
 int
@@ -324,7 +371,7 @@ xfs_attr_remove_args(
 
 	if (!xfs_inode_hasattr(dp)) {
 		error = -ENOATTR;
-	} else if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL) {
+	} else if (dp->i_afp->if_format == XFS_DINODE_FMT_LOCAL) {
 		ASSERT(dp->i_afp->if_flags & XFS_IFINLINE);
 		error = xfs_attr_shortform_remove(args);
 	} else if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
@@ -336,188 +383,140 @@ xfs_attr_remove_args(
 	return error;
 }
 
+/*
+ * Note: If args->value is NULL the attribute will be removed, just like the
+ * Linux ->setattr API.
+ */
 int
 xfs_attr_set(
-	struct xfs_inode	*dp,
-	const unsigned char	*name,
-	size_t			namelen,
-	unsigned char		*value,
-	int			valuelen,
-	int			flags)
+	struct xfs_da_args	*args)
 {
+	struct xfs_inode	*dp = args->dp;
 	struct xfs_mount	*mp = dp->i_mount;
-	struct xfs_da_args	args;
 	struct xfs_trans_res	tres;
-	int			rsvd = (flags & ATTR_ROOT) != 0;
+	bool			rsvd = (args->attr_filter & XFS_ATTR_ROOT);
 	int			error, local;
-
-	XFS_STATS_INC(mp, xs_attr_set);
+	unsigned int		total;
 
 	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
 		return -EIO;
 
-	error = xfs_attr_args_init(&args, dp, name, namelen, flags);
-	if (error)
-		return error;
-
-	args.value = value;
-	args.valuelen = valuelen;
-	args.op_flags = XFS_DA_OP_ADDNAME | XFS_DA_OP_OKNOENT;
-	args.total = xfs_attr_calc_size(&args, &local);
-
 	error = xfs_qm_dqattach(dp);
 	if (error)
 		return error;
 
-	/*
-	 * If the inode doesn't have an attribute fork, add one.
-	 * (inode must not be locked when we call this routine)
-	 */
-	if (XFS_IFORK_Q(dp) == 0) {
-		int sf_size = sizeof(xfs_attr_sf_hdr_t) +
-			XFS_ATTR_SF_ENTSIZE_BYNAME(args.namelen, valuelen);
-
-		error = xfs_bmap_add_attrfork(dp, sf_size, rsvd);
-		if (error)
-			return error;
-	}
-
-	tres.tr_logres = M_RES(mp)->tr_attrsetm.tr_logres +
-			 M_RES(mp)->tr_attrsetrt.tr_logres * args.total;
-	tres.tr_logcount = XFS_ATTRSET_LOG_COUNT;
-	tres.tr_logflags = XFS_TRANS_PERM_LOG_RES;
-
-	/*
-	 * Root fork attributes can use reserved data blocks for this
-	 * operation if necessary
-	 */
-	error = xfs_trans_alloc(mp, &tres, args.total, 0,
-			rsvd ? XFS_TRANS_RESERVE : 0, &args.trans);
-	if (error)
-		return error;
-
-	xfs_ilock(dp, XFS_ILOCK_EXCL);
-	error = xfs_trans_reserve_quota_nblks(args.trans, dp, args.total, 0,
-				rsvd ? XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES :
-				       XFS_QMOPT_RES_REGBLKS);
-	if (error)
-		goto out_trans_cancel;
-
-	xfs_trans_ijoin(args.trans, dp, 0);
-	error = xfs_attr_set_args(&args);
-	if (error)
-		goto out_trans_cancel;
-	if (!args.trans) {
-		/* shortform attribute has already been committed */
-		goto out_unlock;
-	}
-
-	/*
-	 * 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)
-		xfs_trans_set_sync(args.trans);
-
-	if ((flags & ATTR_KERNOTIME) == 0)
-		xfs_trans_ichgtime(args.trans, dp, XFS_ICHGTIME_CHG);
+	args->geo = mp->m_attr_geo;
+	args->whichfork = XFS_ATTR_FORK;
+	args->hashval = xfs_da_hashname(args->name, args->namelen);
 
 	/*
-	 * Commit the last in the sequence of transactions.
+	 * We have no control over the attribute names that userspace passes us
+	 * to remove, so we have to allow the name lookup prior to attribute
+	 * removal to fail as well.
 	 */
-	xfs_trans_log_inode(args.trans, dp, XFS_ILOG_CORE);
-	error = xfs_trans_commit(args.trans);
-out_unlock:
-	xfs_iunlock(dp, XFS_ILOCK_EXCL);
-	return error;
-
-out_trans_cancel:
-	if (args.trans)
-		xfs_trans_cancel(args.trans);
-	goto out_unlock;
-}
+	args->op_flags = XFS_DA_OP_OKNOENT;
 
-/*
- * Generic handler routine to remove a name from an attribute list.
- * Transitions attribute list from Btree to shortform as necessary.
- */
-int
-xfs_attr_remove(
-	struct xfs_inode	*dp,
-	const unsigned char	*name,
-	size_t			namelen,
-	int			flags)
-{
-	struct xfs_mount	*mp = dp->i_mount;
-	struct xfs_da_args	args;
-	int			error;
+	if (args->value) {
+		XFS_STATS_INC(mp, xs_attr_set);
 
-	XFS_STATS_INC(mp, xs_attr_remove);
+		args->op_flags |= XFS_DA_OP_ADDNAME;
+		args->total = xfs_attr_calc_size(args, &local);
 
-	if (XFS_FORCED_SHUTDOWN(dp->i_mount))
-		return -EIO;
+		/*
+		 * If the inode doesn't have an attribute fork, add one.
+		 * (inode must not be locked when we call this routine)
+		 */
+		if (XFS_IFORK_Q(dp) == 0) {
+			int sf_size = sizeof(struct xfs_attr_sf_hdr) +
+				XFS_ATTR_SF_ENTSIZE_BYNAME(args->namelen,
+						args->valuelen);
 
-	error = xfs_attr_args_init(&args, dp, name, namelen, flags);
-	if (error)
-		return error;
+			error = xfs_bmap_add_attrfork(dp, sf_size, rsvd);
+			if (error)
+				return error;
+		}
 
-	/*
-	 * we have no control over the attribute names that userspace passes us
-	 * to remove, so we have to allow the name lookup prior to attribute
-	 * removal to fail.
-	 */
-	args.op_flags = XFS_DA_OP_OKNOENT;
+		tres.tr_logres = M_RES(mp)->tr_attrsetm.tr_logres +
+				 M_RES(mp)->tr_attrsetrt.tr_logres *
+					args->total;
+		tres.tr_logcount = XFS_ATTRSET_LOG_COUNT;
+		tres.tr_logflags = XFS_TRANS_PERM_LOG_RES;
+		total = args->total;
+	} else {
+		XFS_STATS_INC(mp, xs_attr_remove);
 
-	error = xfs_qm_dqattach(dp);
-	if (error)
-		return error;
+		tres = M_RES(mp)->tr_attrrm;
+		total = XFS_ATTRRM_SPACE_RES(mp);
+	}
 
 	/*
 	 * Root fork attributes can use reserved data blocks for this
 	 * operation if necessary
 	 */
-	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_attrrm,
-			XFS_ATTRRM_SPACE_RES(mp), 0,
-			(flags & ATTR_ROOT) ? XFS_TRANS_RESERVE : 0,
-			&args.trans);
+	error = xfs_trans_alloc(mp, &tres, total, 0,
+			rsvd ? XFS_TRANS_RESERVE : 0, &args->trans);
 	if (error)
 		return error;
 
 	xfs_ilock(dp, XFS_ILOCK_EXCL);
-	/*
-	 * No need to make quota reservations here. We expect to release some
-	 * blocks not allocate in the common case.
-	 */
-	xfs_trans_ijoin(args.trans, dp, 0);
+	xfs_trans_ijoin(args->trans, dp, 0);
+	if (args->value) {
+		unsigned int	quota_flags = XFS_QMOPT_RES_REGBLKS;
+
+		if (rsvd)
+			quota_flags |= XFS_QMOPT_FORCE_RES;
+		error = xfs_trans_reserve_quota_nblks(args->trans, dp,
+				args->total, 0, quota_flags);
+		if (error)
+			goto out_trans_cancel;
 
-	error = xfs_attr_remove_args(&args);
-	if (error)
-		goto out;
+		error = xfs_has_attr(args);
+		if (error == -EEXIST && (args->attr_flags & XATTR_CREATE))
+			goto out_trans_cancel;
+		if (error == -ENOATTR && (args->attr_flags & XATTR_REPLACE))
+			goto out_trans_cancel;
+		if (error != -ENOATTR && error != -EEXIST)
+			goto out_trans_cancel;
+
+		error = xfs_attr_set_args(args);
+		if (error)
+			goto out_trans_cancel;
+		/* shortform attribute has already been committed */
+		if (!args->trans)
+			goto out_unlock;
+	} else {
+		error = xfs_has_attr(args);
+		if (error != -EEXIST)
+			goto out_trans_cancel;
+
+		error = xfs_attr_remove_args(args);
+		if (error)
+			goto out_trans_cancel;
+	}
 
 	/*
 	 * 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)
-		xfs_trans_set_sync(args.trans);
+		xfs_trans_set_sync(args->trans);
 
-	if ((flags & ATTR_KERNOTIME) == 0)
-		xfs_trans_ichgtime(args.trans, dp, XFS_ICHGTIME_CHG);
+	if (!(args->op_flags & XFS_DA_OP_NOTIME))
+		xfs_trans_ichgtime(args->trans, dp, XFS_ICHGTIME_CHG);
 
 	/*
 	 * Commit the last in the sequence of transactions.
 	 */
-	xfs_trans_log_inode(args.trans, dp, XFS_ILOG_CORE);
-	error = xfs_trans_commit(args.trans);
+	xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE);
+	error = xfs_trans_commit(args->trans);
+out_unlock:
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);
-
 	return error;
 
-out:
-	if (args.trans)
-		xfs_trans_cancel(args.trans);
-	xfs_iunlock(dp, XFS_ILOCK_EXCL);
-	return error;
+out_trans_cancel:
+	if (args->trans)
+		xfs_trans_cancel(args->trans);
+	goto out_unlock;
 }
 
 /*========================================================================
@@ -536,10 +535,10 @@ xfs_attr_shortform_addname(xfs_da_args_t *args)
 	trace_xfs_attr_sf_addname(args);
 
 	retval = xfs_attr_shortform_lookup(args);
-	if ((args->flags & ATTR_REPLACE) && (retval == -ENOATTR)) {
+	if (retval == -ENOATTR && (args->attr_flags & XATTR_REPLACE))
 		return retval;
-	} else if (retval == -EEXIST) {
-		if (args->flags & ATTR_CREATE)
+	if (retval == -EEXIST) {
+		if (args->attr_flags & XATTR_CREATE)
 			return retval;
 		retval = xfs_attr_shortform_remove(args);
 		if (retval)
@@ -549,7 +548,7 @@ xfs_attr_shortform_addname(xfs_da_args_t *args)
 		 * that the leaf format add routine won't trip over the attr
 		 * not being around.
 		 */
-		args->flags &= ~ATTR_REPLACE;
+		args->attr_flags &= ~XATTR_REPLACE;
 	}
 
 	if (args->namelen >= XFS_ATTR_SF_ENTSIZE_MAX ||
@@ -572,54 +571,65 @@ xfs_attr_shortform_addname(xfs_da_args_t *args)
  * External routines when attribute list is one block
  *========================================================================*/
 
+/* Store info about a remote block */
+STATIC void
+xfs_attr_save_rmt_blk(
+	struct xfs_da_args	*args)
+{
+	args->blkno2 = args->blkno;
+	args->index2 = args->index;
+	args->rmtblkno2 = args->rmtblkno;
+	args->rmtblkcnt2 = args->rmtblkcnt;
+	args->rmtvaluelen2 = args->rmtvaluelen;
+}
+
+/* Set stored info about a remote block */
+STATIC void
+xfs_attr_restore_rmt_blk(
+	struct xfs_da_args	*args)
+{
+	args->blkno = args->blkno2;
+	args->index = args->index2;
+	args->rmtblkno = args->rmtblkno2;
+	args->rmtblkcnt = args->rmtblkcnt2;
+	args->rmtvaluelen = args->rmtvaluelen2;
+}
+
 /*
- * Add a name to the leaf attribute list structure
+ * Tries to add an attribute to an inode in leaf form
  *
- * This leaf block cannot have a "remote" value, we only call this routine
- * if bmap_one_block() says there is only one block (ie: no remote blks).
+ * This function is meant to execute as part of a delayed operation and leaves
+ * the transaction handling to the caller.  On success the attribute is added
+ * and the inode and transaction are left dirty.  If there is not enough space,
+ * the attr data is converted to node format and -ENOSPC is returned. Caller is
+ * responsible for handling the dirty inode and transaction or adding the attr
+ * in node format.
  */
 STATIC int
-xfs_attr_leaf_addname(
-	struct xfs_da_args	*args)
+xfs_attr_leaf_try_add(
+	struct xfs_da_args	*args,
+	struct xfs_buf		*bp)
 {
-	struct xfs_inode	*dp;
-	struct xfs_buf		*bp;
-	int			retval, error, forkoff;
-
-	trace_xfs_attr_leaf_addname(args);
-
-	/*
-	 * Read the (only) block in the attribute list in.
-	 */
-	dp = args->dp;
-	args->blkno = 0;
-	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, &bp);
-	if (error)
-		return error;
+	int			retval;
 
 	/*
 	 * Look up the given attribute in the leaf block.  Figure out if
 	 * the given flags produce an error or call for an atomic rename.
 	 */
-	retval = xfs_attr3_leaf_lookup_int(bp, args);
-	if ((args->flags & ATTR_REPLACE) && (retval == -ENOATTR)) {
-		xfs_trans_brelse(args->trans, bp);
+	retval = xfs_attr_leaf_hasname(args, &bp);
+	if (retval != -ENOATTR && retval != -EEXIST)
 		return retval;
-	} else if (retval == -EEXIST) {
-		if (args->flags & ATTR_CREATE) {	/* pure create op */
-			xfs_trans_brelse(args->trans, bp);
-			return retval;
-		}
+	if (retval == -ENOATTR && (args->attr_flags & XATTR_REPLACE))
+		goto out_brelse;
+	if (retval == -EEXIST) {
+		if (args->attr_flags & XATTR_CREATE)
+			goto out_brelse;
 
 		trace_xfs_attr_leaf_replace(args);
 
 		/* save the attribute state for later removal*/
 		args->op_flags |= XFS_DA_OP_RENAME;	/* an atomic rename */
-		args->blkno2 = args->blkno;		/* set 2nd entry info*/
-		args->index2 = args->index;
-		args->rmtblkno2 = args->rmtblkno;
-		args->rmtblkcnt2 = args->rmtblkcnt;
-		args->rmtvaluelen2 = args->rmtvaluelen;
+		xfs_attr_save_rmt_blk(args);
 
 		/*
 		 * clear the remote attr state now that it is saved so that the
@@ -632,37 +642,35 @@ xfs_attr_leaf_addname(
 	}
 
 	/*
-	 * Add the attribute to the leaf block, transitioning to a Btree
-	 * if required.
+	 * Add the attribute to the leaf block
 	 */
-	retval = xfs_attr3_leaf_add(bp, args);
-	if (retval == -ENOSPC) {
-		/*
-		 * Promote the attribute list to the Btree format, then
-		 * Commit that transaction so that the node_addname() call
-		 * can manage its own transactions.
-		 */
-		error = xfs_attr3_leaf_to_node(args);
-		if (error)
-			return error;
-		error = xfs_defer_finish(&args->trans);
-		if (error)
-			return error;
+	return xfs_attr3_leaf_add(bp, args);
 
-		/*
-		 * Commit the current trans (including the inode) and start
-		 * a new one.
-		 */
-		error = xfs_trans_roll_inode(&args->trans, dp);
-		if (error)
-			return error;
+out_brelse:
+	xfs_trans_brelse(args->trans, bp);
+	return retval;
+}
 
-		/*
-		 * Fob the whole rest of the problem off on the Btree code.
-		 */
-		error = xfs_attr_node_addname(args);
+
+/*
+ * Add a name to the leaf attribute list structure
+ *
+ * This leaf block cannot have a "remote" value, we only call this routine
+ * if bmap_one_block() says there is only one block (ie: no remote blks).
+ */
+STATIC int
+xfs_attr_leaf_addname(
+	struct xfs_da_args	*args)
+{
+	int			error, forkoff;
+	struct xfs_buf		*bp = NULL;
+	struct xfs_inode	*dp = args->dp;
+
+	trace_xfs_attr_leaf_addname(args);
+
+	error = xfs_attr_leaf_try_add(args, bp);
+	if (error)
 		return error;
-	}
 
 	/*
 	 * Commit the transaction that added the attr name so that
@@ -684,71 +692,92 @@ xfs_attr_leaf_addname(
 			return error;
 	}
 
-	/*
-	 * If this is an atomic rename operation, we must "flip" the
-	 * incomplete flags on the "new" and "old" attribute/value pairs
-	 * so that one disappears and one appears atomically.  Then we
-	 * must remove the "old" attribute/value pair.
-	 */
-	if (args->op_flags & XFS_DA_OP_RENAME) {
+	if (!(args->op_flags & XFS_DA_OP_RENAME)) {
 		/*
-		 * In a separate transaction, set the incomplete flag on the
-		 * "old" attr and clear the incomplete flag on the "new" attr.
+		 * Added a "remote" value, just clear the incomplete flag.
 		 */
-		error = xfs_attr3_leaf_flipflags(args);
+		if (args->rmtblkno > 0)
+			error = xfs_attr3_leaf_clearflag(args);
+
+		return error;
+	}
+
+	/*
+	 * If this is an atomic rename operation, we must "flip" the incomplete
+	 * flags on the "new" and "old" attribute/value pairs so that one
+	 * disappears and one appears atomically.  Then we must remove the "old"
+	 * attribute/value pair.
+	 *
+	 * In a separate transaction, set the incomplete flag on the "old" attr
+	 * and clear the incomplete flag on the "new" attr.
+	 */
+
+	error = xfs_attr3_leaf_flipflags(args);
+	if (error)
+		return error;
+	/*
+	 * Commit the flag value change and start the next trans in series.
+	 */
+	error = xfs_trans_roll_inode(&args->trans, args->dp);
+	if (error)
+		return error;
+
+	/*
+	 * Dismantle the "old" attribute/value pair by removing a "remote" value
+	 * (if it exists).
+	 */
+	xfs_attr_restore_rmt_blk(args);
+
+	if (args->rmtblkno) {
+		error = xfs_attr_rmtval_invalidate(args);
 		if (error)
 			return error;
 
-		/*
-		 * Dismantle the "old" attribute/value pair by removing
-		 * a "remote" value (if it exists).
-		 */
-		args->index = args->index2;
-		args->blkno = args->blkno2;
-		args->rmtblkno = args->rmtblkno2;
-		args->rmtblkcnt = args->rmtblkcnt2;
-		args->rmtvaluelen = args->rmtvaluelen2;
-		if (args->rmtblkno) {
-			error = xfs_attr_rmtval_remove(args);
-			if (error)
-				return error;
-		}
-
-		/*
-		 * Read in the block containing the "old" attr, then
-		 * remove the "old" attr from that block (neat, huh!)
-		 */
-		error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno,
-					   &bp);
+		error = xfs_attr_rmtval_remove(args);
 		if (error)
 			return error;
+	}
 
-		xfs_attr3_leaf_remove(bp, args);
+	/*
+	 * Read in the block containing the "old" attr, then remove the "old"
+	 * attr from that block (neat, huh!)
+	 */
+	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno,
+				   &bp);
+	if (error)
+		return error;
 
-		/*
-		 * If the result is small enough, shrink it all into the inode.
-		 */
-		if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) {
-			error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
-			/* bp is gone due to xfs_da_shrink_inode */
-			if (error)
-				return error;
-			error = xfs_defer_finish(&args->trans);
-			if (error)
-				return error;
-		}
+	xfs_attr3_leaf_remove(bp, args);
 
-		/*
-		 * Commit the remove and start the next trans in series.
-		 */
-		error = xfs_trans_roll_inode(&args->trans, dp);
+	/*
+	 * If the result is small enough, shrink it all into the inode.
+	 */
+	forkoff = xfs_attr_shortform_allfit(bp, dp);
+	if (forkoff)
+		error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
+		/* bp is gone due to xfs_da_shrink_inode */
+
+	return error;
+}
+
+/*
+ * Return EEXIST if attr is found, or ENOATTR if not
+ */
+STATIC int
+xfs_attr_leaf_hasname(
+	struct xfs_da_args	*args,
+	struct xfs_buf		**bp)
+{
+	int                     error = 0;
+
+	error = xfs_attr3_leaf_read(args->trans, args->dp, 0, bp);
+	if (error)
+		return error;
+
+	error = xfs_attr3_leaf_lookup_int(*bp, args);
+	if (error != -ENOATTR && error != -EEXIST)
+		xfs_trans_brelse(args->trans, *bp);
 
-	} else if (args->rmtblkno > 0) {
-		/*
-		 * Added a "remote" value, just clear the incomplete flag.
-		 */
-		error = xfs_attr3_leaf_clearflag(args);
-	}
 	return error;
 }
 
@@ -772,31 +801,25 @@ xfs_attr_leaf_removename(
 	 * Remove the attribute.
 	 */
 	dp = args->dp;
-	args->blkno = 0;
-	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, &bp);
-	if (error)
-		return error;
 
-	error = xfs_attr3_leaf_lookup_int(bp, args);
+	error = xfs_attr_leaf_hasname(args, &bp);
+
 	if (error == -ENOATTR) {
 		xfs_trans_brelse(args->trans, bp);
 		return error;
-	}
+	} else if (error != -EEXIST)
+		return error;
 
 	xfs_attr3_leaf_remove(bp, args);
 
 	/*
 	 * If the result is small enough, shrink it all into the inode.
 	 */
-	if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) {
-		error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
+	forkoff = xfs_attr_shortform_allfit(bp, dp);
+	if (forkoff)
+		return xfs_attr3_leaf_to_shortform(bp, args, forkoff);
 		/* bp is gone due to xfs_da_shrink_inode */
-		if (error)
-			return error;
-		error = xfs_defer_finish(&args->trans);
-		if (error)
-			return error;
-	}
+
 	return 0;
 }
 
@@ -816,21 +839,53 @@ xfs_attr_leaf_get(xfs_da_args_t *args)
 
 	trace_xfs_attr_leaf_get(args);
 
-	args->blkno = 0;
-	error = xfs_attr3_leaf_read(args->trans, args->dp, args->blkno, &bp);
-	if (error)
-		return error;
+	error = xfs_attr_leaf_hasname(args, &bp);
 
-	error = xfs_attr3_leaf_lookup_int(bp, args);
-	if (error != -EEXIST)  {
+	if (error == -ENOATTR)  {
 		xfs_trans_brelse(args->trans, bp);
 		return error;
-	}
+	} else if (error != -EEXIST)
+		return error;
+
+
 	error = xfs_attr3_leaf_getvalue(bp, args);
 	xfs_trans_brelse(args->trans, bp);
 	return error;
 }
 
+/*
+ * Return EEXIST if attr is found, or ENOATTR if not
+ * statep: If not null is set to point at the found state.  Caller will
+ *         be responsible for freeing the state in this case.
+ */
+STATIC int
+xfs_attr_node_hasname(
+	struct xfs_da_args	*args,
+	struct xfs_da_state	**statep)
+{
+	struct xfs_da_state	*state;
+	int			retval, error;
+
+	state = xfs_da_state_alloc(args);
+	if (statep != NULL)
+		*statep = NULL;
+
+	/*
+	 * Search to see if name exists, and get back a pointer to it.
+	 */
+	error = xfs_da3_node_lookup_int(state, &retval);
+	if (error) {
+		xfs_da_state_free(state);
+		return error;
+	}
+
+	if (statep != NULL)
+		*statep = state;
+	else
+		xfs_da_state_free(state);
+	return retval;
+}
+
 /*========================================================================
  * External routines when attribute list size > geo->blksize
  *========================================================================*/
@@ -852,7 +907,6 @@ xfs_attr_node_addname(
 	struct xfs_da_state	*state;
 	struct xfs_da_state_blk	*blk;
 	struct xfs_inode	*dp;
-	struct xfs_mount	*mp;
 	int			retval, error;
 
 	trace_xfs_attr_node_addname(args);
@@ -861,36 +915,28 @@ xfs_attr_node_addname(
 	 * Fill in bucket of arguments/results/context to carry around.
 	 */
 	dp = args->dp;
-	mp = dp->i_mount;
 restart:
-	state = xfs_da_state_alloc();
-	state->args = args;
-	state->mp = mp;
-
 	/*
 	 * Search to see if name already exists, and get back a pointer
 	 * to where it should go.
 	 */
-	error = xfs_da3_node_lookup_int(state, &retval);
-	if (error)
+	retval = xfs_attr_node_hasname(args, &state);
+	if (retval != -ENOATTR && retval != -EEXIST)
 		goto out;
+
 	blk = &state->path.blk[ state->path.active-1 ];
 	ASSERT(blk->magic == XFS_ATTR_LEAF_MAGIC);
-	if ((args->flags & ATTR_REPLACE) && (retval == -ENOATTR)) {
+	if (retval == -ENOATTR && (args->attr_flags & XATTR_REPLACE))
 		goto out;
-	} else if (retval == -EEXIST) {
-		if (args->flags & ATTR_CREATE)
+	if (retval == -EEXIST) {
+		if (args->attr_flags & XATTR_CREATE)
 			goto out;
 
 		trace_xfs_attr_node_replace(args);
 
 		/* save the attribute state for later removal*/
 		args->op_flags |= XFS_DA_OP_RENAME;	/* atomic rename op */
-		args->blkno2 = args->blkno;		/* set 2nd entry info*/
-		args->index2 = args->index;
-		args->rmtblkno2 = args->rmtblkno;
-		args->rmtblkcnt2 = args->rmtblkcnt;
-		args->rmtvaluelen2 = args->rmtvaluelen;
+		xfs_attr_save_rmt_blk(args);
 
 		/*
 		 * clear the remote attr state now that it is saved so that the
@@ -976,82 +1022,75 @@ restart:
 			return error;
 	}
 
-	/*
-	 * If this is an atomic rename operation, we must "flip" the
-	 * incomplete flags on the "new" and "old" attribute/value pairs
-	 * so that one disappears and one appears atomically.  Then we
-	 * must remove the "old" attribute/value pair.
-	 */
-	if (args->op_flags & XFS_DA_OP_RENAME) {
+	if (!(args->op_flags & XFS_DA_OP_RENAME)) {
 		/*
-		 * In a separate transaction, set the incomplete flag on the
-		 * "old" attr and clear the incomplete flag on the "new" attr.
+		 * Added a "remote" value, just clear the incomplete flag.
 		 */
-		error = xfs_attr3_leaf_flipflags(args);
-		if (error)
-			goto out;
+		if (args->rmtblkno > 0)
+			error = xfs_attr3_leaf_clearflag(args);
+		retval = error;
+		goto out;
+	}
 
-		/*
-		 * Dismantle the "old" attribute/value pair by removing
-		 * a "remote" value (if it exists).
-		 */
-		args->index = args->index2;
-		args->blkno = args->blkno2;
-		args->rmtblkno = args->rmtblkno2;
-		args->rmtblkcnt = args->rmtblkcnt2;
-		args->rmtvaluelen = args->rmtvaluelen2;
-		if (args->rmtblkno) {
-			error = xfs_attr_rmtval_remove(args);
-			if (error)
-				return error;
-		}
+	/*
+	 * If this is an atomic rename operation, we must "flip" the incomplete
+	 * flags on the "new" and "old" attribute/value pairs so that one
+	 * disappears and one appears atomically.  Then we must remove the "old"
+	 * attribute/value pair.
+	 *
+	 * In a separate transaction, set the incomplete flag on the "old" attr
+	 * and clear the incomplete flag on the "new" attr.
+	 */
+	error = xfs_attr3_leaf_flipflags(args);
+	if (error)
+		goto out;
+	/*
+	 * Commit the flag value change and start the next trans in series
+	 */
+	error = xfs_trans_roll_inode(&args->trans, args->dp);
+	if (error)
+		goto out;
 
-		/*
-		 * Re-find the "old" attribute entry after any split ops.
-		 * The INCOMPLETE flag means that we will find the "old"
-		 * attr, not the "new" one.
-		 */
-		args->op_flags |= XFS_DA_OP_INCOMPLETE;
-		state = xfs_da_state_alloc();
-		state->args = args;
-		state->mp = mp;
-		state->inleaf = 0;
-		error = xfs_da3_node_lookup_int(state, &retval);
+	/*
+	 * Dismantle the "old" attribute/value pair by removing a "remote" value
+	 * (if it exists).
+	 */
+	xfs_attr_restore_rmt_blk(args);
+
+	if (args->rmtblkno) {
+		error = xfs_attr_rmtval_invalidate(args);
 		if (error)
-			goto out;
+			return error;
 
-		/*
-		 * Remove the name and update the hashvals in the tree.
-		 */
-		blk = &state->path.blk[ state->path.active-1 ];
-		ASSERT(blk->magic == XFS_ATTR_LEAF_MAGIC);
-		error = xfs_attr3_leaf_remove(blk->bp, args);
-		xfs_da3_fixhashpath(state, &state->path);
+		error = xfs_attr_rmtval_remove(args);
+		if (error)
+			return error;
+	}
 
-		/*
-		 * Check to see if the tree needs to be collapsed.
-		 */
-		if (retval && (state->path.active > 1)) {
-			error = xfs_da3_join(state);
-			if (error)
-				goto out;
-			error = xfs_defer_finish(&args->trans);
-			if (error)
-				goto out;
-		}
+	/*
+	 * Re-find the "old" attribute entry after any split ops. The INCOMPLETE
+	 * flag means that we will find the "old" attr, not the "new" one.
+	 */
+	args->attr_filter |= XFS_ATTR_INCOMPLETE;
+	state = xfs_da_state_alloc(args);
+	state->inleaf = 0;
+	error = xfs_da3_node_lookup_int(state, &retval);
+	if (error)
+		goto out;
 
-		/*
-		 * Commit and start the next trans in the chain.
-		 */
-		error = xfs_trans_roll_inode(&args->trans, dp);
-		if (error)
-			goto out;
+	/*
+	 * Remove the name and update the hashvals in the tree.
+	 */
+	blk = &state->path.blk[state->path.active-1];
+	ASSERT(blk->magic == XFS_ATTR_LEAF_MAGIC);
+	error = xfs_attr3_leaf_remove(blk->bp, args);
+	xfs_da3_fixhashpath(state, &state->path);
 
-	} else if (args->rmtblkno > 0) {
-		/*
-		 * Added a "remote" value, just clear the incomplete flag.
-		 */
-		error = xfs_attr3_leaf_clearflag(args);
+	/*
+	 * Check to see if the tree needs to be collapsed.
+	 */
+	if (retval && (state->path.active > 1)) {
+		error = xfs_da3_join(state);
 		if (error)
 			goto out;
 	}
@@ -1066,6 +1105,114 @@ out:
 }
 
 /*
+ * Shrink an attribute from leaf to shortform
+ */
+STATIC int
+xfs_attr_node_shrink(
+	struct xfs_da_args	*args,
+	struct xfs_da_state     *state)
+{
+	struct xfs_inode	*dp = args->dp;
+	int			error, forkoff;
+	struct xfs_buf		*bp;
+
+	/*
+	 * Have to get rid of the copy of this dabuf in the state.
+	 */
+	ASSERT(state->path.active == 1);
+	ASSERT(state->path.blk[0].bp);
+	state->path.blk[0].bp = NULL;
+
+	error = xfs_attr3_leaf_read(args->trans, args->dp, 0, &bp);
+	if (error)
+		return error;
+
+	forkoff = xfs_attr_shortform_allfit(bp, dp);
+	if (forkoff) {
+		error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
+		/* bp is gone due to xfs_da_shrink_inode */
+	} else
+		xfs_trans_brelse(args->trans, bp);
+
+	return error;
+}
+
+/*
+ * Mark an attribute entry INCOMPLETE and save pointers to the relevant buffers
+ * for later deletion of the entry.
+ */
+STATIC int
+xfs_attr_leaf_mark_incomplete(
+	struct xfs_da_args	*args,
+	struct xfs_da_state	*state)
+{
+	int			error;
+
+	/*
+	 * Fill in disk block numbers in the state structure
+	 * so that we can get the buffers back after we commit
+	 * several transactions in the following calls.
+	 */
+	error = xfs_attr_fillstate(state);
+	if (error)
+		return error;
+
+	/*
+	 * Mark the attribute as INCOMPLETE
+	 */
+	return xfs_attr3_leaf_setflag(args);
+}
+
+/*
+ * Initial setup for xfs_attr_node_removename.  Make sure the attr is there and
+ * the blocks are valid.  Attr keys with remote blocks will be marked
+ * incomplete.
+ */
+STATIC
+int xfs_attr_node_removename_setup(
+	struct xfs_da_args	*args,
+	struct xfs_da_state	**state)
+{
+	int			error;
+
+	error = xfs_attr_node_hasname(args, state);
+	if (error != -EEXIST)
+		return error;
+
+	ASSERT((*state)->path.blk[(*state)->path.active - 1].bp != NULL);
+	ASSERT((*state)->path.blk[(*state)->path.active - 1].magic ==
+		XFS_ATTR_LEAF_MAGIC);
+
+	if (args->rmtblkno > 0) {
+		error = xfs_attr_leaf_mark_incomplete(args, *state);
+		if (error)
+			return error;
+
+		return xfs_attr_rmtval_invalidate(args);
+	}
+
+	return 0;
+}
+
+STATIC int
+xfs_attr_node_remove_rmt(
+	struct xfs_da_args	*args,
+	struct xfs_da_state	*state)
+{
+	int			error = 0;
+
+	error = xfs_attr_rmtval_remove(args);
+	if (error)
+		return error;
+
+	/*
+	 * Refill the state structure with buffers, the prior calls released our
+	 * buffers.
+	 */
+	return xfs_attr_refillstate(state);
+}
+
+/*
  * Remove a name from a B-tree attribute list.
  *
  * This will involve walking down the Btree, and may involve joining
@@ -1078,64 +1225,22 @@ xfs_attr_node_removename(
 {
 	struct xfs_da_state	*state;
 	struct xfs_da_state_blk	*blk;
-	struct xfs_inode	*dp;
-	struct xfs_buf		*bp;
-	int			retval, error, forkoff;
+	int			retval, error;
+	struct xfs_inode	*dp = args->dp;
 
 	trace_xfs_attr_node_removename(args);
 
-	/*
-	 * Tie a string around our finger to remind us where we are.
-	 */
-	dp = args->dp;
-	state = xfs_da_state_alloc();
-	state->args = args;
-	state->mp = dp->i_mount;
-
-	/*
-	 * Search to see if name exists, and get back a pointer to it.
-	 */
-	error = xfs_da3_node_lookup_int(state, &retval);
-	if (error || (retval != -EEXIST)) {
-		if (error == 0)
-			error = retval;
+	error = xfs_attr_node_removename_setup(args, &state);
+	if (error)
 		goto out;
-	}
 
 	/*
 	 * If there is an out-of-line value, de-allocate the blocks.
 	 * This is done before we remove the attribute so that we don't
 	 * overflow the maximum size of a transaction and/or hit a deadlock.
 	 */
-	blk = &state->path.blk[ state->path.active-1 ];
-	ASSERT(blk->bp != NULL);
-	ASSERT(blk->magic == XFS_ATTR_LEAF_MAGIC);
 	if (args->rmtblkno > 0) {
-		/*
-		 * Fill in disk block numbers in the state structure
-		 * so that we can get the buffers back after we commit
-		 * several transactions in the following calls.
-		 */
-		error = xfs_attr_fillstate(state);
-		if (error)
-			goto out;
-
-		/*
-		 * Mark the attribute as INCOMPLETE, then bunmapi() the
-		 * remote value.
-		 */
-		error = xfs_attr3_leaf_setflag(args);
-		if (error)
-			goto out;
-		error = xfs_attr_rmtval_remove(args);
-		if (error)
-			goto out;
-
-		/*
-		 * Refill the state structure with buffers, the prior calls
-		 * released our buffers.
-		 */
-		error = xfs_attr_refillstate(state);
+		error = xfs_attr_node_remove_rmt(args, state);
 		if (error)
 			goto out;
 	}
@@ -1169,33 +1274,12 @@ xfs_attr_node_removename(
 	/*
 	 * If the result is small enough, push it all into the inode.
 	 */
-	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK)) {
-		/*
-		 * Have to get rid of the copy of this dabuf in the state.
-		 */
-		ASSERT(state->path.active == 1);
-		ASSERT(state->path.blk[0].bp);
-		state->path.blk[0].bp = NULL;
-
-		error = xfs_attr3_leaf_read(args->trans, args->dp, 0, &bp);
-		if (error)
-			goto out;
-
-		if ((forkoff = xfs_attr_shortform_allfit(bp, dp))) {
-			error = xfs_attr3_leaf_to_shortform(bp, args, forkoff);
-			/* bp is gone due to xfs_da_shrink_inode */
-			if (error)
-				goto out;
-			error = xfs_defer_finish(&args->trans);
-			if (error)
-				goto out;
-		} else
-			xfs_trans_brelse(args->trans, bp);
-	}
-	error = 0;
+	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK))
+		error = xfs_attr_node_shrink(args, state);
 
 out:
-	xfs_da_state_free(state);
+	if (state)
+		xfs_da_state_free(state);
 	return error;
 }
 
@@ -1311,47 +1395,41 @@ xfs_attr_refillstate(xfs_da_state_t *state)
  * Returns 0 on successful retrieval, otherwise an error.
  */
 STATIC int
-xfs_attr_node_get(xfs_da_args_t *args)
+xfs_attr_node_get(
+	struct xfs_da_args	*args)
 {
-	xfs_da_state_t *state;
-	xfs_da_state_blk_t *blk;
-	int error, retval;
-	int i;
+	struct xfs_da_state	*state;
+	struct xfs_da_state_blk	*blk;
+	int			i;
+	int			error;
 
 	trace_xfs_attr_node_get(args);
 
-	state = xfs_da_state_alloc();
-	state->args = args;
-	state->mp = args->dp->i_mount;
-
 	/*
 	 * Search to see if name exists, and get back a pointer to it.
 	 */
-	error = xfs_da3_node_lookup_int(state, &retval);
-	if (error) {
-		retval = error;
-		goto out_release;
-	}
-	if (retval != -EEXIST)
+	error = xfs_attr_node_hasname(args, &state);
+	if (error != -EEXIST)
 		goto out_release;
 
 	/*
 	 * Get the value, local or "remote"
 	 */
 	blk = &state->path.blk[state->path.active - 1];
-	retval = xfs_attr3_leaf_getvalue(blk->bp, args);
+	error = xfs_attr3_leaf_getvalue(blk->bp, args);
 
 	/*
 	 * If not in a transaction, we have to release all the buffers.
 	 */
 out_release:
-	for (i = 0; i < state->path.active; i++) {
+	for (i = 0; state != NULL && i < state->path.active; i++) {
 		xfs_trans_brelse(args->trans, state->path.blk[i].bp);
 		state->path.blk[i].bp = NULL;
 	}
 
-	xfs_da_state_free(state);
-	return retval;
+	if (state)
+		xfs_da_state_free(state);
+	return error;
 }
 
 /* Returns true if the attribute entry name is valid. */
diff --git a/fs/xfs/libxfs/xfs_attr.h b/fs/xfs/libxfs/xfs_attr.h
index 4243b2272642..3e97a935e712 100644
--- a/fs/xfs/libxfs/xfs_attr.h
+++ b/fs/xfs/libxfs/xfs_attr.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000,2002-2003,2005 Silicon Graphics, Inc.
  * All Rights Reserved.
@@ -21,39 +21,6 @@ struct xfs_attr_list_context;
  * as possible so as to fit into the literal area of the inode.
  */
 
-/*========================================================================
- * External interfaces
- *========================================================================*/
-
-
-#define ATTR_DONTFOLLOW	0x0001	/* -- ignored, from IRIX -- */
-#define ATTR_ROOT	0x0002	/* use attrs in root (trusted) namespace */
-#define ATTR_TRUST	0x0004	/* -- unused, from IRIX -- */
-#define ATTR_SECURE	0x0008	/* use attrs in security namespace */
-#define ATTR_CREATE	0x0010	/* pure create: fail if attr already exists */
-#define ATTR_REPLACE	0x0020	/* pure set: fail if attr does not exist */
-
-#define ATTR_KERNOTIME	0x1000	/* [kernel] don't update inode timestamps */
-#define ATTR_KERNOVAL	0x2000	/* [kernel] get attr size only, not value */
-
-#define ATTR_INCOMPLETE	0x4000	/* [kernel] return INCOMPLETE attr keys */
-#define ATTR_ALLOC	0x8000	/* [kernel] allocate xattr buffer on demand */
-
-#define ATTR_KERNEL_FLAGS \
-	(ATTR_KERNOTIME | ATTR_KERNOVAL | ATTR_INCOMPLETE | ATTR_ALLOC)
-
-#define XFS_ATTR_FLAGS \
-	{ ATTR_DONTFOLLOW, 	"DONTFOLLOW" }, \
-	{ ATTR_ROOT,		"ROOT" }, \
-	{ ATTR_TRUST,		"TRUST" }, \
-	{ ATTR_SECURE,		"SECURE" }, \
-	{ ATTR_CREATE,		"CREATE" }, \
-	{ ATTR_REPLACE,		"REPLACE" }, \
-	{ ATTR_KERNOTIME,	"KERNOTIME" }, \
-	{ ATTR_KERNOVAL,	"KERNOVAL" }, \
-	{ ATTR_INCOMPLETE,	"INCOMPLETE" }, \
-	{ ATTR_ALLOC,		"ALLOC" }
-
 /*
  * The maximum size (into the kernel or returned from the kernel) of an
  * attribute value or the buffer used for an attr_list() call.  Larger
@@ -62,45 +29,16 @@ struct xfs_attr_list_context;
 #define	ATTR_MAX_VALUELEN	(64*1024)	/* max length of a value */
 
 /*
- * Define how lists of attribute names are returned to the user from
- * the attr_list() call.  A large, 32bit aligned, buffer is passed in
- * along with its size.  We put an array of offsets at the top that each
- * reference an attrlist_ent_t and pack the attrlist_ent_t's at the bottom.
- */
-typedef struct attrlist {
-	__s32	al_count;	/* number of entries in attrlist */
-	__s32	al_more;	/* T/F: more attrs (do call again) */
-	__s32	al_offset[1];	/* byte offsets of attrs [var-sized] */
-} attrlist_t;
-
-/*
- * Show the interesting info about one attribute.  This is what the
- * al_offset[i] entry points to.
- */
-typedef struct attrlist_ent {	/* data from attr_list() */
-	__u32	a_valuelen;	/* number bytes in value of attr */
-	char	a_name[1];	/* attr name (NULL terminated) */
-} attrlist_ent_t;
-
-/*
- * Given a pointer to the (char*) buffer containing the attr_list() result,
- * and an index, return a pointer to the indicated attribute in the buffer.
- */
-#define	ATTR_ENTRY(buffer, index)		\
-	((attrlist_ent_t *)			\
-	 &((char *)buffer)[ ((attrlist_t *)(buffer))->al_offset[index] ])
-
-/*
  * Kernel-internal version of the attrlist cursor.
  */
-typedef struct attrlist_cursor_kern {
+struct xfs_attrlist_cursor_kern {
 	__u32	hashval;	/* hash value of next entry to add */
 	__u32	blkno;		/* block containing entry (suggestion) */
 	__u32	offset;		/* offset in list of equal-hashvals */
 	__u16	pad1;		/* padding to match user-level */
 	__u8	pad2;		/* padding to match user-level */
 	__u8	initted;	/* T/F: cursor has been initialized */
-} attrlist_cursor_kern_t;
+};
 
 
 /*========================================================================
@@ -112,27 +50,28 @@ typedef struct attrlist_cursor_kern {
 typedef void (*put_listent_func_t)(struct xfs_attr_list_context *, int,
 			      unsigned char *, int, int);
 
-typedef struct xfs_attr_list_context {
-	struct xfs_trans		*tp;
-	struct xfs_inode		*dp;		/* inode */
-	struct attrlist_cursor_kern	*cursor;	/* position in list */
-	char				*alist;		/* output buffer */
+struct xfs_attr_list_context {
+	struct xfs_trans	*tp;
+	struct xfs_inode	*dp;		/* inode */
+	struct xfs_attrlist_cursor_kern cursor;	/* position in list */
+	void			*buffer;	/* output buffer */
 
 	/*
 	 * Abort attribute list iteration if non-zero.  Can be used to pass
 	 * error values to the xfs_attr_list caller.
 	 */
-	int				seen_enough;
+	int			seen_enough;
+	bool			allow_incomplete;
 
-	ssize_t				count;		/* num used entries */
-	int				dupcnt;		/* count dup hashvals seen */
-	int				bufsize;	/* total buffer size */
-	int				firstu;		/* first used byte in buffer */
-	int				flags;		/* from VOP call */
-	int				resynch;	/* T/F: resynch with cursor */
-	put_listent_func_t		put_listent;	/* list output fmt function */
-	int				index;		/* index into output buffer */
-} xfs_attr_list_context_t;
+	ssize_t			count;		/* num used entries */
+	int			dupcnt;		/* count dup hashvals seen */
+	int			bufsize;	/* total buffer size */
+	int			firstu;		/* first used byte in buffer */
+	unsigned int		attr_filter;	/* XFS_ATTR_{ROOT,SECURE} */
+	int			resynch;	/* T/F: resynch with cursor */
+	put_listent_func_t	put_listent;	/* list output fmt function */
+	int			index;		/* index into output buffer */
+};
 
 
 /*========================================================================
@@ -143,21 +82,15 @@ typedef struct xfs_attr_list_context {
  * Overall external interface routines.
  */
 int xfs_attr_inactive(struct xfs_inode *dp);
-int xfs_attr_list_int_ilocked(struct xfs_attr_list_context *);
-int xfs_attr_list_int(struct xfs_attr_list_context *);
+int xfs_attr_list_ilocked(struct xfs_attr_list_context *);
+int xfs_attr_list(struct xfs_attr_list_context *);
 int xfs_inode_hasattr(struct xfs_inode *ip);
-int xfs_attr_get_ilocked(struct xfs_inode *ip, struct xfs_da_args *args);
-int xfs_attr_get(struct xfs_inode *ip, const unsigned char *name,
-		 size_t namelen, unsigned char **value, int *valuelenp,
-		 int flags);
-int xfs_attr_set(struct xfs_inode *dp, const unsigned char *name,
-		 size_t namelen, unsigned char *value, int valuelen, int flags);
+int xfs_attr_get_ilocked(struct xfs_da_args *args);
+int xfs_attr_get(struct xfs_da_args *args);
+int xfs_attr_set(struct xfs_da_args *args);
 int xfs_attr_set_args(struct xfs_da_args *args);
-int xfs_attr_remove(struct xfs_inode *dp, const unsigned char *name,
-		    size_t namelen, int flags);
+int xfs_has_attr(struct xfs_da_args *args);
 int xfs_attr_remove_args(struct xfs_da_args *args);
-int xfs_attr_list(struct xfs_inode *dp, char *buffer, int bufsize,
-		  int flags, struct attrlist_cursor_kern *cursor);
 bool xfs_attr_namecheck(const void *name, size_t length);
 
 #endif	/* __XFS_ATTR_H__ */
diff --git a/fs/xfs/libxfs/xfs_attr_leaf.c b/fs/xfs/libxfs/xfs_attr_leaf.c
index fed537a4353d..8623c815164a 100644
--- a/fs/xfs/libxfs/xfs_attr_leaf.c
+++ b/fs/xfs/libxfs/xfs_attr_leaf.c
@@ -309,14 +309,6 @@ xfs_attr3_leaf_verify(
 		return fa;
 
 	/*
-	 * In recovery there is a transient state where count == 0 is valid
-	 * because we may have transitioned an empty shortform attr to a leaf
-	 * if the attr didn't fit in shortform.
-	 */
-	if (!xfs_log_in_recovery(mp) && ichdr.count == 0)
-		return __this_address;
-
-	/*
 	 * firstused is the block offset of the first name info structure.
 	 * Make sure it doesn't go off the block or crash into the header.
 	 */
@@ -331,6 +323,13 @@ xfs_attr3_leaf_verify(
 	    (char *)bp->b_addr + ichdr.firstused)
 		return __this_address;
 
+	/*
+	 * NOTE: This verifier historically failed empty leaf buffers because
+	 * we expect the fork to be in another format. Empty attr fork format
+	 * conversions are possible during xattr set, however, and format
+	 * conversion is not atomic with the xattr set that triggers it. We
+	 * cannot assume leaf blocks are non-empty until that is addressed.
+	*/
 	buf_end = (char *)bp->b_addr + mp->m_attr_geo->blksize;
 	for (i = 0, ent = entries; i < ichdr.count; ent++, i++) {
 		fa = xfs_attr3_leaf_verify_entry(mp, buf_end, leaf, &ichdr,
@@ -445,14 +444,25 @@ xfs_attr3_leaf_read(
  * Namespace helper routines
  *========================================================================*/
 
-/*
- * If namespace bits don't match return 0.
- * If all match then return 1.
- */
-STATIC int
-xfs_attr_namesp_match(int arg_flags, int ondisk_flags)
+static bool
+xfs_attr_match(
+	struct xfs_da_args	*args,
+	uint8_t			namelen,
+	unsigned char		*name,
+	int			flags)
 {
-	return XFS_ATTR_NSP_ONDISK(ondisk_flags) == XFS_ATTR_NSP_ARGS_TO_ONDISK(arg_flags);
+	if (args->namelen != namelen)
+		return false;
+	if (memcmp(args->name, name, namelen) != 0)
+		return false;
+	/*
+	 * If we are looking for incomplete entries, show only those, else only
+	 * show complete entries.
+	 */
+	if (args->attr_filter !=
+	    (flags & (XFS_ATTR_NSP_ONDISK_MASK | XFS_ATTR_INCOMPLETE)))
+		return false;
+	return true;
 }
 
 static int
@@ -464,7 +474,7 @@ xfs_attr_copy_value(
 	/*
 	 * No copy if all we have to do is get the length
 	 */
-	if (args->flags & ATTR_KERNOVAL) {
+	if (!args->valuelen) {
 		args->valuelen = valuelen;
 		return 0;
 	}
@@ -477,8 +487,8 @@ xfs_attr_copy_value(
 		return -ERANGE;
 	}
 
-	if (args->op_flags & XFS_DA_OP_ALLOCVAL) {
-		args->value = kmem_alloc_large(valuelen, 0);
+	if (!args->value) {
+		args->value = kmem_alloc_large(valuelen, KM_NOLOCKDEP);
 		if (!args->value)
 			return -ENOMEM;
 	}
@@ -526,9 +536,9 @@ xfs_attr_shortform_bytesfit(
 	int			offset;
 
 	/* rounded down */
-	offset = (XFS_LITINO(mp, dp->i_d.di_version) - bytes) >> 3;
+	offset = (XFS_LITINO(mp) - bytes) >> 3;
 
-	if (dp->i_d.di_format == XFS_DINODE_FMT_DEV) {
+	if (dp->i_df.if_format == XFS_DINODE_FMT_DEV) {
 		minforkoff = roundup(sizeof(xfs_dev_t), 8) >> 3;
 		return (offset >= minforkoff) ? minforkoff : 0;
 	}
@@ -556,7 +566,7 @@ xfs_attr_shortform_bytesfit(
 
 	dsize = dp->i_df.if_bytes;
 
-	switch (dp->i_d.di_format) {
+	switch (dp->i_df.if_format) {
 	case XFS_DINODE_FMT_EXTENTS:
 		/*
 		 * If there is no attr fork and the data fork is extents, 
@@ -593,8 +603,7 @@ xfs_attr_shortform_bytesfit(
 	minforkoff = roundup(minforkoff, 8) >> 3;
 
 	/* attr fork btree root can have at least this many key/ptr pairs */
-	maxforkoff = XFS_LITINO(mp, dp->i_d.di_version) -
-			XFS_BMDR_SPACE_CALC(MINABTPTRS);
+	maxforkoff = XFS_LITINO(mp) - XFS_BMDR_SPACE_CALC(MINABTPTRS);
 	maxforkoff = maxforkoff >> 3;	/* rounded down */
 
 	if (offset >= maxforkoff)
@@ -626,22 +635,19 @@ xfs_sbversion_add_attr2(xfs_mount_t *mp, xfs_trans_t *tp)
  * Create the initial contents of a shortform attribute list.
  */
 void
-xfs_attr_shortform_create(xfs_da_args_t *args)
+xfs_attr_shortform_create(
+	struct xfs_da_args	*args)
 {
-	xfs_attr_sf_hdr_t *hdr;
-	xfs_inode_t *dp;
-	struct xfs_ifork *ifp;
+	struct xfs_inode	*dp = args->dp;
+	struct xfs_ifork	*ifp = dp->i_afp;
+	struct xfs_attr_sf_hdr	*hdr;
 
 	trace_xfs_attr_sf_create(args);
 
-	dp = args->dp;
-	ASSERT(dp != NULL);
-	ifp = dp->i_afp;
-	ASSERT(ifp != NULL);
 	ASSERT(ifp->if_bytes == 0);
-	if (dp->i_d.di_aformat == XFS_DINODE_FMT_EXTENTS) {
+	if (ifp->if_format == XFS_DINODE_FMT_EXTENTS) {
 		ifp->if_flags &= ~XFS_IFEXTENTS;	/* just in case */
-		dp->i_d.di_aformat = XFS_DINODE_FMT_LOCAL;
+		ifp->if_format = XFS_DINODE_FMT_LOCAL;
 		ifp->if_flags |= XFS_IFINLINE;
 	} else {
 		ASSERT(ifp->if_flags & XFS_IFINLINE);
@@ -654,18 +660,65 @@ xfs_attr_shortform_create(xfs_da_args_t *args)
 }
 
 /*
+ * Return -EEXIST if attr is found, or -ENOATTR if not
+ * args:  args containing attribute name and namelen
+ * sfep:  If not null, pointer will be set to the last attr entry found on
+	  -EEXIST.  On -ENOATTR pointer is left at the last entry in the list
+ * basep: If not null, pointer is set to the byte offset of the entry in the
+ *	  list on -EEXIST.  On -ENOATTR, pointer is left at the byte offset of
+ *	  the last entry in the list
+ */
+int
+xfs_attr_sf_findname(
+	struct xfs_da_args	 *args,
+	struct xfs_attr_sf_entry **sfep,
+	unsigned int		 *basep)
+{
+	struct xfs_attr_shortform *sf;
+	struct xfs_attr_sf_entry *sfe;
+	unsigned int		base = sizeof(struct xfs_attr_sf_hdr);
+	int			size = 0;
+	int			end;
+	int			i;
+
+	sf = (struct xfs_attr_shortform *)args->dp->i_afp->if_u1.if_data;
+	sfe = &sf->list[0];
+	end = sf->hdr.count;
+	for (i = 0; i < end; sfe = XFS_ATTR_SF_NEXTENTRY(sfe),
+			     base += size, i++) {
+		size = XFS_ATTR_SF_ENTSIZE(sfe);
+		if (!xfs_attr_match(args, sfe->namelen, sfe->nameval,
+				    sfe->flags))
+			continue;
+		break;
+	}
+
+	if (sfep != NULL)
+		*sfep = sfe;
+
+	if (basep != NULL)
+		*basep = base;
+
+	if (i == end)
+		return -ENOATTR;
+	return -EEXIST;
+}
+
+/*
  * Add a name/value pair to the shortform attribute list.
  * Overflow from the inode has already been checked for.
  */
 void
-xfs_attr_shortform_add(xfs_da_args_t *args, int forkoff)
+xfs_attr_shortform_add(
+	struct xfs_da_args		*args,
+	int				forkoff)
 {
-	xfs_attr_shortform_t *sf;
-	xfs_attr_sf_entry_t *sfe;
-	int i, offset, size;
-	xfs_mount_t *mp;
-	xfs_inode_t *dp;
-	struct xfs_ifork *ifp;
+	struct xfs_attr_shortform	*sf;
+	struct xfs_attr_sf_entry	*sfe;
+	int				offset, size;
+	struct xfs_mount		*mp;
+	struct xfs_inode		*dp;
+	struct xfs_ifork		*ifp;
 
 	trace_xfs_attr_sf_add(args);
 
@@ -676,18 +729,8 @@ xfs_attr_shortform_add(xfs_da_args_t *args, int forkoff)
 	ifp = dp->i_afp;
 	ASSERT(ifp->if_flags & XFS_IFINLINE);
 	sf = (xfs_attr_shortform_t *)ifp->if_u1.if_data;
-	sfe = &sf->list[0];
-	for (i = 0; i < sf->hdr.count; sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
-#ifdef DEBUG
-		if (sfe->namelen != args->namelen)
-			continue;
-		if (memcmp(args->name, sfe->nameval, args->namelen) != 0)
-			continue;
-		if (!xfs_attr_namesp_match(args->flags, sfe->flags))
-			continue;
+	if (xfs_attr_sf_findname(args, &sfe, NULL) == -EEXIST)
 		ASSERT(0);
-#endif
-	}
 
 	offset = (char *)sfe - (char *)sf;
 	size = XFS_ATTR_SF_ENTSIZE_BYNAME(args->namelen, args->valuelen);
@@ -697,7 +740,7 @@ xfs_attr_shortform_add(xfs_da_args_t *args, int forkoff)
 
 	sfe->namelen = args->namelen;
 	sfe->valuelen = args->valuelen;
-	sfe->flags = XFS_ATTR_NSP_ARGS_TO_ONDISK(args->flags);
+	sfe->flags = args->attr_filter;
 	memcpy(sfe->nameval, args->name, args->namelen);
 	memcpy(&sfe->nameval[args->namelen], args->value, args->valuelen);
 	sf->hdr.count++;
@@ -716,13 +759,12 @@ xfs_attr_fork_remove(
 	struct xfs_inode	*ip,
 	struct xfs_trans	*tp)
 {
-	xfs_idestroy_fork(ip, XFS_ATTR_FORK);
-	ip->i_d.di_forkoff = 0;
-	ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
-
-	ASSERT(ip->i_d.di_anextents == 0);
-	ASSERT(ip->i_afp == NULL);
+	ASSERT(ip->i_afp->if_nextents == 0);
 
+	xfs_idestroy_fork(ip->i_afp);
+	kmem_cache_free(xfs_ifork_zone, ip->i_afp);
+	ip->i_afp = NULL;
+	ip->i_d.di_forkoff = 0;
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 }
 
@@ -730,35 +772,27 @@ xfs_attr_fork_remove(
  * Remove an attribute from the shortform attribute list structure.
  */
 int
-xfs_attr_shortform_remove(xfs_da_args_t *args)
+xfs_attr_shortform_remove(
+	struct xfs_da_args		*args)
 {
-	xfs_attr_shortform_t *sf;
-	xfs_attr_sf_entry_t *sfe;
-	int base, size=0, end, totsize, i;
-	xfs_mount_t *mp;
-	xfs_inode_t *dp;
+	struct xfs_attr_shortform	*sf;
+	struct xfs_attr_sf_entry	*sfe;
+	int				size = 0, end, totsize;
+	unsigned int			base;
+	struct xfs_mount		*mp;
+	struct xfs_inode		*dp;
+	int				error;
 
 	trace_xfs_attr_sf_remove(args);
 
 	dp = args->dp;
 	mp = dp->i_mount;
-	base = sizeof(xfs_attr_sf_hdr_t);
 	sf = (xfs_attr_shortform_t *)dp->i_afp->if_u1.if_data;
-	sfe = &sf->list[0];
-	end = sf->hdr.count;
-	for (i = 0; i < end; sfe = XFS_ATTR_SF_NEXTENTRY(sfe),
-					base += size, i++) {
-		size = XFS_ATTR_SF_ENTSIZE(sfe);
-		if (sfe->namelen != args->namelen)
-			continue;
-		if (memcmp(sfe->nameval, args->name, args->namelen) != 0)
-			continue;
-		if (!xfs_attr_namesp_match(args->flags, sfe->flags))
-			continue;
-		break;
-	}
-	if (i == end)
-		return -ENOATTR;
+
+	error = xfs_attr_sf_findname(args, &sfe, &base);
+	if (error != -EEXIST)
+		return error;
+	size = XFS_ATTR_SF_ENTSIZE(sfe);
 
 	/*
 	 * Fix up the attribute fork data, covering the hole
@@ -776,7 +810,7 @@ xfs_attr_shortform_remove(xfs_da_args_t *args)
 	totsize -= size;
 	if (totsize == sizeof(xfs_attr_sf_hdr_t) &&
 	    (mp->m_flags & XFS_MOUNT_ATTR2) &&
-	    (dp->i_d.di_format != XFS_DINODE_FMT_BTREE) &&
+	    (dp->i_df.if_format != XFS_DINODE_FMT_BTREE) &&
 	    !(args->op_flags & XFS_DA_OP_ADDNAME)) {
 		xfs_attr_fork_remove(dp, args->trans);
 	} else {
@@ -786,7 +820,7 @@ xfs_attr_shortform_remove(xfs_da_args_t *args)
 		ASSERT(totsize > sizeof(xfs_attr_sf_hdr_t) ||
 				(args->op_flags & XFS_DA_OP_ADDNAME) ||
 				!(mp->m_flags & XFS_MOUNT_ATTR2) ||
-				dp->i_d.di_format == XFS_DINODE_FMT_BTREE);
+				dp->i_df.if_format == XFS_DINODE_FMT_BTREE);
 		xfs_trans_log_inode(args->trans, dp,
 					XFS_ILOG_CORE | XFS_ILOG_ADATA);
 	}
@@ -816,13 +850,9 @@ xfs_attr_shortform_lookup(xfs_da_args_t *args)
 	sfe = &sf->list[0];
 	for (i = 0; i < sf->hdr.count;
 				sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
-		if (sfe->namelen != args->namelen)
-			continue;
-		if (memcmp(args->name, sfe->nameval, args->namelen) != 0)
-			continue;
-		if (!xfs_attr_namesp_match(args->flags, sfe->flags))
-			continue;
-		return -EEXIST;
+		if (xfs_attr_match(args, sfe->namelen, sfe->nameval,
+				sfe->flags))
+			return -EEXIST;
 	}
 	return -ENOATTR;
 }
@@ -830,9 +860,9 @@ xfs_attr_shortform_lookup(xfs_da_args_t *args)
 /*
  * Retrieve the attribute value and length.
  *
- * If ATTR_KERNOVAL is specified, only the length needs to be returned.
- * Unlike a lookup, we only return an error if the attribute does not
- * exist or we can't retrieve the value.
+ * If args->valuelen is zero, only the length needs to be returned.  Unlike a
+ * lookup, we only return an error if the attribute does not exist or we can't
+ * retrieve the value.
  */
 int
 xfs_attr_shortform_getvalue(
@@ -847,14 +877,10 @@ xfs_attr_shortform_getvalue(
 	sfe = &sf->list[0];
 	for (i = 0; i < sf->hdr.count;
 				sfe = XFS_ATTR_SF_NEXTENTRY(sfe), i++) {
-		if (sfe->namelen != args->namelen)
-			continue;
-		if (memcmp(args->name, sfe->nameval, args->namelen) != 0)
-			continue;
-		if (!xfs_attr_namesp_match(args->flags, sfe->flags))
-			continue;
-		return xfs_attr_copy_value(args, &sfe->nameval[args->namelen],
-						sfe->valuelen);
+		if (xfs_attr_match(args, sfe->namelen, sfe->nameval,
+				sfe->flags))
+			return xfs_attr_copy_value(args,
+				&sfe->nameval[args->namelen], sfe->valuelen);
 	}
 	return -ENOATTR;
 }
@@ -918,7 +944,7 @@ xfs_attr_shortform_to_leaf(
 		nargs.valuelen = sfe->valuelen;
 		nargs.hashval = xfs_da_hashname(sfe->nameval,
 						sfe->namelen);
-		nargs.flags = XFS_ATTR_NSP_ONDISK_TO_ARGS(sfe->flags);
+		nargs.attr_filter = sfe->flags & XFS_ATTR_NSP_ONDISK_MASK;
 		error = xfs_attr3_leaf_lookup_int(bp, &nargs); /* set a->index */
 		ASSERT(error == -ENOATTR);
 		error = xfs_attr3_leaf_add(bp, &nargs);
@@ -971,7 +997,7 @@ xfs_attr_shortform_allfit(
 				+ be16_to_cpu(name_loc->valuelen);
 	}
 	if ((dp->i_mount->m_flags & XFS_MOUNT_ATTR2) &&
-	    (dp->i_d.di_format != XFS_DINODE_FMT_BTREE) &&
+	    (dp->i_df.if_format != XFS_DINODE_FMT_BTREE) &&
 	    (bytes == sizeof(struct xfs_attr_sf_hdr)))
 		return -1;
 	return xfs_attr_shortform_bytesfit(dp, bytes);
@@ -990,7 +1016,7 @@ xfs_attr_shortform_verify(
 	int				i;
 	int64_t				size;
 
-	ASSERT(ip->i_d.di_aformat == XFS_DINODE_FMT_LOCAL);
+	ASSERT(ip->i_afp->if_format == XFS_DINODE_FMT_LOCAL);
 	ifp = XFS_IFORK_PTR(ip, XFS_ATTR_FORK);
 	sfp = (struct xfs_attr_shortform *)ifp->if_u1.if_data;
 	size = ifp->if_bytes;
@@ -1094,7 +1120,7 @@ xfs_attr3_leaf_to_shortform(
 
 	if (forkoff == -1) {
 		ASSERT(dp->i_mount->m_flags & XFS_MOUNT_ATTR2);
-		ASSERT(dp->i_d.di_format != XFS_DINODE_FMT_BTREE);
+		ASSERT(dp->i_df.if_format != XFS_DINODE_FMT_BTREE);
 		xfs_attr_fork_remove(dp, args->trans);
 		goto out;
 	}
@@ -1124,7 +1150,7 @@ xfs_attr3_leaf_to_shortform(
 		nargs.value = &name_loc->nameval[nargs.namelen];
 		nargs.valuelen = be16_to_cpu(name_loc->valuelen);
 		nargs.hashval = be32_to_cpu(entry->hashval);
-		nargs.flags = XFS_ATTR_NSP_ONDISK_TO_ARGS(entry->flags);
+		nargs.attr_filter = entry->flags & XFS_ATTR_NSP_ONDISK_MASK;
 		xfs_attr_shortform_add(&nargs, forkoff);
 	}
 	error = 0;
@@ -1449,8 +1475,9 @@ xfs_attr3_leaf_add_work(
 	entry->nameidx = cpu_to_be16(ichdr->freemap[mapindex].base +
 				     ichdr->freemap[mapindex].size);
 	entry->hashval = cpu_to_be32(args->hashval);
-	entry->flags = tmp ? XFS_ATTR_LOCAL : 0;
-	entry->flags |= XFS_ATTR_NSP_ARGS_TO_ONDISK(args->flags);
+	entry->flags = args->attr_filter;
+	if (tmp)
+		entry->flags |= XFS_ATTR_LOCAL;
 	if (args->op_flags & XFS_DA_OP_RENAME) {
 		entry->flags |= XFS_ATTR_INCOMPLETE;
 		if ((args->blkno2 == args->blkno) &&
@@ -2346,7 +2373,7 @@ xfs_attr3_leaf_lookup_int(
 	xfs_attr3_leaf_hdr_from_disk(args->geo, &ichdr, leaf);
 	entries = xfs_attr3_leaf_entryp(leaf);
 	if (ichdr.count >= args->geo->blksize / 8) {
-		xfs_buf_corruption_error(bp);
+		xfs_buf_mark_corrupt(bp);
 		return -EFSCORRUPTED;
 	}
 
@@ -2365,11 +2392,11 @@ xfs_attr3_leaf_lookup_int(
 			break;
 	}
 	if (!(probe >= 0 && (!ichdr.count || probe < ichdr.count))) {
-		xfs_buf_corruption_error(bp);
+		xfs_buf_mark_corrupt(bp);
 		return -EFSCORRUPTED;
 	}
 	if (!(span <= 4 || be32_to_cpu(entry->hashval) == hashval)) {
-		xfs_buf_corruption_error(bp);
+		xfs_buf_mark_corrupt(bp);
 		return -EFSCORRUPTED;
 	}
 
@@ -2399,33 +2426,17 @@ xfs_attr3_leaf_lookup_int(
 /*
  * GROT: Add code to remove incomplete entries.
  */
-		/*
-		 * If we are looking for INCOMPLETE entries, show only those.
-		 * If we are looking for complete entries, show only those.
-		 */
-		if (!!(args->op_flags & XFS_DA_OP_INCOMPLETE) !=
-		    !!(entry->flags & XFS_ATTR_INCOMPLETE)) {
-			continue;
-		}
 		if (entry->flags & XFS_ATTR_LOCAL) {
 			name_loc = xfs_attr3_leaf_name_local(leaf, probe);
-			if (name_loc->namelen != args->namelen)
-				continue;
-			if (memcmp(args->name, name_loc->nameval,
-							args->namelen) != 0)
-				continue;
-			if (!xfs_attr_namesp_match(args->flags, entry->flags))
+			if (!xfs_attr_match(args, name_loc->namelen,
+					name_loc->nameval, entry->flags))
 				continue;
 			args->index = probe;
 			return -EEXIST;
 		} else {
 			name_rmt = xfs_attr3_leaf_name_remote(leaf, probe);
-			if (name_rmt->namelen != args->namelen)
-				continue;
-			if (memcmp(args->name, name_rmt->name,
-							args->namelen) != 0)
-				continue;
-			if (!xfs_attr_namesp_match(args->flags, entry->flags))
+			if (!xfs_attr_match(args, name_rmt->namelen,
+					name_rmt->name, entry->flags))
 				continue;
 			args->index = probe;
 			args->rmtvaluelen = be32_to_cpu(name_rmt->valuelen);
@@ -2444,9 +2455,9 @@ xfs_attr3_leaf_lookup_int(
  * Get the value associated with an attribute name from a leaf attribute
  * list structure.
  *
- * If ATTR_KERNOVAL is specified, only the length needs to be returned.
- * Unlike a lookup, we only return an error if the attribute does not
- * exist or we can't retrieve the value.
+ * If args->valuelen is zero, only the length needs to be returned.  Unlike a
+ * lookup, we only return an error if the attribute does not exist or we can't
+ * retrieve the value.
  */
 int
 xfs_attr3_leaf_getvalue(
@@ -2771,10 +2782,7 @@ xfs_attr3_leaf_clearflag(
 			 XFS_DA_LOGRANGE(leaf, name_rmt, sizeof(*name_rmt)));
 	}
 
-	/*
-	 * Commit the flag value change and start the next trans in series.
-	 */
-	return xfs_trans_roll_inode(&args->trans, args->dp);
+	return 0;
 }
 
 /*
@@ -2822,10 +2830,7 @@ xfs_attr3_leaf_setflag(
 			 XFS_DA_LOGRANGE(leaf, name_rmt, sizeof(*name_rmt)));
 	}
 
-	/*
-	 * Commit the flag value change and start the next trans in series.
-	 */
-	return xfs_trans_roll_inode(&args->trans, args->dp);
+	return 0;
 }
 
 /*
@@ -2940,10 +2945,5 @@ xfs_attr3_leaf_flipflags(
 			 XFS_DA_LOGRANGE(leaf2, name_rmt, sizeof(*name_rmt)));
 	}
 
-	/*
-	 * Commit the flag value change and start the next trans in series.
-	 */
-	error = xfs_trans_roll_inode(&args->trans, args->dp);
-
-	return error;
+	return 0;
 }
diff --git a/fs/xfs/libxfs/xfs_attr_leaf.h b/fs/xfs/libxfs/xfs_attr_leaf.h
index 73615b1dd1a8..9b1c59f40a26 100644
--- a/fs/xfs/libxfs/xfs_attr_leaf.h
+++ b/fs/xfs/libxfs/xfs_attr_leaf.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000,2002-2003,2005 Silicon Graphics, Inc.
  * Copyright (c) 2013 Red Hat, Inc.
@@ -8,7 +8,6 @@
 #define	__XFS_ATTR_LEAF_H__
 
 struct attrlist;
-struct attrlist_cursor_kern;
 struct xfs_attr_list_context;
 struct xfs_da_args;
 struct xfs_da_state;
@@ -53,6 +52,9 @@ int	xfs_attr_shortform_getvalue(struct xfs_da_args *args);
 int	xfs_attr_shortform_to_leaf(struct xfs_da_args *args,
 			struct xfs_buf **leaf_bp);
 int	xfs_attr_shortform_remove(struct xfs_da_args *args);
+int	xfs_attr_sf_findname(struct xfs_da_args *args,
+			     struct xfs_attr_sf_entry **sfep,
+			     unsigned int *basep);
 int	xfs_attr_shortform_allfit(struct xfs_buf *bp, struct xfs_inode *dp);
 int	xfs_attr_shortform_bytesfit(struct xfs_inode *dp, int bytes);
 xfs_failaddr_t xfs_attr_shortform_verify(struct xfs_inode *ip);
diff --git a/fs/xfs/libxfs/xfs_attr_remote.c b/fs/xfs/libxfs/xfs_attr_remote.c
index 8b7f74b3bea2..3f80cede7406 100644
--- a/fs/xfs/libxfs/xfs_attr_remote.c
+++ b/fs/xfs/libxfs/xfs_attr_remote.c
@@ -397,7 +397,7 @@ xfs_attr_rmtval_get(
 
 	trace_xfs_attr_rmtval_get(args);
 
-	ASSERT(!(args->flags & ATTR_KERNOVAL));
+	ASSERT(args->valuelen != 0);
 	ASSERT(args->rmtvaluelen == args->valuelen);
 
 	valuelen = args->rmtvaluelen;
@@ -440,32 +440,23 @@ xfs_attr_rmtval_get(
 }
 
 /*
- * Write the value associated with an attribute into the out-of-line buffer
- * that we have defined for it.
+ * Find a "hole" in the attribute address space large enough for us to drop the
+ * new attribute's value into
  */
-int
-xfs_attr_rmtval_set(
+STATIC int
+xfs_attr_rmt_find_hole(
 	struct xfs_da_args	*args)
 {
 	struct xfs_inode	*dp = args->dp;
 	struct xfs_mount	*mp = dp->i_mount;
-	struct xfs_bmbt_irec	map;
-	xfs_dablk_t		lblkno;
-	xfs_fileoff_t		lfileoff = 0;
-	uint8_t			*src = args->value;
-	int			blkcnt;
-	int			valuelen;
-	int			nmap;
 	int			error;
-	int			offset = 0;
-
-	trace_xfs_attr_rmtval_set(args);
+	int			blkcnt;
+	xfs_fileoff_t		lfileoff = 0;
 
 	/*
-	 * Find a "hole" in the attribute address space large enough for
-	 * us to drop the new attribute's value into. Because CRC enable
-	 * attributes have headers, we can't just do a straight byte to FSB
-	 * conversion and have to take the header space into account.
+	 * Because CRC enable attributes have headers, we can't just do a
+	 * straight byte to FSB conversion and have to take the header space
+	 * into account.
 	 */
 	blkcnt = xfs_attr3_rmt_blocks(mp, args->rmtvaluelen);
 	error = xfs_bmap_first_unused(args->trans, args->dp, blkcnt, &lfileoff,
@@ -473,48 +464,26 @@ xfs_attr_rmtval_set(
 	if (error)
 		return error;
 
-	args->rmtblkno = lblkno = (xfs_dablk_t)lfileoff;
+	args->rmtblkno = (xfs_dablk_t)lfileoff;
 	args->rmtblkcnt = blkcnt;
 
-	/*
-	 * Roll through the "value", allocating blocks on disk as required.
-	 */
-	while (blkcnt > 0) {
-		/*
-		 * Allocate a single extent, up to the size of the value.
-		 *
-		 * Note that we have to consider this a data allocation as we
-		 * write the remote attribute without logging the contents.
-		 * Hence we must ensure that we aren't using blocks that are on
-		 * the busy list so that we don't overwrite blocks which have
-		 * recently been freed but their transactions are not yet
-		 * committed to disk. If we overwrite the contents of a busy
-		 * extent and then crash then the block may not contain the
-		 * correct metadata after log recovery occurs.
-		 */
-		nmap = 1;
-		error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno,
-				  blkcnt, XFS_BMAPI_ATTRFORK, args->total, &map,
-				  &nmap);
-		if (error)
-			return error;
-		error = xfs_defer_finish(&args->trans);
-		if (error)
-			return error;
-
-		ASSERT(nmap == 1);
-		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
-		       (map.br_startblock != HOLESTARTBLOCK));
-		lblkno += map.br_blockcount;
-		blkcnt -= map.br_blockcount;
+	return 0;
+}
 
-		/*
-		 * Start the next trans in the chain.
-		 */
-		error = xfs_trans_roll_inode(&args->trans, dp);
-		if (error)
-			return error;
-	}
+STATIC int
+xfs_attr_rmtval_set_value(
+	struct xfs_da_args	*args)
+{
+	struct xfs_inode	*dp = args->dp;
+	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_bmbt_irec	map;
+	xfs_dablk_t		lblkno;
+	uint8_t			*src = args->value;
+	int			blkcnt;
+	int			valuelen;
+	int			nmap;
+	int			error;
+	int			offset = 0;
 
 	/*
 	 * Roll through the "value", copying the attribute value to the
@@ -595,19 +564,82 @@ xfs_attr_rmtval_stale(
 }
 
 /*
+ * Write the value associated with an attribute into the out-of-line buffer
+ * that we have defined for it.
+ */
+int
+xfs_attr_rmtval_set(
+	struct xfs_da_args	*args)
+{
+	struct xfs_inode	*dp = args->dp;
+	struct xfs_bmbt_irec	map;
+	xfs_dablk_t		lblkno;
+	int			blkcnt;
+	int			nmap;
+	int			error;
+
+	trace_xfs_attr_rmtval_set(args);
+
+	error = xfs_attr_rmt_find_hole(args);
+	if (error)
+		return error;
+
+	blkcnt = args->rmtblkcnt;
+	lblkno = (xfs_dablk_t)args->rmtblkno;
+	/*
+	 * Roll through the "value", allocating blocks on disk as required.
+	 */
+	while (blkcnt > 0) {
+		/*
+		 * Allocate a single extent, up to the size of the value.
+		 *
+		 * Note that we have to consider this a data allocation as we
+		 * write the remote attribute without logging the contents.
+		 * Hence we must ensure that we aren't using blocks that are on
+		 * the busy list so that we don't overwrite blocks which have
+		 * recently been freed but their transactions are not yet
+		 * committed to disk. If we overwrite the contents of a busy
+		 * extent and then crash then the block may not contain the
+		 * correct metadata after log recovery occurs.
+		 */
+		nmap = 1;
+		error = xfs_bmapi_write(args->trans, dp, (xfs_fileoff_t)lblkno,
+				  blkcnt, XFS_BMAPI_ATTRFORK, args->total, &map,
+				  &nmap);
+		if (error)
+			return error;
+		error = xfs_defer_finish(&args->trans);
+		if (error)
+			return error;
+
+		ASSERT(nmap == 1);
+		ASSERT((map.br_startblock != DELAYSTARTBLOCK) &&
+		       (map.br_startblock != HOLESTARTBLOCK));
+		lblkno += map.br_blockcount;
+		blkcnt -= map.br_blockcount;
+
+		/*
+		 * Start the next trans in the chain.
+		 */
+		error = xfs_trans_roll_inode(&args->trans, dp);
+		if (error)
+			return error;
+	}
+
+	return xfs_attr_rmtval_set_value(args);
+}
+
+/*
  * Remove the value associated with an attribute by deleting the
  * out-of-line buffer that it is stored on.
  */
 int
-xfs_attr_rmtval_remove(
+xfs_attr_rmtval_invalidate(
 	struct xfs_da_args	*args)
 {
 	xfs_dablk_t		lblkno;
 	int			blkcnt;
 	int			error;
-	int			done;
-
-	trace_xfs_attr_rmtval_remove(args);
 
 	/*
 	 * Roll through the "value", invalidating the attribute value's blocks.
@@ -635,21 +667,29 @@ xfs_attr_rmtval_remove(
 		lblkno += map.br_blockcount;
 		blkcnt -= map.br_blockcount;
 	}
+	return 0;
+}
+
+/*
+ * Remove the value associated with an attribute by deleting the
+ * out-of-line buffer that it is stored on.
+ */
+int
+xfs_attr_rmtval_remove(
+	struct xfs_da_args      *args)
+{
+	int			error;
+	int			retval;
+
+	trace_xfs_attr_rmtval_remove(args);
 
 	/*
 	 * Keep de-allocating extents until the remote-value region is gone.
 	 */
-	lblkno = args->rmtblkno;
-	blkcnt = args->rmtblkcnt;
-	done = 0;
-	while (!done) {
-		error = xfs_bunmapi(args->trans, args->dp, lblkno, blkcnt,
-				    XFS_BMAPI_ATTRFORK, 1, &done);
-		if (error)
-			return error;
-		error = xfs_defer_finish(&args->trans);
-		if (error)
-			return error;
+	do {
+		retval = __xfs_attr_rmtval_remove(args);
+		if (retval && retval != -EAGAIN)
+			return retval;
 
 		/*
 		 * Close out trans and start the next one in the chain.
@@ -657,6 +697,36 @@ xfs_attr_rmtval_remove(
 		error = xfs_trans_roll_inode(&args->trans, args->dp);
 		if (error)
 			return error;
-	}
+	} while (retval == -EAGAIN);
+
 	return 0;
 }
+
+/*
+ * Remove the value associated with an attribute by deleting the out-of-line
+ * buffer that it is stored on. Returns EAGAIN for the caller to refresh the
+ * transaction and re-call the function
+ */
+int
+__xfs_attr_rmtval_remove(
+	struct xfs_da_args	*args)
+{
+	int			error, done;
+
+	/*
+	 * Unmap value blocks for this attr.
+	 */
+	error = xfs_bunmapi(args->trans, args->dp, args->rmtblkno,
+			    args->rmtblkcnt, XFS_BMAPI_ATTRFORK, 1, &done);
+	if (error)
+		return error;
+
+	error = xfs_defer_finish(&args->trans);
+	if (error)
+		return error;
+
+	if (!done)
+		return -EAGAIN;
+
+	return error;
+}
diff --git a/fs/xfs/libxfs/xfs_attr_remote.h b/fs/xfs/libxfs/xfs_attr_remote.h
index 6fb4572845ce..9eee615da156 100644
--- a/fs/xfs/libxfs/xfs_attr_remote.h
+++ b/fs/xfs/libxfs/xfs_attr_remote.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2013 Red Hat, Inc.
  * All Rights Reserved.
@@ -13,5 +13,6 @@ int xfs_attr_rmtval_set(struct xfs_da_args *args);
 int xfs_attr_rmtval_remove(struct xfs_da_args *args);
 int xfs_attr_rmtval_stale(struct xfs_inode *ip, struct xfs_bmbt_irec *map,
 		xfs_buf_flags_t incore_flags);
-
+int xfs_attr_rmtval_invalidate(struct xfs_da_args *args);
+int __xfs_attr_rmtval_remove(struct xfs_da_args *args);
 #endif /* __XFS_ATTR_REMOTE_H__ */
diff --git a/fs/xfs/libxfs/xfs_attr_sf.h b/fs/xfs/libxfs/xfs_attr_sf.h
index aafa4fe70624..bb004fb7944a 100644
--- a/fs/xfs/libxfs/xfs_attr_sf.h
+++ b/fs/xfs/libxfs/xfs_attr_sf.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000,2002,2005 Silicon Graphics, Inc.
  * All Rights Reserved.
diff --git a/fs/xfs/libxfs/xfs_bit.h b/fs/xfs/libxfs/xfs_bit.h
index 99017b8df292..a04f266ae644 100644
--- a/fs/xfs/libxfs/xfs_bit.h
+++ b/fs/xfs/libxfs/xfs_bit.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000,2002,2005 Silicon Graphics, Inc.
  * All Rights Reserved.
diff --git a/fs/xfs/libxfs/xfs_bmap.c b/fs/xfs/libxfs/xfs_bmap.c
index 9a6d7a84689a..9c40d5971035 100644
--- a/fs/xfs/libxfs/xfs_bmap.c
+++ b/fs/xfs/libxfs/xfs_bmap.c
@@ -61,10 +61,10 @@ xfs_bmap_compute_maxlevels(
 	int		sz;		/* root block size */
 
 	/*
-	 * The maximum number of extents in a file, hence the maximum
-	 * number of leaf entries, is controlled by the type of di_nextents
-	 * (a signed 32-bit number, xfs_extnum_t), or by di_anextents
-	 * (a signed 16-bit number, xfs_aextnum_t).
+	 * The maximum number of extents in a file, hence the maximum number of
+	 * leaf entries, is controlled by the size of the on-disk extent count,
+	 * either a signed 32-bit number for the data fork, or a signed 16-bit
+	 * number for the attr fork.
 	 *
 	 * Note that we can no longer assume that if we are in ATTR1 that
 	 * the fork offset of all the inodes will be
@@ -120,10 +120,11 @@ xfs_bmbt_lookup_first(
  */
 static inline bool xfs_bmap_needs_btree(struct xfs_inode *ip, int whichfork)
 {
+	struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork);
+
 	return whichfork != XFS_COW_FORK &&
-		XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
-		XFS_IFORK_NEXTENTS(ip, whichfork) >
-			XFS_IFORK_MAXEXT(ip, whichfork);
+		ifp->if_format == XFS_DINODE_FMT_EXTENTS &&
+		ifp->if_nextents > XFS_IFORK_MAXEXT(ip, whichfork);
 }
 
 /*
@@ -131,10 +132,11 @@ static inline bool xfs_bmap_needs_btree(struct xfs_inode *ip, int whichfork)
  */
 static inline bool xfs_bmap_wants_extents(struct xfs_inode *ip, int whichfork)
 {
+	struct xfs_ifork *ifp = XFS_IFORK_PTR(ip, whichfork);
+
 	return whichfork != XFS_COW_FORK &&
-		XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE &&
-		XFS_IFORK_NEXTENTS(ip, whichfork) <=
-			XFS_IFORK_MAXEXT(ip, whichfork);
+		ifp->if_format == XFS_DINODE_FMT_BTREE &&
+		ifp->if_nextents <= XFS_IFORK_MAXEXT(ip, whichfork);
 }
 
 /*
@@ -193,14 +195,12 @@ xfs_default_attroffset(
 	struct xfs_mount	*mp = ip->i_mount;
 	uint			offset;
 
-	if (mp->m_sb.sb_inodesize == 256) {
-		offset = XFS_LITINO(mp, ip->i_d.di_version) -
-				XFS_BMDR_SPACE_CALC(MINABTPTRS);
-	} else {
+	if (mp->m_sb.sb_inodesize == 256)
+		offset = XFS_LITINO(mp) - XFS_BMDR_SPACE_CALC(MINABTPTRS);
+	else
 		offset = XFS_BMDR_SPACE_CALC(6 * MINABTPTRS);
-	}
 
-	ASSERT(offset < XFS_LITINO(mp, ip->i_d.di_version));
+	ASSERT(offset < XFS_LITINO(mp));
 	return offset;
 }
 
@@ -215,8 +215,8 @@ xfs_bmap_forkoff_reset(
 	int		whichfork)
 {
 	if (whichfork == XFS_ATTR_FORK &&
-	    ip->i_d.di_format != XFS_DINODE_FMT_DEV &&
-	    ip->i_d.di_format != XFS_DINODE_FMT_BTREE) {
+	    ip->i_df.if_format != XFS_DINODE_FMT_DEV &&
+	    ip->i_df.if_format != XFS_DINODE_FMT_BTREE) {
 		uint	dfl_forkoff = xfs_default_attroffset(ip) >> 3;
 
 		if (dfl_forkoff > ip->i_d.di_forkoff)
@@ -317,31 +317,28 @@ xfs_bmap_check_leaf_extents(
 	xfs_inode_t		*ip,		/* incore inode pointer */
 	int			whichfork)	/* data or attr fork */
 {
+	struct xfs_mount	*mp = ip->i_mount;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 	struct xfs_btree_block	*block;	/* current btree block */
 	xfs_fsblock_t		bno;	/* block # of "block" */
 	xfs_buf_t		*bp;	/* buffer for "block" */
 	int			error;	/* error return value */
 	xfs_extnum_t		i=0, j;	/* index into the extents list */
-	struct xfs_ifork	*ifp;	/* fork structure */
 	int			level;	/* btree level, for checking */
-	xfs_mount_t		*mp;	/* file system mount structure */
 	__be64			*pp;	/* pointer to block address */
 	xfs_bmbt_rec_t		*ep;	/* pointer to current extent */
 	xfs_bmbt_rec_t		last = {0, 0}; /* last extent in prev block */
 	xfs_bmbt_rec_t		*nextp;	/* pointer to next extent */
 	int			bp_release = 0;
 
-	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE) {
+	if (ifp->if_format != XFS_DINODE_FMT_BTREE)
 		return;
-	}
 
 	/* skip large extent count inodes */
-	if (ip->i_d.di_nextents > 10000)
+	if (ip->i_df.if_nextents > 10000)
 		return;
 
 	bno = NULLFSBLOCK;
-	mp = ip->i_mount;
-	ifp = XFS_IFORK_PTR(ip, whichfork);
 	block = ifp->if_broot;
 	/*
 	 * Root level must use BMAP_BROOT_PTR_ADDR macro to get ptr out.
@@ -556,7 +553,8 @@ __xfs_bmap_add_free(
 #endif
 	ASSERT(xfs_bmap_free_item_zone != NULL);
 
-	new = kmem_zone_alloc(xfs_bmap_free_item_zone, 0);
+	new = kmem_cache_alloc(xfs_bmap_free_item_zone,
+			       GFP_KERNEL | __GFP_NOFAIL);
 	new->xefi_startblock = bno;
 	new->xefi_blockcount = (xfs_extlen_t)len;
 	if (oinfo)
@@ -606,7 +604,7 @@ xfs_bmap_btree_to_extents(
 	ASSERT(cur);
 	ASSERT(whichfork != XFS_COW_FORK);
 	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
+	ASSERT(ifp->if_format == XFS_DINODE_FMT_BTREE);
 	ASSERT(be16_to_cpu(rblock->bb_level) == 1);
 	ASSERT(be16_to_cpu(rblock->bb_numrecs) == 1);
 	ASSERT(xfs_bmbt_maxrecs(mp, ifp->if_broot_bytes, 0) == 1);
@@ -634,7 +632,7 @@ xfs_bmap_btree_to_extents(
 	xfs_iroot_realloc(ip, -1, whichfork);
 	ASSERT(ifp->if_broot == NULL);
 	ASSERT((ifp->if_flags & XFS_IFBROOT) == 0);
-	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
+	ifp->if_format = XFS_DINODE_FMT_EXTENTS;
 	*logflagsp |= XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
 	return 0;
 }
@@ -670,7 +668,7 @@ xfs_bmap_extents_to_btree(
 	mp = ip->i_mount;
 	ASSERT(whichfork != XFS_COW_FORK);
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS);
+	ASSERT(ifp->if_format == XFS_DINODE_FMT_EXTENTS);
 
 	/*
 	 * Make space in the inode incore. This needs to be undone if we fail
@@ -690,11 +688,11 @@ xfs_bmap_extents_to_btree(
 	 * Need a cursor.  Can't allocate until bb_level is filled in.
 	 */
 	cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-	cur->bc_private.b.flags = wasdel ? XFS_BTCUR_BPRV_WASDEL : 0;
+	cur->bc_ino.flags = wasdel ? XFS_BTCUR_BMBT_WASDEL : 0;
 	/*
 	 * Convert to a btree with two levels, one record in root.
 	 */
-	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_BTREE);
+	ifp->if_format = XFS_DINODE_FMT_BTREE;
 	memset(&args, 0, sizeof(args));
 	args.tp = tp;
 	args.mp = mp;
@@ -727,7 +725,7 @@ xfs_bmap_extents_to_btree(
 	ASSERT(tp->t_firstblock == NULLFSBLOCK ||
 	       args.agno >= XFS_FSB_TO_AGNO(mp, tp->t_firstblock));
 	tp->t_firstblock = args.fsbno;
-	cur->bc_private.b.allocated++;
+	cur->bc_ino.allocated++;
 	ip->i_d.di_nblocks++;
 	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, 1L);
 	error = xfs_trans_get_buf(tp, mp->m_ddev_targp,
@@ -752,7 +750,7 @@ xfs_bmap_extents_to_btree(
 		xfs_bmbt_disk_set_all(arp, &rec);
 		cnt++;
 	}
-	ASSERT(cnt == XFS_IFORK_NEXTENTS(ip, whichfork));
+	ASSERT(cnt == ifp->if_nextents);
 	xfs_btree_set_numrecs(ablock, cnt);
 
 	/*
@@ -780,7 +778,7 @@ out_unreserve_dquot:
 	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, -1L);
 out_root_realloc:
 	xfs_iroot_realloc(ip, -1, whichfork);
-	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
+	ifp->if_format = XFS_DINODE_FMT_EXTENTS;
 	ASSERT(ifp->if_broot == NULL);
 	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
 
@@ -802,16 +800,16 @@ xfs_bmap_local_to_extents_empty(
 	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 
 	ASSERT(whichfork != XFS_COW_FORK);
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
+	ASSERT(ifp->if_format == XFS_DINODE_FMT_LOCAL);
 	ASSERT(ifp->if_bytes == 0);
-	ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) == 0);
+	ASSERT(ifp->if_nextents == 0);
 
 	xfs_bmap_forkoff_reset(ip, whichfork);
 	ifp->if_flags &= ~XFS_IFINLINE;
 	ifp->if_flags |= XFS_IFEXTENTS;
 	ifp->if_u1.if_root = NULL;
 	ifp->if_height = 0;
-	XFS_IFORK_FMT_SET(ip, whichfork, XFS_DINODE_FMT_EXTENTS);
+	ifp->if_format = XFS_DINODE_FMT_EXTENTS;
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 }
 
@@ -842,7 +840,7 @@ xfs_bmap_local_to_extents(
 	 */
 	ASSERT(!(S_ISREG(VFS_I(ip)->i_mode) && whichfork == XFS_DATA_FORK));
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
+	ASSERT(ifp->if_format == XFS_DINODE_FMT_LOCAL);
 
 	if (!ifp->if_bytes) {
 		xfs_bmap_local_to_extents_empty(tp, ip, whichfork);
@@ -909,7 +907,7 @@ xfs_bmap_local_to_extents(
 	xfs_iext_first(ifp, &icur);
 	xfs_iext_insert(ip, &icur, &rec, 0);
 
-	XFS_IFORK_NEXT_SET(ip, whichfork, 1);
+	ifp->if_nextents = 1;
 	ip->i_d.di_nblocks = 1;
 	xfs_trans_mod_dquot_byino(tp, ip,
 		XFS_TRANS_DQ_BCOUNT, 1L);
@@ -953,7 +951,7 @@ xfs_bmap_add_attrfork_btree(
 			xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 			return -ENOSPC;
 		}
-		cur->bc_private.b.allocated = 0;
+		cur->bc_ino.allocated = 0;
 		xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 	}
 	return 0;
@@ -974,13 +972,14 @@ xfs_bmap_add_attrfork_extents(
 	xfs_btree_cur_t		*cur;		/* bmap btree cursor */
 	int			error;		/* error return value */
 
-	if (ip->i_d.di_nextents * sizeof(xfs_bmbt_rec_t) <= XFS_IFORK_DSIZE(ip))
+	if (ip->i_df.if_nextents * sizeof(struct xfs_bmbt_rec) <=
+	    XFS_IFORK_DSIZE(ip))
 		return 0;
 	cur = NULL;
 	error = xfs_bmap_extents_to_btree(tp, ip, &cur, 0, flags,
 					  XFS_DATA_FORK);
 	if (cur) {
-		cur->bc_private.b.allocated = 0;
+		cur->bc_ino.allocated = 0;
 		xfs_btree_del_cursor(cur, error);
 	}
 	return error;
@@ -1035,7 +1034,7 @@ xfs_bmap_set_attrforkoff(
 	int			size,
 	int			*version)
 {
-	switch (ip->i_d.di_format) {
+	switch (ip->i_df.if_format) {
 	case XFS_DINODE_FMT_DEV:
 		ip->i_d.di_forkoff = roundup(sizeof(xfs_dev_t), 8) >> 3;
 		break;
@@ -1093,17 +1092,6 @@ xfs_bmap_add_attrfork(
 		goto trans_cancel;
 	if (XFS_IFORK_Q(ip))
 		goto trans_cancel;
-	if (XFS_IS_CORRUPT(mp, ip->i_d.di_anextents != 0)) {
-		error = -EFSCORRUPTED;
-		goto trans_cancel;
-	}
-	if (ip->i_d.di_aformat != XFS_DINODE_FMT_EXTENTS) {
-		/*
-		 * For inodes coming from pre-6.2 filesystems.
-		 */
-		ASSERT(ip->i_d.di_aformat == 0);
-		ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
-	}
 
 	xfs_trans_ijoin(tp, ip, 0);
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
@@ -1111,10 +1099,14 @@ xfs_bmap_add_attrfork(
 	if (error)
 		goto trans_cancel;
 	ASSERT(ip->i_afp == NULL);
-	ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, 0);
+
+	ip->i_afp = kmem_cache_zalloc(xfs_ifork_zone,
+				      GFP_KERNEL | __GFP_NOFAIL);
+
+	ip->i_afp->if_format = XFS_DINODE_FMT_EXTENTS;
 	ip->i_afp->if_flags = XFS_IFEXTENTS;
 	logflags = 0;
-	switch (ip->i_d.di_format) {
+	switch (ip->i_df.if_format) {
 	case XFS_DINODE_FMT_LOCAL:
 		error = xfs_bmap_add_attrfork_local(tp, ip, &logflags);
 		break;
@@ -1178,20 +1170,20 @@ xfs_iread_bmbt_block(
 {
 	struct xfs_iread_state	*ir = priv;
 	struct xfs_mount	*mp = cur->bc_mp;
-	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	struct xfs_inode	*ip = cur->bc_ino.ip;
 	struct xfs_btree_block	*block;
 	struct xfs_buf		*bp;
 	struct xfs_bmbt_rec	*frp;
 	xfs_extnum_t		num_recs;
 	xfs_extnum_t		j;
-	int			whichfork = cur->bc_private.b.whichfork;
+	int			whichfork = cur->bc_ino.whichfork;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 
 	block = xfs_btree_get_block(cur, level, &bp);
 
 	/* Abort if we find more records than nextents. */
 	num_recs = xfs_btree_get_numrecs(block);
-	if (unlikely(ir->loaded + num_recs >
-		     XFS_IFORK_NEXTENTS(ip, whichfork))) {
+	if (unlikely(ir->loaded + num_recs > ifp->if_nextents)) {
 		xfs_warn(ip->i_mount, "corrupt dinode %llu, (btree extents).",
 				(unsigned long long)ip->i_ino);
 		xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__, block,
@@ -1217,7 +1209,7 @@ xfs_iread_bmbt_block(
 				xfs_bmap_fork_to_state(whichfork));
 		trace_xfs_read_extent(ip, &ir->icur,
 				xfs_bmap_fork_to_state(whichfork), _THIS_IP_);
-		xfs_iext_next(XFS_IFORK_PTR(ip, whichfork), &ir->icur);
+		xfs_iext_next(ifp, &ir->icur);
 	}
 
 	return 0;
@@ -1240,9 +1232,7 @@ xfs_iread_extents(
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 
-	if (XFS_IS_CORRUPT(mp,
-			   XFS_IFORK_FORMAT(ip, whichfork) !=
-			   XFS_DINODE_FMT_BTREE)) {
+	if (XFS_IS_CORRUPT(mp, ifp->if_format != XFS_DINODE_FMT_BTREE)) {
 		error = -EFSCORRUPTED;
 		goto out;
 	}
@@ -1256,8 +1246,7 @@ xfs_iread_extents(
 	if (error)
 		goto out;
 
-	if (XFS_IS_CORRUPT(mp,
-			   ir.loaded != XFS_IFORK_NEXTENTS(ip, whichfork))) {
+	if (XFS_IS_CORRUPT(mp, ir.loaded != ifp->if_nextents)) {
 		error = -EFSCORRUPTED;
 		goto out;
 	}
@@ -1291,14 +1280,13 @@ xfs_bmap_first_unused(
 	xfs_fileoff_t		lowest, max;
 	int			error;
 
-	ASSERT(xfs_ifork_has_extents(ip, whichfork) ||
-	       XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL);
-
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
+	if (ifp->if_format == XFS_DINODE_FMT_LOCAL) {
 		*first_unused = 0;
 		return 0;
 	}
 
+	ASSERT(xfs_ifork_has_extents(ifp));
+
 	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
 		error = xfs_iread_extents(tp, ip, whichfork);
 		if (error)
@@ -1339,7 +1327,7 @@ xfs_bmap_last_before(
 	struct xfs_iext_cursor	icur;
 	int			error;
 
-	switch (XFS_IFORK_FORMAT(ip, whichfork)) {
+	switch (ifp->if_format) {
 	case XFS_DINODE_FMT_LOCAL:
 		*last_block = 0;
 		return 0;
@@ -1438,16 +1426,17 @@ xfs_bmap_last_offset(
 	xfs_fileoff_t		*last_block,
 	int			whichfork)
 {
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 	struct xfs_bmbt_irec	rec;
 	int			is_empty;
 	int			error;
 
 	*last_block = 0;
 
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL)
+	if (ifp->if_format == XFS_DINODE_FMT_LOCAL)
 		return 0;
 
-	if (XFS_IS_CORRUPT(ip->i_mount, !xfs_ifork_has_extents(ip, whichfork)))
+	if (XFS_IS_CORRUPT(ip->i_mount, !xfs_ifork_has_extents(ifp)))
 		return -EFSCORRUPTED;
 
 	error = xfs_bmap_last_extent(NULL, ip, whichfork, &rec, &is_empty);
@@ -1465,23 +1454,22 @@ xfs_bmap_last_offset(
  */
 int					/* 1=>1 block, 0=>otherwise */
 xfs_bmap_one_block(
-	xfs_inode_t	*ip,		/* incore inode */
-	int		whichfork)	/* data or attr fork */
+	struct xfs_inode	*ip,		/* incore inode */
+	int			whichfork)	/* data or attr fork */
 {
-	struct xfs_ifork *ifp;		/* inode fork pointer */
-	int		rval;		/* return value */
-	xfs_bmbt_irec_t	s;		/* internal version of extent */
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
+	int			rval;		/* return value */
+	struct xfs_bmbt_irec	s;		/* internal version of extent */
 	struct xfs_iext_cursor icur;
 
 #ifndef DEBUG
 	if (whichfork == XFS_DATA_FORK)
 		return XFS_ISIZE(ip) == ip->i_mount->m_sb.sb_blocksize;
 #endif	/* !DEBUG */
-	if (XFS_IFORK_NEXTENTS(ip, whichfork) != 1)
+	if (ifp->if_nextents != 1)
 		return 0;
-	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
+	if (ifp->if_format != XFS_DINODE_FMT_EXTENTS)
 		return 0;
-	ifp = XFS_IFORK_PTR(ip, whichfork);
 	ASSERT(ifp->if_flags & XFS_IFEXTENTS);
 	xfs_iext_first(ifp, &icur);
 	xfs_iext_get_extent(ifp, &icur, &s);
@@ -1503,10 +1491,11 @@ xfs_bmap_add_extent_delay_real(
 	struct xfs_bmalloca	*bma,
 	int			whichfork)
 {
+	struct xfs_mount	*mp = bma->ip->i_mount;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(bma->ip, whichfork);
 	struct xfs_bmbt_irec	*new = &bma->got;
 	int			error;	/* error return value */
 	int			i;	/* temp state */
-	struct xfs_ifork	*ifp;	/* inode fork pointer */
 	xfs_fileoff_t		new_endoff;	/* end offset of new entry */
 	xfs_bmbt_irec_t		r[3];	/* neighbor extent entries */
 					/* left is 0, right is 1, prev is 2 */
@@ -1516,19 +1505,12 @@ xfs_bmap_add_extent_delay_real(
 	xfs_filblks_t		da_old; /* old count del alloc blocks used */
 	xfs_filblks_t		temp=0;	/* value for da_new calculations */
 	int			tmp_rval;	/* partial logging flags */
-	struct xfs_mount	*mp;
-	xfs_extnum_t		*nextents;
 	struct xfs_bmbt_irec	old;
 
-	mp = bma->ip->i_mount;
-	ifp = XFS_IFORK_PTR(bma->ip, whichfork);
 	ASSERT(whichfork != XFS_ATTR_FORK);
-	nextents = (whichfork == XFS_COW_FORK ? &bma->ip->i_cnextents :
-						&bma->ip->i_d.di_nextents);
-
 	ASSERT(!isnullstartblock(new->br_startblock));
 	ASSERT(!bma->cur ||
-	       (bma->cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL));
+	       (bma->cur->bc_ino.flags & XFS_BTCUR_BMBT_WASDEL));
 
 	XFS_STATS_INC(mp, xs_add_exlist);
 
@@ -1616,7 +1598,7 @@ xfs_bmap_add_extent_delay_real(
 		xfs_iext_remove(bma->ip, &bma->icur, state);
 		xfs_iext_prev(ifp, &bma->icur);
 		xfs_iext_update_extent(bma->ip, state, &bma->icur, &LEFT);
-		(*nextents)--;
+		ifp->if_nextents--;
 
 		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -1720,8 +1702,8 @@ xfs_bmap_add_extent_delay_real(
 		PREV.br_startblock = new->br_startblock;
 		PREV.br_state = new->br_state;
 		xfs_iext_update_extent(bma->ip, state, &bma->icur, &PREV);
+		ifp->if_nextents++;
 
-		(*nextents)++;
 		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -1786,7 +1768,8 @@ xfs_bmap_add_extent_delay_real(
 		 * The left neighbor is not contiguous.
 		 */
 		xfs_iext_update_extent(bma->ip, state, &bma->icur, new);
-		(*nextents)++;
+		ifp->if_nextents++;
+
 		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -1818,7 +1801,7 @@ xfs_bmap_add_extent_delay_real(
 		temp = PREV.br_blockcount - new->br_blockcount;
 		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
 			startblockval(PREV.br_startblock) -
-			(bma->cur ? bma->cur->bc_private.b.allocated : 0));
+			(bma->cur ? bma->cur->bc_ino.allocated : 0));
 
 		PREV.br_startoff = new_endoff;
 		PREV.br_blockcount = temp;
@@ -1872,7 +1855,8 @@ xfs_bmap_add_extent_delay_real(
 		 * The right neighbor is not contiguous.
 		 */
 		xfs_iext_update_extent(bma->ip, state, &bma->icur, new);
-		(*nextents)++;
+		ifp->if_nextents++;
+
 		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -1904,7 +1888,7 @@ xfs_bmap_add_extent_delay_real(
 		temp = PREV.br_blockcount - new->br_blockcount;
 		da_new = XFS_FILBLKS_MIN(xfs_bmap_worst_indlen(bma->ip, temp),
 			startblockval(PREV.br_startblock) -
-			(bma->cur ? bma->cur->bc_private.b.allocated : 0));
+			(bma->cur ? bma->cur->bc_ino.allocated : 0));
 
 		PREV.br_startblock = nullstartblock(da_new);
 		PREV.br_blockcount = temp;
@@ -1957,7 +1941,7 @@ xfs_bmap_add_extent_delay_real(
 		xfs_iext_next(ifp, &bma->icur);
 		xfs_iext_insert(bma->ip, &bma->icur, &RIGHT, state);
 		xfs_iext_insert(bma->ip, &bma->icur, &LEFT, state);
-		(*nextents)++;
+		ifp->if_nextents++;
 
 		if (bma->cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
@@ -2025,8 +2009,8 @@ xfs_bmap_add_extent_delay_real(
 		xfs_mod_delalloc(mp, (int64_t)da_new - da_old);
 
 	if (bma->cur) {
-		da_new += bma->cur->bc_private.b.allocated;
-		bma->cur->bc_private.b.allocated = 0;
+		da_new += bma->cur->bc_ino.allocated;
+		bma->cur->bc_ino.allocated = 0;
 	}
 
 	/* adjust for changes in reserved delayed indirect blocks */
@@ -2161,8 +2145,7 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_iext_remove(ip, icur, state);
 		xfs_iext_prev(ifp, icur);
 		xfs_iext_update_extent(ip, state, icur, &LEFT);
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-				XFS_IFORK_NEXTENTS(ip, whichfork) - 2);
+		ifp->if_nextents -= 2;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2214,8 +2197,7 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_iext_remove(ip, icur, state);
 		xfs_iext_prev(ifp, icur);
 		xfs_iext_update_extent(ip, state, icur, &LEFT);
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-				XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
+		ifp->if_nextents--;
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2257,9 +2239,8 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_iext_remove(ip, icur, state);
 		xfs_iext_prev(ifp, icur);
 		xfs_iext_update_extent(ip, state, icur, &PREV);
+		ifp->if_nextents--;
 
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-				XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2366,8 +2347,8 @@ xfs_bmap_add_extent_unwritten_real(
 
 		xfs_iext_update_extent(ip, state, icur, &PREV);
 		xfs_iext_insert(ip, icur, new, state);
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-				XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
+		ifp->if_nextents++;
+
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2442,9 +2423,8 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_iext_update_extent(ip, state, icur, &PREV);
 		xfs_iext_next(ifp, icur);
 		xfs_iext_insert(ip, icur, new, state);
+		ifp->if_nextents++;
 
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-				XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2495,9 +2475,8 @@ xfs_bmap_add_extent_unwritten_real(
 		xfs_iext_next(ifp, icur);
 		xfs_iext_insert(ip, icur, &r[1], state);
 		xfs_iext_insert(ip, icur, &r[0], state);
+		ifp->if_nextents += 2;
 
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-				XFS_IFORK_NEXTENTS(ip, whichfork) + 2);
 		if (cur == NULL)
 			rval = XFS_ILOG_CORE | XFS_ILOG_DEXT;
 		else {
@@ -2573,7 +2552,7 @@ xfs_bmap_add_extent_unwritten_real(
 
 	/* clear out the allocated field, done with it now in any case. */
 	if (cur) {
-		cur->bc_private.b.allocated = 0;
+		cur->bc_ino.allocated = 0;
 		*curp = cur;
 	}
 
@@ -2752,7 +2731,7 @@ xfs_bmap_add_extent_hole_real(
 	struct xfs_bmbt_irec	old;
 
 	ASSERT(!isnullstartblock(new->br_startblock));
-	ASSERT(!cur || !(cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL));
+	ASSERT(!cur || !(cur->bc_ino.flags & XFS_BTCUR_BMBT_WASDEL));
 
 	XFS_STATS_INC(mp, xs_add_exlist);
 
@@ -2812,9 +2791,8 @@ xfs_bmap_add_extent_hole_real(
 		xfs_iext_remove(ip, icur, state);
 		xfs_iext_prev(ifp, icur);
 		xfs_iext_update_extent(ip, state, icur, &left);
+		ifp->if_nextents--;
 
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-			XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
 		if (cur == NULL) {
 			rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
 		} else {
@@ -2912,8 +2890,8 @@ xfs_bmap_add_extent_hole_real(
 		 * Insert a new entry.
 		 */
 		xfs_iext_insert(ip, icur, new, state);
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-			XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
+		ifp->if_nextents++;
+
 		if (cur == NULL) {
 			rval = XFS_ILOG_CORE | xfs_ilog_fext(whichfork);
 		} else {
@@ -2955,7 +2933,7 @@ xfs_bmap_add_extent_hole_real(
 
 	/* clear out the allocated field, done with it now in any case. */
 	if (cur)
-		cur->bc_private.b.allocated = 0;
+		cur->bc_ino.allocated = 0;
 
 	xfs_bmap_check_leaf_extents(cur, ip, whichfork);
 done:
@@ -3893,7 +3871,8 @@ xfs_bmapi_read(
 	int			flags)
 {
 	struct xfs_mount	*mp = ip->i_mount;
-	struct xfs_ifork	*ifp;
+	int			whichfork = xfs_bmapi_whichfork(flags);
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 	struct xfs_bmbt_irec	got;
 	xfs_fileoff_t		obno;
 	xfs_fileoff_t		end;
@@ -3901,48 +3880,23 @@ xfs_bmapi_read(
 	int			error;
 	bool			eof = false;
 	int			n = 0;
-	int			whichfork = xfs_bmapi_whichfork(flags);
 
 	ASSERT(*nmap >= 1);
-	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK|XFS_BMAPI_ENTIRE|
-			   XFS_BMAPI_COWFORK)));
+	ASSERT(!(flags & ~(XFS_BMAPI_ATTRFORK | XFS_BMAPI_ENTIRE)));
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_SHARED|XFS_ILOCK_EXCL));
 
-	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ip, whichfork)) ||
-	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
+	if (WARN_ON_ONCE(!ifp))
+		return -EFSCORRUPTED;
+
+	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
+	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT))
 		return -EFSCORRUPTED;
-	}
 
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
 
 	XFS_STATS_INC(mp, xs_blk_mapr);
 
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (!ifp) {
-		/* No CoW fork?  Return a hole. */
-		if (whichfork == XFS_COW_FORK) {
-			mval->br_startoff = bno;
-			mval->br_startblock = HOLESTARTBLOCK;
-			mval->br_blockcount = len;
-			mval->br_state = XFS_EXT_NORM;
-			*nmap = 1;
-			return 0;
-		}
-
-		/*
-		 * A missing attr ifork implies that the inode says we're in
-		 * extents or btree format but failed to pass the inode fork
-		 * verifier while trying to load it.  Treat that as a file
-		 * corruption too.
-		 */
-#ifdef DEBUG
-		xfs_alert(mp, "%s: inode %llu missing fork %d",
-				__func__, ip->i_ino, whichfork);
-#endif /* DEBUG */
-		return -EFSCORRUPTED;
-	}
-
 	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
 		error = xfs_iread_extents(NULL, ip, whichfork);
 		if (error)
@@ -4187,25 +4141,15 @@ xfs_bmapi_allocate(
 	bma->nallocs++;
 
 	if (bma->cur)
-		bma->cur->bc_private.b.flags =
-			bma->wasdel ? XFS_BTCUR_BPRV_WASDEL : 0;
+		bma->cur->bc_ino.flags =
+			bma->wasdel ? XFS_BTCUR_BMBT_WASDEL : 0;
 
 	bma->got.br_startoff = bma->offset;
 	bma->got.br_startblock = bma->blkno;
 	bma->got.br_blockcount = bma->length;
 	bma->got.br_state = XFS_EXT_NORM;
 
-	/*
-	 * In the data fork, a wasdelay extent has been initialized, so
-	 * shouldn't be flagged as unwritten.
-	 *
-	 * For the cow fork, however, we convert delalloc reservations
-	 * (extents allocated for speculative preallocation) to
-	 * allocated unwritten extents, and only convert the unwritten
-	 * extents to real extents when we're about to write the data.
-	 */
-	if ((!bma->wasdel || (bma->flags & XFS_BMAPI_COWFORK)) &&
-	    (bma->flags & XFS_BMAPI_PREALLOC))
+	if (bma->flags & XFS_BMAPI_PREALLOC)
 		bma->got.br_state = XFS_EXT_UNWRITTEN;
 
 	if (bma->wasdel)
@@ -4319,11 +4263,13 @@ xfs_bmapi_minleft(
 	struct xfs_inode	*ip,
 	int			fork)
 {
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, fork);
+
 	if (tp && tp->t_firstblock != NULLFSBLOCK)
 		return 0;
-	if (XFS_IFORK_FORMAT(ip, fork) != XFS_DINODE_FMT_BTREE)
+	if (ifp->if_format != XFS_DINODE_FMT_BTREE)
 		return 1;
-	return be16_to_cpu(XFS_IFORK_PTR(ip, fork)->if_broot->bb_level) + 1;
+	return be16_to_cpu(ifp->if_broot->bb_level) + 1;
 }
 
 /*
@@ -4338,11 +4284,13 @@ xfs_bmapi_finish(
 	int			whichfork,
 	int			error)
 {
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(bma->ip, whichfork);
+
 	if ((bma->logflags & xfs_ilog_fext(whichfork)) &&
-	    XFS_IFORK_FORMAT(bma->ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
+	    ifp->if_format != XFS_DINODE_FMT_EXTENTS)
 		bma->logflags &= ~xfs_ilog_fext(whichfork);
 	else if ((bma->logflags & xfs_ilog_fbroot(whichfork)) &&
-		 XFS_IFORK_FORMAT(bma->ip, whichfork) != XFS_DINODE_FMT_BTREE)
+		 ifp->if_format != XFS_DINODE_FMT_BTREE)
 		bma->logflags &= ~xfs_ilog_fbroot(whichfork);
 
 	if (bma->logflags)
@@ -4374,13 +4322,13 @@ xfs_bmapi_write(
 		.total		= total,
 	};
 	struct xfs_mount	*mp = ip->i_mount;
-	struct xfs_ifork	*ifp;
+	int			whichfork = xfs_bmapi_whichfork(flags);
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 	xfs_fileoff_t		end;		/* end of mapped file region */
 	bool			eof = false;	/* after the end of extents */
 	int			error;		/* error return */
 	int			n;		/* current extent index */
 	xfs_fileoff_t		obno;		/* old block number (offset) */
-	int			whichfork;	/* data or attr fork */
 
 #ifdef DEBUG
 	xfs_fileoff_t		orig_bno;	/* original block number value */
@@ -4395,13 +4343,12 @@ xfs_bmapi_write(
 	orig_mval = mval;
 	orig_nmap = *nmap;
 #endif
-	whichfork = xfs_bmapi_whichfork(flags);
 
 	ASSERT(*nmap >= 1);
 	ASSERT(*nmap <= XFS_BMAP_MAX_NMAP);
 	ASSERT(tp != NULL);
 	ASSERT(len > 0);
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL);
+	ASSERT(ifp->if_format != XFS_DINODE_FMT_LOCAL);
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 	ASSERT(!(flags & XFS_BMAPI_REMAP));
 
@@ -4417,7 +4364,7 @@ xfs_bmapi_write(
 	ASSERT((flags & (XFS_BMAPI_PREALLOC | XFS_BMAPI_ZERO)) !=
 			(XFS_BMAPI_PREALLOC | XFS_BMAPI_ZERO));
 
-	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ip, whichfork)) ||
+	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
 		return -EFSCORRUPTED;
 	}
@@ -4425,8 +4372,6 @@ xfs_bmapi_write(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
 
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-
 	XFS_STATS_INC(mp, xs_blk_mapw);
 
 	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
@@ -4536,9 +4481,8 @@ xfs_bmapi_write(
 	if (error)
 		goto error0;
 
-	ASSERT(XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE ||
-	       XFS_IFORK_NEXTENTS(ip, whichfork) >
-		XFS_IFORK_MAXEXT(ip, whichfork));
+	ASSERT(ifp->if_format != XFS_DINODE_FMT_BTREE ||
+	       ifp->if_nextents > XFS_IFORK_MAXEXT(ip, whichfork));
 	xfs_bmapi_finish(&bma, whichfork, 0);
 	xfs_bmap_validate_ret(orig_bno, orig_len, orig_flags, orig_mval,
 		orig_nmap, *nmap);
@@ -4613,8 +4557,23 @@ xfs_bmapi_convert_delalloc(
 	bma.offset = bma.got.br_startoff;
 	bma.length = max_t(xfs_filblks_t, bma.got.br_blockcount, MAXEXTLEN);
 	bma.minleft = xfs_bmapi_minleft(tp, ip, whichfork);
+
+	/*
+	 * When we're converting the delalloc reservations backing dirty pages
+	 * in the page cache, we must be careful about how we create the new
+	 * extents:
+	 *
+	 * New CoW fork extents are created unwritten, turned into real extents
+	 * when we're about to write the data to disk, and mapped into the data
+	 * fork after the write finishes.  End of story.
+	 *
+	 * New data fork extents must be mapped in as unwritten and converted
+	 * to real extents after the write succeeds to avoid exposing stale
+	 * disk contents if we crash.
+	 */
+	bma.flags = XFS_BMAPI_PREALLOC;
 	if (whichfork == XFS_COW_FORK)
-		bma.flags = XFS_BMAPI_COWFORK | XFS_BMAPI_PREALLOC;
+		bma.flags |= XFS_BMAPI_COWFORK;
 
 	if (!xfs_iext_peek_prev_extent(ifp, &bma.icur, &bma.prev))
 		bma.prev.br_startoff = NULLFILEOFF;
@@ -4684,7 +4643,7 @@ xfs_bmapi_remap(
 	ASSERT((flags & (XFS_BMAPI_ATTRFORK | XFS_BMAPI_PREALLOC)) !=
 			(XFS_BMAPI_ATTRFORK | XFS_BMAPI_PREALLOC));
 
-	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ip, whichfork)) ||
+	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
 		return -EFSCORRUPTED;
 	}
@@ -4709,7 +4668,7 @@ xfs_bmapi_remap(
 
 	if (ifp->if_flags & XFS_IFBROOT) {
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_private.b.flags = 0;
+		cur->bc_ino.flags = 0;
 	}
 
 	got.br_startoff = bno;
@@ -4728,9 +4687,9 @@ xfs_bmapi_remap(
 	error = xfs_bmap_btree_to_extents(tp, ip, cur, &logflags, whichfork);
 
 error0:
-	if (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS)
+	if (ip->i_df.if_format != XFS_DINODE_FMT_EXTENTS)
 		logflags &= ~XFS_ILOG_DEXT;
-	else if (ip->i_d.di_format != XFS_DINODE_FMT_BTREE)
+	else if (ip->i_df.if_format != XFS_DINODE_FMT_BTREE)
 		logflags &= ~XFS_ILOG_DBROOT;
 
 	if (logflags)
@@ -5080,9 +5039,8 @@ xfs_bmap_del_extent_real(
 	 * conversion to btree format, since the transaction will be dirty then.
 	 */
 	if (tp->t_blk_res == 0 &&
-	    XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_EXTENTS &&
-	    XFS_IFORK_NEXTENTS(ip, whichfork) >=
-			XFS_IFORK_MAXEXT(ip, whichfork) &&
+	    ifp->if_format == XFS_DINODE_FMT_EXTENTS &&
+	    ifp->if_nextents >= XFS_IFORK_MAXEXT(ip, whichfork) &&
 	    del->br_startoff > got.br_startoff && del_endoff < got_endoff)
 		return -ENOSPC;
 
@@ -5134,8 +5092,8 @@ xfs_bmap_del_extent_real(
 		 */
 		xfs_iext_remove(ip, icur, state);
 		xfs_iext_prev(ifp, icur);
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-			XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
+		ifp->if_nextents--;
+
 		flags |= XFS_ILOG_CORE;
 		if (!cur) {
 			flags |= xfs_ilog_fext(whichfork);
@@ -5243,8 +5201,8 @@ xfs_bmap_del_extent_real(
 			}
 		} else
 			flags |= xfs_ilog_fext(whichfork);
-		XFS_IFORK_NEXT_SET(ip, whichfork,
-			XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
+
+		ifp->if_nextents++;
 		xfs_iext_next(ifp, icur);
 		xfs_iext_insert(ip, icur, &new, state);
 		break;
@@ -5324,7 +5282,7 @@ __xfs_bunmapi(
 	whichfork = xfs_bmapi_whichfork(flags);
 	ASSERT(whichfork != XFS_COW_FORK);
 	ifp = XFS_IFORK_PTR(ip, whichfork);
-	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ip, whichfork)))
+	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)))
 		return -EFSCORRUPTED;
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
@@ -5362,9 +5320,9 @@ __xfs_bunmapi(
 
 	logflags = 0;
 	if (ifp->if_flags & XFS_IFBROOT) {
-		ASSERT(XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_BTREE);
+		ASSERT(ifp->if_format == XFS_DINODE_FMT_BTREE);
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_private.b.flags = 0;
+		cur->bc_ino.flags = 0;
 	} else
 		cur = NULL;
 
@@ -5607,10 +5565,10 @@ error0:
 	 * logging the extent records if we've converted to btree format.
 	 */
 	if ((logflags & xfs_ilog_fext(whichfork)) &&
-	    XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_EXTENTS)
+	    ifp->if_format != XFS_DINODE_FMT_EXTENTS)
 		logflags &= ~xfs_ilog_fext(whichfork);
 	else if ((logflags & xfs_ilog_fbroot(whichfork)) &&
-		 XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_BTREE)
+		 ifp->if_format != XFS_DINODE_FMT_BTREE)
 		logflags &= ~xfs_ilog_fbroot(whichfork);
 	/*
 	 * Log inode even in the error case, if the transaction
@@ -5620,7 +5578,7 @@ error0:
 		xfs_trans_log_inode(tp, ip, logflags);
 	if (cur) {
 		if (!error)
-			cur->bc_private.b.allocated = 0;
+			cur->bc_ino.allocated = 0;
 		xfs_btree_del_cursor(cur, error);
 	}
 	return error;
@@ -5692,6 +5650,7 @@ xfs_bmse_merge(
 	struct xfs_btree_cur		*cur,
 	int				*logflags)	/* output */
 {
+	struct xfs_ifork		*ifp = XFS_IFORK_PTR(ip, whichfork);
 	struct xfs_bmbt_irec		new;
 	xfs_filblks_t			blockcount;
 	int				error, i;
@@ -5710,8 +5669,7 @@ xfs_bmse_merge(
 	 * Update the on-disk extent count, the btree if necessary and log the
 	 * inode.
 	 */
-	XFS_IFORK_NEXT_SET(ip, whichfork,
-			   XFS_IFORK_NEXTENTS(ip, whichfork) - 1);
+	ifp->if_nextents--;
 	*logflags |= XFS_ILOG_CORE;
 	if (!cur) {
 		*logflags |= XFS_ILOG_DEXT;
@@ -5749,7 +5707,7 @@ xfs_bmse_merge(
 
 done:
 	xfs_iext_remove(ip, icur, 0);
-	xfs_iext_prev(XFS_IFORK_PTR(ip, whichfork), icur);
+	xfs_iext_prev(ifp, icur);
 	xfs_iext_update_extent(ip, xfs_bmap_fork_to_state(whichfork), icur,
 			&new);
 
@@ -5821,7 +5779,7 @@ xfs_bmap_collapse_extents(
 	int			error = 0;
 	int			logflags = 0;
 
-	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ip, whichfork)) ||
+	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
 		return -EFSCORRUPTED;
 	}
@@ -5839,7 +5797,7 @@ xfs_bmap_collapse_extents(
 
 	if (ifp->if_flags & XFS_IFBROOT) {
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_private.b.flags = 0;
+		cur->bc_ino.flags = 0;
 	}
 
 	if (!xfs_iext_lookup_extent(ip, ifp, *next_fsb, &icur, &got)) {
@@ -5938,7 +5896,7 @@ xfs_bmap_insert_extents(
 	int			error = 0;
 	int			logflags = 0;
 
-	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ip, whichfork)) ||
+	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
 		return -EFSCORRUPTED;
 	}
@@ -5956,7 +5914,7 @@ xfs_bmap_insert_extents(
 
 	if (ifp->if_flags & XFS_IFBROOT) {
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_private.b.flags = 0;
+		cur->bc_ino.flags = 0;
 	}
 
 	if (*next_fsb == NULLFSBLOCK) {
@@ -6025,25 +5983,25 @@ del_cursor:
  * @split_fsb is a block where the extents is split.  If split_fsb lies in a
  * hole or the first block of extents, just return 0.
  */
-STATIC int
-xfs_bmap_split_extent_at(
+int
+xfs_bmap_split_extent(
 	struct xfs_trans	*tp,
 	struct xfs_inode	*ip,
 	xfs_fileoff_t		split_fsb)
 {
 	int				whichfork = XFS_DATA_FORK;
+	struct xfs_ifork		*ifp = XFS_IFORK_PTR(ip, whichfork);
 	struct xfs_btree_cur		*cur = NULL;
 	struct xfs_bmbt_irec		got;
 	struct xfs_bmbt_irec		new; /* split extent */
 	struct xfs_mount		*mp = ip->i_mount;
-	struct xfs_ifork		*ifp;
 	xfs_fsblock_t			gotblkcnt; /* new block count for got */
 	struct xfs_iext_cursor		icur;
 	int				error = 0;
 	int				logflags = 0;
 	int				i = 0;
 
-	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ip, whichfork)) ||
+	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ifp)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
 		return -EFSCORRUPTED;
 	}
@@ -6051,7 +6009,6 @@ xfs_bmap_split_extent_at(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		return -EIO;
 
-	ifp = XFS_IFORK_PTR(ip, whichfork);
 	if (!(ifp->if_flags & XFS_IFEXTENTS)) {
 		/* Read in all the extents */
 		error = xfs_iread_extents(tp, ip, whichfork);
@@ -6074,7 +6031,7 @@ xfs_bmap_split_extent_at(
 
 	if (ifp->if_flags & XFS_IFBROOT) {
 		cur = xfs_bmbt_init_cursor(mp, tp, ip, whichfork);
-		cur->bc_private.b.flags = 0;
+		cur->bc_ino.flags = 0;
 		error = xfs_bmbt_lookup_eq(cur, &got, &i);
 		if (error)
 			goto del_cursor;
@@ -6099,8 +6056,7 @@ xfs_bmap_split_extent_at(
 	/* Add new extent */
 	xfs_iext_next(ifp, &icur);
 	xfs_iext_insert(ip, &icur, &new, 0);
-	XFS_IFORK_NEXT_SET(ip, whichfork,
-			   XFS_IFORK_NEXTENTS(ip, whichfork) + 1);
+	ifp->if_nextents++;
 
 	if (cur) {
 		error = xfs_bmbt_lookup_eq(cur, &new, &i);
@@ -6133,7 +6089,7 @@ xfs_bmap_split_extent_at(
 
 del_cursor:
 	if (cur) {
-		cur->bc_private.b.allocated = 0;
+		cur->bc_ino.allocated = 0;
 		xfs_btree_del_cursor(cur, error);
 	}
 
@@ -6142,34 +6098,6 @@ del_cursor:
 	return error;
 }
 
-int
-xfs_bmap_split_extent(
-	struct xfs_inode        *ip,
-	xfs_fileoff_t           split_fsb)
-{
-	struct xfs_mount        *mp = ip->i_mount;
-	struct xfs_trans        *tp;
-	int                     error;
-
-	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write,
-			XFS_DIOSTRAT_SPACE_RES(mp, 0), 0, 0, &tp);
-	if (error)
-		return error;
-
-	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
-
-	error = xfs_bmap_split_extent_at(tp, ip, split_fsb);
-	if (error)
-		goto out;
-
-	return xfs_trans_commit(tp);
-
-out:
-	xfs_trans_cancel(tp);
-	return error;
-}
-
 /* Deferred mapping is only for real extents in the data fork. */
 static bool
 xfs_bmap_is_update_needed(
diff --git a/fs/xfs/libxfs/xfs_bmap.h b/fs/xfs/libxfs/xfs_bmap.h
index 14d25e0b7d9c..e1bd484e5548 100644
--- a/fs/xfs/libxfs/xfs_bmap.h
+++ b/fs/xfs/libxfs/xfs_bmap.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000-2006 Silicon Graphics, Inc.
  * All Rights Reserved.
@@ -158,17 +158,22 @@ static inline int xfs_bmapi_whichfork(int bmapi_flags)
 	{ BMAP_ATTRFORK,	"ATTR" }, \
 	{ BMAP_COWFORK,		"COW" }
 
+/* Return true if the extent is an allocated extent, written or not. */
+static inline bool xfs_bmap_is_real_extent(struct xfs_bmbt_irec *irec)
+{
+	return irec->br_startblock != HOLESTARTBLOCK &&
+		irec->br_startblock != DELAYSTARTBLOCK &&
+		!isnullstartblock(irec->br_startblock);
+}
 
 /*
  * Return true if the extent is a real, allocated extent, or false if it is  a
  * delayed allocation, and unwritten extent or a hole.
  */
-static inline bool xfs_bmap_is_real_extent(struct xfs_bmbt_irec *irec)
+static inline bool xfs_bmap_is_written_extent(struct xfs_bmbt_irec *irec)
 {
-	return irec->br_state != XFS_EXT_UNWRITTEN &&
-		irec->br_startblock != HOLESTARTBLOCK &&
-		irec->br_startblock != DELAYSTARTBLOCK &&
-		!isnullstartblock(irec->br_startblock);
+	return xfs_bmap_is_real_extent(irec) &&
+	       irec->br_state != XFS_EXT_UNWRITTEN;
 }
 
 /*
@@ -222,7 +227,8 @@ int	xfs_bmap_can_insert_extents(struct xfs_inode *ip, xfs_fileoff_t off,
 int	xfs_bmap_insert_extents(struct xfs_trans *tp, struct xfs_inode *ip,
 		xfs_fileoff_t *next_fsb, xfs_fileoff_t offset_shift_fsb,
 		bool *done, xfs_fileoff_t stop_fsb);
-int	xfs_bmap_split_extent(struct xfs_inode *ip, xfs_fileoff_t split_offset);
+int	xfs_bmap_split_extent(struct xfs_trans *tp, struct xfs_inode *ip,
+		xfs_fileoff_t split_offset);
 int	xfs_bmapi_reserve_delalloc(struct xfs_inode *ip, int whichfork,
 		xfs_fileoff_t off, xfs_filblks_t len, xfs_filblks_t prealloc,
 		struct xfs_bmbt_irec *got, struct xfs_iext_cursor *cur,
diff --git a/fs/xfs/libxfs/xfs_bmap_btree.c b/fs/xfs/libxfs/xfs_bmap_btree.c
index ffe608d2a2d9..ecec604e6e4d 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.c
+++ b/fs/xfs/libxfs/xfs_bmap_btree.c
@@ -166,13 +166,13 @@ xfs_bmbt_dup_cursor(
 	struct xfs_btree_cur	*new;
 
 	new = xfs_bmbt_init_cursor(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.b.ip, cur->bc_private.b.whichfork);
+			cur->bc_ino.ip, cur->bc_ino.whichfork);
 
 	/*
 	 * Copy the firstblock, dfops, and flags values,
 	 * since init cursor doesn't get them.
 	 */
-	new->bc_private.b.flags = cur->bc_private.b.flags;
+	new->bc_ino.flags = cur->bc_ino.flags;
 
 	return new;
 }
@@ -183,12 +183,12 @@ xfs_bmbt_update_cursor(
 	struct xfs_btree_cur	*dst)
 {
 	ASSERT((dst->bc_tp->t_firstblock != NULLFSBLOCK) ||
-	       (dst->bc_private.b.ip->i_d.di_flags & XFS_DIFLAG_REALTIME));
+	       (dst->bc_ino.ip->i_d.di_flags & XFS_DIFLAG_REALTIME));
 
-	dst->bc_private.b.allocated += src->bc_private.b.allocated;
+	dst->bc_ino.allocated += src->bc_ino.allocated;
 	dst->bc_tp->t_firstblock = src->bc_tp->t_firstblock;
 
-	src->bc_private.b.allocated = 0;
+	src->bc_ino.allocated = 0;
 }
 
 STATIC int
@@ -205,8 +205,8 @@ xfs_bmbt_alloc_block(
 	args.tp = cur->bc_tp;
 	args.mp = cur->bc_mp;
 	args.fsbno = cur->bc_tp->t_firstblock;
-	xfs_rmap_ino_bmbt_owner(&args.oinfo, cur->bc_private.b.ip->i_ino,
-			cur->bc_private.b.whichfork);
+	xfs_rmap_ino_bmbt_owner(&args.oinfo, cur->bc_ino.ip->i_ino,
+			cur->bc_ino.whichfork);
 
 	if (args.fsbno == NULLFSBLOCK) {
 		args.fsbno = be64_to_cpu(start->l);
@@ -230,7 +230,7 @@ xfs_bmbt_alloc_block(
 	}
 
 	args.minlen = args.maxlen = args.prod = 1;
-	args.wasdel = cur->bc_private.b.flags & XFS_BTCUR_BPRV_WASDEL;
+	args.wasdel = cur->bc_ino.flags & XFS_BTCUR_BMBT_WASDEL;
 	if (!args.wasdel && args.tp->t_blk_res == 0) {
 		error = -ENOSPC;
 		goto error0;
@@ -259,10 +259,10 @@ xfs_bmbt_alloc_block(
 
 	ASSERT(args.len == 1);
 	cur->bc_tp->t_firstblock = args.fsbno;
-	cur->bc_private.b.allocated++;
-	cur->bc_private.b.ip->i_d.di_nblocks++;
-	xfs_trans_log_inode(args.tp, cur->bc_private.b.ip, XFS_ILOG_CORE);
-	xfs_trans_mod_dquot_byino(args.tp, cur->bc_private.b.ip,
+	cur->bc_ino.allocated++;
+	cur->bc_ino.ip->i_d.di_nblocks++;
+	xfs_trans_log_inode(args.tp, cur->bc_ino.ip, XFS_ILOG_CORE);
+	xfs_trans_mod_dquot_byino(args.tp, cur->bc_ino.ip,
 			XFS_TRANS_DQ_BCOUNT, 1L);
 
 	new->l = cpu_to_be64(args.fsbno);
@@ -280,12 +280,12 @@ xfs_bmbt_free_block(
 	struct xfs_buf		*bp)
 {
 	struct xfs_mount	*mp = cur->bc_mp;
-	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	struct xfs_inode	*ip = cur->bc_ino.ip;
 	struct xfs_trans	*tp = cur->bc_tp;
 	xfs_fsblock_t		fsbno = XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(bp));
 	struct xfs_owner_info	oinfo;
 
-	xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, cur->bc_private.b.whichfork);
+	xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, cur->bc_ino.whichfork);
 	xfs_bmap_add_free(cur->bc_tp, fsbno, 1, &oinfo);
 	ip->i_d.di_nblocks--;
 
@@ -302,8 +302,8 @@ xfs_bmbt_get_minrecs(
 	if (level == cur->bc_nlevels - 1) {
 		struct xfs_ifork	*ifp;
 
-		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
-				    cur->bc_private.b.whichfork);
+		ifp = XFS_IFORK_PTR(cur->bc_ino.ip,
+				    cur->bc_ino.whichfork);
 
 		return xfs_bmbt_maxrecs(cur->bc_mp,
 					ifp->if_broot_bytes, level == 0) / 2;
@@ -320,8 +320,8 @@ xfs_bmbt_get_maxrecs(
 	if (level == cur->bc_nlevels - 1) {
 		struct xfs_ifork	*ifp;
 
-		ifp = XFS_IFORK_PTR(cur->bc_private.b.ip,
-				    cur->bc_private.b.whichfork);
+		ifp = XFS_IFORK_PTR(cur->bc_ino.ip,
+				    cur->bc_ino.whichfork);
 
 		return xfs_bmbt_maxrecs(cur->bc_mp,
 					ifp->if_broot_bytes, level == 0);
@@ -347,7 +347,7 @@ 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_private.b.forksize, level == 0);
+	return xfs_bmdr_maxrecs(cur->bc_ino.forksize, level == 0);
 }
 
 STATIC void
@@ -552,7 +552,7 @@ xfs_bmbt_init_cursor(
 	struct xfs_btree_cur	*cur;
 	ASSERT(whichfork != XFS_COW_FORK);
 
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
+	cur = kmem_cache_zalloc(xfs_btree_cur_zone, GFP_NOFS | __GFP_NOFAIL);
 
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
@@ -566,11 +566,11 @@ xfs_bmbt_init_cursor(
 	if (xfs_sb_version_hascrc(&mp->m_sb))
 		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
 
-	cur->bc_private.b.forksize = XFS_IFORK_SIZE(ip, whichfork);
-	cur->bc_private.b.ip = ip;
-	cur->bc_private.b.allocated = 0;
-	cur->bc_private.b.flags = 0;
-	cur->bc_private.b.whichfork = whichfork;
+	cur->bc_ino.forksize = XFS_IFORK_SIZE(ip, whichfork);
+	cur->bc_ino.ip = ip;
+	cur->bc_ino.allocated = 0;
+	cur->bc_ino.flags = 0;
+	cur->bc_ino.whichfork = whichfork;
 
 	return cur;
 }
@@ -636,15 +636,12 @@ xfs_bmbt_change_owner(
 
 	ASSERT(tp || buffer_list);
 	ASSERT(!(tp && buffer_list));
-	if (whichfork == XFS_DATA_FORK)
-		ASSERT(ip->i_d.di_format == XFS_DINODE_FMT_BTREE);
-	else
-		ASSERT(ip->i_d.di_aformat == XFS_DINODE_FMT_BTREE);
+	ASSERT(XFS_IFORK_PTR(ip, whichfork)->if_format == XFS_DINODE_FMT_BTREE);
 
 	cur = xfs_bmbt_init_cursor(ip->i_mount, tp, ip, whichfork);
 	if (!cur)
 		return -ENOMEM;
-	cur->bc_private.b.flags |= XFS_BTCUR_BPRV_INVALID_OWNER;
+	cur->bc_ino.flags |= XFS_BTCUR_BMBT_INVALID_OWNER;
 
 	error = xfs_btree_change_owner(cur, new_owner, buffer_list);
 	xfs_btree_del_cursor(cur, error);
diff --git a/fs/xfs/libxfs/xfs_bmap_btree.h b/fs/xfs/libxfs/xfs_bmap_btree.h
index 29b407d053b4..72bf74c79fb9 100644
--- a/fs/xfs/libxfs/xfs_bmap_btree.h
+++ b/fs/xfs/libxfs/xfs_bmap_btree.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000,2002-2005 Silicon Graphics, Inc.
  * All Rights Reserved.
diff --git a/fs/xfs/libxfs/xfs_btree.c b/fs/xfs/libxfs/xfs_btree.c
index fd300dc93ca4..2d25bab68764 100644
--- a/fs/xfs/libxfs/xfs_btree.c
+++ b/fs/xfs/libxfs/xfs_btree.c
@@ -20,6 +20,7 @@
 #include "xfs_trace.h"
 #include "xfs_alloc.h"
 #include "xfs_log.h"
+#include "xfs_btree_staging.h"
 
 /*
  * Cursor allocation zone.
@@ -214,7 +215,7 @@ xfs_btree_check_sptr(
 {
 	if (level <= 0)
 		return false;
-	return xfs_verify_agbno(cur->bc_mp, cur->bc_private.a.agno, agbno);
+	return xfs_verify_agbno(cur->bc_mp, cur->bc_ag.agno, agbno);
 }
 
 /*
@@ -234,8 +235,8 @@ xfs_btree_check_ptr(
 			return 0;
 		xfs_err(cur->bc_mp,
 "Inode %llu fork %d: Corrupt btree %d pointer at level %d index %d.",
-				cur->bc_private.b.ip->i_ino,
-				cur->bc_private.b.whichfork, cur->bc_btnum,
+				cur->bc_ino.ip->i_ino,
+				cur->bc_ino.whichfork, cur->bc_btnum,
 				level, index);
 	} else {
 		if (xfs_btree_check_sptr(cur, be32_to_cpu((&ptr->s)[index]),
@@ -243,7 +244,7 @@ xfs_btree_check_ptr(
 			return 0;
 		xfs_err(cur->bc_mp,
 "AG %u: Corrupt btree %d pointer at level %d index %d.",
-				cur->bc_private.a.agno, cur->bc_btnum,
+				cur->bc_ag.agno, cur->bc_btnum,
 				level, index);
 	}
 
@@ -378,10 +379,12 @@ xfs_btree_del_cursor(
 	 * allocated indirect blocks' accounting.
 	 */
 	ASSERT(cur->bc_btnum != XFS_BTNUM_BMAP ||
-	       cur->bc_private.b.allocated == 0);
+	       cur->bc_ino.allocated == 0);
 	/*
 	 * Free the cursor.
 	 */
+	if (unlikely(cur->bc_flags & XFS_BTREE_STAGING))
+		kmem_free((void *)cur->bc_ops);
 	kmem_cache_free(xfs_btree_cur_zone, cur);
 }
 
@@ -642,6 +645,17 @@ xfs_btree_ptr_addr(
 		((char *)block + xfs_btree_ptr_offset(cur, n, level));
 }
 
+struct xfs_ifork *
+xfs_btree_ifork_ptr(
+	struct xfs_btree_cur	*cur)
+{
+	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+
+	if (cur->bc_flags & XFS_BTREE_STAGING)
+		return cur->bc_ino.ifake->if_fork;
+	return XFS_IFORK_PTR(cur->bc_ino.ip, cur->bc_ino.whichfork);
+}
+
 /*
  * Get the root block which is stored in the inode.
  *
@@ -652,9 +666,8 @@ STATIC struct xfs_btree_block *
 xfs_btree_get_iroot(
 	struct xfs_btree_cur	*cur)
 {
-	struct xfs_ifork	*ifp;
+	struct xfs_ifork	*ifp = xfs_btree_ifork_ptr(cur);
 
-	ifp = XFS_IFORK_PTR(cur->bc_private.b.ip, cur->bc_private.b.whichfork);
 	return (struct xfs_btree_block *)ifp->if_broot;
 }
 
@@ -881,13 +894,13 @@ xfs_btree_readahead_sblock(
 
 
 	if ((lr & XFS_BTCUR_LEFTRA) && left != NULLAGBLOCK) {
-		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
+		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_ag.agno,
 				     left, 1, cur->bc_ops->buf_ops);
 		rval++;
 	}
 
 	if ((lr & XFS_BTCUR_RIGHTRA) && right != NULLAGBLOCK) {
-		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_private.a.agno,
+		xfs_btree_reada_bufs(cur->bc_mp, cur->bc_ag.agno,
 				     right, 1, cur->bc_ops->buf_ops);
 		rval++;
 	}
@@ -945,7 +958,7 @@ xfs_btree_ptr_to_daddr(
 		*daddr = XFS_FSB_TO_DADDR(cur->bc_mp, fsbno);
 	} else {
 		agbno = be32_to_cpu(ptr->s);
-		*daddr = XFS_AGB_TO_DADDR(cur->bc_mp, cur->bc_private.a.agno,
+		*daddr = XFS_AGB_TO_DADDR(cur->bc_mp, cur->bc_ag.agno,
 				agbno);
 	}
 
@@ -1014,7 +1027,7 @@ xfs_btree_ptr_is_null(
 		return ptr->s == cpu_to_be32(NULLAGBLOCK);
 }
 
-STATIC void
+void
 xfs_btree_set_ptr_null(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*ptr)
@@ -1050,7 +1063,7 @@ xfs_btree_get_sibling(
 	}
 }
 
-STATIC void
+void
 xfs_btree_set_sibling(
 	struct xfs_btree_cur	*cur,
 	struct xfs_btree_block	*block,
@@ -1128,7 +1141,7 @@ xfs_btree_init_block(
 				 btnum, level, numrecs, owner, 0);
 }
 
-STATIC void
+void
 xfs_btree_init_block_cur(
 	struct xfs_btree_cur	*cur,
 	struct xfs_buf		*bp,
@@ -1144,9 +1157,9 @@ xfs_btree_init_block_cur(
 	 * code.
 	 */
 	if (cur->bc_flags & XFS_BTREE_LONG_PTRS)
-		owner = cur->bc_private.b.ip->i_ino;
+		owner = cur->bc_ino.ip->i_ino;
 	else
-		owner = cur->bc_private.a.agno;
+		owner = cur->bc_ag.agno;
 
 	xfs_btree_init_block_int(cur->bc_mp, XFS_BUF_TO_BLOCK(bp), bp->b_bn,
 				 cur->bc_btnum, level, numrecs,
@@ -1220,7 +1233,7 @@ xfs_btree_set_refs(
 	}
 }
 
-STATIC int
+int
 xfs_btree_get_buf_block(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*ptr,
@@ -1280,7 +1293,7 @@ xfs_btree_read_buf_block(
 /*
  * Copy keys from one btree block to another.
  */
-STATIC void
+void
 xfs_btree_copy_keys(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_key	*dst_key,
@@ -1308,11 +1321,11 @@ xfs_btree_copy_recs(
 /*
  * Copy block pointers from one btree block to another.
  */
-STATIC void
+void
 xfs_btree_copy_ptrs(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*dst_ptr,
-	union xfs_btree_ptr	*src_ptr,
+	const union xfs_btree_ptr *src_ptr,
 	int			numptrs)
 {
 	ASSERT(numptrs >= 0);
@@ -1393,8 +1406,8 @@ xfs_btree_log_keys(
 				  xfs_btree_key_offset(cur, first),
 				  xfs_btree_key_offset(cur, last + 1) - 1);
 	} else {
-		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
-				xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_ino.ip,
+				xfs_ilog_fbroot(cur->bc_ino.whichfork));
 	}
 }
 
@@ -1436,8 +1449,8 @@ xfs_btree_log_ptrs(
 				xfs_btree_ptr_offset(cur, first, level),
 				xfs_btree_ptr_offset(cur, last + 1, level) - 1);
 	} else {
-		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
-			xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_ino.ip,
+			xfs_ilog_fbroot(cur->bc_ino.whichfork));
 	}
 
 }
@@ -1505,8 +1518,8 @@ xfs_btree_log_block(
 		xfs_trans_buf_set_type(cur->bc_tp, bp, XFS_BLFT_BTREE_BUF);
 		xfs_trans_log_buf(cur->bc_tp, bp, first, last);
 	} else {
-		xfs_trans_log_inode(cur->bc_tp, cur->bc_private.b.ip,
-			xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+		xfs_trans_log_inode(cur->bc_tp, cur->bc_ino.ip,
+			xfs_ilog_fbroot(cur->bc_ino.whichfork));
 	}
 }
 
@@ -1743,10 +1756,10 @@ xfs_btree_lookup_get_block(
 
 	/* Check the inode owner since the verifiers don't. */
 	if (xfs_sb_version_hascrc(&cur->bc_mp->m_sb) &&
-	    !(cur->bc_private.b.flags & XFS_BTCUR_BPRV_INVALID_OWNER) &&
+	    !(cur->bc_ino.flags & XFS_BTCUR_BMBT_INVALID_OWNER) &&
 	    (cur->bc_flags & XFS_BTREE_LONG_PTRS) &&
 	    be64_to_cpu((*blkp)->bb_u.l.bb_owner) !=
-			cur->bc_private.b.ip->i_ino)
+			cur->bc_ino.ip->i_ino)
 		goto out_bad;
 
 	/* Did we get the level we were looking for? */
@@ -1762,7 +1775,7 @@ xfs_btree_lookup_get_block(
 
 out_bad:
 	*blkp = NULL;
-	xfs_buf_corruption_error(bp);
+	xfs_buf_mark_corrupt(bp);
 	xfs_trans_brelse(cur->bc_tp, bp);
 	return -EFSCORRUPTED;
 }
@@ -2938,9 +2951,9 @@ xfs_btree_new_iroot(
 
 	xfs_btree_copy_ptrs(cur, pp, &nptr, 1);
 
-	xfs_iroot_realloc(cur->bc_private.b.ip,
+	xfs_iroot_realloc(cur->bc_ino.ip,
 			  1 - xfs_btree_get_numrecs(cblock),
-			  cur->bc_private.b.whichfork);
+			  cur->bc_ino.whichfork);
 
 	xfs_btree_setbuf(cur, level, cbp);
 
@@ -2953,7 +2966,7 @@ xfs_btree_new_iroot(
 	xfs_btree_log_ptrs(cur, cbp, 1, be16_to_cpu(cblock->bb_numrecs));
 
 	*logflags |=
-		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_private.b.whichfork);
+		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_ino.whichfork);
 	*stat = 1;
 	return 0;
 error0:
@@ -3105,11 +3118,11 @@ xfs_btree_make_block_unfull(
 
 	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
 	    level == cur->bc_nlevels - 1) {
-		struct xfs_inode *ip = cur->bc_private.b.ip;
+		struct xfs_inode *ip = cur->bc_ino.ip;
 
 		if (numrecs < cur->bc_ops->get_dmaxrecs(cur, level)) {
 			/* A root block that can be made bigger. */
-			xfs_iroot_realloc(ip, 1, cur->bc_private.b.whichfork);
+			xfs_iroot_realloc(ip, 1, cur->bc_ino.whichfork);
 			*stat = 1;
 		} else {
 			/* A root block that needs replacing */
@@ -3455,8 +3468,8 @@ STATIC int
 xfs_btree_kill_iroot(
 	struct xfs_btree_cur	*cur)
 {
-	int			whichfork = cur->bc_private.b.whichfork;
-	struct xfs_inode	*ip = cur->bc_private.b.ip;
+	int			whichfork = cur->bc_ino.whichfork;
+	struct xfs_inode	*ip = cur->bc_ino.ip;
 	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 	struct xfs_btree_block	*block;
 	struct xfs_btree_block	*cblock;
@@ -3514,8 +3527,8 @@ xfs_btree_kill_iroot(
 
 	index = numrecs - cur->bc_ops->get_maxrecs(cur, level);
 	if (index) {
-		xfs_iroot_realloc(cur->bc_private.b.ip, index,
-				  cur->bc_private.b.whichfork);
+		xfs_iroot_realloc(cur->bc_ino.ip, index,
+				  cur->bc_ino.whichfork);
 		block = ifp->if_broot;
 	}
 
@@ -3544,7 +3557,7 @@ xfs_btree_kill_iroot(
 	cur->bc_bufs[level - 1] = NULL;
 	be16_add_cpu(&block->bb_level, -1);
 	xfs_trans_log_inode(cur->bc_tp, ip,
-		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_private.b.whichfork));
+		XFS_ILOG_CORE | xfs_ilog_fbroot(cur->bc_ino.whichfork));
 	cur->bc_nlevels--;
 out0:
 	return 0;
@@ -3712,8 +3725,8 @@ xfs_btree_delrec(
 	 */
 	if (level == cur->bc_nlevels - 1) {
 		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
-			xfs_iroot_realloc(cur->bc_private.b.ip, -1,
-					  cur->bc_private.b.whichfork);
+			xfs_iroot_realloc(cur->bc_ino.ip, -1,
+					  cur->bc_ino.whichfork);
 
 			error = xfs_btree_kill_iroot(cur);
 			if (error)
diff --git a/fs/xfs/libxfs/xfs_btree.h b/fs/xfs/libxfs/xfs_btree.h
index 3eff7c321d43..10e50cbacacf 100644
--- a/fs/xfs/libxfs/xfs_btree.h
+++ b/fs/xfs/libxfs/xfs_btree.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
  * All Rights Reserved.
@@ -10,6 +10,7 @@ struct xfs_buf;
 struct xfs_inode;
 struct xfs_mount;
 struct xfs_trans;
+struct xfs_ifork;
 
 extern kmem_zone_t	*xfs_btree_cur_zone;
 
@@ -177,15 +178,37 @@ union xfs_btree_irec {
 	struct xfs_refcount_irec	rc;
 };
 
-/* Per-AG btree private information. */
-union xfs_btree_cur_private {
-	struct {
-		unsigned long	nr_ops;		/* # record updates */
-		int		shape_changes;	/* # of extent splits */
-	} refc;
-	struct {
-		bool		active;		/* allocation cursor state */
-	} abt;
+/* Per-AG btree information. */
+struct xfs_btree_cur_ag {
+	union {
+		struct xfs_buf		*agbp;
+		struct xbtree_afakeroot	*afake;	/* for staging cursor */
+	};
+	xfs_agnumber_t		agno;
+	union {
+		struct {
+			unsigned long nr_ops;	/* # record updates */
+			int	shape_changes;	/* # of extent splits */
+		} refc;
+		struct {
+			bool	active;		/* allocation cursor state */
+		} abt;
+	};
+};
+
+/* Btree-in-inode cursor information */
+struct xfs_btree_cur_ino {
+	struct xfs_inode		*ip;
+	struct xbtree_ifakeroot		*ifake;	/* for staging cursor */
+	int				allocated;
+	short				forksize;
+	char				whichfork;
+	char				flags;
+/* We are converting a delalloc reservation */
+#define	XFS_BTCUR_BMBT_WASDEL		(1 << 0)
+
+/* For extent swap, ignore owner check in verifier */
+#define	XFS_BTCUR_BMBT_INVALID_OWNER	(1 << 1)
 };
 
 /*
@@ -209,21 +232,9 @@ typedef struct xfs_btree_cur
 	xfs_btnum_t	bc_btnum;	/* identifies which btree type */
 	int		bc_statoff;	/* offset of btre stats array */
 	union {
-		struct {			/* needed for BNO, CNT, INO */
-			struct xfs_buf	*agbp;	/* agf/agi buffer pointer */
-			xfs_agnumber_t	agno;	/* ag number */
-			union xfs_btree_cur_private	priv;
-		} a;
-		struct {			/* needed for BMAP */
-			struct xfs_inode *ip;	/* pointer to our inode */
-			int		allocated;	/* count of alloced */
-			short		forksize;	/* fork's inode space */
-			char		whichfork;	/* data or attr fork */
-			char		flags;		/* flags */
-#define	XFS_BTCUR_BPRV_WASDEL		(1<<0)		/* was delayed */
-#define	XFS_BTCUR_BPRV_INVALID_OWNER	(1<<1)		/* for ext swap */
-		} b;
-	}		bc_private;	/* per-btree type data */
+		struct xfs_btree_cur_ag	bc_ag;
+		struct xfs_btree_cur_ino bc_ino;
+	};
 } xfs_btree_cur_t;
 
 /* cursor flags */
@@ -232,6 +243,12 @@ typedef struct xfs_btree_cur
 #define XFS_BTREE_LASTREC_UPDATE	(1<<2)	/* track last rec externally */
 #define XFS_BTREE_CRC_BLOCKS		(1<<3)	/* uses extended btree blocks */
 #define XFS_BTREE_OVERLAPPING		(1<<4)	/* overlapping intervals */
+/*
+ * The root of this btree is a fakeroot structure so that we can stage a btree
+ * rebuild without leaving it accessible via primary metadata.  The ops struct
+ * is dynamically allocated and must be freed when the cursor is deleted.
+ */
+#define XFS_BTREE_STAGING		(1<<5)
 
 
 #define	XFS_BTREE_NOERROR	0
@@ -494,6 +511,7 @@ union xfs_btree_key *xfs_btree_high_key_from_key(struct xfs_btree_cur *cur,
 int xfs_btree_has_record(struct xfs_btree_cur *cur, union xfs_btree_irec *low,
 		union xfs_btree_irec *high, bool *exists);
 bool xfs_btree_has_more_records(struct xfs_btree_cur *cur);
+struct xfs_ifork *xfs_btree_ifork_ptr(struct xfs_btree_cur *cur);
 
 /* Does this cursor point to the last block in the given level? */
 static inline bool
@@ -512,4 +530,20 @@ xfs_btree_islastblock(
 	return block->bb_u.s.bb_rightsib == cpu_to_be32(NULLAGBLOCK);
 }
 
+void xfs_btree_set_ptr_null(struct xfs_btree_cur *cur,
+		union xfs_btree_ptr *ptr);
+int xfs_btree_get_buf_block(struct xfs_btree_cur *cur, union xfs_btree_ptr *ptr,
+		struct xfs_btree_block **block, struct xfs_buf **bpp);
+void xfs_btree_set_sibling(struct xfs_btree_cur *cur,
+		struct xfs_btree_block *block, union xfs_btree_ptr *ptr,
+		int lr);
+void xfs_btree_init_block_cur(struct xfs_btree_cur *cur,
+		struct xfs_buf *bp, int level, int numrecs);
+void xfs_btree_copy_ptrs(struct xfs_btree_cur *cur,
+		union xfs_btree_ptr *dst_ptr,
+		const union xfs_btree_ptr *src_ptr, int numptrs);
+void xfs_btree_copy_keys(struct xfs_btree_cur *cur,
+		union xfs_btree_key *dst_key, union xfs_btree_key *src_key,
+		int numkeys);
+
 #endif	/* __XFS_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_btree_staging.c b/fs/xfs/libxfs/xfs_btree_staging.c
new file mode 100644
index 000000000000..f464a7c7cf22
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_btree_staging.c
@@ -0,0 +1,879 @@
+// SPDX-License-Identifier: GPL-2.0-or-later
+/*
+ * Copyright (C) 2020 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_bit.h"
+#include "xfs_mount.h"
+#include "xfs_inode.h"
+#include "xfs_trans.h"
+#include "xfs_btree.h"
+#include "xfs_trace.h"
+#include "xfs_btree_staging.h"
+
+/*
+ * Staging Cursors and Fake Roots for Btrees
+ * =========================================
+ *
+ * A staging btree cursor is a special type of btree cursor that callers must
+ * use to construct a new btree index using the btree bulk loader code.  The
+ * bulk loading code uses the staging btree cursor to abstract the details of
+ * initializing new btree blocks and filling them with records or key/ptr
+ * pairs.  Regular btree operations (e.g. queries and modifications) are not
+ * supported with staging cursors, and callers must not invoke them.
+ *
+ * Fake root structures contain all the information about a btree that is under
+ * construction by the bulk loading code.  Staging btree cursors point to fake
+ * root structures instead of the usual AG header or inode structure.
+ *
+ * Callers are expected to initialize a fake root structure and pass it into
+ * the _stage_cursor function for a specific btree type.  When bulk loading is
+ * complete, callers should call the _commit_staged_btree function for that
+ * specific btree type to commit the new btree into the filesystem.
+ */
+
+/*
+ * Don't allow staging cursors to be duplicated because they're supposed to be
+ * kept private to a single thread.
+ */
+STATIC struct xfs_btree_cur *
+xfs_btree_fakeroot_dup_cursor(
+	struct xfs_btree_cur	*cur)
+{
+	ASSERT(0);
+	return NULL;
+}
+
+/*
+ * Don't allow block allocation for a staging cursor, because staging cursors
+ * do not support regular btree modifications.
+ *
+ * Bulk loading uses a separate callback to obtain new blocks from a
+ * preallocated list, which prevents ENOSPC failures during loading.
+ */
+STATIC int
+xfs_btree_fakeroot_alloc_block(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*start_bno,
+	union xfs_btree_ptr	*new_bno,
+	int			*stat)
+{
+	ASSERT(0);
+	return -EFSCORRUPTED;
+}
+
+/*
+ * Don't allow block freeing for a staging cursor, because staging cursors
+ * do not support regular btree modifications.
+ */
+STATIC int
+xfs_btree_fakeroot_free_block(
+	struct xfs_btree_cur	*cur,
+	struct xfs_buf		*bp)
+{
+	ASSERT(0);
+	return -EFSCORRUPTED;
+}
+
+/* Initialize a pointer to the root block from the fakeroot. */
+STATIC void
+xfs_btree_fakeroot_init_ptr_from_cur(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr)
+{
+	struct xbtree_afakeroot	*afake;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+
+	afake = cur->bc_ag.afake;
+	ptr->s = cpu_to_be32(afake->af_root);
+}
+
+/*
+ * Bulk Loading for AG Btrees
+ * ==========================
+ *
+ * For a btree rooted in an AG header, pass a xbtree_afakeroot structure to the
+ * staging cursor.  Callers should initialize this to zero.
+ *
+ * The _stage_cursor() function for a specific btree type should call
+ * xfs_btree_stage_afakeroot to set up the in-memory cursor as a staging
+ * cursor.  The corresponding _commit_staged_btree() function should log the
+ * new root and call xfs_btree_commit_afakeroot() to transform the staging
+ * cursor into a regular btree cursor.
+ */
+
+/* Update the btree root information for a per-AG fake root. */
+STATIC void
+xfs_btree_afakeroot_set_root(
+	struct xfs_btree_cur	*cur,
+	union xfs_btree_ptr	*ptr,
+	int			inc)
+{
+	struct xbtree_afakeroot	*afake = cur->bc_ag.afake;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+	afake->af_root = be32_to_cpu(ptr->s);
+	afake->af_levels += inc;
+}
+
+/*
+ * Initialize a AG-rooted btree cursor with the given AG btree fake root.
+ * The btree cursor's bc_ops will be overridden as needed to make the staging
+ * functionality work.
+ */
+void
+xfs_btree_stage_afakeroot(
+	struct xfs_btree_cur		*cur,
+	struct xbtree_afakeroot		*afake)
+{
+	struct xfs_btree_ops		*nops;
+
+	ASSERT(!(cur->bc_flags & XFS_BTREE_STAGING));
+	ASSERT(!(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE));
+	ASSERT(cur->bc_tp == NULL);
+
+	nops = kmem_alloc(sizeof(struct xfs_btree_ops), KM_NOFS);
+	memcpy(nops, cur->bc_ops, sizeof(struct xfs_btree_ops));
+	nops->alloc_block = xfs_btree_fakeroot_alloc_block;
+	nops->free_block = xfs_btree_fakeroot_free_block;
+	nops->init_ptr_from_cur = xfs_btree_fakeroot_init_ptr_from_cur;
+	nops->set_root = xfs_btree_afakeroot_set_root;
+	nops->dup_cursor = xfs_btree_fakeroot_dup_cursor;
+
+	cur->bc_ag.afake = afake;
+	cur->bc_nlevels = afake->af_levels;
+	cur->bc_ops = nops;
+	cur->bc_flags |= XFS_BTREE_STAGING;
+}
+
+/*
+ * Transform an AG-rooted staging btree cursor back into a regular cursor by
+ * substituting a real btree root for the fake one and restoring normal btree
+ * cursor ops.  The caller must log the btree root change prior to calling
+ * this.
+ */
+void
+xfs_btree_commit_afakeroot(
+	struct xfs_btree_cur		*cur,
+	struct xfs_trans		*tp,
+	struct xfs_buf			*agbp,
+	const struct xfs_btree_ops	*ops)
+{
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+	ASSERT(cur->bc_tp == NULL);
+
+	trace_xfs_btree_commit_afakeroot(cur);
+
+	kmem_free((void *)cur->bc_ops);
+	cur->bc_ag.agbp = agbp;
+	cur->bc_ops = ops;
+	cur->bc_flags &= ~XFS_BTREE_STAGING;
+	cur->bc_tp = tp;
+}
+
+/*
+ * Bulk Loading for Inode-Rooted Btrees
+ * ====================================
+ *
+ * For a btree rooted in an inode fork, pass a xbtree_ifakeroot structure to
+ * the staging cursor.  This structure should be initialized as follows:
+ *
+ * - if_fork_size field should be set to the number of bytes available to the
+ *   fork in the inode.
+ *
+ * - if_fork should point to a freshly allocated struct xfs_ifork.
+ *
+ * - if_format should be set to the appropriate fork type (e.g.
+ *   XFS_DINODE_FMT_BTREE).
+ *
+ * All other fields must be zero.
+ *
+ * The _stage_cursor() function for a specific btree type should call
+ * xfs_btree_stage_ifakeroot to set up the in-memory cursor as a staging
+ * cursor.  The corresponding _commit_staged_btree() function should log the
+ * new root and call xfs_btree_commit_ifakeroot() to transform the staging
+ * cursor into a regular btree cursor.
+ */
+
+/*
+ * Initialize an inode-rooted btree cursor with the given inode btree fake
+ * root.  The btree cursor's bc_ops will be overridden as needed to make the
+ * staging functionality work.  If new_ops is not NULL, these new ops will be
+ * passed out to the caller for further overriding.
+ */
+void
+xfs_btree_stage_ifakeroot(
+	struct xfs_btree_cur		*cur,
+	struct xbtree_ifakeroot		*ifake,
+	struct xfs_btree_ops		**new_ops)
+{
+	struct xfs_btree_ops		*nops;
+
+	ASSERT(!(cur->bc_flags & XFS_BTREE_STAGING));
+	ASSERT(cur->bc_flags & XFS_BTREE_ROOT_IN_INODE);
+	ASSERT(cur->bc_tp == NULL);
+
+	nops = kmem_alloc(sizeof(struct xfs_btree_ops), KM_NOFS);
+	memcpy(nops, cur->bc_ops, sizeof(struct xfs_btree_ops));
+	nops->alloc_block = xfs_btree_fakeroot_alloc_block;
+	nops->free_block = xfs_btree_fakeroot_free_block;
+	nops->init_ptr_from_cur = xfs_btree_fakeroot_init_ptr_from_cur;
+	nops->dup_cursor = xfs_btree_fakeroot_dup_cursor;
+
+	cur->bc_ino.ifake = ifake;
+	cur->bc_nlevels = ifake->if_levels;
+	cur->bc_ops = nops;
+	cur->bc_flags |= XFS_BTREE_STAGING;
+
+	if (new_ops)
+		*new_ops = nops;
+}
+
+/*
+ * Transform an inode-rooted staging btree cursor back into a regular cursor by
+ * substituting a real btree root for the fake one and restoring normal btree
+ * cursor ops.  The caller must log the btree root change prior to calling
+ * this.
+ */
+void
+xfs_btree_commit_ifakeroot(
+	struct xfs_btree_cur		*cur,
+	struct xfs_trans		*tp,
+	int				whichfork,
+	const struct xfs_btree_ops	*ops)
+{
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+	ASSERT(cur->bc_tp == NULL);
+
+	trace_xfs_btree_commit_ifakeroot(cur);
+
+	kmem_free((void *)cur->bc_ops);
+	cur->bc_ino.ifake = NULL;
+	cur->bc_ino.whichfork = whichfork;
+	cur->bc_ops = ops;
+	cur->bc_flags &= ~XFS_BTREE_STAGING;
+	cur->bc_tp = tp;
+}
+
+/*
+ * Bulk Loading of Staged Btrees
+ * =============================
+ *
+ * This interface is used with a staged btree cursor to create a totally new
+ * btree with a large number of records (i.e. more than what would fit in a
+ * single root block).  When the creation is complete, the new root can be
+ * linked atomically into the filesystem by committing the staged cursor.
+ *
+ * Creation of a new btree proceeds roughly as follows:
+ *
+ * The first step is to initialize an appropriate fake btree root structure and
+ * then construct a staged btree cursor.  Refer to the block comments about
+ * "Bulk Loading for AG Btrees" and "Bulk Loading for Inode-Rooted Btrees" for
+ * more information about how to do this.
+ *
+ * The second step is to initialize a struct xfs_btree_bload context as
+ * documented in the structure definition.
+ *
+ * The third step is to call xfs_btree_bload_compute_geometry to compute the
+ * height of and the number of blocks needed to construct the btree.  See the
+ * section "Computing the Geometry of the New Btree" for details about this
+ * computation.
+ *
+ * In step four, the caller must allocate xfs_btree_bload.nr_blocks blocks and
+ * save them for later use by ->claim_block().  Bulk loading requires all
+ * blocks to be allocated beforehand to avoid ENOSPC failures midway through a
+ * rebuild, and to minimize seek distances of the new btree.
+ *
+ * Step five is to call xfs_btree_bload() to start constructing the btree.
+ *
+ * The final step is to commit the staging btree cursor, which logs the new
+ * btree root and turns the staging cursor into a regular cursor.  The caller
+ * is responsible for cleaning up the previous btree blocks, if any.
+ *
+ * Computing the Geometry of the New Btree
+ * =======================================
+ *
+ * The number of items placed in each btree block is computed via the following
+ * algorithm: For leaf levels, the number of items for the level is nr_records
+ * in the bload structure.  For node levels, the number of items for the level
+ * is the number of blocks in the next lower level of the tree.  For each
+ * level, the desired number of items per block is defined as:
+ *
+ * desired = max(minrecs, maxrecs - slack factor)
+ *
+ * The number of blocks for the level is defined to be:
+ *
+ * blocks = floor(nr_items / desired)
+ *
+ * Note this is rounded down so that the npb calculation below will never fall
+ * below minrecs.  The number of items that will actually be loaded into each
+ * btree block is defined as:
+ *
+ * npb =  nr_items / blocks
+ *
+ * Some of the leftmost blocks in the level will contain one extra record as
+ * needed to handle uneven division.  If the number of records in any block
+ * would exceed maxrecs for that level, blocks is incremented and npb is
+ * recalculated.
+ *
+ * In other words, we compute the number of blocks needed to satisfy a given
+ * loading level, then spread the items as evenly as possible.
+ *
+ * The height and number of fs blocks required to create the btree are computed
+ * and returned via btree_height and nr_blocks.
+ */
+
+/*
+ * Put a btree block that we're loading onto the ordered list and release it.
+ * The btree blocks will be written to disk when bulk loading is finished.
+ */
+static void
+xfs_btree_bload_drop_buf(
+	struct list_head	*buffers_list,
+	struct xfs_buf		**bpp)
+{
+	if (*bpp == NULL)
+		return;
+
+	if (!xfs_buf_delwri_queue(*bpp, buffers_list))
+		ASSERT(0);
+
+	xfs_buf_relse(*bpp);
+	*bpp = NULL;
+}
+
+/*
+ * Allocate and initialize one btree block for bulk loading.
+ *
+ * The new btree block will have its level and numrecs fields set to the values
+ * of the level and nr_this_block parameters, respectively.
+ *
+ * The caller should ensure that ptrp, bpp, and blockp refer to the left
+ * sibling of the new block, if there is any.  On exit, ptrp, bpp, and blockp
+ * will all point to the new block.
+ */
+STATIC int
+xfs_btree_bload_prep_block(
+	struct xfs_btree_cur		*cur,
+	struct xfs_btree_bload		*bbl,
+	struct list_head		*buffers_list,
+	unsigned int			level,
+	unsigned int			nr_this_block,
+	union xfs_btree_ptr		*ptrp, /* in/out */
+	struct xfs_buf			**bpp, /* in/out */
+	struct xfs_btree_block		**blockp, /* in/out */
+	void				*priv)
+{
+	union xfs_btree_ptr		new_ptr;
+	struct xfs_buf			*new_bp;
+	struct xfs_btree_block		*new_block;
+	int				ret;
+
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    level == cur->bc_nlevels - 1) {
+		struct xfs_ifork	*ifp = xfs_btree_ifork_ptr(cur);
+		size_t			new_size;
+
+		ASSERT(*bpp == NULL);
+
+		/* Allocate a new incore btree root block. */
+		new_size = bbl->iroot_size(cur, nr_this_block, priv);
+		ifp->if_broot = kmem_zalloc(new_size, 0);
+		ifp->if_broot_bytes = (int)new_size;
+		ifp->if_flags |= XFS_IFBROOT;
+
+		/* Initialize it and send it out. */
+		xfs_btree_init_block_int(cur->bc_mp, ifp->if_broot,
+				XFS_BUF_DADDR_NULL, cur->bc_btnum, level,
+				nr_this_block, cur->bc_ino.ip->i_ino,
+				cur->bc_flags);
+
+		*bpp = NULL;
+		*blockp = ifp->if_broot;
+		xfs_btree_set_ptr_null(cur, ptrp);
+		return 0;
+	}
+
+	/* Claim one of the caller's preallocated blocks. */
+	xfs_btree_set_ptr_null(cur, &new_ptr);
+	ret = bbl->claim_block(cur, &new_ptr, priv);
+	if (ret)
+		return ret;
+
+	ASSERT(!xfs_btree_ptr_is_null(cur, &new_ptr));
+
+	ret = xfs_btree_get_buf_block(cur, &new_ptr, &new_block, &new_bp);
+	if (ret)
+		return ret;
+
+	/*
+	 * The previous block (if any) is the left sibling of the new block,
+	 * so set its right sibling pointer to the new block and drop it.
+	 */
+	if (*blockp)
+		xfs_btree_set_sibling(cur, *blockp, &new_ptr, XFS_BB_RIGHTSIB);
+	xfs_btree_bload_drop_buf(buffers_list, bpp);
+
+	/* Initialize the new btree block. */
+	xfs_btree_init_block_cur(cur, new_bp, level, nr_this_block);
+	xfs_btree_set_sibling(cur, new_block, ptrp, XFS_BB_LEFTSIB);
+
+	/* Set the out parameters. */
+	*bpp = new_bp;
+	*blockp = new_block;
+	xfs_btree_copy_ptrs(cur, ptrp, &new_ptr, 1);
+	return 0;
+}
+
+/* Load one leaf block. */
+STATIC int
+xfs_btree_bload_leaf(
+	struct xfs_btree_cur		*cur,
+	unsigned int			recs_this_block,
+	xfs_btree_bload_get_record_fn	get_record,
+	struct xfs_btree_block		*block,
+	void				*priv)
+{
+	unsigned int			j;
+	int				ret;
+
+	/* Fill the leaf block with records. */
+	for (j = 1; j <= recs_this_block; j++) {
+		union xfs_btree_rec	*block_rec;
+
+		ret = get_record(cur, priv);
+		if (ret)
+			return ret;
+		block_rec = xfs_btree_rec_addr(cur, j, block);
+		cur->bc_ops->init_rec_from_cur(cur, block_rec);
+	}
+
+	return 0;
+}
+
+/*
+ * Load one node block with key/ptr pairs.
+ *
+ * child_ptr must point to a block within the next level down in the tree.  A
+ * key/ptr entry will be created in the new node block to the block pointed to
+ * by child_ptr.  On exit, child_ptr points to the next block on the child
+ * level that needs processing.
+ */
+STATIC int
+xfs_btree_bload_node(
+	struct xfs_btree_cur	*cur,
+	unsigned int		recs_this_block,
+	union xfs_btree_ptr	*child_ptr,
+	struct xfs_btree_block	*block)
+{
+	unsigned int		j;
+	int			ret;
+
+	/* Fill the node block with keys and pointers. */
+	for (j = 1; j <= recs_this_block; j++) {
+		union xfs_btree_key	child_key;
+		union xfs_btree_ptr	*block_ptr;
+		union xfs_btree_key	*block_key;
+		struct xfs_btree_block	*child_block;
+		struct xfs_buf		*child_bp;
+
+		ASSERT(!xfs_btree_ptr_is_null(cur, child_ptr));
+
+		ret = xfs_btree_get_buf_block(cur, child_ptr, &child_block,
+				&child_bp);
+		if (ret)
+			return ret;
+
+		block_ptr = xfs_btree_ptr_addr(cur, j, block);
+		xfs_btree_copy_ptrs(cur, block_ptr, child_ptr, 1);
+
+		block_key = xfs_btree_key_addr(cur, j, block);
+		xfs_btree_get_keys(cur, child_block, &child_key);
+		xfs_btree_copy_keys(cur, block_key, &child_key, 1);
+
+		xfs_btree_get_sibling(cur, child_block, child_ptr,
+				XFS_BB_RIGHTSIB);
+		xfs_buf_relse(child_bp);
+	}
+
+	return 0;
+}
+
+/*
+ * Compute the maximum number of records (or keyptrs) per block that we want to
+ * install at this level in the btree.  Caller is responsible for having set
+ * @cur->bc_ino.forksize to the desired fork size, if appropriate.
+ */
+STATIC unsigned int
+xfs_btree_bload_max_npb(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_bload	*bbl,
+	unsigned int		level)
+{
+	unsigned int		ret;
+
+	if (level == cur->bc_nlevels - 1 && cur->bc_ops->get_dmaxrecs)
+		return cur->bc_ops->get_dmaxrecs(cur, level);
+
+	ret = cur->bc_ops->get_maxrecs(cur, level);
+	if (level == 0)
+		ret -= bbl->leaf_slack;
+	else
+		ret -= bbl->node_slack;
+	return ret;
+}
+
+/*
+ * Compute the desired number of records (or keyptrs) per block that we want to
+ * install at this level in the btree, which must be somewhere between minrecs
+ * and max_npb.  The caller is free to install fewer records per block.
+ */
+STATIC unsigned int
+xfs_btree_bload_desired_npb(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_bload	*bbl,
+	unsigned int		level)
+{
+	unsigned int		npb = xfs_btree_bload_max_npb(cur, bbl, level);
+
+	/* Root blocks are not subject to minrecs rules. */
+	if (level == cur->bc_nlevels - 1)
+		return max(1U, npb);
+
+	return max_t(unsigned int, cur->bc_ops->get_minrecs(cur, level), npb);
+}
+
+/*
+ * Compute the number of records to be stored in each block at this level and
+ * the number of blocks for this level.  For leaf levels, we must populate an
+ * empty root block even if there are no records, so we have to have at least
+ * one block.
+ */
+STATIC void
+xfs_btree_bload_level_geometry(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_bload	*bbl,
+	unsigned int		level,
+	uint64_t		nr_this_level,
+	unsigned int		*avg_per_block,
+	uint64_t		*blocks,
+	uint64_t		*blocks_with_extra)
+{
+	uint64_t		npb;
+	uint64_t		dontcare;
+	unsigned int		desired_npb;
+	unsigned int		maxnr;
+
+	maxnr = cur->bc_ops->get_maxrecs(cur, level);
+
+	/*
+	 * Compute the number of blocks we need to fill each block with the
+	 * desired number of records/keyptrs per block.  Because desired_npb
+	 * could be minrecs, we use regular integer division (which rounds
+	 * the block count down) so that in the next step the effective # of
+	 * items per block will never be less than desired_npb.
+	 */
+	desired_npb = xfs_btree_bload_desired_npb(cur, bbl, level);
+	*blocks = div64_u64_rem(nr_this_level, desired_npb, &dontcare);
+	*blocks = max(1ULL, *blocks);
+
+	/*
+	 * Compute the number of records that we will actually put in each
+	 * block, assuming that we want to spread the records evenly between
+	 * the blocks.  Take care that the effective # of items per block (npb)
+	 * won't exceed maxrecs even for the blocks that get an extra record,
+	 * since desired_npb could be maxrecs, and in the previous step we
+	 * rounded the block count down.
+	 */
+	npb = div64_u64_rem(nr_this_level, *blocks, blocks_with_extra);
+	if (npb > maxnr || (npb == maxnr && *blocks_with_extra > 0)) {
+		(*blocks)++;
+		npb = div64_u64_rem(nr_this_level, *blocks, blocks_with_extra);
+	}
+
+	*avg_per_block = min_t(uint64_t, npb, nr_this_level);
+
+	trace_xfs_btree_bload_level_geometry(cur, level, nr_this_level,
+			*avg_per_block, desired_npb, *blocks,
+			*blocks_with_extra);
+}
+
+/*
+ * Ensure a slack value is appropriate for the btree.
+ *
+ * If the slack value is negative, set slack so that we fill the block to
+ * halfway between minrecs and maxrecs.  Make sure the slack is never so large
+ * that we can underflow minrecs.
+ */
+static void
+xfs_btree_bload_ensure_slack(
+	struct xfs_btree_cur	*cur,
+	int			*slack,
+	int			level)
+{
+	int			maxr;
+	int			minr;
+
+	maxr = cur->bc_ops->get_maxrecs(cur, level);
+	minr = cur->bc_ops->get_minrecs(cur, level);
+
+	/*
+	 * If slack is negative, automatically set slack so that we load the
+	 * btree block approximately halfway between minrecs and maxrecs.
+	 * Generally, this will net us 75% loading.
+	 */
+	if (*slack < 0)
+		*slack = maxr - ((maxr + minr) >> 1);
+
+	*slack = min(*slack, maxr - minr);
+}
+
+/*
+ * Prepare a btree cursor for a bulk load operation by computing the geometry
+ * fields in bbl.  Caller must ensure that the btree cursor is a staging
+ * cursor.  This function can be called multiple times.
+ */
+int
+xfs_btree_bload_compute_geometry(
+	struct xfs_btree_cur	*cur,
+	struct xfs_btree_bload	*bbl,
+	uint64_t		nr_records)
+{
+	uint64_t		nr_blocks = 0;
+	uint64_t		nr_this_level;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+
+	/*
+	 * Make sure that the slack values make sense for traditional leaf and
+	 * node blocks.  Inode-rooted btrees will return different minrecs and
+	 * maxrecs values for the root block (bc_nlevels == level - 1).  We're
+	 * checking levels 0 and 1 here, so set bc_nlevels such that the btree
+	 * code doesn't interpret either as the root level.
+	 */
+	cur->bc_nlevels = XFS_BTREE_MAXLEVELS - 1;
+	xfs_btree_bload_ensure_slack(cur, &bbl->leaf_slack, 0);
+	xfs_btree_bload_ensure_slack(cur, &bbl->node_slack, 1);
+
+	bbl->nr_records = nr_this_level = nr_records;
+	for (cur->bc_nlevels = 1; cur->bc_nlevels < XFS_BTREE_MAXLEVELS;) {
+		uint64_t	level_blocks;
+		uint64_t	dontcare64;
+		unsigned int	level = cur->bc_nlevels - 1;
+		unsigned int	avg_per_block;
+
+		xfs_btree_bload_level_geometry(cur, bbl, level, nr_this_level,
+				&avg_per_block, &level_blocks, &dontcare64);
+
+		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
+			/*
+			 * If all the items we want to store at this level
+			 * would fit in the inode root block, then we have our
+			 * btree root and are done.
+			 *
+			 * Note that bmap btrees forbid records in the root.
+			 */
+			if (level != 0 && nr_this_level <= avg_per_block) {
+				nr_blocks++;
+				break;
+			}
+
+			/*
+			 * Otherwise, we have to store all the items for this
+			 * level in traditional btree blocks and therefore need
+			 * another level of btree to point to those blocks.
+			 *
+			 * We have to re-compute the geometry for each level of
+			 * an inode-rooted btree because the geometry differs
+			 * between a btree root in an inode fork and a
+			 * traditional btree block.
+			 *
+			 * This distinction is made in the btree code based on
+			 * whether level == bc_nlevels - 1.  Based on the
+			 * previous root block size check against the root
+			 * block geometry, we know that we aren't yet ready to
+			 * populate the root.  Increment bc_nevels and
+			 * recalculate the geometry for a traditional
+			 * block-based btree level.
+			 */
+			cur->bc_nlevels++;
+			xfs_btree_bload_level_geometry(cur, bbl, level,
+					nr_this_level, &avg_per_block,
+					&level_blocks, &dontcare64);
+		} else {
+			/*
+			 * If all the items we want to store at this level
+			 * would fit in a single root block, we're done.
+			 */
+			if (nr_this_level <= avg_per_block) {
+				nr_blocks++;
+				break;
+			}
+
+			/* Otherwise, we need another level of btree. */
+			cur->bc_nlevels++;
+		}
+
+		nr_blocks += level_blocks;
+		nr_this_level = level_blocks;
+	}
+
+	if (cur->bc_nlevels == XFS_BTREE_MAXLEVELS)
+		return -EOVERFLOW;
+
+	bbl->btree_height = cur->bc_nlevels;
+	if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+		bbl->nr_blocks = nr_blocks - 1;
+	else
+		bbl->nr_blocks = nr_blocks;
+	return 0;
+}
+
+/* Bulk load a btree given the parameters and geometry established in bbl. */
+int
+xfs_btree_bload(
+	struct xfs_btree_cur		*cur,
+	struct xfs_btree_bload		*bbl,
+	void				*priv)
+{
+	struct list_head		buffers_list;
+	union xfs_btree_ptr		child_ptr;
+	union xfs_btree_ptr		ptr;
+	struct xfs_buf			*bp = NULL;
+	struct xfs_btree_block		*block = NULL;
+	uint64_t			nr_this_level = bbl->nr_records;
+	uint64_t			blocks;
+	uint64_t			i;
+	uint64_t			blocks_with_extra;
+	uint64_t			total_blocks = 0;
+	unsigned int			avg_per_block;
+	unsigned int			level = 0;
+	int				ret;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+
+	INIT_LIST_HEAD(&buffers_list);
+	cur->bc_nlevels = bbl->btree_height;
+	xfs_btree_set_ptr_null(cur, &child_ptr);
+	xfs_btree_set_ptr_null(cur, &ptr);
+
+	xfs_btree_bload_level_geometry(cur, bbl, level, nr_this_level,
+			&avg_per_block, &blocks, &blocks_with_extra);
+
+	/* Load each leaf block. */
+	for (i = 0; i < blocks; i++) {
+		unsigned int		nr_this_block = avg_per_block;
+
+		/*
+		 * Due to rounding, btree blocks will not be evenly populated
+		 * in most cases.  blocks_with_extra tells us how many blocks
+		 * will receive an extra record to distribute the excess across
+		 * the current level as evenly as possible.
+		 */
+		if (i < blocks_with_extra)
+			nr_this_block++;
+
+		ret = xfs_btree_bload_prep_block(cur, bbl, &buffers_list, level,
+				nr_this_block, &ptr, &bp, &block, priv);
+		if (ret)
+			goto out;
+
+		trace_xfs_btree_bload_block(cur, level, i, blocks, &ptr,
+				nr_this_block);
+
+		ret = xfs_btree_bload_leaf(cur, nr_this_block, bbl->get_record,
+				block, priv);
+		if (ret)
+			goto out;
+
+		/*
+		 * Record the leftmost leaf pointer so we know where to start
+		 * with the first node level.
+		 */
+		if (i == 0)
+			xfs_btree_copy_ptrs(cur, &child_ptr, &ptr, 1);
+	}
+	total_blocks += blocks;
+	xfs_btree_bload_drop_buf(&buffers_list, &bp);
+
+	/* Populate the internal btree nodes. */
+	for (level = 1; level < cur->bc_nlevels; level++) {
+		union xfs_btree_ptr	first_ptr;
+
+		nr_this_level = blocks;
+		block = NULL;
+		xfs_btree_set_ptr_null(cur, &ptr);
+
+		xfs_btree_bload_level_geometry(cur, bbl, level, nr_this_level,
+				&avg_per_block, &blocks, &blocks_with_extra);
+
+		/* Load each node block. */
+		for (i = 0; i < blocks; i++) {
+			unsigned int	nr_this_block = avg_per_block;
+
+			if (i < blocks_with_extra)
+				nr_this_block++;
+
+			ret = xfs_btree_bload_prep_block(cur, bbl,
+					&buffers_list, level, nr_this_block,
+					&ptr, &bp, &block, priv);
+			if (ret)
+				goto out;
+
+			trace_xfs_btree_bload_block(cur, level, i, blocks,
+					&ptr, nr_this_block);
+
+			ret = xfs_btree_bload_node(cur, nr_this_block,
+					&child_ptr, block);
+			if (ret)
+				goto out;
+
+			/*
+			 * Record the leftmost node pointer so that we know
+			 * where to start the next node level above this one.
+			 */
+			if (i == 0)
+				xfs_btree_copy_ptrs(cur, &first_ptr, &ptr, 1);
+		}
+		total_blocks += blocks;
+		xfs_btree_bload_drop_buf(&buffers_list, &bp);
+		xfs_btree_copy_ptrs(cur, &child_ptr, &first_ptr, 1);
+	}
+
+	/* Initialize the new root. */
+	if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) {
+		ASSERT(xfs_btree_ptr_is_null(cur, &ptr));
+		cur->bc_ino.ifake->if_levels = cur->bc_nlevels;
+		cur->bc_ino.ifake->if_blocks = total_blocks - 1;
+	} else {
+		cur->bc_ag.afake->af_root = be32_to_cpu(ptr.s);
+		cur->bc_ag.afake->af_levels = cur->bc_nlevels;
+		cur->bc_ag.afake->af_blocks = total_blocks;
+	}
+
+	/*
+	 * Write the new blocks to disk.  If the ordered list isn't empty after
+	 * that, then something went wrong and we have to fail.  This should
+	 * never happen, but we'll check anyway.
+	 */
+	ret = xfs_buf_delwri_submit(&buffers_list);
+	if (ret)
+		goto out;
+	if (!list_empty(&buffers_list)) {
+		ASSERT(list_empty(&buffers_list));
+		ret = -EIO;
+	}
+
+out:
+	xfs_buf_delwri_cancel(&buffers_list);
+	if (bp)
+		xfs_buf_relse(bp);
+	return ret;
+}
diff --git a/fs/xfs/libxfs/xfs_btree_staging.h b/fs/xfs/libxfs/xfs_btree_staging.h
new file mode 100644
index 000000000000..f0d2976050ae
--- /dev/null
+++ b/fs/xfs/libxfs/xfs_btree_staging.h
@@ -0,0 +1,123 @@
+/* SPDX-License-Identifier: GPL-2.0-or-later */
+/*
+ * Copyright (C) 2020 Oracle.  All Rights Reserved.
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ */
+#ifndef __XFS_BTREE_STAGING_H__
+#define __XFS_BTREE_STAGING_H__
+
+/* Fake root for an AG-rooted btree. */
+struct xbtree_afakeroot {
+	/* AG block number of the new btree root. */
+	xfs_agblock_t		af_root;
+
+	/* Height of the new btree. */
+	unsigned int		af_levels;
+
+	/* Number of blocks used by the btree. */
+	unsigned int		af_blocks;
+};
+
+/* Cursor interactions with fake roots for AG-rooted btrees. */
+void xfs_btree_stage_afakeroot(struct xfs_btree_cur *cur,
+		struct xbtree_afakeroot *afake);
+void xfs_btree_commit_afakeroot(struct xfs_btree_cur *cur, struct xfs_trans *tp,
+		struct xfs_buf *agbp, const struct xfs_btree_ops *ops);
+
+/* Fake root for an inode-rooted btree. */
+struct xbtree_ifakeroot {
+	/* Fake inode fork. */
+	struct xfs_ifork	*if_fork;
+
+	/* Number of blocks used by the btree. */
+	int64_t			if_blocks;
+
+	/* Height of the new btree. */
+	unsigned int		if_levels;
+
+	/* Number of bytes available for this fork in the inode. */
+	unsigned int		if_fork_size;
+
+	/* Fork format. */
+	unsigned int		if_format;
+
+	/* Number of records. */
+	unsigned int		if_extents;
+};
+
+/* Cursor interactions with fake roots for inode-rooted btrees. */
+void xfs_btree_stage_ifakeroot(struct xfs_btree_cur *cur,
+		struct xbtree_ifakeroot *ifake,
+		struct xfs_btree_ops **new_ops);
+void xfs_btree_commit_ifakeroot(struct xfs_btree_cur *cur, struct xfs_trans *tp,
+		int whichfork, const struct xfs_btree_ops *ops);
+
+/* Bulk loading of staged btrees. */
+typedef int (*xfs_btree_bload_get_record_fn)(struct xfs_btree_cur *cur, void *priv);
+typedef int (*xfs_btree_bload_claim_block_fn)(struct xfs_btree_cur *cur,
+		union xfs_btree_ptr *ptr, void *priv);
+typedef size_t (*xfs_btree_bload_iroot_size_fn)(struct xfs_btree_cur *cur,
+		unsigned int nr_this_level, void *priv);
+
+struct xfs_btree_bload {
+	/*
+	 * This function will be called nr_records times to load records into
+	 * the btree.  The function does this by setting the cursor's bc_rec
+	 * field in in-core format.  Records must be returned in sort order.
+	 */
+	xfs_btree_bload_get_record_fn	get_record;
+
+	/*
+	 * This function will be called nr_blocks times to obtain a pointer
+	 * to a new btree block on disk.  Callers must preallocate all space
+	 * for the new btree before calling xfs_btree_bload, and this function
+	 * is what claims that reservation.
+	 */
+	xfs_btree_bload_claim_block_fn	claim_block;
+
+	/*
+	 * This function should return the size of the in-core btree root
+	 * block.  It is only necessary for XFS_BTREE_ROOT_IN_INODE btree
+	 * types.
+	 */
+	xfs_btree_bload_iroot_size_fn	iroot_size;
+
+	/*
+	 * The caller should set this to the number of records that will be
+	 * stored in the new btree.
+	 */
+	uint64_t			nr_records;
+
+	/*
+	 * Number of free records to leave in each leaf block.  If the caller
+	 * sets this to -1, the slack value will be calculated to be halfway
+	 * between maxrecs and minrecs.  This typically leaves the block 75%
+	 * full.  Note that slack values are not enforced on inode root blocks.
+	 */
+	int				leaf_slack;
+
+	/*
+	 * Number of free key/ptrs pairs to leave in each node block.  This
+	 * field has the same semantics as leaf_slack.
+	 */
+	int				node_slack;
+
+	/*
+	 * The xfs_btree_bload_compute_geometry function will set this to the
+	 * number of btree blocks needed to store nr_records records.
+	 */
+	uint64_t			nr_blocks;
+
+	/*
+	 * The xfs_btree_bload_compute_geometry function will set this to the
+	 * height of the new btree.
+	 */
+	unsigned int			btree_height;
+};
+
+int xfs_btree_bload_compute_geometry(struct xfs_btree_cur *cur,
+		struct xfs_btree_bload *bbl, uint64_t nr_records);
+int xfs_btree_bload(struct xfs_btree_cur *cur, struct xfs_btree_bload *bbl,
+		void *priv);
+
+#endif	/* __XFS_BTREE_STAGING_H__ */
diff --git a/fs/xfs/libxfs/xfs_da_btree.c b/fs/xfs/libxfs/xfs_da_btree.c
index 875e04f82541..e46bc03365db 100644
--- a/fs/xfs/libxfs/xfs_da_btree.c
+++ b/fs/xfs/libxfs/xfs_da_btree.c
@@ -78,10 +78,16 @@ kmem_zone_t *xfs_da_state_zone;	/* anchor for state struct zone */
  * Allocate a dir-state structure.
  * We don't put them on the stack since they're large.
  */
-xfs_da_state_t *
-xfs_da_state_alloc(void)
+struct xfs_da_state *
+xfs_da_state_alloc(
+	struct xfs_da_args	*args)
 {
-	return kmem_zone_zalloc(xfs_da_state_zone, KM_NOFS);
+	struct xfs_da_state	*state;
+
+	state = kmem_cache_zalloc(xfs_da_state_zone, GFP_NOFS | __GFP_NOFAIL);
+	state->args = args;
+	state->mp = args->dp->i_mount;
+	return state;
 }
 
 /*
@@ -590,7 +596,7 @@ xfs_da3_split(
 	node = oldblk->bp->b_addr;
 	if (node->hdr.info.forw) {
 		if (be32_to_cpu(node->hdr.info.forw) != addblk->blkno) {
-			xfs_buf_corruption_error(oldblk->bp);
+			xfs_buf_mark_corrupt(oldblk->bp);
 			error = -EFSCORRUPTED;
 			goto out;
 		}
@@ -603,7 +609,7 @@ xfs_da3_split(
 	node = oldblk->bp->b_addr;
 	if (node->hdr.info.back) {
 		if (be32_to_cpu(node->hdr.info.back) != addblk->blkno) {
-			xfs_buf_corruption_error(oldblk->bp);
+			xfs_buf_mark_corrupt(oldblk->bp);
 			error = -EFSCORRUPTED;
 			goto out;
 		}
@@ -1624,7 +1630,7 @@ xfs_da3_node_lookup_int(
 		}
 
 		if (magic != XFS_DA_NODE_MAGIC && magic != XFS_DA3_NODE_MAGIC) {
-			xfs_buf_corruption_error(blk->bp);
+			xfs_buf_mark_corrupt(blk->bp);
 			return -EFSCORRUPTED;
 		}
 
@@ -1639,7 +1645,7 @@ xfs_da3_node_lookup_int(
 
 		/* Tree taller than we can handle; bail out! */
 		if (nodehdr.level >= XFS_DA_NODE_MAXDEPTH) {
-			xfs_buf_corruption_error(blk->bp);
+			xfs_buf_mark_corrupt(blk->bp);
 			return -EFSCORRUPTED;
 		}
 
@@ -1647,7 +1653,7 @@ xfs_da3_node_lookup_int(
 		if (blkno == args->geo->leafblk)
 			expected_level = nodehdr.level - 1;
 		else if (expected_level != nodehdr.level) {
-			xfs_buf_corruption_error(blk->bp);
+			xfs_buf_mark_corrupt(blk->bp);
 			return -EFSCORRUPTED;
 		} else
 			expected_level--;
@@ -1986,7 +1992,8 @@ xfs_da3_path_shift(
 	ASSERT(path != NULL);
 	ASSERT((path->active > 0) && (path->active < XFS_DA_NODE_MAXDEPTH));
 	level = (path->active-1) - 1;	/* skip bottom layer in path */
-	for (blk = &path->blk[level]; level >= 0; blk--, level--) {
+	for (; level >= 0; level--) {
+		blk = &path->blk[level];
 		xfs_da3_node_hdr_from_disk(dp->i_mount, &nodehdr,
 					   blk->bp->b_addr);
 
@@ -2520,8 +2527,10 @@ xfs_dabuf_map(
 	 */
 	if (nirecs > 1) {
 		map = kmem_zalloc(nirecs * sizeof(struct xfs_buf_map), KM_NOFS);
-		if (!map)
+		if (!map) {
+			error = -ENOMEM;
 			goto out_free_irecs;
+		}
 		*mapp = map;
 	}
 
diff --git a/fs/xfs/libxfs/xfs_da_btree.h b/fs/xfs/libxfs/xfs_da_btree.h
index 0f4fbb0889ff..ad5dd324631a 100644
--- a/fs/xfs/libxfs/xfs_da_btree.h
+++ b/fs/xfs/libxfs/xfs_da_btree.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000,2002,2005 Silicon Graphics, Inc.
  * Copyright (c) 2013 Red Hat, Inc.
@@ -57,9 +57,10 @@ typedef struct xfs_da_args {
 	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 */
-	uint8_t		*value;		/* set of bytes (maybe contain NULLs) */
+	void		*value;		/* set of bytes (maybe contain NULLs) */
 	int		valuelen;	/* length of value */
-	int		flags;		/* argument flags (eg: ATTR_NOCREATE) */
+	unsigned int	attr_filter;	/* XFS_ATTR_{ROOT,SECURE,INCOMPLETE} */
+	unsigned int	attr_flags;	/* XATTR_{CREATE,REPLACE} */
 	xfs_dahash_t	hashval;	/* hash value of name */
 	xfs_ino_t	inumber;	/* input/output inode number */
 	struct xfs_inode *dp;		/* directory inode to manipulate */
@@ -88,8 +89,7 @@ typedef struct xfs_da_args {
 #define XFS_DA_OP_ADDNAME	0x0004	/* this is an add operation */
 #define XFS_DA_OP_OKNOENT	0x0008	/* lookup/add op, ENOENT ok, else die */
 #define XFS_DA_OP_CILOOKUP	0x0010	/* lookup to return CI name if found */
-#define XFS_DA_OP_ALLOCVAL	0x0020	/* lookup to alloc buffer if found  */
-#define XFS_DA_OP_INCOMPLETE	0x0040	/* lookup INCOMPLETE attr keys */
+#define XFS_DA_OP_NOTIME	0x0020	/* don't update inode timestamps */
 
 #define XFS_DA_OP_FLAGS \
 	{ XFS_DA_OP_JUSTCHECK,	"JUSTCHECK" }, \
@@ -97,8 +97,7 @@ typedef struct xfs_da_args {
 	{ XFS_DA_OP_ADDNAME,	"ADDNAME" }, \
 	{ XFS_DA_OP_OKNOENT,	"OKNOENT" }, \
 	{ XFS_DA_OP_CILOOKUP,	"CILOOKUP" }, \
-	{ XFS_DA_OP_ALLOCVAL,	"ALLOCVAL" }, \
-	{ XFS_DA_OP_INCOMPLETE,	"INCOMPLETE" }
+	{ XFS_DA_OP_NOTIME,	"NOTIME" }
 
 /*
  * Storage for holding state during Btree searches and split/join ops.
@@ -220,7 +219,7 @@ enum xfs_dacmp xfs_da_compname(struct xfs_da_args *args,
 				const unsigned char *name, int len);
 
 
-xfs_da_state_t *xfs_da_state_alloc(void);
+struct xfs_da_state *xfs_da_state_alloc(struct xfs_da_args *args);
 void xfs_da_state_free(xfs_da_state_t *state);
 
 void	xfs_da3_node_hdr_from_disk(struct xfs_mount *mp,
diff --git a/fs/xfs/libxfs/xfs_da_format.h b/fs/xfs/libxfs/xfs_da_format.h
index 734837a9b51a..059ac108b1b3 100644
--- a/fs/xfs/libxfs/xfs_da_format.h
+++ b/fs/xfs/libxfs/xfs_da_format.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
  * Copyright (c) 2013 Red Hat, Inc.
@@ -692,19 +692,7 @@ struct xfs_attr3_leafblock {
 #define XFS_ATTR_ROOT		(1 << XFS_ATTR_ROOT_BIT)
 #define XFS_ATTR_SECURE		(1 << XFS_ATTR_SECURE_BIT)
 #define XFS_ATTR_INCOMPLETE	(1 << XFS_ATTR_INCOMPLETE_BIT)
-
-/*
- * Conversion macros for converting namespace bits from argument flags
- * to ondisk flags.
- */
-#define XFS_ATTR_NSP_ARGS_MASK		(ATTR_ROOT | ATTR_SECURE)
 #define XFS_ATTR_NSP_ONDISK_MASK	(XFS_ATTR_ROOT | XFS_ATTR_SECURE)
-#define XFS_ATTR_NSP_ONDISK(flags)	((flags) & XFS_ATTR_NSP_ONDISK_MASK)
-#define XFS_ATTR_NSP_ARGS(flags)	((flags) & XFS_ATTR_NSP_ARGS_MASK)
-#define XFS_ATTR_NSP_ARGS_TO_ONDISK(x)	(((x) & ATTR_ROOT ? XFS_ATTR_ROOT : 0) |\
-					 ((x) & ATTR_SECURE ? XFS_ATTR_SECURE : 0))
-#define XFS_ATTR_NSP_ONDISK_TO_ARGS(x)	(((x) & XFS_ATTR_ROOT ? ATTR_ROOT : 0) |\
-					 ((x) & XFS_ATTR_SECURE ? ATTR_SECURE : 0))
 
 /*
  * Alignment for namelist and valuelist entries (since they are mixed
diff --git a/fs/xfs/libxfs/xfs_defer.c b/fs/xfs/libxfs/xfs_defer.c
index 22557527cfdb..d8f586256add 100644
--- a/fs/xfs/libxfs/xfs_defer.c
+++ b/fs/xfs/libxfs/xfs_defer.c
@@ -178,6 +178,18 @@ static const struct xfs_defer_op_type *defer_op_types[] = {
 	[XFS_DEFER_OPS_TYPE_AGFL_FREE]	= &xfs_agfl_free_defer_type,
 };
 
+static void
+xfs_defer_create_intent(
+	struct xfs_trans		*tp,
+	struct xfs_defer_pending	*dfp,
+	bool				sort)
+{
+	const struct xfs_defer_op_type	*ops = defer_op_types[dfp->dfp_type];
+
+	dfp->dfp_intent = ops->create_intent(tp, &dfp->dfp_work,
+			dfp->dfp_count, sort);
+}
+
 /*
  * For each pending item in the intake list, log its intent item and the
  * associated extents, then add the entire intake list to the end of
@@ -187,17 +199,11 @@ STATIC void
 xfs_defer_create_intents(
 	struct xfs_trans		*tp)
 {
-	struct list_head		*li;
 	struct xfs_defer_pending	*dfp;
-	const struct xfs_defer_op_type	*ops;
 
 	list_for_each_entry(dfp, &tp->t_dfops, dfp_list) {
-		ops = defer_op_types[dfp->dfp_type];
-		dfp->dfp_intent = ops->create_intent(tp, dfp->dfp_count);
 		trace_xfs_defer_create_intent(tp->t_mountp, dfp);
-		list_sort(tp->t_mountp, &dfp->dfp_work, ops->diff_items);
-		list_for_each(li, &dfp->dfp_work)
-			ops->log_item(tp, dfp->dfp_intent, li);
+		xfs_defer_create_intent(tp, dfp, true);
 	}
 }
 
@@ -234,10 +240,13 @@ xfs_defer_trans_roll(
 	struct xfs_log_item		*lip;
 	struct xfs_buf			*bplist[XFS_DEFER_OPS_NR_BUFS];
 	struct xfs_inode		*iplist[XFS_DEFER_OPS_NR_INODES];
+	unsigned int			ordered = 0; /* bitmap */
 	int				bpcount = 0, ipcount = 0;
 	int				i;
 	int				error;
 
+	BUILD_BUG_ON(NBBY * sizeof(ordered) < XFS_DEFER_OPS_NR_BUFS);
+
 	list_for_each_entry(lip, &tp->t_items, li_trans) {
 		switch (lip->li_type) {
 		case XFS_LI_BUF:
@@ -248,7 +257,10 @@ xfs_defer_trans_roll(
 					ASSERT(0);
 					return -EFSCORRUPTED;
 				}
-				xfs_trans_dirty_buf(tp, bli->bli_buf);
+				if (bli->bli_flags & XFS_BLI_ORDERED)
+					ordered |= (1U << bpcount);
+				else
+					xfs_trans_dirty_buf(tp, bli->bli_buf);
 				bplist[bpcount++] = bli->bli_buf;
 			}
 			break;
@@ -289,6 +301,8 @@ xfs_defer_trans_roll(
 	/* Rejoin the buffers and dirty them so the log moves forward. */
 	for (i = 0; i < bpcount; i++) {
 		xfs_trans_bjoin(tp, bplist[i]);
+		if (ordered & (1U << i))
+			xfs_trans_ordered_buf(tp, bplist[i]);
 		xfs_trans_bhold(tp, bplist[i]);
 	}
 
@@ -346,6 +360,53 @@ xfs_defer_cancel_list(
 }
 
 /*
+ * Log an intent-done item for the first pending intent, and finish the work
+ * items.
+ */
+static int
+xfs_defer_finish_one(
+	struct xfs_trans		*tp,
+	struct xfs_defer_pending	*dfp)
+{
+	const struct xfs_defer_op_type	*ops = defer_op_types[dfp->dfp_type];
+	struct xfs_btree_cur		*state = NULL;
+	struct list_head		*li, *n;
+	int				error;
+
+	trace_xfs_defer_pending_finish(tp->t_mountp, dfp);
+
+	dfp->dfp_done = ops->create_done(tp, dfp->dfp_intent, dfp->dfp_count);
+	list_for_each_safe(li, n, &dfp->dfp_work) {
+		list_del(li);
+		dfp->dfp_count--;
+		error = ops->finish_item(tp, dfp->dfp_done, li, &state);
+		if (error == -EAGAIN) {
+			/*
+			 * Caller wants a fresh transaction; put the work item
+			 * back on the list and log a new log intent item to
+			 * replace the old one.  See "Requesting a Fresh
+			 * Transaction while Finishing Deferred Work" above.
+			 */
+			list_add(li, &dfp->dfp_work);
+			dfp->dfp_count++;
+			dfp->dfp_done = NULL;
+			xfs_defer_create_intent(tp, dfp, false);
+		}
+
+		if (error)
+			goto out;
+	}
+
+	/* Done with the dfp, free it. */
+	list_del(&dfp->dfp_list);
+	kmem_free(dfp);
+out:
+	if (ops->finish_cleanup)
+		ops->finish_cleanup(tp, state, error);
+	return error;
+}
+
+/*
  * Finish all the pending work.  This involves logging intent items for
  * any work items that wandered in since the last transaction roll (if
  * one has even happened), rolling the transaction, and finishing the
@@ -358,11 +419,7 @@ xfs_defer_finish_noroll(
 	struct xfs_trans		**tp)
 {
 	struct xfs_defer_pending	*dfp;
-	struct list_head		*li;
-	struct list_head		*n;
-	void				*state;
 	int				error = 0;
-	const struct xfs_defer_op_type	*ops;
 	LIST_HEAD(dop_pending);
 
 	ASSERT((*tp)->t_flags & XFS_TRANS_PERM_LOG_RES);
@@ -371,87 +428,30 @@ xfs_defer_finish_noroll(
 
 	/* Until we run out of pending work to finish... */
 	while (!list_empty(&dop_pending) || !list_empty(&(*tp)->t_dfops)) {
-		/* log intents and pull in intake items */
 		xfs_defer_create_intents(*tp);
 		list_splice_tail_init(&(*tp)->t_dfops, &dop_pending);
 
-		/*
-		 * Roll the transaction.
-		 */
 		error = xfs_defer_trans_roll(tp);
 		if (error)
-			goto out;
+			goto out_shutdown;
 
-		/* Log an intent-done item for the first pending item. */
 		dfp = list_first_entry(&dop_pending, struct xfs_defer_pending,
 				       dfp_list);
-		ops = defer_op_types[dfp->dfp_type];
-		trace_xfs_defer_pending_finish((*tp)->t_mountp, dfp);
-		dfp->dfp_done = ops->create_done(*tp, dfp->dfp_intent,
-				dfp->dfp_count);
-
-		/* Finish the work items. */
-		state = NULL;
-		list_for_each_safe(li, n, &dfp->dfp_work) {
-			list_del(li);
-			dfp->dfp_count--;
-			error = ops->finish_item(*tp, li, dfp->dfp_done,
-					&state);
-			if (error == -EAGAIN) {
-				/*
-				 * Caller wants a fresh transaction;
-				 * put the work item back on the list
-				 * and jump out.
-				 */
-				list_add(li, &dfp->dfp_work);
-				dfp->dfp_count++;
-				break;
-			} else if (error) {
-				/*
-				 * Clean up after ourselves and jump out.
-				 * xfs_defer_cancel will take care of freeing
-				 * all these lists and stuff.
-				 */
-				if (ops->finish_cleanup)
-					ops->finish_cleanup(*tp, state, error);
-				goto out;
-			}
-		}
-		if (error == -EAGAIN) {
-			/*
-			 * Caller wants a fresh transaction, so log a
-			 * new log intent item to replace the old one
-			 * and roll the transaction.  See "Requesting
-			 * a Fresh Transaction while Finishing
-			 * Deferred Work" above.
-			 */
-			dfp->dfp_intent = ops->create_intent(*tp,
-					dfp->dfp_count);
-			dfp->dfp_done = NULL;
-			list_for_each(li, &dfp->dfp_work)
-				ops->log_item(*tp, dfp->dfp_intent, li);
-		} else {
-			/* Done with the dfp, free it. */
-			list_del(&dfp->dfp_list);
-			kmem_free(dfp);
-		}
-
-		if (ops->finish_cleanup)
-			ops->finish_cleanup(*tp, state, error);
-	}
-
-out:
-	if (error) {
-		xfs_defer_trans_abort(*tp, &dop_pending);
-		xfs_force_shutdown((*tp)->t_mountp, SHUTDOWN_CORRUPT_INCORE);
-		trace_xfs_defer_finish_error(*tp, error);
-		xfs_defer_cancel_list((*tp)->t_mountp, &dop_pending);
-		xfs_defer_cancel(*tp);
-		return error;
+		error = xfs_defer_finish_one(*tp, dfp);
+		if (error && error != -EAGAIN)
+			goto out_shutdown;
 	}
 
 	trace_xfs_defer_finish_done(*tp, _RET_IP_);
 	return 0;
+
+out_shutdown:
+	xfs_defer_trans_abort(*tp, &dop_pending);
+	xfs_force_shutdown((*tp)->t_mountp, SHUTDOWN_CORRUPT_INCORE);
+	trace_xfs_defer_finish_error(*tp, error);
+	xfs_defer_cancel_list((*tp)->t_mountp, &dop_pending);
+	xfs_defer_cancel(*tp);
+	return error;
 }
 
 int
diff --git a/fs/xfs/libxfs/xfs_defer.h b/fs/xfs/libxfs/xfs_defer.h
index 7c28d7608ac6..6b2ca580f2b0 100644
--- a/fs/xfs/libxfs/xfs_defer.h
+++ b/fs/xfs/libxfs/xfs_defer.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0+
+/* SPDX-License-Identifier: GPL-2.0+ */
 /*
  * Copyright (C) 2016 Oracle.  All Rights Reserved.
  * Author: Darrick J. Wong <darrick.wong@oracle.com>
@@ -6,6 +6,7 @@
 #ifndef __XFS_DEFER_H__
 #define	__XFS_DEFER_H__
 
+struct xfs_btree_cur;
 struct xfs_defer_op_type;
 
 /*
@@ -28,8 +29,8 @@ enum xfs_defer_ops_type {
 struct xfs_defer_pending {
 	struct list_head		dfp_list;	/* pending items */
 	struct list_head		dfp_work;	/* work items */
-	void				*dfp_intent;	/* log intent item */
-	void				*dfp_done;	/* log done item */
+	struct xfs_log_item		*dfp_intent;	/* log intent item */
+	struct xfs_log_item		*dfp_done;	/* log done item */
 	unsigned int			dfp_count;	/* # extent items */
 	enum xfs_defer_ops_type		dfp_type;
 };
@@ -43,15 +44,16 @@ void xfs_defer_move(struct xfs_trans *dtp, struct xfs_trans *stp);
 
 /* Description of a deferred type. */
 struct xfs_defer_op_type {
-	void (*abort_intent)(void *);
-	void *(*create_done)(struct xfs_trans *, void *, unsigned int);
-	int (*finish_item)(struct xfs_trans *, struct list_head *, void *,
-			void **);
-	void (*finish_cleanup)(struct xfs_trans *, void *, int);
-	void (*cancel_item)(struct list_head *);
-	int (*diff_items)(void *, struct list_head *, struct list_head *);
-	void *(*create_intent)(struct xfs_trans *, uint);
-	void (*log_item)(struct xfs_trans *, void *, struct list_head *);
+	struct xfs_log_item *(*create_intent)(struct xfs_trans *tp,
+			struct list_head *items, unsigned int count, bool sort);
+	void (*abort_intent)(struct xfs_log_item *intent);
+	struct xfs_log_item *(*create_done)(struct xfs_trans *tp,
+			struct xfs_log_item *intent, unsigned int count);
+	int (*finish_item)(struct xfs_trans *tp, struct xfs_log_item *done,
+			struct list_head *item, struct xfs_btree_cur **state);
+	void (*finish_cleanup)(struct xfs_trans *tp,
+			struct xfs_btree_cur *state, int error);
+	void (*cancel_item)(struct list_head *item);
 	unsigned int		max_items;
 };
 
diff --git a/fs/xfs/libxfs/xfs_dir2.c b/fs/xfs/libxfs/xfs_dir2.c
index dd6fcaaea318..612a9c5e41b1 100644
--- a/fs/xfs/libxfs/xfs_dir2.c
+++ b/fs/xfs/libxfs/xfs_dir2.c
@@ -278,7 +278,7 @@ xfs_dir_createname(
 	if (!inum)
 		args->op_flags |= XFS_DA_OP_JUSTCHECK;
 
-	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) {
+	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
 		rval = xfs_dir2_sf_addname(args);
 		goto out_free;
 	}
@@ -373,7 +373,7 @@ xfs_dir_lookup(
 		args->op_flags |= XFS_DA_OP_CILOOKUP;
 
 	lock_mode = xfs_ilock_data_map_shared(dp);
-	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) {
+	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
 		rval = xfs_dir2_sf_lookup(args);
 		goto out_check_rval;
 	}
@@ -443,7 +443,7 @@ xfs_dir_removename(
 	args->whichfork = XFS_DATA_FORK;
 	args->trans = tp;
 
-	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) {
+	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
 		rval = xfs_dir2_sf_removename(args);
 		goto out_free;
 	}
@@ -504,7 +504,7 @@ xfs_dir_replace(
 	args->whichfork = XFS_DATA_FORK;
 	args->trans = tp;
 
-	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL) {
+	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
 		rval = xfs_dir2_sf_replace(args);
 		goto out_free;
 	}
diff --git a/fs/xfs/libxfs/xfs_dir2.h b/fs/xfs/libxfs/xfs_dir2.h
index 033777e282f2..e55378640b05 100644
--- a/fs/xfs/libxfs/xfs_dir2.h
+++ b/fs/xfs/libxfs/xfs_dir2.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
  * All Rights Reserved.
diff --git a/fs/xfs/libxfs/xfs_dir2_block.c b/fs/xfs/libxfs/xfs_dir2_block.c
index d6ced59b9567..5b59d3f7746b 100644
--- a/fs/xfs/libxfs/xfs_dir2_block.c
+++ b/fs/xfs/libxfs/xfs_dir2_block.c
@@ -114,6 +114,23 @@ const struct xfs_buf_ops xfs_dir3_block_buf_ops = {
 	.verify_struct = xfs_dir3_block_verify,
 };
 
+static xfs_failaddr_t
+xfs_dir3_block_header_check(
+	struct xfs_inode	*dp,
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = dp->i_mount;
+
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		struct xfs_dir3_blk_hdr *hdr3 = bp->b_addr;
+
+		if (be64_to_cpu(hdr3->owner) != dp->i_ino)
+			return __this_address;
+	}
+
+	return NULL;
+}
+
 int
 xfs_dir3_block_read(
 	struct xfs_trans	*tp,
@@ -121,12 +138,24 @@ xfs_dir3_block_read(
 	struct xfs_buf		**bpp)
 {
 	struct xfs_mount	*mp = dp->i_mount;
+	xfs_failaddr_t		fa;
 	int			err;
 
 	err = xfs_da_read_buf(tp, dp, mp->m_dir_geo->datablk, 0, bpp,
 				XFS_DATA_FORK, &xfs_dir3_block_buf_ops);
-	if (!err && tp && *bpp)
-		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_BLOCK_BUF);
+	if (err || !*bpp)
+		return err;
+
+	/* Check things that we can't do in the verifier. */
+	fa = xfs_dir3_block_header_check(dp, *bpp);
+	if (fa) {
+		__xfs_buf_mark_corrupt(*bpp, fa);
+		xfs_trans_brelse(tp, *bpp);
+		*bpp = NULL;
+		return -EFSCORRUPTED;
+	}
+
+	xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_BLOCK_BUF);
 	return err;
 }
 
@@ -1075,7 +1104,7 @@ xfs_dir2_sf_to_block(
 	ASSERT(ifp->if_bytes == dp->i_d.di_size);
 	ASSERT(ifp->if_u1.if_data != NULL);
 	ASSERT(dp->i_d.di_size >= xfs_dir2_sf_hdr_size(oldsfp->i8count));
-	ASSERT(dp->i_d.di_nextents == 0);
+	ASSERT(dp->i_df.if_nextents == 0);
 
 	/*
 	 * Copy the directory into a temporary buffer.
diff --git a/fs/xfs/libxfs/xfs_dir2_data.c b/fs/xfs/libxfs/xfs_dir2_data.c
index b9eba8213180..375b3edb2ad2 100644
--- a/fs/xfs/libxfs/xfs_dir2_data.c
+++ b/fs/xfs/libxfs/xfs_dir2_data.c
@@ -394,6 +394,22 @@ static const struct xfs_buf_ops xfs_dir3_data_reada_buf_ops = {
 	.verify_write = xfs_dir3_data_write_verify,
 };
 
+static xfs_failaddr_t
+xfs_dir3_data_header_check(
+	struct xfs_inode	*dp,
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = dp->i_mount;
+
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		struct xfs_dir3_data_hdr *hdr3 = bp->b_addr;
+
+		if (be64_to_cpu(hdr3->hdr.owner) != dp->i_ino)
+			return __this_address;
+	}
+
+	return NULL;
+}
 
 int
 xfs_dir3_data_read(
@@ -403,12 +419,24 @@ xfs_dir3_data_read(
 	unsigned int		flags,
 	struct xfs_buf		**bpp)
 {
+	xfs_failaddr_t		fa;
 	int			err;
 
 	err = xfs_da_read_buf(tp, dp, bno, flags, bpp, XFS_DATA_FORK,
 			&xfs_dir3_data_buf_ops);
-	if (!err && tp && *bpp)
-		xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_DATA_BUF);
+	if (err || !*bpp)
+		return err;
+
+	/* Check things that we can't do in the verifier. */
+	fa = xfs_dir3_data_header_check(dp, *bpp);
+	if (fa) {
+		__xfs_buf_mark_corrupt(*bpp, fa);
+		xfs_trans_brelse(tp, *bpp);
+		*bpp = NULL;
+		return -EFSCORRUPTED;
+	}
+
+	xfs_trans_buf_set_type(tp, *bpp, XFS_BLFT_DIR_DATA_BUF);
 	return err;
 }
 
diff --git a/fs/xfs/libxfs/xfs_dir2_leaf.c b/fs/xfs/libxfs/xfs_dir2_leaf.c
index a131b520aac7..95d2a3f92d75 100644
--- a/fs/xfs/libxfs/xfs_dir2_leaf.c
+++ b/fs/xfs/libxfs/xfs_dir2_leaf.c
@@ -1383,7 +1383,7 @@ xfs_dir2_leaf_removename(
 	ltp = xfs_dir2_leaf_tail_p(geo, leaf);
 	bestsp = xfs_dir2_leaf_bests_p(ltp);
 	if (be16_to_cpu(bestsp[db]) != oldbest) {
-		xfs_buf_corruption_error(lbp);
+		xfs_buf_mark_corrupt(lbp);
 		return -EFSCORRUPTED;
 	}
 	/*
diff --git a/fs/xfs/libxfs/xfs_dir2_node.c b/fs/xfs/libxfs/xfs_dir2_node.c
index a0cc5e240306..5d51265d29d6 100644
--- a/fs/xfs/libxfs/xfs_dir2_node.c
+++ b/fs/xfs/libxfs/xfs_dir2_node.c
@@ -194,6 +194,8 @@ xfs_dir3_free_header_check(
 			return __this_address;
 		if (be32_to_cpu(hdr3->nvalid) < be32_to_cpu(hdr3->nused))
 			return __this_address;
+		if (be64_to_cpu(hdr3->hdr.owner) != dp->i_ino)
+			return __this_address;
 	} else {
 		struct xfs_dir2_free_hdr *hdr = bp->b_addr;
 
@@ -226,8 +228,9 @@ __xfs_dir3_free_read(
 	/* Check things that we can't do in the verifier. */
 	fa = xfs_dir3_free_header_check(dp, fbno, *bpp);
 	if (fa) {
-		xfs_verifier_error(*bpp, -EFSCORRUPTED, fa);
+		__xfs_buf_mark_corrupt(*bpp, fa);
 		xfs_trans_brelse(tp, *bpp);
+		*bpp = NULL;
 		return -EFSCORRUPTED;
 	}
 
@@ -439,7 +442,7 @@ xfs_dir2_leaf_to_node(
 	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 	if (be32_to_cpu(ltp->bestcount) >
 				(uint)dp->i_d.di_size / args->geo->blksize) {
-		xfs_buf_corruption_error(lbp);
+		xfs_buf_mark_corrupt(lbp);
 		return -EFSCORRUPTED;
 	}
 
@@ -513,7 +516,7 @@ xfs_dir2_leafn_add(
 	 * into other peoples memory
 	 */
 	if (index < 0) {
-		xfs_buf_corruption_error(bp);
+		xfs_buf_mark_corrupt(bp);
 		return -EFSCORRUPTED;
 	}
 
@@ -800,7 +803,7 @@ xfs_dir2_leafn_lookup_for_entry(
 
 	xfs_dir3_leaf_check(dp, bp);
 	if (leafhdr.count <= 0) {
-		xfs_buf_corruption_error(bp);
+		xfs_buf_mark_corrupt(bp);
 		return -EFSCORRUPTED;
 	}
 
@@ -2012,9 +2015,7 @@ xfs_dir2_node_addname(
 	/*
 	 * Allocate and initialize the state (btree cursor).
 	 */
-	state = xfs_da_state_alloc();
-	state->args = args;
-	state->mp = args->dp->i_mount;
+	state = xfs_da_state_alloc(args);
 	/*
 	 * Look up the name.  We're not supposed to find it, but
 	 * this gives us the insertion point.
@@ -2083,9 +2084,8 @@ xfs_dir2_node_lookup(
 	/*
 	 * Allocate and initialize the btree cursor.
 	 */
-	state = xfs_da_state_alloc();
-	state->args = args;
-	state->mp = args->dp->i_mount;
+	state = xfs_da_state_alloc(args);
+
 	/*
 	 * Fill in the path to the entry in the cursor.
 	 */
@@ -2136,9 +2136,7 @@ xfs_dir2_node_removename(
 	/*
 	 * Allocate and initialize the btree cursor.
 	 */
-	state = xfs_da_state_alloc();
-	state->args = args;
-	state->mp = args->dp->i_mount;
+	state = xfs_da_state_alloc(args);
 
 	/* Look up the entry we're deleting, set up the cursor. */
 	error = xfs_da3_node_lookup_int(state, &rval);
@@ -2203,9 +2201,7 @@ xfs_dir2_node_replace(
 	/*
 	 * Allocate and initialize the btree cursor.
 	 */
-	state = xfs_da_state_alloc();
-	state->args = args;
-	state->mp = args->dp->i_mount;
+	state = xfs_da_state_alloc(args);
 
 	/*
 	 * We have to save new inode number and ftype since
diff --git a/fs/xfs/libxfs/xfs_dir2_priv.h b/fs/xfs/libxfs/xfs_dir2_priv.h
index 01ee0b926572..44c6a77cba05 100644
--- a/fs/xfs/libxfs/xfs_dir2_priv.h
+++ b/fs/xfs/libxfs/xfs_dir2_priv.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000-2001,2005 Silicon Graphics, Inc.
  * All Rights Reserved.
diff --git a/fs/xfs/libxfs/xfs_dir2_sf.c b/fs/xfs/libxfs/xfs_dir2_sf.c
index 7b7f6fb2ea3b..2463b5d73447 100644
--- a/fs/xfs/libxfs/xfs_dir2_sf.c
+++ b/fs/xfs/libxfs/xfs_dir2_sf.c
@@ -343,7 +343,7 @@ xfs_dir2_block_to_sf(
 	 */
 	ASSERT(dp->i_df.if_bytes == 0);
 	xfs_init_local_fork(dp, XFS_DATA_FORK, sfp, size);
-	dp->i_d.di_format = XFS_DINODE_FMT_LOCAL;
+	dp->i_df.if_format = XFS_DINODE_FMT_LOCAL;
 	dp->i_d.di_size = size;
 
 	logflags |= XFS_ILOG_DDATA;
@@ -710,11 +710,11 @@ xfs_dir2_sf_verify(
 	struct xfs_inode		*ip)
 {
 	struct xfs_mount		*mp = ip->i_mount;
+	struct xfs_ifork		*ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
 	struct xfs_dir2_sf_hdr		*sfp;
 	struct xfs_dir2_sf_entry	*sfep;
 	struct xfs_dir2_sf_entry	*next_sfep;
 	char				*endp;
-	struct xfs_ifork		*ifp;
 	xfs_ino_t			ino;
 	int				i;
 	int				i8count;
@@ -723,9 +723,8 @@ xfs_dir2_sf_verify(
 	int				error;
 	uint8_t				filetype;
 
-	ASSERT(ip->i_d.di_format == XFS_DINODE_FMT_LOCAL);
+	ASSERT(ifp->if_format == XFS_DINODE_FMT_LOCAL);
 
-	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
 	sfp = (struct xfs_dir2_sf_hdr *)ifp->if_u1.if_data;
 	size = ifp->if_bytes;
 
@@ -827,9 +826,9 @@ xfs_dir2_sf_create(
 	 * If it's currently a zero-length extent file,
 	 * convert it to local format.
 	 */
-	if (dp->i_d.di_format == XFS_DINODE_FMT_EXTENTS) {
+	if (dp->i_df.if_format == XFS_DINODE_FMT_EXTENTS) {
 		dp->i_df.if_flags &= ~XFS_IFEXTENTS;	/* just in case */
-		dp->i_d.di_format = XFS_DINODE_FMT_LOCAL;
+		dp->i_df.if_format = XFS_DINODE_FMT_LOCAL;
 		xfs_trans_log_inode(args->trans, dp, XFS_ILOG_CORE);
 		dp->i_df.if_flags |= XFS_IFINLINE;
 	}
@@ -1027,7 +1026,7 @@ xfs_dir2_sf_replace_needblock(
 	int			newsize;
 	struct xfs_dir2_sf_hdr	*sfp;
 
-	if (dp->i_d.di_format != XFS_DINODE_FMT_LOCAL)
+	if (dp->i_df.if_format != XFS_DINODE_FMT_LOCAL)
 		return false;
 
 	sfp = (struct xfs_dir2_sf_hdr *)dp->i_df.if_u1.if_data;
diff --git a/fs/xfs/libxfs/xfs_dquot_buf.c b/fs/xfs/libxfs/xfs_dquot_buf.c
index bedc1e752b60..5a2db00b9d5f 100644
--- a/fs/xfs/libxfs/xfs_dquot_buf.c
+++ b/fs/xfs/libxfs/xfs_dquot_buf.c
@@ -37,9 +37,10 @@ xfs_failaddr_t
 xfs_dquot_verify(
 	struct xfs_mount	*mp,
 	struct xfs_disk_dquot	*ddq,
-	xfs_dqid_t		id,
-	uint			type)	/* used only during quotacheck */
+	xfs_dqid_t		id)	/* used only during quotacheck */
 {
+	__u8			ddq_type;
+
 	/*
 	 * We can encounter an uninitialized dquot buffer for 2 reasons:
 	 * 1. If we crash while deleting the quotainode(s), and those blks got
@@ -60,11 +61,12 @@ xfs_dquot_verify(
 	if (ddq->d_version != XFS_DQUOT_VERSION)
 		return __this_address;
 
-	if (type && ddq->d_flags != type)
+	if (ddq->d_type & ~XFS_DQTYPE_ANY)
 		return __this_address;
-	if (ddq->d_flags != XFS_DQ_USER &&
-	    ddq->d_flags != XFS_DQ_PROJ &&
-	    ddq->d_flags != XFS_DQ_GROUP)
+	ddq_type = ddq->d_type & XFS_DQTYPE_REC_MASK;
+	if (ddq_type != XFS_DQTYPE_USER &&
+	    ddq_type != XFS_DQTYPE_PROJ &&
+	    ddq_type != XFS_DQTYPE_GROUP)
 		return __this_address;
 
 	if (id != -1 && id != be32_to_cpu(ddq->d_id))
@@ -95,14 +97,13 @@ xfs_failaddr_t
 xfs_dqblk_verify(
 	struct xfs_mount	*mp,
 	struct xfs_dqblk	*dqb,
-	xfs_dqid_t	 	id,
-	uint		 	type)	/* used only during quotacheck */
+	xfs_dqid_t		id)	/* used only during quotacheck */
 {
 	if (xfs_sb_version_hascrc(&mp->m_sb) &&
 	    !uuid_equal(&dqb->dd_uuid, &mp->m_sb.sb_meta_uuid))
 		return __this_address;
 
-	return xfs_dquot_verify(mp, &dqb->dd_diskdq, id, type);
+	return xfs_dquot_verify(mp, &dqb->dd_diskdq, id);
 }
 
 /*
@@ -113,7 +114,7 @@ xfs_dqblk_repair(
 	struct xfs_mount	*mp,
 	struct xfs_dqblk	*dqb,
 	xfs_dqid_t		id,
-	uint			type)
+	xfs_dqtype_t		type)
 {
 	/*
 	 * Typically, a repair is only requested by quotacheck.
@@ -123,7 +124,7 @@ xfs_dqblk_repair(
 
 	dqb->dd_diskdq.d_magic = cpu_to_be16(XFS_DQUOT_MAGIC);
 	dqb->dd_diskdq.d_version = XFS_DQUOT_VERSION;
-	dqb->dd_diskdq.d_flags = type;
+	dqb->dd_diskdq.d_type = type;
 	dqb->dd_diskdq.d_id = cpu_to_be32(id);
 
 	if (xfs_sb_version_hascrc(&mp->m_sb)) {
@@ -205,7 +206,7 @@ xfs_dquot_buf_verify(
 		if (i == 0)
 			id = be32_to_cpu(ddq->d_id);
 
-		fa = xfs_dqblk_verify(mp, &dqb[i], id + i, 0);
+		fa = xfs_dqblk_verify(mp, &dqb[i], id + i);
 		if (fa) {
 			if (!readahead)
 				xfs_buf_verifier_error(bp, -EFSCORRUPTED,
diff --git a/fs/xfs/libxfs/xfs_errortag.h b/fs/xfs/libxfs/xfs_errortag.h
index 79e6c4fb1d8a..53b305dea381 100644
--- a/fs/xfs/libxfs/xfs_errortag.h
+++ b/fs/xfs/libxfs/xfs_errortag.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0+
+/* SPDX-License-Identifier: GPL-2.0+ */
 /*
  * Copyright (c) 2000-2002,2005 Silicon Graphics, Inc.
  * Copyright (C) 2017 Oracle.
@@ -55,7 +55,8 @@
 #define XFS_ERRTAG_FORCE_SCRUB_REPAIR			32
 #define XFS_ERRTAG_FORCE_SUMMARY_RECALC			33
 #define XFS_ERRTAG_IUNLINK_FALLBACK			34
-#define XFS_ERRTAG_MAX					35
+#define XFS_ERRTAG_BUF_IOERROR				35
+#define XFS_ERRTAG_MAX					36
 
 /*
  * Random factors for above tags, 1 means always, 2 means 1/2 time, etc.
@@ -95,5 +96,6 @@
 #define XFS_RANDOM_FORCE_SCRUB_REPAIR			1
 #define XFS_RANDOM_FORCE_SUMMARY_RECALC			1
 #define XFS_RANDOM_IUNLINK_FALLBACK			(XFS_RANDOM_DEFAULT/10)
+#define XFS_RANDOM_BUF_IOERROR				XFS_RANDOM_DEFAULT
 
 #endif /* __XFS_ERRORTAG_H_ */
diff --git a/fs/xfs/libxfs/xfs_format.h b/fs/xfs/libxfs/xfs_format.h
index 77e9fa385980..31b7ece985bb 100644
--- a/fs/xfs/libxfs/xfs_format.h
+++ b/fs/xfs/libxfs/xfs_format.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0
+/* SPDX-License-Identifier: GPL-2.0 */
 /*
  * Copyright (c) 2000-2005 Silicon Graphics, Inc.
  * All Rights Reserved.
@@ -497,6 +497,23 @@ static inline bool xfs_sb_version_hascrc(struct xfs_sb *sbp)
 	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5;
 }
 
+/*
+ * v5 file systems support V3 inodes only, earlier file systems support
+ * v2 and v1 inodes.
+ */
+static inline bool xfs_sb_version_has_v3inode(struct xfs_sb *sbp)
+{
+	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5;
+}
+
+static inline bool xfs_dinode_good_version(struct xfs_sb *sbp,
+		uint8_t version)
+{
+	if (xfs_sb_version_has_v3inode(sbp))
+		return version == 3;
+	return version == 1 || version == 2;
+}
+
 static inline bool xfs_sb_version_has_pquotino(struct xfs_sb *sbp)
 {
 	return XFS_SB_VERSION_NUM(sbp) == XFS_SB_VERSION_5;
@@ -560,7 +577,6 @@ xfs_is_quota_inode(struct xfs_sb *sbp, xfs_ino_t ino)
 
 #define XFS_SB_DADDR		((xfs_daddr_t)0) /* daddr in filesystem/ag */
 #define	XFS_SB_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_SB_DADDR)
-#define XFS_BUF_TO_SBP(bp)	((xfs_dsb_t *)((bp)->b_addr))
 
 #define	XFS_HDR_BLOCK(mp,d)	((xfs_agblock_t)XFS_BB_TO_FSBT(mp,d))
 #define	XFS_DADDR_TO_FSB(mp,d)	XFS_AGB_TO_FSB(mp, \
@@ -707,7 +723,6 @@ typedef struct xfs_agf {
 /* disk block (xfs_daddr_t) in the AG */
 #define XFS_AGF_DADDR(mp)	((xfs_daddr_t)(1 << (mp)->m_sectbb_log))
 #define	XFS_AGF_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_AGF_DADDR(mp))
-#define	XFS_BUF_TO_AGF(bp)	((xfs_agf_t *)((bp)->b_addr))
 
 /*
  * Size of the unlinked inode hash table in the agi.
@@ -775,7 +790,6 @@ typedef struct xfs_agi {
 /* disk block (xfs_daddr_t) in the AG */
 #define XFS_AGI_DADDR(mp)	((xfs_daddr_t)(2 << (mp)->m_sectbb_log))
 #define	XFS_AGI_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_AGI_DADDR(mp))
-#define	XFS_BUF_TO_AGI(bp)	((xfs_agi_t *)((bp)->b_addr))
 
 /*
  * The third a.g. block contains the a.g. freelist, an array
@@ -783,21 +797,15 @@ typedef struct xfs_agi {
  */
 #define XFS_AGFL_DADDR(mp)	((xfs_daddr_t)(3 << (mp)->m_sectbb_log))
 #define	XFS_AGFL_BLOCK(mp)	XFS_HDR_BLOCK(mp, XFS_AGFL_DADDR(mp))
-#define	XFS_BUF_TO_AGFL(bp)	((xfs_agfl_t *)((bp)->b_addr))
-
-#define XFS_BUF_TO_AGFL_BNO(mp, bp) \
-	(xfs_sb_version_hascrc(&((mp)->m_sb)) ? \
-		&(XFS_BUF_TO_AGFL(bp)->agfl_bno[0]) : \
-		(__be32 *)(bp)->b_addr)
+#define	XFS_BUF_TO_AGFL(bp)	((struct xfs_agfl *)((bp)->b_addr))
 
-typedef struct xfs_agfl {
+struct xfs_agfl {
 	__be32		agfl_magicnum;
 	__be32		agfl_seqno;
 	uuid_t		agfl_uuid;
 	__be64		agfl_lsn;
 	__be32		agfl_crc;
-	__be32		agfl_bno[];	/* actually xfs_agfl_size(mp) */
-} __attribute__((packed)) xfs_agfl_t;
+} __attribute__((packed));
 
 #define XFS_AGFL_CRC_OFF	offsetof(struct xfs_agfl, agfl_crc)
 
@@ -946,23 +954,22 @@ enum xfs_dinode_fmt {
 /*
  * Inode size for given fs.
  */
-#define XFS_LITINO(mp, version) \
-	((int)(((mp)->m_sb.sb_inodesize) - xfs_dinode_size(version)))
+#define XFS_DINODE_SIZE(sbp) \
+	(xfs_sb_version_has_v3inode(sbp) ? \
+		sizeof(struct xfs_dinode) : \
+		offsetof(struct xfs_dinode, di_crc))
+#define XFS_LITINO(mp) \
+	((mp)->m_sb.sb_inodesize - XFS_DINODE_SIZE(&(mp)->m_sb))
 
 /*
  * Inode data & attribute fork sizes, per inode.
  */
-#define XFS_DFORK_Q(dip)		((dip)->di_forkoff != 0)
 #define XFS_DFORK_BOFF(dip)		((int)((dip)->di_forkoff << 3))
 
 #define XFS_DFORK_DSIZE(dip,mp) \
-	(XFS_DFORK_Q(dip) ? \
-		XFS_DFORK_BOFF(dip) : \
-		XFS_LITINO(mp, (dip)->di_version))
+	((dip)->di_forkoff ? XFS_DFORK_BOFF(dip) : XFS_LITINO(mp))
 #define XFS_DFORK_ASIZE(dip,mp) \
-	(XFS_DFORK_Q(dip) ? \
-		XFS_LITINO(mp, (dip)->di_version) - XFS_DFORK_BOFF(dip) : \
-		0)
+	((dip)->di_forkoff ? XFS_LITINO(mp) - XFS_DFORK_BOFF(dip) : 0)
 #define XFS_DFORK_SIZE(dip,mp,w) \
 	((w) == XFS_DATA_FORK ? \
 		XFS_DFORK_DSIZE(dip, mp) : \
@@ -1142,16 +1149,26 @@ static inline void xfs_dinode_put_rdev(struct xfs_dinode *dip, xfs_dev_t rdev)
 #define XFS_DQUOT_MAGIC		0x4451		/* 'DQ' */
 #define XFS_DQUOT_VERSION	(uint8_t)0x01	/* latest version number */
 
+#define XFS_DQTYPE_USER		0x01		/* user dquot record */
+#define XFS_DQTYPE_PROJ		0x02		/* project dquot record */
+#define XFS_DQTYPE_GROUP	0x04		/* group dquot record */
+
+/* bitmask to determine if this is a user/group/project dquot */
+#define XFS_DQTYPE_REC_MASK	(XFS_DQTYPE_USER | \
+				 XFS_DQTYPE_PROJ | \
+				 XFS_DQTYPE_GROUP)
+
+#define XFS_DQTYPE_ANY		(XFS_DQTYPE_REC_MASK)
+
 /*
- * This is the main portion of the on-disk representation of quota
- * information for a user. This is the q_core of the struct xfs_dquot that
- * is kept in kernel memory. We pad this with some more expansion room
- * to construct the on disk structure.
+ * This is the main portion of the on-disk representation of quota information
+ * for a user.  We pad this with some more expansion room to construct the on
+ * disk structure.
  */
 struct xfs_disk_dquot {
 	__be16		d_magic;	/* dquot magic = XFS_DQUOT_MAGIC */
 	__u8		d_version;	/* dquot version */
-	__u8		d_flags;	/* XFS_DQ_USER/PROJ/GROUP */
+	__u8		d_type;		/* XFS_DQTYPE_USER/PROJ/GROUP */
 	__be32		d_id;		/* user,project,group id */
 	__be64		d_blk_hardlimit;/* absolute limit on disk blks */
 	__be64		d_blk_softlimit;/* preferred limit on disk blks */
@@ -1192,6 +1209,22 @@ typedef struct xfs_dqblk {
 #define XFS_DQUOT_CRC_OFF	offsetof(struct xfs_dqblk, dd_crc)
 
 /*
+ * This defines the unit of allocation of dquots.
+ *
+ * Currently, it is just one file system block, and a 4K blk contains 30
+ * (136 * 30 = 4080) dquots. It's probably not worth trying to make
+ * this more dynamic.
+ *
+ * However, if this number is changed, we have to make sure that we don't
+ * implicitly assume that we do allocations in chunks of a single filesystem
+ * block in the dquot/xqm code.
+ *
+ * This is part of the ondisk format because the structure size is not a power
+ * of two, which leaves slack at the end of the disk block.
+ */
+#define XFS_DQUOT_CLUSTER_SIZE_FSB	(xfs_filblks_t)1
+
+/*
  * Remote symlink format and access functions.
  */
 #define XFS_SYMLINK_MAGIC	0x58534c4d	/* XSLM */
@@ -1673,7 +1706,7 @@ struct xfs_acl_entry {
 
 struct xfs_acl {
 	__be32			acl_cnt;
-	struct xfs_acl_entry	acl_entry[0];
+	struct xfs_acl_entry	acl_entry[];
 };
 
 /*
diff --git a/fs/xfs/libxfs/xfs_fs.h b/fs/xfs/libxfs/xfs_fs.h
index ef95ca07d084..84bcffa87753 100644
--- a/fs/xfs/libxfs/xfs_fs.h
+++ b/fs/xfs/libxfs/xfs_fs.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: LGPL-2.1
+/* SPDX-License-Identifier: LGPL-2.1 */
 /*
  * Copyright (c) 1995-2005 Silicon Graphics, Inc.
  * All Rights Reserved.
@@ -568,10 +568,40 @@ typedef struct xfs_fsop_setdm_handlereq {
 	struct fsdmidata		__user *data;	/* DMAPI data	*/
 } xfs_fsop_setdm_handlereq_t;
 
+/*
+ * Flags passed in xfs_attr_multiop.am_flags for the attr ioctl interface.
+ *
+ * NOTE: Must match the values declared in libattr without the XFS_IOC_ prefix.
+ */
+#define XFS_IOC_ATTR_ROOT	0x0002	/* use attrs in root namespace */
+#define XFS_IOC_ATTR_SECURE	0x0008	/* use attrs in security namespace */
+#define XFS_IOC_ATTR_CREATE	0x0010	/* fail if attr already exists */
+#define XFS_IOC_ATTR_REPLACE	0x0020	/* fail if attr does not exist */
+
 typedef struct xfs_attrlist_cursor {
 	__u32		opaque[4];
 } xfs_attrlist_cursor_t;
 
+/*
+ * Define how lists of attribute names are returned to userspace from the
+ * XFS_IOC_ATTRLIST_BY_HANDLE ioctl.  struct xfs_attrlist is the header at the
+ * beginning of the returned buffer, and a each entry in al_offset contains the
+ * relative offset of an xfs_attrlist_ent containing the actual entry.
+ *
+ * NOTE: struct xfs_attrlist must match struct attrlist defined in libattr, and
+ * struct xfs_attrlist_ent must match struct attrlist_ent defined in libattr.
+ */
+struct xfs_attrlist {
+	__s32	al_count;	/* number of entries in attrlist */
+	__s32	al_more;	/* T/F: more attrs (do call again) */
+	__s32	al_offset[1];	/* byte offsets of attrs [var-sized] */
+};
+
+struct xfs_attrlist_ent {	/* data from attr_list() */
+	__u32	a_valuelen;	/* number bytes in value of attr */
+	char	a_name[1];	/* attr name (NULL terminated) */
+};
+
 typedef struct xfs_fsop_attrlist_handlereq {
 	struct xfs_fsop_handlereq	hreq; /* handle interface structure */
 	struct xfs_attrlist_cursor	pos; /* opaque cookie, list offset */
@@ -589,7 +619,7 @@ typedef struct xfs_attr_multiop {
 	void		__user *am_attrname;
 	void		__user *am_attrvalue;
 	__u32		am_length;
-	__u32		am_flags;
+	__u32		am_flags; /* XFS_IOC_ATTR_* */
 } xfs_attr_multiop_t;
 
 typedef struct xfs_fsop_attrmulti_handlereq {
diff --git a/fs/xfs/libxfs/xfs_health.h b/fs/xfs/libxfs/xfs_health.h
index 272005ac8c88..99e796256c5d 100644
--- a/fs/xfs/libxfs/xfs_health.h
+++ b/fs/xfs/libxfs/xfs_health.h
@@ -1,4 +1,4 @@
-// SPDX-License-Identifier: GPL-2.0+
+/* SPDX-License-Identifier: GPL-2.0+ */
 /*
  * Copyright (C) 2019 Oracle.  All Rights Reserved.
  * Author: Darrick J. Wong <darrick.wong@oracle.com>
diff --git a/fs/xfs/libxfs/xfs_ialloc.c b/fs/xfs/libxfs/xfs_ialloc.c
index bf161e930f1d..f742a96a2fe1 100644
--- a/fs/xfs/libxfs/xfs_ialloc.c
+++ b/fs/xfs/libxfs/xfs_ialloc.c
@@ -105,7 +105,7 @@ xfs_inobt_get_rec(
 	int				*stat)
 {
 	struct xfs_mount		*mp = cur->bc_mp;
-	xfs_agnumber_t			agno = cur->bc_private.a.agno;
+	xfs_agnumber_t			agno = cur->bc_ag.agno;
 	union xfs_btree_rec		*rec;
 	int				error;
 	uint64_t			realfree;
@@ -177,7 +177,7 @@ xfs_inobt_insert(
 	xfs_btnum_t		btnum)
 {
 	struct xfs_btree_cur	*cur;
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agbp);
+	struct xfs_agi		*agi = agbp->b_addr;
 	xfs_agnumber_t		agno = be32_to_cpu(agi->agi_seqno);
 	xfs_agino_t		thisino;
 	int			i;
@@ -304,7 +304,7 @@ xfs_ialloc_inode_init(
 	 * That means for v3 inode we log the entire buffer rather than just the
 	 * inode cores.
 	 */
-	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+	if (xfs_sb_version_has_v3inode(&mp->m_sb)) {
 		version = 3;
 		ino = XFS_AGINO_TO_INO(mp, agno, XFS_AGB_TO_AGINO(mp, agbno));
 
@@ -339,7 +339,7 @@ xfs_ialloc_inode_init(
 		xfs_buf_zero(fbuf, 0, BBTOB(fbuf->b_length));
 		for (i = 0; i < M_IGEO(mp)->inodes_per_cluster; i++) {
 			int	ioffset = i << mp->m_sb.sb_inodelog;
-			uint	isize = xfs_dinode_size(version);
+			uint	isize = XFS_DINODE_SIZE(&mp->m_sb);
 
 			free = xfs_make_iptr(mp, fbuf, i);
 			free->di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
@@ -525,7 +525,7 @@ xfs_inobt_insert_sprec(
 	bool				merge)	/* merge or replace */
 {
 	struct xfs_btree_cur		*cur;
-	struct xfs_agi			*agi = XFS_BUF_TO_AGI(agbp);
+	struct xfs_agi			*agi = agbp->b_addr;
 	xfs_agnumber_t			agno = be32_to_cpu(agi->agi_seqno);
 	int				error;
 	int				i;
@@ -658,7 +658,7 @@ xfs_ialloc_ag_alloc(
 	 * chunk of inodes.  If the filesystem is striped, this will fill
 	 * an entire stripe unit with inodes.
 	 */
-	agi = XFS_BUF_TO_AGI(agbp);
+	agi = agbp->b_addr;
 	newino = be32_to_cpu(agi->agi_newino);
 	agno = be32_to_cpu(agi->agi_seqno);
 	args.agbno = XFS_AGINO_TO_AGBNO(args.mp, newino) +
@@ -888,10 +888,9 @@ sparse_alloc:
 	 */
 	be32_add_cpu(&agi->agi_count, newlen);
 	be32_add_cpu(&agi->agi_freecount, newlen);
-	pag = xfs_perag_get(args.mp, agno);
+	pag = agbp->b_pag;
 	pag->pagi_freecount += newlen;
 	pag->pagi_count += newlen;
-	xfs_perag_put(pag);
 	agi->agi_newino = cpu_to_be32(newino);
 
 	/*
@@ -1130,11 +1129,11 @@ xfs_dialloc_ag_inobt(
 	xfs_ino_t		*inop)
 {
 	struct xfs_mount	*mp = tp->t_mountp;
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agbp);
+	struct xfs_agi		*agi = agbp->b_addr;
 	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_perag	*pag = agbp->b_pag;
 	struct xfs_btree_cur	*cur, *tcur;
 	struct xfs_inobt_rec_incore rec, trec;
 	xfs_ino_t		ino;
@@ -1143,8 +1142,6 @@ xfs_dialloc_ag_inobt(
 	int			i, j;
 	int			searchdistance = 10;
 
-	pag = xfs_perag_get(mp, agno);
-
 	ASSERT(pag->pagi_init);
 	ASSERT(pag->pagi_inodeok);
 	ASSERT(pag->pagi_freecount > 0);
@@ -1384,14 +1381,12 @@ alloc_inode:
 
 	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
 	xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
-	xfs_perag_put(pag);
 	*inop = ino;
 	return 0;
 error1:
 	xfs_btree_del_cursor(tcur, XFS_BTREE_ERROR);
 error0:
 	xfs_btree_del_cursor(cur, XFS_BTREE_ERROR);
-	xfs_perag_put(pag);
 	return error;
 }
 
@@ -1583,11 +1578,10 @@ xfs_dialloc_ag(
 	xfs_ino_t		*inop)
 {
 	struct xfs_mount		*mp = tp->t_mountp;
-	struct xfs_agi			*agi = XFS_BUF_TO_AGI(agbp);
+	struct xfs_agi			*agi = agbp->b_addr;
 	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;
@@ -1599,8 +1593,6 @@ xfs_dialloc_ag(
 	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.
@@ -1667,7 +1659,7 @@ xfs_dialloc_ag(
 	 */
 	be32_add_cpu(&agi->agi_freecount, -1);
 	xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
-	pag->pagi_freecount--;
+	agbp->b_pag->pagi_freecount--;
 
 	xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -1);
 
@@ -1680,7 +1672,6 @@ xfs_dialloc_ag(
 
 	xfs_btree_del_cursor(icur, XFS_BTREE_NOERROR);
 	xfs_btree_del_cursor(cur, XFS_BTREE_NOERROR);
-	xfs_perag_put(pag);
 	*inop = ino;
 	return 0;
 
@@ -1688,7 +1679,6 @@ 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;
 }
 
@@ -1943,9 +1933,8 @@ xfs_difree_inobt(
 	struct xfs_icluster		*xic,
 	struct xfs_inobt_rec_incore	*orec)
 {
-	struct xfs_agi			*agi = XFS_BUF_TO_AGI(agbp);
+	struct xfs_agi			*agi = agbp->b_addr;
 	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;
@@ -2007,6 +1996,8 @@ xfs_difree_inobt(
 	if (!(mp->m_flags & XFS_MOUNT_IKEEP) &&
 	    rec.ir_free == XFS_INOBT_ALL_FREE &&
 	    mp->m_sb.sb_inopblock <= XFS_INODES_PER_CHUNK) {
+		struct xfs_perag	*pag = agbp->b_pag;
+
 		xic->deleted = true;
 		xic->first_ino = XFS_AGINO_TO_INO(mp, agno, rec.ir_startino);
 		xic->alloc = xfs_inobt_irec_to_allocmask(&rec);
@@ -2020,10 +2011,8 @@ xfs_difree_inobt(
 		be32_add_cpu(&agi->agi_count, -ilen);
 		be32_add_cpu(&agi->agi_freecount, -(ilen - 1));
 		xfs_ialloc_log_agi(tp, agbp, XFS_AGI_COUNT | XFS_AGI_FREECOUNT);
-		pag = xfs_perag_get(mp, agno);
 		pag->pagi_freecount -= ilen - 1;
 		pag->pagi_count -= ilen;
-		xfs_perag_put(pag);
 		xfs_trans_mod_sb(tp, XFS_TRANS_SB_ICOUNT, -ilen);
 		xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, -(ilen - 1));
 
@@ -2049,9 +2038,7 @@ xfs_difree_inobt(
 		 */
 		be32_add_cpu(&agi->agi_freecount, 1);
 		xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREECOUNT);
-		pag = xfs_perag_get(mp, agno);
-		pag->pagi_freecount++;
-		xfs_perag_put(pag);
+		agbp->b_pag->pagi_freecount++;
 		xfs_trans_mod_sb(tp, XFS_TRANS_SB_IFREE, 1);
 	}
 
@@ -2079,7 +2066,7 @@ xfs_difree_finobt(
 	xfs_agino_t			agino,
 	struct xfs_inobt_rec_incore	*ibtrec) /* inobt record */
 {
-	struct xfs_agi			*agi = XFS_BUF_TO_AGI(agbp);
+	struct xfs_agi			*agi = agbp->b_addr;
 	xfs_agnumber_t			agno = be32_to_cpu(agi->agi_seqno);
 	struct xfs_btree_cur		*cur;
 	struct xfs_inobt_rec_incore	rec;
@@ -2489,9 +2476,8 @@ xfs_ialloc_log_agi(
 		sizeof(xfs_agi_t)
 	};
 #ifdef DEBUG
-	xfs_agi_t		*agi;	/* allocation group header */
+	struct xfs_agi		*agi = bp->b_addr;
 
-	agi = XFS_BUF_TO_AGI(bp);
 	ASSERT(agi->agi_magicnum == cpu_to_be32(XFS_AGI_MAGIC));
 #endif
 
@@ -2523,14 +2509,13 @@ xfs_agi_verify(
 	struct xfs_buf	*bp)
 {
 	struct xfs_mount *mp = bp->b_mount;
-	struct xfs_agi	*agi = XFS_BUF_TO_AGI(bp);
+	struct xfs_agi	*agi = bp->b_addr;
 	int		i;
 
 	if (xfs_sb_version_hascrc(&mp->m_sb)) {
 		if (!uuid_equal(&agi->agi_uuid, &mp->m_sb.sb_meta_uuid))
 			return __this_address;
-		if (!xfs_log_check_lsn(mp,
-				be64_to_cpu(XFS_BUF_TO_AGI(bp)->agi_lsn)))
+		if (!xfs_log_check_lsn(mp, be64_to_cpu(agi->agi_lsn)))
 			return __this_address;
 	}
 
@@ -2593,6 +2578,7 @@ xfs_agi_write_verify(
 {
 	struct xfs_mount	*mp = bp->b_mount;
 	struct xfs_buf_log_item	*bip = bp->b_log_item;
+	struct xfs_agi		*agi = bp->b_addr;
 	xfs_failaddr_t		fa;
 
 	fa = xfs_agi_verify(bp);
@@ -2605,7 +2591,7 @@ xfs_agi_write_verify(
 		return;
 
 	if (bip)
-		XFS_BUF_TO_AGI(bp)->agi_lsn = cpu_to_be64(bip->bli_item.li_lsn);
+		agi->agi_lsn = cpu_to_be64(bip->bli_item.li_lsn);
 	xfs_buf_update_cksum(bp, XFS_AGI_CRC_OFF);
 }
 
@@ -2661,8 +2647,8 @@ xfs_ialloc_read_agi(
 	if (error)
 		return error;
 
-	agi = XFS_BUF_TO_AGI(*bpp);
-	pag = xfs_perag_get(mp, agno);
+	agi = (*bpp)->b_addr;
+	pag = (*bpp)->b_pag;
 	if (!pag->pagi_init) {
 		pag->pagi_freecount = be32_to_cpu(agi->agi_freecount);
 		pag->pagi_count = be32_to_cpu(agi->agi_count);
@@ -2675,7 +2661,6 @@ xfs_ialloc_read_agi(
 	 */
 	ASSERT(pag->pagi_freecount == be32_to_cpu(agi->agi_freecount) ||
 		XFS_FORCED_SHUTDOWN(mp));
-	xfs_perag_put(pag);
 	return 0;
 }
 
@@ -2873,7 +2858,7 @@ xfs_ialloc_setup_geometry(
 	 * cannot change the behavior.
 	 */
 	igeo->inode_cluster_size_raw = XFS_INODE_BIG_CLUSTER_SIZE;
-	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+	if (xfs_sb_version_has_v3inode(&mp->m_sb)) {
 		int	new_size = igeo->inode_cluster_size_raw;
 
 		new_size *= mp->m_sb.sb_inodesize / XFS_DINODE_MIN_SIZE;
diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.c b/fs/xfs/libxfs/xfs_ialloc_btree.c
index b82992f795aa..3c8aebc36e64 100644
--- a/fs/xfs/libxfs/xfs_ialloc_btree.c
+++ b/fs/xfs/libxfs/xfs_ialloc_btree.c
@@ -12,6 +12,7 @@
 #include "xfs_bit.h"
 #include "xfs_mount.h"
 #include "xfs_btree.h"
+#include "xfs_btree_staging.h"
 #include "xfs_ialloc.h"
 #include "xfs_ialloc_btree.h"
 #include "xfs_alloc.h"
@@ -20,7 +21,6 @@
 #include "xfs_trans.h"
 #include "xfs_rmap.h"
 
-
 STATIC int
 xfs_inobt_get_minrecs(
 	struct xfs_btree_cur	*cur,
@@ -34,7 +34,7 @@ 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_ag.agbp, cur->bc_ag.agno,
 			cur->bc_btnum);
 }
 
@@ -44,8 +44,8 @@ xfs_inobt_set_root(
 	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);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agi		*agi = agbp->b_addr;
 
 	agi->agi_root = nptr->s;
 	be32_add_cpu(&agi->agi_level, inc);
@@ -58,8 +58,8 @@ xfs_finobt_set_root(
 	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);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agi		*agi = agbp->b_addr;
 
 	agi->agi_free_root = nptr->s;
 	be32_add_cpu(&agi->agi_free_level, inc);
@@ -83,7 +83,7 @@ __xfs_inobt_alloc_block(
 	args.tp = cur->bc_tp;
 	args.mp = cur->bc_mp;
 	args.oinfo = XFS_RMAP_OINFO_INOBT;
-	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_private.a.agno, sbno);
+	args.fsbno = XFS_AGB_TO_FSB(args.mp, cur->bc_ag.agno, sbno);
 	args.minlen = 1;
 	args.maxlen = 1;
 	args.prod = 1;
@@ -212,9 +212,9 @@ xfs_inobt_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);
+	struct xfs_agi		*agi = cur->bc_ag.agbp->b_addr;
 
-	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agi->agi_seqno));
+	ASSERT(cur->bc_ag.agno == be32_to_cpu(agi->agi_seqno));
 
 	ptr->s = agi->agi_root;
 }
@@ -224,9 +224,9 @@ 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);
+	struct xfs_agi		*agi = cur->bc_ag.agbp->b_addr;
 
-	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agi->agi_seqno));
+	ASSERT(cur->bc_ag.agno == be32_to_cpu(agi->agi_seqno));
 	ptr->s = agi->agi_free_root;
 }
 
@@ -400,32 +400,27 @@ static const struct xfs_btree_ops xfs_finobt_ops = {
 };
 
 /*
- * Allocate a new inode btree cursor.
+ * Initialize a new inode btree cursor.
  */
-struct xfs_btree_cur *				/* new inode btree cursor */
-xfs_inobt_init_cursor(
+static struct xfs_btree_cur *
+xfs_inobt_init_common(
 	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_btnum_t		btnum)		/* ialloc or free ino btree */
 {
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agbp);
 	struct xfs_btree_cur	*cur;
 
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
-
+	cur = kmem_cache_zalloc(xfs_btree_cur_zone, GFP_NOFS | __GFP_NOFAIL);
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
 	cur->bc_btnum = btnum;
 	if (btnum == XFS_BTNUM_INO) {
-		cur->bc_nlevels = be32_to_cpu(agi->agi_level);
-		cur->bc_ops = &xfs_inobt_ops;
 		cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_ibt_2);
+		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_statoff = XFS_STATS_CALC_INDEX(xs_fibt_2);
+		cur->bc_ops = &xfs_finobt_ops;
 	}
 
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
@@ -433,12 +428,75 @@ xfs_inobt_init_cursor(
 	if (xfs_sb_version_hascrc(&mp->m_sb))
 		cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
 
-	cur->bc_private.a.agbp = agbp;
-	cur->bc_private.a.agno = agno;
+	cur->bc_ag.agno = agno;
+	return cur;
+}
+
+/* Create an inode btree cursor. */
+struct xfs_btree_cur *
+xfs_inobt_init_cursor(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
+	xfs_agnumber_t		agno,
+	xfs_btnum_t		btnum)
+{
+	struct xfs_btree_cur	*cur;
+	struct xfs_agi		*agi = agbp->b_addr;
 
+	cur = xfs_inobt_init_common(mp, tp, agno, btnum);
+	if (btnum == XFS_BTNUM_INO)
+		cur->bc_nlevels = be32_to_cpu(agi->agi_level);
+	else
+		cur->bc_nlevels = be32_to_cpu(agi->agi_free_level);
+	cur->bc_ag.agbp = agbp;
 	return cur;
 }
 
+/* Create an inode btree cursor with a fake root for staging. */
+struct xfs_btree_cur *
+xfs_inobt_stage_cursor(
+	struct xfs_mount	*mp,
+	struct xbtree_afakeroot	*afake,
+	xfs_agnumber_t		agno,
+	xfs_btnum_t		btnum)
+{
+	struct xfs_btree_cur	*cur;
+
+	cur = xfs_inobt_init_common(mp, NULL, agno, btnum);
+	xfs_btree_stage_afakeroot(cur, afake);
+	return cur;
+}
+
+/*
+ * Install a new inobt btree root.  Caller is responsible for invalidating
+ * and freeing the old btree blocks.
+ */
+void
+xfs_inobt_commit_staged_btree(
+	struct xfs_btree_cur	*cur,
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp)
+{
+	struct xfs_agi		*agi = agbp->b_addr;
+	struct xbtree_afakeroot	*afake = cur->bc_ag.afake;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+
+	if (cur->bc_btnum == XFS_BTNUM_INO) {
+		agi->agi_root = cpu_to_be32(afake->af_root);
+		agi->agi_level = cpu_to_be32(afake->af_levels);
+		xfs_ialloc_log_agi(tp, agbp, XFS_AGI_ROOT | XFS_AGI_LEVEL);
+		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_inobt_ops);
+	} else {
+		agi->agi_free_root = cpu_to_be32(afake->af_root);
+		agi->agi_free_level = cpu_to_be32(afake->af_levels);
+		xfs_ialloc_log_agi(tp, agbp, XFS_AGI_FREE_ROOT |
+					     XFS_AGI_FREE_LEVEL);
+		xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_finobt_ops);
+	}
+}
+
 /*
  * Calculate number of records in an inobt btree block.
  */
diff --git a/fs/xfs/libxfs/xfs_ialloc_btree.h b/fs/xfs/libxfs/xfs_ialloc_btree.h
index 951305ecaae1..35bbd978c272 100644
--- a/fs/xfs/libxfs/xfs_ialloc_btree.h
+++ b/fs/xfs/libxfs/xfs_ialloc_btree.h
@@ -48,6 +48,9 @@ struct xfs_mount;
 extern struct xfs_btree_cur *xfs_inobt_init_cursor(struct xfs_mount *,
 		struct xfs_trans *, struct xfs_buf *, xfs_agnumber_t,
 		xfs_btnum_t);
+struct xfs_btree_cur *xfs_inobt_stage_cursor(struct xfs_mount *mp,
+		struct xbtree_afakeroot *afake, xfs_agnumber_t agno,
+		xfs_btnum_t btnum);
 extern int xfs_inobt_maxrecs(struct xfs_mount *, int, int);
 
 /* ir_holemask to inode allocation bitmap conversion */
@@ -68,4 +71,7 @@ int xfs_inobt_cur(struct xfs_mount *mp, struct xfs_trans *tp,
 		xfs_agnumber_t agno, xfs_btnum_t btnum,
 		struct xfs_btree_cur **curpp, struct xfs_buf **agi_bpp);
 
+void xfs_inobt_commit_staged_btree(struct xfs_btree_cur *cur,
+		struct xfs_trans *tp, struct xfs_buf *agbp);
+
 #endif	/* __XFS_IALLOC_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_inode_buf.c b/fs/xfs/libxfs/xfs_inode_buf.c
index 8afacfe4be0a..8d5dd08eab75 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.c
+++ b/fs/xfs/libxfs/xfs_inode_buf.c
@@ -21,41 +21,6 @@
 #include <linux/iversion.h>
 
 /*
- * Check that none of the inode's in the buffer have a next
- * unlinked field of 0.
- */
-#if defined(DEBUG)
-void
-xfs_inobp_check(
-	xfs_mount_t	*mp,
-	xfs_buf_t	*bp)
-{
-	int		i;
-	xfs_dinode_t	*dip;
-
-	for (i = 0; i < M_IGEO(mp)->inodes_per_cluster; i++) {
-		dip = xfs_buf_offset(bp, i * mp->m_sb.sb_inodesize);
-		if (!dip->di_next_unlinked)  {
-			xfs_alert(mp,
-	"Detected bogus zero next_unlinked field in inode %d buffer 0x%llx.",
-				i, (long long)bp->b_bn);
-		}
-	}
-}
-#endif
-
-bool
-xfs_dinode_good_version(
-	struct xfs_mount *mp,
-	__u8		version)
-{
-	if (xfs_sb_version_hascrc(&mp->m_sb))
-		return version == 3;
-
-	return version == 1 || version == 2;
-}
-
-/*
  * If we are doing readahead on an inode buffer, we might be in log recovery
  * reading an inode allocation buffer that hasn't yet been replayed, and hence
  * has not had the inode cores stamped into it. Hence for readahead, the buffer
@@ -64,10 +29,10 @@ xfs_dinode_good_version(
  * If the readahead buffer is invalid, we need to mark it with an error and
  * clear the DONE status of the buffer so that a followup read will re-read it
  * from disk. We don't report the error otherwise to avoid warnings during log
- * recovery and we don't get unnecssary panics on debug kernels. We use EIO here
+ * recovery and we don't get unnecessary panics on debug kernels. We use EIO here
  * because all we want to do is say readahead failed; there is no-one to report
  * the error to, so this will distinguish it from a non-ra verifier failure.
- * Changes to this readahead error behavour also need to be reflected in
+ * Changes to this readahead error behaviour also need to be reflected in
  * xfs_dquot_buf_readahead_verify().
  */
 static void
@@ -93,7 +58,7 @@ xfs_inode_buf_verify(
 		dip = xfs_buf_offset(bp, (i << mp->m_sb.sb_inodelog));
 		unlinked_ino = be32_to_cpu(dip->di_next_unlinked);
 		di_ok = xfs_verify_magic16(bp, dip->di_magic) &&
-			xfs_dinode_good_version(mp, dip->di_version) &&
+			xfs_dinode_good_version(&mp->m_sb, dip->di_version) &&
 			xfs_verify_agino_or_null(mp, agno, unlinked_ino);
 		if (unlikely(XFS_TEST_ERROR(!di_ok, mp,
 						XFS_ERRTAG_ITOBP_INOTOBP))) {
@@ -172,8 +137,7 @@ xfs_imap_to_bp(
 	struct xfs_imap		*imap,
 	struct xfs_dinode       **dipp,
 	struct xfs_buf		**bpp,
-	uint			buf_flags,
-	uint			iget_flags)
+	uint			buf_flags)
 {
 	struct xfs_buf		*bp;
 	int			error;
@@ -183,48 +147,63 @@ xfs_imap_to_bp(
 				   (int)imap->im_len, buf_flags, &bp,
 				   &xfs_inode_buf_ops);
 	if (error) {
-		if (error == -EAGAIN) {
-			ASSERT(buf_flags & XBF_TRYLOCK);
-			return error;
-		}
-		xfs_warn(mp, "%s: xfs_trans_read_buf() returned error %d.",
-			__func__, error);
+		ASSERT(error != -EAGAIN || (buf_flags & XBF_TRYLOCK));
 		return error;
 	}
 
 	*bpp = bp;
-	*dipp = xfs_buf_offset(bp, imap->im_boffset);
+	if (dipp)
+		*dipp = xfs_buf_offset(bp, imap->im_boffset);
 	return 0;
 }
 
-void
+int
 xfs_inode_from_disk(
 	struct xfs_inode	*ip,
 	struct xfs_dinode	*from)
 {
 	struct xfs_icdinode	*to = &ip->i_d;
 	struct inode		*inode = VFS_I(ip);
+	int			error;
+	xfs_failaddr_t		fa;
+
+	ASSERT(ip->i_cowfp == NULL);
+	ASSERT(ip->i_afp == NULL);
+
+	fa = xfs_dinode_verify(ip->i_mount, ip->i_ino, from);
+	if (fa) {
+		xfs_inode_verifier_error(ip, -EFSCORRUPTED, "dinode", from,
+				sizeof(*from), fa);
+		return -EFSCORRUPTED;
+	}
 
+	/*
+	 * First get the permanent information that is needed to allocate an
+	 * inode. If the inode is unused, mode is zero and we shouldn't mess
+	 * with the uninitialized part of it.
+	 */
+	to->di_flushiter = be16_to_cpu(from->di_flushiter);
+	inode->i_generation = be32_to_cpu(from->di_gen);
+	inode->i_mode = be16_to_cpu(from->di_mode);
+	if (!inode->i_mode)
+		return 0;
 
 	/*
 	 * Convert v1 inodes immediately to v2 inode format as this is the
 	 * minimum inode version format we support in the rest of the code.
+	 * They will also be unconditionally written back to disk as v2 inodes.
 	 */
-	to->di_version = from->di_version;
-	if (to->di_version == 1) {
+	if (unlikely(from->di_version == 1)) {
 		set_nlink(inode, be16_to_cpu(from->di_onlink));
 		to->di_projid = 0;
-		to->di_version = 2;
 	} else {
 		set_nlink(inode, be32_to_cpu(from->di_nlink));
 		to->di_projid = (prid_t)be16_to_cpu(from->di_projid_hi) << 16 |
 					be16_to_cpu(from->di_projid_lo);
 	}
 
-	to->di_format = from->di_format;
-	to->di_uid = be32_to_cpu(from->di_uid);
-	to->di_gid = be32_to_cpu(from->di_gid);
-	to->di_flushiter = be16_to_cpu(from->di_flushiter);
+	i_uid_write(inode, be32_to_cpu(from->di_uid));
+	i_gid_write(inode, be32_to_cpu(from->di_gid));
 
 	/*
 	 * Time is signed, so need to convert to signed 32 bit before
@@ -238,21 +217,16 @@ xfs_inode_from_disk(
 	inode->i_mtime.tv_nsec = (int)be32_to_cpu(from->di_mtime.t_nsec);
 	inode->i_ctime.tv_sec = (int)be32_to_cpu(from->di_ctime.t_sec);
 	inode->i_ctime.tv_nsec = (int)be32_to_cpu(from->di_ctime.t_nsec);
-	inode->i_generation = be32_to_cpu(from->di_gen);
-	inode->i_mode = be16_to_cpu(from->di_mode);
 
 	to->di_size = be64_to_cpu(from->di_size);
 	to->di_nblocks = be64_to_cpu(from->di_nblocks);
 	to->di_extsize = be32_to_cpu(from->di_extsize);
-	to->di_nextents = be32_to_cpu(from->di_nextents);
-	to->di_anextents = be16_to_cpu(from->di_anextents);
 	to->di_forkoff = from->di_forkoff;
-	to->di_aformat	= from->di_aformat;
 	to->di_dmevmask	= be32_to_cpu(from->di_dmevmask);
 	to->di_dmstate	= be16_to_cpu(from->di_dmstate);
 	to->di_flags	= be16_to_cpu(from->di_flags);
 
-	if (to->di_version == 3) {
+	if (xfs_sb_version_has_v3inode(&ip->i_mount->m_sb)) {
 		inode_set_iversion_queried(inode,
 					   be64_to_cpu(from->di_changecount));
 		to->di_crtime.tv_sec = be32_to_cpu(from->di_crtime.t_sec);
@@ -260,6 +234,22 @@ xfs_inode_from_disk(
 		to->di_flags2 = be64_to_cpu(from->di_flags2);
 		to->di_cowextsize = be32_to_cpu(from->di_cowextsize);
 	}
+
+	error = xfs_iformat_data_fork(ip, from);
+	if (error)
+		return error;
+	if (from->di_forkoff) {
+		error = xfs_iformat_attr_fork(ip, from);
+		if (error)
+			goto out_destroy_data_fork;
+	}
+	if (xfs_is_reflink_inode(ip))
+		xfs_ifork_init_cow(ip);
+	return 0;
+
+out_destroy_data_fork:
+	xfs_idestroy_fork(&ip->i_df);
+	return error;
 }
 
 void
@@ -274,10 +264,9 @@ xfs_inode_to_disk(
 	to->di_magic = cpu_to_be16(XFS_DINODE_MAGIC);
 	to->di_onlink = 0;
 
-	to->di_version = from->di_version;
-	to->di_format = from->di_format;
-	to->di_uid = cpu_to_be32(from->di_uid);
-	to->di_gid = cpu_to_be32(from->di_gid);
+	to->di_format = xfs_ifork_format(&ip->i_df);
+	to->di_uid = cpu_to_be32(i_uid_read(inode));
+	to->di_gid = cpu_to_be32(i_gid_read(inode));
 	to->di_projid_lo = cpu_to_be16(from->di_projid & 0xffff);
 	to->di_projid_hi = cpu_to_be16(from->di_projid >> 16);
 
@@ -295,15 +284,16 @@ xfs_inode_to_disk(
 	to->di_size = cpu_to_be64(from->di_size);
 	to->di_nblocks = cpu_to_be64(from->di_nblocks);
 	to->di_extsize = cpu_to_be32(from->di_extsize);
-	to->di_nextents = cpu_to_be32(from->di_nextents);
-	to->di_anextents = cpu_to_be16(from->di_anextents);
+	to->di_nextents = cpu_to_be32(xfs_ifork_nextents(&ip->i_df));
+	to->di_anextents = cpu_to_be16(xfs_ifork_nextents(ip->i_afp));
 	to->di_forkoff = from->di_forkoff;
-	to->di_aformat = from->di_aformat;
+	to->di_aformat = xfs_ifork_format(ip->i_afp);
 	to->di_dmevmask = cpu_to_be32(from->di_dmevmask);
 	to->di_dmstate = cpu_to_be16(from->di_dmstate);
 	to->di_flags = cpu_to_be16(from->di_flags);
 
-	if (from->di_version == 3) {
+	if (xfs_sb_version_has_v3inode(&ip->i_mount->m_sb)) {
+		to->di_version = 3;
 		to->di_changecount = cpu_to_be64(inode_peek_iversion(inode));
 		to->di_crtime.t_sec = cpu_to_be32(from->di_crtime.tv_sec);
 		to->di_crtime.t_nsec = cpu_to_be32(from->di_crtime.tv_nsec);
@@ -315,6 +305,7 @@ xfs_inode_to_disk(
 		uuid_copy(&to->di_uuid, &ip->i_mount->m_sb.sb_meta_uuid);
 		to->di_flushiter = 0;
 	} else {
+		to->di_version = 2;
 		to->di_flushiter = cpu_to_be16(from->di_flushiter);
 	}
 }
@@ -417,7 +408,7 @@ xfs_dinode_verify_forkoff(
 	struct xfs_dinode	*dip,
 	struct xfs_mount	*mp)
 {
-	if (!XFS_DFORK_Q(dip))
+	if (!dip->di_forkoff)
 		return NULL;
 
 	switch (dip->di_format)  {
@@ -428,7 +419,7 @@ xfs_dinode_verify_forkoff(
 	case XFS_DINODE_FMT_LOCAL:	/* fall through ... */
 	case XFS_DINODE_FMT_EXTENTS:    /* fall through ... */
 	case XFS_DINODE_FMT_BTREE:
-		if (dip->di_forkoff >= (XFS_LITINO(mp, dip->di_version) >> 3))
+		if (dip->di_forkoff >= (XFS_LITINO(mp) >> 3))
 			return __this_address;
 		break;
 	default:
@@ -454,7 +445,7 @@ xfs_dinode_verify(
 
 	/* Verify v3 integrity information first */
 	if (dip->di_version >= 3) {
-		if (!xfs_sb_version_hascrc(&mp->m_sb))
+		if (!xfs_sb_version_has_v3inode(&mp->m_sb))
 			return __this_address;
 		if (!xfs_verify_cksum((char *)dip, mp->m_sb.sb_inodesize,
 				      XFS_DINODE_CRC_OFF))
@@ -520,7 +511,7 @@ xfs_dinode_verify(
 		return __this_address;
 	}
 
-	if (XFS_DFORK_Q(dip)) {
+	if (dip->di_forkoff) {
 		fa = xfs_dinode_verify_fork(dip, mp, XFS_ATTR_FORK);
 		if (fa)
 			return fa;
@@ -597,125 +588,6 @@ xfs_dinode_calc_crc(
 }
 
 /*
- * Read the disk inode attributes into the in-core inode structure.
- *
- * For version 5 superblocks, if we are initialising a new inode and we are not
- * utilising the XFS_MOUNT_IKEEP inode cluster mode, we can simple build the new
- * inode core with a random generation number. If we are keeping inodes around,
- * we need to read the inode cluster to get the existing generation number off
- * disk. Further, if we are using version 4 superblocks (i.e. v1/v2 inode
- * format) then log recovery is dependent on the di_flushiter field being
- * initialised from the current on-disk value and hence we must also read the
- * inode off disk.
- */
-int
-xfs_iread(
-	xfs_mount_t	*mp,
-	xfs_trans_t	*tp,
-	xfs_inode_t	*ip,
-	uint		iget_flags)
-{
-	xfs_buf_t	*bp;
-	xfs_dinode_t	*dip;
-	xfs_failaddr_t	fa;
-	int		error;
-
-	/*
-	 * Fill in the location information in the in-core inode.
-	 */
-	error = xfs_imap(mp, tp, ip->i_ino, &ip->i_imap, iget_flags);
-	if (error)
-		return error;
-
-	/* shortcut IO on inode allocation if possible */
-	if ((iget_flags & XFS_IGET_CREATE) &&
-	    xfs_sb_version_hascrc(&mp->m_sb) &&
-	    !(mp->m_flags & XFS_MOUNT_IKEEP)) {
-		VFS_I(ip)->i_generation = prandom_u32();
-		ip->i_d.di_version = 3;
-		return 0;
-	}
-
-	/*
-	 * Get pointers to the on-disk inode and the buffer containing it.
-	 */
-	error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &bp, 0, iget_flags);
-	if (error)
-		return error;
-
-	/* even unallocated inodes are verified */
-	fa = xfs_dinode_verify(mp, ip->i_ino, dip);
-	if (fa) {
-		xfs_inode_verifier_error(ip, -EFSCORRUPTED, "dinode", dip,
-				sizeof(*dip), fa);
-		error = -EFSCORRUPTED;
-		goto out_brelse;
-	}
-
-	/*
-	 * If the on-disk inode is already linked to a directory
-	 * entry, copy all of the inode into the in-core inode.
-	 * xfs_iformat_fork() handles copying in the inode format
-	 * specific information.
-	 * Otherwise, just get the truly permanent information.
-	 */
-	if (dip->di_mode) {
-		xfs_inode_from_disk(ip, dip);
-		error = xfs_iformat_fork(ip, dip);
-		if (error)  {
-#ifdef DEBUG
-			xfs_alert(mp, "%s: xfs_iformat() returned error %d",
-				__func__, error);
-#endif /* DEBUG */
-			goto out_brelse;
-		}
-	} else {
-		/*
-		 * Partial initialisation of the in-core inode. Just the bits
-		 * that xfs_ialloc won't overwrite or relies on being correct.
-		 */
-		ip->i_d.di_version = dip->di_version;
-		VFS_I(ip)->i_generation = be32_to_cpu(dip->di_gen);
-		ip->i_d.di_flushiter = be16_to_cpu(dip->di_flushiter);
-
-		/*
-		 * Make sure to pull in the mode here as well in
-		 * case the inode is released without being used.
-		 * This ensures that xfs_inactive() will see that
-		 * the inode is already free and not try to mess
-		 * with the uninitialized part of it.
-		 */
-		VFS_I(ip)->i_mode = 0;
-	}
-
-	ASSERT(ip->i_d.di_version >= 2);
-	ip->i_delayed_blks = 0;
-
-	/*
-	 * Mark the buffer containing the inode as something to keep
-	 * around for a while.  This helps to keep recently accessed
-	 * meta-data in-core longer.
-	 */
-	xfs_buf_set_ref(bp, XFS_INO_REF);
-
-	/*
-	 * Use xfs_trans_brelse() to release the buffer containing the on-disk
-	 * inode, because it was acquired with xfs_trans_read_buf() in
-	 * xfs_imap_to_bp() above.  If tp is NULL, this is just a normal
-	 * brelse().  If we're within a transaction, then xfs_trans_brelse()
-	 * will only release the buffer if it is not dirty within the
-	 * transaction.  It will be OK to release the buffer in this case,
-	 * because inodes on disk are never destroyed and we will be locking the
-	 * new in-core inode before putting it in the cache where other
-	 * processes can find it.  Thus we don't have to worry about the inode
-	 * being changed just because we released the buffer.
-	 */
- out_brelse:
-	xfs_trans_brelse(tp, bp);
-	return error;
-}
-
-/*
  * Validate di_extsize hint.
  *
  * The rules are documented at xfs_ioctl_setattr_check_extsize().
diff --git a/fs/xfs/libxfs/xfs_inode_buf.h b/fs/xfs/libxfs/xfs_inode_buf.h
index fd94b1078722..6b08b9d060c2 100644
--- a/fs/xfs/libxfs/xfs_inode_buf.h
+++ b/fs/xfs/libxfs/xfs_inode_buf.h
@@ -16,19 +16,12 @@ struct xfs_dinode;
  * format specific structures at the appropriate time.
  */
 struct xfs_icdinode {
-	int8_t		di_version;	/* inode version */
-	int8_t		di_format;	/* format of di_c data */
 	uint16_t	di_flushiter;	/* incremented on flush */
-	uint32_t	di_uid;		/* owner's user id */
-	uint32_t	di_gid;		/* owner's group id */
 	uint32_t	di_projid;	/* owner's project id */
 	xfs_fsize_t	di_size;	/* number of bytes in file */
 	xfs_rfsblock_t	di_nblocks;	/* # of direct & btree blocks used */
 	xfs_extlen_t	di_extsize;	/* basic/minimum extent size for file */
-	xfs_extnum_t	di_nextents;	/* number of extents in data fork */
-	xfs_aextnum_t	di_anextents;	/* number of extents in attribute fork*/
 	uint8_t		di_forkoff;	/* attr fork offs, <<3 for 64b align */
-	int8_t		di_aformat;	/* format of attr fork's data */
 	uint32_t	di_dmevmask;	/* DMIG event mask */
 	uint16_t	di_dmstate;	/* DMIG state info */
 	uint16_t	di_flags;	/* random flags, XFS_DIFLAG_... */
@@ -51,24 +44,14 @@ struct xfs_imap {
 
 int	xfs_imap_to_bp(struct xfs_mount *, struct xfs_trans *,
 		       struct xfs_imap *, struct xfs_dinode **,
-		       struct xfs_buf **, uint, uint);
-int	xfs_iread(struct xfs_mount *, struct xfs_trans *,
-		  struct xfs_inode *, uint);
+		       struct xfs_buf **, uint);
 void	xfs_dinode_calc_crc(struct xfs_mount *, struct xfs_dinode *);
 void	xfs_inode_to_disk(struct xfs_inode *ip, struct xfs_dinode *to,
 			  xfs_lsn_t lsn);
-void	xfs_inode_from_disk(struct xfs_inode *ip, struct xfs_dinode *from);
+int	xfs_inode_from_disk(struct xfs_inode *ip, struct xfs_dinode *from);
 void	xfs_log_dinode_to_disk(struct xfs_log_dinode *from,
 			       struct xfs_dinode *to);
 
-bool	xfs_dinode_good_version(struct xfs_mount *mp, __u8 version);
-
-#if defined(DEBUG)
-void	xfs_inobp_check(struct xfs_mount *, struct xfs_buf *);
-#else
-#define	xfs_inobp_check(mp, bp)
-#endif /* DEBUG */
-
 xfs_failaddr_t xfs_dinode_verify(struct xfs_mount *mp, xfs_ino_t ino,
 			   struct xfs_dinode *dip);
 xfs_failaddr_t xfs_inode_validate_extsize(struct xfs_mount *mp,
diff --git a/fs/xfs/libxfs/xfs_inode_fork.c b/fs/xfs/libxfs/xfs_inode_fork.c
index ad2b9c313fd2..0cf853d42d62 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.c
+++ b/fs/xfs/libxfs/xfs_inode_fork.c
@@ -26,110 +26,6 @@
 
 kmem_zone_t *xfs_ifork_zone;
 
-STATIC int xfs_iformat_local(xfs_inode_t *, xfs_dinode_t *, int, int);
-STATIC int xfs_iformat_extents(xfs_inode_t *, xfs_dinode_t *, int);
-STATIC int xfs_iformat_btree(xfs_inode_t *, xfs_dinode_t *, int);
-
-/*
- * Copy inode type and data and attr format specific information from the
- * on-disk inode to the in-core inode and fork structures.  For fifos, devices,
- * and sockets this means set i_rdev to the proper value.  For files,
- * directories, and symlinks this means to bring in the in-line data or extent
- * pointers as well as the attribute fork.  For a fork in B-tree format, only
- * the root is immediately brought in-core.  The rest will be read in later when
- * first referenced (see xfs_iread_extents()).
- */
-int
-xfs_iformat_fork(
-	struct xfs_inode	*ip,
-	struct xfs_dinode	*dip)
-{
-	struct inode		*inode = VFS_I(ip);
-	struct xfs_attr_shortform *atp;
-	int			size;
-	int			error = 0;
-	xfs_fsize_t             di_size;
-
-	switch (inode->i_mode & S_IFMT) {
-	case S_IFIFO:
-	case S_IFCHR:
-	case S_IFBLK:
-	case S_IFSOCK:
-		ip->i_d.di_size = 0;
-		inode->i_rdev = xfs_to_linux_dev_t(xfs_dinode_get_rdev(dip));
-		break;
-
-	case S_IFREG:
-	case S_IFLNK:
-	case S_IFDIR:
-		switch (dip->di_format) {
-		case XFS_DINODE_FMT_LOCAL:
-			di_size = be64_to_cpu(dip->di_size);
-			size = (int)di_size;
-			error = xfs_iformat_local(ip, dip, XFS_DATA_FORK, size);
-			break;
-		case XFS_DINODE_FMT_EXTENTS:
-			error = xfs_iformat_extents(ip, dip, XFS_DATA_FORK);
-			break;
-		case XFS_DINODE_FMT_BTREE:
-			error = xfs_iformat_btree(ip, dip, XFS_DATA_FORK);
-			break;
-		default:
-			xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__,
-					dip, sizeof(*dip), __this_address);
-			return -EFSCORRUPTED;
-		}
-		break;
-
-	default:
-		xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__, dip,
-				sizeof(*dip), __this_address);
-		return -EFSCORRUPTED;
-	}
-	if (error)
-		return error;
-
-	if (xfs_is_reflink_inode(ip)) {
-		ASSERT(ip->i_cowfp == NULL);
-		xfs_ifork_init_cow(ip);
-	}
-
-	if (!XFS_DFORK_Q(dip))
-		return 0;
-
-	ASSERT(ip->i_afp == NULL);
-	ip->i_afp = kmem_zone_zalloc(xfs_ifork_zone, KM_NOFS);
-
-	switch (dip->di_aformat) {
-	case XFS_DINODE_FMT_LOCAL:
-		atp = (xfs_attr_shortform_t *)XFS_DFORK_APTR(dip);
-		size = be16_to_cpu(atp->hdr.totsize);
-
-		error = xfs_iformat_local(ip, dip, XFS_ATTR_FORK, size);
-		break;
-	case XFS_DINODE_FMT_EXTENTS:
-		error = xfs_iformat_extents(ip, dip, XFS_ATTR_FORK);
-		break;
-	case XFS_DINODE_FMT_BTREE:
-		error = xfs_iformat_btree(ip, dip, XFS_ATTR_FORK);
-		break;
-	default:
-		xfs_inode_verifier_error(ip, error, __func__, dip,
-				sizeof(*dip), __this_address);
-		error = -EFSCORRUPTED;
-		break;
-	}
-	if (error) {
-		kmem_cache_free(xfs_ifork_zone, ip->i_afp);
-		ip->i_afp = NULL;
-		if (ip->i_cowfp)
-			kmem_cache_free(xfs_ifork_zone, ip->i_cowfp);
-		ip->i_cowfp = NULL;
-		xfs_idestroy_fork(ip, XFS_DATA_FORK);
-	}
-	return error;
-}
-
 void
 xfs_init_local_fork(
 	struct xfs_inode	*ip,
@@ -183,7 +79,7 @@ xfs_iformat_local(
 	 */
 	if (unlikely(size > XFS_DFORK_SIZE(dip, ip->i_mount, whichfork))) {
 		xfs_warn(ip->i_mount,
-	"corrupt inode %Lu (bad size %d for local fork, size = %d).",
+	"corrupt inode %Lu (bad size %d for local fork, size = %zd).",
 			(unsigned long long) ip->i_ino, size,
 			XFS_DFORK_SIZE(dip, ip->i_mount, whichfork));
 		xfs_inode_verifier_error(ip, -EFSCORRUPTED,
@@ -292,12 +188,11 @@ xfs_iformat_btree(
 	 * or the number of extents is greater than the number of
 	 * blocks.
 	 */
-	if (unlikely(XFS_IFORK_NEXTENTS(ip, whichfork) <=
-					XFS_IFORK_MAXEXT(ip, whichfork) ||
+	if (unlikely(ifp->if_nextents <= XFS_IFORK_MAXEXT(ip, whichfork) ||
 		     nrecs == 0 ||
 		     XFS_BMDR_SPACE_CALC(nrecs) >
 					XFS_DFORK_SIZE(dip, mp, whichfork) ||
-		     XFS_IFORK_NEXTENTS(ip, whichfork) > ip->i_d.di_nblocks) ||
+		     ifp->if_nextents > ip->i_d.di_nblocks) ||
 		     level == 0 || level > XFS_BTREE_MAXLEVELS) {
 		xfs_warn(mp, "corrupt inode %Lu (btree).",
 					(unsigned long long) ip->i_ino);
@@ -325,6 +220,110 @@ xfs_iformat_btree(
 	return 0;
 }
 
+int
+xfs_iformat_data_fork(
+	struct xfs_inode	*ip,
+	struct xfs_dinode	*dip)
+{
+	struct inode		*inode = VFS_I(ip);
+	int			error;
+
+	/*
+	 * Initialize the extent count early, as the per-format routines may
+	 * depend on it.
+	 */
+	ip->i_df.if_format = dip->di_format;
+	ip->i_df.if_nextents = be32_to_cpu(dip->di_nextents);
+
+	switch (inode->i_mode & S_IFMT) {
+	case S_IFIFO:
+	case S_IFCHR:
+	case S_IFBLK:
+	case S_IFSOCK:
+		ip->i_d.di_size = 0;
+		inode->i_rdev = xfs_to_linux_dev_t(xfs_dinode_get_rdev(dip));
+		return 0;
+	case S_IFREG:
+	case S_IFLNK:
+	case S_IFDIR:
+		switch (ip->i_df.if_format) {
+		case XFS_DINODE_FMT_LOCAL:
+			error = xfs_iformat_local(ip, dip, XFS_DATA_FORK,
+					be64_to_cpu(dip->di_size));
+			if (!error)
+				error = xfs_ifork_verify_local_data(ip);
+			return error;
+		case XFS_DINODE_FMT_EXTENTS:
+			return xfs_iformat_extents(ip, dip, XFS_DATA_FORK);
+		case XFS_DINODE_FMT_BTREE:
+			return xfs_iformat_btree(ip, dip, XFS_DATA_FORK);
+		default:
+			xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__,
+					dip, sizeof(*dip), __this_address);
+			return -EFSCORRUPTED;
+		}
+		break;
+	default:
+		xfs_inode_verifier_error(ip, -EFSCORRUPTED, __func__, dip,
+				sizeof(*dip), __this_address);
+		return -EFSCORRUPTED;
+	}
+}
+
+static uint16_t
+xfs_dfork_attr_shortform_size(
+	struct xfs_dinode		*dip)
+{
+	struct xfs_attr_shortform	*atp =
+		(struct xfs_attr_shortform *)XFS_DFORK_APTR(dip);
+
+	return be16_to_cpu(atp->hdr.totsize);
+}
+
+int
+xfs_iformat_attr_fork(
+	struct xfs_inode	*ip,
+	struct xfs_dinode	*dip)
+{
+	int			error = 0;
+
+	/*
+	 * Initialize the extent count early, as the per-format routines may
+	 * depend on it.
+	 */
+	ip->i_afp = kmem_cache_zalloc(xfs_ifork_zone, GFP_NOFS | __GFP_NOFAIL);
+	ip->i_afp->if_format = dip->di_aformat;
+	if (unlikely(ip->i_afp->if_format == 0)) /* pre IRIX 6.2 file system */
+		ip->i_afp->if_format = XFS_DINODE_FMT_EXTENTS;
+	ip->i_afp->if_nextents = be16_to_cpu(dip->di_anextents);
+
+	switch (ip->i_afp->if_format) {
+	case XFS_DINODE_FMT_LOCAL:
+		error = xfs_iformat_local(ip, dip, XFS_ATTR_FORK,
+				xfs_dfork_attr_shortform_size(dip));
+		if (!error)
+			error = xfs_ifork_verify_local_attr(ip);
+		break;
+	case XFS_DINODE_FMT_EXTENTS:
+		error = xfs_iformat_extents(ip, dip, XFS_ATTR_FORK);
+		break;
+	case XFS_DINODE_FMT_BTREE:
+		error = xfs_iformat_btree(ip, dip, XFS_ATTR_FORK);
+		break;
+	default:
+		xfs_inode_verifier_error(ip, error, __func__, dip,
+				sizeof(*dip), __this_address);
+		error = -EFSCORRUPTED;
+		break;
+	}
+
+	if (error) {
+		kmem_cache_free(xfs_ifork_zone, ip->i_afp);
+		ip->i_afp = NULL;
+	}
+	return error;
+}
+
 /*
  * Reallocate the space for if_broot based on the number of records
  * being added or deleted as indicated in rec_diff.  Move the records
@@ -504,38 +503,24 @@ xfs_idata_realloc(
 
 void
 xfs_idestroy_fork(
-	xfs_inode_t	*ip,
-	int		whichfork)
+	struct xfs_ifork	*ifp)
 {
-	struct xfs_ifork	*ifp;
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
 	if (ifp->if_broot != NULL) {
 		kmem_free(ifp->if_broot);
 		ifp->if_broot = NULL;
 	}
 
 	/*
-	 * If the format is local, then we can't have an extents
-	 * array so just look for an inline data array.  If we're
-	 * not local then we may or may not have an extents list,
-	 * so check and free it up if we do.
+	 * If the format is local, then we can't have an extents array so just
+	 * look for an inline data array.  If we're not local then we may or may
+	 * not have an extents list, so check and free it up if we do.
 	 */
-	if (XFS_IFORK_FORMAT(ip, whichfork) == XFS_DINODE_FMT_LOCAL) {
-		if (ifp->if_u1.if_data != NULL) {
-			kmem_free(ifp->if_u1.if_data);
-			ifp->if_u1.if_data = NULL;
-		}
-	} else if ((ifp->if_flags & XFS_IFEXTENTS) && ifp->if_height) {
-		xfs_iext_destroy(ifp);
-	}
-
-	if (whichfork == XFS_ATTR_FORK) {
-		kmem_cache_free(xfs_ifork_zone, ip->i_afp);
-		ip->i_afp = NULL;
-	} else if (whichfork == XFS_COW_FORK) {
-		kmem_cache_free(xfs_ifork_zone, ip->i_cowfp);
-		ip->i_cowfp = NULL;
+	if (ifp->if_format == XFS_DINODE_FMT_LOCAL) {
+		kmem_free(ifp->if_u1.if_data);
+		ifp->if_u1.if_data = NULL;
+	} else if (ifp->if_flags & XFS_IFEXTENTS) {
+		if (ifp->if_height)
+			xfs_iext_destroy(ifp);
 	}
 }
 
@@ -592,7 +577,7 @@ void
 xfs_iflush_fork(
 	xfs_inode_t		*ip,
 	xfs_dinode_t		*dip,
-	xfs_inode_log_item_t	*iip,
+	struct xfs_inode_log_item *iip,
 	int			whichfork)
 {
 	char			*cp;
@@ -618,7 +603,7 @@ xfs_iflush_fork(
 	}
 	cp = XFS_DFORK_PTR(dip, whichfork);
 	mp = ip->i_mount;
-	switch (XFS_IFORK_FORMAT(ip, whichfork)) {
+	switch (ifp->if_format) {
 	case XFS_DINODE_FMT_LOCAL:
 		if ((iip->ili_fields & dataflag[whichfork]) &&
 		    (ifp->if_bytes > 0)) {
@@ -633,7 +618,7 @@ xfs_iflush_fork(
 		       !(iip->ili_fields & extflag[whichfork]));
 		if ((iip->ili_fields & extflag[whichfork]) &&
 		    (ifp->if_bytes > 0)) {
-			ASSERT(XFS_IFORK_NEXTENTS(ip, whichfork) > 0);
+			ASSERT(ifp->if_nextents > 0);
 			(void)xfs_iextents_copy(ip, (xfs_bmbt_rec_t *)cp,
 				whichfork);
 		}
@@ -688,51 +673,58 @@ xfs_ifork_init_cow(
 	if (ip->i_cowfp)
 		return;
 
-	ip->i_cowfp = kmem_zone_zalloc(xfs_ifork_zone,
-				       KM_NOFS);
+	ip->i_cowfp = kmem_cache_zalloc(xfs_ifork_zone,
+				       GFP_NOFS | __GFP_NOFAIL);
 	ip->i_cowfp->if_flags = XFS_IFEXTENTS;
-	ip->i_cformat = XFS_DINODE_FMT_EXTENTS;
-	ip->i_cnextents = 0;
+	ip->i_cowfp->if_format = XFS_DINODE_FMT_EXTENTS;
 }
 
-/* Default fork content verifiers. */
-struct xfs_ifork_ops xfs_default_ifork_ops = {
-	.verify_attr	= xfs_attr_shortform_verify,
-	.verify_dir	= xfs_dir2_sf_verify,
-	.verify_symlink	= xfs_symlink_shortform_verify,
-};
-
 /* Verify the inline contents of the data fork of an inode. */
-xfs_failaddr_t
-xfs_ifork_verify_data(
-	struct xfs_inode	*ip,
-	struct xfs_ifork_ops	*ops)
+int
+xfs_ifork_verify_local_data(
+	struct xfs_inode	*ip)
 {
-	/* Non-local data fork, we're done. */
-	if (ip->i_d.di_format != XFS_DINODE_FMT_LOCAL)
-		return NULL;
+	xfs_failaddr_t		fa = NULL;
 
-	/* Check the inline data fork if there is one. */
 	switch (VFS_I(ip)->i_mode & S_IFMT) {
 	case S_IFDIR:
-		return ops->verify_dir(ip);
+		fa = xfs_dir2_sf_verify(ip);
+		break;
 	case S_IFLNK:
-		return ops->verify_symlink(ip);
+		fa = xfs_symlink_shortform_verify(ip);
+		break;
 	default:
-		return NULL;
+		break;
 	}
+
+	if (fa) {
+		xfs_inode_verifier_error(ip, -EFSCORRUPTED, "data fork",
+				ip->i_df.if_u1.if_data, ip->i_df.if_bytes, fa);
+		return -EFSCORRUPTED;
+	}
+
+	return 0;
 }
 
 /* Verify the inline contents of the attr fork of an inode. */
-xfs_failaddr_t
-xfs_ifork_verify_attr(
-	struct xfs_inode	*ip,
-	struct xfs_ifork_ops	*ops)
+int
+xfs_ifork_verify_local_attr(
+	struct xfs_inode	*ip)
 {
-	/* There has to be an attr fork allocated if aformat is local. */
-	if (ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)
-		return NULL;
-	if (!XFS_IFORK_PTR(ip, XFS_ATTR_FORK))
-		return __this_address;
-	return ops->verify_attr(ip);
+	struct xfs_ifork	*ifp = ip->i_afp;
+	xfs_failaddr_t		fa;
+
+	if (!ifp)
+		fa = __this_address;
+	else
+		fa = xfs_attr_shortform_verify(ip);
+
+	if (fa) {
+		xfs_inode_verifier_error(ip, -EFSCORRUPTED, "attr fork",
+				ifp ? ifp->if_u1.if_data : NULL,
+				ifp ? ifp->if_bytes : 0, fa);
+		return -EFSCORRUPTED;
+	}
+
+	return 0;
 }
diff --git a/fs/xfs/libxfs/xfs_inode_fork.h b/fs/xfs/libxfs/xfs_inode_fork.h
index 500333d0101e..a4953e95c4f3 100644
--- a/fs/xfs/libxfs/xfs_inode_fork.h
+++ b/fs/xfs/libxfs/xfs_inode_fork.h
@@ -23,6 +23,8 @@ struct xfs_ifork {
 	} if_u1;
 	short			if_broot_bytes;	/* bytes allocated for root */
 	unsigned char		if_flags;	/* per-fork flags */
+	int8_t			if_format;	/* format of this fork */
+	xfs_extnum_t		if_nextents;	/* # of extents in this fork */
 };
 
 /*
@@ -46,57 +48,45 @@ struct xfs_ifork {
 			(ip)->i_afp : \
 			(ip)->i_cowfp))
 #define XFS_IFORK_DSIZE(ip) \
-	(XFS_IFORK_Q(ip) ? \
-		XFS_IFORK_BOFF(ip) : \
-		XFS_LITINO((ip)->i_mount, (ip)->i_d.di_version))
+	(XFS_IFORK_Q(ip) ? XFS_IFORK_BOFF(ip) : XFS_LITINO((ip)->i_mount))
 #define XFS_IFORK_ASIZE(ip) \
-	(XFS_IFORK_Q(ip) ? \
-		XFS_LITINO((ip)->i_mount, (ip)->i_d.di_version) - \
-			XFS_IFORK_BOFF(ip) : \
-		0)
+	(XFS_IFORK_Q(ip) ? XFS_LITINO((ip)->i_mount) - XFS_IFORK_BOFF(ip) : 0)
 #define XFS_IFORK_SIZE(ip,w) \
 	((w) == XFS_DATA_FORK ? \
 		XFS_IFORK_DSIZE(ip) : \
 		((w) == XFS_ATTR_FORK ? \
 			XFS_IFORK_ASIZE(ip) : \
 			0))
-#define XFS_IFORK_FORMAT(ip,w) \
-	((w) == XFS_DATA_FORK ? \
-		(ip)->i_d.di_format : \
-		((w) == XFS_ATTR_FORK ? \
-			(ip)->i_d.di_aformat : \
-			(ip)->i_cformat))
-#define XFS_IFORK_FMT_SET(ip,w,n) \
-	((w) == XFS_DATA_FORK ? \
-		((ip)->i_d.di_format = (n)) : \
-		((w) == XFS_ATTR_FORK ? \
-			((ip)->i_d.di_aformat = (n)) : \
-			((ip)->i_cformat = (n))))
-#define XFS_IFORK_NEXTENTS(ip,w) \
-	((w) == XFS_DATA_FORK ? \
-		(ip)->i_d.di_nextents : \
-		((w) == XFS_ATTR_FORK ? \
-			(ip)->i_d.di_anextents : \
-			(ip)->i_cnextents))
-#define XFS_IFORK_NEXT_SET(ip,w,n) \
-	((w) == XFS_DATA_FORK ? \
-		((ip)->i_d.di_nextents = (n)) : \
-		((w) == XFS_ATTR_FORK ? \
-			((ip)->i_d.di_anextents = (n)) : \
-			((ip)->i_cnextents = (n))))
 #define XFS_IFORK_MAXEXT(ip, w) \
 	(XFS_IFORK_SIZE(ip, w) / sizeof(xfs_bmbt_rec_t))
 
-#define xfs_ifork_has_extents(ip, w) \
-	(XFS_IFORK_FORMAT((ip), (w)) == XFS_DINODE_FMT_EXTENTS || \
-	 XFS_IFORK_FORMAT((ip), (w)) == XFS_DINODE_FMT_BTREE)
+static inline bool xfs_ifork_has_extents(struct xfs_ifork *ifp)
+{
+	return ifp->if_format == XFS_DINODE_FMT_EXTENTS ||
+		ifp->if_format == XFS_DINODE_FMT_BTREE;
+}
+
+static inline xfs_extnum_t xfs_ifork_nextents(struct xfs_ifork *ifp)
+{
+	if (!ifp)
+		return 0;
+	return ifp->if_nextents;
+}
+
+static inline int8_t xfs_ifork_format(struct xfs_ifork *ifp)
+{
+	if (!ifp)
+		return XFS_DINODE_FMT_EXTENTS;
+	return ifp->if_format;
+}
 
 struct xfs_ifork *xfs_iext_state_to_fork(struct xfs_inode *ip, int state);
 
-int		xfs_iformat_fork(struct xfs_inode *, struct xfs_dinode *);
+int		xfs_iformat_data_fork(struct xfs_inode *, struct xfs_dinode *);
+int		xfs_iformat_attr_fork(struct xfs_inode *, struct xfs_dinode *);
 void		xfs_iflush_fork(struct xfs_inode *, struct xfs_dinode *,
 				struct xfs_inode_log_item *, int);
-void		xfs_idestroy_fork(struct xfs_inode *, int);
+void		xfs_idestroy_fork(struct xfs_ifork *ifp);
 void		xfs_idata_realloc(struct xfs_inode *ip, int64_t byte_diff,
 				int whichfork);
 void		xfs_iroot_realloc(struct xfs_inode *, int, int);
@@ -180,18 +170,7 @@ extern struct kmem_zone	*xfs_ifork_zone;
 
 extern void xfs_ifork_init_cow(struct xfs_inode *ip);
 
-typedef xfs_failaddr_t (*xfs_ifork_verifier_t)(struct xfs_inode *);
-
-struct xfs_ifork_ops {
-	xfs_ifork_verifier_t	verify_symlink;
-	xfs_ifork_verifier_t	verify_dir;
-	xfs_ifork_verifier_t	verify_attr;
-};
-extern struct xfs_ifork_ops	xfs_default_ifork_ops;
-
-xfs_failaddr_t xfs_ifork_verify_data(struct xfs_inode *ip,
-		struct xfs_ifork_ops *ops);
-xfs_failaddr_t xfs_ifork_verify_attr(struct xfs_inode *ip,
-		struct xfs_ifork_ops *ops);
+int xfs_ifork_verify_local_data(struct xfs_inode *ip);
+int xfs_ifork_verify_local_attr(struct xfs_inode *ip);
 
 #endif	/* __XFS_INODE_FORK_H__ */
diff --git a/fs/xfs/libxfs/xfs_log_format.h b/fs/xfs/libxfs/xfs_log_format.h
index 9bac0d2e56dc..e3400c9c71cd 100644
--- a/fs/xfs/libxfs/xfs_log_format.h
+++ b/fs/xfs/libxfs/xfs_log_format.h
@@ -424,12 +424,10 @@ struct xfs_log_dinode {
 	/* structure must be padded to 64 bit alignment */
 };
 
-static inline uint xfs_log_dinode_size(int version)
-{
-	if (version == 3)
-		return sizeof(struct xfs_log_dinode);
-	return offsetof(struct xfs_log_dinode, di_next_unlinked);
-}
+#define xfs_log_dinode_size(mp)						\
+	(xfs_sb_version_has_v3inode(&(mp)->m_sb) ?			\
+		sizeof(struct xfs_log_dinode) :				\
+		offsetof(struct xfs_log_dinode, di_next_unlinked))
 
 /*
  * Buffer Log Format definitions
diff --git a/fs/xfs/libxfs/xfs_log_recover.h b/fs/xfs/libxfs/xfs_log_recover.h
index 3bf671637a91..641132d0e39d 100644
--- a/fs/xfs/libxfs/xfs_log_recover.h
+++ b/fs/xfs/libxfs/xfs_log_recover.h
@@ -7,6 +7,73 @@
 #define __XFS_LOG_RECOVER_H__
 
 /*
+ * Each log item type (XFS_LI_*) gets its own xlog_recover_item_ops to
+ * define how recovery should work for that type of log item.
+ */
+struct xlog_recover_item;
+
+/* Sorting hat for log items as they're read in. */
+enum xlog_recover_reorder {
+	XLOG_REORDER_BUFFER_LIST,
+	XLOG_REORDER_ITEM_LIST,
+	XLOG_REORDER_INODE_BUFFER_LIST,
+	XLOG_REORDER_CANCEL_LIST,
+};
+
+struct xlog_recover_item_ops {
+	uint16_t	item_type;	/* XFS_LI_* type code. */
+
+	/*
+	 * Help sort recovered log items into the order required to replay them
+	 * correctly.  Log item types that always use XLOG_REORDER_ITEM_LIST do
+	 * not have to supply a function here.  See the comment preceding
+	 * xlog_recover_reorder_trans for more details about what the return
+	 * values mean.
+	 */
+	enum xlog_recover_reorder (*reorder)(struct xlog_recover_item *item);
+
+	/* Start readahead for pass2, if provided. */
+	void (*ra_pass2)(struct xlog *log, struct xlog_recover_item *item);
+
+	/* Do whatever work we need to do for pass1, if provided. */
+	int (*commit_pass1)(struct xlog *log, struct xlog_recover_item *item);
+
+	/*
+	 * This function should do whatever work is needed for pass2 of log
+	 * recovery, if provided.
+	 *
+	 * If the recovered item is an intent item, this function should parse
+	 * the recovered item to construct an in-core log intent item and
+	 * insert it into the AIL.  The in-core log intent item should have 1
+	 * refcount so that the item is freed either (a) when we commit the
+	 * recovered log item for the intent-done item; (b) replay the work and
+	 * log a new intent-done item; or (c) recovery fails and we have to
+	 * abort.
+	 *
+	 * If the recovered item is an intent-done item, this function should
+	 * parse the recovered item to find the id of the corresponding intent
+	 * log item.  Next, it should find the in-core log intent item in the
+	 * AIL and release it.
+	 */
+	int (*commit_pass2)(struct xlog *log, struct list_head *buffer_list,
+			    struct xlog_recover_item *item, xfs_lsn_t lsn);
+};
+
+extern const struct xlog_recover_item_ops xlog_icreate_item_ops;
+extern const struct xlog_recover_item_ops xlog_buf_item_ops;
+extern const struct xlog_recover_item_ops xlog_inode_item_ops;
+extern const struct xlog_recover_item_ops xlog_dquot_item_ops;
+extern const struct xlog_recover_item_ops xlog_quotaoff_item_ops;
+extern const struct xlog_recover_item_ops xlog_bui_item_ops;
+extern const struct xlog_recover_item_ops xlog_bud_item_ops;
+extern const struct xlog_recover_item_ops xlog_efi_item_ops;
+extern const struct xlog_recover_item_ops xlog_efd_item_ops;
+extern const struct xlog_recover_item_ops xlog_rui_item_ops;
+extern const struct xlog_recover_item_ops xlog_rud_item_ops;
+extern const struct xlog_recover_item_ops xlog_cui_item_ops;
+extern const struct xlog_recover_item_ops xlog_cud_item_ops;
+
+/*
  * Macros, structures, prototypes for internal log manager use.
  */
 
@@ -22,13 +89,13 @@
 /*
  * item headers are in ri_buf[0].  Additional buffers follow.
  */
-typedef struct xlog_recover_item {
+struct xlog_recover_item {
 	struct list_head	ri_list;
-	int			ri_type;
 	int			ri_cnt;	/* count of regions found */
 	int			ri_total;	/* total regions */
-	xfs_log_iovec_t		*ri_buf;	/* ptr to regions buffer */
-} xlog_recover_item_t;
+	struct xfs_log_iovec	*ri_buf;	/* ptr to regions buffer */
+	const struct xlog_recover_item_ops *ri_ops;
+};
 
 struct xlog_recover {
 	struct hlist_node	r_list;
@@ -51,4 +118,12 @@ struct xlog_recover {
 #define	XLOG_RECOVER_PASS1	1
 #define	XLOG_RECOVER_PASS2	2
 
+void xlog_buf_readahead(struct xlog *log, xfs_daddr_t blkno, uint len,
+		const struct xfs_buf_ops *ops);
+bool xlog_is_buffer_cancelled(struct xlog *log, xfs_daddr_t blkno, uint len);
+void xlog_recover_iodone(struct xfs_buf *bp);
+
+void xlog_recover_release_intent(struct xlog *log, unsigned short intent_type,
+		uint64_t intent_id);
+
 #endif	/* __XFS_LOG_RECOVER_H__ */
diff --git a/fs/xfs/libxfs/xfs_quota_defs.h b/fs/xfs/libxfs/xfs_quota_defs.h
index b2113b17e53c..076bdc7037ee 100644
--- a/fs/xfs/libxfs/xfs_quota_defs.h
+++ b/fs/xfs/libxfs/xfs_quota_defs.h
@@ -18,23 +18,22 @@
 typedef uint64_t	xfs_qcnt_t;
 typedef uint16_t	xfs_qwarncnt_t;
 
+typedef uint8_t		xfs_dqtype_t;
+
+#define XFS_DQTYPE_STRINGS \
+	{ XFS_DQTYPE_USER,	"USER" }, \
+	{ XFS_DQTYPE_PROJ,	"PROJ" }, \
+	{ XFS_DQTYPE_GROUP,	"GROUP" }
+
 /*
  * flags for q_flags field in the dquot.
  */
-#define XFS_DQ_USER		0x0001		/* a user quota */
-#define XFS_DQ_PROJ		0x0002		/* project quota */
-#define XFS_DQ_GROUP		0x0004		/* a group quota */
-#define XFS_DQ_DIRTY		0x0008		/* dquot is dirty */
-#define XFS_DQ_FREEING		0x0010		/* dquot is being torn down */
-
-#define XFS_DQ_ALLTYPES		(XFS_DQ_USER|XFS_DQ_PROJ|XFS_DQ_GROUP)
+#define XFS_DQFLAG_DIRTY	(1 << 0)	/* dquot is dirty */
+#define XFS_DQFLAG_FREEING	(1 << 1)	/* dquot is being torn down */
 
-#define XFS_DQ_FLAGS \
-	{ XFS_DQ_USER,		"USER" }, \
-	{ XFS_DQ_PROJ,		"PROJ" }, \
-	{ XFS_DQ_GROUP,		"GROUP" }, \
-	{ XFS_DQ_DIRTY,		"DIRTY" }, \
-	{ XFS_DQ_FREEING,	"FREEING" }
+#define XFS_DQFLAG_STRINGS \
+	{ XFS_DQFLAG_DIRTY,	"DIRTY" }, \
+	{ XFS_DQFLAG_FREEING,	"FREEING" }
 
 /*
  * We have the possibility of all three quota types being active at once, and
@@ -100,7 +99,6 @@ typedef uint16_t	xfs_qwarncnt_t;
 #define XFS_QMOPT_FORCE_RES	0x0000010 /* ignore quota limits */
 #define XFS_QMOPT_SBVERSION	0x0000040 /* change superblock version num */
 #define XFS_QMOPT_GQUOTA	0x0002000 /* group dquot requested */
-#define XFS_QMOPT_ENOSPC	0x0004000 /* enospc instead of edquot (prj) */
 
 /*
  * flags to xfs_trans_mod_dquot to indicate which field needs to be
@@ -138,11 +136,11 @@ typedef uint16_t	xfs_qwarncnt_t;
 #define XFS_QMOPT_RESBLK_MASK	(XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_RES_RTBLKS)
 
 extern xfs_failaddr_t xfs_dquot_verify(struct xfs_mount *mp,
-		struct xfs_disk_dquot *ddq, xfs_dqid_t id, uint type);
+		struct xfs_disk_dquot *ddq, xfs_dqid_t id);
 extern xfs_failaddr_t xfs_dqblk_verify(struct xfs_mount *mp,
-		struct xfs_dqblk *dqb, xfs_dqid_t id, uint type);
+		struct xfs_dqblk *dqb, xfs_dqid_t id);
 extern int xfs_calc_dquots_per_chunk(unsigned int nbblks);
 extern void xfs_dqblk_repair(struct xfs_mount *mp, struct xfs_dqblk *dqb,
-		xfs_dqid_t id, uint type);
+		xfs_dqid_t id, xfs_dqtype_t type);
 
 #endif	/* __XFS_QUOTA_H__ */
diff --git a/fs/xfs/libxfs/xfs_refcount.c b/fs/xfs/libxfs/xfs_refcount.c
index 6e1665f2cb67..2076627243b0 100644
--- a/fs/xfs/libxfs/xfs_refcount.c
+++ b/fs/xfs/libxfs/xfs_refcount.c
@@ -46,7 +46,7 @@ xfs_refcount_lookup_le(
 	xfs_agblock_t		bno,
 	int			*stat)
 {
-	trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_private.a.agno, bno,
+	trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_ag.agno, bno,
 			XFS_LOOKUP_LE);
 	cur->bc_rec.rc.rc_startblock = bno;
 	cur->bc_rec.rc.rc_blockcount = 0;
@@ -63,7 +63,7 @@ xfs_refcount_lookup_ge(
 	xfs_agblock_t		bno,
 	int			*stat)
 {
-	trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_private.a.agno, bno,
+	trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_ag.agno, bno,
 			XFS_LOOKUP_GE);
 	cur->bc_rec.rc.rc_startblock = bno;
 	cur->bc_rec.rc.rc_blockcount = 0;
@@ -80,7 +80,7 @@ xfs_refcount_lookup_eq(
 	xfs_agblock_t		bno,
 	int			*stat)
 {
-	trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_private.a.agno, bno,
+	trace_xfs_refcount_lookup(cur->bc_mp, cur->bc_ag.agno, bno,
 			XFS_LOOKUP_LE);
 	cur->bc_rec.rc.rc_startblock = bno;
 	cur->bc_rec.rc.rc_blockcount = 0;
@@ -108,7 +108,7 @@ xfs_refcount_get_rec(
 	int				*stat)
 {
 	struct xfs_mount		*mp = cur->bc_mp;
-	xfs_agnumber_t			agno = cur->bc_private.a.agno;
+	xfs_agnumber_t			agno = cur->bc_ag.agno;
 	union xfs_btree_rec		*rec;
 	int				error;
 	xfs_agblock_t			realstart;
@@ -119,7 +119,7 @@ xfs_refcount_get_rec(
 
 	xfs_refcount_btrec_to_irec(rec, irec);
 
-	agno = cur->bc_private.a.agno;
+	agno = cur->bc_ag.agno;
 	if (irec->rc_blockcount == 0 || irec->rc_blockcount > MAXREFCEXTLEN)
 		goto out_bad_rec;
 
@@ -144,7 +144,7 @@ xfs_refcount_get_rec(
 	if (irec->rc_refcount == 0 || irec->rc_refcount > MAXREFCOUNT)
 		goto out_bad_rec;
 
-	trace_xfs_refcount_get(cur->bc_mp, cur->bc_private.a.agno, irec);
+	trace_xfs_refcount_get(cur->bc_mp, cur->bc_ag.agno, irec);
 	return 0;
 
 out_bad_rec:
@@ -169,14 +169,14 @@ xfs_refcount_update(
 	union xfs_btree_rec	rec;
 	int			error;
 
-	trace_xfs_refcount_update(cur->bc_mp, cur->bc_private.a.agno, irec);
+	trace_xfs_refcount_update(cur->bc_mp, cur->bc_ag.agno, irec);
 	rec.refc.rc_startblock = cpu_to_be32(irec->rc_startblock);
 	rec.refc.rc_blockcount = cpu_to_be32(irec->rc_blockcount);
 	rec.refc.rc_refcount = cpu_to_be32(irec->rc_refcount);
 	error = xfs_btree_update(cur, &rec);
 	if (error)
 		trace_xfs_refcount_update_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -193,7 +193,7 @@ xfs_refcount_insert(
 {
 	int				error;
 
-	trace_xfs_refcount_insert(cur->bc_mp, cur->bc_private.a.agno, irec);
+	trace_xfs_refcount_insert(cur->bc_mp, cur->bc_ag.agno, irec);
 	cur->bc_rec.rc.rc_startblock = irec->rc_startblock;
 	cur->bc_rec.rc.rc_blockcount = irec->rc_blockcount;
 	cur->bc_rec.rc.rc_refcount = irec->rc_refcount;
@@ -208,7 +208,7 @@ xfs_refcount_insert(
 out_error:
 	if (error)
 		trace_xfs_refcount_insert_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -234,7 +234,7 @@ xfs_refcount_delete(
 		error = -EFSCORRUPTED;
 		goto out_error;
 	}
-	trace_xfs_refcount_delete(cur->bc_mp, cur->bc_private.a.agno, &irec);
+	trace_xfs_refcount_delete(cur->bc_mp, cur->bc_ag.agno, &irec);
 	error = xfs_btree_delete(cur, i);
 	if (XFS_IS_CORRUPT(cur->bc_mp, *i != 1)) {
 		error = -EFSCORRUPTED;
@@ -246,7 +246,7 @@ xfs_refcount_delete(
 out_error:
 	if (error)
 		trace_xfs_refcount_delete_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -366,7 +366,7 @@ xfs_refcount_split_extent(
 		return 0;
 
 	*shape_changed = true;
-	trace_xfs_refcount_split_extent(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_refcount_split_extent(cur->bc_mp, cur->bc_ag.agno,
 			&rcext, agbno);
 
 	/* Establish the right extent. */
@@ -391,7 +391,7 @@ xfs_refcount_split_extent(
 
 out_error:
 	trace_xfs_refcount_split_extent_error(cur->bc_mp,
-			cur->bc_private.a.agno, error, _RET_IP_);
+			cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -411,7 +411,7 @@ xfs_refcount_merge_center_extents(
 	int				found_rec;
 
 	trace_xfs_refcount_merge_center_extents(cur->bc_mp,
-			cur->bc_private.a.agno, left, center, right);
+			cur->bc_ag.agno, left, center, right);
 
 	/*
 	 * Make sure the center and right extents are not in the btree.
@@ -468,7 +468,7 @@ xfs_refcount_merge_center_extents(
 
 out_error:
 	trace_xfs_refcount_merge_center_extents_error(cur->bc_mp,
-			cur->bc_private.a.agno, error, _RET_IP_);
+			cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -487,7 +487,7 @@ xfs_refcount_merge_left_extent(
 	int				found_rec;
 
 	trace_xfs_refcount_merge_left_extent(cur->bc_mp,
-			cur->bc_private.a.agno, left, cleft);
+			cur->bc_ag.agno, left, cleft);
 
 	/* If the extent at agbno (cleft) wasn't synthesized, remove it. */
 	if (cleft->rc_refcount > 1) {
@@ -530,7 +530,7 @@ xfs_refcount_merge_left_extent(
 
 out_error:
 	trace_xfs_refcount_merge_left_extent_error(cur->bc_mp,
-			cur->bc_private.a.agno, error, _RET_IP_);
+			cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -548,7 +548,7 @@ xfs_refcount_merge_right_extent(
 	int				found_rec;
 
 	trace_xfs_refcount_merge_right_extent(cur->bc_mp,
-			cur->bc_private.a.agno, cright, right);
+			cur->bc_ag.agno, cright, right);
 
 	/*
 	 * If the extent ending at agbno+aglen (cright) wasn't synthesized,
@@ -594,7 +594,7 @@ xfs_refcount_merge_right_extent(
 
 out_error:
 	trace_xfs_refcount_merge_right_extent_error(cur->bc_mp,
-			cur->bc_private.a.agno, error, _RET_IP_);
+			cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -679,13 +679,13 @@ xfs_refcount_find_left_extents(
 		cleft->rc_blockcount = aglen;
 		cleft->rc_refcount = 1;
 	}
-	trace_xfs_refcount_find_left_extent(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_refcount_find_left_extent(cur->bc_mp, cur->bc_ag.agno,
 			left, cleft, agbno);
 	return error;
 
 out_error:
 	trace_xfs_refcount_find_left_extent_error(cur->bc_mp,
-			cur->bc_private.a.agno, error, _RET_IP_);
+			cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -768,13 +768,13 @@ xfs_refcount_find_right_extents(
 		cright->rc_blockcount = aglen;
 		cright->rc_refcount = 1;
 	}
-	trace_xfs_refcount_find_right_extent(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_refcount_find_right_extent(cur->bc_mp, cur->bc_ag.agno,
 			cright, right, agbno + aglen);
 	return error;
 
 out_error:
 	trace_xfs_refcount_find_right_extent_error(cur->bc_mp,
-			cur->bc_private.a.agno, error, _RET_IP_);
+			cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -883,7 +883,7 @@ xfs_refcount_still_have_space(
 {
 	unsigned long			overhead;
 
-	overhead = cur->bc_private.a.priv.refc.shape_changes *
+	overhead = cur->bc_ag.refc.shape_changes *
 			xfs_allocfree_log_count(cur->bc_mp, 1);
 	overhead *= cur->bc_mp->m_sb.sb_blocksize;
 
@@ -891,17 +891,17 @@ xfs_refcount_still_have_space(
 	 * Only allow 2 refcount extent updates per transaction if the
 	 * refcount continue update "error" has been injected.
 	 */
-	if (cur->bc_private.a.priv.refc.nr_ops > 2 &&
+	if (cur->bc_ag.refc.nr_ops > 2 &&
 	    XFS_TEST_ERROR(false, cur->bc_mp,
 			XFS_ERRTAG_REFCOUNT_CONTINUE_UPDATE))
 		return false;
 
-	if (cur->bc_private.a.priv.refc.nr_ops == 0)
+	if (cur->bc_ag.refc.nr_ops == 0)
 		return true;
 	else if (overhead > cur->bc_tp->t_log_res)
 		return false;
 	return  cur->bc_tp->t_log_res - overhead >
-		cur->bc_private.a.priv.refc.nr_ops * XFS_REFCOUNT_ITEM_OVERHEAD;
+		cur->bc_ag.refc.nr_ops * XFS_REFCOUNT_ITEM_OVERHEAD;
 }
 
 /*
@@ -952,7 +952,7 @@ xfs_refcount_adjust_extents(
 					ext.rc_startblock - *agbno);
 			tmp.rc_refcount = 1 + adj;
 			trace_xfs_refcount_modify_extent(cur->bc_mp,
-					cur->bc_private.a.agno, &tmp);
+					cur->bc_ag.agno, &tmp);
 
 			/*
 			 * Either cover the hole (increment) or
@@ -968,10 +968,10 @@ xfs_refcount_adjust_extents(
 					error = -EFSCORRUPTED;
 					goto out_error;
 				}
-				cur->bc_private.a.priv.refc.nr_ops++;
+				cur->bc_ag.refc.nr_ops++;
 			} else {
 				fsbno = XFS_AGB_TO_FSB(cur->bc_mp,
-						cur->bc_private.a.agno,
+						cur->bc_ag.agno,
 						tmp.rc_startblock);
 				xfs_bmap_add_free(cur->bc_tp, fsbno,
 						  tmp.rc_blockcount, oinfo);
@@ -998,12 +998,12 @@ xfs_refcount_adjust_extents(
 			goto skip;
 		ext.rc_refcount += adj;
 		trace_xfs_refcount_modify_extent(cur->bc_mp,
-				cur->bc_private.a.agno, &ext);
+				cur->bc_ag.agno, &ext);
 		if (ext.rc_refcount > 1) {
 			error = xfs_refcount_update(cur, &ext);
 			if (error)
 				goto out_error;
-			cur->bc_private.a.priv.refc.nr_ops++;
+			cur->bc_ag.refc.nr_ops++;
 		} else if (ext.rc_refcount == 1) {
 			error = xfs_refcount_delete(cur, &found_rec);
 			if (error)
@@ -1012,11 +1012,11 @@ xfs_refcount_adjust_extents(
 				error = -EFSCORRUPTED;
 				goto out_error;
 			}
-			cur->bc_private.a.priv.refc.nr_ops++;
+			cur->bc_ag.refc.nr_ops++;
 			goto advloop;
 		} else {
 			fsbno = XFS_AGB_TO_FSB(cur->bc_mp,
-					cur->bc_private.a.agno,
+					cur->bc_ag.agno,
 					ext.rc_startblock);
 			xfs_bmap_add_free(cur->bc_tp, fsbno, ext.rc_blockcount,
 					  oinfo);
@@ -1035,7 +1035,7 @@ advloop:
 	return error;
 out_error:
 	trace_xfs_refcount_modify_extent_error(cur->bc_mp,
-			cur->bc_private.a.agno, error, _RET_IP_);
+			cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -1057,10 +1057,10 @@ xfs_refcount_adjust(
 	*new_agbno = agbno;
 	*new_aglen = aglen;
 	if (adj == XFS_REFCOUNT_ADJUST_INCREASE)
-		trace_xfs_refcount_increase(cur->bc_mp, cur->bc_private.a.agno,
+		trace_xfs_refcount_increase(cur->bc_mp, cur->bc_ag.agno,
 				agbno, aglen);
 	else
-		trace_xfs_refcount_decrease(cur->bc_mp, cur->bc_private.a.agno,
+		trace_xfs_refcount_decrease(cur->bc_mp, cur->bc_ag.agno,
 				agbno, aglen);
 
 	/*
@@ -1088,7 +1088,7 @@ xfs_refcount_adjust(
 	if (shape_changed)
 		shape_changes++;
 	if (shape_changes)
-		cur->bc_private.a.priv.refc.shape_changes++;
+		cur->bc_ag.refc.shape_changes++;
 
 	/* Now that we've taken care of the ends, adjust the middle extents */
 	error = xfs_refcount_adjust_extents(cur, new_agbno, new_aglen,
@@ -1099,7 +1099,7 @@ xfs_refcount_adjust(
 	return 0;
 
 out_error:
-	trace_xfs_refcount_adjust_error(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_refcount_adjust_error(cur->bc_mp, cur->bc_ag.agno,
 			error, _RET_IP_);
 	return error;
 }
@@ -1115,7 +1115,7 @@ xfs_refcount_finish_one_cleanup(
 
 	if (rcur == NULL)
 		return;
-	agbp = rcur->bc_private.a.agbp;
+	agbp = rcur->bc_ag.agbp;
 	xfs_btree_del_cursor(rcur, error);
 	if (error)
 		xfs_trans_brelse(tp, agbp);
@@ -1165,9 +1165,9 @@ xfs_refcount_finish_one(
 	 * the startblock, get one now.
 	 */
 	rcur = *pcur;
-	if (rcur != NULL && rcur->bc_private.a.agno != agno) {
-		nr_ops = rcur->bc_private.a.priv.refc.nr_ops;
-		shape_changes = rcur->bc_private.a.priv.refc.shape_changes;
+	if (rcur != NULL && rcur->bc_ag.agno != agno) {
+		nr_ops = rcur->bc_ag.refc.nr_ops;
+		shape_changes = rcur->bc_ag.refc.shape_changes;
 		xfs_refcount_finish_one_cleanup(tp, rcur, 0);
 		rcur = NULL;
 		*pcur = NULL;
@@ -1183,8 +1183,8 @@ xfs_refcount_finish_one(
 			error = -ENOMEM;
 			goto out_cur;
 		}
-		rcur->bc_private.a.priv.refc.nr_ops = nr_ops;
-		rcur->bc_private.a.priv.refc.shape_changes = shape_changes;
+		rcur->bc_ag.refc.nr_ops = nr_ops;
+		rcur->bc_ag.refc.shape_changes = shape_changes;
 	}
 	*pcur = rcur;
 
@@ -1303,7 +1303,7 @@ xfs_refcount_find_shared(
 	int				have;
 	int				error;
 
-	trace_xfs_refcount_find_shared(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_refcount_find_shared(cur->bc_mp, cur->bc_ag.agno,
 			agbno, aglen);
 
 	/* By default, skip the whole range */
@@ -1383,12 +1383,12 @@ xfs_refcount_find_shared(
 
 done:
 	trace_xfs_refcount_find_shared_result(cur->bc_mp,
-			cur->bc_private.a.agno, *fbno, *flen);
+			cur->bc_ag.agno, *fbno, *flen);
 
 out_error:
 	if (error)
 		trace_xfs_refcount_find_shared_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -1485,7 +1485,7 @@ xfs_refcount_adjust_cow_extents(
 		tmp.rc_blockcount = aglen;
 		tmp.rc_refcount = 1;
 		trace_xfs_refcount_modify_extent(cur->bc_mp,
-				cur->bc_private.a.agno, &tmp);
+				cur->bc_ag.agno, &tmp);
 
 		error = xfs_refcount_insert(cur, &tmp,
 				&found_tmp);
@@ -1513,7 +1513,7 @@ xfs_refcount_adjust_cow_extents(
 
 		ext.rc_refcount = 0;
 		trace_xfs_refcount_modify_extent(cur->bc_mp,
-				cur->bc_private.a.agno, &ext);
+				cur->bc_ag.agno, &ext);
 		error = xfs_refcount_delete(cur, &found_rec);
 		if (error)
 			goto out_error;
@@ -1529,7 +1529,7 @@ xfs_refcount_adjust_cow_extents(
 	return error;
 out_error:
 	trace_xfs_refcount_modify_extent_error(cur->bc_mp,
-			cur->bc_private.a.agno, error, _RET_IP_);
+			cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -1575,7 +1575,7 @@ xfs_refcount_adjust_cow(
 	return 0;
 
 out_error:
-	trace_xfs_refcount_adjust_cow_error(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_refcount_adjust_cow_error(cur->bc_mp, cur->bc_ag.agno,
 			error, _RET_IP_);
 	return error;
 }
@@ -1589,7 +1589,7 @@ __xfs_refcount_cow_alloc(
 	xfs_agblock_t		agbno,
 	xfs_extlen_t		aglen)
 {
-	trace_xfs_refcount_cow_increase(rcur->bc_mp, rcur->bc_private.a.agno,
+	trace_xfs_refcount_cow_increase(rcur->bc_mp, rcur->bc_ag.agno,
 			agbno, aglen);
 
 	/* Add refcount btree reservation */
@@ -1606,7 +1606,7 @@ __xfs_refcount_cow_free(
 	xfs_agblock_t		agbno,
 	xfs_extlen_t		aglen)
 {
-	trace_xfs_refcount_cow_decrease(rcur->bc_mp, rcur->bc_private.a.agno,
+	trace_xfs_refcount_cow_decrease(rcur->bc_mp, rcur->bc_ag.agno,
 			agbno, aglen);
 
 	/* Remove refcount btree reservation */
diff --git a/fs/xfs/libxfs/xfs_refcount_btree.c b/fs/xfs/libxfs/xfs_refcount_btree.c
index 38529dbacd55..a6ac60ae9421 100644
--- a/fs/xfs/libxfs/xfs_refcount_btree.c
+++ b/fs/xfs/libxfs/xfs_refcount_btree.c
@@ -12,6 +12,7 @@
 #include "xfs_sb.h"
 #include "xfs_mount.h"
 #include "xfs_btree.h"
+#include "xfs_btree_staging.h"
 #include "xfs_refcount_btree.h"
 #include "xfs_alloc.h"
 #include "xfs_error.h"
@@ -25,7 +26,7 @@ xfs_refcountbt_dup_cursor(
 	struct xfs_btree_cur	*cur)
 {
 	return xfs_refcountbt_init_cursor(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agbp, cur->bc_private.a.agno);
+			cur->bc_ag.agbp, cur->bc_ag.agno);
 }
 
 STATIC void
@@ -34,17 +35,15 @@ xfs_refcountbt_set_root(
 	union xfs_btree_ptr	*ptr,
 	int			inc)
 {
-	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
-	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
-	struct xfs_perag	*pag = xfs_perag_get(cur->bc_mp, seqno);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agf		*agf = agbp->b_addr;
+	struct xfs_perag	*pag = agbp->b_pag;
 
 	ASSERT(ptr->s != 0);
 
 	agf->agf_refcount_root = ptr->s;
 	be32_add_cpu(&agf->agf_refcount_level, inc);
 	pag->pagf_refcount_level += inc;
-	xfs_perag_put(pag);
 
 	xfs_alloc_log_agf(cur->bc_tp, agbp,
 			XFS_AGF_REFCOUNT_ROOT | XFS_AGF_REFCOUNT_LEVEL);
@@ -57,8 +56,8 @@ xfs_refcountbt_alloc_block(
 	union xfs_btree_ptr	*new,
 	int			*stat)
 {
-	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agf		*agf = agbp->b_addr;
 	struct xfs_alloc_arg	args;		/* block allocation args */
 	int			error;		/* error return value */
 
@@ -66,7 +65,7 @@ xfs_refcountbt_alloc_block(
 	args.tp = cur->bc_tp;
 	args.mp = cur->bc_mp;
 	args.type = XFS_ALLOCTYPE_NEAR_BNO;
-	args.fsbno = XFS_AGB_TO_FSB(cur->bc_mp, cur->bc_private.a.agno,
+	args.fsbno = XFS_AGB_TO_FSB(cur->bc_mp, cur->bc_ag.agno,
 			xfs_refc_block(args.mp));
 	args.oinfo = XFS_RMAP_OINFO_REFC;
 	args.minlen = args.maxlen = args.prod = 1;
@@ -75,13 +74,13 @@ xfs_refcountbt_alloc_block(
 	error = xfs_alloc_vextent(&args);
 	if (error)
 		goto out_error;
-	trace_xfs_refcountbt_alloc_block(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_refcountbt_alloc_block(cur->bc_mp, cur->bc_ag.agno,
 			args.agbno, 1);
 	if (args.fsbno == NULLFSBLOCK) {
 		*stat = 0;
 		return 0;
 	}
-	ASSERT(args.agno == cur->bc_private.a.agno);
+	ASSERT(args.agno == cur->bc_ag.agno);
 	ASSERT(args.len == 1);
 
 	new->s = cpu_to_be32(args.agbno);
@@ -101,12 +100,12 @@ xfs_refcountbt_free_block(
 	struct xfs_buf		*bp)
 {
 	struct xfs_mount	*mp = cur->bc_mp;
-	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agf		*agf = agbp->b_addr;
 	xfs_fsblock_t		fsbno = XFS_DADDR_TO_FSB(mp, XFS_BUF_ADDR(bp));
 	int			error;
 
-	trace_xfs_refcountbt_free_block(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_refcountbt_free_block(cur->bc_mp, cur->bc_ag.agno,
 			XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno), 1);
 	be32_add_cpu(&agf->agf_refcount_blocks, -1);
 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_REFCOUNT_BLOCKS);
@@ -169,9 +168,9 @@ xfs_refcountbt_init_ptr_from_cur(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*ptr)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
+	struct xfs_agf		*agf = cur->bc_ag.agbp->b_addr;
 
-	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno));
+	ASSERT(cur->bc_ag.agno == be32_to_cpu(agf->agf_seqno));
 
 	ptr->s = agf->agf_refcount_root;
 }
@@ -311,42 +310,91 @@ static const struct xfs_btree_ops xfs_refcountbt_ops = {
 };
 
 /*
- * Allocate a new refcount btree cursor.
+ * Initialize a new refcount btree cursor.
  */
-struct xfs_btree_cur *
-xfs_refcountbt_init_cursor(
+static struct xfs_btree_cur *
+xfs_refcountbt_init_common(
 	struct xfs_mount	*mp,
 	struct xfs_trans	*tp,
-	struct xfs_buf		*agbp,
 	xfs_agnumber_t		agno)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
 	struct xfs_btree_cur	*cur;
 
 	ASSERT(agno != NULLAGNUMBER);
 	ASSERT(agno < mp->m_sb.sb_agcount);
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
 
+	cur = kmem_cache_zalloc(xfs_btree_cur_zone, GFP_NOFS | __GFP_NOFAIL);
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
 	cur->bc_btnum = XFS_BTNUM_REFC;
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
-	cur->bc_ops = &xfs_refcountbt_ops;
 	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_refcbt_2);
 
-	cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level);
-
-	cur->bc_private.a.agbp = agbp;
-	cur->bc_private.a.agno = agno;
+	cur->bc_ag.agno = agno;
 	cur->bc_flags |= XFS_BTREE_CRC_BLOCKS;
 
-	cur->bc_private.a.priv.refc.nr_ops = 0;
-	cur->bc_private.a.priv.refc.shape_changes = 0;
+	cur->bc_ag.refc.nr_ops = 0;
+	cur->bc_ag.refc.shape_changes = 0;
+	cur->bc_ops = &xfs_refcountbt_ops;
+	return cur;
+}
+
+/* Create a btree cursor. */
+struct xfs_btree_cur *
+xfs_refcountbt_init_cursor(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
+	xfs_agnumber_t		agno)
+{
+	struct xfs_agf		*agf = agbp->b_addr;
+	struct xfs_btree_cur	*cur;
 
+	cur = xfs_refcountbt_init_common(mp, tp, agno);
+	cur->bc_nlevels = be32_to_cpu(agf->agf_refcount_level);
+	cur->bc_ag.agbp = agbp;
+	return cur;
+}
+
+/* Create a btree cursor with a fake root for staging. */
+struct xfs_btree_cur *
+xfs_refcountbt_stage_cursor(
+	struct xfs_mount	*mp,
+	struct xbtree_afakeroot	*afake,
+	xfs_agnumber_t		agno)
+{
+	struct xfs_btree_cur	*cur;
+
+	cur = xfs_refcountbt_init_common(mp, NULL, agno);
+	xfs_btree_stage_afakeroot(cur, afake);
 	return cur;
 }
 
 /*
+ * Swap in the new btree root.  Once we pass this point the newly rebuilt btree
+ * is in place and we have to kill off all the old btree blocks.
+ */
+void
+xfs_refcountbt_commit_staged_btree(
+	struct xfs_btree_cur	*cur,
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp)
+{
+	struct xfs_agf		*agf = agbp->b_addr;
+	struct xbtree_afakeroot	*afake = cur->bc_ag.afake;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+
+	agf->agf_refcount_root = cpu_to_be32(afake->af_root);
+	agf->agf_refcount_level = cpu_to_be32(afake->af_levels);
+	agf->agf_refcount_blocks = cpu_to_be32(afake->af_blocks);
+	xfs_alloc_log_agf(tp, agbp, XFS_AGF_REFCOUNT_BLOCKS |
+				    XFS_AGF_REFCOUNT_ROOT |
+				    XFS_AGF_REFCOUNT_LEVEL);
+	xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_refcountbt_ops);
+}
+
+/*
  * Calculate the number of records in a refcount btree block.
  */
 int
@@ -420,7 +468,7 @@ xfs_refcountbt_calc_reserves(
 	if (error)
 		return error;
 
-	agf = XFS_BUF_TO_AGF(agbp);
+	agf = agbp->b_addr;
 	agblocks = be32_to_cpu(agf->agf_length);
 	tree_len = be32_to_cpu(agf->agf_refcount_blocks);
 	xfs_trans_brelse(tp, agbp);
diff --git a/fs/xfs/libxfs/xfs_refcount_btree.h b/fs/xfs/libxfs/xfs_refcount_btree.h
index ba416f71c824..69dc515db671 100644
--- a/fs/xfs/libxfs/xfs_refcount_btree.h
+++ b/fs/xfs/libxfs/xfs_refcount_btree.h
@@ -13,6 +13,7 @@
 struct xfs_buf;
 struct xfs_btree_cur;
 struct xfs_mount;
+struct xbtree_afakeroot;
 
 /*
  * Btree block header size
@@ -46,6 +47,8 @@ struct xfs_mount;
 extern struct xfs_btree_cur *xfs_refcountbt_init_cursor(struct xfs_mount *mp,
 		struct xfs_trans *tp, struct xfs_buf *agbp,
 		xfs_agnumber_t agno);
+struct xfs_btree_cur *xfs_refcountbt_stage_cursor(struct xfs_mount *mp,
+		struct xbtree_afakeroot *afake, xfs_agnumber_t agno);
 extern int xfs_refcountbt_maxrecs(int blocklen, bool leaf);
 extern void xfs_refcountbt_compute_maxlevels(struct xfs_mount *mp);
 
@@ -58,4 +61,7 @@ extern int xfs_refcountbt_calc_reserves(struct xfs_mount *mp,
 		struct xfs_trans *tp, xfs_agnumber_t agno, xfs_extlen_t *ask,
 		xfs_extlen_t *used);
 
+void xfs_refcountbt_commit_staged_btree(struct xfs_btree_cur *cur,
+		struct xfs_trans *tp, struct xfs_buf *agbp);
+
 #endif	/* __XFS_REFCOUNT_BTREE_H__ */
diff --git a/fs/xfs/libxfs/xfs_rmap.c b/fs/xfs/libxfs/xfs_rmap.c
index ff9412f113c4..27c39268c31f 100644
--- a/fs/xfs/libxfs/xfs_rmap.c
+++ b/fs/xfs/libxfs/xfs_rmap.c
@@ -79,7 +79,7 @@ xfs_rmap_update(
 	union xfs_btree_rec	rec;
 	int			error;
 
-	trace_xfs_rmap_update(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_rmap_update(cur->bc_mp, cur->bc_ag.agno,
 			irec->rm_startblock, irec->rm_blockcount,
 			irec->rm_owner, irec->rm_offset, irec->rm_flags);
 
@@ -91,7 +91,7 @@ xfs_rmap_update(
 	error = xfs_btree_update(cur, &rec);
 	if (error)
 		trace_xfs_rmap_update_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -107,7 +107,7 @@ xfs_rmap_insert(
 	int			i;
 	int			error;
 
-	trace_xfs_rmap_insert(rcur->bc_mp, rcur->bc_private.a.agno, agbno,
+	trace_xfs_rmap_insert(rcur->bc_mp, rcur->bc_ag.agno, agbno,
 			len, owner, offset, flags);
 
 	error = xfs_rmap_lookup_eq(rcur, agbno, len, owner, offset, flags, &i);
@@ -133,7 +133,7 @@ xfs_rmap_insert(
 done:
 	if (error)
 		trace_xfs_rmap_insert_error(rcur->bc_mp,
-				rcur->bc_private.a.agno, error, _RET_IP_);
+				rcur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -149,7 +149,7 @@ xfs_rmap_delete(
 	int			i;
 	int			error;
 
-	trace_xfs_rmap_delete(rcur->bc_mp, rcur->bc_private.a.agno, agbno,
+	trace_xfs_rmap_delete(rcur->bc_mp, rcur->bc_ag.agno, agbno,
 			len, owner, offset, flags);
 
 	error = xfs_rmap_lookup_eq(rcur, agbno, len, owner, offset, flags, &i);
@@ -170,7 +170,7 @@ xfs_rmap_delete(
 done:
 	if (error)
 		trace_xfs_rmap_delete_error(rcur->bc_mp,
-				rcur->bc_private.a.agno, error, _RET_IP_);
+				rcur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -197,7 +197,7 @@ xfs_rmap_get_rec(
 	int			*stat)
 {
 	struct xfs_mount	*mp = cur->bc_mp;
-	xfs_agnumber_t		agno = cur->bc_private.a.agno;
+	xfs_agnumber_t		agno = cur->bc_ag.agno;
 	union xfs_btree_rec	*rec;
 	int			error;
 
@@ -260,7 +260,7 @@ xfs_rmap_find_left_neighbor_helper(
 	struct xfs_find_left_neighbor_info	*info = priv;
 
 	trace_xfs_rmap_find_left_neighbor_candidate(cur->bc_mp,
-			cur->bc_private.a.agno, rec->rm_startblock,
+			cur->bc_ag.agno, rec->rm_startblock,
 			rec->rm_blockcount, rec->rm_owner, rec->rm_offset,
 			rec->rm_flags);
 
@@ -312,7 +312,7 @@ xfs_rmap_find_left_neighbor(
 	info.stat = stat;
 
 	trace_xfs_rmap_find_left_neighbor_query(cur->bc_mp,
-			cur->bc_private.a.agno, bno, 0, owner, offset, flags);
+			cur->bc_ag.agno, bno, 0, owner, offset, flags);
 
 	error = xfs_rmap_query_range(cur, &info.high, &info.high,
 			xfs_rmap_find_left_neighbor_helper, &info);
@@ -320,7 +320,7 @@ xfs_rmap_find_left_neighbor(
 		error = 0;
 	if (*stat)
 		trace_xfs_rmap_find_left_neighbor_result(cur->bc_mp,
-				cur->bc_private.a.agno, irec->rm_startblock,
+				cur->bc_ag.agno, irec->rm_startblock,
 				irec->rm_blockcount, irec->rm_owner,
 				irec->rm_offset, irec->rm_flags);
 	return error;
@@ -336,7 +336,7 @@ xfs_rmap_lookup_le_range_helper(
 	struct xfs_find_left_neighbor_info	*info = priv;
 
 	trace_xfs_rmap_lookup_le_range_candidate(cur->bc_mp,
-			cur->bc_private.a.agno, rec->rm_startblock,
+			cur->bc_ag.agno, rec->rm_startblock,
 			rec->rm_blockcount, rec->rm_owner, rec->rm_offset,
 			rec->rm_flags);
 
@@ -385,14 +385,14 @@ xfs_rmap_lookup_le_range(
 	info.stat = stat;
 
 	trace_xfs_rmap_lookup_le_range(cur->bc_mp,
-			cur->bc_private.a.agno, bno, 0, owner, offset, flags);
+			cur->bc_ag.agno, bno, 0, owner, offset, flags);
 	error = xfs_rmap_query_range(cur, &info.high, &info.high,
 			xfs_rmap_lookup_le_range_helper, &info);
 	if (error == -ECANCELED)
 		error = 0;
 	if (*stat)
 		trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
-				cur->bc_private.a.agno, irec->rm_startblock,
+				cur->bc_ag.agno, irec->rm_startblock,
 				irec->rm_blockcount, irec->rm_owner,
 				irec->rm_offset, irec->rm_flags);
 	return error;
@@ -498,7 +498,7 @@ xfs_rmap_unmap(
 			(flags & XFS_RMAP_BMBT_BLOCK);
 	if (unwritten)
 		flags |= XFS_RMAP_UNWRITTEN;
-	trace_xfs_rmap_unmap(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_unmap(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 
 	/*
@@ -522,7 +522,7 @@ xfs_rmap_unmap(
 		goto out_error;
 	}
 	trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
-			cur->bc_private.a.agno, ltrec.rm_startblock,
+			cur->bc_ag.agno, ltrec.rm_startblock,
 			ltrec.rm_blockcount, ltrec.rm_owner,
 			ltrec.rm_offset, ltrec.rm_flags);
 	ltoff = ltrec.rm_offset;
@@ -588,7 +588,7 @@ xfs_rmap_unmap(
 
 	if (ltrec.rm_startblock == bno && ltrec.rm_blockcount == len) {
 		/* exact match, simply remove the record from rmap tree */
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, cur->bc_ag.agno,
 				ltrec.rm_startblock, ltrec.rm_blockcount,
 				ltrec.rm_owner, ltrec.rm_offset,
 				ltrec.rm_flags);
@@ -666,7 +666,7 @@ xfs_rmap_unmap(
 		else
 			cur->bc_rec.r.rm_offset = offset + len;
 		cur->bc_rec.r.rm_flags = flags;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_insert(mp, cur->bc_ag.agno,
 				cur->bc_rec.r.rm_startblock,
 				cur->bc_rec.r.rm_blockcount,
 				cur->bc_rec.r.rm_owner,
@@ -678,11 +678,11 @@ xfs_rmap_unmap(
 	}
 
 out_done:
-	trace_xfs_rmap_unmap_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_unmap_done(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 out_error:
 	if (error)
-		trace_xfs_rmap_unmap_error(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_unmap_error(mp, cur->bc_ag.agno,
 				error, _RET_IP_);
 	return error;
 }
@@ -773,7 +773,7 @@ xfs_rmap_map(
 			(flags & XFS_RMAP_BMBT_BLOCK);
 	if (unwritten)
 		flags |= XFS_RMAP_UNWRITTEN;
-	trace_xfs_rmap_map(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_map(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 	ASSERT(!xfs_rmap_should_skip_owner_update(oinfo));
 
@@ -795,7 +795,7 @@ xfs_rmap_map(
 			goto out_error;
 		}
 		trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
-				cur->bc_private.a.agno, ltrec.rm_startblock,
+				cur->bc_ag.agno, ltrec.rm_startblock,
 				ltrec.rm_blockcount, ltrec.rm_owner,
 				ltrec.rm_offset, ltrec.rm_flags);
 
@@ -831,7 +831,7 @@ xfs_rmap_map(
 			goto out_error;
 		}
 		trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
-			cur->bc_private.a.agno, gtrec.rm_startblock,
+			cur->bc_ag.agno, gtrec.rm_startblock,
 			gtrec.rm_blockcount, gtrec.rm_owner,
 			gtrec.rm_offset, gtrec.rm_flags);
 		if (!xfs_rmap_is_mergeable(&gtrec, owner, flags))
@@ -870,7 +870,7 @@ xfs_rmap_map(
 			 * result: |rrrrrrrrrrrrrrrrrrrrrrrrrrrrr|
 			 */
 			ltrec.rm_blockcount += gtrec.rm_blockcount;
-			trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+			trace_xfs_rmap_delete(mp, cur->bc_ag.agno,
 					gtrec.rm_startblock,
 					gtrec.rm_blockcount,
 					gtrec.rm_owner,
@@ -921,7 +921,7 @@ xfs_rmap_map(
 		cur->bc_rec.r.rm_owner = owner;
 		cur->bc_rec.r.rm_offset = offset;
 		cur->bc_rec.r.rm_flags = flags;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, len,
+		trace_xfs_rmap_insert(mp, cur->bc_ag.agno, bno, len,
 			owner, offset, flags);
 		error = xfs_btree_insert(cur, &i);
 		if (error)
@@ -932,11 +932,11 @@ xfs_rmap_map(
 		}
 	}
 
-	trace_xfs_rmap_map_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_map_done(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 out_error:
 	if (error)
-		trace_xfs_rmap_map_error(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_map_error(mp, cur->bc_ag.agno,
 				error, _RET_IP_);
 	return error;
 }
@@ -1010,7 +1010,7 @@ xfs_rmap_convert(
 			(flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK))));
 	oldext = unwritten ? XFS_RMAP_UNWRITTEN : 0;
 	new_endoff = offset + len;
-	trace_xfs_rmap_convert(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_convert(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 
 	/*
@@ -1034,7 +1034,7 @@ xfs_rmap_convert(
 		goto done;
 	}
 	trace_xfs_rmap_lookup_le_range_result(cur->bc_mp,
-			cur->bc_private.a.agno, PREV.rm_startblock,
+			cur->bc_ag.agno, PREV.rm_startblock,
 			PREV.rm_blockcount, PREV.rm_owner,
 			PREV.rm_offset, PREV.rm_flags);
 
@@ -1076,7 +1076,7 @@ xfs_rmap_convert(
 			goto done;
 		}
 		trace_xfs_rmap_find_left_neighbor_result(cur->bc_mp,
-				cur->bc_private.a.agno, LEFT.rm_startblock,
+				cur->bc_ag.agno, LEFT.rm_startblock,
 				LEFT.rm_blockcount, LEFT.rm_owner,
 				LEFT.rm_offset, LEFT.rm_flags);
 		if (LEFT.rm_startblock + LEFT.rm_blockcount == bno &&
@@ -1114,7 +1114,7 @@ xfs_rmap_convert(
 			goto done;
 		}
 		trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
-				cur->bc_private.a.agno, RIGHT.rm_startblock,
+				cur->bc_ag.agno, RIGHT.rm_startblock,
 				RIGHT.rm_blockcount, RIGHT.rm_owner,
 				RIGHT.rm_offset, RIGHT.rm_flags);
 		if (bno + len == RIGHT.rm_startblock &&
@@ -1132,7 +1132,7 @@ xfs_rmap_convert(
 	     RIGHT.rm_blockcount > XFS_RMAP_LEN_MAX)
 		state &= ~RMAP_RIGHT_CONTIG;
 
-	trace_xfs_rmap_convert_state(mp, cur->bc_private.a.agno, state,
+	trace_xfs_rmap_convert_state(mp, cur->bc_ag.agno, state,
 			_RET_IP_);
 
 	/* reset the cursor back to PREV */
@@ -1162,7 +1162,7 @@ xfs_rmap_convert(
 			error = -EFSCORRUPTED;
 			goto done;
 		}
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, cur->bc_ag.agno,
 				RIGHT.rm_startblock, RIGHT.rm_blockcount,
 				RIGHT.rm_owner, RIGHT.rm_offset,
 				RIGHT.rm_flags);
@@ -1180,7 +1180,7 @@ xfs_rmap_convert(
 			error = -EFSCORRUPTED;
 			goto done;
 		}
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, cur->bc_ag.agno,
 				PREV.rm_startblock, PREV.rm_blockcount,
 				PREV.rm_owner, PREV.rm_offset,
 				PREV.rm_flags);
@@ -1210,7 +1210,7 @@ xfs_rmap_convert(
 		 * Setting all of a previous oldext extent to newext.
 		 * The left neighbor is contiguous, the right is not.
 		 */
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, cur->bc_ag.agno,
 				PREV.rm_startblock, PREV.rm_blockcount,
 				PREV.rm_owner, PREV.rm_offset,
 				PREV.rm_flags);
@@ -1247,7 +1247,7 @@ xfs_rmap_convert(
 			error = -EFSCORRUPTED;
 			goto done;
 		}
-		trace_xfs_rmap_delete(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_delete(mp, cur->bc_ag.agno,
 				RIGHT.rm_startblock, RIGHT.rm_blockcount,
 				RIGHT.rm_owner, RIGHT.rm_offset,
 				RIGHT.rm_flags);
@@ -1326,7 +1326,7 @@ xfs_rmap_convert(
 		NEW.rm_blockcount = len;
 		NEW.rm_flags = newext;
 		cur->bc_rec.r = NEW;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno,
+		trace_xfs_rmap_insert(mp, cur->bc_ag.agno, bno,
 				len, owner, offset, newext);
 		error = xfs_btree_insert(cur, &i);
 		if (error)
@@ -1383,7 +1383,7 @@ xfs_rmap_convert(
 		NEW.rm_blockcount = len;
 		NEW.rm_flags = newext;
 		cur->bc_rec.r = NEW;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno,
+		trace_xfs_rmap_insert(mp, cur->bc_ag.agno, bno,
 				len, owner, offset, newext);
 		error = xfs_btree_insert(cur, &i);
 		if (error)
@@ -1414,7 +1414,7 @@ xfs_rmap_convert(
 		NEW = PREV;
 		NEW.rm_blockcount = offset - PREV.rm_offset;
 		cur->bc_rec.r = NEW;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno,
+		trace_xfs_rmap_insert(mp, cur->bc_ag.agno,
 				NEW.rm_startblock, NEW.rm_blockcount,
 				NEW.rm_owner, NEW.rm_offset,
 				NEW.rm_flags);
@@ -1441,7 +1441,7 @@ xfs_rmap_convert(
 		/* new middle extent - newext */
 		cur->bc_rec.r.rm_flags &= ~XFS_RMAP_UNWRITTEN;
 		cur->bc_rec.r.rm_flags |= newext;
-		trace_xfs_rmap_insert(mp, cur->bc_private.a.agno, bno, len,
+		trace_xfs_rmap_insert(mp, cur->bc_ag.agno, bno, len,
 				owner, offset, newext);
 		error = xfs_btree_insert(cur, &i);
 		if (error)
@@ -1465,12 +1465,12 @@ xfs_rmap_convert(
 		ASSERT(0);
 	}
 
-	trace_xfs_rmap_convert_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_convert_done(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 done:
 	if (error)
 		trace_xfs_rmap_convert_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -1506,7 +1506,7 @@ xfs_rmap_convert_shared(
 			(flags & (XFS_RMAP_ATTR_FORK | XFS_RMAP_BMBT_BLOCK))));
 	oldext = unwritten ? XFS_RMAP_UNWRITTEN : 0;
 	new_endoff = offset + len;
-	trace_xfs_rmap_convert(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_convert(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 
 	/*
@@ -1573,7 +1573,7 @@ xfs_rmap_convert_shared(
 			goto done;
 		}
 		trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
-				cur->bc_private.a.agno, RIGHT.rm_startblock,
+				cur->bc_ag.agno, RIGHT.rm_startblock,
 				RIGHT.rm_blockcount, RIGHT.rm_owner,
 				RIGHT.rm_offset, RIGHT.rm_flags);
 		if (xfs_rmap_is_mergeable(&RIGHT, owner, newext))
@@ -1589,7 +1589,7 @@ xfs_rmap_convert_shared(
 	     RIGHT.rm_blockcount > XFS_RMAP_LEN_MAX)
 		state &= ~RMAP_RIGHT_CONTIG;
 
-	trace_xfs_rmap_convert_state(mp, cur->bc_private.a.agno, state,
+	trace_xfs_rmap_convert_state(mp, cur->bc_ag.agno, state,
 			_RET_IP_);
 	/*
 	 * Switch out based on the FILLING and CONTIG state bits.
@@ -1880,12 +1880,12 @@ xfs_rmap_convert_shared(
 		ASSERT(0);
 	}
 
-	trace_xfs_rmap_convert_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_convert_done(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 done:
 	if (error)
 		trace_xfs_rmap_convert_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -1923,7 +1923,7 @@ xfs_rmap_unmap_shared(
 	xfs_owner_info_unpack(oinfo, &owner, &offset, &flags);
 	if (unwritten)
 		flags |= XFS_RMAP_UNWRITTEN;
-	trace_xfs_rmap_unmap(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_unmap(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 
 	/*
@@ -2072,12 +2072,12 @@ xfs_rmap_unmap_shared(
 			goto out_error;
 	}
 
-	trace_xfs_rmap_unmap_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_unmap_done(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 out_error:
 	if (error)
 		trace_xfs_rmap_unmap_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -2112,7 +2112,7 @@ xfs_rmap_map_shared(
 	xfs_owner_info_unpack(oinfo, &owner, &offset, &flags);
 	if (unwritten)
 		flags |= XFS_RMAP_UNWRITTEN;
-	trace_xfs_rmap_map(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_map(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 
 	/* Is there a left record that abuts our range? */
@@ -2138,7 +2138,7 @@ xfs_rmap_map_shared(
 			goto out_error;
 		}
 		trace_xfs_rmap_find_right_neighbor_result(cur->bc_mp,
-			cur->bc_private.a.agno, gtrec.rm_startblock,
+			cur->bc_ag.agno, gtrec.rm_startblock,
 			gtrec.rm_blockcount, gtrec.rm_owner,
 			gtrec.rm_offset, gtrec.rm_flags);
 
@@ -2231,12 +2231,12 @@ xfs_rmap_map_shared(
 			goto out_error;
 	}
 
-	trace_xfs_rmap_map_done(mp, cur->bc_private.a.agno, bno, len,
+	trace_xfs_rmap_map_done(mp, cur->bc_ag.agno, bno, len,
 			unwritten, oinfo);
 out_error:
 	if (error)
 		trace_xfs_rmap_map_error(cur->bc_mp,
-				cur->bc_private.a.agno, error, _RET_IP_);
+				cur->bc_ag.agno, error, _RET_IP_);
 	return error;
 }
 
@@ -2336,7 +2336,7 @@ xfs_rmap_finish_one_cleanup(
 
 	if (rcur == NULL)
 		return;
-	agbp = rcur->bc_private.a.agbp;
+	agbp = rcur->bc_ag.agbp;
 	xfs_btree_del_cursor(rcur, error);
 	if (error)
 		xfs_trans_brelse(tp, agbp);
@@ -2386,7 +2386,7 @@ xfs_rmap_finish_one(
 	 * the startblock, get one now.
 	 */
 	rcur = *pcur;
-	if (rcur != NULL && rcur->bc_private.a.agno != agno) {
+	if (rcur != NULL && rcur->bc_ag.agno != agno) {
 		xfs_rmap_finish_one_cleanup(tp, rcur, 0);
 		rcur = NULL;
 		*pcur = NULL;
@@ -2694,7 +2694,6 @@ struct xfs_rmap_key_state {
 	uint64_t			owner;
 	uint64_t			offset;
 	unsigned int			flags;
-	bool				has_rmap;
 };
 
 /* For each rmap given, figure out if it doesn't match the key we want. */
@@ -2709,7 +2708,6 @@ xfs_rmap_has_other_keys_helper(
 	if (rks->owner == rec->rm_owner && rks->offset == rec->rm_offset &&
 	    ((rks->flags & rec->rm_flags) & XFS_RMAP_KEY_FLAGS) == rks->flags)
 		return 0;
-	rks->has_rmap = true;
 	return -ECANCELED;
 }
 
@@ -2731,7 +2729,7 @@ xfs_rmap_has_other_keys(
 	int				error;
 
 	xfs_owner_info_unpack(oinfo, &rks.owner, &rks.offset, &rks.flags);
-	rks.has_rmap = false;
+	*has_rmap = false;
 
 	low.rm_startblock = bno;
 	memset(&high, 0xFF, sizeof(high));
@@ -2739,11 +2737,12 @@ xfs_rmap_has_other_keys(
 
 	error = xfs_rmap_query_range(cur, &low, &high,
 			xfs_rmap_has_other_keys_helper, &rks);
-	if (error < 0)
-		return error;
+	if (error == -ECANCELED) {
+		*has_rmap = true;
+		return 0;
+	}
 
-	*has_rmap = rks.has_rmap;
-	return 0;
+	return error;
 }
 
 const struct xfs_owner_info XFS_RMAP_OINFO_SKIP_UPDATE = {
diff --git a/fs/xfs/libxfs/xfs_rmap_btree.c b/fs/xfs/libxfs/xfs_rmap_btree.c
index fc78efa52c94..beb81c84a937 100644
--- a/fs/xfs/libxfs/xfs_rmap_btree.c
+++ b/fs/xfs/libxfs/xfs_rmap_btree.c
@@ -14,6 +14,7 @@
 #include "xfs_trans.h"
 #include "xfs_alloc.h"
 #include "xfs_btree.h"
+#include "xfs_btree_staging.h"
 #include "xfs_rmap.h"
 #include "xfs_rmap_btree.h"
 #include "xfs_trace.h"
@@ -51,7 +52,7 @@ xfs_rmapbt_dup_cursor(
 	struct xfs_btree_cur	*cur)
 {
 	return xfs_rmapbt_init_cursor(cur->bc_mp, cur->bc_tp,
-			cur->bc_private.a.agbp, cur->bc_private.a.agno);
+			cur->bc_ag.agbp, cur->bc_ag.agno);
 }
 
 STATIC void
@@ -60,18 +61,16 @@ xfs_rmapbt_set_root(
 	union xfs_btree_ptr	*ptr,
 	int			inc)
 {
-	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
-	xfs_agnumber_t		seqno = be32_to_cpu(agf->agf_seqno);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agf		*agf = agbp->b_addr;
 	int			btnum = cur->bc_btnum;
-	struct xfs_perag	*pag = xfs_perag_get(cur->bc_mp, seqno);
+	struct xfs_perag	*pag = agbp->b_pag;
 
 	ASSERT(ptr->s != 0);
 
 	agf->agf_roots[btnum] = ptr->s;
 	be32_add_cpu(&agf->agf_levels[btnum], inc);
 	pag->pagf_levels[btnum] += inc;
-	xfs_perag_put(pag);
 
 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS);
 }
@@ -83,25 +82,25 @@ xfs_rmapbt_alloc_block(
 	union xfs_btree_ptr	*new,
 	int			*stat)
 {
-	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agf		*agf = agbp->b_addr;
 	int			error;
 	xfs_agblock_t		bno;
 
 	/* Allocate the new block from the freelist. If we can't, give up.  */
-	error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_private.a.agbp,
+	error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_ag.agbp,
 				       &bno, 1);
 	if (error)
 		return error;
 
-	trace_xfs_rmapbt_alloc_block(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_rmapbt_alloc_block(cur->bc_mp, cur->bc_ag.agno,
 			bno, 1);
 	if (bno == NULLAGBLOCK) {
 		*stat = 0;
 		return 0;
 	}
 
-	xfs_extent_busy_reuse(cur->bc_mp, cur->bc_private.a.agno, bno, 1,
+	xfs_extent_busy_reuse(cur->bc_mp, cur->bc_ag.agno, bno, 1,
 			false);
 
 	xfs_trans_agbtree_delta(cur->bc_tp, 1);
@@ -109,7 +108,7 @@ xfs_rmapbt_alloc_block(
 	be32_add_cpu(&agf->agf_rmap_blocks, 1);
 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_RMAP_BLOCKS);
 
-	xfs_ag_resv_rmapbt_alloc(cur->bc_mp, cur->bc_private.a.agno);
+	xfs_ag_resv_rmapbt_alloc(cur->bc_mp, cur->bc_ag.agno);
 
 	*stat = 1;
 	return 0;
@@ -120,13 +119,14 @@ xfs_rmapbt_free_block(
 	struct xfs_btree_cur	*cur,
 	struct xfs_buf		*bp)
 {
-	struct xfs_buf		*agbp = cur->bc_private.a.agbp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+	struct xfs_buf		*agbp = cur->bc_ag.agbp;
+	struct xfs_agf		*agf = agbp->b_addr;
+	struct xfs_perag	*pag;
 	xfs_agblock_t		bno;
 	int			error;
 
 	bno = xfs_daddr_to_agbno(cur->bc_mp, XFS_BUF_ADDR(bp));
-	trace_xfs_rmapbt_free_block(cur->bc_mp, cur->bc_private.a.agno,
+	trace_xfs_rmapbt_free_block(cur->bc_mp, cur->bc_ag.agno,
 			bno, 1);
 	be32_add_cpu(&agf->agf_rmap_blocks, -1);
 	xfs_alloc_log_agf(cur->bc_tp, agbp, XFS_AGF_RMAP_BLOCKS);
@@ -138,8 +138,8 @@ xfs_rmapbt_free_block(
 			      XFS_EXTENT_BUSY_SKIP_DISCARD);
 	xfs_trans_agbtree_delta(cur->bc_tp, -1);
 
-	xfs_ag_resv_rmapbt_free(cur->bc_mp, cur->bc_private.a.agno);
-
+	pag = cur->bc_ag.agbp->b_pag;
+	xfs_ag_resv_free_extent(pag, XFS_AG_RESV_RMAPBT, NULL, 1);
 	return 0;
 }
 
@@ -215,9 +215,9 @@ xfs_rmapbt_init_ptr_from_cur(
 	struct xfs_btree_cur	*cur,
 	union xfs_btree_ptr	*ptr)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(cur->bc_private.a.agbp);
+	struct xfs_agf		*agf = cur->bc_ag.agbp->b_addr;
 
-	ASSERT(cur->bc_private.a.agno == be32_to_cpu(agf->agf_seqno));
+	ASSERT(cur->bc_ag.agno == be32_to_cpu(agf->agf_seqno));
 
 	ptr->s = agf->agf_roots[cur->bc_btnum];
 }
@@ -448,37 +448,83 @@ static const struct xfs_btree_ops xfs_rmapbt_ops = {
 	.recs_inorder		= xfs_rmapbt_recs_inorder,
 };
 
-/*
- * Allocate a new allocation btree cursor.
- */
-struct xfs_btree_cur *
-xfs_rmapbt_init_cursor(
+static struct xfs_btree_cur *
+xfs_rmapbt_init_common(
 	struct xfs_mount	*mp,
 	struct xfs_trans	*tp,
-	struct xfs_buf		*agbp,
 	xfs_agnumber_t		agno)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
 	struct xfs_btree_cur	*cur;
 
-	cur = kmem_zone_zalloc(xfs_btree_cur_zone, KM_NOFS);
+	cur = kmem_cache_zalloc(xfs_btree_cur_zone, GFP_NOFS | __GFP_NOFAIL);
 	cur->bc_tp = tp;
 	cur->bc_mp = mp;
 	/* Overlapping btree; 2 keys per pointer. */
 	cur->bc_btnum = XFS_BTNUM_RMAP;
 	cur->bc_flags = XFS_BTREE_CRC_BLOCKS | XFS_BTREE_OVERLAPPING;
 	cur->bc_blocklog = mp->m_sb.sb_blocklog;
+	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_rmap_2);
+	cur->bc_ag.agno = agno;
 	cur->bc_ops = &xfs_rmapbt_ops;
+
+	return cur;
+}
+
+/* Create a new reverse mapping btree cursor. */
+struct xfs_btree_cur *
+xfs_rmapbt_init_cursor(
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
+	xfs_agnumber_t		agno)
+{
+	struct xfs_agf		*agf = agbp->b_addr;
+	struct xfs_btree_cur	*cur;
+
+	cur = xfs_rmapbt_init_common(mp, tp, agno);
 	cur->bc_nlevels = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]);
-	cur->bc_statoff = XFS_STATS_CALC_INDEX(xs_rmap_2);
+	cur->bc_ag.agbp = agbp;
+	return cur;
+}
 
-	cur->bc_private.a.agbp = agbp;
-	cur->bc_private.a.agno = agno;
+/* Create a new reverse mapping btree cursor with a fake root for staging. */
+struct xfs_btree_cur *
+xfs_rmapbt_stage_cursor(
+	struct xfs_mount	*mp,
+	struct xbtree_afakeroot	*afake,
+	xfs_agnumber_t		agno)
+{
+	struct xfs_btree_cur	*cur;
 
+	cur = xfs_rmapbt_init_common(mp, NULL, agno);
+	xfs_btree_stage_afakeroot(cur, afake);
 	return cur;
 }
 
 /*
+ * Install a new reverse mapping btree root.  Caller is responsible for
+ * invalidating and freeing the old btree blocks.
+ */
+void
+xfs_rmapbt_commit_staged_btree(
+	struct xfs_btree_cur	*cur,
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp)
+{
+	struct xfs_agf		*agf = agbp->b_addr;
+	struct xbtree_afakeroot	*afake = cur->bc_ag.afake;
+
+	ASSERT(cur->bc_flags & XFS_BTREE_STAGING);
+
+	agf->agf_roots[cur->bc_btnum] = cpu_to_be32(afake->af_root);
+	agf->agf_levels[cur->bc_btnum] = cpu_to_be32(afake->af_levels);
+	agf->agf_rmap_blocks = cpu_to_be32(afake->af_blocks);
+	xfs_alloc_log_agf(tp, agbp, XFS_AGF_ROOTS | XFS_AGF_LEVELS |
+				    XFS_AGF_RMAP_BLOCKS);
+	xfs_btree_commit_afakeroot(cur, tp, agbp, &xfs_rmapbt_ops);
+}
+
+/*
  * Calculate number of records in an rmap btree block.
  */
 int
@@ -569,7 +615,7 @@ xfs_rmapbt_calc_reserves(
 	if (error)
 		return error;
 
-	agf = XFS_BUF_TO_AGF(agbp);
+	agf = agbp->b_addr;
 	agblocks = be32_to_cpu(agf->agf_length);
 	tree_len = be32_to_cpu(agf->agf_rmap_blocks);
 	xfs_trans_brelse(tp, agbp);
diff --git a/fs/xfs/libxfs/xfs_rmap_btree.h b/fs/xfs/libxfs/xfs_rmap_btree.h
index 820d668b063d..115c3455a734 100644
--- a/fs/xfs/libxfs/xfs_rmap_btree.h
+++ b/fs/xfs/libxfs/xfs_rmap_btree.h
@@ -9,6 +9,7 @@
 struct xfs_buf;
 struct xfs_btree_cur;
 struct xfs_mount;
+struct xbtree_afakeroot;
 
 /* rmaps only exist on crc enabled filesystems */
 #define XFS_RMAP_BLOCK_LEN	XFS_BTREE_SBLOCK_CRC_LEN
@@ -43,6 +44,10 @@ struct xfs_mount;
 struct xfs_btree_cur *xfs_rmapbt_init_cursor(struct xfs_mount *mp,
 				struct xfs_trans *tp, struct xfs_buf *bp,
 				xfs_agnumber_t agno);
+struct xfs_btree_cur *xfs_rmapbt_stage_cursor(struct xfs_mount *mp,
+		struct xbtree_afakeroot *afake, xfs_agnumber_t agno);
+void xfs_rmapbt_commit_staged_btree(struct xfs_btree_cur *cur,
+		struct xfs_trans *tp, struct xfs_buf *agbp);
 int xfs_rmapbt_maxrecs(int blocklen, int leaf);
 extern void xfs_rmapbt_compute_maxlevels(struct xfs_mount *mp);
 
diff --git a/fs/xfs/libxfs/xfs_rtbitmap.c b/fs/xfs/libxfs/xfs_rtbitmap.c
index f42c74cb8be5..1d9fa8a300f1 100644
--- a/fs/xfs/libxfs/xfs_rtbitmap.c
+++ b/fs/xfs/libxfs/xfs_rtbitmap.c
@@ -66,11 +66,11 @@ xfs_rtbuf_get(
 
 	ip = issum ? mp->m_rsumip : mp->m_rbmip;
 
-	error = xfs_bmapi_read(ip, block, 1, &map, &nmap, XFS_DATA_FORK);
+	error = xfs_bmapi_read(ip, block, 1, &map, &nmap, 0);
 	if (error)
 		return error;
 
-	if (XFS_IS_CORRUPT(mp, nmap == 0 || !xfs_bmap_is_real_extent(&map)))
+	if (XFS_IS_CORRUPT(mp, nmap == 0 || !xfs_bmap_is_written_extent(&map)))
 		return -EFSCORRUPTED;
 
 	ASSERT(map.br_startblock != NULLFSBLOCK);
diff --git a/fs/xfs/libxfs/xfs_sb.c b/fs/xfs/libxfs/xfs_sb.c
index 2f60fc3c99a0..ae9aaf1f34bf 100644
--- a/fs/xfs/libxfs/xfs_sb.c
+++ b/fs/xfs/libxfs/xfs_sb.c
@@ -220,7 +220,7 @@ xfs_validate_sb_common(
 	struct xfs_buf		*bp,
 	struct xfs_sb		*sbp)
 {
-	struct xfs_dsb		*dsb = XFS_BUF_TO_SBP(bp);
+	struct xfs_dsb		*dsb = bp->b_addr;
 	uint32_t		agcount = 0;
 	uint32_t		rem;
 
@@ -243,7 +243,7 @@ xfs_validate_sb_common(
 	} else if (sbp->sb_qflags & (XFS_PQUOTA_ENFD | XFS_GQUOTA_ENFD |
 				XFS_PQUOTA_CHKD | XFS_GQUOTA_CHKD)) {
 			xfs_notice(mp,
-"Superblock earlier than Version 5 has XFS_[PQ]UOTA_{ENFD|CHKD} bits.");
+"Superblock earlier than Version 5 has XFS_{P|G}QUOTA_{ENFD|CHKD} bits.");
 			return -EFSCORRUPTED;
 	}
 
@@ -328,6 +328,38 @@ xfs_validate_sb_common(
 		return -EFSCORRUPTED;
 	}
 
+	/* Validate the realtime geometry; stolen from xfs_repair */
+	if (sbp->sb_rextsize * sbp->sb_blocksize > XFS_MAX_RTEXTSIZE ||
+	    sbp->sb_rextsize * sbp->sb_blocksize < XFS_MIN_RTEXTSIZE) {
+		xfs_notice(mp,
+			"realtime extent sanity check failed");
+		return -EFSCORRUPTED;
+	}
+
+	if (sbp->sb_rblocks == 0) {
+		if (sbp->sb_rextents != 0 || sbp->sb_rbmblocks != 0 ||
+		    sbp->sb_rextslog != 0 || sbp->sb_frextents != 0) {
+			xfs_notice(mp,
+				"realtime zeroed geometry check failed");
+			return -EFSCORRUPTED;
+		}
+	} else {
+		uint64_t	rexts;
+		uint64_t	rbmblocks;
+
+		rexts = div_u64(sbp->sb_rblocks, sbp->sb_rextsize);
+		rbmblocks = howmany_64(sbp->sb_rextents,
+				       NBBY * sbp->sb_blocksize);
+
+		if (sbp->sb_rextents != rexts ||
+		    sbp->sb_rextslog != xfs_highbit32(sbp->sb_rextents) ||
+		    sbp->sb_rbmblocks != rbmblocks) {
+			xfs_notice(mp,
+				"realtime geometry sanity check failed");
+			return -EFSCORRUPTED;
+		}
+	}
+
 	if (sbp->sb_unit) {
 		if (!xfs_sb_version_hasdalign(sbp) ||
 		    sbp->sb_unit > sbp->sb_width ||
@@ -568,7 +600,7 @@ xfs_sb_quota_to_disk(
 	 * disk. If neither are active, we should NULL the inode.
 	 *
 	 * In all cases, the separate pquotino must remain 0 because it
-	 * it beyond the "end" of the valid non-pquotino superblock.
+	 * is beyond the "end" of the valid non-pquotino superblock.
 	 */
 	if (from->sb_qflags & XFS_GQUOTA_ACCT)
 		to->sb_gquotino = cpu_to_be64(from->sb_gquotino);
@@ -681,7 +713,7 @@ xfs_sb_read_verify(
 {
 	struct xfs_sb		sb;
 	struct xfs_mount	*mp = bp->b_mount;
-	struct xfs_dsb		*dsb = XFS_BUF_TO_SBP(bp);
+	struct xfs_dsb		*dsb = bp->b_addr;
 	int			error;
 
 	/*
@@ -707,7 +739,7 @@ xfs_sb_read_verify(
 	 * Check all the superblock fields.  Don't byteswap the xquota flags
 	 * because _verify_common checks the on-disk values.
 	 */
-	__xfs_sb_from_disk(&sb, XFS_BUF_TO_SBP(bp), false);
+	__xfs_sb_from_disk(&sb, dsb, false);
 	error = xfs_validate_sb_common(mp, bp, &sb);
 	if (error)
 		goto out_error;
@@ -730,7 +762,7 @@ static void
 xfs_sb_quiet_read_verify(
 	struct xfs_buf	*bp)
 {
-	struct xfs_dsb	*dsb = XFS_BUF_TO_SBP(bp);
+	struct xfs_dsb	*dsb = bp->b_addr;
 
 	if (dsb->sb_magicnum == cpu_to_be32(XFS_SB_MAGIC)) {
 		/* XFS filesystem, verify noisily! */
@@ -748,13 +780,14 @@ xfs_sb_write_verify(
 	struct xfs_sb		sb;
 	struct xfs_mount	*mp = bp->b_mount;
 	struct xfs_buf_log_item	*bip = bp->b_log_item;
+	struct xfs_dsb		*dsb = bp->b_addr;
 	int			error;
 
 	/*
 	 * Check all the superblock fields.  Don't byteswap the xquota flags
 	 * because _verify_common checks the on-disk values.
 	 */
-	__xfs_sb_from_disk(&sb, XFS_BUF_TO_SBP(bp), false);
+	__xfs_sb_from_disk(&sb, dsb, false);
 	error = xfs_validate_sb_common(mp, bp, &sb);
 	if (error)
 		goto out_error;
@@ -766,7 +799,7 @@ xfs_sb_write_verify(
 		return;
 
 	if (bip)
-		XFS_BUF_TO_SBP(bp)->sb_lsn = cpu_to_be64(bip->bli_item.li_lsn);
+		dsb->sb_lsn = cpu_to_be64(bip->bli_item.li_lsn);
 
 	xfs_buf_update_cksum(bp, XFS_SB_CRC_OFF);
 	return;
@@ -927,7 +960,7 @@ xfs_log_sb(
 	mp->m_sb.sb_ifree = percpu_counter_sum(&mp->m_ifree);
 	mp->m_sb.sb_fdblocks = percpu_counter_sum(&mp->m_fdblocks);
 
-	xfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb);
+	xfs_sb_to_disk(bp->b_addr, &mp->m_sb);
 	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_SB_BUF);
 	xfs_trans_log_buf(tp, bp, 0, sizeof(struct xfs_dsb) - 1);
 }
@@ -1007,7 +1040,7 @@ xfs_update_secondary_sbs(
 		bp->b_ops = &xfs_sb_buf_ops;
 		xfs_buf_oneshot(bp);
 		xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
-		xfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb);
+		xfs_sb_to_disk(bp->b_addr, &mp->m_sb);
 		xfs_buf_delwri_queue(bp, &buffer_list);
 		xfs_buf_relse(bp);
 
diff --git a/fs/xfs/libxfs/xfs_shared.h b/fs/xfs/libxfs/xfs_shared.h
index c45acbd3add9..708feb8eac76 100644
--- a/fs/xfs/libxfs/xfs_shared.h
+++ b/fs/xfs/libxfs/xfs_shared.h
@@ -65,6 +65,7 @@ void	xfs_log_get_max_trans_res(struct xfs_mount *mp,
 #define XFS_TRANS_DQ_DIRTY	0x10	/* at least one dquot in trx dirty */
 #define XFS_TRANS_RESERVE	0x20    /* OK to use reserved data blocks */
 #define XFS_TRANS_NO_WRITECOUNT 0x40	/* do not elevate SB writecount */
+#define XFS_TRANS_RES_FDBLKS	0x80	/* reserve newly freed blocks */
 /*
  * LOWMODE is used by the allocator to activate the lowspace algorithm - when
  * free space is running low the extent allocator may choose to allocate an
diff --git a/fs/xfs/libxfs/xfs_symlink_remote.c b/fs/xfs/libxfs/xfs_symlink_remote.c
index 3b8260ca7d1b..594bc447a7dd 100644
--- a/fs/xfs/libxfs/xfs_symlink_remote.c
+++ b/fs/xfs/libxfs/xfs_symlink_remote.c
@@ -204,16 +204,12 @@ xfs_failaddr_t
 xfs_symlink_shortform_verify(
 	struct xfs_inode	*ip)
 {
-	char			*sfp;
-	char			*endp;
-	struct xfs_ifork	*ifp;
-	int			size;
-
-	ASSERT(ip->i_d.di_format == XFS_DINODE_FMT_LOCAL);
-	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
-	sfp = (char *)ifp->if_u1.if_data;
-	size = ifp->if_bytes;
-	endp = sfp + size;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	char			*sfp = (char *)ifp->if_u1.if_data;
+	int			size = ifp->if_bytes;
+	char			*endp = sfp + size;
+
+	ASSERT(ifp->if_format == XFS_DINODE_FMT_LOCAL);
 
 	/*
 	 * Zero length symlinks should never occur in memory as they are
diff --git a/fs/xfs/libxfs/xfs_trans_inode.c b/fs/xfs/libxfs/xfs_trans_inode.c
index 2b8ccb5b975d..b7e222befb08 100644
--- a/fs/xfs/libxfs/xfs_trans_inode.c
+++ b/fs/xfs/libxfs/xfs_trans_inode.c
@@ -8,6 +8,8 @@
 #include "xfs_shared.h"
 #include "xfs_format.h"
 #include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
 #include "xfs_inode.h"
 #include "xfs_trans.h"
 #include "xfs_trans_priv.h"
@@ -27,7 +29,7 @@ xfs_trans_ijoin(
 	struct xfs_inode	*ip,
 	uint			lock_flags)
 {
-	xfs_inode_log_item_t	*iip;
+	struct xfs_inode_log_item *iip;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 	if (ip->i_itemp == NULL)
@@ -36,6 +38,7 @@ xfs_trans_ijoin(
 
 	ASSERT(iip->ili_lock_flags == 0);
 	iip->ili_lock_flags = lock_flags;
+	ASSERT(!xfs_iflags_test(ip, XFS_ISTALE));
 
 	/*
 	 * Get a log_item_desc to point at the new item.
@@ -71,24 +74,35 @@ xfs_trans_ichgtime(
 }
 
 /*
- * This is called to mark the fields indicated in fieldmask as needing
- * to be logged when the transaction is committed.  The inode must
- * already be associated with the given transaction.
+ * This is called to mark the fields indicated in fieldmask as needing to be
+ * logged when the transaction is committed.  The inode must already be
+ * associated with the given transaction.
  *
- * The values for fieldmask are defined in xfs_inode_item.h.  We always
- * log all of the core inode if any of it has changed, and we always log
- * all of the inline data/extents/b-tree root if any of them has changed.
+ * The values for fieldmask are defined in xfs_inode_item.h.  We always log all
+ * of the core inode if any of it has changed, and we always log all of the
+ * inline data/extents/b-tree root if any of them has changed.
+ *
+ * Grab and pin the cluster buffer associated with this inode to avoid RMW
+ * cycles at inode writeback time. Avoid the need to add error handling to every
+ * xfs_trans_log_inode() call by shutting down on read error.  This will cause
+ * transactions to fail and everything to error out, just like if we return a
+ * read error in a dirty transaction and cancel it.
  */
 void
 xfs_trans_log_inode(
-	xfs_trans_t	*tp,
-	xfs_inode_t	*ip,
-	uint		flags)
+	struct xfs_trans	*tp,
+	struct xfs_inode	*ip,
+	uint			flags)
 {
-	struct inode	*inode = VFS_I(ip);
+	struct xfs_inode_log_item *iip = ip->i_itemp;
+	struct inode		*inode = VFS_I(ip);
+	uint			iversion_flags = 0;
 
-	ASSERT(ip->i_itemp != NULL);
+	ASSERT(iip);
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	ASSERT(!xfs_iflags_test(ip, XFS_ISTALE));
+
+	tp->t_flags |= XFS_TRANS_DIRTY;
 
 	/*
 	 * Don't bother with i_lock for the I_DIRTY_TIME check here, as races
@@ -96,22 +110,13 @@ xfs_trans_log_inode(
 	 * to log the timestamps, or will clear already cleared fields in the
 	 * worst case.
 	 */
-	if (inode->i_state & (I_DIRTY_TIME | I_DIRTY_TIME_EXPIRED)) {
+	if (inode->i_state & I_DIRTY_TIME) {
 		spin_lock(&inode->i_lock);
-		inode->i_state &= ~(I_DIRTY_TIME | I_DIRTY_TIME_EXPIRED);
+		inode->i_state &= ~I_DIRTY_TIME;
 		spin_unlock(&inode->i_lock);
 	}
 
 	/*
-	 * Record the specific change for fdatasync optimisation. This
-	 * allows fdatasync to skip log forces for inodes that are only
-	 * timestamp dirty. We do this before the change count so that
-	 * the core being logged in this case does not impact on fdatasync
-	 * behaviour.
-	 */
-	ip->i_itemp->ili_fsync_fields |= flags;
-
-	/*
 	 * First time we log the inode in a transaction, bump the inode change
 	 * counter if it is configured for this to occur. While we have the
 	 * inode locked exclusively for metadata modification, we can usually
@@ -120,23 +125,64 @@ xfs_trans_log_inode(
 	 * set however, then go ahead and bump the i_version counter
 	 * unconditionally.
 	 */
-	if (!test_and_set_bit(XFS_LI_DIRTY, &ip->i_itemp->ili_item.li_flags) &&
-	    IS_I_VERSION(VFS_I(ip))) {
-		if (inode_maybe_inc_iversion(VFS_I(ip), flags & XFS_ILOG_CORE))
-			flags |= XFS_ILOG_CORE;
+	if (!test_and_set_bit(XFS_LI_DIRTY, &iip->ili_item.li_flags)) {
+		if (IS_I_VERSION(inode) &&
+		    inode_maybe_inc_iversion(inode, flags & XFS_ILOG_CORE))
+			iversion_flags = XFS_ILOG_CORE;
 	}
 
-	tp->t_flags |= XFS_TRANS_DIRTY;
+	/*
+	 * Record the specific change for fdatasync optimisation. This allows
+	 * fdatasync to skip log forces for inodes that are only timestamp
+	 * dirty.
+	 */
+	spin_lock(&iip->ili_lock);
+	iip->ili_fsync_fields |= flags;
+
+	if (!iip->ili_item.li_buf) {
+		struct xfs_buf	*bp;
+		int		error;
+
+		/*
+		 * We hold the ILOCK here, so this inode is not going to be
+		 * flushed while we are here. Further, because there is no
+		 * buffer attached to the item, we know that there is no IO in
+		 * progress, so nothing will clear the ili_fields while we read
+		 * in the buffer. Hence we can safely drop the spin lock and
+		 * read the buffer knowing that the state will not change from
+		 * here.
+		 */
+		spin_unlock(&iip->ili_lock);
+		error = xfs_imap_to_bp(ip->i_mount, tp, &ip->i_imap, NULL,
+					&bp, 0);
+		if (error) {
+			xfs_force_shutdown(ip->i_mount, SHUTDOWN_META_IO_ERROR);
+			return;
+		}
+
+		/*
+		 * We need an explicit buffer reference for the log item but
+		 * don't want the buffer to remain attached to the transaction.
+		 * Hold the buffer but release the transaction reference once
+		 * we've attached the inode log item to the buffer log item
+		 * list.
+		 */
+		xfs_buf_hold(bp);
+		spin_lock(&iip->ili_lock);
+		iip->ili_item.li_buf = bp;
+		bp->b_flags |= _XBF_INODES;
+		list_add_tail(&iip->ili_item.li_bio_list, &bp->b_li_list);
+		xfs_trans_brelse(tp, bp);
+	}
 
 	/*
-	 * Always OR in the bits from the ili_last_fields field.
-	 * This is to coordinate with the xfs_iflush() and xfs_iflush_done()
-	 * routines in the eventual clearing of the ili_fields bits.
-	 * See the big comment in xfs_iflush() for an explanation of
-	 * this coordination mechanism.
+	 * Always OR in the bits from the ili_last_fields field.  This is to
+	 * coordinate with the xfs_iflush() and xfs_iflush_done() routines in
+	 * the eventual clearing of the ili_fields bits.  See the big comment in
+	 * xfs_iflush() for an explanation of this coordination mechanism.
 	 */
-	flags |= ip->i_itemp->ili_last_fields;
-	ip->i_itemp->ili_fields |= flags;
+	iip->ili_fields |= (flags | iip->ili_last_fields | iversion_flags);
+	spin_unlock(&iip->ili_lock);
 }
 
 int
diff --git a/fs/xfs/libxfs/xfs_trans_resv.c b/fs/xfs/libxfs/xfs_trans_resv.c
index 7a9c04920505..d1a0848cb52e 100644
--- a/fs/xfs/libxfs/xfs_trans_resv.c
+++ b/fs/xfs/libxfs/xfs_trans_resv.c
@@ -187,7 +187,7 @@ xfs_calc_inode_chunk_res(
 			       XFS_FSB_TO_B(mp, 1));
 	if (alloc) {
 		/* icreate tx uses ordered buffers */
-		if (xfs_sb_version_hascrc(&mp->m_sb))
+		if (xfs_sb_version_has_v3inode(&mp->m_sb))
 			return res;
 		size = XFS_FSB_TO_B(mp, 1);
 	}
diff --git a/fs/xfs/libxfs/xfs_trans_space.h b/fs/xfs/libxfs/xfs_trans_space.h
index 88221c7a04cc..c6df01a2a158 100644
--- a/fs/xfs/libxfs/xfs_trans_space.h
+++ b/fs/xfs/libxfs/xfs_trans_space.h
@@ -57,7 +57,7 @@
 	XFS_DAREMOVE_SPACE_RES(mp, XFS_DATA_FORK)
 #define	XFS_IALLOC_SPACE_RES(mp)	\
 	(M_IGEO(mp)->ialloc_blks + \
-	 (xfs_sb_version_hasfinobt(&mp->m_sb) ? 2 : 1 * \
+	 ((xfs_sb_version_hasfinobt(&mp->m_sb) ? 2 : 1) * \
 	  (M_IGEO(mp)->inobt_maxlevels - 1)))
 
 /*
diff --git a/fs/xfs/scrub/agheader.c b/fs/xfs/scrub/agheader.c
index ba0f747c82e8..e9bcf1faa183 100644
--- a/fs/xfs/scrub/agheader.c
+++ b/fs/xfs/scrub/agheader.c
@@ -92,7 +92,7 @@ xchk_superblock(
 	if (!xchk_process_error(sc, agno, XFS_SB_BLOCK(mp), &error))
 		return error;
 
-	sb = XFS_BUF_TO_SBP(bp);
+	sb = bp->b_addr;
 
 	/*
 	 * Verify the geometries match.  Fields that are permanently
@@ -358,7 +358,7 @@ static inline void
 xchk_agf_xref_freeblks(
 	struct xfs_scrub	*sc)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	struct xfs_agf		*agf = sc->sa.agf_bp->b_addr;
 	xfs_extlen_t		blocks = 0;
 	int			error;
 
@@ -378,7 +378,7 @@ static inline void
 xchk_agf_xref_cntbt(
 	struct xfs_scrub	*sc)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	struct xfs_agf		*agf = sc->sa.agf_bp->b_addr;
 	xfs_agblock_t		agbno;
 	xfs_extlen_t		blocks;
 	int			have;
@@ -410,7 +410,7 @@ STATIC void
 xchk_agf_xref_btreeblks(
 	struct xfs_scrub	*sc)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	struct xfs_agf		*agf = sc->sa.agf_bp->b_addr;
 	struct xfs_mount	*mp = sc->mp;
 	xfs_agblock_t		blocks;
 	xfs_agblock_t		btreeblks;
@@ -456,7 +456,7 @@ static inline void
 xchk_agf_xref_refcblks(
 	struct xfs_scrub	*sc)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	struct xfs_agf		*agf = sc->sa.agf_bp->b_addr;
 	xfs_agblock_t		blocks;
 	int			error;
 
@@ -525,7 +525,7 @@ xchk_agf(
 		goto out;
 	xchk_buffer_recheck(sc, sc->sa.agf_bp);
 
-	agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	agf = sc->sa.agf_bp->b_addr;
 
 	/* Check the AG length */
 	eoag = be32_to_cpu(agf->agf_length);
@@ -711,7 +711,7 @@ xchk_agfl(
 		goto out;
 
 	/* Allocate buffer to ensure uniqueness of AGFL entries. */
-	agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	agf = sc->sa.agf_bp->b_addr;
 	agflcount = be32_to_cpu(agf->agf_flcount);
 	if (agflcount > xfs_agfl_size(sc->mp)) {
 		xchk_block_set_corrupt(sc, sc->sa.agf_bp);
@@ -728,7 +728,7 @@ xchk_agfl(
 	}
 
 	/* Check the blocks in the AGFL. */
-	error = xfs_agfl_walk(sc->mp, XFS_BUF_TO_AGF(sc->sa.agf_bp),
+	error = xfs_agfl_walk(sc->mp, sc->sa.agf_bp->b_addr,
 			sc->sa.agfl_bp, xchk_agfl_block, &sai);
 	if (error == -ECANCELED) {
 		error = 0;
@@ -765,7 +765,7 @@ static inline void
 xchk_agi_xref_icounts(
 	struct xfs_scrub	*sc)
 {
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(sc->sa.agi_bp);
+	struct xfs_agi		*agi = sc->sa.agi_bp->b_addr;
 	xfs_agino_t		icount;
 	xfs_agino_t		freecount;
 	int			error;
@@ -834,7 +834,7 @@ xchk_agi(
 		goto out;
 	xchk_buffer_recheck(sc, sc->sa.agi_bp);
 
-	agi = XFS_BUF_TO_AGI(sc->sa.agi_bp);
+	agi = sc->sa.agi_bp->b_addr;
 
 	/* Check the AG length */
 	eoag = be32_to_cpu(agi->agi_length);
diff --git a/fs/xfs/scrub/agheader_repair.c b/fs/xfs/scrub/agheader_repair.c
index d5e6db9af434..bca2ab1d4be9 100644
--- a/fs/xfs/scrub/agheader_repair.c
+++ b/fs/xfs/scrub/agheader_repair.c
@@ -49,7 +49,7 @@ xrep_superblock(
 
 	/* Copy AG 0's superblock to this one. */
 	xfs_buf_zero(bp, 0, BBTOB(bp->b_length));
-	xfs_sb_to_disk(XFS_BUF_TO_SBP(bp), &mp->m_sb);
+	xfs_sb_to_disk(bp->b_addr, &mp->m_sb);
 
 	/* Write this to disk. */
 	xfs_trans_buf_set_type(sc->tp, bp, XFS_BLFT_SB_BUF);
@@ -140,7 +140,7 @@ xrep_agf_find_btrees(
 	struct xrep_find_ag_btree	*fab,
 	struct xfs_buf			*agfl_bp)
 {
-	struct xfs_agf			*old_agf = XFS_BUF_TO_AGF(agf_bp);
+	struct xfs_agf			*old_agf = agf_bp->b_addr;
 	int				error;
 
 	/* Go find the root data. */
@@ -181,7 +181,7 @@ xrep_agf_init_header(
 	struct xfs_agf		*old_agf)
 {
 	struct xfs_mount	*mp = sc->mp;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agf_bp);
+	struct xfs_agf		*agf = agf_bp->b_addr;
 
 	memcpy(old_agf, agf, sizeof(*old_agf));
 	memset(agf, 0, BBTOB(agf_bp->b_length));
@@ -238,7 +238,7 @@ xrep_agf_calc_from_btrees(
 {
 	struct xrep_agf_allocbt	raa = { .sc = sc };
 	struct xfs_btree_cur	*cur = NULL;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agf_bp);
+	struct xfs_agf		*agf = agf_bp->b_addr;
 	struct xfs_mount	*mp = sc->mp;
 	xfs_agblock_t		btreeblks;
 	xfs_agblock_t		blocks;
@@ -302,7 +302,7 @@ xrep_agf_commit_new(
 	struct xfs_buf		*agf_bp)
 {
 	struct xfs_perag	*pag;
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agf_bp);
+	struct xfs_agf		*agf = agf_bp->b_addr;
 
 	/* Trigger fdblocks recalculation */
 	xfs_force_summary_recalc(sc->mp);
@@ -376,7 +376,7 @@ xrep_agf(
 	if (error)
 		return error;
 	agf_bp->b_ops = &xfs_agf_buf_ops;
-	agf = XFS_BUF_TO_AGF(agf_bp);
+	agf = agf_bp->b_addr;
 
 	/*
 	 * Load the AGFL so that we can screen out OWN_AG blocks that are on
@@ -395,7 +395,7 @@ xrep_agf(
 	 * Spot-check the AGFL blocks; if they're obviously corrupt then
 	 * there's nothing we can do but bail out.
 	 */
-	error = xfs_agfl_walk(sc->mp, XFS_BUF_TO_AGF(agf_bp), agfl_bp,
+	error = xfs_agfl_walk(sc->mp, agf_bp->b_addr, agfl_bp,
 			xrep_agf_check_agfl_block, sc);
 	if (error)
 		return error;
@@ -429,10 +429,10 @@ out_revert:
 
 struct xrep_agfl {
 	/* Bitmap of other OWN_AG metadata blocks. */
-	struct xfs_bitmap	agmetablocks;
+	struct xbitmap		agmetablocks;
 
 	/* Bitmap of free space. */
-	struct xfs_bitmap	*freesp;
+	struct xbitmap		*freesp;
 
 	struct xfs_scrub	*sc;
 };
@@ -453,14 +453,14 @@ xrep_agfl_walk_rmap(
 
 	/* Record all the OWN_AG blocks. */
 	if (rec->rm_owner == XFS_RMAP_OWN_AG) {
-		fsb = XFS_AGB_TO_FSB(cur->bc_mp, cur->bc_private.a.agno,
+		fsb = XFS_AGB_TO_FSB(cur->bc_mp, cur->bc_ag.agno,
 				rec->rm_startblock);
-		error = xfs_bitmap_set(ra->freesp, fsb, rec->rm_blockcount);
+		error = xbitmap_set(ra->freesp, fsb, rec->rm_blockcount);
 		if (error)
 			return error;
 	}
 
-	return xfs_bitmap_set_btcur_path(&ra->agmetablocks, cur);
+	return xbitmap_set_btcur_path(&ra->agmetablocks, cur);
 }
 
 /*
@@ -476,19 +476,17 @@ STATIC int
 xrep_agfl_collect_blocks(
 	struct xfs_scrub	*sc,
 	struct xfs_buf		*agf_bp,
-	struct xfs_bitmap	*agfl_extents,
+	struct xbitmap		*agfl_extents,
 	xfs_agblock_t		*flcount)
 {
 	struct xrep_agfl	ra;
 	struct xfs_mount	*mp = sc->mp;
 	struct xfs_btree_cur	*cur;
-	struct xfs_bitmap_range	*br;
-	struct xfs_bitmap_range	*n;
 	int			error;
 
 	ra.sc = sc;
 	ra.freesp = agfl_extents;
-	xfs_bitmap_init(&ra.agmetablocks);
+	xbitmap_init(&ra.agmetablocks);
 
 	/* Find all space used by the free space btrees & rmapbt. */
 	cur = xfs_rmapbt_init_cursor(mp, sc->tp, agf_bp, sc->sa.agno);
@@ -500,7 +498,7 @@ xrep_agfl_collect_blocks(
 	/* Find all blocks currently being used by the bnobt. */
 	cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp, sc->sa.agno,
 			XFS_BTNUM_BNO);
-	error = xfs_bitmap_set_btblocks(&ra.agmetablocks, cur);
+	error = xbitmap_set_btblocks(&ra.agmetablocks, cur);
 	if (error)
 		goto err;
 	xfs_btree_del_cursor(cur, error);
@@ -508,7 +506,7 @@ xrep_agfl_collect_blocks(
 	/* Find all blocks currently being used by the cntbt. */
 	cur = xfs_allocbt_init_cursor(mp, sc->tp, agf_bp, sc->sa.agno,
 			XFS_BTNUM_CNT);
-	error = xfs_bitmap_set_btblocks(&ra.agmetablocks, cur);
+	error = xbitmap_set_btblocks(&ra.agmetablocks, cur);
 	if (error)
 		goto err;
 
@@ -518,8 +516,8 @@ xrep_agfl_collect_blocks(
 	 * Drop the freesp meta blocks that are in use by btrees.
 	 * The remaining blocks /should/ be AGFL blocks.
 	 */
-	error = xfs_bitmap_disunion(agfl_extents, &ra.agmetablocks);
-	xfs_bitmap_destroy(&ra.agmetablocks);
+	error = xbitmap_disunion(agfl_extents, &ra.agmetablocks);
+	xbitmap_destroy(&ra.agmetablocks);
 	if (error)
 		return error;
 
@@ -527,18 +525,12 @@ xrep_agfl_collect_blocks(
 	 * Calculate the new AGFL size.  If we found more blocks than fit in
 	 * the AGFL we'll free them later.
 	 */
-	*flcount = 0;
-	for_each_xfs_bitmap_extent(br, n, agfl_extents) {
-		*flcount += br->len;
-		if (*flcount > xfs_agfl_size(mp))
-			break;
-	}
-	if (*flcount > xfs_agfl_size(mp))
-		*flcount = xfs_agfl_size(mp);
+	*flcount = min_t(uint64_t, xbitmap_hweight(agfl_extents),
+			 xfs_agfl_size(mp));
 	return 0;
 
 err:
-	xfs_bitmap_destroy(&ra.agmetablocks);
+	xbitmap_destroy(&ra.agmetablocks);
 	xfs_btree_del_cursor(cur, error);
 	return error;
 }
@@ -550,7 +542,7 @@ xrep_agfl_update_agf(
 	struct xfs_buf		*agf_bp,
 	xfs_agblock_t		flcount)
 {
-	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agf_bp);
+	struct xfs_agf		*agf = agf_bp->b_addr;
 
 	ASSERT(flcount <= xfs_agfl_size(sc->mp));
 
@@ -573,13 +565,13 @@ STATIC void
 xrep_agfl_init_header(
 	struct xfs_scrub	*sc,
 	struct xfs_buf		*agfl_bp,
-	struct xfs_bitmap	*agfl_extents,
+	struct xbitmap		*agfl_extents,
 	xfs_agblock_t		flcount)
 {
 	struct xfs_mount	*mp = sc->mp;
 	__be32			*agfl_bno;
-	struct xfs_bitmap_range	*br;
-	struct xfs_bitmap_range	*n;
+	struct xbitmap_range	*br;
+	struct xbitmap_range	*n;
 	struct xfs_agfl		*agfl;
 	xfs_agblock_t		agbno;
 	unsigned int		fl_off;
@@ -602,8 +594,8 @@ xrep_agfl_init_header(
 	 * step.
 	 */
 	fl_off = 0;
-	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agfl_bp);
-	for_each_xfs_bitmap_extent(br, n, agfl_extents) {
+	agfl_bno = xfs_buf_to_agfl_bno(agfl_bp);
+	for_each_xbitmap_extent(br, n, agfl_extents) {
 		agbno = XFS_FSB_TO_AGBNO(mp, br->start);
 
 		trace_xrep_agfl_insert(mp, sc->sa.agno, agbno, br->len);
@@ -637,7 +629,7 @@ int
 xrep_agfl(
 	struct xfs_scrub	*sc)
 {
-	struct xfs_bitmap	agfl_extents;
+	struct xbitmap		agfl_extents;
 	struct xfs_mount	*mp = sc->mp;
 	struct xfs_buf		*agf_bp;
 	struct xfs_buf		*agfl_bp;
@@ -649,7 +641,7 @@ xrep_agfl(
 		return -EOPNOTSUPP;
 
 	xchk_perag_get(sc->mp, &sc->sa);
-	xfs_bitmap_init(&agfl_extents);
+	xbitmap_init(&agfl_extents);
 
 	/*
 	 * Read the AGF so that we can query the rmapbt.  We hope that there's
@@ -696,10 +688,10 @@ xrep_agfl(
 		goto err;
 
 	/* Dump any AGFL overflow. */
-	return xrep_reap_extents(sc, &agfl_extents, &XFS_RMAP_OINFO_AG,
+	error = xrep_reap_extents(sc, &agfl_extents, &XFS_RMAP_OINFO_AG,
 			XFS_AG_RESV_AGFL);
 err:
-	xfs_bitmap_destroy(&agfl_extents);
+	xbitmap_destroy(&agfl_extents);
 	return error;
 }
 
@@ -761,7 +753,7 @@ xrep_agi_init_header(
 	struct xfs_buf		*agi_bp,
 	struct xfs_agi		*old_agi)
 {
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agi_bp);
+	struct xfs_agi		*agi = agi_bp->b_addr;
 	struct xfs_mount	*mp = sc->mp;
 
 	memcpy(old_agi, agi, sizeof(*old_agi));
@@ -807,7 +799,7 @@ xrep_agi_calc_from_btrees(
 	struct xfs_buf		*agi_bp)
 {
 	struct xfs_btree_cur	*cur;
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agi_bp);
+	struct xfs_agi		*agi = agi_bp->b_addr;
 	struct xfs_mount	*mp = sc->mp;
 	xfs_agino_t		count;
 	xfs_agino_t		freecount;
@@ -835,7 +827,7 @@ xrep_agi_commit_new(
 	struct xfs_buf		*agi_bp)
 {
 	struct xfs_perag	*pag;
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agi_bp);
+	struct xfs_agi		*agi = agi_bp->b_addr;
 
 	/* Trigger inode count recalculation */
 	xfs_force_summary_recalc(sc->mp);
@@ -892,7 +884,7 @@ xrep_agi(
 	if (error)
 		return error;
 	agi_bp->b_ops = &xfs_agi_buf_ops;
-	agi = XFS_BUF_TO_AGI(agi_bp);
+	agi = agi_bp->b_addr;
 
 	/* Find the AGI btree roots. */
 	error = xrep_agi_find_btrees(sc, fab);
diff --git a/fs/xfs/scrub/alloc.c b/fs/xfs/scrub/alloc.c
index 5533e48e605d..73d924e47565 100644
--- a/fs/xfs/scrub/alloc.c
+++ b/fs/xfs/scrub/alloc.c
@@ -94,7 +94,7 @@ xchk_allocbt_rec(
 	union xfs_btree_rec	*rec)
 {
 	struct xfs_mount	*mp = bs->cur->bc_mp;
-	xfs_agnumber_t		agno = bs->cur->bc_private.a.agno;
+	xfs_agnumber_t		agno = bs->cur->bc_ag.agno;
 	xfs_agblock_t		bno;
 	xfs_extlen_t		len;
 
diff --git a/fs/xfs/scrub/attr.c b/fs/xfs/scrub/attr.c
index d9f0dd444b80..9faddb334a2c 100644
--- a/fs/xfs/scrub/attr.c
+++ b/fs/xfs/scrub/attr.c
@@ -98,7 +98,7 @@ struct xchk_xattr {
 /*
  * Check that an extended attribute key can be looked up by hash.
  *
- * We use the XFS attribute list iterator (i.e. xfs_attr_list_int_ilocked)
+ * We use the XFS attribute list iterator (i.e. xfs_attr_list_ilocked)
  * to call this function for every attribute key in an inode.  Once
  * we're here, we load the attribute value to see if any errors happen,
  * or if we get more or less data than we expected.
@@ -147,11 +147,8 @@ xchk_xattr_listent(
 		return;
 	}
 
-	args.flags = ATTR_KERNOTIME;
-	if (flags & XFS_ATTR_ROOT)
-		args.flags |= ATTR_ROOT;
-	else if (flags & XFS_ATTR_SECURE)
-		args.flags |= ATTR_SECURE;
+	args.op_flags = XFS_DA_OP_NOTIME;
+	args.attr_filter = flags & XFS_ATTR_NSP_ONDISK_MASK;
 	args.geo = context->dp->i_mount->m_attr_geo;
 	args.whichfork = XFS_ATTR_FORK;
 	args.dp = context->dp;
@@ -162,7 +159,10 @@ xchk_xattr_listent(
 	args.value = xchk_xattr_valuebuf(sx->sc);
 	args.valuelen = valuelen;
 
-	error = xfs_attr_get_ilocked(context->dp, &args);
+	error = xfs_attr_get_ilocked(&args);
+	/* ENODATA means the hash lookup failed and the attr is bad */
+	if (error == -ENODATA)
+		error = -EFSCORRUPTED;
 	if (!xchk_fblock_process_error(sx->sc, XFS_ATTR_FORK, args.blkno,
 			&error))
 		goto fail_xref;
@@ -474,7 +474,6 @@ xchk_xattr(
 	struct xfs_scrub		*sc)
 {
 	struct xchk_xattr		sx;
-	struct attrlist_cursor_kern	cursor = { 0 };
 	xfs_dablk_t			last_checked = -1U;
 	int				error = 0;
 
@@ -493,11 +492,10 @@ xchk_xattr(
 
 	/* Check that every attr key can also be looked up by hash. */
 	sx.context.dp = sc->ip;
-	sx.context.cursor = &cursor;
 	sx.context.resynch = 1;
 	sx.context.put_listent = xchk_xattr_listent;
 	sx.context.tp = sc->tp;
-	sx.context.flags = ATTR_INCOMPLETE;
+	sx.context.allow_incomplete = true;
 	sx.sc = sc;
 
 	/*
@@ -516,7 +514,7 @@ xchk_xattr(
 	 * iteration, which doesn't really follow the usual buffer
 	 * locking order.
 	 */
-	error = xfs_attr_list_int_ilocked(&sx.context);
+	error = xfs_attr_list_ilocked(&sx.context);
 	if (!xchk_fblock_process_error(sc, XFS_ATTR_FORK, 0, &error))
 		goto out;
 
diff --git a/fs/xfs/scrub/bitmap.c b/fs/xfs/scrub/bitmap.c
index 18a684e18a69..f88694f22d05 100644
--- a/fs/xfs/scrub/bitmap.c
+++ b/fs/xfs/scrub/bitmap.c
@@ -18,14 +18,14 @@
  * This is the logical equivalent of bitmap |= mask(start, len).
  */
 int
-xfs_bitmap_set(
-	struct xfs_bitmap	*bitmap,
+xbitmap_set(
+	struct xbitmap		*bitmap,
 	uint64_t		start,
 	uint64_t		len)
 {
-	struct xfs_bitmap_range	*bmr;
+	struct xbitmap_range	*bmr;
 
-	bmr = kmem_alloc(sizeof(struct xfs_bitmap_range), KM_MAYFAIL);
+	bmr = kmem_alloc(sizeof(struct xbitmap_range), KM_MAYFAIL);
 	if (!bmr)
 		return -ENOMEM;
 
@@ -39,13 +39,13 @@ xfs_bitmap_set(
 
 /* Free everything related to this bitmap. */
 void
-xfs_bitmap_destroy(
-	struct xfs_bitmap	*bitmap)
+xbitmap_destroy(
+	struct xbitmap		*bitmap)
 {
-	struct xfs_bitmap_range	*bmr;
-	struct xfs_bitmap_range	*n;
+	struct xbitmap_range	*bmr;
+	struct xbitmap_range	*n;
 
-	for_each_xfs_bitmap_extent(bmr, n, bitmap) {
+	for_each_xbitmap_extent(bmr, n, bitmap) {
 		list_del(&bmr->list);
 		kmem_free(bmr);
 	}
@@ -53,24 +53,24 @@ xfs_bitmap_destroy(
 
 /* Set up a per-AG block bitmap. */
 void
-xfs_bitmap_init(
-	struct xfs_bitmap	*bitmap)
+xbitmap_init(
+	struct xbitmap		*bitmap)
 {
 	INIT_LIST_HEAD(&bitmap->list);
 }
 
 /* Compare two btree extents. */
 static int
-xfs_bitmap_range_cmp(
+xbitmap_range_cmp(
 	void			*priv,
 	struct list_head	*a,
 	struct list_head	*b)
 {
-	struct xfs_bitmap_range	*ap;
-	struct xfs_bitmap_range	*bp;
+	struct xbitmap_range	*ap;
+	struct xbitmap_range	*bp;
 
-	ap = container_of(a, struct xfs_bitmap_range, list);
-	bp = container_of(b, struct xfs_bitmap_range, list);
+	ap = container_of(a, struct xbitmap_range, list);
+	bp = container_of(b, struct xbitmap_range, list);
 
 	if (ap->start > bp->start)
 		return 1;
@@ -96,14 +96,14 @@ xfs_bitmap_range_cmp(
 #define LEFT_ALIGNED	(1 << 0)
 #define RIGHT_ALIGNED	(1 << 1)
 int
-xfs_bitmap_disunion(
-	struct xfs_bitmap	*bitmap,
-	struct xfs_bitmap	*sub)
+xbitmap_disunion(
+	struct xbitmap		*bitmap,
+	struct xbitmap		*sub)
 {
 	struct list_head	*lp;
-	struct xfs_bitmap_range	*br;
-	struct xfs_bitmap_range	*new_br;
-	struct xfs_bitmap_range	*sub_br;
+	struct xbitmap_range	*br;
+	struct xbitmap_range	*new_br;
+	struct xbitmap_range	*sub_br;
 	uint64_t		sub_start;
 	uint64_t		sub_len;
 	int			state;
@@ -113,8 +113,8 @@ xfs_bitmap_disunion(
 		return 0;
 	ASSERT(!list_empty(&sub->list));
 
-	list_sort(NULL, &bitmap->list, xfs_bitmap_range_cmp);
-	list_sort(NULL, &sub->list, xfs_bitmap_range_cmp);
+	list_sort(NULL, &bitmap->list, xbitmap_range_cmp);
+	list_sort(NULL, &sub->list, xbitmap_range_cmp);
 
 	/*
 	 * Now that we've sorted both lists, we iterate bitmap once, rolling
@@ -124,11 +124,11 @@ xfs_bitmap_disunion(
 	 * list traversal is similar to merge sort, but we're deleting
 	 * instead.  In this manner we avoid O(n^2) operations.
 	 */
-	sub_br = list_first_entry(&sub->list, struct xfs_bitmap_range,
+	sub_br = list_first_entry(&sub->list, struct xbitmap_range,
 			list);
 	lp = bitmap->list.next;
 	while (lp != &bitmap->list) {
-		br = list_entry(lp, struct xfs_bitmap_range, list);
+		br = list_entry(lp, struct xbitmap_range, list);
 
 		/*
 		 * Advance sub_br and/or br until we find a pair that
@@ -181,7 +181,7 @@ xfs_bitmap_disunion(
 			 * Deleting from the middle: add the new right extent
 			 * and then shrink the left extent.
 			 */
-			new_br = kmem_alloc(sizeof(struct xfs_bitmap_range),
+			new_br = kmem_alloc(sizeof(struct xbitmap_range),
 					KM_MAYFAIL);
 			if (!new_br) {
 				error = -ENOMEM;
@@ -247,8 +247,8 @@ out:
  * blocks going from the leaf towards the root.
  */
 int
-xfs_bitmap_set_btcur_path(
-	struct xfs_bitmap	*bitmap,
+xbitmap_set_btcur_path(
+	struct xbitmap		*bitmap,
 	struct xfs_btree_cur	*cur)
 {
 	struct xfs_buf		*bp;
@@ -261,7 +261,7 @@ xfs_bitmap_set_btcur_path(
 		if (!bp)
 			continue;
 		fsb = XFS_DADDR_TO_FSB(cur->bc_mp, bp->b_bn);
-		error = xfs_bitmap_set(bitmap, fsb, 1);
+		error = xbitmap_set(bitmap, fsb, 1);
 		if (error)
 			return error;
 	}
@@ -271,12 +271,12 @@ xfs_bitmap_set_btcur_path(
 
 /* Collect a btree's block in the bitmap. */
 STATIC int
-xfs_bitmap_collect_btblock(
+xbitmap_collect_btblock(
 	struct xfs_btree_cur	*cur,
 	int			level,
 	void			*priv)
 {
-	struct xfs_bitmap	*bitmap = priv;
+	struct xbitmap		*bitmap = priv;
 	struct xfs_buf		*bp;
 	xfs_fsblock_t		fsbno;
 
@@ -285,15 +285,30 @@ xfs_bitmap_collect_btblock(
 		return 0;
 
 	fsbno = XFS_DADDR_TO_FSB(cur->bc_mp, bp->b_bn);
-	return xfs_bitmap_set(bitmap, fsbno, 1);
+	return xbitmap_set(bitmap, fsbno, 1);
 }
 
 /* Walk the btree and mark the bitmap wherever a btree block is found. */
 int
-xfs_bitmap_set_btblocks(
-	struct xfs_bitmap	*bitmap,
+xbitmap_set_btblocks(
+	struct xbitmap		*bitmap,
 	struct xfs_btree_cur	*cur)
 {
-	return xfs_btree_visit_blocks(cur, xfs_bitmap_collect_btblock,
+	return xfs_btree_visit_blocks(cur, xbitmap_collect_btblock,
 			XFS_BTREE_VISIT_ALL, bitmap);
 }
+
+/* How many bits are set in this bitmap? */
+uint64_t
+xbitmap_hweight(
+	struct xbitmap		*bitmap)
+{
+	struct xbitmap_range	*bmr;
+	struct xbitmap_range	*n;
+	uint64_t		ret = 0;
+
+	for_each_xbitmap_extent(bmr, n, bitmap)
+		ret += bmr->len;
+
+	return ret;
+}
diff --git a/fs/xfs/scrub/bitmap.h b/fs/xfs/scrub/bitmap.h
index ae8ecbce6fa6..900646b72de1 100644
--- a/fs/xfs/scrub/bitmap.h
+++ b/fs/xfs/scrub/bitmap.h
@@ -6,31 +6,32 @@
 #ifndef __XFS_SCRUB_BITMAP_H__
 #define __XFS_SCRUB_BITMAP_H__
 
-struct xfs_bitmap_range {
+struct xbitmap_range {
 	struct list_head	list;
 	uint64_t		start;
 	uint64_t		len;
 };
 
-struct xfs_bitmap {
+struct xbitmap {
 	struct list_head	list;
 };
 
-void xfs_bitmap_init(struct xfs_bitmap *bitmap);
-void xfs_bitmap_destroy(struct xfs_bitmap *bitmap);
+void xbitmap_init(struct xbitmap *bitmap);
+void xbitmap_destroy(struct xbitmap *bitmap);
 
-#define for_each_xfs_bitmap_extent(bex, n, bitmap) \
+#define for_each_xbitmap_extent(bex, n, bitmap) \
 	list_for_each_entry_safe((bex), (n), &(bitmap)->list, list)
 
-#define for_each_xfs_bitmap_block(b, bex, n, bitmap) \
+#define for_each_xbitmap_block(b, bex, n, bitmap) \
 	list_for_each_entry_safe((bex), (n), &(bitmap)->list, list) \
-		for ((b) = bex->start; (b) < bex->start + bex->len; (b)++)
+		for ((b) = (bex)->start; (b) < (bex)->start + (bex)->len; (b)++)
 
-int xfs_bitmap_set(struct xfs_bitmap *bitmap, uint64_t start, uint64_t len);
-int xfs_bitmap_disunion(struct xfs_bitmap *bitmap, struct xfs_bitmap *sub);
-int xfs_bitmap_set_btcur_path(struct xfs_bitmap *bitmap,
+int xbitmap_set(struct xbitmap *bitmap, uint64_t start, uint64_t len);
+int xbitmap_disunion(struct xbitmap *bitmap, struct xbitmap *sub);
+int xbitmap_set_btcur_path(struct xbitmap *bitmap,
 		struct xfs_btree_cur *cur);
-int xfs_bitmap_set_btblocks(struct xfs_bitmap *bitmap,
+int xbitmap_set_btblocks(struct xbitmap *bitmap,
 		struct xfs_btree_cur *cur);
+uint64_t xbitmap_hweight(struct xbitmap *bitmap);
 
 #endif	/* __XFS_SCRUB_BITMAP_H__ */
diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
index fa6ea6407992..955302e7cdde 100644
--- a/fs/xfs/scrub/bmap.c
+++ b/fs/xfs/scrub/bmap.c
@@ -45,9 +45,27 @@ xchk_setup_inode_bmap(
 	 */
 	if (S_ISREG(VFS_I(sc->ip)->i_mode) &&
 	    sc->sm->sm_type == XFS_SCRUB_TYPE_BMBTD) {
+		struct address_space	*mapping = VFS_I(sc->ip)->i_mapping;
+
 		inode_dio_wait(VFS_I(sc->ip));
-		error = filemap_write_and_wait(VFS_I(sc->ip)->i_mapping);
-		if (error)
+
+		/*
+		 * Try to flush all incore state to disk before we examine the
+		 * space mappings for the data fork.  Leave accumulated errors
+		 * in the mapping for the writer threads to consume.
+		 *
+		 * On ENOSPC or EIO writeback errors, we continue into the
+		 * extent mapping checks because write failures do not
+		 * necessarily imply anything about the correctness of the file
+		 * metadata.  The metadata and the file data could be on
+		 * completely separate devices; a media failure might only
+		 * affect a subset of the disk, etc.  We can handle delalloc
+		 * extents in the scrubber, so leaving them in memory is fine.
+		 */
+		error = filemap_fdatawrite(mapping);
+		if (!error)
+			error = filemap_fdatawait_keep_errors(mapping);
+		if (error && (error != -ENOSPC && error != -EIO))
 			goto out;
 	}
 
@@ -374,7 +392,7 @@ xchk_bmapbt_rec(
 	struct xfs_bmbt_irec	iext_irec;
 	struct xfs_iext_cursor	icur;
 	struct xchk_bmap_info	*info = bs->private;
-	struct xfs_inode	*ip = bs->cur->bc_private.b.ip;
+	struct xfs_inode	*ip = bs->cur->bc_ino.ip;
 	struct xfs_buf		*bp = NULL;
 	struct xfs_btree_block	*block;
 	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, info->whichfork);
@@ -501,7 +519,7 @@ xchk_bmap_check_rmap(
 			xchk_fblock_set_corrupt(sc, sbcri->whichfork,
 					rec->rm_offset);
 		if (irec.br_startblock != XFS_AGB_TO_FSB(sc->mp,
-				cur->bc_private.a.agno, rec->rm_startblock))
+				cur->bc_ag.agno, rec->rm_startblock))
 			xchk_fblock_set_corrupt(sc, sbcri->whichfork,
 					rec->rm_offset);
 		if (irec.br_blockcount > rec->rm_blockcount)
@@ -566,8 +584,9 @@ xchk_bmap_check_rmaps(
 	struct xfs_scrub	*sc,
 	int			whichfork)
 {
-	loff_t			size;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(sc->ip, whichfork);
 	xfs_agnumber_t		agno;
+	bool			zero_size;
 	int			error;
 
 	if (!xfs_sb_version_hasrmapbt(&sc->mp->m_sb) ||
@@ -579,6 +598,8 @@ xchk_bmap_check_rmaps(
 	if (XFS_IS_REALTIME_INODE(sc->ip) && whichfork == XFS_DATA_FORK)
 		return 0;
 
+	ASSERT(XFS_IFORK_PTR(sc->ip, whichfork) != NULL);
+
 	/*
 	 * Only do this for complex maps that are in btree format, or for
 	 * situations where we would seem to have a size but zero extents.
@@ -586,19 +607,14 @@ xchk_bmap_check_rmaps(
 	 * to flag this bmap as corrupt if there are rmaps that need to be
 	 * reattached.
 	 */
-	switch (whichfork) {
-	case XFS_DATA_FORK:
-		size = i_size_read(VFS_I(sc->ip));
-		break;
-	case XFS_ATTR_FORK:
-		size = XFS_IFORK_Q(sc->ip);
-		break;
-	default:
-		size = 0;
-		break;
-	}
-	if (XFS_IFORK_FORMAT(sc->ip, whichfork) != XFS_DINODE_FMT_BTREE &&
-	    (size == 0 || XFS_IFORK_NEXTENTS(sc->ip, whichfork) > 0))
+
+	if (whichfork == XFS_DATA_FORK)
+		zero_size = i_size_read(VFS_I(sc->ip)) == 0;
+	else
+		zero_size = false;
+
+	if (ifp->if_format != XFS_DINODE_FMT_BTREE &&
+	    (zero_size || ifp->if_nextents > 0))
 		return 0;
 
 	for (agno = 0; agno < sc->mp->m_sb.sb_agcount; agno++) {
@@ -627,12 +643,14 @@ xchk_bmap(
 	struct xchk_bmap_info	info = { NULL };
 	struct xfs_mount	*mp = sc->mp;
 	struct xfs_inode	*ip = sc->ip;
-	struct xfs_ifork	*ifp;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 	xfs_fileoff_t		endoff;
 	struct xfs_iext_cursor	icur;
 	int			error = 0;
 
-	ifp = XFS_IFORK_PTR(ip, whichfork);
+	/* Non-existent forks can be ignored. */
+	if (!ifp)
+		goto out;
 
 	info.is_rt = whichfork == XFS_DATA_FORK && XFS_IS_REALTIME_INODE(ip);
 	info.whichfork = whichfork;
@@ -641,9 +659,6 @@ xchk_bmap(
 
 	switch (whichfork) {
 	case XFS_COW_FORK:
-		/* Non-existent CoW forks are ignorable. */
-		if (!ifp)
-			goto out;
 		/* No CoW forks on non-reflink inodes/filesystems. */
 		if (!xfs_is_reflink_inode(ip)) {
 			xchk_ino_set_corrupt(sc, sc->ip->i_ino);
@@ -651,8 +666,6 @@ xchk_bmap(
 		}
 		break;
 	case XFS_ATTR_FORK:
-		if (!ifp)
-			goto out_check_rmap;
 		if (!xfs_sb_version_hasattr(&mp->m_sb) &&
 		    !xfs_sb_version_hasattr2(&mp->m_sb))
 			xchk_ino_set_corrupt(sc, sc->ip->i_ino);
@@ -663,7 +676,7 @@ xchk_bmap(
 	}
 
 	/* Check the fork values */
-	switch (XFS_IFORK_FORMAT(ip, whichfork)) {
+	switch (ifp->if_format) {
 	case XFS_DINODE_FMT_UUID:
 	case XFS_DINODE_FMT_DEV:
 	case XFS_DINODE_FMT_LOCAL:
@@ -717,7 +730,6 @@ xchk_bmap(
 			goto out;
 	}
 
-out_check_rmap:
 	error = xchk_bmap_check_rmaps(sc, whichfork);
 	if (!xchk_fblock_xref_process_error(sc, whichfork, 0, &error))
 		goto out;
diff --git a/fs/xfs/scrub/dabtree.c b/fs/xfs/scrub/dabtree.c
index 97a15b6f2865..e56786f0a13c 100644
--- a/fs/xfs/scrub/dabtree.c
+++ b/fs/xfs/scrub/dabtree.c
@@ -219,19 +219,21 @@ xchk_da_btree_block_check_sibling(
 	int			direction,
 	xfs_dablk_t		sibling)
 {
+	struct xfs_da_state_path *path = &ds->state->path;
+	struct xfs_da_state_path *altpath = &ds->state->altpath;
 	int			retval;
+	int			plevel;
 	int			error;
 
-	memcpy(&ds->state->altpath, &ds->state->path,
-			sizeof(ds->state->altpath));
+	memcpy(altpath, path, sizeof(ds->state->altpath));
 
 	/*
 	 * If the pointer is null, we shouldn't be able to move the upper
 	 * level pointer anywhere.
 	 */
 	if (sibling == 0) {
-		error = xfs_da3_path_shift(ds->state, &ds->state->altpath,
-				direction, false, &retval);
+		error = xfs_da3_path_shift(ds->state, altpath, direction,
+				false, &retval);
 		if (error == 0 && retval == 0)
 			xchk_da_set_corrupt(ds, level);
 		error = 0;
@@ -239,27 +241,33 @@ xchk_da_btree_block_check_sibling(
 	}
 
 	/* Move the alternate cursor one block in the direction given. */
-	error = xfs_da3_path_shift(ds->state, &ds->state->altpath,
-			direction, false, &retval);
+	error = xfs_da3_path_shift(ds->state, altpath, direction, false,
+			&retval);
 	if (!xchk_da_process_error(ds, level, &error))
-		return error;
+		goto out;
 	if (retval) {
 		xchk_da_set_corrupt(ds, level);
-		return error;
+		goto out;
 	}
-	if (ds->state->altpath.blk[level].bp)
-		xchk_buffer_recheck(ds->sc,
-				ds->state->altpath.blk[level].bp);
+	if (altpath->blk[level].bp)
+		xchk_buffer_recheck(ds->sc, altpath->blk[level].bp);
 
 	/* Compare upper level pointer to sibling pointer. */
-	if (ds->state->altpath.blk[level].blkno != sibling)
+	if (altpath->blk[level].blkno != sibling)
 		xchk_da_set_corrupt(ds, level);
-	if (ds->state->altpath.blk[level].bp) {
-		xfs_trans_brelse(ds->dargs.trans,
-				ds->state->altpath.blk[level].bp);
-		ds->state->altpath.blk[level].bp = NULL;
-	}
+
 out:
+	/* Free all buffers in the altpath that aren't referenced from path. */
+	for (plevel = 0; plevel < altpath->active; plevel++) {
+		if (altpath->blk[plevel].bp == NULL ||
+		    (plevel < path->active &&
+		     altpath->blk[plevel].bp == path->blk[plevel].bp))
+			continue;
+
+		xfs_trans_brelse(ds->dargs.trans, altpath->blk[plevel].bp);
+		altpath->blk[plevel].bp = NULL;
+	}
+
 	return error;
 }
 
@@ -460,7 +468,7 @@ xchk_da_btree(
 	int				error;
 
 	/* Skip short format data structures; no btree to scan. */
-	if (!xfs_ifork_has_extents(sc->ip, whichfork))
+	if (!xfs_ifork_has_extents(XFS_IFORK_PTR(sc->ip, whichfork)))
 		return 0;
 
 	/* Set up initial da state. */
@@ -468,9 +476,7 @@ xchk_da_btree(
 	ds.dargs.whichfork = whichfork;
 	ds.dargs.trans = sc->tp;
 	ds.dargs.op_flags = XFS_DA_OP_OKNOENT;
-	ds.state = xfs_da_state_alloc();
-	ds.state->args = &ds.dargs;
-	ds.state->mp = mp;
+	ds.state = xfs_da_state_alloc(&ds.dargs);
 	ds.sc = sc;
 	ds.private = private;
 	if (whichfork == XFS_ATTR_FORK) {
diff --git a/fs/xfs/scrub/dir.c b/fs/xfs/scrub/dir.c
index 266da4e4bde6..7c432997edad 100644
--- a/fs/xfs/scrub/dir.c
+++ b/fs/xfs/scrub/dir.c
@@ -155,6 +155,9 @@ xchk_dir_actor(
 	xname.type = XFS_DIR3_FT_UNKNOWN;
 
 	error = xfs_dir_lookup(sdc->sc->tp, ip, &xname, &lookup_ino, NULL);
+	/* ENOENT means the hash lookup failed and the dir is corrupt */
+	if (error == -ENOENT)
+		error = -EFSCORRUPTED;
 	if (!xchk_fblock_process_error(sdc->sc, XFS_DATA_FORK, offset,
 			&error))
 		goto out;
@@ -500,7 +503,7 @@ xchk_directory_leaf1_bestfree(
 	/* Read the free space block. */
 	error = xfs_dir3_leaf_read(sc->tp, sc->ip, lblk, &bp);
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
-		goto out;
+		return error;
 	xchk_buffer_recheck(sc, bp);
 
 	leaf = bp->b_addr;
@@ -565,9 +568,10 @@ xchk_directory_leaf1_bestfree(
 		xchk_directory_check_freesp(sc, lblk, dbp, best);
 		xfs_trans_brelse(sc->tp, dbp);
 		if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
-			goto out;
+			break;
 	}
 out:
+	xfs_trans_brelse(sc->tp, bp);
 	return error;
 }
 
@@ -589,7 +593,7 @@ xchk_directory_free_bestfree(
 	/* Read the free space block */
 	error = xfs_dir2_free_read(sc->tp, sc->ip, lblk, &bp);
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
-		goto out;
+		return error;
 	xchk_buffer_recheck(sc, bp);
 
 	if (xfs_sb_version_hascrc(&sc->mp->m_sb)) {
@@ -612,7 +616,7 @@ xchk_directory_free_bestfree(
 				0, &dbp);
 		if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, lblk,
 				&error))
-			break;
+			goto out;
 		xchk_directory_check_freesp(sc, lblk, dbp, best);
 		xfs_trans_brelse(sc->tp, dbp);
 	}
@@ -620,6 +624,7 @@ xchk_directory_free_bestfree(
 	if (freehdr.nused + stale != freehdr.nvalid)
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
 out:
+	xfs_trans_brelse(sc->tp, bp);
 	return error;
 }
 
@@ -630,7 +635,7 @@ xchk_directory_blocks(
 {
 	struct xfs_bmbt_irec	got;
 	struct xfs_da_args	args;
-	struct xfs_ifork	*ifp;
+	struct xfs_ifork	*ifp = XFS_IFORK_PTR(sc->ip, XFS_DATA_FORK);
 	struct xfs_mount	*mp = sc->mp;
 	xfs_fileoff_t		leaf_lblk;
 	xfs_fileoff_t		free_lblk;
@@ -642,11 +647,10 @@ xchk_directory_blocks(
 	int			error;
 
 	/* Ignore local format directories. */
-	if (sc->ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS &&
-	    sc->ip->i_d.di_format != XFS_DINODE_FMT_BTREE)
+	if (ifp->if_format != XFS_DINODE_FMT_EXTENTS &&
+	    ifp->if_format != XFS_DINODE_FMT_BTREE)
 		return 0;
 
-	ifp = XFS_IFORK_PTR(sc->ip, XFS_DATA_FORK);
 	lblk = XFS_B_TO_FSB(mp, XFS_DIR2_DATA_OFFSET);
 	leaf_lblk = XFS_B_TO_FSB(mp, XFS_DIR2_LEAF_OFFSET);
 	free_lblk = XFS_B_TO_FSB(mp, XFS_DIR2_FREE_OFFSET);
diff --git a/fs/xfs/scrub/ialloc.c b/fs/xfs/scrub/ialloc.c
index 681758704fda..6517d67e8d51 100644
--- a/fs/xfs/scrub/ialloc.c
+++ b/fs/xfs/scrub/ialloc.c
@@ -104,7 +104,7 @@ xchk_iallocbt_chunk(
 	xfs_extlen_t			len)
 {
 	struct xfs_mount		*mp = bs->cur->bc_mp;
-	xfs_agnumber_t			agno = bs->cur->bc_private.a.agno;
+	xfs_agnumber_t			agno = bs->cur->bc_ag.agno;
 	xfs_agblock_t			bno;
 
 	bno = XFS_AGINO_TO_AGBNO(mp, agino);
@@ -164,7 +164,7 @@ xchk_iallocbt_check_cluster_ifree(
 	 * the record, compute which fs inode we're talking about.
 	 */
 	agino = irec->ir_startino + irec_ino;
-	fsino = XFS_AGINO_TO_INO(mp, bs->cur->bc_private.a.agno, agino);
+	fsino = XFS_AGINO_TO_INO(mp, bs->cur->bc_ag.agno, agino);
 	irec_free = (irec->ir_free & XFS_INOBT_MASK(irec_ino));
 
 	if (be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC ||
@@ -215,7 +215,7 @@ xchk_iallocbt_check_cluster(
 	struct xfs_dinode		*dip;
 	struct xfs_buf			*cluster_bp;
 	unsigned int			nr_inodes;
-	xfs_agnumber_t			agno = bs->cur->bc_private.a.agno;
+	xfs_agnumber_t			agno = bs->cur->bc_ag.agno;
 	xfs_agblock_t			agbno;
 	unsigned int			cluster_index;
 	uint16_t			cluster_mask = 0;
@@ -278,8 +278,7 @@ xchk_iallocbt_check_cluster(
 			&XFS_RMAP_OINFO_INODES);
 
 	/* Grab the inode cluster buffer. */
-	error = xfs_imap_to_bp(mp, bs->cur->bc_tp, &imap, &dip, &cluster_bp,
-			0, 0);
+	error = xfs_imap_to_bp(mp, bs->cur->bc_tp, &imap, &dip, &cluster_bp, 0);
 	if (!xchk_btree_xref_process_error(bs->sc, bs->cur, 0, &error))
 		return error;
 
@@ -426,7 +425,7 @@ xchk_iallocbt_rec(
 	struct xchk_iallocbt		*iabt = bs->private;
 	struct xfs_inobt_rec_incore	irec;
 	uint64_t			holes;
-	xfs_agnumber_t			agno = bs->cur->bc_private.a.agno;
+	xfs_agnumber_t			agno = bs->cur->bc_ag.agno;
 	xfs_agino_t			agino;
 	xfs_extlen_t			len;
 	int				holecount;
diff --git a/fs/xfs/scrub/parent.c b/fs/xfs/scrub/parent.c
index 5705adc43a75..855aa8bcab64 100644
--- a/fs/xfs/scrub/parent.c
+++ b/fs/xfs/scrub/parent.c
@@ -90,7 +90,7 @@ xchk_parent_count_parent_dentries(
 	 * if there is one.
 	 */
 	lock_mode = xfs_ilock_data_map_shared(parent);
-	if (parent->i_d.di_nextents > 0)
+	if (parent->i_df.if_nextents > 0)
 		error = xfs_dir3_data_readahead(parent, 0, 0);
 	xfs_iunlock(parent, lock_mode);
 	if (error)
diff --git a/fs/xfs/scrub/quota.c b/fs/xfs/scrub/quota.c
index 905a34558361..e34ca20ae8e4 100644
--- a/fs/xfs/scrub/quota.c
+++ b/fs/xfs/scrub/quota.c
@@ -18,17 +18,17 @@
 #include "scrub/common.h"
 
 /* Convert a scrub type code to a DQ flag, or return 0 if error. */
-static inline uint
+static inline xfs_dqtype_t
 xchk_quota_to_dqtype(
 	struct xfs_scrub	*sc)
 {
 	switch (sc->sm->sm_type) {
 	case XFS_SCRUB_TYPE_UQUOTA:
-		return XFS_DQ_USER;
+		return XFS_DQTYPE_USER;
 	case XFS_SCRUB_TYPE_GQUOTA:
-		return XFS_DQ_GROUP;
+		return XFS_DQTYPE_GROUP;
 	case XFS_SCRUB_TYPE_PQUOTA:
-		return XFS_DQ_PROJ;
+		return XFS_DQTYPE_PROJ;
 	default:
 		return 0;
 	}
@@ -40,7 +40,7 @@ xchk_setup_quota(
 	struct xfs_scrub	*sc,
 	struct xfs_inode	*ip)
 {
-	uint			dqtype;
+	xfs_dqtype_t		dqtype;
 	int			error;
 
 	if (!XFS_IS_QUOTA_RUNNING(sc->mp) || !XFS_IS_QUOTA_ON(sc->mp))
@@ -73,26 +73,15 @@ struct xchk_quota_info {
 STATIC int
 xchk_quota_item(
 	struct xfs_dquot	*dq,
-	uint			dqtype,
+	xfs_dqtype_t		dqtype,
 	void			*priv)
 {
 	struct xchk_quota_info	*sqi = priv;
 	struct xfs_scrub	*sc = sqi->sc;
 	struct xfs_mount	*mp = sc->mp;
-	struct xfs_disk_dquot	*d = &dq->q_core;
 	struct xfs_quotainfo	*qi = mp->m_quotainfo;
 	xfs_fileoff_t		offset;
-	unsigned long long	bsoft;
-	unsigned long long	isoft;
-	unsigned long long	rsoft;
-	unsigned long long	bhard;
-	unsigned long long	ihard;
-	unsigned long long	rhard;
-	unsigned long long	bcount;
-	unsigned long long	icount;
-	unsigned long long	rcount;
 	xfs_ino_t		fs_icount;
-	xfs_dqid_t		id = be32_to_cpu(d->d_id);
 	int			error = 0;
 
 	if (xchk_should_terminate(sc, &error))
@@ -102,27 +91,11 @@ xchk_quota_item(
 	 * Except for the root dquot, the actual dquot we got must either have
 	 * the same or higher id as we saw before.
 	 */
-	offset = id / qi->qi_dqperchunk;
-	if (id && id <= sqi->last_id)
+	offset = dq->q_id / qi->qi_dqperchunk;
+	if (dq->q_id && dq->q_id <= sqi->last_id)
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
 
-	sqi->last_id = id;
-
-	/* Did we get the dquot type we wanted? */
-	if (dqtype != (d->d_flags & XFS_DQ_ALLTYPES))
-		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
-
-	if (d->d_pad0 != cpu_to_be32(0) || d->d_pad != cpu_to_be16(0))
-		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
-
-	/* Check the limits. */
-	bhard = be64_to_cpu(d->d_blk_hardlimit);
-	ihard = be64_to_cpu(d->d_ino_hardlimit);
-	rhard = be64_to_cpu(d->d_rtb_hardlimit);
-
-	bsoft = be64_to_cpu(d->d_blk_softlimit);
-	isoft = be64_to_cpu(d->d_ino_softlimit);
-	rsoft = be64_to_cpu(d->d_rtb_softlimit);
+	sqi->last_id = dq->q_id;
 
 	/*
 	 * Warn if the hard limits are larger than the fs.
@@ -132,25 +105,22 @@ xchk_quota_item(
 	 * Complain about corruption if the soft limit is greater than
 	 * the hard limit.
 	 */
-	if (bhard > mp->m_sb.sb_dblocks)
+	if (dq->q_blk.hardlimit > mp->m_sb.sb_dblocks)
 		xchk_fblock_set_warning(sc, XFS_DATA_FORK, offset);
-	if (bsoft > bhard)
+	if (dq->q_blk.softlimit > dq->q_blk.hardlimit)
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
 
-	if (ihard > M_IGEO(mp)->maxicount)
+	if (dq->q_ino.hardlimit > M_IGEO(mp)->maxicount)
 		xchk_fblock_set_warning(sc, XFS_DATA_FORK, offset);
-	if (isoft > ihard)
+	if (dq->q_ino.softlimit > dq->q_ino.hardlimit)
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
 
-	if (rhard > mp->m_sb.sb_rblocks)
+	if (dq->q_rtb.hardlimit > mp->m_sb.sb_rblocks)
 		xchk_fblock_set_warning(sc, XFS_DATA_FORK, offset);
-	if (rsoft > rhard)
+	if (dq->q_rtb.softlimit > dq->q_rtb.hardlimit)
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
 
 	/* Check the resource counts. */
-	bcount = be64_to_cpu(d->d_bcount);
-	icount = be64_to_cpu(d->d_icount);
-	rcount = be64_to_cpu(d->d_rtbcount);
 	fs_icount = percpu_counter_sum(&mp->m_icount);
 
 	/*
@@ -159,15 +129,15 @@ xchk_quota_item(
 	 * if there are no quota limits.
 	 */
 	if (xfs_sb_version_hasreflink(&mp->m_sb)) {
-		if (mp->m_sb.sb_dblocks < bcount)
+		if (mp->m_sb.sb_dblocks < dq->q_blk.count)
 			xchk_fblock_set_warning(sc, XFS_DATA_FORK,
 					offset);
 	} else {
-		if (mp->m_sb.sb_dblocks < bcount)
+		if (mp->m_sb.sb_dblocks < dq->q_blk.count)
 			xchk_fblock_set_corrupt(sc, XFS_DATA_FORK,
 					offset);
 	}
-	if (icount > fs_icount || rcount > mp->m_sb.sb_rblocks)
+	if (dq->q_ino.count > fs_icount || dq->q_rtb.count > mp->m_sb.sb_rblocks)
 		xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
 
 	/*
@@ -175,13 +145,22 @@ xchk_quota_item(
 	 * lower limit than the actual usage.  However, we flag it for
 	 * admin review.
 	 */
-	if (id != 0 && bhard != 0 && bcount > bhard)
+	if (dq->q_id == 0)
+		goto out;
+
+	if (dq->q_blk.hardlimit != 0 &&
+	    dq->q_blk.count > dq->q_blk.hardlimit)
 		xchk_fblock_set_warning(sc, XFS_DATA_FORK, offset);
-	if (id != 0 && ihard != 0 && icount > ihard)
+
+	if (dq->q_ino.hardlimit != 0 &&
+	    dq->q_ino.count > dq->q_ino.hardlimit)
 		xchk_fblock_set_warning(sc, XFS_DATA_FORK, offset);
-	if (id != 0 && rhard != 0 && rcount > rhard)
+
+	if (dq->q_rtb.hardlimit != 0 &&
+	    dq->q_rtb.count > dq->q_rtb.hardlimit)
 		xchk_fblock_set_warning(sc, XFS_DATA_FORK, offset);
 
+out:
 	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
 		return -EFSCORRUPTED;
 
@@ -235,7 +214,7 @@ xchk_quota(
 	struct xchk_quota_info	sqi;
 	struct xfs_mount	*mp = sc->mp;
 	struct xfs_quotainfo	*qi = mp->m_quotainfo;
-	uint			dqtype;
+	xfs_dqtype_t		dqtype;
 	int			error = 0;
 
 	dqtype = xchk_quota_to_dqtype(sc);
diff --git a/fs/xfs/scrub/refcount.c b/fs/xfs/scrub/refcount.c
index 0cab11a5d390..beaeb6fa3119 100644
--- a/fs/xfs/scrub/refcount.c
+++ b/fs/xfs/scrub/refcount.c
@@ -336,7 +336,7 @@ xchk_refcountbt_rec(
 {
 	struct xfs_mount	*mp = bs->cur->bc_mp;
 	xfs_agblock_t		*cow_blocks = bs->private;
-	xfs_agnumber_t		agno = bs->cur->bc_private.a.agno;
+	xfs_agnumber_t		agno = bs->cur->bc_ag.agno;
 	xfs_agblock_t		bno;
 	xfs_extlen_t		len;
 	xfs_nlink_t		refcount;
diff --git a/fs/xfs/scrub/repair.c b/fs/xfs/scrub/repair.c
index e489d7a8446a..25e86c71e7b9 100644
--- a/fs/xfs/scrub/repair.c
+++ b/fs/xfs/scrub/repair.c
@@ -208,8 +208,10 @@ xrep_calc_ag_resblks(
 	/* Now grab the block counters from the AGF. */
 	error = xfs_alloc_read_agf(mp, NULL, sm->sm_agno, 0, &bp);
 	if (!error) {
-		aglen = be32_to_cpu(XFS_BUF_TO_AGF(bp)->agf_length);
-		freelen = be32_to_cpu(XFS_BUF_TO_AGF(bp)->agf_freeblks);
+		struct xfs_agf	*agf = bp->b_addr;
+
+		aglen = be32_to_cpu(agf->agf_length);
+		freelen = be32_to_cpu(agf->agf_freeblks);
 		usedlen = aglen - freelen;
 		xfs_buf_relse(bp);
 	}
@@ -434,10 +436,10 @@ xrep_init_btblock(
 int
 xrep_invalidate_blocks(
 	struct xfs_scrub	*sc,
-	struct xfs_bitmap	*bitmap)
+	struct xbitmap		*bitmap)
 {
-	struct xfs_bitmap_range	*bmr;
-	struct xfs_bitmap_range	*n;
+	struct xbitmap_range	*bmr;
+	struct xbitmap_range	*n;
 	struct xfs_buf		*bp;
 	xfs_fsblock_t		fsbno;
 
@@ -449,7 +451,7 @@ xrep_invalidate_blocks(
 	 * because we never own those; and if we can't TRYLOCK the buffer we
 	 * assume it's owned by someone else.
 	 */
-	for_each_xfs_bitmap_block(fsbno, bmr, n, bitmap) {
+	for_each_xbitmap_block(fsbno, bmr, n, bitmap) {
 		/* Skip AG headers and post-EOFS blocks */
 		if (!xfs_verify_fsbno(sc->mp, fsbno))
 			continue;
@@ -595,18 +597,18 @@ out_free:
 int
 xrep_reap_extents(
 	struct xfs_scrub		*sc,
-	struct xfs_bitmap		*bitmap,
+	struct xbitmap			*bitmap,
 	const struct xfs_owner_info	*oinfo,
 	enum xfs_ag_resv_type		type)
 {
-	struct xfs_bitmap_range		*bmr;
-	struct xfs_bitmap_range		*n;
+	struct xbitmap_range		*bmr;
+	struct xbitmap_range		*n;
 	xfs_fsblock_t			fsbno;
 	int				error = 0;
 
 	ASSERT(xfs_sb_version_hasrmapbt(&sc->mp->m_sb));
 
-	for_each_xfs_bitmap_block(fsbno, bmr, n, bitmap) {
+	for_each_xbitmap_block(fsbno, bmr, n, bitmap) {
 		ASSERT(sc->ip != NULL ||
 		       XFS_FSB_TO_AGNO(sc->mp, fsbno) == sc->sa.agno);
 		trace_xrep_dispose_btree_extent(sc->mp,
@@ -615,11 +617,9 @@ xrep_reap_extents(
 
 		error = xrep_reap_block(sc, fsbno, oinfo, type);
 		if (error)
-			goto out;
+			break;
 	}
 
-out:
-	xfs_bitmap_destroy(bitmap);
 	return error;
 }
 
@@ -879,7 +879,7 @@ xrep_find_ag_btree_roots(
 
 	ri.sc = sc;
 	ri.btree_info = btree_info;
-	ri.agf = XFS_BUF_TO_AGF(agf_bp);
+	ri.agf = agf_bp->b_addr;
 	ri.agfl_bp = agfl_bp;
 	for (fab = btree_info; fab->buf_ops; fab++) {
 		ASSERT(agfl_bp || fab->rmap_owner != XFS_RMAP_OWN_AG);
@@ -899,11 +899,11 @@ xrep_find_ag_btree_roots(
 void
 xrep_force_quotacheck(
 	struct xfs_scrub	*sc,
-	uint			dqtype)
+	xfs_dqtype_t		type)
 {
 	uint			flag;
 
-	flag = xfs_quota_chkd_flag(dqtype);
+	flag = xfs_quota_chkd_flag(type);
 	if (!(flag & sc->mp->m_qflags))
 		return;
 
@@ -939,11 +939,11 @@ xrep_ino_dqattach(
 "inode %llu repair encountered quota error %d, quotacheck forced.",
 				(unsigned long long)sc->ip->i_ino, error);
 		if (XFS_IS_UQUOTA_ON(sc->mp) && !sc->ip->i_udquot)
-			xrep_force_quotacheck(sc, XFS_DQ_USER);
+			xrep_force_quotacheck(sc, XFS_DQTYPE_USER);
 		if (XFS_IS_GQUOTA_ON(sc->mp) && !sc->ip->i_gdquot)
-			xrep_force_quotacheck(sc, XFS_DQ_GROUP);
+			xrep_force_quotacheck(sc, XFS_DQTYPE_GROUP);
 		if (XFS_IS_PQUOTA_ON(sc->mp) && !sc->ip->i_pdquot)
-			xrep_force_quotacheck(sc, XFS_DQ_PROJ);
+			xrep_force_quotacheck(sc, XFS_DQTYPE_PROJ);
 		/* fall through */
 	case -ESRCH:
 		error = 0;
diff --git a/fs/xfs/scrub/repair.h b/fs/xfs/scrub/repair.h
index c3422403b169..fe77de01abe0 100644
--- a/fs/xfs/scrub/repair.h
+++ b/fs/xfs/scrub/repair.h
@@ -6,6 +6,8 @@
 #ifndef __XFS_SCRUB_REPAIR_H__
 #define __XFS_SCRUB_REPAIR_H__
 
+#include "xfs_quota_defs.h"
+
 static inline int xrep_notsupported(struct xfs_scrub *sc)
 {
 	return -EOPNOTSUPP;
@@ -28,11 +30,11 @@ int xrep_init_btblock(struct xfs_scrub *sc, xfs_fsblock_t fsb,
 		struct xfs_buf **bpp, xfs_btnum_t btnum,
 		const struct xfs_buf_ops *ops);
 
-struct xfs_bitmap;
+struct xbitmap;
 
 int xrep_fix_freelist(struct xfs_scrub *sc, bool can_shrink);
-int xrep_invalidate_blocks(struct xfs_scrub *sc, struct xfs_bitmap *btlist);
-int xrep_reap_extents(struct xfs_scrub *sc, struct xfs_bitmap *exlist,
+int xrep_invalidate_blocks(struct xfs_scrub *sc, struct xbitmap *btlist);
+int xrep_reap_extents(struct xfs_scrub *sc, struct xbitmap *exlist,
 		const struct xfs_owner_info *oinfo, enum xfs_ag_resv_type type);
 
 struct xrep_find_ag_btree {
@@ -49,7 +51,7 @@ struct xrep_find_ag_btree {
 
 int xrep_find_ag_btree_roots(struct xfs_scrub *sc, struct xfs_buf *agf_bp,
 		struct xrep_find_ag_btree *btree_info, struct xfs_buf *agfl_bp);
-void xrep_force_quotacheck(struct xfs_scrub *sc, uint dqtype);
+void xrep_force_quotacheck(struct xfs_scrub *sc, xfs_dqtype_t type);
 int xrep_ino_dqattach(struct xfs_scrub *sc);
 
 /* Metadata repairers */
diff --git a/fs/xfs/scrub/rmap.c b/fs/xfs/scrub/rmap.c
index 8d4cefd761c1..f4fcb4719f41 100644
--- a/fs/xfs/scrub/rmap.c
+++ b/fs/xfs/scrub/rmap.c
@@ -92,7 +92,7 @@ xchk_rmapbt_rec(
 {
 	struct xfs_mount	*mp = bs->cur->bc_mp;
 	struct xfs_rmap_irec	irec;
-	xfs_agnumber_t		agno = bs->cur->bc_private.a.agno;
+	xfs_agnumber_t		agno = bs->cur->bc_ag.agno;
 	bool			non_inode;
 	bool			is_unwritten;
 	bool			is_bmbt;
diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c
index c642bc206c41..76e4ffe0315b 100644
--- a/fs/xfs/scrub/rtbitmap.c
+++ b/fs/xfs/scrub/rtbitmap.c
@@ -13,6 +13,7 @@
 #include "xfs_trans.h"
 #include "xfs_rtalloc.h"
 #include "xfs_inode.h"
+#include "xfs_bmap.h"
 #include "scrub/scrub.h"
 #include "scrub/common.h"
 
@@ -58,6 +59,41 @@ xchk_rtbitmap_rec(
 	return 0;
 }
 
+/* Make sure the entire rtbitmap file is mapped with written extents. */
+STATIC int
+xchk_rtbitmap_check_extents(
+	struct xfs_scrub	*sc)
+{
+	struct xfs_mount	*mp = sc->mp;
+	struct xfs_bmbt_irec	map;
+	xfs_rtblock_t		off;
+	int			nmap;
+	int			error = 0;
+
+	for (off = 0; off < mp->m_sb.sb_rbmblocks;) {
+		if (xchk_should_terminate(sc, &error) ||
+		    (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+			break;
+
+		/* Make sure we have a written extent. */
+		nmap = 1;
+		error = xfs_bmapi_read(mp->m_rbmip, off,
+				mp->m_sb.sb_rbmblocks - off, &map, &nmap,
+				XFS_DATA_FORK);
+		if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, off, &error))
+			break;
+
+		if (nmap != 1 || !xfs_bmap_is_written_extent(&map)) {
+			xchk_fblock_set_corrupt(sc, XFS_DATA_FORK, off);
+			break;
+		}
+
+		off += map.br_blockcount;
+	}
+
+	return error;
+}
+
 /* Scrub the realtime bitmap. */
 int
 xchk_rtbitmap(
@@ -65,11 +101,22 @@ xchk_rtbitmap(
 {
 	int			error;
 
+	/* Is the size of the rtbitmap correct? */
+	if (sc->mp->m_rbmip->i_d.di_size !=
+	    XFS_FSB_TO_B(sc->mp, sc->mp->m_sb.sb_rbmblocks)) {
+		xchk_ino_set_corrupt(sc, sc->mp->m_rbmip->i_ino);
+		return 0;
+	}
+
 	/* Invoke the fork scrubber. */
 	error = xchk_metadata_inode_forks(sc);
 	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
 		return error;
 
+	error = xchk_rtbitmap_check_extents(sc);
+	if (error || (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		return error;
+
 	error = xfs_rtalloc_query_all(sc->tp, xchk_rtbitmap_rec, sc);
 	if (!xchk_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
 		goto out;
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
index f1775bb19313..8ebf35b115ce 100644
--- a/fs/xfs/scrub/scrub.c
+++ b/fs/xfs/scrub/scrub.c
@@ -168,6 +168,7 @@ xchk_teardown(
 			xfs_irele(sc->ip);
 		sc->ip = NULL;
 	}
+	sb_end_write(sc->mp->m_super);
 	if (sc->flags & XCHK_REAPING_DISABLED)
 		xchk_start_reaping(sc);
 	if (sc->flags & XCHK_HAS_QUOTAOFFLOCK) {
@@ -490,6 +491,14 @@ xfs_scrub_metadata(
 	sc.ops = &meta_scrub_ops[sm->sm_type];
 	sc.sick_mask = xchk_health_mask_for_scrub_type(sm->sm_type);
 retry_op:
+	/*
+	 * If freeze runs concurrently with a scrub, the freeze can be delayed
+	 * indefinitely as we walk the filesystem and iterate over metadata
+	 * buffers.  Freeze quiesces the log (which waits for the buffer LRU to
+	 * be emptied) and that won't happen while checking is running.
+	 */
+	sb_start_write(mp->m_super);
+
 	/* Set up for the operation. */
 	error = sc.ops->setup(&sc, ip);
 	if (error)
diff --git a/fs/xfs/scrub/trace.c b/fs/xfs/scrub/trace.c
index 9eaab2eb5ed3..2c6c248be823 100644
--- a/fs/xfs/scrub/trace.c
+++ b/fs/xfs/scrub/trace.c
@@ -24,9 +24,9 @@ xchk_btree_cur_fsbno(
 		return XFS_DADDR_TO_FSB(cur->bc_mp, cur->bc_bufs[level]->b_bn);
 	else if (level == cur->bc_nlevels - 1 &&
 		 cur->bc_flags & XFS_BTREE_LONG_PTRS)
-		return XFS_INO_TO_FSB(cur->bc_mp, cur->bc_private.b.ip->i_ino);
+		return XFS_INO_TO_FSB(cur->bc_mp, cur->bc_ino.ip->i_ino);
 	else if (!(cur->bc_flags & XFS_BTREE_LONG_PTRS))
-		return XFS_AGB_TO_FSB(cur->bc_mp, cur->bc_private.a.agno, 0);
+		return XFS_AGB_TO_FSB(cur->bc_mp, cur->bc_ag.agno, 0);
 	return NULLFSBLOCK;
 }
 
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
index 096203119934..e46f5cef90da 100644
--- a/fs/xfs/scrub/trace.h
+++ b/fs/xfs/scrub/trace.h
@@ -379,7 +379,7 @@ TRACE_EVENT(xchk_ifork_btree_op_error,
 		xfs_fsblock_t fsbno = xchk_btree_cur_fsbno(cur, level);
 		__entry->dev = sc->mp->m_super->s_dev;
 		__entry->ino = sc->ip->i_ino;
-		__entry->whichfork = cur->bc_private.b.whichfork;
+		__entry->whichfork = cur->bc_ino.whichfork;
 		__entry->type = sc->sm->sm_type;
 		__entry->btnum = cur->bc_btnum;
 		__entry->level = level;
@@ -459,7 +459,7 @@ TRACE_EVENT(xchk_ifork_btree_error,
 		xfs_fsblock_t fsbno = xchk_btree_cur_fsbno(cur, level);
 		__entry->dev = sc->mp->m_super->s_dev;
 		__entry->ino = sc->ip->i_ino;
-		__entry->whichfork = cur->bc_private.b.whichfork;
+		__entry->whichfork = cur->bc_ino.whichfork;
 		__entry->type = sc->sm->sm_type;
 		__entry->btnum = cur->bc_btnum;
 		__entry->level = level;
diff --git a/fs/xfs/xfs_acl.c b/fs/xfs/xfs_acl.c
index cd743fad8478..d4c687b5cd06 100644
--- a/fs/xfs/xfs_acl.c
+++ b/fs/xfs/xfs_acl.c
@@ -14,6 +14,8 @@
 #include "xfs_trace.h"
 #include "xfs_error.h"
 #include "xfs_acl.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
 
 #include <linux/posix_acl_xattr.h>
 
@@ -67,10 +69,12 @@ xfs_acl_from_disk(
 
 		switch (acl_e->e_tag) {
 		case ACL_USER:
-			acl_e->e_uid = xfs_uid_to_kuid(be32_to_cpu(ace->ae_id));
+			acl_e->e_uid = make_kuid(&init_user_ns,
+						 be32_to_cpu(ace->ae_id));
 			break;
 		case ACL_GROUP:
-			acl_e->e_gid = xfs_gid_to_kgid(be32_to_cpu(ace->ae_id));
+			acl_e->e_gid = make_kgid(&init_user_ns,
+						 be32_to_cpu(ace->ae_id));
 			break;
 		case ACL_USER_OBJ:
 		case ACL_GROUP_OBJ:
@@ -103,10 +107,12 @@ xfs_acl_to_disk(struct xfs_acl *aclp, const struct posix_acl *acl)
 		ace->ae_tag = cpu_to_be32(acl_e->e_tag);
 		switch (acl_e->e_tag) {
 		case ACL_USER:
-			ace->ae_id = cpu_to_be32(xfs_kuid_to_uid(acl_e->e_uid));
+			ace->ae_id = cpu_to_be32(
+					from_kuid(&init_user_ns, acl_e->e_uid));
 			break;
 		case ACL_GROUP:
-			ace->ae_id = cpu_to_be32(xfs_kgid_to_gid(acl_e->e_gid));
+			ace->ae_id = cpu_to_be32(
+					from_kgid(&init_user_ns, acl_e->e_gid));
 			break;
 		default:
 			ace->ae_id = cpu_to_be32(ACL_UNDEFINED_ID);
@@ -120,102 +126,86 @@ xfs_acl_to_disk(struct xfs_acl *aclp, const struct posix_acl *acl)
 struct posix_acl *
 xfs_get_acl(struct inode *inode, int type)
 {
-	struct xfs_inode *ip = XFS_I(inode);
-	struct posix_acl *acl = NULL;
-	struct xfs_acl *xfs_acl = NULL;
-	unsigned char *ea_name;
-	int error;
-	int len;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_mount	*mp = ip->i_mount;
+	struct posix_acl	*acl = NULL;
+	struct xfs_da_args	args = {
+		.dp		= ip,
+		.attr_filter	= XFS_ATTR_ROOT,
+		.valuelen	= XFS_ACL_MAX_SIZE(mp),
+	};
+	int			error;
 
 	trace_xfs_get_acl(ip);
 
 	switch (type) {
 	case ACL_TYPE_ACCESS:
-		ea_name = SGI_ACL_FILE;
+		args.name = SGI_ACL_FILE;
 		break;
 	case ACL_TYPE_DEFAULT:
-		ea_name = SGI_ACL_DEFAULT;
+		args.name = SGI_ACL_DEFAULT;
 		break;
 	default:
 		BUG();
 	}
+	args.namelen = strlen(args.name);
 
 	/*
-	 * If we have a cached ACLs value just return it, not need to
-	 * go out to the disk.
+	 * If the attribute doesn't exist make sure we have a negative cache
+	 * entry, for any other error assume it is transient.
 	 */
-	len = XFS_ACL_MAX_SIZE(ip->i_mount);
-	error = xfs_attr_get(ip, ea_name, strlen(ea_name),
-				(unsigned char **)&xfs_acl, &len,
-				ATTR_ALLOC | ATTR_ROOT);
-	if (error) {
-		/*
-		 * If the attribute doesn't exist make sure we have a negative
-		 * cache entry, for any other error assume it is transient.
-		 */
-		if (error != -ENOATTR)
-			acl = ERR_PTR(error);
-	} else  {
-		acl = xfs_acl_from_disk(ip->i_mount, xfs_acl, len,
-					XFS_ACL_MAX_ENTRIES(ip->i_mount));
-		kmem_free(xfs_acl);
+	error = xfs_attr_get(&args);
+	if (!error) {
+		acl = xfs_acl_from_disk(mp, args.value, args.valuelen,
+					XFS_ACL_MAX_ENTRIES(mp));
+	} else if (error != -ENOATTR) {
+		acl = ERR_PTR(error);
 	}
+
+	kmem_free(args.value);
 	return acl;
 }
 
 int
 __xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type)
 {
-	struct xfs_inode *ip = XFS_I(inode);
-	unsigned char *ea_name;
-	int error;
+	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_da_args	args = {
+		.dp		= ip,
+		.attr_filter	= XFS_ATTR_ROOT,
+	};
+	int			error;
 
 	switch (type) {
 	case ACL_TYPE_ACCESS:
-		ea_name = SGI_ACL_FILE;
+		args.name = SGI_ACL_FILE;
 		break;
 	case ACL_TYPE_DEFAULT:
 		if (!S_ISDIR(inode->i_mode))
 			return acl ? -EACCES : 0;
-		ea_name = SGI_ACL_DEFAULT;
+		args.name = SGI_ACL_DEFAULT;
 		break;
 	default:
 		return -EINVAL;
 	}
+	args.namelen = strlen(args.name);
 
 	if (acl) {
-		struct xfs_acl *xfs_acl;
-		int len = XFS_ACL_MAX_SIZE(ip->i_mount);
-
-		xfs_acl = kmem_zalloc_large(len, 0);
-		if (!xfs_acl)
+		args.valuelen = XFS_ACL_SIZE(acl->a_count);
+		args.value = kmem_zalloc_large(args.valuelen, 0);
+		if (!args.value)
 			return -ENOMEM;
-
-		xfs_acl_to_disk(xfs_acl, acl);
-
-		/* subtract away the unused acl entries */
-		len -= sizeof(struct xfs_acl_entry) *
-			 (XFS_ACL_MAX_ENTRIES(ip->i_mount) - acl->a_count);
-
-		error = xfs_attr_set(ip, ea_name, strlen(ea_name),
-				     (unsigned char *)xfs_acl, len, ATTR_ROOT);
-
-		kmem_free(xfs_acl);
-	} else {
-		/*
-		 * A NULL ACL argument means we want to remove the ACL.
-		 */
-		error = xfs_attr_remove(ip, ea_name,
-					strlen(ea_name),
-					ATTR_ROOT);
-
-		/*
-		 * If the attribute didn't exist to start with that's fine.
-		 */
-		if (error == -ENOATTR)
-			error = 0;
+		xfs_acl_to_disk(args.value, acl);
 	}
 
+	error = xfs_attr_set(&args);
+	kmem_free(args.value);
+
+	/*
+	 * If the attribute didn't exist to start with that's fine.
+	 */
+	if (!acl && error == -ENOATTR)
+		error = 0;
 	if (!error)
 		set_cached_acl(inode, type, acl);
 	return error;
@@ -275,3 +265,19 @@ xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type)
 
 	return error;
 }
+
+/*
+ * Invalidate any cached ACLs if the user has bypassed the ACL interface.
+ * We don't validate the content whatsoever so it is caller responsibility to
+ * provide data in valid format and ensure i_mode is consistent.
+ */
+void
+xfs_forget_acl(
+	struct inode		*inode,
+	const char		*name)
+{
+	if (!strcmp(name, SGI_ACL_FILE))
+		forget_cached_acl(inode, ACL_TYPE_ACCESS);
+	else if (!strcmp(name, SGI_ACL_DEFAULT))
+		forget_cached_acl(inode, ACL_TYPE_DEFAULT);
+}
diff --git a/fs/xfs/xfs_acl.h b/fs/xfs/xfs_acl.h
index 94615e34bc86..c042c0868016 100644
--- a/fs/xfs/xfs_acl.h
+++ b/fs/xfs/xfs_acl.h
@@ -13,14 +13,16 @@ struct posix_acl;
 extern struct posix_acl *xfs_get_acl(struct inode *inode, int type);
 extern int xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type);
 extern int __xfs_set_acl(struct inode *inode, struct posix_acl *acl, int type);
+void xfs_forget_acl(struct inode *inode, const char *name);
 #else
 static inline struct posix_acl *xfs_get_acl(struct inode *inode, int type)
 {
 	return NULL;
 }
 # define xfs_set_acl					NULL
+static inline void xfs_forget_acl(struct inode *inode, const char *name)
+{
+}
 #endif /* CONFIG_XFS_POSIX_ACL */
 
-extern void xfs_forget_acl(struct inode *inode, const char *name, int xflags);
-
 #endif	/* __XFS_ACL_H__ */
diff --git a/fs/xfs/xfs_aops.c b/fs/xfs/xfs_aops.c
index 58e937be24ce..b35611882ff9 100644
--- a/fs/xfs/xfs_aops.c
+++ b/fs/xfs/xfs_aops.c
@@ -382,7 +382,7 @@ xfs_map_blocks(
 	 */
 retry:
 	xfs_ilock(ip, XFS_ILOCK_SHARED);
-	ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
+	ASSERT(ip->i_df.if_format != XFS_DINODE_FMT_BTREE ||
 	       (ip->i_df.if_flags & XFS_IFEXTENTS));
 
 	/*
@@ -539,7 +539,7 @@ xfs_discard_page(
 	if (XFS_FORCED_SHUTDOWN(mp))
 		goto out_invalidate;
 
-	xfs_alert(mp,
+	xfs_alert_ratelimited(mp,
 		"page discard on page "PTR_FMT", inode 0x%llx, offset %llu.",
 			page, ip->i_ino, offset);
 
@@ -621,14 +621,11 @@ xfs_vm_readpage(
 	return iomap_readpage(page, &xfs_read_iomap_ops);
 }
 
-STATIC int
-xfs_vm_readpages(
-	struct file		*unused,
-	struct address_space	*mapping,
-	struct list_head	*pages,
-	unsigned		nr_pages)
+STATIC void
+xfs_vm_readahead(
+	struct readahead_control	*rac)
 {
-	return iomap_readpages(mapping, pages, nr_pages, &xfs_read_iomap_ops);
+	iomap_readahead(rac, &xfs_read_iomap_ops);
 }
 
 static int
@@ -644,7 +641,7 @@ xfs_iomap_swapfile_activate(
 
 const struct address_space_operations xfs_address_space_operations = {
 	.readpage		= xfs_vm_readpage,
-	.readpages		= xfs_vm_readpages,
+	.readahead		= xfs_vm_readahead,
 	.writepage		= xfs_vm_writepage,
 	.writepages		= xfs_vm_writepages,
 	.set_page_dirty		= iomap_set_page_dirty,
diff --git a/fs/xfs/xfs_attr_inactive.c b/fs/xfs/xfs_attr_inactive.c
index bbfa6ba84dcd..bfad669e6b2f 100644
--- a/fs/xfs/xfs_attr_inactive.c
+++ b/fs/xfs/xfs_attr_inactive.c
@@ -145,8 +145,8 @@ xfs_attr3_node_inactive(
 	 * Since this code is recursive (gasp!) we must protect ourselves.
 	 */
 	if (level > XFS_DA_NODE_MAXDEPTH) {
+		xfs_buf_mark_corrupt(bp);
 		xfs_trans_brelse(*trans, bp);	/* no locks for later trans */
-		xfs_buf_corruption_error(bp);
 		return -EFSCORRUPTED;
 	}
 
@@ -194,7 +194,7 @@ xfs_attr3_node_inactive(
 			error = xfs_attr3_leaf_inactive(trans, dp, child_bp);
 			break;
 		default:
-			xfs_buf_corruption_error(child_bp);
+			xfs_buf_mark_corrupt(child_bp);
 			xfs_trans_brelse(*trans, child_bp);
 			error = -EFSCORRUPTED;
 			break;
@@ -289,7 +289,7 @@ xfs_attr3_root_inactive(
 		break;
 	default:
 		error = -EFSCORRUPTED;
-		xfs_buf_corruption_error(bp);
+		xfs_buf_mark_corrupt(bp);
 		xfs_trans_brelse(*trans, bp);
 		break;
 	}
@@ -367,7 +367,7 @@ xfs_attr_inactive(
 	 * removal below.
 	 */
 	if (xfs_inode_hasattr(dp) &&
-	    dp->i_d.di_aformat != XFS_DINODE_FMT_LOCAL) {
+	    dp->i_afp->if_format != XFS_DINODE_FMT_LOCAL) {
 		error = xfs_attr3_root_inactive(&trans, dp);
 		if (error)
 			goto out_cancel;
@@ -388,8 +388,11 @@ out_cancel:
 	xfs_trans_cancel(trans);
 out_destroy_fork:
 	/* kill the in-core attr fork before we drop the inode lock */
-	if (dp->i_afp)
-		xfs_idestroy_fork(dp, XFS_ATTR_FORK);
+	if (dp->i_afp) {
+		xfs_idestroy_fork(dp->i_afp);
+		kmem_cache_free(xfs_ifork_zone, dp->i_afp);
+		dp->i_afp = NULL;
+	}
 	if (lock_mode)
 		xfs_iunlock(dp, lock_mode);
 	return error;
diff --git a/fs/xfs/xfs_attr_list.c b/fs/xfs/xfs_attr_list.c
index d37743bdf274..50f922cad91a 100644
--- a/fs/xfs/xfs_attr_list.c
+++ b/fs/xfs/xfs_attr_list.c
@@ -44,7 +44,7 @@ xfs_attr_shortform_compare(const void *a, const void *b)
 /*
  * Copy out entries of shortform attribute lists for attr_list().
  * Shortform attribute lists are not stored in hashval sorted order.
- * If the output buffer is not large enough to hold them all, then we
+ * If the output buffer is not large enough to hold them all, then
  * we have to calculate each entries' hashvalue and sort them before
  * we can begin returning them to the user.
  */
@@ -52,24 +52,19 @@ static int
 xfs_attr_shortform_list(
 	struct xfs_attr_list_context	*context)
 {
-	struct attrlist_cursor_kern	*cursor;
+	struct xfs_attrlist_cursor_kern	*cursor = &context->cursor;
+	struct xfs_inode		*dp = context->dp;
 	struct xfs_attr_sf_sort		*sbuf, *sbp;
 	struct xfs_attr_shortform	*sf;
 	struct xfs_attr_sf_entry	*sfe;
-	struct xfs_inode		*dp;
 	int				sbsize, nsbuf, count, i;
 	int				error = 0;
 
-	ASSERT(context != NULL);
-	dp = context->dp;
-	ASSERT(dp != NULL);
 	ASSERT(dp->i_afp != NULL);
 	sf = (xfs_attr_shortform_t *)dp->i_afp->if_u1.if_data;
 	ASSERT(sf != NULL);
 	if (!sf->hdr.count)
 		return 0;
-	cursor = context->cursor;
-	ASSERT(cursor != NULL);
 
 	trace_xfs_attr_list_sf(context);
 
@@ -205,7 +200,7 @@ out:
 STATIC int
 xfs_attr_node_list_lookup(
 	struct xfs_attr_list_context	*context,
-	struct attrlist_cursor_kern	*cursor,
+	struct xfs_attrlist_cursor_kern	*cursor,
 	struct xfs_buf			**pbp)
 {
 	struct xfs_da3_icnode_hdr	nodehdr;
@@ -279,7 +274,7 @@ xfs_attr_node_list_lookup(
 	return 0;
 
 out_corruptbuf:
-	xfs_buf_corruption_error(bp);
+	xfs_buf_mark_corrupt(bp);
 	xfs_trans_brelse(tp, bp);
 	return -EFSCORRUPTED;
 }
@@ -288,8 +283,8 @@ STATIC int
 xfs_attr_node_list(
 	struct xfs_attr_list_context	*context)
 {
+	struct xfs_attrlist_cursor_kern	*cursor = &context->cursor;
 	struct xfs_attr3_icleaf_hdr	leafhdr;
-	struct attrlist_cursor_kern	*cursor;
 	struct xfs_attr_leafblock	*leaf;
 	struct xfs_da_intnode		*node;
 	struct xfs_buf			*bp;
@@ -299,7 +294,6 @@ xfs_attr_node_list(
 
 	trace_xfs_attr_node_list(context);
 
-	cursor = context->cursor;
 	cursor->initted = 1;
 
 	/*
@@ -394,7 +388,7 @@ xfs_attr3_leaf_list_int(
 	struct xfs_buf			*bp,
 	struct xfs_attr_list_context	*context)
 {
-	struct attrlist_cursor_kern	*cursor;
+	struct xfs_attrlist_cursor_kern	*cursor = &context->cursor;
 	struct xfs_attr_leafblock	*leaf;
 	struct xfs_attr3_icleaf_hdr	ichdr;
 	struct xfs_attr_leaf_entry	*entries;
@@ -408,7 +402,6 @@ xfs_attr3_leaf_list_int(
 	xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &ichdr, leaf);
 	entries = xfs_attr3_leaf_entryp(leaf);
 
-	cursor = context->cursor;
 	cursor->initted = 1;
 
 	/*
@@ -452,8 +445,8 @@ xfs_attr3_leaf_list_int(
 		}
 
 		if ((entry->flags & XFS_ATTR_INCOMPLETE) &&
-		    !(context->flags & ATTR_INCOMPLETE))
-			continue;		/* skip incomplete entries */
+		    !context->allow_incomplete)
+			continue;
 
 		if (entry->flags & XFS_ATTR_LOCAL) {
 			xfs_attr_leaf_name_local_t *name_loc;
@@ -488,14 +481,15 @@ xfs_attr3_leaf_list_int(
  * Copy out attribute entries for attr_list(), for leaf attribute lists.
  */
 STATIC int
-xfs_attr_leaf_list(xfs_attr_list_context_t *context)
+xfs_attr_leaf_list(
+	struct xfs_attr_list_context	*context)
 {
-	int error;
-	struct xfs_buf *bp;
+	struct xfs_buf			*bp;
+	int				error;
 
 	trace_xfs_attr_leaf_list(context);
 
-	context->cursor->blkno = 0;
+	context->cursor.blkno = 0;
 	error = xfs_attr3_leaf_read(context->tp, context->dp, 0, &bp);
 	if (error)
 		return error;
@@ -506,7 +500,7 @@ xfs_attr_leaf_list(xfs_attr_list_context_t *context)
 }
 
 int
-xfs_attr_list_int_ilocked(
+xfs_attr_list_ilocked(
 	struct xfs_attr_list_context	*context)
 {
 	struct xfs_inode		*dp = context->dp;
@@ -518,20 +512,20 @@ xfs_attr_list_int_ilocked(
 	 */
 	if (!xfs_inode_hasattr(dp))
 		return 0;
-	else if (dp->i_d.di_aformat == XFS_DINODE_FMT_LOCAL)
+	if (dp->i_afp->if_format == XFS_DINODE_FMT_LOCAL)
 		return xfs_attr_shortform_list(context);
-	else if (xfs_bmap_one_block(dp, XFS_ATTR_FORK))
+	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK))
 		return xfs_attr_leaf_list(context);
 	return xfs_attr_node_list(context);
 }
 
 int
-xfs_attr_list_int(
-	xfs_attr_list_context_t *context)
+xfs_attr_list(
+	struct xfs_attr_list_context	*context)
 {
-	int error;
-	xfs_inode_t *dp = context->dp;
-	uint		lock_mode;
+	struct xfs_inode		*dp = context->dp;
+	uint				lock_mode;
+	int				error;
 
 	XFS_STATS_INC(dp->i_mount, xs_attr_list);
 
@@ -539,130 +533,7 @@ xfs_attr_list_int(
 		return -EIO;
 
 	lock_mode = xfs_ilock_attr_map_shared(dp);
-	error = xfs_attr_list_int_ilocked(context);
+	error = xfs_attr_list_ilocked(context);
 	xfs_iunlock(dp, lock_mode);
 	return error;
 }
-
-#define	ATTR_ENTBASESIZE		/* minimum bytes used by an attr */ \
-	(((struct attrlist_ent *) 0)->a_name - (char *) 0)
-#define	ATTR_ENTSIZE(namelen)		/* actual bytes used by an attr */ \
-	((ATTR_ENTBASESIZE + (namelen) + 1 + sizeof(uint32_t)-1) \
-	 & ~(sizeof(uint32_t)-1))
-
-/*
- * Format an attribute and copy it out to the user's buffer.
- * Take care to check values and protect against them changing later,
- * we may be reading them directly out of a user buffer.
- */
-STATIC void
-xfs_attr_put_listent(
-	xfs_attr_list_context_t *context,
-	int		flags,
-	unsigned char	*name,
-	int		namelen,
-	int		valuelen)
-{
-	struct attrlist *alist = (struct attrlist *)context->alist;
-	attrlist_ent_t *aep;
-	int arraytop;
-
-	ASSERT(!context->seen_enough);
-	ASSERT(!(context->flags & ATTR_KERNOVAL));
-	ASSERT(context->count >= 0);
-	ASSERT(context->count < (ATTR_MAX_VALUELEN/8));
-	ASSERT(context->firstu >= sizeof(*alist));
-	ASSERT(context->firstu <= context->bufsize);
-
-	/*
-	 * Only list entries in the right namespace.
-	 */
-	if (((context->flags & ATTR_SECURE) == 0) !=
-	    ((flags & XFS_ATTR_SECURE) == 0))
-		return;
-	if (((context->flags & ATTR_ROOT) == 0) !=
-	    ((flags & XFS_ATTR_ROOT) == 0))
-		return;
-
-	arraytop = sizeof(*alist) +
-			context->count * sizeof(alist->al_offset[0]);
-	context->firstu -= ATTR_ENTSIZE(namelen);
-	if (context->firstu < arraytop) {
-		trace_xfs_attr_list_full(context);
-		alist->al_more = 1;
-		context->seen_enough = 1;
-		return;
-	}
-
-	aep = (attrlist_ent_t *)&context->alist[context->firstu];
-	aep->a_valuelen = valuelen;
-	memcpy(aep->a_name, name, namelen);
-	aep->a_name[namelen] = 0;
-	alist->al_offset[context->count++] = context->firstu;
-	alist->al_count = context->count;
-	trace_xfs_attr_list_add(context);
-	return;
-}
-
-/*
- * Generate a list of extended attribute names and optionally
- * also value lengths.  Positive return value follows the XFS
- * convention of being an error, zero or negative return code
- * is the length of the buffer returned (negated), indicating
- * success.
- */
-int
-xfs_attr_list(
-	xfs_inode_t	*dp,
-	char		*buffer,
-	int		bufsize,
-	int		flags,
-	attrlist_cursor_kern_t *cursor)
-{
-	xfs_attr_list_context_t context;
-	struct attrlist *alist;
-	int error;
-
-	/*
-	 * Validate the cursor.
-	 */
-	if (cursor->pad1 || cursor->pad2)
-		return -EINVAL;
-	if ((cursor->initted == 0) &&
-	    (cursor->hashval || cursor->blkno || cursor->offset))
-		return -EINVAL;
-
-	/* Only internal consumers can retrieve incomplete attrs. */
-	if (flags & ATTR_INCOMPLETE)
-		return -EINVAL;
-
-	/*
-	 * Check for a properly aligned buffer.
-	 */
-	if (((long)buffer) & (sizeof(int)-1))
-		return -EFAULT;
-	if (flags & ATTR_KERNOVAL)
-		bufsize = 0;
-
-	/*
-	 * Initialize the output buffer.
-	 */
-	memset(&context, 0, sizeof(context));
-	context.dp = dp;
-	context.cursor = cursor;
-	context.resynch = 1;
-	context.flags = flags;
-	context.alist = buffer;
-	context.bufsize = (bufsize & ~(sizeof(int)-1));  /* align */
-	context.firstu = context.bufsize;
-	context.put_listent = xfs_attr_put_listent;
-
-	alist = (struct attrlist *)context.alist;
-	alist->al_count = 0;
-	alist->al_more = 0;
-	alist->al_offset[0] = context.bufsize;
-
-	error = xfs_attr_list_int(&context);
-	ASSERT(error <= 0);
-	return error;
-}
diff --git a/fs/xfs/xfs_bmap_item.c b/fs/xfs/xfs_bmap_item.c
index ee6f4229cebc..ec3691372e7c 100644
--- a/fs/xfs/xfs_bmap_item.c
+++ b/fs/xfs/xfs_bmap_item.c
@@ -22,16 +22,20 @@
 #include "xfs_bmap_btree.h"
 #include "xfs_trans_space.h"
 #include "xfs_error.h"
+#include "xfs_log_priv.h"
+#include "xfs_log_recover.h"
 
 kmem_zone_t	*xfs_bui_zone;
 kmem_zone_t	*xfs_bud_zone;
 
+static const struct xfs_item_ops xfs_bui_item_ops;
+
 static inline struct xfs_bui_log_item *BUI_ITEM(struct xfs_log_item *lip)
 {
 	return container_of(lip, struct xfs_bui_log_item, bui_item);
 }
 
-void
+STATIC void
 xfs_bui_item_free(
 	struct xfs_bui_log_item	*buip)
 {
@@ -45,13 +49,13 @@ xfs_bui_item_free(
  * committed vs unpin operations in bulk insert operations. Hence the reference
  * count to ensure only the last caller frees the BUI.
  */
-void
+STATIC void
 xfs_bui_release(
 	struct xfs_bui_log_item	*buip)
 {
 	ASSERT(atomic_read(&buip->bui_refcount) > 0);
 	if (atomic_dec_and_test(&buip->bui_refcount)) {
-		xfs_trans_ail_remove(&buip->bui_item, SHUTDOWN_LOG_IO_ERROR);
+		xfs_trans_ail_delete(&buip->bui_item, SHUTDOWN_LOG_IO_ERROR);
 		xfs_bui_item_free(buip);
 	}
 }
@@ -124,24 +128,17 @@ xfs_bui_item_release(
 	xfs_bui_release(BUI_ITEM(lip));
 }
 
-static const struct xfs_item_ops xfs_bui_item_ops = {
-	.iop_size	= xfs_bui_item_size,
-	.iop_format	= xfs_bui_item_format,
-	.iop_unpin	= xfs_bui_item_unpin,
-	.iop_release	= xfs_bui_item_release,
-};
-
 /*
  * Allocate and initialize an bui item with the given number of extents.
  */
-struct xfs_bui_log_item *
+STATIC struct xfs_bui_log_item *
 xfs_bui_init(
 	struct xfs_mount		*mp)
 
 {
 	struct xfs_bui_log_item		*buip;
 
-	buip = kmem_zone_zalloc(xfs_bui_zone, 0);
+	buip = kmem_cache_zalloc(xfs_bui_zone, GFP_KERNEL | __GFP_NOFAIL);
 
 	xfs_log_item_init(mp, &buip->bui_item, XFS_LI_BUI, &xfs_bui_item_ops);
 	buip->bui_format.bui_nextents = XFS_BUI_MAX_FAST_EXTENTS;
@@ -218,7 +215,7 @@ xfs_trans_get_bud(
 {
 	struct xfs_bud_log_item		*budp;
 
-	budp = kmem_zone_zalloc(xfs_bud_zone, 0);
+	budp = kmem_cache_zalloc(xfs_bud_zone, GFP_KERNEL | __GFP_NOFAIL);
 	xfs_log_item_init(tp->t_mountp, &budp->bud_item, XFS_LI_BUD,
 			  &xfs_bud_item_ops);
 	budp->bud_buip = buip;
@@ -278,27 +275,6 @@ xfs_bmap_update_diff_items(
 	return ba->bi_owner->i_ino - bb->bi_owner->i_ino;
 }
 
-/* Get an BUI. */
-STATIC void *
-xfs_bmap_update_create_intent(
-	struct xfs_trans		*tp,
-	unsigned int			count)
-{
-	struct xfs_bui_log_item		*buip;
-
-	ASSERT(count == XFS_BUI_MAX_FAST_EXTENTS);
-	ASSERT(tp != NULL);
-
-	buip = xfs_bui_init(tp->t_mountp);
-	ASSERT(buip != NULL);
-
-	/*
-	 * Get a log_item_desc to point at the new item.
-	 */
-	xfs_trans_add_item(tp, &buip->bui_item);
-	return buip;
-}
-
 /* Set the map extent flags for this mapping. */
 static void
 xfs_trans_set_bmap_flags(
@@ -326,16 +302,12 @@ xfs_trans_set_bmap_flags(
 STATIC void
 xfs_bmap_update_log_item(
 	struct xfs_trans		*tp,
-	void				*intent,
-	struct list_head		*item)
+	struct xfs_bui_log_item		*buip,
+	struct xfs_bmap_intent		*bmap)
 {
-	struct xfs_bui_log_item		*buip = intent;
-	struct xfs_bmap_intent		*bmap;
 	uint				next_extent;
 	struct xfs_map_extent		*map;
 
-	bmap = container_of(item, struct xfs_bmap_intent, bi_list);
-
 	tp->t_flags |= XFS_TRANS_DIRTY;
 	set_bit(XFS_LI_DIRTY, &buip->bui_item.li_flags);
 
@@ -355,23 +327,44 @@ xfs_bmap_update_log_item(
 			bmap->bi_bmap.br_state);
 }
 
+static struct xfs_log_item *
+xfs_bmap_update_create_intent(
+	struct xfs_trans		*tp,
+	struct list_head		*items,
+	unsigned int			count,
+	bool				sort)
+{
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_bui_log_item		*buip = xfs_bui_init(mp);
+	struct xfs_bmap_intent		*bmap;
+
+	ASSERT(count == XFS_BUI_MAX_FAST_EXTENTS);
+
+	xfs_trans_add_item(tp, &buip->bui_item);
+	if (sort)
+		list_sort(mp, items, xfs_bmap_update_diff_items);
+	list_for_each_entry(bmap, items, bi_list)
+		xfs_bmap_update_log_item(tp, buip, bmap);
+	return &buip->bui_item;
+}
+
 /* Get an BUD so we can process all the deferred rmap updates. */
-STATIC void *
+static struct xfs_log_item *
 xfs_bmap_update_create_done(
 	struct xfs_trans		*tp,
-	void				*intent,
+	struct xfs_log_item		*intent,
 	unsigned int			count)
 {
-	return xfs_trans_get_bud(tp, intent);
+	return &xfs_trans_get_bud(tp, BUI_ITEM(intent))->bud_item;
 }
 
 /* Process a deferred rmap update. */
 STATIC int
 xfs_bmap_update_finish_item(
 	struct xfs_trans		*tp,
+	struct xfs_log_item		*done,
 	struct list_head		*item,
-	void				*done_item,
-	void				**state)
+	struct xfs_btree_cur		**state)
 {
 	struct xfs_bmap_intent		*bmap;
 	xfs_filblks_t			count;
@@ -379,7 +372,7 @@ xfs_bmap_update_finish_item(
 
 	bmap = container_of(item, struct xfs_bmap_intent, bi_list);
 	count = bmap->bi_bmap.br_blockcount;
-	error = xfs_trans_log_finish_bmap_update(tp, done_item,
+	error = xfs_trans_log_finish_bmap_update(tp, BUD_ITEM(done),
 			bmap->bi_type,
 			bmap->bi_owner, bmap->bi_whichfork,
 			bmap->bi_bmap.br_startoff,
@@ -398,9 +391,9 @@ xfs_bmap_update_finish_item(
 /* Abort all pending BUIs. */
 STATIC void
 xfs_bmap_update_abort_intent(
-	void				*intent)
+	struct xfs_log_item		*intent)
 {
-	xfs_bui_release(intent);
+	xfs_bui_release(BUI_ITEM(intent));
 }
 
 /* Cancel a deferred rmap update. */
@@ -416,10 +409,8 @@ xfs_bmap_update_cancel_item(
 
 const struct xfs_defer_op_type xfs_bmap_update_defer_type = {
 	.max_items	= XFS_BUI_MAX_FAST_EXTENTS,
-	.diff_items	= xfs_bmap_update_diff_items,
 	.create_intent	= xfs_bmap_update_create_intent,
 	.abort_intent	= xfs_bmap_update_abort_intent,
-	.log_item	= xfs_bmap_update_log_item,
 	.create_done	= xfs_bmap_update_create_done,
 	.finish_item	= xfs_bmap_update_finish_item,
 	.cancel_item	= xfs_bmap_update_cancel_item,
@@ -429,32 +420,30 @@ const struct xfs_defer_op_type xfs_bmap_update_defer_type = {
  * Process a bmap update intent item that was recovered from the log.
  * We need to update some inode's bmbt.
  */
-int
-xfs_bui_recover(
-	struct xfs_trans		*parent_tp,
-	struct xfs_bui_log_item		*buip)
+STATIC int
+xfs_bui_item_recover(
+	struct xfs_log_item		*lip,
+	struct xfs_trans		*parent_tp)
 {
-	int				error = 0;
-	unsigned int			bui_type;
+	struct xfs_bmbt_irec		irec;
+	struct xfs_bui_log_item		*buip = BUI_ITEM(lip);
+	struct xfs_trans		*tp;
+	struct xfs_inode		*ip = NULL;
+	struct xfs_mount		*mp = parent_tp->t_mountp;
 	struct xfs_map_extent		*bmap;
+	struct xfs_bud_log_item		*budp;
 	xfs_fsblock_t			startblock_fsb;
 	xfs_fsblock_t			inode_fsb;
 	xfs_filblks_t			count;
-	bool				op_ok;
-	struct xfs_bud_log_item		*budp;
+	xfs_exntst_t			state;
 	enum xfs_bmap_intent_type	type;
+	bool				op_ok;
+	unsigned int			bui_type;
 	int				whichfork;
-	xfs_exntst_t			state;
-	struct xfs_trans		*tp;
-	struct xfs_inode		*ip = NULL;
-	struct xfs_bmbt_irec		irec;
-	struct xfs_mount		*mp = parent_tp->t_mountp;
-
-	ASSERT(!test_bit(XFS_BUI_RECOVERED, &buip->bui_flags));
+	int				error = 0;
 
 	/* Only one mapping operation per BUI... */
 	if (buip->bui_format.bui_nextents != XFS_BUI_MAX_FAST_EXTENTS) {
-		set_bit(XFS_BUI_RECOVERED, &buip->bui_flags);
 		xfs_bui_release(buip);
 		return -EFSCORRUPTED;
 	}
@@ -488,7 +477,6 @@ xfs_bui_recover(
 		 * This will pull the BUI from the AIL and
 		 * free the memory associated with it.
 		 */
-		set_bit(XFS_BUI_RECOVERED, &buip->bui_flags);
 		xfs_bui_release(buip);
 		return -EFSCORRUPTED;
 	}
@@ -546,7 +534,6 @@ xfs_bui_recover(
 		xfs_bmap_unmap_extent(tp, ip, &irec);
 	}
 
-	set_bit(XFS_BUI_RECOVERED, &buip->bui_flags);
 	xfs_defer_move(parent_tp, tp);
 	error = xfs_trans_commit(tp);
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
@@ -563,3 +550,121 @@ err_inode:
 	}
 	return error;
 }
+
+STATIC bool
+xfs_bui_item_match(
+	struct xfs_log_item	*lip,
+	uint64_t		intent_id)
+{
+	return BUI_ITEM(lip)->bui_format.bui_id == intent_id;
+}
+
+static const struct xfs_item_ops xfs_bui_item_ops = {
+	.iop_size	= xfs_bui_item_size,
+	.iop_format	= xfs_bui_item_format,
+	.iop_unpin	= xfs_bui_item_unpin,
+	.iop_release	= xfs_bui_item_release,
+	.iop_recover	= xfs_bui_item_recover,
+	.iop_match	= xfs_bui_item_match,
+};
+
+/*
+ * Copy an BUI format buffer from the given buf, and into the destination
+ * BUI format structure.  The BUI/BUD items were designed not to need any
+ * special alignment handling.
+ */
+static int
+xfs_bui_copy_format(
+	struct xfs_log_iovec		*buf,
+	struct xfs_bui_log_format	*dst_bui_fmt)
+{
+	struct xfs_bui_log_format	*src_bui_fmt;
+	uint				len;
+
+	src_bui_fmt = buf->i_addr;
+	len = xfs_bui_log_format_sizeof(src_bui_fmt->bui_nextents);
+
+	if (buf->i_len == len) {
+		memcpy(dst_bui_fmt, src_bui_fmt, len);
+		return 0;
+	}
+	XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, NULL);
+	return -EFSCORRUPTED;
+}
+
+/*
+ * This routine is called to create an in-core extent bmap update
+ * item from the bui format structure which was logged on disk.
+ * It allocates an in-core bui, copies the extents from the format
+ * structure into it, and adds the bui to the AIL with the given
+ * LSN.
+ */
+STATIC int
+xlog_recover_bui_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	int				error;
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_bui_log_item		*buip;
+	struct xfs_bui_log_format	*bui_formatp;
+
+	bui_formatp = item->ri_buf[0].i_addr;
+
+	if (bui_formatp->bui_nextents != XFS_BUI_MAX_FAST_EXTENTS) {
+		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, log->l_mp);
+		return -EFSCORRUPTED;
+	}
+	buip = xfs_bui_init(mp);
+	error = xfs_bui_copy_format(&item->ri_buf[0], &buip->bui_format);
+	if (error) {
+		xfs_bui_item_free(buip);
+		return error;
+	}
+	atomic_set(&buip->bui_next_extent, bui_formatp->bui_nextents);
+	/*
+	 * Insert the intent into the AIL directly and drop one reference so
+	 * that finishing or canceling the work will drop the other.
+	 */
+	xfs_trans_ail_insert(log->l_ailp, &buip->bui_item, lsn);
+	xfs_bui_release(buip);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_bui_item_ops = {
+	.item_type		= XFS_LI_BUI,
+	.commit_pass2		= xlog_recover_bui_commit_pass2,
+};
+
+/*
+ * This routine is called when an BUD format structure is found in a committed
+ * transaction in the log. Its purpose is to cancel the corresponding BUI if it
+ * was still in the log. To do this it searches the AIL for the BUI with an id
+ * equal to that in the BUD format structure. If we find it we drop the BUD
+ * reference, which removes the BUI from the AIL and frees it.
+ */
+STATIC int
+xlog_recover_bud_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_bud_log_format	*bud_formatp;
+
+	bud_formatp = item->ri_buf[0].i_addr;
+	if (item->ri_buf[0].i_len != sizeof(struct xfs_bud_log_format)) {
+		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, log->l_mp);
+		return -EFSCORRUPTED;
+	}
+
+	xlog_recover_release_intent(log, XFS_LI_BUI, bud_formatp->bud_bui_id);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_bud_item_ops = {
+	.item_type		= XFS_LI_BUD,
+	.commit_pass2		= xlog_recover_bud_commit_pass2,
+};
diff --git a/fs/xfs/xfs_bmap_item.h b/fs/xfs/xfs_bmap_item.h
index ad479cc73de8..b9be62f8bd52 100644
--- a/fs/xfs/xfs_bmap_item.h
+++ b/fs/xfs/xfs_bmap_item.h
@@ -33,11 +33,6 @@ struct kmem_zone;
 #define	XFS_BUI_MAX_FAST_EXTENTS	1
 
 /*
- * Define BUI flag bits. Manipulated by set/clear/test_bit operators.
- */
-#define	XFS_BUI_RECOVERED		1
-
-/*
  * This is the "bmap update intent" log item.  It is used to log the fact that
  * some reverse mappings need to change.  It is used in conjunction with the
  * "bmap update done" log item described below.
@@ -49,7 +44,6 @@ struct xfs_bui_log_item {
 	struct xfs_log_item		bui_item;
 	atomic_t			bui_refcount;
 	atomic_t			bui_next_extent;
-	unsigned long			bui_flags;	/* misc flags */
 	struct xfs_bui_log_format	bui_format;
 };
 
@@ -74,9 +68,4 @@ struct xfs_bud_log_item {
 extern struct kmem_zone	*xfs_bui_zone;
 extern struct kmem_zone	*xfs_bud_zone;
 
-struct xfs_bui_log_item *xfs_bui_init(struct xfs_mount *);
-void xfs_bui_item_free(struct xfs_bui_log_item *);
-void xfs_bui_release(struct xfs_bui_log_item *);
-int xfs_bui_recover(struct xfs_trans *parent_tp, struct xfs_bui_log_item *buip);
-
 #endif	/* __XFS_BMAP_ITEM_H__ */
diff --git a/fs/xfs/xfs_bmap_util.c b/fs/xfs/xfs_bmap_util.c
index e62fb5216341..73cafc843cd7 100644
--- a/fs/xfs/xfs_bmap_util.c
+++ b/fs/xfs/xfs_bmap_util.c
@@ -126,7 +126,7 @@ xfs_bmap_rtalloc(
 	 * pick an extent that will space things out in the rt area.
 	 */
 	if (ap->eof && ap->offset == 0) {
-		xfs_rtblock_t uninitialized_var(rtx); /* realtime extent no */
+		xfs_rtblock_t rtx; /* realtime extent no */
 
 		error = xfs_rtpick_extent(mp, ap->tp, ralen, &rtx);
 		if (error)
@@ -223,7 +223,7 @@ xfs_bmap_count_blocks(
 	if (!ifp)
 		return 0;
 
-	switch (XFS_IFORK_FORMAT(ip, whichfork)) {
+	switch (ifp->if_format) {
 	case XFS_DINODE_FMT_BTREE:
 		if (!(ifp->if_flags & XFS_IFEXTENTS)) {
 			error = xfs_iread_extents(tp, ip, whichfork);
@@ -449,7 +449,7 @@ xfs_getbmap(
 		break;
 	}
 
-	switch (XFS_IFORK_FORMAT(ip, whichfork)) {
+	switch (ifp->if_format) {
 	case XFS_DINODE_FMT_EXTENTS:
 	case XFS_DINODE_FMT_BTREE:
 		break;
@@ -1062,7 +1062,6 @@ xfs_collapse_file_space(
 	int			error;
 	xfs_fileoff_t		next_fsb = XFS_B_TO_FSB(mp, offset + len);
 	xfs_fileoff_t		shift_fsb = XFS_B_TO_FSB(mp, len);
-	uint			resblks = XFS_DIOSTRAT_SPACE_RES(mp, 0);
 	bool			done = false;
 
 	ASSERT(xfs_isilocked(ip, XFS_IOLOCK_EXCL));
@@ -1078,32 +1077,34 @@ xfs_collapse_file_space(
 	if (error)
 		return error;
 
-	while (!error && !done) {
-		error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0, 0,
-					&tp);
-		if (error)
-			break;
+	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, 0, 0, 0, &tp);
+	if (error)
+		return error;
 
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-		error = xfs_trans_reserve_quota(tp, mp, ip->i_udquot,
-				ip->i_gdquot, ip->i_pdquot, resblks, 0,
-				XFS_QMOPT_RES_REGBLKS);
-		if (error)
-			goto out_trans_cancel;
-		xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, ip, 0);
 
+	while (!done) {
 		error = xfs_bmap_collapse_extents(tp, ip, &next_fsb, shift_fsb,
 				&done);
 		if (error)
 			goto out_trans_cancel;
+		if (done)
+			break;
 
-		error = xfs_trans_commit(tp);
+		/* finish any deferred frees and roll the transaction */
+		error = xfs_defer_finish(&tp);
+		if (error)
+			goto out_trans_cancel;
 	}
 
+	error = xfs_trans_commit(tp);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	return error;
 
 out_trans_cancel:
 	xfs_trans_cancel(tp);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	return error;
 }
 
@@ -1146,35 +1147,41 @@ xfs_insert_file_space(
 	if (error)
 		return error;
 
+	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write,
+			XFS_DIOSTRAT_SPACE_RES(mp, 0), 0, 0, &tp);
+	if (error)
+		return error;
+
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	xfs_trans_ijoin(tp, ip, 0);
+
 	/*
 	 * The extent shifting code works on extent granularity. So, if stop_fsb
 	 * is not the starting block of extent, we need to split the extent at
 	 * stop_fsb.
 	 */
-	error = xfs_bmap_split_extent(ip, stop_fsb);
+	error = xfs_bmap_split_extent(tp, ip, stop_fsb);
 	if (error)
-		return error;
+		goto out_trans_cancel;
 
-	while (!error && !done) {
-		error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, 0, 0, 0,
-					&tp);
+	do {
+		error = xfs_trans_roll_inode(&tp, ip);
 		if (error)
-			break;
+			goto out_trans_cancel;
 
-		xfs_ilock(ip, XFS_ILOCK_EXCL);
-		xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 		error = xfs_bmap_insert_extents(tp, ip, &next_fsb, shift_fsb,
 				&done, stop_fsb);
 		if (error)
 			goto out_trans_cancel;
+	} while (!done);
 
-		error = xfs_trans_commit(tp);
-	}
-
+	error = xfs_trans_commit(tp);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	return error;
 
 out_trans_cancel:
 	xfs_trans_cancel(tp);
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	return error;
 }
 
@@ -1203,17 +1210,26 @@ xfs_swap_extents_check_format(
 	struct xfs_inode	*ip,	/* target inode */
 	struct xfs_inode	*tip)	/* tmp inode */
 {
+	struct xfs_ifork	*ifp = &ip->i_df;
+	struct xfs_ifork	*tifp = &tip->i_df;
+
+	/* User/group/project quota ids must match if quotas are enforced. */
+	if (XFS_IS_QUOTA_ON(ip->i_mount) &&
+	    (!uid_eq(VFS_I(ip)->i_uid, VFS_I(tip)->i_uid) ||
+	     !gid_eq(VFS_I(ip)->i_gid, VFS_I(tip)->i_gid) ||
+	     ip->i_d.di_projid != tip->i_d.di_projid))
+		return -EINVAL;
 
 	/* Should never get a local format */
-	if (ip->i_d.di_format == XFS_DINODE_FMT_LOCAL ||
-	    tip->i_d.di_format == XFS_DINODE_FMT_LOCAL)
+	if (ifp->if_format == XFS_DINODE_FMT_LOCAL ||
+	    tifp->if_format == XFS_DINODE_FMT_LOCAL)
 		return -EINVAL;
 
 	/*
 	 * if the target inode has less extents that then temporary inode then
 	 * why did userspace call us?
 	 */
-	if (ip->i_d.di_nextents < tip->i_d.di_nextents)
+	if (ifp->if_nextents < tifp->if_nextents)
 		return -EINVAL;
 
 	/*
@@ -1228,20 +1244,18 @@ xfs_swap_extents_check_format(
 	 * form then we will end up with the target inode in the wrong format
 	 * as we already know there are less extents in the temp inode.
 	 */
-	if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
-	    tip->i_d.di_format == XFS_DINODE_FMT_BTREE)
+	if (ifp->if_format == XFS_DINODE_FMT_EXTENTS &&
+	    tifp->if_format == XFS_DINODE_FMT_BTREE)
 		return -EINVAL;
 
 	/* Check temp in extent form to max in target */
-	if (tip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
-	    XFS_IFORK_NEXTENTS(tip, XFS_DATA_FORK) >
-			XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK))
+	if (tifp->if_format == XFS_DINODE_FMT_EXTENTS &&
+	    tifp->if_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK))
 		return -EINVAL;
 
 	/* Check target in extent form to max in temp */
-	if (ip->i_d.di_format == XFS_DINODE_FMT_EXTENTS &&
-	    XFS_IFORK_NEXTENTS(ip, XFS_DATA_FORK) >
-			XFS_IFORK_MAXEXT(tip, XFS_DATA_FORK))
+	if (ifp->if_format == XFS_DINODE_FMT_EXTENTS &&
+	    ifp->if_nextents > XFS_IFORK_MAXEXT(tip, XFS_DATA_FORK))
 		return -EINVAL;
 
 	/*
@@ -1253,22 +1267,20 @@ xfs_swap_extents_check_format(
 	 * (a common defrag case) which will occur when the temp inode is in
 	 * extent format...
 	 */
-	if (tip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
+	if (tifp->if_format == XFS_DINODE_FMT_BTREE) {
 		if (XFS_IFORK_Q(ip) &&
-		    XFS_BMAP_BMDR_SPACE(tip->i_df.if_broot) > XFS_IFORK_BOFF(ip))
+		    XFS_BMAP_BMDR_SPACE(tifp->if_broot) > XFS_IFORK_BOFF(ip))
 			return -EINVAL;
-		if (XFS_IFORK_NEXTENTS(tip, XFS_DATA_FORK) <=
-		    XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK))
+		if (tifp->if_nextents <= XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK))
 			return -EINVAL;
 	}
 
 	/* Reciprocal target->temp btree format checks */
-	if (ip->i_d.di_format == XFS_DINODE_FMT_BTREE) {
+	if (ifp->if_format == XFS_DINODE_FMT_BTREE) {
 		if (XFS_IFORK_Q(tip) &&
 		    XFS_BMAP_BMDR_SPACE(ip->i_df.if_broot) > XFS_IFORK_BOFF(tip))
 			return -EINVAL;
-		if (XFS_IFORK_NEXTENTS(ip, XFS_DATA_FORK) <=
-		    XFS_IFORK_MAXEXT(tip, XFS_DATA_FORK))
+		if (ifp->if_nextents <= XFS_IFORK_MAXEXT(tip, XFS_DATA_FORK))
 			return -EINVAL;
 	}
 
@@ -1420,15 +1432,15 @@ xfs_swap_extent_forks(
 	/*
 	 * Count the number of extended attribute blocks
 	 */
-	if ( ((XFS_IFORK_Q(ip) != 0) && (ip->i_d.di_anextents > 0)) &&
-	     (ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
+	if (XFS_IFORK_Q(ip) && ip->i_afp->if_nextents > 0 &&
+	    ip->i_afp->if_format != XFS_DINODE_FMT_LOCAL) {
 		error = xfs_bmap_count_blocks(tp, ip, XFS_ATTR_FORK, &junk,
 				&aforkblks);
 		if (error)
 			return error;
 	}
-	if ( ((XFS_IFORK_Q(tip) != 0) && (tip->i_d.di_anextents > 0)) &&
-	     (tip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL)) {
+	if (XFS_IFORK_Q(tip) && tip->i_afp->if_nextents > 0 &&
+	    tip->i_afp->if_format != XFS_DINODE_FMT_LOCAL) {
 		error = xfs_bmap_count_blocks(tp, tip, XFS_ATTR_FORK, &junk,
 				&taforkblks);
 		if (error)
@@ -1442,12 +1454,12 @@ xfs_swap_extent_forks(
 	 * event of a crash. Set the owner change log flags now and leave the
 	 * bmbt scan as the last step.
 	 */
-	if (ip->i_d.di_version == 3 &&
-	    ip->i_d.di_format == XFS_DINODE_FMT_BTREE)
-		(*target_log_flags) |= XFS_ILOG_DOWNER;
-	if (tip->i_d.di_version == 3 &&
-	    tip->i_d.di_format == XFS_DINODE_FMT_BTREE)
-		(*src_log_flags) |= XFS_ILOG_DOWNER;
+	if (xfs_sb_version_has_v3inode(&ip->i_mount->m_sb)) {
+		if (ip->i_df.if_format == XFS_DINODE_FMT_BTREE)
+			(*target_log_flags) |= XFS_ILOG_DOWNER;
+		if (tip->i_df.if_format == XFS_DINODE_FMT_BTREE)
+			(*src_log_flags) |= XFS_ILOG_DOWNER;
+	}
 
 	/*
 	 * Swap the data forks of the inodes
@@ -1461,9 +1473,6 @@ xfs_swap_extent_forks(
 	ip->i_d.di_nblocks = tip->i_d.di_nblocks - taforkblks + aforkblks;
 	tip->i_d.di_nblocks = tmp + taforkblks - aforkblks;
 
-	swap(ip->i_d.di_nextents, tip->i_d.di_nextents);
-	swap(ip->i_d.di_format, tip->i_d.di_format);
-
 	/*
 	 * The extents in the source inode could still contain speculative
 	 * preallocation beyond EOF (e.g. the file is open but not modified
@@ -1477,24 +1486,24 @@ xfs_swap_extent_forks(
 	tip->i_delayed_blks = ip->i_delayed_blks;
 	ip->i_delayed_blks = 0;
 
-	switch (ip->i_d.di_format) {
+	switch (ip->i_df.if_format) {
 	case XFS_DINODE_FMT_EXTENTS:
 		(*src_log_flags) |= XFS_ILOG_DEXT;
 		break;
 	case XFS_DINODE_FMT_BTREE:
-		ASSERT(ip->i_d.di_version < 3 ||
+		ASSERT(!xfs_sb_version_has_v3inode(&ip->i_mount->m_sb) ||
 		       (*src_log_flags & XFS_ILOG_DOWNER));
 		(*src_log_flags) |= XFS_ILOG_DBROOT;
 		break;
 	}
 
-	switch (tip->i_d.di_format) {
+	switch (tip->i_df.if_format) {
 	case XFS_DINODE_FMT_EXTENTS:
 		(*target_log_flags) |= XFS_ILOG_DEXT;
 		break;
 	case XFS_DINODE_FMT_BTREE:
 		(*target_log_flags) |= XFS_ILOG_DBROOT;
-		ASSERT(tip->i_d.di_version < 3 ||
+		ASSERT(!xfs_sb_version_has_v3inode(&ip->i_mount->m_sb) ||
 		       (*target_log_flags & XFS_ILOG_DOWNER));
 		break;
 	}
@@ -1558,6 +1567,7 @@ xfs_swap_extents(
 	int			lock_flags;
 	uint64_t		f;
 	int			resblks = 0;
+	unsigned int		flags = 0;
 
 	/*
 	 * Lock the inodes against other IO, page faults and truncate to
@@ -1599,7 +1609,7 @@ xfs_swap_extents(
 	if (xfs_inode_has_cow_data(tip)) {
 		error = xfs_reflink_cancel_cow_range(tip, 0, NULLFILEOFF, true);
 		if (error)
-			return error;
+			goto out_unlock;
 	}
 
 	/*
@@ -1608,9 +1618,9 @@ xfs_swap_extents(
 	 * performed with log redo items!
 	 */
 	if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
-		int		w	= XFS_DATA_FORK;
-		uint32_t	ipnext	= XFS_IFORK_NEXTENTS(ip, w);
-		uint32_t	tipnext	= XFS_IFORK_NEXTENTS(tip, w);
+		int		w = XFS_DATA_FORK;
+		uint32_t	ipnext = ip->i_df.if_nextents;
+		uint32_t	tipnext	= tip->i_df.if_nextents;
 
 		/*
 		 * Conceptually this shouldn't affect the shape of either bmbt,
@@ -1621,17 +1631,16 @@ xfs_swap_extents(
 		resblks +=  XFS_SWAP_RMAP_SPACE_RES(mp, tipnext, w);
 
 		/*
-		 * Handle the corner case where either inode might straddle the
-		 * btree format boundary. If so, the inode could bounce between
-		 * btree <-> extent format on unmap -> remap cycles, freeing and
-		 * allocating a bmapbt block each time.
+		 * If either inode straddles a bmapbt block allocation boundary,
+		 * the rmapbt algorithm triggers repeated allocs and frees as
+		 * extents are remapped. This can exhaust the block reservation
+		 * prematurely and cause shutdown. Return freed blocks to the
+		 * transaction reservation to counter this behavior.
 		 */
-		if (ipnext == (XFS_IFORK_MAXEXT(ip, w) + 1))
-			resblks += XFS_IFORK_MAXEXT(ip, w);
-		if (tipnext == (XFS_IFORK_MAXEXT(tip, w) + 1))
-			resblks += XFS_IFORK_MAXEXT(tip, w);
+		flags |= XFS_TRANS_RES_FDBLKS;
 	}
-	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0, 0, &tp);
+	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0, flags,
+				&tp);
 	if (error)
 		goto out_unlock;
 
@@ -1710,10 +1719,11 @@ xfs_swap_extents(
 
 	/* Swap the cow forks. */
 	if (xfs_sb_version_hasreflink(&mp->m_sb)) {
-		ASSERT(ip->i_cformat == XFS_DINODE_FMT_EXTENTS);
-		ASSERT(tip->i_cformat == XFS_DINODE_FMT_EXTENTS);
+		ASSERT(!ip->i_cowfp ||
+		       ip->i_cowfp->if_format == XFS_DINODE_FMT_EXTENTS);
+		ASSERT(!tip->i_cowfp ||
+		       tip->i_cowfp->if_format == XFS_DINODE_FMT_EXTENTS);
 
-		swap(ip->i_cnextents, tip->i_cnextents);
 		swap(ip->i_cowfp, tip->i_cowfp);
 
 		if (ip->i_cowfp && ip->i_cowfp->if_bytes)
diff --git a/fs/xfs/xfs_buf.c b/fs/xfs/xfs_buf.c
index 217e4f82a44a..d4cdcb6fb2fe 100644
--- a/fs/xfs/xfs_buf.c
+++ b/fs/xfs/xfs_buf.c
@@ -14,6 +14,9 @@
 #include "xfs_mount.h"
 #include "xfs_trace.h"
 #include "xfs_log.h"
+#include "xfs_log_recover.h"
+#include "xfs_trans.h"
+#include "xfs_buf_item.h"
 #include "xfs_errortag.h"
 #include "xfs_error.h"
 
@@ -211,9 +214,7 @@ _xfs_buf_alloc(
 	int			i;
 
 	*bpp = NULL;
-	bp = kmem_zone_zalloc(xfs_buf_zone, KM_NOFS);
-	if (unlikely(!bp))
-		return -ENOMEM;
+	bp = kmem_cache_zalloc(xfs_buf_zone, GFP_NOFS | __GFP_NOFAIL);
 
 	/*
 	 * We don't want certain flags to appear in b_flags unless they are
@@ -327,6 +328,9 @@ xfs_buf_free(
 
 			__free_page(page);
 		}
+		if (current->reclaim_state)
+			current->reclaim_state->reclaimed_slab +=
+							bp->b_page_count;
 	} else if (bp->b_flags & _XBF_KMEM)
 		kmem_free(bp->b_addr);
 	_xfs_buf_free_pages(bp);
@@ -474,7 +478,7 @@ _xfs_buf_map_pages(
 		nofs_flag = memalloc_nofs_save();
 		do {
 			bp->b_addr = vm_map_ram(bp->b_pages, bp->b_page_count,
-						-1, PAGE_KERNEL);
+						-1);
 			if (bp->b_addr)
 				break;
 			vm_unmap_aliases();
@@ -652,7 +656,6 @@ found:
 	 */
 	if (bp->b_flags & XBF_STALE) {
 		ASSERT((bp->b_flags & _XBF_DELWRI_Q) == 0);
-		ASSERT(bp->b_iodone == NULL);
 		bp->b_flags &= _XBF_KMEM | _XBF_PAGES;
 		bp->b_ops = NULL;
 	}
@@ -727,8 +730,9 @@ found:
 	if (!bp->b_addr) {
 		error = _xfs_buf_map_pages(bp, flags);
 		if (unlikely(error)) {
-			xfs_warn(target->bt_mount,
-				"%s: failed to map pagesn", __func__);
+			xfs_warn_ratelimited(target->bt_mount,
+				"%s: failed to map %u pages", __func__,
+				bp->b_page_count);
 			xfs_buf_relse(bp);
 			return error;
 		}
@@ -1187,21 +1191,39 @@ xfs_buf_ioend(
 	if (!bp->b_error && bp->b_io_error)
 		xfs_buf_ioerror(bp, bp->b_io_error);
 
-	/* Only validate buffers that were read without errors */
-	if (read && !bp->b_error && bp->b_ops) {
-		ASSERT(!bp->b_iodone);
-		bp->b_ops->verify_read(bp);
+	if (read) {
+		if (!bp->b_error && bp->b_ops)
+			bp->b_ops->verify_read(bp);
+		if (!bp->b_error)
+			bp->b_flags |= XBF_DONE;
+		xfs_buf_ioend_finish(bp);
+		return;
 	}
 
-	if (!bp->b_error)
+	if (!bp->b_error) {
+		bp->b_flags &= ~XBF_WRITE_FAIL;
 		bp->b_flags |= XBF_DONE;
+	}
 
-	if (bp->b_iodone)
-		(*(bp->b_iodone))(bp);
-	else if (bp->b_flags & XBF_ASYNC)
-		xfs_buf_relse(bp);
-	else
-		complete(&bp->b_iowait);
+	/*
+	 * If this is a log recovery buffer, we aren't doing transactional IO
+	 * yet so we need to let it handle IO completions.
+	 */
+	if (bp->b_flags & _XBF_LOGRECOVERY) {
+		xlog_recover_iodone(bp);
+		return;
+	}
+
+	if (bp->b_flags & _XBF_INODES) {
+		xfs_buf_inode_iodone(bp);
+		return;
+	}
+
+	if (bp->b_flags & _XBF_DQUOTS) {
+		xfs_buf_dquot_iodone(bp);
+		return;
+	}
+	xfs_buf_iodone(bp);
 }
 
 static void
@@ -1238,10 +1260,26 @@ xfs_buf_ioerror_alert(
 	struct xfs_buf		*bp,
 	xfs_failaddr_t		func)
 {
-	xfs_alert(bp->b_mount,
-"metadata I/O error in \"%pS\" at daddr 0x%llx len %d error %d",
-			func, (uint64_t)XFS_BUF_ADDR(bp), bp->b_length,
-			-bp->b_error);
+	xfs_buf_alert_ratelimited(bp, "XFS: metadata IO error",
+		"metadata I/O error in \"%pS\" at daddr 0x%llx len %d error %d",
+				  func, (uint64_t)XFS_BUF_ADDR(bp),
+				  bp->b_length, -bp->b_error);
+}
+
+/*
+ * To simulate an I/O failure, the buffer must be locked and held with at least
+ * three references. The LRU reference is dropped by the stale call. The buf
+ * item reference is dropped via ioend processing. The third reference is owned
+ * by the caller and is dropped on I/O completion if the buffer is XBF_ASYNC.
+ */
+void
+xfs_buf_ioend_fail(
+	struct xfs_buf	*bp)
+{
+	bp->b_flags &= ~XBF_DONE;
+	xfs_buf_stale(bp);
+	xfs_buf_ioerror(bp, -EIO);
+	xfs_buf_ioend(bp);
 }
 
 int
@@ -1254,7 +1292,7 @@ xfs_bwrite(
 
 	bp->b_flags |= XBF_WRITE;
 	bp->b_flags &= ~(XBF_ASYNC | XBF_READ | _XBF_DELWRI_Q |
-			 XBF_WRITE_FAIL | XBF_DONE);
+			 XBF_DONE);
 
 	error = xfs_buf_submit(bp);
 	if (error)
@@ -1268,6 +1306,11 @@ xfs_buf_bio_end_io(
 {
 	struct xfs_buf		*bp = (struct xfs_buf *)bio->bi_private;
 
+	if (!bio->bi_status &&
+	    (bp->b_flags & XBF_WRITE) && (bp->b_flags & XBF_ASYNC) &&
+	    XFS_TEST_ERROR(false, bp->b_mount, XFS_ERRTAG_BUF_IOERROR))
+		bio->bi_status = BLK_STS_IOERR;
+
 	/*
 	 * don't overwrite existing errors - otherwise we can lose errors on
 	 * buffers that require multiple bios to complete.
@@ -1476,10 +1519,7 @@ __xfs_buf_submit(
 
 	/* on shutdown we stale and complete the buffer immediately */
 	if (XFS_FORCED_SHUTDOWN(bp->b_mount)) {
-		xfs_buf_ioerror(bp, -EIO);
-		bp->b_flags &= ~XBF_DONE;
-		xfs_buf_stale(bp);
-		xfs_buf_ioend(bp);
+		xfs_buf_ioend_fail(bp);
 		return -EIO;
 	}
 
@@ -1573,6 +1613,28 @@ xfs_buf_zero(
 }
 
 /*
+ * Log a message about and stale a buffer that a caller has decided is corrupt.
+ *
+ * This function should be called for the kinds of metadata corruption that
+ * cannot be detect from a verifier, such as incorrect inter-block relationship
+ * data.  Do /not/ call this function from a verifier function.
+ *
+ * The buffer must be XBF_DONE prior to the call.  Afterwards, the buffer will
+ * be marked stale, but b_error will not be set.  The caller is responsible for
+ * releasing the buffer or fixing it.
+ */
+void
+__xfs_buf_mark_corrupt(
+	struct xfs_buf		*bp,
+	xfs_failaddr_t		fa)
+{
+	ASSERT(bp->b_flags & XBF_DONE);
+
+	xfs_buf_corruption_error(bp, fa);
+	xfs_buf_stale(bp);
+}
+
+/*
  *	Handling of buffer targets (buftargs).
  */
 
@@ -1616,7 +1678,8 @@ xfs_wait_buftarg(
 	struct xfs_buftarg	*btp)
 {
 	LIST_HEAD(dispose);
-	int loop = 0;
+	int			loop = 0;
+	bool			write_fail = false;
 
 	/*
 	 * First wait on the buftarg I/O count for all in-flight buffers to be
@@ -1644,17 +1707,29 @@ xfs_wait_buftarg(
 			bp = list_first_entry(&dispose, struct xfs_buf, b_lru);
 			list_del_init(&bp->b_lru);
 			if (bp->b_flags & XBF_WRITE_FAIL) {
-				xfs_alert(btp->bt_mount,
+				write_fail = true;
+				xfs_buf_alert_ratelimited(bp,
+					"XFS: Corruption Alert",
 "Corruption Alert: Buffer at daddr 0x%llx had permanent write failures!",
 					(long long)bp->b_bn);
-				xfs_alert(btp->bt_mount,
-"Please run xfs_repair to determine the extent of the problem.");
 			}
 			xfs_buf_rele(bp);
 		}
 		if (loop++ != 0)
 			delay(100);
 	}
+
+	/*
+	 * If one or more failed buffers were freed, that means dirty metadata
+	 * was thrown away. This should only ever happen after I/O completion
+	 * handling has elevated I/O error(s) to permanent failures and shuts
+	 * down the fs.
+	 */
+	if (write_fail) {
+		ASSERT(XFS_FORCED_SHUTDOWN(btp->bt_mount));
+		xfs_alert(btp->bt_mount,
+	      "Please run xfs_repair to determine the extent of the problem.");
+	}
 }
 
 static enum lru_status
@@ -1787,6 +1862,13 @@ xfs_alloc_buftarg(
 	btp->bt_bdev = bdev;
 	btp->bt_daxdev = dax_dev;
 
+	/*
+	 * Buffer IO error rate limiting. Limit it to no more than 10 messages
+	 * per 30 seconds so as to not spam logs too much on repeated errors.
+	 */
+	ratelimit_state_init(&btp->bt_ioerror_rl, 30 * HZ,
+			     DEFAULT_RATELIMIT_BURST);
+
 	if (xfs_setsize_buftarg_early(btp, bdev))
 		goto error_free;
 
@@ -1957,7 +2039,7 @@ xfs_buf_delwri_submit_buffers(
 		 * synchronously. Otherwise, drop the buffer from the delwri
 		 * queue and submit async.
 		 */
-		bp->b_flags &= ~(_XBF_DELWRI_Q | XBF_WRITE_FAIL);
+		bp->b_flags &= ~_XBF_DELWRI_Q;
 		bp->b_flags |= XBF_WRITE;
 		if (wait_list) {
 			bp->b_flags &= ~XBF_ASYNC;
@@ -2091,9 +2173,11 @@ xfs_buf_delwri_pushbuf(
 int __init
 xfs_buf_init(void)
 {
-	xfs_buf_zone = kmem_cache_create("xfs_buf",
-					 sizeof(struct xfs_buf), 0,
-					 SLAB_HWCACHE_ALIGN, NULL);
+	xfs_buf_zone = kmem_cache_create("xfs_buf", sizeof(struct xfs_buf), 0,
+					 SLAB_HWCACHE_ALIGN |
+					 SLAB_RECLAIM_ACCOUNT |
+					 SLAB_MEM_SPREAD,
+					 NULL);
 	if (!xfs_buf_zone)
 		goto out;
 
diff --git a/fs/xfs/xfs_buf.h b/fs/xfs/xfs_buf.h
index d79a1fe5d738..755b652e695a 100644
--- a/fs/xfs/xfs_buf.h
+++ b/fs/xfs/xfs_buf.h
@@ -18,6 +18,7 @@
 /*
  *	Base types
  */
+struct xfs_buf;
 
 #define XFS_BUF_DADDR_NULL	((xfs_daddr_t) (-1LL))
 
@@ -30,15 +31,20 @@
 #define XBF_STALE	 (1 << 6) /* buffer has been staled, do not find it */
 #define XBF_WRITE_FAIL	 (1 << 7) /* async writes have failed on this buffer */
 
-/* flags used only as arguments to access routines */
-#define XBF_TRYLOCK	 (1 << 16)/* lock requested, but do not wait */
-#define XBF_UNMAPPED	 (1 << 17)/* do not map the buffer */
+/* buffer type flags for write callbacks */
+#define _XBF_INODES	 (1 << 16)/* inode buffer */
+#define _XBF_DQUOTS	 (1 << 17)/* dquot buffer */
+#define _XBF_LOGRECOVERY	 (1 << 18)/* log recovery buffer */
 
 /* flags used only internally */
 #define _XBF_PAGES	 (1 << 20)/* backed by refcounted pages */
 #define _XBF_KMEM	 (1 << 21)/* backed by heap memory */
 #define _XBF_DELWRI_Q	 (1 << 22)/* buffer on a delwri queue */
 
+/* flags used only as arguments to access routines */
+#define XBF_TRYLOCK	 (1 << 30)/* lock requested, but do not wait */
+#define XBF_UNMAPPED	 (1 << 31)/* do not map the buffer */
+
 typedef unsigned int xfs_buf_flags_t;
 
 #define XFS_BUF_FLAGS \
@@ -50,12 +56,15 @@ typedef unsigned int xfs_buf_flags_t;
 	{ XBF_DONE,		"DONE" }, \
 	{ XBF_STALE,		"STALE" }, \
 	{ XBF_WRITE_FAIL,	"WRITE_FAIL" }, \
-	{ XBF_TRYLOCK,		"TRYLOCK" },	/* should never be set */\
-	{ XBF_UNMAPPED,		"UNMAPPED" },	/* ditto */\
+	{ _XBF_INODES,		"INODES" }, \
+	{ _XBF_DQUOTS,		"DQUOTS" }, \
+	{ _XBF_LOGRECOVERY,		"LOG_RECOVERY" }, \
 	{ _XBF_PAGES,		"PAGES" }, \
 	{ _XBF_KMEM,		"KMEM" }, \
-	{ _XBF_DELWRI_Q,	"DELWRI_Q" }
-
+	{ _XBF_DELWRI_Q,	"DELWRI_Q" }, \
+	/* The following interface flags should never be set */ \
+	{ XBF_TRYLOCK,		"TRYLOCK" }, \
+	{ XBF_UNMAPPED,		"UNMAPPED" }
 
 /*
  * Internal state flags.
@@ -91,12 +100,9 @@ typedef struct xfs_buftarg {
 	struct list_lru		bt_lru;
 
 	struct percpu_counter	bt_io_count;
+	struct ratelimit_state	bt_ioerror_rl;
 } xfs_buftarg_t;
 
-struct xfs_buf;
-typedef void (*xfs_buf_iodone_t)(struct xfs_buf *);
-
-
 #define XB_PAGES	2
 
 struct xfs_buf_map {
@@ -149,7 +155,6 @@ typedef struct xfs_buf {
 	xfs_buftarg_t		*b_target;	/* buffer target (device) */
 	void			*b_addr;	/* virtual address of buffer */
 	struct work_struct	b_ioend_work;
-	xfs_buf_iodone_t	b_iodone;	/* I/O completion function */
 	struct completion	b_iowait;	/* queue for I/O waiters */
 	struct xfs_buf_log_item	*b_log_item;
 	struct list_head	b_li_list;	/* Log items list head */
@@ -256,13 +261,28 @@ extern void xfs_buf_unlock(xfs_buf_t *);
 #define xfs_buf_islocked(bp) \
 	((bp)->b_sema.count <= 0)
 
+static inline void xfs_buf_relse(xfs_buf_t *bp)
+{
+	xfs_buf_unlock(bp);
+	xfs_buf_rele(bp);
+}
+
 /* Buffer Read and Write Routines */
 extern int xfs_bwrite(struct xfs_buf *bp);
 extern void xfs_buf_ioend(struct xfs_buf *bp);
+static inline void xfs_buf_ioend_finish(struct xfs_buf *bp)
+{
+	if (bp->b_flags & XBF_ASYNC)
+		xfs_buf_relse(bp);
+	else
+		complete(&bp->b_iowait);
+}
+
 extern void __xfs_buf_ioerror(struct xfs_buf *bp, int error,
 		xfs_failaddr_t failaddr);
 #define xfs_buf_ioerror(bp, err) __xfs_buf_ioerror((bp), (err), __this_address)
 extern void xfs_buf_ioerror_alert(struct xfs_buf *bp, xfs_failaddr_t fa);
+void xfs_buf_ioend_fail(struct xfs_buf *);
 
 extern int __xfs_buf_submit(struct xfs_buf *bp, bool);
 static inline int xfs_buf_submit(struct xfs_buf *bp)
@@ -272,6 +292,8 @@ static inline int xfs_buf_submit(struct xfs_buf *bp)
 }
 
 void xfs_buf_zero(struct xfs_buf *bp, size_t boff, size_t bsize);
+void __xfs_buf_mark_corrupt(struct xfs_buf *bp, xfs_failaddr_t fa);
+#define xfs_buf_mark_corrupt(bp) __xfs_buf_mark_corrupt((bp), __this_address)
 
 /* Buffer Utility Routines */
 extern void *xfs_buf_offset(struct xfs_buf *, size_t);
@@ -320,12 +342,6 @@ static inline int xfs_buf_ispinned(struct xfs_buf *bp)
 	return atomic_read(&bp->b_pin_count);
 }
 
-static inline void xfs_buf_relse(xfs_buf_t *bp)
-{
-	xfs_buf_unlock(bp);
-	xfs_buf_rele(bp);
-}
-
 static inline int
 xfs_buf_verify_cksum(struct xfs_buf *bp, unsigned long cksum_offset)
 {
diff --git a/fs/xfs/xfs_buf_item.c b/fs/xfs/xfs_buf_item.c
index 663810e6cd59..408d1b572d3f 100644
--- a/fs/xfs/xfs_buf_item.c
+++ b/fs/xfs/xfs_buf_item.c
@@ -12,8 +12,13 @@
 #include "xfs_bit.h"
 #include "xfs_mount.h"
 #include "xfs_trans.h"
-#include "xfs_buf_item.h"
 #include "xfs_trans_priv.h"
+#include "xfs_buf_item.h"
+#include "xfs_inode.h"
+#include "xfs_inode_item.h"
+#include "xfs_quota.h"
+#include "xfs_dquot_item.h"
+#include "xfs_dquot.h"
 #include "xfs_trace.h"
 #include "xfs_log.h"
 
@@ -25,7 +30,7 @@ static inline struct xfs_buf_log_item *BUF_ITEM(struct xfs_log_item *lip)
 	return container_of(lip, struct xfs_buf_log_item, bli_item);
 }
 
-STATIC void	xfs_buf_do_callbacks(struct xfs_buf *bp);
+static void xfs_buf_item_done(struct xfs_buf *bp);
 
 /* Is this log iovec plausibly large enough to contain the buffer log format? */
 bool
@@ -122,7 +127,7 @@ xfs_buf_item_size_segment(
  * stretch of non-contiguous chunks to be logged.  Contiguous chunks are logged
  * in a single iovec.
  *
- * Discontiguous buffers need a format structure per region that that is being
+ * Discontiguous buffers need a format structure per region that is being
  * logged. This makes the changes in the buffer appear to log recovery as though
  * they came from separate buffers, just like would occur if multiple buffers
  * were used instead of a single discontiguous buffer. This enables
@@ -345,7 +350,7 @@ xfs_buf_item_format(
 	 * occurs during recovery.
 	 */
 	if (bip->bli_flags & XFS_BLI_INODE_BUF) {
-		if (xfs_sb_version_hascrc(&lip->li_mountp->m_sb) ||
+		if (xfs_sb_version_has_v3inode(&lip->li_mountp->m_sb) ||
 		    !((bip->bli_flags & XFS_BLI_INODE_ALLOC_BUF) &&
 		      xfs_log_item_in_current_chkpt(lip)))
 			bip->__bli_format.blf_flags |= XFS_BLF_INODE_BUF;
@@ -410,7 +415,6 @@ xfs_buf_item_unpin(
 {
 	struct xfs_buf_log_item	*bip = BUF_ITEM(lip);
 	xfs_buf_t		*bp = bip->bli_buf;
-	struct xfs_ail		*ailp = lip->li_ailp;
 	int			stale = bip->bli_flags & XFS_BLI_STALE;
 	int			freed;
 
@@ -452,58 +456,33 @@ xfs_buf_item_unpin(
 		}
 
 		/*
-		 * If we get called here because of an IO error, we may
-		 * or may not have the item on the AIL. xfs_trans_ail_delete()
-		 * will take care of that situation.
-		 * xfs_trans_ail_delete() drops the AIL lock.
+		 * If we get called here because of an IO error, we may or may
+		 * not have the item on the AIL. xfs_trans_ail_delete() will
+		 * take care of that situation. xfs_trans_ail_delete() drops
+		 * the AIL lock.
 		 */
 		if (bip->bli_flags & XFS_BLI_STALE_INODE) {
-			xfs_buf_do_callbacks(bp);
-			bp->b_log_item = NULL;
-			list_del_init(&bp->b_li_list);
-			bp->b_iodone = NULL;
+			xfs_buf_item_done(bp);
+			xfs_iflush_done(bp);
+			ASSERT(list_empty(&bp->b_li_list));
 		} else {
-			spin_lock(&ailp->ail_lock);
-			xfs_trans_ail_delete(ailp, lip, SHUTDOWN_LOG_IO_ERROR);
+			xfs_trans_ail_delete(lip, SHUTDOWN_LOG_IO_ERROR);
 			xfs_buf_item_relse(bp);
 			ASSERT(bp->b_log_item == NULL);
 		}
 		xfs_buf_relse(bp);
 	} else if (freed && remove) {
 		/*
-		 * There are currently two references to the buffer - the active
-		 * LRU reference and the buf log item. What we are about to do
-		 * here - simulate a failed IO completion - requires 3
-		 * references.
-		 *
-		 * The LRU reference is removed by the xfs_buf_stale() call. The
-		 * buf item reference is removed by the xfs_buf_iodone()
-		 * callback that is run by xfs_buf_do_callbacks() during ioend
-		 * processing (via the bp->b_iodone callback), and then finally
-		 * the ioend processing will drop the IO reference if the buffer
-		 * is marked XBF_ASYNC.
-		 *
-		 * Hence we need to take an additional reference here so that IO
-		 * completion processing doesn't free the buffer prematurely.
+		 * The buffer must be locked and held by the caller to simulate
+		 * an async I/O failure.
 		 */
 		xfs_buf_lock(bp);
 		xfs_buf_hold(bp);
 		bp->b_flags |= XBF_ASYNC;
-		xfs_buf_ioerror(bp, -EIO);
-		bp->b_flags &= ~XBF_DONE;
-		xfs_buf_stale(bp);
-		xfs_buf_ioend(bp);
+		xfs_buf_ioend_fail(bp);
 	}
 }
 
-/*
- * Buffer IO error rate limiting. Limit it to no more than 10 messages per 30
- * seconds so as to not spam logs too much on repeated detection of the same
- * buffer being bad..
- */
-
-static DEFINE_RATELIMIT_STATE(xfs_buf_write_fail_rl_state, 30 * HZ, 10);
-
 STATIC uint
 xfs_buf_item_push(
 	struct xfs_log_item	*lip,
@@ -533,11 +512,10 @@ xfs_buf_item_push(
 	trace_xfs_buf_item_push(bip);
 
 	/* has a previous flush failed due to IO errors? */
-	if ((bp->b_flags & XBF_WRITE_FAIL) &&
-	    ___ratelimit(&xfs_buf_write_fail_rl_state, "XFS: Failing async write")) {
-		xfs_warn(bp->b_mount,
-"Failing async write on buffer block 0x%llx. Retrying async write.",
-			 (long long)bp->b_bn);
+	if (bp->b_flags & XBF_WRITE_FAIL) {
+		xfs_buf_alert_ratelimited(bp, "XFS: Failing async write",
+	    "Failing async write on buffer block 0x%llx. Retrying async write.",
+					  (long long)bp->b_bn);
 	}
 
 	if (!xfs_buf_delwri_queue(bp, buffer_list))
@@ -584,7 +562,7 @@ xfs_buf_item_put(
 	 * state.
 	 */
 	if (aborted)
-		xfs_trans_ail_remove(lip, SHUTDOWN_LOG_IO_ERROR);
+		xfs_trans_ail_delete(lip, 0);
 	xfs_buf_item_relse(bip->bli_buf);
 	return true;
 }
@@ -760,7 +738,7 @@ xfs_buf_item_init(
 		return 0;
 	}
 
-	bip = kmem_zone_zalloc(xfs_buf_item_zone, 0);
+	bip = kmem_cache_zalloc(xfs_buf_item_zone, GFP_KERNEL | __GFP_NOFAIL);
 	xfs_log_item_init(mp, &bip->bli_item, XFS_LI_BUF, &xfs_buf_item_ops);
 	bip->bli_buf = bp;
 
@@ -962,11 +940,7 @@ xfs_buf_item_free(
 }
 
 /*
- * This is called when the buf log item is no longer needed.  It should
- * free the buf log item associated with the given buffer and clear
- * the buffer's pointer to the buf log item.  If there are no more
- * items in the list, clear the b_iodone field of the buffer (see
- * xfs_buf_attach_iodone() below).
+ * xfs_buf_item_relse() is called when the buf log item is no longer needed.
  */
 void
 xfs_buf_item_relse(
@@ -978,137 +952,28 @@ xfs_buf_item_relse(
 	ASSERT(!test_bit(XFS_LI_IN_AIL, &bip->bli_item.li_flags));
 
 	bp->b_log_item = NULL;
-	if (list_empty(&bp->b_li_list))
-		bp->b_iodone = NULL;
-
 	xfs_buf_rele(bp);
 	xfs_buf_item_free(bip);
 }
 
-
-/*
- * Add the given log item with its callback to the list of callbacks
- * to be called when the buffer's I/O completes.  If it is not set
- * already, set the buffer's b_iodone() routine to be
- * xfs_buf_iodone_callbacks() and link the log item into the list of
- * items rooted at b_li_list.
- */
-void
-xfs_buf_attach_iodone(
-	struct xfs_buf		*bp,
-	void			(*cb)(struct xfs_buf *, struct xfs_log_item *),
-	struct xfs_log_item	*lip)
-{
-	ASSERT(xfs_buf_islocked(bp));
-
-	lip->li_cb = cb;
-	list_add_tail(&lip->li_bio_list, &bp->b_li_list);
-
-	ASSERT(bp->b_iodone == NULL ||
-	       bp->b_iodone == xfs_buf_iodone_callbacks);
-	bp->b_iodone = xfs_buf_iodone_callbacks;
-}
-
-/*
- * We can have many callbacks on a buffer. Running the callbacks individually
- * can cause a lot of contention on the AIL lock, so we allow for a single
- * callback to be able to scan the remaining items in bp->b_li_list for other
- * items of the same type and callback to be processed in the first call.
- *
- * As a result, the loop walking the callback list below will also modify the
- * list. it removes the first item from the list and then runs the callback.
- * The loop then restarts from the new first item int the list. This allows the
- * callback to scan and modify the list attached to the buffer and we don't
- * have to care about maintaining a next item pointer.
- */
-STATIC void
-xfs_buf_do_callbacks(
-	struct xfs_buf		*bp)
-{
-	struct xfs_buf_log_item *blip = bp->b_log_item;
-	struct xfs_log_item	*lip;
-
-	/* If there is a buf_log_item attached, run its callback */
-	if (blip) {
-		lip = &blip->bli_item;
-		lip->li_cb(bp, lip);
-	}
-
-	while (!list_empty(&bp->b_li_list)) {
-		lip = list_first_entry(&bp->b_li_list, struct xfs_log_item,
-				       li_bio_list);
-
-		/*
-		 * Remove the item from the list, so we don't have any
-		 * confusion if the item is added to another buf.
-		 * Don't touch the log item after calling its
-		 * callback, because it could have freed itself.
-		 */
-		list_del_init(&lip->li_bio_list);
-		lip->li_cb(bp, lip);
-	}
-}
-
 /*
- * Invoke the error state callback for each log item affected by the failed I/O.
- *
- * If a metadata buffer write fails with a non-permanent error, the buffer is
- * eventually resubmitted and so the completion callbacks are not run. The error
- * state may need to be propagated to the log items attached to the buffer,
- * however, so the next AIL push of the item knows hot to handle it correctly.
+ * Decide if we're going to retry the write after a failure, and prepare
+ * the buffer for retrying the write.
  */
-STATIC void
-xfs_buf_do_callbacks_fail(
-	struct xfs_buf		*bp)
-{
-	struct xfs_log_item	*lip;
-	struct xfs_ail		*ailp;
-
-	/*
-	 * Buffer log item errors are handled directly by xfs_buf_item_push()
-	 * and xfs_buf_iodone_callback_error, and they have no IO error
-	 * callbacks. Check only for items in b_li_list.
-	 */
-	if (list_empty(&bp->b_li_list))
-		return;
-
-	lip = list_first_entry(&bp->b_li_list, struct xfs_log_item,
-			li_bio_list);
-	ailp = lip->li_ailp;
-	spin_lock(&ailp->ail_lock);
-	list_for_each_entry(lip, &bp->b_li_list, li_bio_list) {
-		if (lip->li_ops->iop_error)
-			lip->li_ops->iop_error(lip, bp);
-	}
-	spin_unlock(&ailp->ail_lock);
-}
-
 static bool
-xfs_buf_iodone_callback_error(
+xfs_buf_ioerror_fail_without_retry(
 	struct xfs_buf		*bp)
 {
-	struct xfs_buf_log_item	*bip = bp->b_log_item;
-	struct xfs_log_item	*lip;
-	struct xfs_mount	*mp;
+	struct xfs_mount	*mp = bp->b_mount;
 	static ulong		lasttime;
 	static xfs_buftarg_t	*lasttarg;
-	struct xfs_error_cfg	*cfg;
-
-	/*
-	 * The failed buffer might not have a buf_log_item attached or the
-	 * log_item list might be empty. Get the mp from the available
-	 * xfs_log_item
-	 */
-	lip = list_first_entry_or_null(&bp->b_li_list, struct xfs_log_item,
-				       li_bio_list);
-	mp = lip ? lip->li_mountp : bip->bli_item.li_mountp;
 
 	/*
 	 * If we've already decided to shutdown the filesystem because of
 	 * I/O errors, there's no point in giving this a retry.
 	 */
 	if (XFS_FORCED_SHUTDOWN(mp))
-		goto out_stale;
+		return true;
 
 	if (bp->b_target != lasttarg ||
 	    time_after(jiffies, (lasttime + 5*HZ))) {
@@ -1119,171 +984,240 @@ xfs_buf_iodone_callback_error(
 
 	/* synchronous writes will have callers process the error */
 	if (!(bp->b_flags & XBF_ASYNC))
-		goto out_stale;
-
-	trace_xfs_buf_item_iodone_async(bp, _RET_IP_);
-	ASSERT(bp->b_iodone != NULL);
-
-	cfg = xfs_error_get_cfg(mp, XFS_ERR_METADATA, bp->b_error);
+		return true;
+	return false;
+}
 
-	/*
-	 * If the write was asynchronous then no one will be looking for the
-	 * error.  If this is the first failure of this type, clear the error
-	 * state and write the buffer out again. This means we always retry an
-	 * async write failure at least once, but we also need to set the buffer
-	 * up to behave correctly now for repeated failures.
-	 */
-	if (!(bp->b_flags & (XBF_STALE | XBF_WRITE_FAIL)) ||
-	     bp->b_last_error != bp->b_error) {
-		bp->b_flags |= (XBF_WRITE | XBF_DONE | XBF_WRITE_FAIL);
-		bp->b_last_error = bp->b_error;
-		if (cfg->retry_timeout != XFS_ERR_RETRY_FOREVER &&
-		    !bp->b_first_retry_time)
-			bp->b_first_retry_time = jiffies;
+static bool
+xfs_buf_ioerror_retry(
+	struct xfs_buf		*bp,
+	struct xfs_error_cfg	*cfg)
+{
+	if ((bp->b_flags & (XBF_STALE | XBF_WRITE_FAIL)) &&
+	    bp->b_last_error == bp->b_error)
+		return false;
 
-		xfs_buf_ioerror(bp, 0);
-		xfs_buf_submit(bp);
-		return true;
-	}
+	bp->b_flags |= (XBF_WRITE | XBF_DONE | XBF_WRITE_FAIL);
+	bp->b_last_error = bp->b_error;
+	if (cfg->retry_timeout != XFS_ERR_RETRY_FOREVER &&
+	    !bp->b_first_retry_time)
+		bp->b_first_retry_time = jiffies;
+	return true;
+}
 
-	/*
-	 * Repeated failure on an async write. Take action according to the
-	 * error configuration we have been set up to use.
-	 */
+/*
+ * Account for this latest trip around the retry handler, and decide if
+ * we've failed enough times to constitute a permanent failure.
+ */
+static bool
+xfs_buf_ioerror_permanent(
+	struct xfs_buf		*bp,
+	struct xfs_error_cfg	*cfg)
+{
+	struct xfs_mount	*mp = bp->b_mount;
 
 	if (cfg->max_retries != XFS_ERR_RETRY_FOREVER &&
 	    ++bp->b_retries > cfg->max_retries)
-			goto permanent_error;
+		return true;
 	if (cfg->retry_timeout != XFS_ERR_RETRY_FOREVER &&
 	    time_after(jiffies, cfg->retry_timeout + bp->b_first_retry_time))
-			goto permanent_error;
+		return true;
 
 	/* At unmount we may treat errors differently */
 	if ((mp->m_flags & XFS_MOUNT_UNMOUNTING) && mp->m_fail_unmount)
-		goto permanent_error;
+		return true;
 
-	/*
-	 * Still a transient error, run IO completion failure callbacks and let
-	 * the higher layers retry the buffer.
-	 */
-	xfs_buf_do_callbacks_fail(bp);
-	xfs_buf_ioerror(bp, 0);
-	xfs_buf_relse(bp);
-	return true;
+	return false;
+}
+
+/*
+ * On a sync write or shutdown we just want to stale the buffer and let the
+ * caller handle the error in bp->b_error appropriately.
+ *
+ * If the write was asynchronous then no one will be looking for the error.  If
+ * this is the first failure of this type, clear the error state and write the
+ * buffer out again. This means we always retry an async write failure at least
+ * once, but we also need to set the buffer up to behave correctly now for
+ * repeated failures.
+ *
+ * If we get repeated async write failures, then we take action according to the
+ * error configuration we have been set up to use.
+ *
+ * Multi-state return value:
+ *
+ * XBF_IOERROR_FINISH: clear IO error retry state and run callback completions
+ * XBF_IOERROR_DONE: resubmitted immediately, do not run any completions
+ * XBF_IOERROR_FAIL: transient error, run failure callback completions and then
+ *    release the buffer
+ */
+enum {
+	XBF_IOERROR_FINISH,
+	XBF_IOERROR_DONE,
+	XBF_IOERROR_FAIL,
+};
+
+static int
+xfs_buf_iodone_error(
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+	struct xfs_error_cfg	*cfg;
+
+	if (xfs_buf_ioerror_fail_without_retry(bp))
+		goto out_stale;
+
+	trace_xfs_buf_item_iodone_async(bp, _RET_IP_);
+
+	cfg = xfs_error_get_cfg(mp, XFS_ERR_METADATA, bp->b_error);
+	if (xfs_buf_ioerror_retry(bp, cfg)) {
+		xfs_buf_ioerror(bp, 0);
+		xfs_buf_submit(bp);
+		return XBF_IOERROR_DONE;
+	}
 
 	/*
 	 * Permanent error - we need to trigger a shutdown if we haven't already
 	 * to indicate that inconsistency will result from this action.
 	 */
-permanent_error:
-	xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+	if (xfs_buf_ioerror_permanent(bp, cfg)) {
+		xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
+		goto out_stale;
+	}
+
+	/* Still considered a transient error. Caller will schedule retries. */
+	return XBF_IOERROR_FAIL;
+
 out_stale:
 	xfs_buf_stale(bp);
 	bp->b_flags |= XBF_DONE;
 	trace_xfs_buf_error_relse(bp, _RET_IP_);
-	return false;
+	return XBF_IOERROR_FINISH;
 }
 
-/*
- * This is the iodone() function for buffers which have had callbacks attached
- * to them by xfs_buf_attach_iodone(). We need to iterate the items on the
- * callback list, mark the buffer as having no more callbacks and then push the
- * buffer through IO completion processing.
- */
-void
-xfs_buf_iodone_callbacks(
+static void
+xfs_buf_item_done(
 	struct xfs_buf		*bp)
 {
-	/*
-	 * If there is an error, process it. Some errors require us
-	 * to run callbacks after failure processing is done so we
-	 * detect that and take appropriate action.
-	 */
-	if (bp->b_error && xfs_buf_iodone_callback_error(bp))
+	struct xfs_buf_log_item	*bip = bp->b_log_item;
+
+	if (!bip)
 		return;
 
 	/*
-	 * Successful IO or permanent error. Either way, we can clear the
-	 * retry state here in preparation for the next error that may occur.
+	 * If we are forcibly shutting down, this may well be off the AIL
+	 * already. That's because we simulate the log-committed callbacks to
+	 * unpin these buffers. Or we may never have put this item on AIL
+	 * because of the transaction was aborted forcibly.
+	 * xfs_trans_ail_delete() takes care of these.
+	 *
+	 * Either way, AIL is useless if we're forcing a shutdown.
 	 */
+	xfs_trans_ail_delete(&bip->bli_item, SHUTDOWN_CORRUPT_INCORE);
+	bp->b_log_item = NULL;
+	xfs_buf_item_free(bip);
+	xfs_buf_rele(bp);
+}
+
+static inline void
+xfs_buf_clear_ioerror_retry_state(
+	struct xfs_buf		*bp)
+{
 	bp->b_last_error = 0;
 	bp->b_retries = 0;
 	bp->b_first_retry_time = 0;
-
-	xfs_buf_do_callbacks(bp);
-	bp->b_log_item = NULL;
-	list_del_init(&bp->b_li_list);
-	bp->b_iodone = NULL;
-	xfs_buf_ioend(bp);
 }
 
 /*
- * This is the iodone() function for buffers which have been
- * logged.  It is called when they are eventually flushed out.
- * It should remove the buf item from the AIL, and free the buf item.
- * It is called by xfs_buf_iodone_callbacks() above which will take
- * care of cleaning up the buffer itself.
+ * Inode buffer iodone callback function.
  */
 void
-xfs_buf_iodone(
-	struct xfs_buf		*bp,
-	struct xfs_log_item	*lip)
+xfs_buf_inode_iodone(
+	struct xfs_buf		*bp)
 {
-	struct xfs_ail		*ailp = lip->li_ailp;
+	if (bp->b_error) {
+		struct xfs_log_item *lip;
+		int ret = xfs_buf_iodone_error(bp);
+
+		if (ret == XBF_IOERROR_FINISH)
+			goto finish_iodone;
+		if (ret == XBF_IOERROR_DONE)
+			return;
+		ASSERT(ret == XBF_IOERROR_FAIL);
+		list_for_each_entry(lip, &bp->b_li_list, li_bio_list) {
+			set_bit(XFS_LI_FAILED, &lip->li_flags);
+		}
+		xfs_buf_ioerror(bp, 0);
+		xfs_buf_relse(bp);
+		return;
+	}
 
-	ASSERT(BUF_ITEM(lip)->bli_buf == bp);
+finish_iodone:
+	xfs_buf_clear_ioerror_retry_state(bp);
+	xfs_buf_item_done(bp);
+	xfs_iflush_done(bp);
+	xfs_buf_ioend_finish(bp);
+}
 
-	xfs_buf_rele(bp);
+/*
+ * Dquot buffer iodone callback function.
+ */
+void
+xfs_buf_dquot_iodone(
+	struct xfs_buf		*bp)
+{
+	if (bp->b_error) {
+		struct xfs_log_item *lip;
+		int ret = xfs_buf_iodone_error(bp);
+
+		if (ret == XBF_IOERROR_FINISH)
+			goto finish_iodone;
+		if (ret == XBF_IOERROR_DONE)
+			return;
+		ASSERT(ret == XBF_IOERROR_FAIL);
+		spin_lock(&bp->b_mount->m_ail->ail_lock);
+		list_for_each_entry(lip, &bp->b_li_list, li_bio_list) {
+			xfs_set_li_failed(lip, bp);
+		}
+		spin_unlock(&bp->b_mount->m_ail->ail_lock);
+		xfs_buf_ioerror(bp, 0);
+		xfs_buf_relse(bp);
+		return;
+	}
 
-	/*
-	 * If we are forcibly shutting down, this may well be
-	 * off the AIL already. That's because we simulate the
-	 * log-committed callbacks to unpin these buffers. Or we may never
-	 * have put this item on AIL because of the transaction was
-	 * aborted forcibly. xfs_trans_ail_delete() takes care of these.
-	 *
-	 * Either way, AIL is useless if we're forcing a shutdown.
-	 */
-	spin_lock(&ailp->ail_lock);
-	xfs_trans_ail_delete(ailp, lip, SHUTDOWN_CORRUPT_INCORE);
-	xfs_buf_item_free(BUF_ITEM(lip));
+finish_iodone:
+	xfs_buf_clear_ioerror_retry_state(bp);
+	/* a newly allocated dquot buffer might have a log item attached */
+	xfs_buf_item_done(bp);
+	xfs_dquot_done(bp);
+	xfs_buf_ioend_finish(bp);
 }
 
 /*
- * Requeue a failed buffer for writeback.
+ * Dirty buffer iodone callback function.
  *
- * We clear the log item failed state here as well, but we have to be careful
- * about reference counts because the only active reference counts on the buffer
- * may be the failed log items. Hence if we clear the log item failed state
- * before queuing the buffer for IO we can release all active references to
- * the buffer and free it, leading to use after free problems in
- * xfs_buf_delwri_queue. It makes no difference to the buffer or log items which
- * order we process them in - the buffer is locked, and we own the buffer list
- * so nothing on them is going to change while we are performing this action.
- *
- * Hence we can safely queue the buffer for IO before we clear the failed log
- * item state, therefore  always having an active reference to the buffer and
- * avoiding the transient zero-reference state that leads to use-after-free.
- *
- * Return true if the buffer was added to the buffer list, false if it was
- * already on the buffer list.
+ * Note that for things like remote attribute buffers, there may not be a buffer
+ * log item here, so processing the buffer log item must remain be optional.
  */
-bool
-xfs_buf_resubmit_failed_buffers(
-	struct xfs_buf		*bp,
-	struct list_head	*buffer_list)
+void
+xfs_buf_iodone(
+	struct xfs_buf		*bp)
 {
-	struct xfs_log_item	*lip;
-	bool			ret;
-
-	ret = xfs_buf_delwri_queue(bp, buffer_list);
-
-	/*
-	 * XFS_LI_FAILED set/clear is protected by ail_lock, caller of this
-	 * function already have it acquired
-	 */
-	list_for_each_entry(lip, &bp->b_li_list, li_bio_list)
-		xfs_clear_li_failed(lip);
+	if (bp->b_error) {
+		int ret = xfs_buf_iodone_error(bp);
+
+		if (ret == XBF_IOERROR_FINISH)
+			goto finish_iodone;
+		if (ret == XBF_IOERROR_DONE)
+			return;
+		ASSERT(ret == XBF_IOERROR_FAIL);
+		ASSERT(list_empty(&bp->b_li_list));
+		xfs_buf_ioerror(bp, 0);
+		xfs_buf_relse(bp);
+		return;
+	}
 
-	return ret;
+finish_iodone:
+	xfs_buf_clear_ioerror_retry_state(bp);
+	xfs_buf_item_done(bp);
+	xfs_buf_ioend_finish(bp);
 }
diff --git a/fs/xfs/xfs_buf_item.h b/fs/xfs/xfs_buf_item.h
index 30114b510332..23507cbb4c41 100644
--- a/fs/xfs/xfs_buf_item.h
+++ b/fs/xfs/xfs_buf_item.h
@@ -54,13 +54,9 @@ void	xfs_buf_item_relse(struct xfs_buf *);
 bool	xfs_buf_item_put(struct xfs_buf_log_item *);
 void	xfs_buf_item_log(struct xfs_buf_log_item *, uint, uint);
 bool	xfs_buf_item_dirty_format(struct xfs_buf_log_item *);
-void	xfs_buf_attach_iodone(struct xfs_buf *,
-			      void(*)(struct xfs_buf *, struct xfs_log_item *),
-			      struct xfs_log_item *);
-void	xfs_buf_iodone_callbacks(struct xfs_buf *);
-void	xfs_buf_iodone(struct xfs_buf *, struct xfs_log_item *);
-bool	xfs_buf_resubmit_failed_buffers(struct xfs_buf *,
-					struct list_head *);
+void	xfs_buf_inode_iodone(struct xfs_buf *);
+void	xfs_buf_dquot_iodone(struct xfs_buf *);
+void	xfs_buf_iodone(struct xfs_buf *);
 bool	xfs_buf_log_check_iovec(struct xfs_log_iovec *iovec);
 
 extern kmem_zone_t	*xfs_buf_item_zone;
diff --git a/fs/xfs/xfs_buf_item_recover.c b/fs/xfs/xfs_buf_item_recover.c
new file mode 100644
index 000000000000..8f0457d67d77
--- /dev/null
+++ b/fs/xfs/xfs_buf_item_recover.c
@@ -0,0 +1,982 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_bit.h"
+#include "xfs_mount.h"
+#include "xfs_trans.h"
+#include "xfs_buf_item.h"
+#include "xfs_trans_priv.h"
+#include "xfs_trace.h"
+#include "xfs_log.h"
+#include "xfs_log_priv.h"
+#include "xfs_log_recover.h"
+#include "xfs_error.h"
+#include "xfs_inode.h"
+#include "xfs_dir2.h"
+#include "xfs_quota.h"
+
+/*
+ * This structure is used during recovery to record the buf log items which
+ * have been canceled and should not be replayed.
+ */
+struct xfs_buf_cancel {
+	xfs_daddr_t		bc_blkno;
+	uint			bc_len;
+	int			bc_refcount;
+	struct list_head	bc_list;
+};
+
+static struct xfs_buf_cancel *
+xlog_find_buffer_cancelled(
+	struct xlog		*log,
+	xfs_daddr_t		blkno,
+	uint			len)
+{
+	struct list_head	*bucket;
+	struct xfs_buf_cancel	*bcp;
+
+	if (!log->l_buf_cancel_table)
+		return NULL;
+
+	bucket = XLOG_BUF_CANCEL_BUCKET(log, blkno);
+	list_for_each_entry(bcp, bucket, bc_list) {
+		if (bcp->bc_blkno == blkno && bcp->bc_len == len)
+			return bcp;
+	}
+
+	return NULL;
+}
+
+static bool
+xlog_add_buffer_cancelled(
+	struct xlog		*log,
+	xfs_daddr_t		blkno,
+	uint			len)
+{
+	struct xfs_buf_cancel	*bcp;
+
+	/*
+	 * If we find an existing cancel record, this indicates that the buffer
+	 * was cancelled multiple times.  To ensure that during pass 2 we keep
+	 * the record in the table until we reach its last occurrence in the
+	 * log, a reference count is kept to tell how many times we expect to
+	 * see this record during the second pass.
+	 */
+	bcp = xlog_find_buffer_cancelled(log, blkno, len);
+	if (bcp) {
+		bcp->bc_refcount++;
+		return false;
+	}
+
+	bcp = kmem_alloc(sizeof(struct xfs_buf_cancel), 0);
+	bcp->bc_blkno = blkno;
+	bcp->bc_len = len;
+	bcp->bc_refcount = 1;
+	list_add_tail(&bcp->bc_list, XLOG_BUF_CANCEL_BUCKET(log, blkno));
+	return true;
+}
+
+/*
+ * Check if there is and entry for blkno, len in the buffer cancel record table.
+ */
+bool
+xlog_is_buffer_cancelled(
+	struct xlog		*log,
+	xfs_daddr_t		blkno,
+	uint			len)
+{
+	return xlog_find_buffer_cancelled(log, blkno, len) != NULL;
+}
+
+/*
+ * Check if there is and entry for blkno, len in the buffer cancel record table,
+ * and decremented the reference count on it if there is one.
+ *
+ * Remove the cancel record once the refcount hits zero, so that if the same
+ * buffer is re-used again after its last cancellation we actually replay the
+ * changes made at that point.
+ */
+static bool
+xlog_put_buffer_cancelled(
+	struct xlog		*log,
+	xfs_daddr_t		blkno,
+	uint			len)
+{
+	struct xfs_buf_cancel	*bcp;
+
+	bcp = xlog_find_buffer_cancelled(log, blkno, len);
+	if (!bcp) {
+		ASSERT(0);
+		return false;
+	}
+
+	if (--bcp->bc_refcount == 0) {
+		list_del(&bcp->bc_list);
+		kmem_free(bcp);
+	}
+	return true;
+}
+
+/* log buffer item recovery */
+
+/*
+ * Sort buffer items for log recovery.  Most buffer items should end up on the
+ * buffer list and are recovered first, with the following exceptions:
+ *
+ * 1. XFS_BLF_CANCEL buffers must be processed last because some log items
+ *    might depend on the incor ecancellation record, and replaying a cancelled
+ *    buffer item can remove the incore record.
+ *
+ * 2. XFS_BLF_INODE_BUF buffers are handled after most regular items so that
+ *    we replay di_next_unlinked only after flushing the inode 'free' state
+ *    to the inode buffer.
+ *
+ * See xlog_recover_reorder_trans for more details.
+ */
+STATIC enum xlog_recover_reorder
+xlog_recover_buf_reorder(
+	struct xlog_recover_item	*item)
+{
+	struct xfs_buf_log_format	*buf_f = item->ri_buf[0].i_addr;
+
+	if (buf_f->blf_flags & XFS_BLF_CANCEL)
+		return XLOG_REORDER_CANCEL_LIST;
+	if (buf_f->blf_flags & XFS_BLF_INODE_BUF)
+		return XLOG_REORDER_INODE_BUFFER_LIST;
+	return XLOG_REORDER_BUFFER_LIST;
+}
+
+STATIC void
+xlog_recover_buf_ra_pass2(
+	struct xlog                     *log,
+	struct xlog_recover_item        *item)
+{
+	struct xfs_buf_log_format	*buf_f = item->ri_buf[0].i_addr;
+
+	xlog_buf_readahead(log, buf_f->blf_blkno, buf_f->blf_len, NULL);
+}
+
+/*
+ * Build up the table of buf cancel records so that we don't replay cancelled
+ * data in the second pass.
+ */
+static int
+xlog_recover_buf_commit_pass1(
+	struct xlog			*log,
+	struct xlog_recover_item	*item)
+{
+	struct xfs_buf_log_format	*bf = item->ri_buf[0].i_addr;
+
+	if (!xfs_buf_log_check_iovec(&item->ri_buf[0])) {
+		xfs_err(log->l_mp, "bad buffer log item size (%d)",
+				item->ri_buf[0].i_len);
+		return -EFSCORRUPTED;
+	}
+
+	if (!(bf->blf_flags & XFS_BLF_CANCEL))
+		trace_xfs_log_recover_buf_not_cancel(log, bf);
+	else if (xlog_add_buffer_cancelled(log, bf->blf_blkno, bf->blf_len))
+		trace_xfs_log_recover_buf_cancel_add(log, bf);
+	else
+		trace_xfs_log_recover_buf_cancel_ref_inc(log, bf);
+	return 0;
+}
+
+/*
+ * Validate the recovered buffer is of the correct type and attach the
+ * appropriate buffer operations to them for writeback. Magic numbers are in a
+ * few places:
+ *	the first 16 bits of the buffer (inode buffer, dquot buffer),
+ *	the first 32 bits of the buffer (most blocks),
+ *	inside a struct xfs_da_blkinfo at the start of the buffer.
+ */
+static void
+xlog_recover_validate_buf_type(
+	struct xfs_mount		*mp,
+	struct xfs_buf			*bp,
+	struct xfs_buf_log_format	*buf_f,
+	xfs_lsn_t			current_lsn)
+{
+	struct xfs_da_blkinfo		*info = bp->b_addr;
+	uint32_t			magic32;
+	uint16_t			magic16;
+	uint16_t			magicda;
+	char				*warnmsg = NULL;
+
+	/*
+	 * We can only do post recovery validation on items on CRC enabled
+	 * fielsystems as we need to know when the buffer was written to be able
+	 * to determine if we should have replayed the item. If we replay old
+	 * metadata over a newer buffer, then it will enter a temporarily
+	 * inconsistent state resulting in verification failures. Hence for now
+	 * just avoid the verification stage for non-crc filesystems
+	 */
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return;
+
+	magic32 = be32_to_cpu(*(__be32 *)bp->b_addr);
+	magic16 = be16_to_cpu(*(__be16*)bp->b_addr);
+	magicda = be16_to_cpu(info->magic);
+	switch (xfs_blft_from_flags(buf_f)) {
+	case XFS_BLFT_BTREE_BUF:
+		switch (magic32) {
+		case XFS_ABTB_CRC_MAGIC:
+		case XFS_ABTB_MAGIC:
+			bp->b_ops = &xfs_bnobt_buf_ops;
+			break;
+		case XFS_ABTC_CRC_MAGIC:
+		case XFS_ABTC_MAGIC:
+			bp->b_ops = &xfs_cntbt_buf_ops;
+			break;
+		case XFS_IBT_CRC_MAGIC:
+		case XFS_IBT_MAGIC:
+			bp->b_ops = &xfs_inobt_buf_ops;
+			break;
+		case XFS_FIBT_CRC_MAGIC:
+		case XFS_FIBT_MAGIC:
+			bp->b_ops = &xfs_finobt_buf_ops;
+			break;
+		case XFS_BMAP_CRC_MAGIC:
+		case XFS_BMAP_MAGIC:
+			bp->b_ops = &xfs_bmbt_buf_ops;
+			break;
+		case XFS_RMAP_CRC_MAGIC:
+			bp->b_ops = &xfs_rmapbt_buf_ops;
+			break;
+		case XFS_REFC_CRC_MAGIC:
+			bp->b_ops = &xfs_refcountbt_buf_ops;
+			break;
+		default:
+			warnmsg = "Bad btree block magic!";
+			break;
+		}
+		break;
+	case XFS_BLFT_AGF_BUF:
+		if (magic32 != XFS_AGF_MAGIC) {
+			warnmsg = "Bad AGF block magic!";
+			break;
+		}
+		bp->b_ops = &xfs_agf_buf_ops;
+		break;
+	case XFS_BLFT_AGFL_BUF:
+		if (magic32 != XFS_AGFL_MAGIC) {
+			warnmsg = "Bad AGFL block magic!";
+			break;
+		}
+		bp->b_ops = &xfs_agfl_buf_ops;
+		break;
+	case XFS_BLFT_AGI_BUF:
+		if (magic32 != XFS_AGI_MAGIC) {
+			warnmsg = "Bad AGI block magic!";
+			break;
+		}
+		bp->b_ops = &xfs_agi_buf_ops;
+		break;
+	case XFS_BLFT_UDQUOT_BUF:
+	case XFS_BLFT_PDQUOT_BUF:
+	case XFS_BLFT_GDQUOT_BUF:
+#ifdef CONFIG_XFS_QUOTA
+		if (magic16 != XFS_DQUOT_MAGIC) {
+			warnmsg = "Bad DQUOT block magic!";
+			break;
+		}
+		bp->b_ops = &xfs_dquot_buf_ops;
+#else
+		xfs_alert(mp,
+	"Trying to recover dquots without QUOTA support built in!");
+		ASSERT(0);
+#endif
+		break;
+	case XFS_BLFT_DINO_BUF:
+		if (magic16 != XFS_DINODE_MAGIC) {
+			warnmsg = "Bad INODE block magic!";
+			break;
+		}
+		bp->b_ops = &xfs_inode_buf_ops;
+		break;
+	case XFS_BLFT_SYMLINK_BUF:
+		if (magic32 != XFS_SYMLINK_MAGIC) {
+			warnmsg = "Bad symlink block magic!";
+			break;
+		}
+		bp->b_ops = &xfs_symlink_buf_ops;
+		break;
+	case XFS_BLFT_DIR_BLOCK_BUF:
+		if (magic32 != XFS_DIR2_BLOCK_MAGIC &&
+		    magic32 != XFS_DIR3_BLOCK_MAGIC) {
+			warnmsg = "Bad dir block magic!";
+			break;
+		}
+		bp->b_ops = &xfs_dir3_block_buf_ops;
+		break;
+	case XFS_BLFT_DIR_DATA_BUF:
+		if (magic32 != XFS_DIR2_DATA_MAGIC &&
+		    magic32 != XFS_DIR3_DATA_MAGIC) {
+			warnmsg = "Bad dir data magic!";
+			break;
+		}
+		bp->b_ops = &xfs_dir3_data_buf_ops;
+		break;
+	case XFS_BLFT_DIR_FREE_BUF:
+		if (magic32 != XFS_DIR2_FREE_MAGIC &&
+		    magic32 != XFS_DIR3_FREE_MAGIC) {
+			warnmsg = "Bad dir3 free magic!";
+			break;
+		}
+		bp->b_ops = &xfs_dir3_free_buf_ops;
+		break;
+	case XFS_BLFT_DIR_LEAF1_BUF:
+		if (magicda != XFS_DIR2_LEAF1_MAGIC &&
+		    magicda != XFS_DIR3_LEAF1_MAGIC) {
+			warnmsg = "Bad dir leaf1 magic!";
+			break;
+		}
+		bp->b_ops = &xfs_dir3_leaf1_buf_ops;
+		break;
+	case XFS_BLFT_DIR_LEAFN_BUF:
+		if (magicda != XFS_DIR2_LEAFN_MAGIC &&
+		    magicda != XFS_DIR3_LEAFN_MAGIC) {
+			warnmsg = "Bad dir leafn magic!";
+			break;
+		}
+		bp->b_ops = &xfs_dir3_leafn_buf_ops;
+		break;
+	case XFS_BLFT_DA_NODE_BUF:
+		if (magicda != XFS_DA_NODE_MAGIC &&
+		    magicda != XFS_DA3_NODE_MAGIC) {
+			warnmsg = "Bad da node magic!";
+			break;
+		}
+		bp->b_ops = &xfs_da3_node_buf_ops;
+		break;
+	case XFS_BLFT_ATTR_LEAF_BUF:
+		if (magicda != XFS_ATTR_LEAF_MAGIC &&
+		    magicda != XFS_ATTR3_LEAF_MAGIC) {
+			warnmsg = "Bad attr leaf magic!";
+			break;
+		}
+		bp->b_ops = &xfs_attr3_leaf_buf_ops;
+		break;
+	case XFS_BLFT_ATTR_RMT_BUF:
+		if (magic32 != XFS_ATTR3_RMT_MAGIC) {
+			warnmsg = "Bad attr remote magic!";
+			break;
+		}
+		bp->b_ops = &xfs_attr3_rmt_buf_ops;
+		break;
+	case XFS_BLFT_SB_BUF:
+		if (magic32 != XFS_SB_MAGIC) {
+			warnmsg = "Bad SB block magic!";
+			break;
+		}
+		bp->b_ops = &xfs_sb_buf_ops;
+		break;
+#ifdef CONFIG_XFS_RT
+	case XFS_BLFT_RTBITMAP_BUF:
+	case XFS_BLFT_RTSUMMARY_BUF:
+		/* no magic numbers for verification of RT buffers */
+		bp->b_ops = &xfs_rtbuf_ops;
+		break;
+#endif /* CONFIG_XFS_RT */
+	default:
+		xfs_warn(mp, "Unknown buffer type %d!",
+			 xfs_blft_from_flags(buf_f));
+		break;
+	}
+
+	/*
+	 * Nothing else to do in the case of a NULL current LSN as this means
+	 * the buffer is more recent than the change in the log and will be
+	 * skipped.
+	 */
+	if (current_lsn == NULLCOMMITLSN)
+		return;
+
+	if (warnmsg) {
+		xfs_warn(mp, warnmsg);
+		ASSERT(0);
+	}
+
+	/*
+	 * We must update the metadata LSN of the buffer as it is written out to
+	 * ensure that older transactions never replay over this one and corrupt
+	 * the buffer. This can occur if log recovery is interrupted at some
+	 * point after the current transaction completes, at which point a
+	 * subsequent mount starts recovery from the beginning.
+	 *
+	 * Write verifiers update the metadata LSN from log items attached to
+	 * the buffer. Therefore, initialize a bli purely to carry the LSN to
+	 * the verifier. We'll clean it up in our ->iodone() callback.
+	 */
+	if (bp->b_ops) {
+		struct xfs_buf_log_item	*bip;
+
+		bp->b_flags |= _XBF_LOGRECOVERY;
+		xfs_buf_item_init(bp, mp);
+		bip = bp->b_log_item;
+		bip->bli_item.li_lsn = current_lsn;
+	}
+}
+
+/*
+ * Perform a 'normal' buffer recovery.  Each logged region of the
+ * buffer should be copied over the corresponding region in the
+ * given buffer.  The bitmap in the buf log format structure indicates
+ * where to place the logged data.
+ */
+STATIC void
+xlog_recover_do_reg_buffer(
+	struct xfs_mount		*mp,
+	struct xlog_recover_item	*item,
+	struct xfs_buf			*bp,
+	struct xfs_buf_log_format	*buf_f,
+	xfs_lsn_t			current_lsn)
+{
+	int			i;
+	int			bit;
+	int			nbits;
+	xfs_failaddr_t		fa;
+	const size_t		size_disk_dquot = sizeof(struct xfs_disk_dquot);
+
+	trace_xfs_log_recover_buf_reg_buf(mp->m_log, buf_f);
+
+	bit = 0;
+	i = 1;  /* 0 is the buf format structure */
+	while (1) {
+		bit = xfs_next_bit(buf_f->blf_data_map,
+				   buf_f->blf_map_size, bit);
+		if (bit == -1)
+			break;
+		nbits = xfs_contig_bits(buf_f->blf_data_map,
+					buf_f->blf_map_size, bit);
+		ASSERT(nbits > 0);
+		ASSERT(item->ri_buf[i].i_addr != NULL);
+		ASSERT(item->ri_buf[i].i_len % XFS_BLF_CHUNK == 0);
+		ASSERT(BBTOB(bp->b_length) >=
+		       ((uint)bit << XFS_BLF_SHIFT) + (nbits << XFS_BLF_SHIFT));
+
+		/*
+		 * The dirty regions logged in the buffer, even though
+		 * contiguous, may span multiple chunks. This is because the
+		 * dirty region may span a physical page boundary in a buffer
+		 * and hence be split into two separate vectors for writing into
+		 * the log. Hence we need to trim nbits back to the length of
+		 * the current region being copied out of the log.
+		 */
+		if (item->ri_buf[i].i_len < (nbits << XFS_BLF_SHIFT))
+			nbits = item->ri_buf[i].i_len >> XFS_BLF_SHIFT;
+
+		/*
+		 * Do a sanity check if this is a dquot buffer. Just checking
+		 * the first dquot in the buffer should do. XXXThis is
+		 * probably a good thing to do for other buf types also.
+		 */
+		fa = NULL;
+		if (buf_f->blf_flags &
+		   (XFS_BLF_UDQUOT_BUF|XFS_BLF_PDQUOT_BUF|XFS_BLF_GDQUOT_BUF)) {
+			if (item->ri_buf[i].i_addr == NULL) {
+				xfs_alert(mp,
+					"XFS: NULL dquot in %s.", __func__);
+				goto next;
+			}
+			if (item->ri_buf[i].i_len < size_disk_dquot) {
+				xfs_alert(mp,
+					"XFS: dquot too small (%d) in %s.",
+					item->ri_buf[i].i_len, __func__);
+				goto next;
+			}
+			fa = xfs_dquot_verify(mp, item->ri_buf[i].i_addr, -1);
+			if (fa) {
+				xfs_alert(mp,
+	"dquot corrupt at %pS trying to replay into block 0x%llx",
+					fa, bp->b_bn);
+				goto next;
+			}
+		}
+
+		memcpy(xfs_buf_offset(bp,
+			(uint)bit << XFS_BLF_SHIFT),	/* dest */
+			item->ri_buf[i].i_addr,		/* source */
+			nbits<<XFS_BLF_SHIFT);		/* length */
+ next:
+		i++;
+		bit += nbits;
+	}
+
+	/* Shouldn't be any more regions */
+	ASSERT(i == item->ri_total);
+
+	xlog_recover_validate_buf_type(mp, bp, buf_f, current_lsn);
+}
+
+/*
+ * Perform a dquot buffer recovery.
+ * Simple algorithm: if we have found a QUOTAOFF log item of the same type
+ * (ie. USR or GRP), then just toss this buffer away; don't recover it.
+ * Else, treat it as a regular buffer and do recovery.
+ *
+ * Return false if the buffer was tossed and true if we recovered the buffer to
+ * indicate to the caller if the buffer needs writing.
+ */
+STATIC bool
+xlog_recover_do_dquot_buffer(
+	struct xfs_mount		*mp,
+	struct xlog			*log,
+	struct xlog_recover_item	*item,
+	struct xfs_buf			*bp,
+	struct xfs_buf_log_format	*buf_f)
+{
+	uint			type;
+
+	trace_xfs_log_recover_buf_dquot_buf(log, buf_f);
+
+	/*
+	 * Filesystems are required to send in quota flags at mount time.
+	 */
+	if (!mp->m_qflags)
+		return false;
+
+	type = 0;
+	if (buf_f->blf_flags & XFS_BLF_UDQUOT_BUF)
+		type |= XFS_DQTYPE_USER;
+	if (buf_f->blf_flags & XFS_BLF_PDQUOT_BUF)
+		type |= XFS_DQTYPE_PROJ;
+	if (buf_f->blf_flags & XFS_BLF_GDQUOT_BUF)
+		type |= XFS_DQTYPE_GROUP;
+	/*
+	 * This type of quotas was turned off, so ignore this buffer
+	 */
+	if (log->l_quotaoffs_flag & type)
+		return false;
+
+	xlog_recover_do_reg_buffer(mp, item, bp, buf_f, NULLCOMMITLSN);
+	return true;
+}
+
+/*
+ * Perform recovery for a buffer full of inodes.  In these buffers, the only
+ * data which should be recovered is that which corresponds to the
+ * di_next_unlinked pointers in the on disk inode structures.  The rest of the
+ * data for the inodes is always logged through the inodes themselves rather
+ * than the inode buffer and is recovered in xlog_recover_inode_pass2().
+ *
+ * The only time when buffers full of inodes are fully recovered is when the
+ * buffer is full of newly allocated inodes.  In this case the buffer will
+ * not be marked as an inode buffer and so will be sent to
+ * xlog_recover_do_reg_buffer() below during recovery.
+ */
+STATIC int
+xlog_recover_do_inode_buffer(
+	struct xfs_mount		*mp,
+	struct xlog_recover_item	*item,
+	struct xfs_buf			*bp,
+	struct xfs_buf_log_format	*buf_f)
+{
+	int				i;
+	int				item_index = 0;
+	int				bit = 0;
+	int				nbits = 0;
+	int				reg_buf_offset = 0;
+	int				reg_buf_bytes = 0;
+	int				next_unlinked_offset;
+	int				inodes_per_buf;
+	xfs_agino_t			*logged_nextp;
+	xfs_agino_t			*buffer_nextp;
+
+	trace_xfs_log_recover_buf_inode_buf(mp->m_log, buf_f);
+
+	/*
+	 * Post recovery validation only works properly on CRC enabled
+	 * filesystems.
+	 */
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		bp->b_ops = &xfs_inode_buf_ops;
+
+	inodes_per_buf = BBTOB(bp->b_length) >> mp->m_sb.sb_inodelog;
+	for (i = 0; i < inodes_per_buf; i++) {
+		next_unlinked_offset = (i * mp->m_sb.sb_inodesize) +
+			offsetof(xfs_dinode_t, di_next_unlinked);
+
+		while (next_unlinked_offset >=
+		       (reg_buf_offset + reg_buf_bytes)) {
+			/*
+			 * The next di_next_unlinked field is beyond
+			 * the current logged region.  Find the next
+			 * logged region that contains or is beyond
+			 * the current di_next_unlinked field.
+			 */
+			bit += nbits;
+			bit = xfs_next_bit(buf_f->blf_data_map,
+					   buf_f->blf_map_size, bit);
+
+			/*
+			 * If there are no more logged regions in the
+			 * buffer, then we're done.
+			 */
+			if (bit == -1)
+				return 0;
+
+			nbits = xfs_contig_bits(buf_f->blf_data_map,
+						buf_f->blf_map_size, bit);
+			ASSERT(nbits > 0);
+			reg_buf_offset = bit << XFS_BLF_SHIFT;
+			reg_buf_bytes = nbits << XFS_BLF_SHIFT;
+			item_index++;
+		}
+
+		/*
+		 * If the current logged region starts after the current
+		 * di_next_unlinked field, then move on to the next
+		 * di_next_unlinked field.
+		 */
+		if (next_unlinked_offset < reg_buf_offset)
+			continue;
+
+		ASSERT(item->ri_buf[item_index].i_addr != NULL);
+		ASSERT((item->ri_buf[item_index].i_len % XFS_BLF_CHUNK) == 0);
+		ASSERT((reg_buf_offset + reg_buf_bytes) <= BBTOB(bp->b_length));
+
+		/*
+		 * The current logged region contains a copy of the
+		 * current di_next_unlinked field.  Extract its value
+		 * and copy it to the buffer copy.
+		 */
+		logged_nextp = item->ri_buf[item_index].i_addr +
+				next_unlinked_offset - reg_buf_offset;
+		if (XFS_IS_CORRUPT(mp, *logged_nextp == 0)) {
+			xfs_alert(mp,
+		"Bad inode buffer log record (ptr = "PTR_FMT", bp = "PTR_FMT"). "
+		"Trying to replay bad (0) inode di_next_unlinked field.",
+				item, bp);
+			return -EFSCORRUPTED;
+		}
+
+		buffer_nextp = xfs_buf_offset(bp, next_unlinked_offset);
+		*buffer_nextp = *logged_nextp;
+
+		/*
+		 * If necessary, recalculate the CRC in the on-disk inode. We
+		 * have to leave the inode in a consistent state for whoever
+		 * reads it next....
+		 */
+		xfs_dinode_calc_crc(mp,
+				xfs_buf_offset(bp, i * mp->m_sb.sb_inodesize));
+
+	}
+
+	return 0;
+}
+
+/*
+ * V5 filesystems know the age of the buffer on disk being recovered. We can
+ * have newer objects on disk than we are replaying, and so for these cases we
+ * don't want to replay the current change as that will make the buffer contents
+ * temporarily invalid on disk.
+ *
+ * The magic number might not match the buffer type we are going to recover
+ * (e.g. reallocated blocks), so we ignore the xfs_buf_log_format flags.  Hence
+ * extract the LSN of the existing object in the buffer based on it's current
+ * magic number.  If we don't recognise the magic number in the buffer, then
+ * return a LSN of -1 so that the caller knows it was an unrecognised block and
+ * so can recover the buffer.
+ *
+ * Note: we cannot rely solely on magic number matches to determine that the
+ * buffer has a valid LSN - we also need to verify that it belongs to this
+ * filesystem, so we need to extract the object's LSN and compare it to that
+ * which we read from the superblock. If the UUIDs don't match, then we've got a
+ * stale metadata block from an old filesystem instance that we need to recover
+ * over the top of.
+ */
+static xfs_lsn_t
+xlog_recover_get_buf_lsn(
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp)
+{
+	uint32_t		magic32;
+	uint16_t		magic16;
+	uint16_t		magicda;
+	void			*blk = bp->b_addr;
+	uuid_t			*uuid;
+	xfs_lsn_t		lsn = -1;
+
+	/* v4 filesystems always recover immediately */
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		goto recover_immediately;
+
+	magic32 = be32_to_cpu(*(__be32 *)blk);
+	switch (magic32) {
+	case XFS_ABTB_CRC_MAGIC:
+	case XFS_ABTC_CRC_MAGIC:
+	case XFS_ABTB_MAGIC:
+	case XFS_ABTC_MAGIC:
+	case XFS_RMAP_CRC_MAGIC:
+	case XFS_REFC_CRC_MAGIC:
+	case XFS_IBT_CRC_MAGIC:
+	case XFS_IBT_MAGIC: {
+		struct xfs_btree_block *btb = blk;
+
+		lsn = be64_to_cpu(btb->bb_u.s.bb_lsn);
+		uuid = &btb->bb_u.s.bb_uuid;
+		break;
+	}
+	case XFS_BMAP_CRC_MAGIC:
+	case XFS_BMAP_MAGIC: {
+		struct xfs_btree_block *btb = blk;
+
+		lsn = be64_to_cpu(btb->bb_u.l.bb_lsn);
+		uuid = &btb->bb_u.l.bb_uuid;
+		break;
+	}
+	case XFS_AGF_MAGIC:
+		lsn = be64_to_cpu(((struct xfs_agf *)blk)->agf_lsn);
+		uuid = &((struct xfs_agf *)blk)->agf_uuid;
+		break;
+	case XFS_AGFL_MAGIC:
+		lsn = be64_to_cpu(((struct xfs_agfl *)blk)->agfl_lsn);
+		uuid = &((struct xfs_agfl *)blk)->agfl_uuid;
+		break;
+	case XFS_AGI_MAGIC:
+		lsn = be64_to_cpu(((struct xfs_agi *)blk)->agi_lsn);
+		uuid = &((struct xfs_agi *)blk)->agi_uuid;
+		break;
+	case XFS_SYMLINK_MAGIC:
+		lsn = be64_to_cpu(((struct xfs_dsymlink_hdr *)blk)->sl_lsn);
+		uuid = &((struct xfs_dsymlink_hdr *)blk)->sl_uuid;
+		break;
+	case XFS_DIR3_BLOCK_MAGIC:
+	case XFS_DIR3_DATA_MAGIC:
+	case XFS_DIR3_FREE_MAGIC:
+		lsn = be64_to_cpu(((struct xfs_dir3_blk_hdr *)blk)->lsn);
+		uuid = &((struct xfs_dir3_blk_hdr *)blk)->uuid;
+		break;
+	case XFS_ATTR3_RMT_MAGIC:
+		/*
+		 * Remote attr blocks are written synchronously, rather than
+		 * being logged. That means they do not contain a valid LSN
+		 * (i.e. transactionally ordered) in them, and hence any time we
+		 * see a buffer to replay over the top of a remote attribute
+		 * block we should simply do so.
+		 */
+		goto recover_immediately;
+	case XFS_SB_MAGIC:
+		/*
+		 * superblock uuids are magic. We may or may not have a
+		 * sb_meta_uuid on disk, but it will be set in the in-core
+		 * superblock. We set the uuid pointer for verification
+		 * according to the superblock feature mask to ensure we check
+		 * the relevant UUID in the superblock.
+		 */
+		lsn = be64_to_cpu(((struct xfs_dsb *)blk)->sb_lsn);
+		if (xfs_sb_version_hasmetauuid(&mp->m_sb))
+			uuid = &((struct xfs_dsb *)blk)->sb_meta_uuid;
+		else
+			uuid = &((struct xfs_dsb *)blk)->sb_uuid;
+		break;
+	default:
+		break;
+	}
+
+	if (lsn != (xfs_lsn_t)-1) {
+		if (!uuid_equal(&mp->m_sb.sb_meta_uuid, uuid))
+			goto recover_immediately;
+		return lsn;
+	}
+
+	magicda = be16_to_cpu(((struct xfs_da_blkinfo *)blk)->magic);
+	switch (magicda) {
+	case XFS_DIR3_LEAF1_MAGIC:
+	case XFS_DIR3_LEAFN_MAGIC:
+	case XFS_DA3_NODE_MAGIC:
+		lsn = be64_to_cpu(((struct xfs_da3_blkinfo *)blk)->lsn);
+		uuid = &((struct xfs_da3_blkinfo *)blk)->uuid;
+		break;
+	default:
+		break;
+	}
+
+	if (lsn != (xfs_lsn_t)-1) {
+		if (!uuid_equal(&mp->m_sb.sb_uuid, uuid))
+			goto recover_immediately;
+		return lsn;
+	}
+
+	/*
+	 * We do individual object checks on dquot and inode buffers as they
+	 * have their own individual LSN records. Also, we could have a stale
+	 * buffer here, so we have to at least recognise these buffer types.
+	 *
+	 * A notd complexity here is inode unlinked list processing - it logs
+	 * the inode directly in the buffer, but we don't know which inodes have
+	 * been modified, and there is no global buffer LSN. Hence we need to
+	 * recover all inode buffer types immediately. This problem will be
+	 * fixed by logical logging of the unlinked list modifications.
+	 */
+	magic16 = be16_to_cpu(*(__be16 *)blk);
+	switch (magic16) {
+	case XFS_DQUOT_MAGIC:
+	case XFS_DINODE_MAGIC:
+		goto recover_immediately;
+	default:
+		break;
+	}
+
+	/* unknown buffer contents, recover immediately */
+
+recover_immediately:
+	return (xfs_lsn_t)-1;
+
+}
+
+/*
+ * This routine replays a modification made to a buffer at runtime.
+ * There are actually two types of buffer, regular and inode, which
+ * are handled differently.  Inode buffers are handled differently
+ * in that we only recover a specific set of data from them, namely
+ * the inode di_next_unlinked fields.  This is because all other inode
+ * data is actually logged via inode records and any data we replay
+ * here which overlaps that may be stale.
+ *
+ * When meta-data buffers are freed at run time we log a buffer item
+ * with the XFS_BLF_CANCEL bit set to indicate that previous copies
+ * of the buffer in the log should not be replayed at recovery time.
+ * This is so that if the blocks covered by the buffer are reused for
+ * file data before we crash we don't end up replaying old, freed
+ * meta-data into a user's file.
+ *
+ * To handle the cancellation of buffer log items, we make two passes
+ * over the log during recovery.  During the first we build a table of
+ * those buffers which have been cancelled, and during the second we
+ * only replay those buffers which do not have corresponding cancel
+ * records in the table.  See xlog_recover_buf_pass[1,2] above
+ * for more details on the implementation of the table of cancel records.
+ */
+STATIC int
+xlog_recover_buf_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			current_lsn)
+{
+	struct xfs_buf_log_format	*buf_f = item->ri_buf[0].i_addr;
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_buf			*bp;
+	int				error;
+	uint				buf_flags;
+	xfs_lsn_t			lsn;
+
+	/*
+	 * In this pass we only want to recover all the buffers which have
+	 * not been cancelled and are not cancellation buffers themselves.
+	 */
+	if (buf_f->blf_flags & XFS_BLF_CANCEL) {
+		if (xlog_put_buffer_cancelled(log, buf_f->blf_blkno,
+				buf_f->blf_len))
+			goto cancelled;
+	} else {
+
+		if (xlog_is_buffer_cancelled(log, buf_f->blf_blkno,
+				buf_f->blf_len))
+			goto cancelled;
+	}
+
+	trace_xfs_log_recover_buf_recover(log, buf_f);
+
+	buf_flags = 0;
+	if (buf_f->blf_flags & XFS_BLF_INODE_BUF)
+		buf_flags |= XBF_UNMAPPED;
+
+	error = xfs_buf_read(mp->m_ddev_targp, buf_f->blf_blkno, buf_f->blf_len,
+			  buf_flags, &bp, NULL);
+	if (error)
+		return error;
+
+	/*
+	 * Recover the buffer only if we get an LSN from it and it's less than
+	 * the lsn of the transaction we are replaying.
+	 *
+	 * Note that we have to be extremely careful of readahead here.
+	 * Readahead does not attach verfiers to the buffers so if we don't
+	 * actually do any replay after readahead because of the LSN we found
+	 * in the buffer if more recent than that current transaction then we
+	 * need to attach the verifier directly. Failure to do so can lead to
+	 * future recovery actions (e.g. EFI and unlinked list recovery) can
+	 * operate on the buffers and they won't get the verifier attached. This
+	 * can lead to blocks on disk having the correct content but a stale
+	 * CRC.
+	 *
+	 * It is safe to assume these clean buffers are currently up to date.
+	 * If the buffer is dirtied by a later transaction being replayed, then
+	 * the verifier will be reset to match whatever recover turns that
+	 * buffer into.
+	 */
+	lsn = xlog_recover_get_buf_lsn(mp, bp);
+	if (lsn && lsn != -1 && XFS_LSN_CMP(lsn, current_lsn) >= 0) {
+		trace_xfs_log_recover_buf_skip(log, buf_f);
+		xlog_recover_validate_buf_type(mp, bp, buf_f, NULLCOMMITLSN);
+		goto out_release;
+	}
+
+	if (buf_f->blf_flags & XFS_BLF_INODE_BUF) {
+		error = xlog_recover_do_inode_buffer(mp, item, bp, buf_f);
+		if (error)
+			goto out_release;
+	} else if (buf_f->blf_flags &
+		  (XFS_BLF_UDQUOT_BUF|XFS_BLF_PDQUOT_BUF|XFS_BLF_GDQUOT_BUF)) {
+		bool	dirty;
+
+		dirty = xlog_recover_do_dquot_buffer(mp, log, item, bp, buf_f);
+		if (!dirty)
+			goto out_release;
+	} else {
+		xlog_recover_do_reg_buffer(mp, item, bp, buf_f, current_lsn);
+	}
+
+	/*
+	 * Perform delayed write on the buffer.  Asynchronous writes will be
+	 * slower when taking into account all the buffers to be flushed.
+	 *
+	 * Also make sure that only inode buffers with good sizes stay in
+	 * the buffer cache.  The kernel moves inodes in buffers of 1 block
+	 * or inode_cluster_size bytes, whichever is bigger.  The inode
+	 * buffers in the log can be a different size if the log was generated
+	 * by an older kernel using unclustered inode buffers or a newer kernel
+	 * running with a different inode cluster size.  Regardless, if
+	 * the inode buffer size isn't max(blocksize, inode_cluster_size)
+	 * for *our* value of inode_cluster_size, then we need to keep
+	 * the buffer out of the buffer cache so that the buffer won't
+	 * overlap with future reads of those inodes.
+	 */
+	if (XFS_DINODE_MAGIC ==
+	    be16_to_cpu(*((__be16 *)xfs_buf_offset(bp, 0))) &&
+	    (BBTOB(bp->b_length) != M_IGEO(log->l_mp)->inode_cluster_size)) {
+		xfs_buf_stale(bp);
+		error = xfs_bwrite(bp);
+	} else {
+		ASSERT(bp->b_mount == mp);
+		bp->b_flags |= _XBF_LOGRECOVERY;
+		xfs_buf_delwri_queue(bp, buffer_list);
+	}
+
+out_release:
+	xfs_buf_relse(bp);
+	return error;
+cancelled:
+	trace_xfs_log_recover_buf_cancel(log, buf_f);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_buf_item_ops = {
+	.item_type		= XFS_LI_BUF,
+	.reorder		= xlog_recover_buf_reorder,
+	.ra_pass2		= xlog_recover_buf_ra_pass2,
+	.commit_pass1		= xlog_recover_buf_commit_pass1,
+	.commit_pass2		= xlog_recover_buf_commit_pass2,
+};
diff --git a/fs/xfs/xfs_dir2_readdir.c b/fs/xfs/xfs_dir2_readdir.c
index 0d3b640cf1cc..66deddd5e296 100644
--- a/fs/xfs/xfs_dir2_readdir.c
+++ b/fs/xfs/xfs_dir2_readdir.c
@@ -147,7 +147,7 @@ xfs_dir2_block_getdents(
 	xfs_off_t		cook;
 	struct xfs_da_geometry	*geo = args->geo;
 	int			lock_mode;
-	unsigned int		offset;
+	unsigned int		offset, next_offset;
 	unsigned int		end;
 
 	/*
@@ -173,9 +173,10 @@ xfs_dir2_block_getdents(
 	 * Loop over the data portion of the block.
 	 * Each object is a real entry (dep) or an unused one (dup).
 	 */
-	offset = geo->data_entry_offset;
 	end = xfs_dir3_data_end_offset(geo, bp->b_addr);
-	while (offset < end) {
+	for (offset = geo->data_entry_offset;
+	     offset < end;
+	     offset = next_offset) {
 		struct xfs_dir2_data_unused	*dup = bp->b_addr + offset;
 		struct xfs_dir2_data_entry	*dep = bp->b_addr + offset;
 		uint8_t filetype;
@@ -184,14 +185,15 @@ xfs_dir2_block_getdents(
 		 * Unused, skip it.
 		 */
 		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
-			offset += be16_to_cpu(dup->length);
+			next_offset = offset + be16_to_cpu(dup->length);
 			continue;
 		}
 
 		/*
 		 * Bump pointer for the next iteration.
 		 */
-		offset += xfs_dir2_data_entsize(dp->i_mount, dep->namelen);
+		next_offset = offset +
+			xfs_dir2_data_entsize(dp->i_mount, dep->namelen);
 
 		/*
 		 * The entry is before the desired starting point, skip it.
@@ -522,7 +524,7 @@ xfs_readdir(
 	args.geo = dp->i_mount->m_dir_geo;
 	args.trans = tp;
 
-	if (dp->i_d.di_format == XFS_DINODE_FMT_LOCAL)
+	if (dp->i_df.if_format == XFS_DINODE_FMT_LOCAL)
 		rval = xfs_dir2_sf_getdents(&args, ctx);
 	else if ((rval = xfs_dir2_isblock(&args, &v)))
 		;
diff --git a/fs/xfs/xfs_discard.c b/fs/xfs/xfs_discard.c
index 0b8350e84d28..f979d0d7e6cd 100644
--- a/fs/xfs/xfs_discard.c
+++ b/fs/xfs/xfs_discard.c
@@ -31,6 +31,7 @@ xfs_trim_extents(
 	struct block_device	*bdev = mp->m_ddev_targp->bt_bdev;
 	struct xfs_btree_cur	*cur;
 	struct xfs_buf		*agbp;
+	struct xfs_agf		*agf;
 	struct xfs_perag	*pag;
 	int			error;
 	int			i;
@@ -47,14 +48,14 @@ xfs_trim_extents(
 	error = xfs_alloc_read_agf(mp, NULL, agno, 0, &agbp);
 	if (error)
 		goto out_put_perag;
+	agf = agbp->b_addr;
 
 	cur = xfs_allocbt_init_cursor(mp, NULL, agbp, agno, XFS_BTNUM_CNT);
 
 	/*
 	 * Look up the longest btree in the AGF and start with it.
 	 */
-	error = xfs_alloc_lookup_ge(cur, 0,
-			    be32_to_cpu(XFS_BUF_TO_AGF(agbp)->agf_longest), &i);
+	error = xfs_alloc_lookup_ge(cur, 0, be32_to_cpu(agf->agf_longest), &i);
 	if (error)
 		goto out_del_cursor;
 
@@ -75,7 +76,7 @@ xfs_trim_extents(
 			error = -EFSCORRUPTED;
 			goto out_del_cursor;
 		}
-		ASSERT(flen <= be32_to_cpu(XFS_BUF_TO_AGF(agbp)->agf_longest));
+		ASSERT(flen <= be32_to_cpu(agf->agf_longest));
 
 		/*
 		 * use daddr format for all range/len calculations as that is
diff --git a/fs/xfs/xfs_dquot.c b/fs/xfs/xfs_dquot.c
index d223e1ae90a6..bcd73b9c2994 100644
--- a/fs/xfs/xfs_dquot.c
+++ b/fs/xfs/xfs_dquot.c
@@ -23,6 +23,7 @@
 #include "xfs_trace.h"
 #include "xfs_log.h"
 #include "xfs_bmap_btree.h"
+#include "xfs_error.h"
 
 /*
  * Lock order:
@@ -66,39 +67,61 @@ xfs_qm_dqdestroy(
  */
 void
 xfs_qm_adjust_dqlimits(
-	struct xfs_mount	*mp,
 	struct xfs_dquot	*dq)
 {
+	struct xfs_mount	*mp = dq->q_mount;
 	struct xfs_quotainfo	*q = mp->m_quotainfo;
-	struct xfs_disk_dquot	*d = &dq->q_core;
 	struct xfs_def_quota	*defq;
 	int			prealloc = 0;
 
-	ASSERT(d->d_id);
-	defq = xfs_get_defquota(dq, q);
+	ASSERT(dq->q_id);
+	defq = xfs_get_defquota(q, xfs_dquot_type(dq));
 
-	if (defq->bsoftlimit && !d->d_blk_softlimit) {
-		d->d_blk_softlimit = cpu_to_be64(defq->bsoftlimit);
+	if (!dq->q_blk.softlimit) {
+		dq->q_blk.softlimit = defq->blk.soft;
 		prealloc = 1;
 	}
-	if (defq->bhardlimit && !d->d_blk_hardlimit) {
-		d->d_blk_hardlimit = cpu_to_be64(defq->bhardlimit);
+	if (!dq->q_blk.hardlimit) {
+		dq->q_blk.hardlimit = defq->blk.hard;
 		prealloc = 1;
 	}
-	if (defq->isoftlimit && !d->d_ino_softlimit)
-		d->d_ino_softlimit = cpu_to_be64(defq->isoftlimit);
-	if (defq->ihardlimit && !d->d_ino_hardlimit)
-		d->d_ino_hardlimit = cpu_to_be64(defq->ihardlimit);
-	if (defq->rtbsoftlimit && !d->d_rtb_softlimit)
-		d->d_rtb_softlimit = cpu_to_be64(defq->rtbsoftlimit);
-	if (defq->rtbhardlimit && !d->d_rtb_hardlimit)
-		d->d_rtb_hardlimit = cpu_to_be64(defq->rtbhardlimit);
+	if (!dq->q_ino.softlimit)
+		dq->q_ino.softlimit = defq->ino.soft;
+	if (!dq->q_ino.hardlimit)
+		dq->q_ino.hardlimit = defq->ino.hard;
+	if (!dq->q_rtb.softlimit)
+		dq->q_rtb.softlimit = defq->rtb.soft;
+	if (!dq->q_rtb.hardlimit)
+		dq->q_rtb.hardlimit = defq->rtb.hard;
 
 	if (prealloc)
 		xfs_dquot_set_prealloc_limits(dq);
 }
 
 /*
+ * Determine if this quota counter is over either limit and set the quota
+ * timers as appropriate.
+ */
+static inline void
+xfs_qm_adjust_res_timer(
+	struct xfs_dquot_res	*res,
+	struct xfs_quota_limits	*qlim)
+{
+	ASSERT(res->hardlimit == 0 || res->softlimit <= res->hardlimit);
+
+	if ((res->softlimit && res->count > res->softlimit) ||
+	    (res->hardlimit && res->count > res->hardlimit)) {
+		if (res->timer == 0)
+			res->timer = ktime_get_real_seconds() + qlim->time;
+	} else {
+		if (res->timer == 0)
+			res->warnings = 0;
+		else
+			res->timer = 0;
+	}
+}
+
+/*
  * Check the limits and timers of a dquot and start or reset timers
  * if necessary.
  * This gets called even when quota enforcement is OFF, which makes our
@@ -113,91 +136,18 @@ xfs_qm_adjust_dqlimits(
  */
 void
 xfs_qm_adjust_dqtimers(
-	struct xfs_mount	*mp,
-	struct xfs_disk_dquot	*d)
+	struct xfs_dquot	*dq)
 {
-	ASSERT(d->d_id);
-
-#ifdef DEBUG
-	if (d->d_blk_hardlimit)
-		ASSERT(be64_to_cpu(d->d_blk_softlimit) <=
-		       be64_to_cpu(d->d_blk_hardlimit));
-	if (d->d_ino_hardlimit)
-		ASSERT(be64_to_cpu(d->d_ino_softlimit) <=
-		       be64_to_cpu(d->d_ino_hardlimit));
-	if (d->d_rtb_hardlimit)
-		ASSERT(be64_to_cpu(d->d_rtb_softlimit) <=
-		       be64_to_cpu(d->d_rtb_hardlimit));
-#endif
-
-	if (!d->d_btimer) {
-		if ((d->d_blk_softlimit &&
-		     (be64_to_cpu(d->d_bcount) >
-		      be64_to_cpu(d->d_blk_softlimit))) ||
-		    (d->d_blk_hardlimit &&
-		     (be64_to_cpu(d->d_bcount) >
-		      be64_to_cpu(d->d_blk_hardlimit)))) {
-			d->d_btimer = cpu_to_be32(ktime_get_real_seconds() +
-					mp->m_quotainfo->qi_btimelimit);
-		} else {
-			d->d_bwarns = 0;
-		}
-	} else {
-		if ((!d->d_blk_softlimit ||
-		     (be64_to_cpu(d->d_bcount) <=
-		      be64_to_cpu(d->d_blk_softlimit))) &&
-		    (!d->d_blk_hardlimit ||
-		    (be64_to_cpu(d->d_bcount) <=
-		     be64_to_cpu(d->d_blk_hardlimit)))) {
-			d->d_btimer = 0;
-		}
-	}
+	struct xfs_mount	*mp = dq->q_mount;
+	struct xfs_quotainfo	*qi = mp->m_quotainfo;
+	struct xfs_def_quota	*defq;
 
-	if (!d->d_itimer) {
-		if ((d->d_ino_softlimit &&
-		     (be64_to_cpu(d->d_icount) >
-		      be64_to_cpu(d->d_ino_softlimit))) ||
-		    (d->d_ino_hardlimit &&
-		     (be64_to_cpu(d->d_icount) >
-		      be64_to_cpu(d->d_ino_hardlimit)))) {
-			d->d_itimer = cpu_to_be32(ktime_get_real_seconds() +
-					mp->m_quotainfo->qi_itimelimit);
-		} else {
-			d->d_iwarns = 0;
-		}
-	} else {
-		if ((!d->d_ino_softlimit ||
-		     (be64_to_cpu(d->d_icount) <=
-		      be64_to_cpu(d->d_ino_softlimit)))  &&
-		    (!d->d_ino_hardlimit ||
-		     (be64_to_cpu(d->d_icount) <=
-		      be64_to_cpu(d->d_ino_hardlimit)))) {
-			d->d_itimer = 0;
-		}
-	}
+	ASSERT(dq->q_id);
+	defq = xfs_get_defquota(qi, xfs_dquot_type(dq));
 
-	if (!d->d_rtbtimer) {
-		if ((d->d_rtb_softlimit &&
-		     (be64_to_cpu(d->d_rtbcount) >
-		      be64_to_cpu(d->d_rtb_softlimit))) ||
-		    (d->d_rtb_hardlimit &&
-		     (be64_to_cpu(d->d_rtbcount) >
-		      be64_to_cpu(d->d_rtb_hardlimit)))) {
-			d->d_rtbtimer = cpu_to_be32(ktime_get_real_seconds() +
-					mp->m_quotainfo->qi_rtbtimelimit);
-		} else {
-			d->d_rtbwarns = 0;
-		}
-	} else {
-		if ((!d->d_rtb_softlimit ||
-		     (be64_to_cpu(d->d_rtbcount) <=
-		      be64_to_cpu(d->d_rtb_softlimit))) &&
-		    (!d->d_rtb_hardlimit ||
-		     (be64_to_cpu(d->d_rtbcount) <=
-		      be64_to_cpu(d->d_rtb_hardlimit)))) {
-			d->d_rtbtimer = 0;
-		}
-	}
+	xfs_qm_adjust_res_timer(&dq->q_blk, &defq->blk);
+	xfs_qm_adjust_res_timer(&dq->q_ino, &defq->ino);
+	xfs_qm_adjust_res_timer(&dq->q_rtb, &defq->rtb);
 }
 
 /*
@@ -205,20 +155,40 @@ xfs_qm_adjust_dqtimers(
  */
 STATIC void
 xfs_qm_init_dquot_blk(
-	xfs_trans_t	*tp,
-	xfs_mount_t	*mp,
-	xfs_dqid_t	id,
-	uint		type,
-	xfs_buf_t	*bp)
+	struct xfs_trans	*tp,
+	struct xfs_mount	*mp,
+	xfs_dqid_t		id,
+	xfs_dqtype_t		type,
+	struct xfs_buf		*bp)
 {
 	struct xfs_quotainfo	*q = mp->m_quotainfo;
-	xfs_dqblk_t	*d;
-	xfs_dqid_t	curid;
-	int		i;
+	struct xfs_dqblk	*d;
+	xfs_dqid_t		curid;
+	unsigned int		qflag;
+	unsigned int		blftype;
+	int			i;
 
 	ASSERT(tp);
 	ASSERT(xfs_buf_islocked(bp));
 
+	switch (type) {
+	case XFS_DQTYPE_USER:
+		qflag = XFS_UQUOTA_CHKD;
+		blftype = XFS_BLF_UDQUOT_BUF;
+		break;
+	case XFS_DQTYPE_PROJ:
+		qflag = XFS_PQUOTA_CHKD;
+		blftype = XFS_BLF_PDQUOT_BUF;
+		break;
+	case XFS_DQTYPE_GROUP:
+		qflag = XFS_GQUOTA_CHKD;
+		blftype = XFS_BLF_GDQUOT_BUF;
+		break;
+	default:
+		ASSERT(0);
+		return;
+	}
+
 	d = bp->b_addr;
 
 	/*
@@ -230,7 +200,7 @@ xfs_qm_init_dquot_blk(
 		d->dd_diskdq.d_magic = cpu_to_be16(XFS_DQUOT_MAGIC);
 		d->dd_diskdq.d_version = XFS_DQUOT_VERSION;
 		d->dd_diskdq.d_id = cpu_to_be32(curid);
-		d->dd_diskdq.d_flags = type;
+		d->dd_diskdq.d_type = type;
 		if (xfs_sb_version_hascrc(&mp->m_sb)) {
 			uuid_copy(&d->dd_uuid, &mp->m_sb.sb_meta_uuid);
 			xfs_update_cksum((char *)d, sizeof(struct xfs_dqblk),
@@ -238,11 +208,28 @@ xfs_qm_init_dquot_blk(
 		}
 	}
 
-	xfs_trans_dquot_buf(tp, bp,
-			    (type & XFS_DQ_USER ? XFS_BLF_UDQUOT_BUF :
-			    ((type & XFS_DQ_PROJ) ? XFS_BLF_PDQUOT_BUF :
-			     XFS_BLF_GDQUOT_BUF)));
-	xfs_trans_log_buf(tp, bp, 0, BBTOB(q->qi_dqchunklen) - 1);
+	xfs_trans_dquot_buf(tp, bp, blftype);
+
+	/*
+	 * quotacheck uses delayed writes to update all the dquots on disk in an
+	 * efficient manner instead of logging the individual dquot changes as
+	 * they are made. However if we log the buffer allocated here and crash
+	 * after quotacheck while the logged initialisation is still in the
+	 * active region of the log, log recovery can replay the dquot buffer
+	 * initialisation over the top of the checked dquots and corrupt quota
+	 * accounting.
+	 *
+	 * To avoid this problem, quotacheck cannot log the initialised buffer.
+	 * We must still dirty the buffer and write it back before the
+	 * allocation transaction clears the log. Therefore, mark the buffer as
+	 * ordered instead of logging it directly. This is safe for quotacheck
+	 * because it detects and repairs allocated but initialized dquot blocks
+	 * in the quota inodes.
+	 */
+	if (!(mp->m_qflags & qflag))
+		xfs_trans_ordered_buf(tp, bp);
+	else
+		xfs_trans_log_buf(tp, bp, 0, BBTOB(q->qi_dqchunklen) - 1);
 }
 
 /*
@@ -255,8 +242,8 @@ xfs_dquot_set_prealloc_limits(struct xfs_dquot *dqp)
 {
 	uint64_t space;
 
-	dqp->q_prealloc_hi_wmark = be64_to_cpu(dqp->q_core.d_blk_hardlimit);
-	dqp->q_prealloc_lo_wmark = be64_to_cpu(dqp->q_core.d_blk_softlimit);
+	dqp->q_prealloc_hi_wmark = dqp->q_blk.hardlimit;
+	dqp->q_prealloc_lo_wmark = dqp->q_blk.softlimit;
 	if (!dqp->q_prealloc_lo_wmark) {
 		dqp->q_prealloc_lo_wmark = dqp->q_prealloc_hi_wmark;
 		do_div(dqp->q_prealloc_lo_wmark, 100);
@@ -286,14 +273,15 @@ xfs_dquot_disk_alloc(
 	struct xfs_trans	*tp = *tpp;
 	struct xfs_mount	*mp = tp->t_mountp;
 	struct xfs_buf		*bp;
-	struct xfs_inode	*quotip = xfs_quota_inode(mp, dqp->dq_flags);
+	xfs_dqtype_t		qtype = xfs_dquot_type(dqp);
+	struct xfs_inode	*quotip = xfs_quota_inode(mp, qtype);
 	int			nmaps = 1;
 	int			error;
 
 	trace_xfs_dqalloc(dqp);
 
 	xfs_ilock(quotip, XFS_ILOCK_EXCL);
-	if (!xfs_this_quota_on(dqp->q_mount, dqp->dq_flags)) {
+	if (!xfs_this_quota_on(dqp->q_mount, qtype)) {
 		/*
 		 * Return if this type of quotas is turned off while we didn't
 		 * have an inode lock
@@ -330,8 +318,7 @@ xfs_dquot_disk_alloc(
 	 * Make a chunk of dquots out of this buffer and log
 	 * the entire thing.
 	 */
-	xfs_qm_init_dquot_blk(tp, mp, be32_to_cpu(dqp->q_core.d_id),
-			      dqp->dq_flags & XFS_DQ_ALLTYPES, bp);
+	xfs_qm_init_dquot_blk(tp, mp, dqp->q_id, qtype, bp);
 	xfs_buf_set_ref(bp, XFS_DQUOT_REF);
 
 	/*
@@ -378,13 +365,14 @@ xfs_dquot_disk_read(
 {
 	struct xfs_bmbt_irec	map;
 	struct xfs_buf		*bp;
-	struct xfs_inode	*quotip = xfs_quota_inode(mp, dqp->dq_flags);
+	xfs_dqtype_t		qtype = xfs_dquot_type(dqp);
+	struct xfs_inode	*quotip = xfs_quota_inode(mp, qtype);
 	uint			lock_mode;
 	int			nmaps = 1;
 	int			error;
 
 	lock_mode = xfs_ilock_data_map_shared(quotip);
-	if (!xfs_this_quota_on(mp, dqp->dq_flags)) {
+	if (!xfs_this_quota_on(mp, qtype)) {
 		/*
 		 * Return if this type of quotas is turned off while we
 		 * didn't have the quota inode lock.
@@ -436,14 +424,14 @@ STATIC struct xfs_dquot *
 xfs_dquot_alloc(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		id,
-	uint			type)
+	xfs_dqtype_t		type)
 {
 	struct xfs_dquot	*dqp;
 
-	dqp = kmem_zone_zalloc(xfs_qm_dqzone, 0);
+	dqp = kmem_cache_zalloc(xfs_qm_dqzone, GFP_KERNEL | __GFP_NOFAIL);
 
-	dqp->dq_flags = type;
-	dqp->q_core.d_id = cpu_to_be32(id);
+	dqp->q_type = type;
+	dqp->q_id = id;
 	dqp->q_mount = mp;
 	INIT_LIST_HEAD(&dqp->q_lru);
 	mutex_init(&dqp->q_qlock);
@@ -468,13 +456,13 @@ xfs_dquot_alloc(
 	 * quotas.
 	 */
 	switch (type) {
-	case XFS_DQ_USER:
+	case XFS_DQTYPE_USER:
 		/* uses the default lock class */
 		break;
-	case XFS_DQ_GROUP:
+	case XFS_DQTYPE_GROUP:
 		lockdep_set_class(&dqp->q_qlock, &xfs_dquot_group_class);
 		break;
-	case XFS_DQ_PROJ:
+	case XFS_DQTYPE_PROJ:
 		lockdep_set_class(&dqp->q_qlock, &xfs_dquot_project_class);
 		break;
 	default:
@@ -489,26 +477,91 @@ xfs_dquot_alloc(
 }
 
 /* Copy the in-core quota fields in from the on-disk buffer. */
-STATIC void
+STATIC int
 xfs_dquot_from_disk(
 	struct xfs_dquot	*dqp,
 	struct xfs_buf		*bp)
 {
 	struct xfs_disk_dquot	*ddqp = bp->b_addr + dqp->q_bufoffset;
 
+	/*
+	 * Ensure that we got the type and ID we were looking for.
+	 * Everything else was checked by the dquot buffer verifier.
+	 */
+	if ((ddqp->d_type & XFS_DQTYPE_REC_MASK) != xfs_dquot_type(dqp) ||
+	    be32_to_cpu(ddqp->d_id) != dqp->q_id) {
+		xfs_alert_tag(bp->b_mount, XFS_PTAG_VERIFIER_ERROR,
+			  "Metadata corruption detected at %pS, quota %u",
+			  __this_address, dqp->q_id);
+		xfs_alert(bp->b_mount, "Unmount and run xfs_repair");
+		return -EFSCORRUPTED;
+	}
+
 	/* copy everything from disk dquot to the incore dquot */
-	memcpy(&dqp->q_core, ddqp, sizeof(struct xfs_disk_dquot));
+	dqp->q_type = ddqp->d_type;
+	dqp->q_blk.hardlimit = be64_to_cpu(ddqp->d_blk_hardlimit);
+	dqp->q_blk.softlimit = be64_to_cpu(ddqp->d_blk_softlimit);
+	dqp->q_ino.hardlimit = be64_to_cpu(ddqp->d_ino_hardlimit);
+	dqp->q_ino.softlimit = be64_to_cpu(ddqp->d_ino_softlimit);
+	dqp->q_rtb.hardlimit = be64_to_cpu(ddqp->d_rtb_hardlimit);
+	dqp->q_rtb.softlimit = be64_to_cpu(ddqp->d_rtb_softlimit);
+
+	dqp->q_blk.count = be64_to_cpu(ddqp->d_bcount);
+	dqp->q_ino.count = be64_to_cpu(ddqp->d_icount);
+	dqp->q_rtb.count = be64_to_cpu(ddqp->d_rtbcount);
+
+	dqp->q_blk.warnings = be16_to_cpu(ddqp->d_bwarns);
+	dqp->q_ino.warnings = be16_to_cpu(ddqp->d_iwarns);
+	dqp->q_rtb.warnings = be16_to_cpu(ddqp->d_rtbwarns);
+
+	dqp->q_blk.timer = be32_to_cpu(ddqp->d_btimer);
+	dqp->q_ino.timer = be32_to_cpu(ddqp->d_itimer);
+	dqp->q_rtb.timer = be32_to_cpu(ddqp->d_rtbtimer);
 
 	/*
 	 * Reservation counters are defined as reservation plus current usage
 	 * to avoid having to add every time.
 	 */
-	dqp->q_res_bcount = be64_to_cpu(ddqp->d_bcount);
-	dqp->q_res_icount = be64_to_cpu(ddqp->d_icount);
-	dqp->q_res_rtbcount = be64_to_cpu(ddqp->d_rtbcount);
+	dqp->q_blk.reserved = dqp->q_blk.count;
+	dqp->q_ino.reserved = dqp->q_ino.count;
+	dqp->q_rtb.reserved = dqp->q_rtb.count;
 
 	/* initialize the dquot speculative prealloc thresholds */
 	xfs_dquot_set_prealloc_limits(dqp);
+	return 0;
+}
+
+/* Copy the in-core quota fields into the on-disk buffer. */
+void
+xfs_dquot_to_disk(
+	struct xfs_disk_dquot	*ddqp,
+	struct xfs_dquot	*dqp)
+{
+	ddqp->d_magic = cpu_to_be16(XFS_DQUOT_MAGIC);
+	ddqp->d_version = XFS_DQUOT_VERSION;
+	ddqp->d_type = dqp->q_type;
+	ddqp->d_id = cpu_to_be32(dqp->q_id);
+	ddqp->d_pad0 = 0;
+	ddqp->d_pad = 0;
+
+	ddqp->d_blk_hardlimit = cpu_to_be64(dqp->q_blk.hardlimit);
+	ddqp->d_blk_softlimit = cpu_to_be64(dqp->q_blk.softlimit);
+	ddqp->d_ino_hardlimit = cpu_to_be64(dqp->q_ino.hardlimit);
+	ddqp->d_ino_softlimit = cpu_to_be64(dqp->q_ino.softlimit);
+	ddqp->d_rtb_hardlimit = cpu_to_be64(dqp->q_rtb.hardlimit);
+	ddqp->d_rtb_softlimit = cpu_to_be64(dqp->q_rtb.softlimit);
+
+	ddqp->d_bcount = cpu_to_be64(dqp->q_blk.count);
+	ddqp->d_icount = cpu_to_be64(dqp->q_ino.count);
+	ddqp->d_rtbcount = cpu_to_be64(dqp->q_rtb.count);
+
+	ddqp->d_bwarns = cpu_to_be16(dqp->q_blk.warnings);
+	ddqp->d_iwarns = cpu_to_be16(dqp->q_ino.warnings);
+	ddqp->d_rtbwarns = cpu_to_be16(dqp->q_rtb.warnings);
+
+	ddqp->d_btimer = cpu_to_be32(dqp->q_blk.timer);
+	ddqp->d_itimer = cpu_to_be32(dqp->q_ino.timer);
+	ddqp->d_rtbtimer = cpu_to_be32(dqp->q_rtb.timer);
 }
 
 /* Allocate and initialize the dquot buffer for this in-core dquot. */
@@ -557,7 +610,7 @@ static int
 xfs_qm_dqread(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		id,
-	uint			type,
+	xfs_dqtype_t		type,
 	bool			can_alloc,
 	struct xfs_dquot	**dqpp)
 {
@@ -582,9 +635,11 @@ xfs_qm_dqread(
 	 * further.
 	 */
 	ASSERT(xfs_buf_islocked(bp));
-	xfs_dquot_from_disk(dqp, bp);
-
+	error = xfs_dquot_from_disk(dqp, bp);
 	xfs_buf_relse(bp);
+	if (error)
+		goto err;
+
 	*dqpp = dqp;
 	return error;
 
@@ -603,7 +658,7 @@ err:
 static int
 xfs_dq_get_next_id(
 	struct xfs_mount	*mp,
-	uint			type,
+	xfs_dqtype_t		type,
 	xfs_dqid_t		*id)
 {
 	struct xfs_inode	*quotip = xfs_quota_inode(mp, type);
@@ -671,7 +726,7 @@ restart:
 	}
 
 	xfs_dqlock(dqp);
-	if (dqp->dq_flags & XFS_DQ_FREEING) {
+	if (dqp->q_flags & XFS_DQFLAG_FREEING) {
 		xfs_dqunlock(dqp);
 		mutex_unlock(&qi->qi_tree_lock);
 		trace_xfs_dqget_freeing(dqp);
@@ -727,21 +782,21 @@ xfs_qm_dqget_cache_insert(
 static int
 xfs_qm_dqget_checks(
 	struct xfs_mount	*mp,
-	uint			type)
+	xfs_dqtype_t		type)
 {
 	if (WARN_ON_ONCE(!XFS_IS_QUOTA_RUNNING(mp)))
 		return -ESRCH;
 
 	switch (type) {
-	case XFS_DQ_USER:
+	case XFS_DQTYPE_USER:
 		if (!XFS_IS_UQUOTA_ON(mp))
 			return -ESRCH;
 		return 0;
-	case XFS_DQ_GROUP:
+	case XFS_DQTYPE_GROUP:
 		if (!XFS_IS_GQUOTA_ON(mp))
 			return -ESRCH;
 		return 0;
-	case XFS_DQ_PROJ:
+	case XFS_DQTYPE_PROJ:
 		if (!XFS_IS_PQUOTA_ON(mp))
 			return -ESRCH;
 		return 0;
@@ -752,14 +807,14 @@ xfs_qm_dqget_checks(
 }
 
 /*
- * Given the file system, id, and type (UDQUOT/GDQUOT), return a a locked
+ * Given the file system, id, and type (UDQUOT/GDQUOT), return a locked
  * dquot, doing an allocation (if requested) as needed.
  */
 int
 xfs_qm_dqget(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		id,
-	uint			type,
+	xfs_dqtype_t		type,
 	bool			can_alloc,
 	struct xfs_dquot	**O_dqpp)
 {
@@ -809,7 +864,7 @@ int
 xfs_qm_dqget_uncached(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		id,
-	uint			type,
+	xfs_dqtype_t		type,
 	struct xfs_dquot	**dqpp)
 {
 	int			error;
@@ -825,14 +880,14 @@ xfs_qm_dqget_uncached(
 xfs_dqid_t
 xfs_qm_id_for_quotatype(
 	struct xfs_inode	*ip,
-	uint			type)
+	xfs_dqtype_t		type)
 {
 	switch (type) {
-	case XFS_DQ_USER:
-		return ip->i_d.di_uid;
-	case XFS_DQ_GROUP:
-		return ip->i_d.di_gid;
-	case XFS_DQ_PROJ:
+	case XFS_DQTYPE_USER:
+		return i_uid_read(VFS_I(ip));
+	case XFS_DQTYPE_GROUP:
+		return i_gid_read(VFS_I(ip));
+	case XFS_DQTYPE_PROJ:
 		return ip->i_d.di_projid;
 	}
 	ASSERT(0);
@@ -847,7 +902,7 @@ xfs_qm_id_for_quotatype(
 int
 xfs_qm_dqget_inode(
 	struct xfs_inode	*ip,
-	uint			type,
+	xfs_dqtype_t		type,
 	bool			can_alloc,
 	struct xfs_dquot	**O_dqpp)
 {
@@ -933,7 +988,7 @@ int
 xfs_qm_dqget_next(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		id,
-	uint			type,
+	xfs_dqtype_t		type,
 	struct xfs_dquot	**dqpp)
 {
 	struct xfs_dquot	*dqp;
@@ -1013,14 +1068,14 @@ xfs_qm_dqrele(
  * from the AIL if it has not been re-logged, and unlocking the dquot's
  * flush lock. This behavior is very similar to that of inodes..
  */
-STATIC void
+static void
 xfs_qm_dqflush_done(
-	struct xfs_buf		*bp,
 	struct xfs_log_item	*lip)
 {
 	struct xfs_dq_logitem	*qip = (struct xfs_dq_logitem *)lip;
 	struct xfs_dquot	*dqp = qip->qli_dquot;
 	struct xfs_ail		*ailp = lip->li_ailp;
+	xfs_lsn_t		tail_lsn;
 
 	/*
 	 * We only want to pull the item from the AIL if its
@@ -1034,16 +1089,13 @@ xfs_qm_dqflush_done(
 	    ((lip->li_lsn == qip->qli_flush_lsn) ||
 	     test_bit(XFS_LI_FAILED, &lip->li_flags))) {
 
-		/* xfs_trans_ail_delete() drops the AIL lock. */
 		spin_lock(&ailp->ail_lock);
+		xfs_clear_li_failed(lip);
 		if (lip->li_lsn == qip->qli_flush_lsn) {
-			xfs_trans_ail_delete(ailp, lip, SHUTDOWN_CORRUPT_INCORE);
+			/* xfs_ail_update_finish() drops the AIL lock */
+			tail_lsn = xfs_ail_delete_one(ailp, lip);
+			xfs_ail_update_finish(ailp, tail_lsn);
 		} else {
-			/*
-			 * Clear the failed state since we are about to drop the
-			 * flush lock
-			 */
-			xfs_clear_li_failed(lip);
 			spin_unlock(&ailp->ail_lock);
 		}
 	}
@@ -1054,6 +1106,48 @@ xfs_qm_dqflush_done(
 	xfs_dqfunlock(dqp);
 }
 
+void
+xfs_dquot_done(
+	struct xfs_buf		*bp)
+{
+	struct xfs_log_item	*lip, *n;
+
+	list_for_each_entry_safe(lip, n, &bp->b_li_list, li_bio_list) {
+		list_del_init(&lip->li_bio_list);
+		xfs_qm_dqflush_done(lip);
+	}
+}
+
+/* Check incore dquot for errors before we flush. */
+static xfs_failaddr_t
+xfs_qm_dqflush_check(
+	struct xfs_dquot	*dqp)
+{
+	xfs_dqtype_t		type = xfs_dquot_type(dqp);
+
+	if (type != XFS_DQTYPE_USER &&
+	    type != XFS_DQTYPE_GROUP &&
+	    type != XFS_DQTYPE_PROJ)
+		return __this_address;
+
+	if (dqp->q_id == 0)
+		return NULL;
+
+	if (dqp->q_blk.softlimit && dqp->q_blk.count > dqp->q_blk.softlimit &&
+	    !dqp->q_blk.timer)
+		return __this_address;
+
+	if (dqp->q_ino.softlimit && dqp->q_ino.count > dqp->q_ino.softlimit &&
+	    !dqp->q_ino.timer)
+		return __this_address;
+
+	if (dqp->q_rtb.softlimit && dqp->q_rtb.count > dqp->q_rtb.softlimit &&
+	    !dqp->q_rtb.timer)
+		return __this_address;
+
+	return NULL;
+}
+
 /*
  * Write a modified dquot to disk.
  * The dquot must be locked and the flush lock too taken by caller.
@@ -1068,9 +1162,9 @@ xfs_qm_dqflush(
 	struct xfs_buf		**bpp)
 {
 	struct xfs_mount	*mp = dqp->q_mount;
+	struct xfs_log_item	*lip = &dqp->q_logitem.qli_item;
 	struct xfs_buf		*bp;
-	struct xfs_dqblk	*dqb;
-	struct xfs_disk_dquot	*ddqp;
+	struct xfs_dqblk	*dqblk;
 	xfs_failaddr_t		fa;
 	int			error;
 
@@ -1084,58 +1178,33 @@ xfs_qm_dqflush(
 	xfs_qm_dqunpin_wait(dqp);
 
 	/*
-	 * This may have been unpinned because the filesystem is shutting
-	 * down forcibly. If that's the case we must not write this dquot
-	 * to disk, because the log record didn't make it to disk.
-	 *
-	 * We also have to remove the log item from the AIL in this case,
-	 * as we wait for an emptry AIL as part of the unmount process.
-	 */
-	if (XFS_FORCED_SHUTDOWN(mp)) {
-		struct xfs_log_item	*lip = &dqp->q_logitem.qli_item;
-		dqp->dq_flags &= ~XFS_DQ_DIRTY;
-
-		xfs_trans_ail_remove(lip, SHUTDOWN_CORRUPT_INCORE);
-
-		error = -EIO;
-		goto out_unlock;
-	}
-
-	/*
 	 * Get the buffer containing the on-disk dquot
 	 */
 	error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp, dqp->q_blkno,
-				   mp->m_quotainfo->qi_dqchunklen, 0, &bp,
-				   &xfs_dquot_buf_ops);
-	if (error)
+				   mp->m_quotainfo->qi_dqchunklen, XBF_TRYLOCK,
+				   &bp, &xfs_dquot_buf_ops);
+	if (error == -EAGAIN)
 		goto out_unlock;
+	if (error)
+		goto out_abort;
 
-	/*
-	 * Calculate the location of the dquot inside the buffer.
-	 */
-	dqb = bp->b_addr + dqp->q_bufoffset;
-	ddqp = &dqb->dd_diskdq;
-
-	/*
-	 * A simple sanity check in case we got a corrupted dquot.
-	 */
-	fa = xfs_dqblk_verify(mp, dqb, be32_to_cpu(ddqp->d_id), 0);
+	fa = xfs_qm_dqflush_check(dqp);
 	if (fa) {
 		xfs_alert(mp, "corrupt dquot ID 0x%x in memory at %pS",
-				be32_to_cpu(ddqp->d_id), fa);
+				dqp->q_id, fa);
 		xfs_buf_relse(bp);
-		xfs_dqfunlock(dqp);
-		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
-		return -EFSCORRUPTED;
+		error = -EFSCORRUPTED;
+		goto out_abort;
 	}
 
-	/* This is the only portion of data that needs to persist */
-	memcpy(ddqp, &dqp->q_core, sizeof(struct xfs_disk_dquot));
+	/* Flush the incore dquot to the ondisk buffer. */
+	dqblk = bp->b_addr + dqp->q_bufoffset;
+	xfs_dquot_to_disk(&dqblk->dd_diskdq, dqp);
 
 	/*
 	 * Clear the dirty field and remember the flush lsn for later use.
 	 */
-	dqp->dq_flags &= ~XFS_DQ_DIRTY;
+	dqp->q_flags &= ~XFS_DQFLAG_DIRTY;
 
 	xfs_trans_ail_copy_lsn(mp->m_ail, &dqp->q_logitem.qli_flush_lsn,
 					&dqp->q_logitem.qli_item.li_lsn);
@@ -1150,17 +1219,17 @@ xfs_qm_dqflush(
 	 * of a dquot without an up-to-date CRC getting to disk.
 	 */
 	if (xfs_sb_version_hascrc(&mp->m_sb)) {
-		dqb->dd_lsn = cpu_to_be64(dqp->q_logitem.qli_item.li_lsn);
-		xfs_update_cksum((char *)dqb, sizeof(struct xfs_dqblk),
+		dqblk->dd_lsn = cpu_to_be64(dqp->q_logitem.qli_item.li_lsn);
+		xfs_update_cksum((char *)dqblk, sizeof(struct xfs_dqblk),
 				 XFS_DQUOT_CRC_OFF);
 	}
 
 	/*
-	 * Attach an iodone routine so that we can remove this dquot from the
-	 * AIL and release the flush lock once the dquot is synced to disk.
+	 * Attach the dquot to the buffer so that we can remove this dquot from
+	 * the AIL and release the flush lock once the dquot is synced to disk.
 	 */
-	xfs_buf_attach_iodone(bp, xfs_qm_dqflush_done,
-				  &dqp->q_logitem.qli_item);
+	bp->b_flags |= _XBF_DQUOTS;
+	list_add_tail(&dqp->q_logitem.qli_item.li_bio_list, &bp->b_li_list);
 
 	/*
 	 * If the buffer is pinned then push on the log so we won't
@@ -1175,9 +1244,13 @@ xfs_qm_dqflush(
 	*bpp = bp;
 	return 0;
 
+out_abort:
+	dqp->q_flags &= ~XFS_DQFLAG_DIRTY;
+	xfs_trans_ail_delete(lip, 0);
+	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
 out_unlock:
 	xfs_dqfunlock(dqp);
-	return -EIO;
+	return error;
 }
 
 /*
@@ -1193,8 +1266,7 @@ xfs_dqlock2(
 {
 	if (d1 && d2) {
 		ASSERT(d1 != d2);
-		if (be32_to_cpu(d1->q_core.d_id) >
-		    be32_to_cpu(d2->q_core.d_id)) {
+		if (d1->q_id > d2->q_id) {
 			mutex_lock(&d2->q_qlock);
 			mutex_lock_nested(&d1->q_qlock, XFS_QLOCK_NESTED);
 		} else {
@@ -1246,7 +1318,7 @@ xfs_qm_exit(void)
 int
 xfs_qm_dqiterate(
 	struct xfs_mount	*mp,
-	uint			dqtype,
+	xfs_dqtype_t		type,
 	xfs_qm_dqiterate_fn	iter_fn,
 	void			*priv)
 {
@@ -1255,16 +1327,15 @@ xfs_qm_dqiterate(
 	int			error;
 
 	do {
-		error = xfs_qm_dqget_next(mp, id, dqtype, &dq);
+		error = xfs_qm_dqget_next(mp, id, type, &dq);
 		if (error == -ENOENT)
 			return 0;
 		if (error)
 			return error;
 
-		error = iter_fn(dq, dqtype, priv);
-		id = be32_to_cpu(dq->q_core.d_id);
+		error = iter_fn(dq, type, priv);
+		id = dq->q_id;
 		xfs_qm_dqput(dq);
-		id++;
 	} while (error == 0 && id != 0);
 
 	return error;
diff --git a/fs/xfs/xfs_dquot.h b/fs/xfs/xfs_dquot.h
index fe3e46df604b..282a65da93c7 100644
--- a/fs/xfs/xfs_dquot.h
+++ b/fs/xfs/xfs_dquot.h
@@ -27,26 +27,53 @@ enum {
 	XFS_QLOWSP_MAX
 };
 
+struct xfs_dquot_res {
+	/* Total resources allocated and reserved. */
+	xfs_qcnt_t		reserved;
+
+	/* Total resources allocated. */
+	xfs_qcnt_t		count;
+
+	/* Absolute and preferred limits. */
+	xfs_qcnt_t		hardlimit;
+	xfs_qcnt_t		softlimit;
+
+	/*
+	 * For root dquots, this is the default grace period, in seconds.
+	 * Otherwise, this is when the quota grace period expires,
+	 * in seconds since the Unix epoch.
+	 */
+	time64_t		timer;
+
+	/*
+	 * For root dquots, this is the maximum number of warnings that will
+	 * be issued for this quota type.  Otherwise, this is the number of
+	 * warnings issued against this quota.  Note that none of this is
+	 * implemented.
+	 */
+	xfs_qwarncnt_t		warnings;
+};
+
 /*
  * The incore dquot structure
  */
 struct xfs_dquot {
-	uint			dq_flags;
 	struct list_head	q_lru;
 	struct xfs_mount	*q_mount;
+	xfs_dqtype_t		q_type;
+	uint16_t		q_flags;
+	xfs_dqid_t		q_id;
 	uint			q_nrefs;
-	xfs_daddr_t		q_blkno;
 	int			q_bufoffset;
+	xfs_daddr_t		q_blkno;
 	xfs_fileoff_t		q_fileoffset;
 
-	struct xfs_disk_dquot	q_core;
+	struct xfs_dquot_res	q_blk;	/* regular blocks */
+	struct xfs_dquot_res	q_ino;	/* inodes */
+	struct xfs_dquot_res	q_rtb;	/* realtime blocks */
+
 	struct xfs_dq_logitem	q_logitem;
-	/* total regular nblks used+reserved */
-	xfs_qcnt_t		q_res_bcount;
-	/* total inos allocd+reserved */
-	xfs_qcnt_t		q_res_icount;
-	/* total realtime blks used+reserved */
-	xfs_qcnt_t		q_res_rtbcount;
+
 	xfs_qcnt_t		q_prealloc_lo_wmark;
 	xfs_qcnt_t		q_prealloc_hi_wmark;
 	int64_t			q_low_space[XFS_QLOWSP_MAX];
@@ -101,34 +128,59 @@ static inline void xfs_dqunlock(struct xfs_dquot *dqp)
 	mutex_unlock(&dqp->q_qlock);
 }
 
-static inline int xfs_this_quota_on(struct xfs_mount *mp, int type)
+static inline int
+xfs_dquot_type(const struct xfs_dquot *dqp)
 {
-	switch (type & XFS_DQ_ALLTYPES) {
-	case XFS_DQ_USER:
+	return dqp->q_type & XFS_DQTYPE_REC_MASK;
+}
+
+static inline int xfs_this_quota_on(struct xfs_mount *mp, xfs_dqtype_t type)
+{
+	switch (type) {
+	case XFS_DQTYPE_USER:
 		return XFS_IS_UQUOTA_ON(mp);
-	case XFS_DQ_GROUP:
+	case XFS_DQTYPE_GROUP:
 		return XFS_IS_GQUOTA_ON(mp);
-	case XFS_DQ_PROJ:
+	case XFS_DQTYPE_PROJ:
 		return XFS_IS_PQUOTA_ON(mp);
 	default:
 		return 0;
 	}
 }
 
-static inline struct xfs_dquot *xfs_inode_dquot(struct xfs_inode *ip, int type)
+static inline struct xfs_dquot *xfs_inode_dquot(
+	struct xfs_inode	*ip,
+	xfs_dqtype_t		type)
 {
-	switch (type & XFS_DQ_ALLTYPES) {
-	case XFS_DQ_USER:
+	switch (type) {
+	case XFS_DQTYPE_USER:
 		return ip->i_udquot;
-	case XFS_DQ_GROUP:
+	case XFS_DQTYPE_GROUP:
 		return ip->i_gdquot;
-	case XFS_DQ_PROJ:
+	case XFS_DQTYPE_PROJ:
 		return ip->i_pdquot;
 	default:
 		return NULL;
 	}
 }
 
+/* Decide if the dquot's limits are actually being enforced. */
+static inline bool
+xfs_dquot_is_enforced(
+	const struct xfs_dquot	*dqp)
+{
+	switch (xfs_dquot_type(dqp)) {
+	case XFS_DQTYPE_USER:
+		return XFS_IS_UQUOTA_ENFORCED(dqp->q_mount);
+	case XFS_DQTYPE_GROUP:
+		return XFS_IS_GQUOTA_ENFORCED(dqp->q_mount);
+	case XFS_DQTYPE_PROJ:
+		return XFS_IS_PQUOTA_ENFORCED(dqp->q_mount);
+	}
+	ASSERT(0);
+	return false;
+}
+
 /*
  * Check whether a dquot is under low free space conditions. We assume the quota
  * is enabled and enforced.
@@ -137,38 +189,35 @@ static inline bool xfs_dquot_lowsp(struct xfs_dquot *dqp)
 {
 	int64_t freesp;
 
-	freesp = be64_to_cpu(dqp->q_core.d_blk_hardlimit) - dqp->q_res_bcount;
+	freesp = dqp->q_blk.hardlimit - dqp->q_blk.reserved;
 	if (freesp < dqp->q_low_space[XFS_QLOWSP_1_PCNT])
 		return true;
 
 	return false;
 }
 
+void xfs_dquot_to_disk(struct xfs_disk_dquot *ddqp, struct xfs_dquot *dqp);
+
 #define XFS_DQ_IS_LOCKED(dqp)	(mutex_is_locked(&((dqp)->q_qlock)))
-#define XFS_DQ_IS_DIRTY(dqp)	((dqp)->dq_flags & XFS_DQ_DIRTY)
-#define XFS_QM_ISUDQ(dqp)	((dqp)->dq_flags & XFS_DQ_USER)
-#define XFS_QM_ISPDQ(dqp)	((dqp)->dq_flags & XFS_DQ_PROJ)
-#define XFS_QM_ISGDQ(dqp)	((dqp)->dq_flags & XFS_DQ_GROUP)
+#define XFS_DQ_IS_DIRTY(dqp)	((dqp)->q_flags & XFS_DQFLAG_DIRTY)
 
 void		xfs_qm_dqdestroy(struct xfs_dquot *dqp);
 int		xfs_qm_dqflush(struct xfs_dquot *dqp, struct xfs_buf **bpp);
 void		xfs_qm_dqunpin_wait(struct xfs_dquot *dqp);
-void		xfs_qm_adjust_dqtimers(struct xfs_mount *mp,
-						struct xfs_disk_dquot *d);
-void		xfs_qm_adjust_dqlimits(struct xfs_mount *mp,
-						struct xfs_dquot *d);
-xfs_dqid_t	xfs_qm_id_for_quotatype(struct xfs_inode *ip, uint type);
+void		xfs_qm_adjust_dqtimers(struct xfs_dquot *d);
+void		xfs_qm_adjust_dqlimits(struct xfs_dquot *d);
+xfs_dqid_t	xfs_qm_id_for_quotatype(struct xfs_inode *ip,
+				xfs_dqtype_t type);
 int		xfs_qm_dqget(struct xfs_mount *mp, xfs_dqid_t id,
-					uint type, bool can_alloc,
-					struct xfs_dquot **dqpp);
-int		xfs_qm_dqget_inode(struct xfs_inode *ip, uint type,
-						bool can_alloc,
-						struct xfs_dquot **dqpp);
+				xfs_dqtype_t type, bool can_alloc,
+				struct xfs_dquot **dqpp);
+int		xfs_qm_dqget_inode(struct xfs_inode *ip, xfs_dqtype_t type,
+				bool can_alloc, struct xfs_dquot **dqpp);
 int		xfs_qm_dqget_next(struct xfs_mount *mp, xfs_dqid_t id,
-					uint type, struct xfs_dquot **dqpp);
+				xfs_dqtype_t type, struct xfs_dquot **dqpp);
 int		xfs_qm_dqget_uncached(struct xfs_mount *mp,
-						xfs_dqid_t id, uint type,
-						struct xfs_dquot **dqpp);
+				xfs_dqid_t id, xfs_dqtype_t type,
+				struct xfs_dquot **dqpp);
 void		xfs_qm_dqput(struct xfs_dquot *dqp);
 
 void		xfs_dqlock2(struct xfs_dquot *, struct xfs_dquot *);
@@ -183,9 +232,9 @@ static inline struct xfs_dquot *xfs_qm_dqhold(struct xfs_dquot *dqp)
 	return dqp;
 }
 
-typedef int (*xfs_qm_dqiterate_fn)(struct xfs_dquot *dq, uint dqtype,
-		void *priv);
-int xfs_qm_dqiterate(struct xfs_mount *mp, uint dqtype,
+typedef int (*xfs_qm_dqiterate_fn)(struct xfs_dquot *dq,
+		xfs_dqtype_t type, void *priv);
+int xfs_qm_dqiterate(struct xfs_mount *mp, xfs_dqtype_t type,
 		xfs_qm_dqiterate_fn iter_fn, void *priv);
 
 #endif /* __XFS_DQUOT_H__ */
diff --git a/fs/xfs/xfs_dquot_item.c b/fs/xfs/xfs_dquot_item.c
index d60647d7197b..8c1fdf37ee8f 100644
--- a/fs/xfs/xfs_dquot_item.c
+++ b/fs/xfs/xfs_dquot_item.c
@@ -45,6 +45,7 @@ xfs_qm_dquot_logitem_format(
 	struct xfs_log_item	*lip,
 	struct xfs_log_vec	*lv)
 {
+	struct xfs_disk_dquot	ddq;
 	struct xfs_dq_logitem	*qlip = DQUOT_ITEM(lip);
 	struct xfs_log_iovec	*vecp = NULL;
 	struct xfs_dq_logformat	*qlf;
@@ -52,14 +53,15 @@ xfs_qm_dquot_logitem_format(
 	qlf = xlog_prepare_iovec(lv, &vecp, XLOG_REG_TYPE_QFORMAT);
 	qlf->qlf_type = XFS_LI_DQUOT;
 	qlf->qlf_size = 2;
-	qlf->qlf_id = be32_to_cpu(qlip->qli_dquot->q_core.d_id);
+	qlf->qlf_id = qlip->qli_dquot->q_id;
 	qlf->qlf_blkno = qlip->qli_dquot->q_blkno;
 	qlf->qlf_len = 1;
 	qlf->qlf_boffset = qlip->qli_dquot->q_bufoffset;
 	xlog_finish_iovec(lv, vecp, sizeof(struct xfs_dq_logformat));
 
-	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_DQUOT,
-			&qlip->qli_dquot->q_core,
+	xfs_dquot_to_disk(&ddq, qlip->qli_dquot);
+
+	xlog_copy_iovec(lv, &vecp, XLOG_REG_TYPE_DQUOT, &ddq,
 			sizeof(struct xfs_disk_dquot));
 }
 
@@ -113,23 +115,6 @@ xfs_qm_dqunpin_wait(
 	wait_event(dqp->q_pinwait, (atomic_read(&dqp->q_pincount) == 0));
 }
 
-/*
- * Callback used to mark a buffer with XFS_LI_FAILED when items in the buffer
- * have been failed during writeback
- *
- * this informs the AIL that the dquot is already flush locked on the next push,
- * and acquires a hold on the buffer to ensure that it isn't reclaimed before
- * dirty data makes it to disk.
- */
-STATIC void
-xfs_dquot_item_error(
-	struct xfs_log_item	*lip,
-	struct xfs_buf		*bp)
-{
-	ASSERT(!completion_done(&DQUOT_ITEM(lip)->qli_dquot->q_flush));
-	xfs_set_li_failed(lip, bp);
-}
-
 STATIC uint
 xfs_qm_dquot_logitem_push(
 	struct xfs_log_item	*lip,
@@ -145,21 +130,6 @@ xfs_qm_dquot_logitem_push(
 	if (atomic_read(&dqp->q_pincount) > 0)
 		return XFS_ITEM_PINNED;
 
-	/*
-	 * The buffer containing this item failed to be written back
-	 * previously. Resubmit the buffer for IO
-	 */
-	if (test_bit(XFS_LI_FAILED, &lip->li_flags)) {
-		if (!xfs_buf_trylock(bp))
-			return XFS_ITEM_LOCKED;
-
-		if (!xfs_buf_resubmit_failed_buffers(bp, buffer_list))
-			rval = XFS_ITEM_FLUSHING;
-
-		xfs_buf_unlock(bp);
-		return rval;
-	}
-
 	if (!xfs_dqlock_nowait(dqp))
 		return XFS_ITEM_LOCKED;
 
@@ -189,7 +159,8 @@ xfs_qm_dquot_logitem_push(
 		if (!xfs_buf_delwri_queue(bp, buffer_list))
 			rval = XFS_ITEM_FLUSHING;
 		xfs_buf_relse(bp);
-	}
+	} else if (error == -EAGAIN)
+		rval = XFS_ITEM_LOCKED;
 
 	spin_lock(&lip->li_ailp->ail_lock);
 out_unlock:
@@ -230,7 +201,6 @@ static const struct xfs_item_ops xfs_dquot_item_ops = {
 	.iop_release	= xfs_qm_dquot_logitem_release,
 	.iop_committing	= xfs_qm_dquot_logitem_committing,
 	.iop_push	= xfs_qm_dquot_logitem_push,
-	.iop_error	= xfs_dquot_item_error
 };
 
 /*
@@ -307,36 +277,62 @@ xfs_qm_qoffend_logitem_committed(
 {
 	struct xfs_qoff_logitem	*qfe = QOFF_ITEM(lip);
 	struct xfs_qoff_logitem	*qfs = qfe->qql_start_lip;
-	struct xfs_ail		*ailp = qfs->qql_item.li_ailp;
 
-	/*
-	 * Delete the qoff-start logitem from the AIL.
-	 * xfs_trans_ail_delete() drops the AIL lock.
-	 */
-	spin_lock(&ailp->ail_lock);
-	xfs_trans_ail_delete(ailp, &qfs->qql_item, SHUTDOWN_LOG_IO_ERROR);
+	xfs_qm_qoff_logitem_relse(qfs);
 
-	kmem_free(qfs->qql_item.li_lv_shadow);
 	kmem_free(lip->li_lv_shadow);
-	kmem_free(qfs);
 	kmem_free(qfe);
 	return (xfs_lsn_t)-1;
 }
 
+STATIC void
+xfs_qm_qoff_logitem_release(
+	struct xfs_log_item	*lip)
+{
+	struct xfs_qoff_logitem	*qoff = QOFF_ITEM(lip);
+
+	if (test_bit(XFS_LI_ABORTED, &lip->li_flags)) {
+		if (qoff->qql_start_lip)
+			xfs_qm_qoff_logitem_relse(qoff->qql_start_lip);
+		xfs_qm_qoff_logitem_relse(qoff);
+	}
+}
+
 static const struct xfs_item_ops xfs_qm_qoffend_logitem_ops = {
 	.iop_size	= xfs_qm_qoff_logitem_size,
 	.iop_format	= xfs_qm_qoff_logitem_format,
 	.iop_committed	= xfs_qm_qoffend_logitem_committed,
 	.iop_push	= xfs_qm_qoff_logitem_push,
+	.iop_release	= xfs_qm_qoff_logitem_release,
 };
 
 static const struct xfs_item_ops xfs_qm_qoff_logitem_ops = {
 	.iop_size	= xfs_qm_qoff_logitem_size,
 	.iop_format	= xfs_qm_qoff_logitem_format,
 	.iop_push	= xfs_qm_qoff_logitem_push,
+	.iop_release	= xfs_qm_qoff_logitem_release,
 };
 
 /*
+ * Delete the quotaoff intent from the AIL and free it. On success,
+ * this should only be called for the start item. It can be used for
+ * either on shutdown or abort.
+ */
+void
+xfs_qm_qoff_logitem_relse(
+	struct xfs_qoff_logitem	*qoff)
+{
+	struct xfs_log_item	*lip = &qoff->qql_item;
+
+	ASSERT(test_bit(XFS_LI_IN_AIL, &lip->li_flags) ||
+	       test_bit(XFS_LI_ABORTED, &lip->li_flags) ||
+	       XFS_FORCED_SHUTDOWN(lip->li_mountp));
+	xfs_trans_ail_delete(lip, 0);
+	kmem_free(lip->li_lv_shadow);
+	kmem_free(qoff);
+}
+
+/*
  * Allocate and initialize an quotaoff item of the correct quota type(s).
  */
 struct xfs_qoff_logitem *
diff --git a/fs/xfs/xfs_dquot_item.h b/fs/xfs/xfs_dquot_item.h
index 3bb19e556ade..2b86a43d7ce2 100644
--- a/fs/xfs/xfs_dquot_item.h
+++ b/fs/xfs/xfs_dquot_item.h
@@ -28,6 +28,7 @@ void xfs_qm_dquot_logitem_init(struct xfs_dquot *dqp);
 struct xfs_qoff_logitem	*xfs_qm_qoff_logitem_init(struct xfs_mount *mp,
 		struct xfs_qoff_logitem *start,
 		uint flags);
+void xfs_qm_qoff_logitem_relse(struct xfs_qoff_logitem *);
 struct xfs_qoff_logitem	*xfs_trans_get_qoff_item(struct xfs_trans *tp,
 		struct xfs_qoff_logitem *startqoff,
 		uint flags);
diff --git a/fs/xfs/xfs_dquot_item_recover.c b/fs/xfs/xfs_dquot_item_recover.c
new file mode 100644
index 000000000000..5875c7e1bd28
--- /dev/null
+++ b/fs/xfs/xfs_dquot_item_recover.c
@@ -0,0 +1,201 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_inode.h"
+#include "xfs_quota.h"
+#include "xfs_trans.h"
+#include "xfs_buf_item.h"
+#include "xfs_trans_priv.h"
+#include "xfs_qm.h"
+#include "xfs_log.h"
+#include "xfs_log_priv.h"
+#include "xfs_log_recover.h"
+
+STATIC void
+xlog_recover_dquot_ra_pass2(
+	struct xlog			*log,
+	struct xlog_recover_item	*item)
+{
+	struct xfs_mount	*mp = log->l_mp;
+	struct xfs_disk_dquot	*recddq;
+	struct xfs_dq_logformat	*dq_f;
+	uint			type;
+
+	if (mp->m_qflags == 0)
+		return;
+
+	recddq = item->ri_buf[1].i_addr;
+	if (recddq == NULL)
+		return;
+	if (item->ri_buf[1].i_len < sizeof(struct xfs_disk_dquot))
+		return;
+
+	type = recddq->d_type & XFS_DQTYPE_REC_MASK;
+	ASSERT(type);
+	if (log->l_quotaoffs_flag & type)
+		return;
+
+	dq_f = item->ri_buf[0].i_addr;
+	ASSERT(dq_f);
+	ASSERT(dq_f->qlf_len == 1);
+
+	xlog_buf_readahead(log, dq_f->qlf_blkno,
+			XFS_FSB_TO_BB(mp, dq_f->qlf_len),
+			&xfs_dquot_buf_ra_ops);
+}
+
+/*
+ * Recover a dquot record
+ */
+STATIC int
+xlog_recover_dquot_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			current_lsn)
+{
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_buf			*bp;
+	struct xfs_disk_dquot		*ddq, *recddq;
+	struct xfs_dq_logformat		*dq_f;
+	xfs_failaddr_t			fa;
+	int				error;
+	uint				type;
+
+	/*
+	 * Filesystems are required to send in quota flags at mount time.
+	 */
+	if (mp->m_qflags == 0)
+		return 0;
+
+	recddq = item->ri_buf[1].i_addr;
+	if (recddq == NULL) {
+		xfs_alert(log->l_mp, "NULL dquot in %s.", __func__);
+		return -EFSCORRUPTED;
+	}
+	if (item->ri_buf[1].i_len < sizeof(struct xfs_disk_dquot)) {
+		xfs_alert(log->l_mp, "dquot too small (%d) in %s.",
+			item->ri_buf[1].i_len, __func__);
+		return -EFSCORRUPTED;
+	}
+
+	/*
+	 * This type of quotas was turned off, so ignore this record.
+	 */
+	type = recddq->d_type & XFS_DQTYPE_REC_MASK;
+	ASSERT(type);
+	if (log->l_quotaoffs_flag & type)
+		return 0;
+
+	/*
+	 * At this point we know that quota was _not_ turned off.
+	 * Since the mount flags are not indicating to us otherwise, this
+	 * must mean that quota is on, and the dquot needs to be replayed.
+	 * Remember that we may not have fully recovered the superblock yet,
+	 * so we can't do the usual trick of looking at the SB quota bits.
+	 *
+	 * The other possibility, of course, is that the quota subsystem was
+	 * removed since the last mount - ENOSYS.
+	 */
+	dq_f = item->ri_buf[0].i_addr;
+	ASSERT(dq_f);
+	fa = xfs_dquot_verify(mp, recddq, dq_f->qlf_id);
+	if (fa) {
+		xfs_alert(mp, "corrupt dquot ID 0x%x in log at %pS",
+				dq_f->qlf_id, fa);
+		return -EFSCORRUPTED;
+	}
+	ASSERT(dq_f->qlf_len == 1);
+
+	/*
+	 * At this point we are assuming that the dquots have been allocated
+	 * and hence the buffer has valid dquots stamped in it. It should,
+	 * therefore, pass verifier validation. If the dquot is bad, then the
+	 * we'll return an error here, so we don't need to specifically check
+	 * the dquot in the buffer after the verifier has run.
+	 */
+	error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp, dq_f->qlf_blkno,
+				   XFS_FSB_TO_BB(mp, dq_f->qlf_len), 0, &bp,
+				   &xfs_dquot_buf_ops);
+	if (error)
+		return error;
+
+	ASSERT(bp);
+	ddq = xfs_buf_offset(bp, dq_f->qlf_boffset);
+
+	/*
+	 * If the dquot has an LSN in it, recover the dquot only if it's less
+	 * than the lsn of the transaction we are replaying.
+	 */
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		struct xfs_dqblk *dqb = (struct xfs_dqblk *)ddq;
+		xfs_lsn_t	lsn = be64_to_cpu(dqb->dd_lsn);
+
+		if (lsn && lsn != -1 && XFS_LSN_CMP(lsn, current_lsn) >= 0) {
+			goto out_release;
+		}
+	}
+
+	memcpy(ddq, recddq, item->ri_buf[1].i_len);
+	if (xfs_sb_version_hascrc(&mp->m_sb)) {
+		xfs_update_cksum((char *)ddq, sizeof(struct xfs_dqblk),
+				 XFS_DQUOT_CRC_OFF);
+	}
+
+	ASSERT(dq_f->qlf_size == 2);
+	ASSERT(bp->b_mount == mp);
+	bp->b_flags |= _XBF_LOGRECOVERY;
+	xfs_buf_delwri_queue(bp, buffer_list);
+
+out_release:
+	xfs_buf_relse(bp);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_dquot_item_ops = {
+	.item_type		= XFS_LI_DQUOT,
+	.ra_pass2		= xlog_recover_dquot_ra_pass2,
+	.commit_pass2		= xlog_recover_dquot_commit_pass2,
+};
+
+/*
+ * Recover QUOTAOFF records. We simply make a note of it in the xlog
+ * structure, so that we know not to do any dquot item or dquot buffer recovery,
+ * of that type.
+ */
+STATIC int
+xlog_recover_quotaoff_commit_pass1(
+	struct xlog			*log,
+	struct xlog_recover_item	*item)
+{
+	struct xfs_qoff_logformat	*qoff_f = item->ri_buf[0].i_addr;
+	ASSERT(qoff_f);
+
+	/*
+	 * The logitem format's flag tells us if this was user quotaoff,
+	 * group/project quotaoff or both.
+	 */
+	if (qoff_f->qf_flags & XFS_UQUOTA_ACCT)
+		log->l_quotaoffs_flag |= XFS_DQTYPE_USER;
+	if (qoff_f->qf_flags & XFS_PQUOTA_ACCT)
+		log->l_quotaoffs_flag |= XFS_DQTYPE_PROJ;
+	if (qoff_f->qf_flags & XFS_GQUOTA_ACCT)
+		log->l_quotaoffs_flag |= XFS_DQTYPE_GROUP;
+
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_quotaoff_item_ops = {
+	.item_type		= XFS_LI_QUOTAOFF,
+	.commit_pass1		= xlog_recover_quotaoff_commit_pass1,
+	/* nothing to commit in pass2 */
+};
diff --git a/fs/xfs/xfs_error.c b/fs/xfs/xfs_error.c
index 331765afc53e..7f6e20899473 100644
--- a/fs/xfs/xfs_error.c
+++ b/fs/xfs/xfs_error.c
@@ -53,6 +53,7 @@ static unsigned int xfs_errortag_random_default[] = {
 	XFS_RANDOM_FORCE_SCRUB_REPAIR,
 	XFS_RANDOM_FORCE_SUMMARY_RECALC,
 	XFS_RANDOM_IUNLINK_FALLBACK,
+	XFS_RANDOM_BUF_IOERROR,
 };
 
 struct xfs_errortag_attr {
@@ -162,6 +163,7 @@ XFS_ERRORTAG_ATTR_RW(buf_lru_ref,	XFS_ERRTAG_BUF_LRU_REF);
 XFS_ERRORTAG_ATTR_RW(force_repair,	XFS_ERRTAG_FORCE_SCRUB_REPAIR);
 XFS_ERRORTAG_ATTR_RW(bad_summary,	XFS_ERRTAG_FORCE_SUMMARY_RECALC);
 XFS_ERRORTAG_ATTR_RW(iunlink_fallback,	XFS_ERRTAG_IUNLINK_FALLBACK);
+XFS_ERRORTAG_ATTR_RW(buf_ioerror,	XFS_ERRTAG_BUF_IOERROR);
 
 static struct attribute *xfs_errortag_attrs[] = {
 	XFS_ERRORTAG_ATTR_LIST(noerror),
@@ -199,6 +201,7 @@ static struct attribute *xfs_errortag_attrs[] = {
 	XFS_ERRORTAG_ATTR_LIST(force_repair),
 	XFS_ERRORTAG_ATTR_LIST(bad_summary),
 	XFS_ERRORTAG_ATTR_LIST(iunlink_fallback),
+	XFS_ERRORTAG_ATTR_LIST(buf_ioerror),
 	NULL,
 };
 
@@ -345,16 +348,19 @@ xfs_corruption_error(
  * Complain about the kinds of metadata corruption that we can't detect from a
  * verifier, such as incorrect inter-block relationship data.  Does not set
  * bp->b_error.
+ *
+ * Call xfs_buf_mark_corrupt, not this function.
  */
 void
 xfs_buf_corruption_error(
-	struct xfs_buf		*bp)
+	struct xfs_buf		*bp,
+	xfs_failaddr_t		fa)
 {
 	struct xfs_mount	*mp = bp->b_mount;
 
 	xfs_alert_tag(mp, XFS_PTAG_VERIFIER_ERROR,
 		  "Metadata corruption detected at %pS, %s block 0x%llx",
-		  __return_address, bp->b_ops->name, bp->b_bn);
+		  fa, bp->b_ops->name, bp->b_bn);
 
 	xfs_alert(mp, "Unmount and run xfs_repair");
 
diff --git a/fs/xfs/xfs_error.h b/fs/xfs/xfs_error.h
index 31a5d321ba9a..1717b7508356 100644
--- a/fs/xfs/xfs_error.h
+++ b/fs/xfs/xfs_error.h
@@ -15,7 +15,7 @@ extern void xfs_corruption_error(const char *tag, int level,
 			struct xfs_mount *mp, const void *buf, size_t bufsize,
 			const char *filename, int linenum,
 			xfs_failaddr_t failaddr);
-void xfs_buf_corruption_error(struct xfs_buf *bp);
+void xfs_buf_corruption_error(struct xfs_buf *bp, xfs_failaddr_t fa);
 extern void xfs_buf_verifier_error(struct xfs_buf *bp, int error,
 			const char *name, const void *buf, size_t bufsz,
 			xfs_failaddr_t failaddr);
diff --git a/fs/xfs/xfs_export.c b/fs/xfs/xfs_export.c
index f1372f9046e3..465fd9e048d4 100644
--- a/fs/xfs/xfs_export.c
+++ b/fs/xfs/xfs_export.c
@@ -15,7 +15,6 @@
 #include "xfs_trans.h"
 #include "xfs_inode_item.h"
 #include "xfs_icache.h"
-#include "xfs_log.h"
 #include "xfs_pnfs.h"
 
 /*
@@ -57,7 +56,7 @@ xfs_fs_encode_fh(
 		fileid_type = FILEID_INO32_GEN_PARENT;
 
 	/*
-	 * If the the filesystem may contain 64bit inode numbers, we need
+	 * If the filesystem may contain 64bit inode numbers, we need
 	 * to use larger file handles that can represent them.
 	 *
 	 * While we only allocate inodes that do not fit into 32 bits any
@@ -221,18 +220,7 @@ STATIC int
 xfs_fs_nfs_commit_metadata(
 	struct inode		*inode)
 {
-	struct xfs_inode	*ip = XFS_I(inode);
-	struct xfs_mount	*mp = ip->i_mount;
-	xfs_lsn_t		lsn = 0;
-
-	xfs_ilock(ip, XFS_ILOCK_SHARED);
-	if (xfs_ipincount(ip))
-		lsn = ip->i_itemp->ili_last_lsn;
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
-
-	if (!lsn)
-		return 0;
-	return xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, NULL);
+	return xfs_log_force_inode(XFS_I(inode));
 }
 
 const struct export_operations xfs_export_operations = {
diff --git a/fs/xfs/xfs_extfree_item.c b/fs/xfs/xfs_extfree_item.c
index 6ea847f6e298..6cb8cd11072a 100644
--- a/fs/xfs/xfs_extfree_item.c
+++ b/fs/xfs/xfs_extfree_item.c
@@ -22,16 +22,20 @@
 #include "xfs_bmap.h"
 #include "xfs_trace.h"
 #include "xfs_error.h"
+#include "xfs_log_priv.h"
+#include "xfs_log_recover.h"
 
 kmem_zone_t	*xfs_efi_zone;
 kmem_zone_t	*xfs_efd_zone;
 
+static const struct xfs_item_ops xfs_efi_item_ops;
+
 static inline struct xfs_efi_log_item *EFI_ITEM(struct xfs_log_item *lip)
 {
 	return container_of(lip, struct xfs_efi_log_item, efi_item);
 }
 
-void
+STATIC void
 xfs_efi_item_free(
 	struct xfs_efi_log_item	*efip)
 {
@@ -49,13 +53,13 @@ xfs_efi_item_free(
  * committed vs unpin operations in bulk insert operations. Hence the reference
  * count to ensure only the last caller frees the EFI.
  */
-void
+STATIC void
 xfs_efi_release(
 	struct xfs_efi_log_item	*efip)
 {
 	ASSERT(atomic_read(&efip->efi_refcount) > 0);
 	if (atomic_dec_and_test(&efip->efi_refcount)) {
-		xfs_trans_ail_remove(&efip->efi_item, SHUTDOWN_LOG_IO_ERROR);
+		xfs_trans_ail_delete(&efip->efi_item, SHUTDOWN_LOG_IO_ERROR);
 		xfs_efi_item_free(efip);
 	}
 }
@@ -139,18 +143,10 @@ xfs_efi_item_release(
 	xfs_efi_release(EFI_ITEM(lip));
 }
 
-static const struct xfs_item_ops xfs_efi_item_ops = {
-	.iop_size	= xfs_efi_item_size,
-	.iop_format	= xfs_efi_item_format,
-	.iop_unpin	= xfs_efi_item_unpin,
-	.iop_release	= xfs_efi_item_release,
-};
-
-
 /*
  * Allocate and initialize an efi item with the given number of extents.
  */
-struct xfs_efi_log_item *
+STATIC struct xfs_efi_log_item *
 xfs_efi_init(
 	struct xfs_mount	*mp,
 	uint			nextents)
@@ -161,11 +157,12 @@ xfs_efi_init(
 
 	ASSERT(nextents > 0);
 	if (nextents > XFS_EFI_MAX_FAST_EXTENTS) {
-		size = (uint)(sizeof(xfs_efi_log_item_t) +
+		size = (uint)(sizeof(struct xfs_efi_log_item) +
 			((nextents - 1) * sizeof(xfs_extent_t)));
 		efip = kmem_zalloc(size, 0);
 	} else {
-		efip = kmem_zone_zalloc(xfs_efi_zone, 0);
+		efip = kmem_cache_zalloc(xfs_efi_zone,
+					 GFP_KERNEL | __GFP_NOFAIL);
 	}
 
 	xfs_log_item_init(mp, &efip->efi_item, XFS_LI_EFI, &xfs_efi_item_ops);
@@ -184,7 +181,7 @@ xfs_efi_init(
  * one of which will be the native format for this kernel.
  * It will handle the conversion of formats if necessary.
  */
-int
+STATIC int
 xfs_efi_copy_format(xfs_log_iovec_t *buf, xfs_efi_log_format_t *dst_efi_fmt)
 {
 	xfs_efi_log_format_t *src_efi_fmt = buf->i_addr;
@@ -336,7 +333,8 @@ xfs_trans_get_efd(
 				(nextents - 1) * sizeof(struct xfs_extent),
 				0);
 	} else {
-		efdp = kmem_zone_zalloc(xfs_efd_zone, 0);
+		efdp = kmem_cache_zalloc(xfs_efd_zone,
+					GFP_KERNEL | __GFP_NOFAIL);
 	}
 
 	xfs_log_item_init(tp->t_mountp, &efdp->efd_item, XFS_LI_EFD,
@@ -412,41 +410,16 @@ xfs_extent_free_diff_items(
 		XFS_FSB_TO_AGNO(mp, rb->xefi_startblock);
 }
 
-/* Get an EFI. */
-STATIC void *
-xfs_extent_free_create_intent(
-	struct xfs_trans		*tp,
-	unsigned int			count)
-{
-	struct xfs_efi_log_item		*efip;
-
-	ASSERT(tp != NULL);
-	ASSERT(count > 0);
-
-	efip = xfs_efi_init(tp->t_mountp, count);
-	ASSERT(efip != NULL);
-
-	/*
-	 * Get a log_item_desc to point at the new item.
-	 */
-	xfs_trans_add_item(tp, &efip->efi_item);
-	return efip;
-}
-
 /* Log a free extent to the intent item. */
 STATIC void
 xfs_extent_free_log_item(
 	struct xfs_trans		*tp,
-	void				*intent,
-	struct list_head		*item)
+	struct xfs_efi_log_item		*efip,
+	struct xfs_extent_free_item	*free)
 {
-	struct xfs_efi_log_item		*efip = intent;
-	struct xfs_extent_free_item	*free;
 	uint				next_extent;
 	struct xfs_extent		*extp;
 
-	free = container_of(item, struct xfs_extent_free_item, xefi_list);
-
 	tp->t_flags |= XFS_TRANS_DIRTY;
 	set_bit(XFS_LI_DIRTY, &efip->efi_item.li_flags);
 
@@ -462,29 +435,50 @@ xfs_extent_free_log_item(
 	extp->ext_len = free->xefi_blockcount;
 }
 
+static struct xfs_log_item *
+xfs_extent_free_create_intent(
+	struct xfs_trans		*tp,
+	struct list_head		*items,
+	unsigned int			count,
+	bool				sort)
+{
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_efi_log_item		*efip = xfs_efi_init(mp, count);
+	struct xfs_extent_free_item	*free;
+
+	ASSERT(count > 0);
+
+	xfs_trans_add_item(tp, &efip->efi_item);
+	if (sort)
+		list_sort(mp, items, xfs_extent_free_diff_items);
+	list_for_each_entry(free, items, xefi_list)
+		xfs_extent_free_log_item(tp, efip, free);
+	return &efip->efi_item;
+}
+
 /* Get an EFD so we can process all the free extents. */
-STATIC void *
+static struct xfs_log_item *
 xfs_extent_free_create_done(
 	struct xfs_trans		*tp,
-	void				*intent,
+	struct xfs_log_item		*intent,
 	unsigned int			count)
 {
-	return xfs_trans_get_efd(tp, intent, count);
+	return &xfs_trans_get_efd(tp, EFI_ITEM(intent), count)->efd_item;
 }
 
 /* Process a free extent. */
 STATIC int
 xfs_extent_free_finish_item(
 	struct xfs_trans		*tp,
+	struct xfs_log_item		*done,
 	struct list_head		*item,
-	void				*done_item,
-	void				**state)
+	struct xfs_btree_cur		**state)
 {
 	struct xfs_extent_free_item	*free;
 	int				error;
 
 	free = container_of(item, struct xfs_extent_free_item, xefi_list);
-	error = xfs_trans_free_extent(tp, done_item,
+	error = xfs_trans_free_extent(tp, EFD_ITEM(done),
 			free->xefi_startblock,
 			free->xefi_blockcount,
 			&free->xefi_oinfo, free->xefi_skip_discard);
@@ -495,9 +489,9 @@ xfs_extent_free_finish_item(
 /* Abort all pending EFIs. */
 STATIC void
 xfs_extent_free_abort_intent(
-	void				*intent)
+	struct xfs_log_item		*intent)
 {
-	xfs_efi_release(intent);
+	xfs_efi_release(EFI_ITEM(intent));
 }
 
 /* Cancel a free extent. */
@@ -513,10 +507,8 @@ xfs_extent_free_cancel_item(
 
 const struct xfs_defer_op_type xfs_extent_free_defer_type = {
 	.max_items	= XFS_EFI_MAX_FAST_EXTENTS,
-	.diff_items	= xfs_extent_free_diff_items,
 	.create_intent	= xfs_extent_free_create_intent,
 	.abort_intent	= xfs_extent_free_abort_intent,
-	.log_item	= xfs_extent_free_log_item,
 	.create_done	= xfs_extent_free_create_done,
 	.finish_item	= xfs_extent_free_finish_item,
 	.cancel_item	= xfs_extent_free_cancel_item,
@@ -529,12 +521,12 @@ const struct xfs_defer_op_type xfs_extent_free_defer_type = {
 STATIC int
 xfs_agfl_free_finish_item(
 	struct xfs_trans		*tp,
+	struct xfs_log_item		*done,
 	struct list_head		*item,
-	void				*done_item,
-	void				**state)
+	struct xfs_btree_cur		**state)
 {
 	struct xfs_mount		*mp = tp->t_mountp;
-	struct xfs_efd_log_item		*efdp = done_item;
+	struct xfs_efd_log_item		*efdp = EFD_ITEM(done);
 	struct xfs_extent_free_item	*free;
 	struct xfs_extent		*extp;
 	struct xfs_buf			*agbp;
@@ -579,10 +571,8 @@ xfs_agfl_free_finish_item(
 /* sub-type with special handling for AGFL deferred frees */
 const struct xfs_defer_op_type xfs_agfl_free_defer_type = {
 	.max_items	= XFS_EFI_MAX_FAST_EXTENTS,
-	.diff_items	= xfs_extent_free_diff_items,
 	.create_intent	= xfs_extent_free_create_intent,
 	.abort_intent	= xfs_extent_free_abort_intent,
-	.log_item	= xfs_extent_free_log_item,
 	.create_done	= xfs_extent_free_create_done,
 	.finish_item	= xfs_agfl_free_finish_item,
 	.cancel_item	= xfs_extent_free_cancel_item,
@@ -592,19 +582,19 @@ const struct xfs_defer_op_type xfs_agfl_free_defer_type = {
  * Process an extent free intent item that was recovered from
  * the log.  We need to free the extents that it describes.
  */
-int
-xfs_efi_recover(
-	struct xfs_mount	*mp,
-	struct xfs_efi_log_item	*efip)
+STATIC int
+xfs_efi_item_recover(
+	struct xfs_log_item		*lip,
+	struct xfs_trans		*parent_tp)
 {
-	struct xfs_efd_log_item	*efdp;
-	struct xfs_trans	*tp;
-	int			i;
-	int			error = 0;
-	xfs_extent_t		*extp;
-	xfs_fsblock_t		startblock_fsb;
-
-	ASSERT(!test_bit(XFS_EFI_RECOVERED, &efip->efi_flags));
+	struct xfs_efi_log_item		*efip = EFI_ITEM(lip);
+	struct xfs_mount		*mp = parent_tp->t_mountp;
+	struct xfs_efd_log_item		*efdp;
+	struct xfs_trans		*tp;
+	struct xfs_extent		*extp;
+	xfs_fsblock_t			startblock_fsb;
+	int				i;
+	int				error = 0;
 
 	/*
 	 * First check the validity of the extents described by the
@@ -623,7 +613,6 @@ xfs_efi_recover(
 			 * This will pull the EFI from the AIL and
 			 * free the memory associated with it.
 			 */
-			set_bit(XFS_EFI_RECOVERED, &efip->efi_flags);
 			xfs_efi_release(efip);
 			return -EFSCORRUPTED;
 		}
@@ -644,7 +633,6 @@ xfs_efi_recover(
 
 	}
 
-	set_bit(XFS_EFI_RECOVERED, &efip->efi_flags);
 	error = xfs_trans_commit(tp);
 	return error;
 
@@ -652,3 +640,93 @@ abort_error:
 	xfs_trans_cancel(tp);
 	return error;
 }
+
+STATIC bool
+xfs_efi_item_match(
+	struct xfs_log_item	*lip,
+	uint64_t		intent_id)
+{
+	return EFI_ITEM(lip)->efi_format.efi_id == intent_id;
+}
+
+static const struct xfs_item_ops xfs_efi_item_ops = {
+	.iop_size	= xfs_efi_item_size,
+	.iop_format	= xfs_efi_item_format,
+	.iop_unpin	= xfs_efi_item_unpin,
+	.iop_release	= xfs_efi_item_release,
+	.iop_recover	= xfs_efi_item_recover,
+	.iop_match	= xfs_efi_item_match,
+};
+
+/*
+ * This routine is called to create an in-core extent free intent
+ * item from the efi format structure which was logged on disk.
+ * It allocates an in-core efi, copies the extents from the format
+ * structure into it, and adds the efi to the AIL with the given
+ * LSN.
+ */
+STATIC int
+xlog_recover_efi_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_efi_log_item		*efip;
+	struct xfs_efi_log_format	*efi_formatp;
+	int				error;
+
+	efi_formatp = item->ri_buf[0].i_addr;
+
+	efip = xfs_efi_init(mp, efi_formatp->efi_nextents);
+	error = xfs_efi_copy_format(&item->ri_buf[0], &efip->efi_format);
+	if (error) {
+		xfs_efi_item_free(efip);
+		return error;
+	}
+	atomic_set(&efip->efi_next_extent, efi_formatp->efi_nextents);
+	/*
+	 * Insert the intent into the AIL directly and drop one reference so
+	 * that finishing or canceling the work will drop the other.
+	 */
+	xfs_trans_ail_insert(log->l_ailp, &efip->efi_item, lsn);
+	xfs_efi_release(efip);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_efi_item_ops = {
+	.item_type		= XFS_LI_EFI,
+	.commit_pass2		= xlog_recover_efi_commit_pass2,
+};
+
+/*
+ * This routine is called when an EFD format structure is found in a committed
+ * transaction in the log. Its purpose is to cancel the corresponding EFI if it
+ * was still in the log. To do this it searches the AIL for the EFI with an id
+ * equal to that in the EFD format structure. If we find it we drop the EFD
+ * reference, which removes the EFI from the AIL and frees it.
+ */
+STATIC int
+xlog_recover_efd_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_efd_log_format	*efd_formatp;
+
+	efd_formatp = item->ri_buf[0].i_addr;
+	ASSERT((item->ri_buf[0].i_len == (sizeof(xfs_efd_log_format_32_t) +
+		((efd_formatp->efd_nextents - 1) * sizeof(xfs_extent_32_t)))) ||
+	       (item->ri_buf[0].i_len == (sizeof(xfs_efd_log_format_64_t) +
+		((efd_formatp->efd_nextents - 1) * sizeof(xfs_extent_64_t)))));
+
+	xlog_recover_release_intent(log, XFS_LI_EFI, efd_formatp->efd_efi_id);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_efd_item_ops = {
+	.item_type		= XFS_LI_EFD,
+	.commit_pass2		= xlog_recover_efd_commit_pass2,
+};
diff --git a/fs/xfs/xfs_extfree_item.h b/fs/xfs/xfs_extfree_item.h
index 16aaab06d4ec..cd2860c875bf 100644
--- a/fs/xfs/xfs_extfree_item.h
+++ b/fs/xfs/xfs_extfree_item.h
@@ -17,11 +17,6 @@ struct kmem_zone;
 #define	XFS_EFI_MAX_FAST_EXTENTS	16
 
 /*
- * Define EFI flag bits. Manipulated by set/clear/test_bit operators.
- */
-#define	XFS_EFI_RECOVERED	1
-
-/*
  * This is the "extent free intention" log item.  It is used to log the fact
  * that some extents need to be free.  It is used in conjunction with the
  * "extent free done" log item described below.
@@ -50,25 +45,24 @@ struct kmem_zone;
  * of commit failure or log I/O errors. Note that the EFD is not inserted in the
  * AIL, so at this point both the EFI and EFD are freed.
  */
-typedef struct xfs_efi_log_item {
+struct xfs_efi_log_item {
 	struct xfs_log_item	efi_item;
 	atomic_t		efi_refcount;
 	atomic_t		efi_next_extent;
-	unsigned long		efi_flags;	/* misc flags */
 	xfs_efi_log_format_t	efi_format;
-} xfs_efi_log_item_t;
+};
 
 /*
  * This is the "extent free done" log item.  It is used to log
  * the fact that some extents earlier mentioned in an efi item
  * have been freed.
  */
-typedef struct xfs_efd_log_item {
+struct xfs_efd_log_item {
 	struct xfs_log_item	efd_item;
-	xfs_efi_log_item_t	*efd_efip;
+	struct xfs_efi_log_item *efd_efip;
 	uint			efd_next_extent;
 	xfs_efd_log_format_t	efd_format;
-} xfs_efd_log_item_t;
+};
 
 /*
  * Max number of extents in fast allocation path.
@@ -78,13 +72,4 @@ typedef struct xfs_efd_log_item {
 extern struct kmem_zone	*xfs_efi_zone;
 extern struct kmem_zone	*xfs_efd_zone;
 
-xfs_efi_log_item_t	*xfs_efi_init(struct xfs_mount *, uint);
-int			xfs_efi_copy_format(xfs_log_iovec_t *buf,
-					    xfs_efi_log_format_t *dst_efi_fmt);
-void			xfs_efi_item_free(xfs_efi_log_item_t *);
-void			xfs_efi_release(struct xfs_efi_log_item *);
-
-int			xfs_efi_recover(struct xfs_mount *mp,
-					struct xfs_efi_log_item *efip);
-
 #endif	/* __XFS_EXTFREE_ITEM_H__ */
diff --git a/fs/xfs/xfs_file.c b/fs/xfs/xfs_file.c
index b8a4a3f29b36..c31cd3be9fb2 100644
--- a/fs/xfs/xfs_file.c
+++ b/fs/xfs/xfs_file.c
@@ -80,19 +80,9 @@ xfs_dir_fsync(
 	int			datasync)
 {
 	struct xfs_inode	*ip = XFS_I(file->f_mapping->host);
-	struct xfs_mount	*mp = ip->i_mount;
-	xfs_lsn_t		lsn = 0;
 
 	trace_xfs_dir_fsync(ip);
-
-	xfs_ilock(ip, XFS_ILOCK_SHARED);
-	if (xfs_ipincount(ip))
-		lsn = ip->i_itemp->ili_last_lsn;
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
-
-	if (!lsn)
-		return 0;
-	return xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, NULL);
+	return xfs_log_force_inode(ip);
 }
 
 STATIC int
@@ -104,6 +94,7 @@ xfs_file_fsync(
 {
 	struct inode		*inode = file->f_mapping->host;
 	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_inode_log_item *iip = ip->i_itemp;
 	struct xfs_mount	*mp = ip->i_mount;
 	int			error = 0;
 	int			log_flushed = 0;
@@ -147,13 +138,15 @@ xfs_file_fsync(
 	xfs_ilock(ip, XFS_ILOCK_SHARED);
 	if (xfs_ipincount(ip)) {
 		if (!datasync ||
-		    (ip->i_itemp->ili_fsync_fields & ~XFS_ILOG_TIMESTAMP))
-			lsn = ip->i_itemp->ili_last_lsn;
+		    (iip->ili_fsync_fields & ~XFS_ILOG_TIMESTAMP))
+			lsn = iip->ili_last_lsn;
 	}
 
 	if (lsn) {
 		error = xfs_log_force_lsn(mp, lsn, XFS_LOG_SYNC, &log_flushed);
-		ip->i_itemp->ili_fsync_fields = 0;
+		spin_lock(&iip->ili_lock);
+		iip->ili_fsync_fields = 0;
+		spin_unlock(&iip->ili_lock);
 	}
 	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 
@@ -515,7 +508,7 @@ xfs_file_dio_aio_write(
 		 */
 		if (xfs_is_cow_inode(ip)) {
 			trace_xfs_reflink_bounce_dio_write(ip, iocb->ki_pos, count);
-			return -EREMCHG;
+			return -ENOTBLK;
 		}
 		iolock = XFS_IOLOCK_EXCL;
 	} else {
@@ -563,8 +556,8 @@ out:
 	xfs_iunlock(ip, iolock);
 
 	/*
-	 * No fallback to buffered IO on errors for XFS, direct IO will either
-	 * complete fully or fail.
+	 * No fallback to buffered IO after short writes for XFS, direct I/O
+	 * will either complete fully or return an error.
 	 */
 	ASSERT(ret < 0 || ret == count);
 	return ret;
@@ -724,7 +717,7 @@ xfs_file_write_iter(
 		 * allow an operation to fall back to buffered mode.
 		 */
 		ret = xfs_file_dio_aio_write(iocb, from);
-		if (ret != -EREMCHG)
+		if (ret != -ENOTBLK)
 			return ret;
 	}
 
@@ -1045,7 +1038,7 @@ xfs_file_remap_range(
 	/* Prepare and then clone file data. */
 	ret = xfs_reflink_remap_prep(file_in, pos_in, file_out, pos_out,
 			&len, remap_flags);
-	if (ret < 0 || len == 0)
+	if (ret || len == 0)
 		return ret;
 
 	trace_xfs_reflink_remap_range(src, pos_in, len, dest, pos_out);
@@ -1069,9 +1062,13 @@ xfs_file_remap_range(
 
 	ret = xfs_reflink_update_dest(dest, pos_out + len, cowextsize,
 			remap_flags);
+	if (ret)
+		goto out_unlock;
 
+	if (mp->m_flags & XFS_MOUNT_WSYNC)
+		xfs_log_force_inode(dest);
 out_unlock:
-	xfs_reflink_remap_unlock(file_in, file_out);
+	xfs_iunlock2_io_mmap(src, dest);
 	if (ret)
 		trace_xfs_reflink_remap_range_error(dest, ret, _RET_IP_);
 	return remapped > 0 ? remapped : ret;
@@ -1086,7 +1083,7 @@ xfs_file_open(
 		return -EFBIG;
 	if (XFS_FORCED_SHUTDOWN(XFS_M(inode->i_sb)))
 		return -EIO;
-	file->f_mode |= FMODE_NOWAIT;
+	file->f_mode |= FMODE_NOWAIT | FMODE_BUF_RASYNC;
 	return 0;
 }
 
@@ -1108,7 +1105,7 @@ xfs_dir_open(
 	 * certain to have the next operation be a read there.
 	 */
 	mode = xfs_ilock_data_map_shared(ip);
-	if (ip->i_d.di_nextents > 0)
+	if (ip->i_df.if_nextents > 0)
 		error = xfs_dir3_data_readahead(ip, 0, 0);
 	xfs_iunlock(ip, mode);
 	return error;
@@ -1179,7 +1176,7 @@ xfs_file_llseek(
  * Locking for serialisation of IO during page faults. This results in a lock
  * ordering of:
  *
- * mmap_sem (MM)
+ * mmap_lock (MM)
  *   sb_start_pagefault(vfs, freeze)
  *     i_mmaplock (XFS - truncate serialisation)
  *       page_lock (MM)
@@ -1269,10 +1266,23 @@ xfs_filemap_pfn_mkwrite(
 	return __xfs_filemap_fault(vmf, PE_SIZE_PTE, true);
 }
 
+static void
+xfs_filemap_map_pages(
+	struct vm_fault		*vmf,
+	pgoff_t			start_pgoff,
+	pgoff_t			end_pgoff)
+{
+	struct inode		*inode = file_inode(vmf->vma->vm_file);
+
+	xfs_ilock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
+	filemap_map_pages(vmf, start_pgoff, end_pgoff);
+	xfs_iunlock(XFS_I(inode), XFS_MMAPLOCK_SHARED);
+}
+
 static const struct vm_operations_struct xfs_file_vm_ops = {
 	.fault		= xfs_filemap_fault,
 	.huge_fault	= xfs_filemap_huge_fault,
-	.map_pages	= filemap_map_pages,
+	.map_pages	= xfs_filemap_map_pages,
 	.page_mkwrite	= xfs_filemap_page_mkwrite,
 	.pfn_mkwrite	= xfs_filemap_pfn_mkwrite,
 };
diff --git a/fs/xfs/xfs_fsmap.c b/fs/xfs/xfs_fsmap.c
index 918456ca29e1..4eebcec4aae6 100644
--- a/fs/xfs/xfs_fsmap.c
+++ b/fs/xfs/xfs_fsmap.c
@@ -344,7 +344,7 @@ xfs_getfsmap_datadev_helper(
 	xfs_fsblock_t			fsb;
 	xfs_daddr_t			rec_daddr;
 
-	fsb = XFS_AGB_TO_FSB(mp, cur->bc_private.a.agno, rec->rm_startblock);
+	fsb = XFS_AGB_TO_FSB(mp, cur->bc_ag.agno, rec->rm_startblock);
 	rec_daddr = XFS_FSB_TO_DADDR(mp, fsb);
 
 	return xfs_getfsmap_helper(cur->bc_tp, info, rec, rec_daddr);
@@ -362,7 +362,7 @@ xfs_getfsmap_datadev_bnobt_helper(
 	struct xfs_rmap_irec		irec;
 	xfs_daddr_t			rec_daddr;
 
-	rec_daddr = XFS_AGB_TO_DADDR(mp, cur->bc_private.a.agno,
+	rec_daddr = XFS_AGB_TO_DADDR(mp, cur->bc_ag.agno,
 			rec->ar_startblock);
 
 	irec.rm_startblock = rec->ar_startblock;
@@ -896,6 +896,14 @@ xfs_getfsmap(
 	info.format_arg = arg;
 	info.head = head;
 
+	/*
+	 * If fsmap runs concurrently with a scrub, the freeze can be delayed
+	 * indefinitely as we walk the rmapbt and iterate over metadata
+	 * buffers.  Freeze quiesces the log (which waits for the buffer LRU to
+	 * be emptied) and that won't happen while we're reading buffers.
+	 */
+	sb_start_write(mp->m_super);
+
 	/* For each device we support... */
 	for (i = 0; i < XFS_GETFSMAP_DEVS; i++) {
 		/* Is this device within the range the user asked for? */
@@ -935,6 +943,7 @@ xfs_getfsmap(
 
 	if (tp)
 		xfs_trans_cancel(tp);
+	sb_end_write(mp->m_super);
 	head->fmh_oflags = FMH_OF_DEV_T;
 	return error;
 }
diff --git a/fs/xfs/xfs_fsops.c b/fs/xfs/xfs_fsops.c
index 3e61d0cc23f8..ef1d5bb88b93 100644
--- a/fs/xfs/xfs_fsops.c
+++ b/fs/xfs/xfs_fsops.c
@@ -504,10 +504,7 @@ xfs_do_force_shutdown(
 	} else if (logerror) {
 		xfs_alert_tag(mp, XFS_PTAG_SHUTDOWN_LOGERROR,
 			"Log I/O Error Detected. Shutting down filesystem");
-	} else if (flags & SHUTDOWN_DEVICE_REQ) {
-		xfs_alert_tag(mp, XFS_PTAG_SHUTDOWN_IOERROR,
-			"All device paths lost. Shutting down filesystem");
-	} else if (!(flags & SHUTDOWN_REMOTE_REQ)) {
+	} else {
 		xfs_alert_tag(mp, XFS_PTAG_SHUTDOWN_IOERROR,
 			"I/O Error Detected. Shutting down filesystem");
 	}
diff --git a/fs/xfs/xfs_icache.c b/fs/xfs/xfs_icache.c
index 8dc2e5414276..101028ebb571 100644
--- a/fs/xfs/xfs_icache.c
+++ b/fs/xfs/xfs_icache.c
@@ -22,6 +22,7 @@
 #include "xfs_dquot_item.h"
 #include "xfs_dquot.h"
 #include "xfs_reflink.h"
+#include "xfs_ialloc.h"
 
 #include <linux/iversion.h>
 
@@ -36,13 +37,11 @@ xfs_inode_alloc(
 	struct xfs_inode	*ip;
 
 	/*
-	 * if this didn't occur in transactions, we could use
-	 * KM_MAYFAIL and return NULL here on ENOMEM. Set the
-	 * code up to do this anyway.
+	 * XXX: If this didn't occur in transactions, we could drop GFP_NOFAIL
+	 * and return NULL here on ENOMEM.
 	 */
-	ip = kmem_zone_alloc(xfs_inode_zone, 0);
-	if (!ip)
-		return NULL;
+	ip = kmem_cache_alloc(xfs_inode_zone, GFP_KERNEL | __GFP_NOFAIL);
+
 	if (inode_init_always(mp->m_super, VFS_I(ip))) {
 		kmem_cache_free(xfs_inode_zone, ip);
 		return NULL;
@@ -62,8 +61,6 @@ xfs_inode_alloc(
 	memset(&ip->i_imap, 0, sizeof(struct xfs_imap));
 	ip->i_afp = NULL;
 	ip->i_cowfp = NULL;
-	ip->i_cnextents = 0;
-	ip->i_cformat = XFS_DINODE_FMT_EXTENTS;
 	memset(&ip->i_df, 0, sizeof(ip->i_df));
 	ip->i_flags = 0;
 	ip->i_delayed_blks = 0;
@@ -88,15 +85,18 @@ xfs_inode_free_callback(
 	case S_IFREG:
 	case S_IFDIR:
 	case S_IFLNK:
-		xfs_idestroy_fork(ip, XFS_DATA_FORK);
+		xfs_idestroy_fork(&ip->i_df);
 		break;
 	}
 
-	if (ip->i_afp)
-		xfs_idestroy_fork(ip, XFS_ATTR_FORK);
-	if (ip->i_cowfp)
-		xfs_idestroy_fork(ip, XFS_COW_FORK);
-
+	if (ip->i_afp) {
+		xfs_idestroy_fork(ip->i_afp);
+		kmem_cache_free(xfs_ifork_zone, ip->i_afp);
+	}
+	if (ip->i_cowfp) {
+		xfs_idestroy_fork(ip->i_cowfp);
+		kmem_cache_free(xfs_ifork_zone, ip->i_cowfp);
+	}
 	if (ip->i_itemp) {
 		ASSERT(!test_bit(XFS_LI_IN_AIL,
 				 &ip->i_itemp->ili_item.li_flags));
@@ -113,6 +113,7 @@ __xfs_inode_free(
 {
 	/* asserts to verify all state is correct here */
 	ASSERT(atomic_read(&ip->i_pincount) == 0);
+	ASSERT(!ip->i_itemp || list_empty(&ip->i_itemp->ili_item.li_bio_list));
 	XFS_STATS_DEC(ip->i_mount, vn_active);
 
 	call_rcu(&VFS_I(ip)->i_rcu, xfs_inode_free_callback);
@@ -139,11 +140,8 @@ xfs_inode_free(
 }
 
 /*
- * Queue a new inode reclaim pass if there are reclaimable inodes and there
- * isn't a reclaim pass already in progress. By default it runs every 5s based
- * on the xfs periodic sync default of 30s. Perhaps this should have it's own
- * tunable, but that can be done if this method proves to be ineffective or too
- * aggressive.
+ * Queue background inode reclaim work if there are reclaimable inodes and there
+ * isn't reclaim work already scheduled or in progress.
  */
 static void
 xfs_reclaim_work_queue(
@@ -158,24 +156,6 @@ xfs_reclaim_work_queue(
 	rcu_read_unlock();
 }
 
-/*
- * This is a fast pass over the inode cache to try to get reclaim moving on as
- * many inodes as possible in a short period of time. It kicks itself every few
- * seconds, as well as being kicked by the inode cache shrinker when memory
- * goes low. It scans as quickly as possible avoiding locked inodes or those
- * already being flushed, and once done schedules a future pass.
- */
-void
-xfs_reclaim_worker(
-	struct work_struct *work)
-{
-	struct xfs_mount *mp = container_of(to_delayed_work(work),
-					struct xfs_mount, m_reclaim_work);
-
-	xfs_reclaim_inodes(mp, SYNC_TRYLOCK);
-	xfs_reclaim_work_queue(mp);
-}
-
 static void
 xfs_perag_set_reclaim_tag(
 	struct xfs_perag	*pag)
@@ -289,6 +269,8 @@ xfs_reinit_inode(
 	uint64_t	version = inode_peek_iversion(inode);
 	umode_t		mode = inode->i_mode;
 	dev_t		dev = inode->i_rdev;
+	kuid_t		uid = inode->i_uid;
+	kgid_t		gid = inode->i_gid;
 
 	error = inode_init_always(mp->m_super, inode);
 
@@ -297,6 +279,8 @@ xfs_reinit_inode(
 	inode_set_iversion_queried(inode, version);
 	inode->i_mode = mode;
 	inode->i_rdev = dev;
+	inode->i_uid = uid;
+	inode->i_gid = gid;
 	return error;
 }
 
@@ -419,6 +403,7 @@ xfs_iget_cache_hit(
 		spin_unlock(&ip->i_flags_lock);
 		rcu_read_unlock();
 
+		ASSERT(!rwsem_is_locked(&inode->i_rwsem));
 		error = xfs_reinit_inode(mp, inode);
 		if (error) {
 			bool wake;
@@ -452,9 +437,6 @@ xfs_iget_cache_hit(
 		ip->i_sick = 0;
 		ip->i_checked = 0;
 
-		ASSERT(!rwsem_is_locked(&inode->i_rwsem));
-		init_rwsem(&inode->i_rwsem);
-
 		spin_unlock(&ip->i_flags_lock);
 		spin_unlock(&pag->pag_ici_lock);
 	} else {
@@ -475,7 +457,7 @@ xfs_iget_cache_hit(
 		xfs_ilock(ip, lock_flags);
 
 	if (!(flags & XFS_IGET_INCORE))
-		xfs_iflags_clear(ip, XFS_ISTALE | XFS_IDONTCACHE);
+		xfs_iflags_clear(ip, XFS_ISTALE);
 	XFS_STATS_INC(mp, xs_ig_found);
 
 	return 0;
@@ -506,18 +488,42 @@ xfs_iget_cache_miss(
 	if (!ip)
 		return -ENOMEM;
 
-	error = xfs_iread(mp, tp, ip, flags);
+	error = xfs_imap(mp, tp, ip->i_ino, &ip->i_imap, flags);
 	if (error)
 		goto out_destroy;
 
-	if (!xfs_inode_verify_forks(ip)) {
-		error = -EFSCORRUPTED;
-		goto out_destroy;
+	/*
+	 * For version 5 superblocks, if we are initialising a new inode and we
+	 * are not utilising the XFS_MOUNT_IKEEP inode cluster mode, we can
+	 * simply build the new inode core with a random generation number.
+	 *
+	 * For version 4 (and older) superblocks, log recovery is dependent on
+	 * the di_flushiter field being initialised from the current on-disk
+	 * value and hence we must also read the inode off disk even when
+	 * initializing new inodes.
+	 */
+	if (xfs_sb_version_has_v3inode(&mp->m_sb) &&
+	    (flags & XFS_IGET_CREATE) && !(mp->m_flags & XFS_MOUNT_IKEEP)) {
+		VFS_I(ip)->i_generation = prandom_u32();
+	} else {
+		struct xfs_dinode	*dip;
+		struct xfs_buf		*bp;
+
+		error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &bp, 0);
+		if (error)
+			goto out_destroy;
+
+		error = xfs_inode_from_disk(ip, dip);
+		if (!error)
+			xfs_buf_set_ref(bp, XFS_INO_REF);
+		xfs_trans_brelse(tp, bp);
+
+		if (error)
+			goto out_destroy;
 	}
 
 	trace_xfs_iget_miss(ip);
 
-
 	/*
 	 * Check the inode free state is valid. This also detects lookup
 	 * racing with unlinks.
@@ -557,7 +563,7 @@ xfs_iget_cache_miss(
 	 */
 	iflags = XFS_INEW;
 	if (flags & XFS_IGET_DONTCACHE)
-		iflags |= XFS_IDONTCACHE;
+		d_mark_dontcache(VFS_I(ip));
 	ip->i_udquot = NULL;
 	ip->i_gdquot = NULL;
 	ip->i_pdquot = NULL;
@@ -590,48 +596,31 @@ out_destroy:
 }
 
 /*
- * Look up an inode by number in the given file system.
- * The inode is looked up in the cache held in each AG.
- * If the inode is found in the cache, initialise the vfs inode
- * if necessary.
+ * Look up an inode by number in the given file system.  The inode is looked up
+ * in the cache held in each AG.  If the inode is found in the cache, initialise
+ * the vfs inode if necessary.
  *
- * If it is not in core, read it in from the file system's device,
- * add it to the cache and initialise the vfs inode.
+ * If it is not in core, read it in from the file system's device, add it to the
+ * cache and initialise the vfs inode.
  *
  * The inode is locked according to the value of the lock_flags parameter.
- * This flag parameter indicates how and if the inode's IO lock and inode lock
- * should be taken.
- *
- * mp -- the mount point structure for the current file system.  It points
- *       to the inode hash table.
- * tp -- a pointer to the current transaction if there is one.  This is
- *       simply passed through to the xfs_iread() call.
- * ino -- the number of the inode desired.  This is the unique identifier
- *        within the file system for the inode being requested.
- * lock_flags -- flags indicating how to lock the inode.  See the comment
- *		 for xfs_ilock() for a list of valid values.
+ * Inode lookup is only done during metadata operations and not as part of the
+ * data IO path. Hence we only allow locking of the XFS_ILOCK during lookup.
  */
 int
 xfs_iget(
-	xfs_mount_t	*mp,
-	xfs_trans_t	*tp,
-	xfs_ino_t	ino,
-	uint		flags,
-	uint		lock_flags,
-	xfs_inode_t	**ipp)
+	struct xfs_mount	*mp,
+	struct xfs_trans	*tp,
+	xfs_ino_t		ino,
+	uint			flags,
+	uint			lock_flags,
+	struct xfs_inode	**ipp)
 {
-	xfs_inode_t	*ip;
-	int		error;
-	xfs_perag_t	*pag;
-	xfs_agino_t	agino;
+	struct xfs_inode	*ip;
+	struct xfs_perag	*pag;
+	xfs_agino_t		agino;
+	int			error;
 
-	/*
-	 * xfs_reclaim_inode() uses the ILOCK to ensure an inode
-	 * doesn't get freed while it's being referenced during a
-	 * radix tree traversal here.  It assumes this function
-	 * aqcuires only the ILOCK (and therefore it has no need to
-	 * involve the IOLOCK in this synchronization).
-	 */
 	ASSERT((lock_flags & (XFS_IOLOCK_EXCL | XFS_IOLOCK_SHARED)) == 0);
 
 	/* reject inode numbers outside existing AGs */
@@ -733,25 +722,22 @@ xfs_icache_inode_is_allocated(
  */
 #define XFS_LOOKUP_BATCH	32
 
-STATIC int
-xfs_inode_ag_walk_grab(
+/*
+ * Decide if the given @ip is eligible to be a part of the inode walk, and
+ * grab it if so.  Returns true if it's ready to go or false if we should just
+ * ignore it.
+ */
+STATIC bool
+xfs_inode_walk_ag_grab(
 	struct xfs_inode	*ip,
 	int			flags)
 {
 	struct inode		*inode = VFS_I(ip);
-	bool			newinos = !!(flags & XFS_AGITER_INEW_WAIT);
+	bool			newinos = !!(flags & XFS_INODE_WALK_INEW_WAIT);
 
 	ASSERT(rcu_read_lock_held());
 
-	/*
-	 * check for stale RCU freed inode
-	 *
-	 * If the inode has been reallocated, it doesn't matter if it's not in
-	 * the AG we are walking - we are walking for writeback, so if it
-	 * passes all the "valid inode" checks and is dirty, then we'll write
-	 * it back anyway.  If it has been reallocated and still being
-	 * initialised, the XFS_INEW check below will catch it.
-	 */
+	/* Check for stale RCU freed inode */
 	spin_lock(&ip->i_flags_lock);
 	if (!ip->i_ino)
 		goto out_unlock_noent;
@@ -764,39 +750,41 @@ xfs_inode_ag_walk_grab(
 
 	/* nothing to sync during shutdown */
 	if (XFS_FORCED_SHUTDOWN(ip->i_mount))
-		return -EFSCORRUPTED;
+		return false;
 
 	/* If we can't grab the inode, it must on it's way to reclaim. */
 	if (!igrab(inode))
-		return -ENOENT;
+		return false;
 
 	/* inode is valid */
-	return 0;
+	return true;
 
 out_unlock_noent:
 	spin_unlock(&ip->i_flags_lock);
-	return -ENOENT;
+	return false;
 }
 
+/*
+ * For a given per-AG structure @pag, grab, @execute, and rele all incore
+ * inodes with the given radix tree @tag.
+ */
 STATIC int
-xfs_inode_ag_walk(
-	struct xfs_mount	*mp,
+xfs_inode_walk_ag(
 	struct xfs_perag	*pag,
-	int			(*execute)(struct xfs_inode *ip, int flags,
-					   void *args),
-	int			flags,
+	int			iter_flags,
+	int			(*execute)(struct xfs_inode *ip, void *args),
 	void			*args,
-	int			tag,
-	int			iter_flags)
+	int			tag)
 {
+	struct xfs_mount	*mp = pag->pag_mount;
 	uint32_t		first_index;
 	int			last_error = 0;
 	int			skipped;
-	int			done;
+	bool			done;
 	int			nr_found;
 
 restart:
-	done = 0;
+	done = false;
 	skipped = 0;
 	first_index = 0;
 	nr_found = 0;
@@ -807,7 +795,7 @@ restart:
 
 		rcu_read_lock();
 
-		if (tag == -1)
+		if (tag == XFS_ICI_NO_TAG)
 			nr_found = radix_tree_gang_lookup(&pag->pag_ici_root,
 					(void **)batch, first_index,
 					XFS_LOOKUP_BATCH);
@@ -829,7 +817,7 @@ restart:
 		for (i = 0; i < nr_found; i++) {
 			struct xfs_inode *ip = batch[i];
 
-			if (done || xfs_inode_ag_walk_grab(ip, iter_flags))
+			if (done || !xfs_inode_walk_ag_grab(ip, iter_flags))
 				batch[i] = NULL;
 
 			/*
@@ -848,7 +836,7 @@ restart:
 				continue;
 			first_index = XFS_INO_TO_AGINO(mp, ip->i_ino + 1);
 			if (first_index < XFS_INO_TO_AGINO(mp, ip->i_ino))
-				done = 1;
+				done = true;
 		}
 
 		/* unlock now we've grabbed the inodes. */
@@ -857,10 +845,10 @@ restart:
 		for (i = 0; i < nr_found; i++) {
 			if (!batch[i])
 				continue;
-			if ((iter_flags & XFS_AGITER_INEW_WAIT) &&
+			if ((iter_flags & XFS_INODE_WALK_INEW_WAIT) &&
 			    xfs_iflags_test(batch[i], XFS_INEW))
 				xfs_inew_wait(batch[i]);
-			error = execute(batch[i], flags, args);
+			error = execute(batch[i], args);
 			xfs_irele(batch[i]);
 			if (error == -EAGAIN) {
 				skipped++;
@@ -885,6 +873,49 @@ restart:
 	return last_error;
 }
 
+/* Fetch the next (possibly tagged) per-AG structure. */
+static inline struct xfs_perag *
+xfs_inode_walk_get_perag(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agno,
+	int			tag)
+{
+	if (tag == XFS_ICI_NO_TAG)
+		return xfs_perag_get(mp, agno);
+	return xfs_perag_get_tag(mp, agno, tag);
+}
+
+/*
+ * Call the @execute function on all incore inodes matching the radix tree
+ * @tag.
+ */
+int
+xfs_inode_walk(
+	struct xfs_mount	*mp,
+	int			iter_flags,
+	int			(*execute)(struct xfs_inode *ip, void *args),
+	void			*args,
+	int			tag)
+{
+	struct xfs_perag	*pag;
+	int			error = 0;
+	int			last_error = 0;
+	xfs_agnumber_t		ag;
+
+	ag = 0;
+	while ((pag = xfs_inode_walk_get_perag(mp, ag, tag))) {
+		ag = pag->pag_agno + 1;
+		error = xfs_inode_walk_ag(pag, iter_flags, execute, args, tag);
+		xfs_perag_put(pag);
+		if (error) {
+			last_error = error;
+			if (error == -EFSCORRUPTED)
+				break;
+		}
+	}
+	return last_error;
+}
+
 /*
  * Background scanning to trim post-EOF preallocated space. This is queued
  * based on the 'speculative_prealloc_lifetime' tunable (5m by default).
@@ -907,7 +938,12 @@ xfs_eofblocks_worker(
 {
 	struct xfs_mount *mp = container_of(to_delayed_work(work),
 				struct xfs_mount, m_eofblocks_work);
+
+	if (!sb_start_write_trylock(mp->m_super))
+		return;
 	xfs_icache_free_eofblocks(mp, NULL);
+	sb_end_write(mp->m_super);
+
 	xfs_queue_eofblocks(mp);
 }
 
@@ -934,235 +970,86 @@ xfs_cowblocks_worker(
 {
 	struct xfs_mount *mp = container_of(to_delayed_work(work),
 				struct xfs_mount, m_cowblocks_work);
-	xfs_icache_free_cowblocks(mp, NULL);
-	xfs_queue_cowblocks(mp);
-}
-
-int
-xfs_inode_ag_iterator_flags(
-	struct xfs_mount	*mp,
-	int			(*execute)(struct xfs_inode *ip, int flags,
-					   void *args),
-	int			flags,
-	void			*args,
-	int			iter_flags)
-{
-	struct xfs_perag	*pag;
-	int			error = 0;
-	int			last_error = 0;
-	xfs_agnumber_t		ag;
-
-	ag = 0;
-	while ((pag = xfs_perag_get(mp, ag))) {
-		ag = pag->pag_agno + 1;
-		error = xfs_inode_ag_walk(mp, pag, execute, flags, args, -1,
-					  iter_flags);
-		xfs_perag_put(pag);
-		if (error) {
-			last_error = error;
-			if (error == -EFSCORRUPTED)
-				break;
-		}
-	}
-	return last_error;
-}
 
-int
-xfs_inode_ag_iterator(
-	struct xfs_mount	*mp,
-	int			(*execute)(struct xfs_inode *ip, int flags,
-					   void *args),
-	int			flags,
-	void			*args)
-{
-	return xfs_inode_ag_iterator_flags(mp, execute, flags, args, 0);
-}
-
-int
-xfs_inode_ag_iterator_tag(
-	struct xfs_mount	*mp,
-	int			(*execute)(struct xfs_inode *ip, int flags,
-					   void *args),
-	int			flags,
-	void			*args,
-	int			tag)
-{
-	struct xfs_perag	*pag;
-	int			error = 0;
-	int			last_error = 0;
-	xfs_agnumber_t		ag;
+	if (!sb_start_write_trylock(mp->m_super))
+		return;
+	xfs_icache_free_cowblocks(mp, NULL);
+	sb_end_write(mp->m_super);
 
-	ag = 0;
-	while ((pag = xfs_perag_get_tag(mp, ag, tag))) {
-		ag = pag->pag_agno + 1;
-		error = xfs_inode_ag_walk(mp, pag, execute, flags, args, tag,
-					  0);
-		xfs_perag_put(pag);
-		if (error) {
-			last_error = error;
-			if (error == -EFSCORRUPTED)
-				break;
-		}
-	}
-	return last_error;
+	xfs_queue_cowblocks(mp);
 }
 
 /*
  * Grab the inode for reclaim exclusively.
- * Return 0 if we grabbed it, non-zero otherwise.
+ *
+ * We have found this inode via a lookup under RCU, so the inode may have
+ * already been freed, or it may be in the process of being recycled by
+ * xfs_iget(). In both cases, the inode will have XFS_IRECLAIM set. If the inode
+ * has been fully recycled by the time we get the i_flags_lock, XFS_IRECLAIMABLE
+ * will not be set. Hence we need to check for both these flag conditions to
+ * avoid inodes that are no longer reclaim candidates.
+ *
+ * Note: checking for other state flags here, under the i_flags_lock or not, is
+ * racy and should be avoided. Those races should be resolved only after we have
+ * ensured that we are able to reclaim this inode and the world can see that we
+ * are going to reclaim it.
+ *
+ * Return true if we grabbed it, false otherwise.
  */
-STATIC int
+static bool
 xfs_reclaim_inode_grab(
-	struct xfs_inode	*ip,
-	int			flags)
+	struct xfs_inode	*ip)
 {
 	ASSERT(rcu_read_lock_held());
 
-	/* quick check for stale RCU freed inode */
-	if (!ip->i_ino)
-		return 1;
-
-	/*
-	 * If we are asked for non-blocking operation, do unlocked checks to
-	 * see if the inode already is being flushed or in reclaim to avoid
-	 * lock traffic.
-	 */
-	if ((flags & SYNC_TRYLOCK) &&
-	    __xfs_iflags_test(ip, XFS_IFLOCK | XFS_IRECLAIM))
-		return 1;
-
-	/*
-	 * The radix tree lock here protects a thread in xfs_iget from racing
-	 * with us starting reclaim on the inode.  Once we have the
-	 * XFS_IRECLAIM flag set it will not touch us.
-	 *
-	 * Due to RCU lookup, we may find inodes that have been freed and only
-	 * have XFS_IRECLAIM set.  Indeed, we may see reallocated inodes that
-	 * aren't candidates for reclaim at all, so we must check the
-	 * XFS_IRECLAIMABLE is set first before proceeding to reclaim.
-	 */
 	spin_lock(&ip->i_flags_lock);
 	if (!__xfs_iflags_test(ip, XFS_IRECLAIMABLE) ||
 	    __xfs_iflags_test(ip, XFS_IRECLAIM)) {
 		/* not a reclaim candidate. */
 		spin_unlock(&ip->i_flags_lock);
-		return 1;
+		return false;
 	}
 	__xfs_iflags_set(ip, XFS_IRECLAIM);
 	spin_unlock(&ip->i_flags_lock);
-	return 0;
+	return true;
 }
 
 /*
- * Inodes in different states need to be treated differently. The following
- * table lists the inode states and the reclaim actions necessary:
- *
- *	inode state	     iflush ret		required action
- *      ---------------      ----------         ---------------
- *	bad			-		reclaim
- *	shutdown		EIO		unpin and reclaim
- *	clean, unpinned		0		reclaim
- *	stale, unpinned		0		reclaim
- *	clean, pinned(*)	0		requeue
- *	stale, pinned		EAGAIN		requeue
- *	dirty, async		-		requeue
- *	dirty, sync		0		reclaim
- *
- * (*) dgc: I don't think the clean, pinned state is possible but it gets
- * handled anyway given the order of checks implemented.
- *
- * Also, because we get the flush lock first, we know that any inode that has
- * been flushed delwri has had the flush completed by the time we check that
- * the inode is clean.
- *
- * Note that because the inode is flushed delayed write by AIL pushing, the
- * flush lock may already be held here and waiting on it can result in very
- * long latencies.  Hence for sync reclaims, where we wait on the flush lock,
- * the caller should push the AIL first before trying to reclaim inodes to
- * minimise the amount of time spent waiting.  For background relaim, we only
- * bother to reclaim clean inodes anyway.
+ * Inode reclaim is non-blocking, so the default action if progress cannot be
+ * made is to "requeue" the inode for reclaim by unlocking it and clearing the
+ * XFS_IRECLAIM flag.  If we are in a shutdown state, we don't care about
+ * blocking anymore and hence we can wait for the inode to be able to reclaim
+ * it.
  *
- * Hence the order of actions after gaining the locks should be:
- *	bad		=> reclaim
- *	shutdown	=> unpin and reclaim
- *	pinned, async	=> requeue
- *	pinned, sync	=> unpin
- *	stale		=> reclaim
- *	clean		=> reclaim
- *	dirty, async	=> requeue
- *	dirty, sync	=> flush, wait and reclaim
+ * We do no IO here - if callers require inodes to be cleaned they must push the
+ * AIL first to trigger writeback of dirty inodes.  This enables writeback to be
+ * done in the background in a non-blocking manner, and enables memory reclaim
+ * to make progress without blocking.
  */
-STATIC int
+static void
 xfs_reclaim_inode(
 	struct xfs_inode	*ip,
-	struct xfs_perag	*pag,
-	int			sync_mode)
+	struct xfs_perag	*pag)
 {
-	struct xfs_buf		*bp = NULL;
 	xfs_ino_t		ino = ip->i_ino; /* for radix_tree_delete */
-	int			error;
 
-restart:
-	error = 0;
-	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	if (!xfs_iflock_nowait(ip)) {
-		if (!(sync_mode & SYNC_WAIT))
-			goto out;
-		xfs_iflock(ip);
-	}
+	if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL))
+		goto out;
+	if (!xfs_iflock_nowait(ip))
+		goto out_iunlock;
 
 	if (XFS_FORCED_SHUTDOWN(ip->i_mount)) {
 		xfs_iunpin_wait(ip);
 		/* xfs_iflush_abort() drops the flush lock */
-		xfs_iflush_abort(ip, false);
+		xfs_iflush_abort(ip);
 		goto reclaim;
 	}
-	if (xfs_ipincount(ip)) {
-		if (!(sync_mode & SYNC_WAIT))
-			goto out_ifunlock;
-		xfs_iunpin_wait(ip);
-	}
-	if (xfs_iflags_test(ip, XFS_ISTALE) || xfs_inode_clean(ip)) {
-		xfs_ifunlock(ip);
-		goto reclaim;
-	}
-
-	/*
-	 * Never flush out dirty data during non-blocking reclaim, as it would
-	 * just contend with AIL pushing trying to do the same job.
-	 */
-	if (!(sync_mode & SYNC_WAIT))
+	if (xfs_ipincount(ip))
+		goto out_ifunlock;
+	if (!xfs_inode_clean(ip))
 		goto out_ifunlock;
 
-	/*
-	 * Now we have an inode that needs flushing.
-	 *
-	 * Note that xfs_iflush will never block on the inode buffer lock, as
-	 * xfs_ifree_cluster() can lock the inode buffer before it locks the
-	 * ip->i_lock, and we are doing the exact opposite here.  As a result,
-	 * doing a blocking xfs_imap_to_bp() to get the cluster buffer would
-	 * result in an ABBA deadlock with xfs_ifree_cluster().
-	 *
-	 * As xfs_ifree_cluser() must gather all inodes that are active in the
-	 * cache to mark them stale, if we hit this case we don't actually want
-	 * to do IO here - we want the inode marked stale so we can simply
-	 * reclaim it.  Hence if we get an EAGAIN error here,  just unlock the
-	 * inode, back off and try again.  Hopefully the next pass through will
-	 * see the stale flag set on the inode.
-	 */
-	error = xfs_iflush(ip, &bp);
-	if (error == -EAGAIN) {
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		/* backoff longer than in xfs_ifree_cluster */
-		delay(2);
-		goto restart;
-	}
-
-	if (!error) {
-		error = xfs_bwrite(bp);
-		xfs_buf_relse(bp);
-	}
-
+	xfs_ifunlock(ip);
 reclaim:
 	ASSERT(!xfs_isiflocked(ip));
 
@@ -1209,23 +1096,17 @@ reclaim:
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_qm_dqdetach(ip);
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	ASSERT(xfs_inode_clean(ip));
 
 	__xfs_inode_free(ip);
-	return error;
+	return;
 
 out_ifunlock:
 	xfs_ifunlock(ip);
+out_iunlock:
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 out:
 	xfs_iflags_clear(ip, XFS_IRECLAIM);
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	/*
-	 * We could return -EAGAIN here to make reclaim rescan the inode tree in
-	 * a short while. However, this just burns CPU time scanning the tree
-	 * waiting for IO to complete and the reclaim work never goes back to
-	 * the idle state. Instead, return 0 to let the next scheduled
-	 * background reclaim attempt to reclaim the inode again.
-	 */
-	return 0;
 }
 
 /*
@@ -1233,23 +1114,19 @@ out:
  * corrupted, we still want to try to reclaim all the inodes. If we don't,
  * then a shut down during filesystem unmount reclaim walk leak all the
  * unreclaimed inodes.
+ *
+ * Returns non-zero if any AGs or inodes were skipped in the reclaim pass
+ * so that callers that want to block until all dirty inodes are written back
+ * and reclaimed can sanely loop.
  */
-STATIC int
+static void
 xfs_reclaim_inodes_ag(
 	struct xfs_mount	*mp,
-	int			flags,
 	int			*nr_to_scan)
 {
 	struct xfs_perag	*pag;
-	int			error = 0;
-	int			last_error = 0;
-	xfs_agnumber_t		ag;
-	int			trylock = flags & SYNC_TRYLOCK;
-	int			skipped;
+	xfs_agnumber_t		ag = 0;
 
-restart:
-	ag = 0;
-	skipped = 0;
 	while ((pag = xfs_perag_get_tag(mp, ag, XFS_ICI_RECLAIM_TAG))) {
 		unsigned long	first_index = 0;
 		int		done = 0;
@@ -1257,16 +1134,7 @@ restart:
 
 		ag = pag->pag_agno + 1;
 
-		if (trylock) {
-			if (!mutex_trylock(&pag->pag_ici_reclaim_lock)) {
-				skipped++;
-				xfs_perag_put(pag);
-				continue;
-			}
-			first_index = pag->pag_ici_reclaim_cursor;
-		} else
-			mutex_lock(&pag->pag_ici_reclaim_lock);
-
+		first_index = READ_ONCE(pag->pag_ici_reclaim_cursor);
 		do {
 			struct xfs_inode *batch[XFS_LOOKUP_BATCH];
 			int	i;
@@ -1290,7 +1158,7 @@ restart:
 			for (i = 0; i < nr_found; i++) {
 				struct xfs_inode *ip = batch[i];
 
-				if (done || xfs_reclaim_inode_grab(ip, flags))
+				if (done || !xfs_reclaim_inode_grab(ip))
 					batch[i] = NULL;
 
 				/*
@@ -1319,59 +1187,39 @@ restart:
 			rcu_read_unlock();
 
 			for (i = 0; i < nr_found; i++) {
-				if (!batch[i])
-					continue;
-				error = xfs_reclaim_inode(batch[i], pag, flags);
-				if (error && last_error != -EFSCORRUPTED)
-					last_error = error;
+				if (batch[i])
+					xfs_reclaim_inode(batch[i], pag);
 			}
 
 			*nr_to_scan -= XFS_LOOKUP_BATCH;
-
 			cond_resched();
-
 		} while (nr_found && !done && *nr_to_scan > 0);
 
-		if (trylock && !done)
-			pag->pag_ici_reclaim_cursor = first_index;
-		else
-			pag->pag_ici_reclaim_cursor = 0;
-		mutex_unlock(&pag->pag_ici_reclaim_lock);
+		if (done)
+			first_index = 0;
+		WRITE_ONCE(pag->pag_ici_reclaim_cursor, first_index);
 		xfs_perag_put(pag);
 	}
-
-	/*
-	 * if we skipped any AG, and we still have scan count remaining, do
-	 * another pass this time using blocking reclaim semantics (i.e
-	 * waiting on the reclaim locks and ignoring the reclaim cursors). This
-	 * ensure that when we get more reclaimers than AGs we block rather
-	 * than spin trying to execute reclaim.
-	 */
-	if (skipped && (flags & SYNC_WAIT) && *nr_to_scan > 0) {
-		trylock = 0;
-		goto restart;
-	}
-	return last_error;
 }
 
-int
+void
 xfs_reclaim_inodes(
-	xfs_mount_t	*mp,
-	int		mode)
+	struct xfs_mount	*mp)
 {
 	int		nr_to_scan = INT_MAX;
 
-	return xfs_reclaim_inodes_ag(mp, mode, &nr_to_scan);
+	while (radix_tree_tagged(&mp->m_perag_tree, XFS_ICI_RECLAIM_TAG)) {
+		xfs_ail_push_all_sync(mp->m_ail);
+		xfs_reclaim_inodes_ag(mp, &nr_to_scan);
+	};
 }
 
 /*
- * Scan a certain number of inodes for reclaim.
- *
- * When called we make sure that there is a background (fast) inode reclaim in
- * progress, while we will throttle the speed of reclaim via doing synchronous
- * reclaim of inodes. That means if we come across dirty inodes, we wait for
- * them to be cleaned, which we hope will not be very long due to the
- * background walker having already kicked the IO off on those dirty inodes.
+ * The shrinker infrastructure determines how many inodes we should scan for
+ * reclaim. We want as many clean inodes ready to reclaim as possible, so we
+ * push the AIL here. We also want to proactively free up memory if we can to
+ * minimise the amount of work memory reclaim has to do so we kick the
+ * background reclaim if it isn't already scheduled.
  */
 long
 xfs_reclaim_inodes_nr(
@@ -1382,7 +1230,8 @@ xfs_reclaim_inodes_nr(
 	xfs_reclaim_work_queue(mp);
 	xfs_ail_push_all(mp->m_ail);
 
-	return xfs_reclaim_inodes_ag(mp, SYNC_TRYLOCK | SYNC_WAIT, &nr_to_scan);
+	xfs_reclaim_inodes_ag(mp, &nr_to_scan);
+	return 0;
 }
 
 /*
@@ -1405,59 +1254,108 @@ xfs_reclaim_inodes_count(
 	return reclaimable;
 }
 
-STATIC int
+STATIC bool
 xfs_inode_match_id(
 	struct xfs_inode	*ip,
 	struct xfs_eofblocks	*eofb)
 {
 	if ((eofb->eof_flags & XFS_EOF_FLAGS_UID) &&
 	    !uid_eq(VFS_I(ip)->i_uid, eofb->eof_uid))
-		return 0;
+		return false;
 
 	if ((eofb->eof_flags & XFS_EOF_FLAGS_GID) &&
 	    !gid_eq(VFS_I(ip)->i_gid, eofb->eof_gid))
-		return 0;
+		return false;
 
 	if ((eofb->eof_flags & XFS_EOF_FLAGS_PRID) &&
 	    ip->i_d.di_projid != eofb->eof_prid)
-		return 0;
+		return false;
 
-	return 1;
+	return true;
 }
 
 /*
  * A union-based inode filtering algorithm. Process the inode if any of the
  * criteria match. This is for global/internal scans only.
  */
-STATIC int
+STATIC bool
 xfs_inode_match_id_union(
 	struct xfs_inode	*ip,
 	struct xfs_eofblocks	*eofb)
 {
 	if ((eofb->eof_flags & XFS_EOF_FLAGS_UID) &&
 	    uid_eq(VFS_I(ip)->i_uid, eofb->eof_uid))
-		return 1;
+		return true;
 
 	if ((eofb->eof_flags & XFS_EOF_FLAGS_GID) &&
 	    gid_eq(VFS_I(ip)->i_gid, eofb->eof_gid))
-		return 1;
+		return true;
 
 	if ((eofb->eof_flags & XFS_EOF_FLAGS_PRID) &&
 	    ip->i_d.di_projid == eofb->eof_prid)
-		return 1;
+		return true;
 
-	return 0;
+	return false;
+}
+
+/*
+ * Is this inode @ip eligible for eof/cow block reclamation, given some
+ * filtering parameters @eofb?  The inode is eligible if @eofb is null or
+ * if the predicate functions match.
+ */
+static bool
+xfs_inode_matches_eofb(
+	struct xfs_inode	*ip,
+	struct xfs_eofblocks	*eofb)
+{
+	bool			match;
+
+	if (!eofb)
+		return true;
+
+	if (eofb->eof_flags & XFS_EOF_FLAGS_UNION)
+		match = xfs_inode_match_id_union(ip, eofb);
+	else
+		match = xfs_inode_match_id(ip, eofb);
+	if (!match)
+		return false;
+
+	/* skip the inode if the file size is too small */
+	if ((eofb->eof_flags & XFS_EOF_FLAGS_MINFILESIZE) &&
+	    XFS_ISIZE(ip) < eofb->eof_min_file_size)
+		return false;
+
+	return true;
+}
+
+/*
+ * This is a fast pass over the inode cache to try to get reclaim moving on as
+ * many inodes as possible in a short period of time. It kicks itself every few
+ * seconds, as well as being kicked by the inode cache shrinker when memory
+ * goes low.
+ */
+void
+xfs_reclaim_worker(
+	struct work_struct *work)
+{
+	struct xfs_mount *mp = container_of(to_delayed_work(work),
+					struct xfs_mount, m_reclaim_work);
+	int		nr_to_scan = INT_MAX;
+
+	xfs_reclaim_inodes_ag(mp, &nr_to_scan);
+	xfs_reclaim_work_queue(mp);
 }
 
 STATIC int
 xfs_inode_free_eofblocks(
 	struct xfs_inode	*ip,
-	int			flags,
 	void			*args)
 {
-	int ret = 0;
-	struct xfs_eofblocks *eofb = args;
-	int match;
+	struct xfs_eofblocks	*eofb = args;
+	bool			wait;
+	int			ret;
+
+	wait = eofb && (eofb->eof_flags & XFS_EOF_FLAGS_SYNC);
 
 	if (!xfs_can_free_eofblocks(ip, false)) {
 		/* inode could be preallocated or append-only */
@@ -1470,62 +1368,34 @@ xfs_inode_free_eofblocks(
 	 * If the mapping is dirty the operation can block and wait for some
 	 * time. Unless we are waiting, skip it.
 	 */
-	if (!(flags & SYNC_WAIT) &&
-	    mapping_tagged(VFS_I(ip)->i_mapping, PAGECACHE_TAG_DIRTY))
+	if (!wait && mapping_tagged(VFS_I(ip)->i_mapping, PAGECACHE_TAG_DIRTY))
 		return 0;
 
-	if (eofb) {
-		if (eofb->eof_flags & XFS_EOF_FLAGS_UNION)
-			match = xfs_inode_match_id_union(ip, eofb);
-		else
-			match = xfs_inode_match_id(ip, eofb);
-		if (!match)
-			return 0;
-
-		/* skip the inode if the file size is too small */
-		if (eofb->eof_flags & XFS_EOF_FLAGS_MINFILESIZE &&
-		    XFS_ISIZE(ip) < eofb->eof_min_file_size)
-			return 0;
-	}
+	if (!xfs_inode_matches_eofb(ip, eofb))
+		return 0;
 
 	/*
 	 * If the caller is waiting, return -EAGAIN to keep the background
 	 * scanner moving and revisit the inode in a subsequent pass.
 	 */
 	if (!xfs_ilock_nowait(ip, XFS_IOLOCK_EXCL)) {
-		if (flags & SYNC_WAIT)
-			ret = -EAGAIN;
-		return ret;
+		if (wait)
+			return -EAGAIN;
+		return 0;
 	}
+
 	ret = xfs_free_eofblocks(ip);
 	xfs_iunlock(ip, XFS_IOLOCK_EXCL);
 
 	return ret;
 }
 
-static int
-__xfs_icache_free_eofblocks(
-	struct xfs_mount	*mp,
-	struct xfs_eofblocks	*eofb,
-	int			(*execute)(struct xfs_inode *ip, int flags,
-					   void *args),
-	int			tag)
-{
-	int flags = SYNC_TRYLOCK;
-
-	if (eofb && (eofb->eof_flags & XFS_EOF_FLAGS_SYNC))
-		flags = SYNC_WAIT;
-
-	return xfs_inode_ag_iterator_tag(mp, execute, flags,
-					 eofb, tag);
-}
-
 int
 xfs_icache_free_eofblocks(
 	struct xfs_mount	*mp,
 	struct xfs_eofblocks	*eofb)
 {
-	return __xfs_icache_free_eofblocks(mp, eofb, xfs_inode_free_eofblocks,
+	return xfs_inode_walk(mp, 0, xfs_inode_free_eofblocks, eofb,
 			XFS_ICI_EOFBLOCKS_TAG);
 }
 
@@ -1552,7 +1422,7 @@ __xfs_inode_free_quota_eofblocks(
 	eofb.eof_flags = XFS_EOF_FLAGS_UNION|XFS_EOF_FLAGS_SYNC;
 
 	if (XFS_IS_UQUOTA_ENFORCED(ip->i_mount)) {
-		dq = xfs_inode_dquot(ip, XFS_DQ_USER);
+		dq = xfs_inode_dquot(ip, XFS_DQTYPE_USER);
 		if (dq && xfs_dquot_lowsp(dq)) {
 			eofb.eof_uid = VFS_I(ip)->i_uid;
 			eofb.eof_flags |= XFS_EOF_FLAGS_UID;
@@ -1561,7 +1431,7 @@ __xfs_inode_free_quota_eofblocks(
 	}
 
 	if (XFS_IS_GQUOTA_ENFORCED(ip->i_mount)) {
-		dq = xfs_inode_dquot(ip, XFS_DQ_GROUP);
+		dq = xfs_inode_dquot(ip, XFS_DQTYPE_GROUP);
 		if (dq && xfs_dquot_lowsp(dq)) {
 			eofb.eof_gid = VFS_I(ip)->i_gid;
 			eofb.eof_flags |= XFS_EOF_FLAGS_GID;
@@ -1742,29 +1612,16 @@ xfs_prep_free_cowblocks(
 STATIC int
 xfs_inode_free_cowblocks(
 	struct xfs_inode	*ip,
-	int			flags,
 	void			*args)
 {
 	struct xfs_eofblocks	*eofb = args;
-	int			match;
 	int			ret = 0;
 
 	if (!xfs_prep_free_cowblocks(ip))
 		return 0;
 
-	if (eofb) {
-		if (eofb->eof_flags & XFS_EOF_FLAGS_UNION)
-			match = xfs_inode_match_id_union(ip, eofb);
-		else
-			match = xfs_inode_match_id(ip, eofb);
-		if (!match)
-			return 0;
-
-		/* skip the inode if the file size is too small */
-		if (eofb->eof_flags & XFS_EOF_FLAGS_MINFILESIZE &&
-		    XFS_ISIZE(ip) < eofb->eof_min_file_size)
-			return 0;
-	}
+	if (!xfs_inode_matches_eofb(ip, eofb))
+		return 0;
 
 	/* Free the CoW blocks */
 	xfs_ilock(ip, XFS_IOLOCK_EXCL);
@@ -1788,7 +1645,7 @@ xfs_icache_free_cowblocks(
 	struct xfs_mount	*mp,
 	struct xfs_eofblocks	*eofb)
 {
-	return __xfs_icache_free_eofblocks(mp, eofb, xfs_inode_free_cowblocks,
+	return xfs_inode_walk(mp, 0, xfs_inode_free_cowblocks, eofb,
 			XFS_ICI_COWBLOCKS_TAG);
 }
 
diff --git a/fs/xfs/xfs_icache.h b/fs/xfs/xfs_icache.h
index 48f1fd2bb6ad..3a4c8b382cd0 100644
--- a/fs/xfs/xfs_icache.h
+++ b/fs/xfs/xfs_icache.h
@@ -17,14 +17,11 @@ struct xfs_eofblocks {
 	__u64		eof_min_file_size;
 };
 
-#define SYNC_WAIT		0x0001	/* wait for i/o to complete */
-#define SYNC_TRYLOCK		0x0002  /* only try to lock inodes */
-
 /*
  * tags for inode radix tree
  */
 #define XFS_ICI_NO_TAG		(-1)	/* special flag for an untagged lookup
-					   in xfs_inode_ag_iterator */
+					   in xfs_inode_walk */
 #define XFS_ICI_RECLAIM_TAG	0	/* inode is to be reclaimed */
 #define XFS_ICI_EOFBLOCKS_TAG	1	/* inode has blocks beyond EOF */
 #define XFS_ICI_COWBLOCKS_TAG	2	/* inode can have cow blocks to gc */
@@ -40,7 +37,7 @@ struct xfs_eofblocks {
 /*
  * flags for AG inode iterator
  */
-#define XFS_AGITER_INEW_WAIT	0x1	/* wait on new inodes */
+#define XFS_INODE_WALK_INEW_WAIT	0x1	/* wait on new inodes */
 
 int xfs_iget(struct xfs_mount *mp, struct xfs_trans *tp, xfs_ino_t ino,
 	     uint flags, uint lock_flags, xfs_inode_t **ipp);
@@ -51,7 +48,7 @@ void xfs_inode_free(struct xfs_inode *ip);
 
 void xfs_reclaim_worker(struct work_struct *work);
 
-int xfs_reclaim_inodes(struct xfs_mount *mp, int mode);
+void xfs_reclaim_inodes(struct xfs_mount *mp);
 int xfs_reclaim_inodes_count(struct xfs_mount *mp);
 long xfs_reclaim_inodes_nr(struct xfs_mount *mp, int nr_to_scan);
 
@@ -71,50 +68,9 @@ int xfs_inode_free_quota_cowblocks(struct xfs_inode *ip);
 void xfs_cowblocks_worker(struct work_struct *);
 void xfs_queue_cowblocks(struct xfs_mount *);
 
-int xfs_inode_ag_iterator(struct xfs_mount *mp,
-	int (*execute)(struct xfs_inode *ip, int flags, void *args),
-	int flags, void *args);
-int xfs_inode_ag_iterator_flags(struct xfs_mount *mp,
-	int (*execute)(struct xfs_inode *ip, int flags, void *args),
-	int flags, void *args, int iter_flags);
-int xfs_inode_ag_iterator_tag(struct xfs_mount *mp,
-	int (*execute)(struct xfs_inode *ip, int flags, void *args),
-	int flags, void *args, int tag);
-
-static inline int
-xfs_fs_eofblocks_from_user(
-	struct xfs_fs_eofblocks		*src,
-	struct xfs_eofblocks		*dst)
-{
-	if (src->eof_version != XFS_EOFBLOCKS_VERSION)
-		return -EINVAL;
-
-	if (src->eof_flags & ~XFS_EOF_FLAGS_VALID)
-		return -EINVAL;
-
-	if (memchr_inv(&src->pad32, 0, sizeof(src->pad32)) ||
-	    memchr_inv(src->pad64, 0, sizeof(src->pad64)))
-		return -EINVAL;
-
-	dst->eof_flags = src->eof_flags;
-	dst->eof_prid = src->eof_prid;
-	dst->eof_min_file_size = src->eof_min_file_size;
-
-	dst->eof_uid = INVALID_UID;
-	if (src->eof_flags & XFS_EOF_FLAGS_UID) {
-		dst->eof_uid = make_kuid(current_user_ns(), src->eof_uid);
-		if (!uid_valid(dst->eof_uid))
-			return -EINVAL;
-	}
-
-	dst->eof_gid = INVALID_GID;
-	if (src->eof_flags & XFS_EOF_FLAGS_GID) {
-		dst->eof_gid = make_kgid(current_user_ns(), src->eof_gid);
-		if (!gid_valid(dst->eof_gid))
-			return -EINVAL;
-	}
-	return 0;
-}
+int xfs_inode_walk(struct xfs_mount *mp, int iter_flags,
+	int (*execute)(struct xfs_inode *ip, void *args),
+	void *args, int tag);
 
 int xfs_icache_inode_is_allocated(struct xfs_mount *mp, struct xfs_trans *tp,
 				  xfs_ino_t ino, bool *inuse);
diff --git a/fs/xfs/xfs_icreate_item.c b/fs/xfs/xfs_icreate_item.c
index 490fee22b878..9b3994b9c716 100644
--- a/fs/xfs/xfs_icreate_item.c
+++ b/fs/xfs/xfs_icreate_item.c
@@ -6,11 +6,19 @@
 #include "xfs.h"
 #include "xfs_fs.h"
 #include "xfs_shared.h"
+#include "xfs_format.h"
 #include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_inode.h"
 #include "xfs_trans.h"
 #include "xfs_trans_priv.h"
 #include "xfs_icreate_item.h"
 #include "xfs_log.h"
+#include "xfs_log_priv.h"
+#include "xfs_log_recover.h"
+#include "xfs_ialloc.h"
+#include "xfs_trace.h"
 
 kmem_zone_t	*xfs_icreate_zone;		/* inode create item zone */
 
@@ -89,7 +97,7 @@ xfs_icreate_log(
 {
 	struct xfs_icreate_item	*icp;
 
-	icp = kmem_zone_zalloc(xfs_icreate_zone, 0);
+	icp = kmem_cache_zalloc(xfs_icreate_zone, GFP_KERNEL | __GFP_NOFAIL);
 
 	xfs_log_item_init(tp->t_mountp, &icp->ic_item, XFS_LI_ICREATE,
 			  &xfs_icreate_item_ops);
@@ -107,3 +115,147 @@ xfs_icreate_log(
 	tp->t_flags |= XFS_TRANS_DIRTY;
 	set_bit(XFS_LI_DIRTY, &icp->ic_item.li_flags);
 }
+
+static enum xlog_recover_reorder
+xlog_recover_icreate_reorder(
+		struct xlog_recover_item *item)
+{
+	/*
+	 * Inode allocation buffers must be replayed before subsequent inode
+	 * items try to modify those buffers.  ICREATE items are the logical
+	 * equivalent of logging a newly initialized inode buffer, so recover
+	 * these at the same time that we recover logged buffers.
+	 */
+	return XLOG_REORDER_BUFFER_LIST;
+}
+
+/*
+ * This routine is called when an inode create format structure is found in a
+ * committed transaction in the log.  It's purpose is to initialise the inodes
+ * being allocated on disk. This requires us to get inode cluster buffers that
+ * match the range to be initialised, stamped with inode templates and written
+ * by delayed write so that subsequent modifications will hit the cached buffer
+ * and only need writing out at the end of recovery.
+ */
+STATIC int
+xlog_recover_icreate_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_icreate_log		*icl;
+	struct xfs_ino_geometry		*igeo = M_IGEO(mp);
+	xfs_agnumber_t			agno;
+	xfs_agblock_t			agbno;
+	unsigned int			count;
+	unsigned int			isize;
+	xfs_agblock_t			length;
+	int				bb_per_cluster;
+	int				cancel_count;
+	int				nbufs;
+	int				i;
+
+	icl = (struct xfs_icreate_log *)item->ri_buf[0].i_addr;
+	if (icl->icl_type != XFS_LI_ICREATE) {
+		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad type");
+		return -EINVAL;
+	}
+
+	if (icl->icl_size != 1) {
+		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad icl size");
+		return -EINVAL;
+	}
+
+	agno = be32_to_cpu(icl->icl_ag);
+	if (agno >= mp->m_sb.sb_agcount) {
+		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad agno");
+		return -EINVAL;
+	}
+	agbno = be32_to_cpu(icl->icl_agbno);
+	if (!agbno || agbno == NULLAGBLOCK || agbno >= mp->m_sb.sb_agblocks) {
+		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad agbno");
+		return -EINVAL;
+	}
+	isize = be32_to_cpu(icl->icl_isize);
+	if (isize != mp->m_sb.sb_inodesize) {
+		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad isize");
+		return -EINVAL;
+	}
+	count = be32_to_cpu(icl->icl_count);
+	if (!count) {
+		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad count");
+		return -EINVAL;
+	}
+	length = be32_to_cpu(icl->icl_length);
+	if (!length || length >= mp->m_sb.sb_agblocks) {
+		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad length");
+		return -EINVAL;
+	}
+
+	/*
+	 * The inode chunk is either full or sparse and we only support
+	 * m_ino_geo.ialloc_min_blks sized sparse allocations at this time.
+	 */
+	if (length != igeo->ialloc_blks &&
+	    length != igeo->ialloc_min_blks) {
+		xfs_warn(log->l_mp,
+			 "%s: unsupported chunk length", __FUNCTION__);
+		return -EINVAL;
+	}
+
+	/* verify inode count is consistent with extent length */
+	if ((count >> mp->m_sb.sb_inopblog) != length) {
+		xfs_warn(log->l_mp,
+			 "%s: inconsistent inode count and chunk length",
+			 __FUNCTION__);
+		return -EINVAL;
+	}
+
+	/*
+	 * The icreate transaction can cover multiple cluster buffers and these
+	 * buffers could have been freed and reused. Check the individual
+	 * buffers for cancellation so we don't overwrite anything written after
+	 * a cancellation.
+	 */
+	bb_per_cluster = XFS_FSB_TO_BB(mp, igeo->blocks_per_cluster);
+	nbufs = length / igeo->blocks_per_cluster;
+	for (i = 0, cancel_count = 0; i < nbufs; i++) {
+		xfs_daddr_t	daddr;
+
+		daddr = XFS_AGB_TO_DADDR(mp, agno,
+				agbno + i * igeo->blocks_per_cluster);
+		if (xlog_is_buffer_cancelled(log, daddr, bb_per_cluster))
+			cancel_count++;
+	}
+
+	/*
+	 * We currently only use icreate for a single allocation at a time. This
+	 * means we should expect either all or none of the buffers to be
+	 * cancelled. Be conservative and skip replay if at least one buffer is
+	 * cancelled, but warn the user that something is awry if the buffers
+	 * are not consistent.
+	 *
+	 * XXX: This must be refined to only skip cancelled clusters once we use
+	 * icreate for multiple chunk allocations.
+	 */
+	ASSERT(!cancel_count || cancel_count == nbufs);
+	if (cancel_count) {
+		if (cancel_count != nbufs)
+			xfs_warn(mp,
+	"WARNING: partial inode chunk cancellation, skipped icreate.");
+		trace_xfs_log_recover_icreate_cancel(log, icl);
+		return 0;
+	}
+
+	trace_xfs_log_recover_icreate_recover(log, icl);
+	return xfs_ialloc_inode_init(mp, NULL, buffer_list, count, agno, agbno,
+				     length, be32_to_cpu(icl->icl_gen));
+}
+
+const struct xlog_recover_item_ops xlog_icreate_item_ops = {
+	.item_type		= XFS_LI_ICREATE,
+	.reorder		= xlog_recover_icreate_reorder,
+	.commit_pass2		= xlog_recover_icreate_commit_pass2,
+};
diff --git a/fs/xfs/xfs_inode.c b/fs/xfs/xfs_inode.c
index c5077e6326c7..c06129cffba9 100644
--- a/fs/xfs/xfs_inode.c
+++ b/fs/xfs/xfs_inode.c
@@ -44,7 +44,6 @@ kmem_zone_t *xfs_inode_zone;
  */
 #define	XFS_ITRUNC_MAX_EXTENTS	2
 
-STATIC int xfs_iflush_int(struct xfs_inode *, struct xfs_buf *);
 STATIC int xfs_iunlink(struct xfs_trans *, struct xfs_inode *);
 STATIC int xfs_iunlink_remove(struct xfs_trans *, struct xfs_inode *);
 
@@ -112,7 +111,7 @@ xfs_ilock_data_map_shared(
 {
 	uint			lock_mode = XFS_ILOCK_SHARED;
 
-	if (ip->i_d.di_format == XFS_DINODE_FMT_BTREE &&
+	if (ip->i_df.if_format == XFS_DINODE_FMT_BTREE &&
 	    (ip->i_df.if_flags & XFS_IFEXTENTS) == 0)
 		lock_mode = XFS_ILOCK_EXCL;
 	xfs_ilock(ip, lock_mode);
@@ -125,7 +124,8 @@ xfs_ilock_attr_map_shared(
 {
 	uint			lock_mode = XFS_ILOCK_SHARED;
 
-	if (ip->i_d.di_aformat == XFS_DINODE_FMT_BTREE &&
+	if (ip->i_afp &&
+	    ip->i_afp->if_format == XFS_DINODE_FMT_BTREE &&
 	    (ip->i_afp->if_flags & XFS_IFEXTENTS) == 0)
 		lock_mode = XFS_ILOCK_EXCL;
 	xfs_ilock(ip, lock_mode);
@@ -144,17 +144,17 @@ xfs_ilock_attr_map_shared(
  *
  * i_rwsem -> i_mmap_lock -> page_lock -> i_ilock
  *
- * mmap_sem locking order:
+ * mmap_lock locking order:
  *
- * i_rwsem -> page lock -> mmap_sem
- * mmap_sem -> i_mmap_lock -> page_lock
+ * i_rwsem -> page lock -> mmap_lock
+ * mmap_lock -> i_mmap_lock -> page_lock
  *
- * The difference in mmap_sem locking order mean that we cannot hold the
+ * The difference in mmap_lock locking order mean that we cannot hold the
  * i_mmap_lock over syscall based read(2)/write(2) based IO. These IO paths can
- * fault in pages during copy in/out (for buffered IO) or require the mmap_sem
+ * fault in pages during copy in/out (for buffered IO) or require the mmap_lock
  * in get_user_pages() to map the user pages into the kernel address space for
  * direct IO. Similarly the i_rwsem cannot be taken inside a page fault because
- * page faults already hold the mmap_sem.
+ * page faults already hold the mmap_lock.
  *
  * Hence to serialise fully against both syscall and mmap based IO, we need to
  * take both the i_rwsem and the i_mmap_lock. These locks should *only* be both
@@ -451,7 +451,7 @@ xfs_lock_inodes(
 	/*
 	 * Currently supports between 2 and 5 inodes with exclusive locking.  We
 	 * support an arbitrary depth of locking here, but absolute limits on
-	 * inodes depend on the the type of locking and the limits placed by
+	 * inodes depend on the type of locking and the limits placed by
 	 * lockdep annotations in xfs_lock_inumorder.  These are all checked by
 	 * the asserts.
 	 */
@@ -801,26 +801,18 @@ xfs_ialloc(
 		return error;
 	ASSERT(ip != NULL);
 	inode = VFS_I(ip);
-
-	/*
-	 * 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;
-
 	inode->i_mode = mode;
 	set_nlink(inode, nlink);
-	ip->i_d.di_uid = xfs_kuid_to_uid(current_fsuid());
-	ip->i_d.di_gid = xfs_kgid_to_gid(current_fsgid());
+	inode->i_uid = current_fsuid();
 	inode->i_rdev = rdev;
 	ip->i_d.di_projid = prid;
 
 	if (pip && XFS_INHERIT_GID(pip)) {
-		ip->i_d.di_gid = pip->i_d.di_gid;
+		inode->i_gid = VFS_I(pip)->i_gid;
 		if ((VFS_I(pip)->i_mode & S_ISGID) && S_ISDIR(mode))
 			inode->i_mode |= S_ISGID;
+	} else {
+		inode->i_gid = current_fsgid();
 	}
 
 	/*
@@ -828,13 +820,12 @@ xfs_ialloc(
 	 * ID or one of the supplementary group IDs, the S_ISGID bit is cleared
 	 * (and only if the irix_sgid_inherit compatibility variable is set).
 	 */
-	if ((irix_sgid_inherit) &&
-	    (inode->i_mode & S_ISGID) &&
-	    (!in_group_p(xfs_gid_to_kgid(ip->i_d.di_gid))))
+	if (irix_sgid_inherit &&
+	    (inode->i_mode & S_ISGID) && !in_group_p(inode->i_gid))
 		inode->i_mode &= ~S_ISGID;
 
 	ip->i_d.di_size = 0;
-	ip->i_d.di_nextents = 0;
+	ip->i_df.if_nextents = 0;
 	ASSERT(ip->i_d.di_nblocks == 0);
 
 	tv = current_time(inode);
@@ -847,21 +838,20 @@ xfs_ialloc(
 	ip->i_d.di_dmstate = 0;
 	ip->i_d.di_flags = 0;
 
-	if (ip->i_d.di_version == 3) {
+	if (xfs_sb_version_has_v3inode(&mp->m_sb)) {
 		inode_set_iversion(inode, 1);
 		ip->i_d.di_flags2 = 0;
 		ip->i_d.di_cowextsize = 0;
 		ip->i_d.di_crtime = tv;
 	}
 
-
 	flags = XFS_ILOG_CORE;
 	switch (mode & S_IFMT) {
 	case S_IFIFO:
 	case S_IFCHR:
 	case S_IFBLK:
 	case S_IFSOCK:
-		ip->i_d.di_format = XFS_DINODE_FMT_DEV;
+		ip->i_df.if_format = XFS_DINODE_FMT_DEV;
 		ip->i_df.if_flags = 0;
 		flags |= XFS_ILOG_DEV;
 		break;
@@ -907,24 +897,17 @@ xfs_ialloc(
 
 			ip->i_d.di_flags |= di_flags;
 		}
-		if (pip &&
-		    (pip->i_d.di_flags2 & XFS_DIFLAG2_ANY) &&
-		    pip->i_d.di_version == 3 &&
-		    ip->i_d.di_version == 3) {
-			uint64_t	di_flags2 = 0;
-
+		if (pip && (pip->i_d.di_flags2 & XFS_DIFLAG2_ANY)) {
 			if (pip->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE) {
-				di_flags2 |= XFS_DIFLAG2_COWEXTSIZE;
+				ip->i_d.di_flags2 |= XFS_DIFLAG2_COWEXTSIZE;
 				ip->i_d.di_cowextsize = pip->i_d.di_cowextsize;
 			}
 			if (pip->i_d.di_flags2 & XFS_DIFLAG2_DAX)
-				di_flags2 |= XFS_DIFLAG2_DAX;
-
-			ip->i_d.di_flags2 |= di_flags2;
+				ip->i_d.di_flags2 |= XFS_DIFLAG2_DAX;
 		}
 		/* FALLTHROUGH */
 	case S_IFLNK:
-		ip->i_d.di_format = XFS_DINODE_FMT_EXTENTS;
+		ip->i_df.if_format = XFS_DINODE_FMT_EXTENTS;
 		ip->i_df.if_flags = XFS_IFEXTENTS;
 		ip->i_df.if_bytes = 0;
 		ip->i_df.if_u1.if_root = NULL;
@@ -932,11 +915,6 @@ xfs_ialloc(
 	default:
 		ASSERT(0);
 	}
-	/*
-	 * Attribute fork settings for new inode.
-	 */
-	ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
-	ip->i_d.di_anextents = 0;
 
 	/*
 	 * Log the new values stuffed into the inode.
@@ -1122,7 +1100,6 @@ xfs_bumplink(
 {
 	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
 
-	ASSERT(ip->i_d.di_version > 1);
 	inc_nlink(VFS_I(ip));
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 }
@@ -1158,8 +1135,7 @@ xfs_create(
 	/*
 	 * Make sure that we have allocated dquot(s) on disk.
 	 */
-	error = xfs_qm_vop_dqalloc(dp, xfs_kuid_to_uid(current_fsuid()),
-					xfs_kgid_to_gid(current_fsgid()), prid,
+	error = xfs_qm_vop_dqalloc(dp, current_fsuid(), current_fsgid(), prid,
 					XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
 					&udqp, &gdqp, &pdqp);
 	if (error)
@@ -1219,8 +1195,7 @@ xfs_create(
 	unlock_dp_on_error = false;
 
 	error = xfs_dir_createname(tp, dp, name, ip->i_ino,
-				   resblks ?
-					resblks - XFS_IALLOC_SPACE_RES(mp) : 0);
+					resblks - XFS_IALLOC_SPACE_RES(mp));
 	if (error) {
 		ASSERT(error != -ENOSPC);
 		goto out_trans_cancel;
@@ -1309,8 +1284,7 @@ xfs_create_tmpfile(
 	/*
 	 * Make sure that we have allocated dquot(s) on disk.
 	 */
-	error = xfs_qm_vop_dqalloc(dp, xfs_kuid_to_uid(current_fsuid()),
-				xfs_kgid_to_gid(current_fsgid()), prid,
+	error = xfs_qm_vop_dqalloc(dp, current_fsuid(), current_fsgid(), prid,
 				XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
 				&udqp, &gdqp, &pdqp);
 	if (error)
@@ -1655,7 +1629,7 @@ xfs_release(
 			return 0;
 		/*
 		 * If we can't get the iolock just skip truncating the blocks
-		 * past EOF because we could deadlock with the mmap_sem
+		 * past EOF because we could deadlock with the mmap_lock
 		 * otherwise. We'll get another chance to drop them once the
 		 * last reference to the inode is dropped, so we'll never leak
 		 * blocks permanently.
@@ -1707,7 +1681,7 @@ xfs_inactive_truncate(
 	if (error)
 		goto error_trans_cancel;
 
-	ASSERT(ip->i_d.di_nextents == 0);
+	ASSERT(ip->i_df.if_nextents == 0);
 
 	error = xfs_trans_commit(tp);
 	if (error)
@@ -1765,10 +1739,31 @@ xfs_inactive_ifree(
 		return error;
 	}
 
+	/*
+	 * We do not hold the inode locked across the entire rolling transaction
+	 * here. We only need to hold it for the first transaction that
+	 * xfs_ifree() builds, which may mark the inode XFS_ISTALE if the
+	 * underlying cluster buffer is freed. Relogging an XFS_ISTALE inode
+	 * here breaks the relationship between cluster buffer invalidation and
+	 * stale inode invalidation on cluster buffer item journal commit
+	 * completion, and can result in leaving dirty stale inodes hanging
+	 * around in memory.
+	 *
+	 * We have no need for serialising this inode operation against other
+	 * operations - we freed the inode and hence reallocation is required
+	 * and that will serialise on reallocating the space the deferops need
+	 * to free. Hence we can unlock the inode on the first commit of
+	 * the transaction rather than roll it right through the deferops. This
+	 * avoids relogging the XFS_ISTALE inode.
+	 *
+	 * We check that xfs_ifree() hasn't grown an internal transaction roll
+	 * by asserting that the inode is still locked when it returns.
+	 */
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	xfs_trans_ijoin(tp, ip, 0);
+	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 
 	error = xfs_ifree(tp, ip);
+	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 	if (error) {
 		/*
 		 * If we fail to free the inode, shut down.  The cancel
@@ -1781,7 +1776,6 @@ xfs_inactive_ifree(
 			xfs_force_shutdown(mp, SHUTDOWN_META_IO_ERROR);
 		}
 		xfs_trans_cancel(tp);
-		xfs_iunlock(ip, XFS_ILOCK_EXCL);
 		return error;
 	}
 
@@ -1799,7 +1793,6 @@ xfs_inactive_ifree(
 		xfs_notice(mp, "%s: xfs_trans_commit returned error %d",
 			__func__, error);
 
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 	return 0;
 }
 
@@ -1857,7 +1850,7 @@ xfs_inactive(
 
 	if (S_ISREG(VFS_I(ip)->i_mode) &&
 	    (ip->i_d.di_size != 0 || XFS_ISIZE(ip) != 0 ||
-	     ip->i_d.di_nextents > 0 || ip->i_delayed_blks > 0))
+	     ip->i_df.if_nextents > 0 || ip->i_delayed_blks > 0))
 		truncate = 1;
 
 	error = xfs_qm_dqattach(ip);
@@ -1883,7 +1876,6 @@ xfs_inactive(
 	}
 
 	ASSERT(!ip->i_afp);
-	ASSERT(ip->i_d.di_anextents == 0);
 	ASSERT(ip->i_d.di_forkoff == 0);
 
 	/*
@@ -2119,7 +2111,7 @@ xfs_iunlink_update_bucket(
 	unsigned int		bucket_index,
 	xfs_agino_t		new_agino)
 {
-	struct xfs_agi		*agi = XFS_BUF_TO_AGI(agibp);
+	struct xfs_agi		*agi = agibp->b_addr;
 	xfs_agino_t		old_value;
 	int			offset;
 
@@ -2135,7 +2127,7 @@ xfs_iunlink_update_bucket(
 	 * head of the list.
 	 */
 	if (old_value == new_agino) {
-		xfs_buf_corruption_error(agibp);
+		xfs_buf_mark_corrupt(agibp);
 		return -EFSCORRUPTED;
 	}
 
@@ -2173,7 +2165,6 @@ xfs_iunlink_update_dinode(
 	xfs_dinode_calc_crc(mp, dip);
 	xfs_trans_inode_buf(tp, ibp);
 	xfs_trans_log_buf(tp, ibp, offset, offset + sizeof(xfs_agino_t) - 1);
-	xfs_inobp_check(mp, ibp);
 }
 
 /* Set an in-core inode's unlinked pointer and return the old value. */
@@ -2193,7 +2184,7 @@ xfs_iunlink_update_inode(
 
 	ASSERT(xfs_verify_agino_or_null(mp, agno, next_agino));
 
-	error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &ibp, 0, 0);
+	error = xfs_imap_to_bp(mp, tp, &ip->i_imap, &dip, &ibp, 0);
 	if (error)
 		return error;
 
@@ -2259,7 +2250,7 @@ xfs_iunlink(
 	error = xfs_read_agi(mp, tp, agno, &agibp);
 	if (error)
 		return error;
-	agi = XFS_BUF_TO_AGI(agibp);
+	agi = agibp->b_addr;
 
 	/*
 	 * Get the index into the agi hash table for the list this inode will
@@ -2269,12 +2260,11 @@ xfs_iunlink(
 	next_agino = be32_to_cpu(agi->agi_unlinked[bucket_index]);
 	if (next_agino == agino ||
 	    !xfs_verify_agino_or_null(mp, agno, next_agino)) {
-		xfs_buf_corruption_error(agibp);
+		xfs_buf_mark_corrupt(agibp);
 		return -EFSCORRUPTED;
 	}
 
 	if (next_agino != NULLAGINO) {
-		struct xfs_perag	*pag;
 		xfs_agino_t		old_agino;
 
 		/*
@@ -2291,9 +2281,7 @@ xfs_iunlink(
 		 * agino has been unlinked, add a backref from the next inode
 		 * back to agino.
 		 */
-		pag = xfs_perag_get(mp, agno);
-		error = xfs_iunlink_add_backref(pag, agino, next_agino);
-		xfs_perag_put(pag);
+		error = xfs_iunlink_add_backref(agibp->b_pag, agino, next_agino);
 		if (error)
 			return error;
 	}
@@ -2323,7 +2311,7 @@ xfs_iunlink_map_ino(
 		return error;
 	}
 
-	error = xfs_imap_to_bp(mp, tp, imap, dipp, bpp, 0, 0);
+	error = xfs_imap_to_bp(mp, tp, imap, dipp, bpp, 0);
 	if (error) {
 		xfs_warn(mp, "%s: xfs_imap_to_bp returned error %d.",
 				__func__, error);
@@ -2429,7 +2417,6 @@ xfs_iunlink_remove(
 	struct xfs_buf		*agibp;
 	struct xfs_buf		*last_ibp;
 	struct xfs_dinode	*last_dip = NULL;
-	struct xfs_perag	*pag = NULL;
 	xfs_agnumber_t		agno = XFS_INO_TO_AGNO(mp, ip->i_ino);
 	xfs_agino_t		agino = XFS_INO_TO_AGINO(mp, ip->i_ino);
 	xfs_agino_t		next_agino;
@@ -2443,7 +2430,7 @@ xfs_iunlink_remove(
 	error = xfs_read_agi(mp, tp, agno, &agibp);
 	if (error)
 		return error;
-	agi = XFS_BUF_TO_AGI(agibp);
+	agi = agibp->b_addr;
 
 	/*
 	 * Get the index into the agi hash table for the list this inode will
@@ -2473,32 +2460,22 @@ xfs_iunlink_remove(
 	 * this inode's backref to point from the next inode.
 	 */
 	if (next_agino != NULLAGINO) {
-		pag = xfs_perag_get(mp, agno);
-		error = xfs_iunlink_change_backref(pag, next_agino,
+		error = xfs_iunlink_change_backref(agibp->b_pag, next_agino,
 				NULLAGINO);
 		if (error)
-			goto out;
+			return error;
 	}
 
-	if (head_agino == agino) {
-		/* Point the head of the list to the next unlinked inode. */
-		error = xfs_iunlink_update_bucket(tp, agno, agibp, bucket_index,
-				next_agino);
-		if (error)
-			goto out;
-	} else {
+	if (head_agino != agino) {
 		struct xfs_imap	imap;
 		xfs_agino_t	prev_agino;
 
-		if (!pag)
-			pag = xfs_perag_get(mp, agno);
-
 		/* We need to search the list for the inode being freed. */
 		error = xfs_iunlink_map_prev(tp, agno, head_agino, agino,
 				&prev_agino, &imap, &last_dip, &last_ibp,
-				pag);
+				agibp->b_pag);
 		if (error)
-			goto out;
+			return error;
 
 		/* Point the previous inode on the list to the next inode. */
 		xfs_iunlink_update_dinode(tp, agno, prev_agino, last_ibp,
@@ -2512,15 +2489,106 @@ xfs_iunlink_remove(
 		 * change_backref takes care of deleting the backref if
 		 * next_agino is NULLAGINO.
 		 */
-		error = xfs_iunlink_change_backref(pag, agino, next_agino);
-		if (error)
-			goto out;
+		return xfs_iunlink_change_backref(agibp->b_pag, agino,
+				next_agino);
 	}
 
-out:
-	if (pag)
-		xfs_perag_put(pag);
-	return error;
+	/* Point the head of the list to the next unlinked inode. */
+	return xfs_iunlink_update_bucket(tp, agno, agibp, bucket_index,
+			next_agino);
+}
+
+/*
+ * Look up the inode number specified and if it is not already marked XFS_ISTALE
+ * mark it stale. We should only find clean inodes in this lookup that aren't
+ * already stale.
+ */
+static void
+xfs_ifree_mark_inode_stale(
+	struct xfs_buf		*bp,
+	struct xfs_inode	*free_ip,
+	xfs_ino_t		inum)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+	struct xfs_perag	*pag = bp->b_pag;
+	struct xfs_inode_log_item *iip;
+	struct xfs_inode	*ip;
+
+retry:
+	rcu_read_lock();
+	ip = radix_tree_lookup(&pag->pag_ici_root, XFS_INO_TO_AGINO(mp, inum));
+
+	/* Inode not in memory, nothing to do */
+	if (!ip) {
+		rcu_read_unlock();
+		return;
+	}
+
+	/*
+	 * because this is an RCU protected lookup, we could find a recently
+	 * freed or even reallocated inode during the lookup. We need to check
+	 * under the i_flags_lock for a valid inode here. Skip it if it is not
+	 * valid, the wrong inode or stale.
+	 */
+	spin_lock(&ip->i_flags_lock);
+	if (ip->i_ino != inum || __xfs_iflags_test(ip, XFS_ISTALE)) {
+		spin_unlock(&ip->i_flags_lock);
+		rcu_read_unlock();
+		return;
+	}
+
+	/*
+	 * Don't try to lock/unlock the current inode, but we _cannot_ skip the
+	 * other inodes that we did not find in the list attached to the buffer
+	 * and are not already marked stale. If we can't lock it, back off and
+	 * retry.
+	 */
+	if (ip != free_ip) {
+		if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) {
+			spin_unlock(&ip->i_flags_lock);
+			rcu_read_unlock();
+			delay(1);
+			goto retry;
+		}
+	}
+	ip->i_flags |= XFS_ISTALE;
+	spin_unlock(&ip->i_flags_lock);
+	rcu_read_unlock();
+
+	/*
+	 * If we can't get the flush lock, the inode is already attached.  All
+	 * we needed to do here is mark the inode stale so buffer IO completion
+	 * will remove it from the AIL.
+	 */
+	iip = ip->i_itemp;
+	if (!xfs_iflock_nowait(ip)) {
+		ASSERT(!list_empty(&iip->ili_item.li_bio_list));
+		ASSERT(iip->ili_last_fields);
+		goto out_iunlock;
+	}
+
+	/*
+	 * Inodes not attached to the buffer can be released immediately.
+	 * Everything else has to go through xfs_iflush_abort() on journal
+	 * commit as the flock synchronises removal of the inode from the
+	 * cluster buffer against inode reclaim.
+	 */
+	if (!iip || list_empty(&iip->ili_item.li_bio_list)) {
+		xfs_ifunlock(ip);
+		goto out_iunlock;
+	}
+
+	/* we have a dirty inode in memory that has not yet been flushed. */
+	spin_lock(&iip->ili_lock);
+	iip->ili_last_fields = iip->ili_fields;
+	iip->ili_fields = 0;
+	iip->ili_fsync_fields = 0;
+	spin_unlock(&iip->ili_lock);
+	ASSERT(iip->ili_last_fields);
+
+out_iunlock:
+	if (ip != free_ip)
+		xfs_iunlock(ip, XFS_ILOCK_EXCL);
 }
 
 /*
@@ -2530,26 +2598,20 @@ out:
  */
 STATIC int
 xfs_ifree_cluster(
-	xfs_inode_t		*free_ip,
-	xfs_trans_t		*tp,
+	struct xfs_inode	*free_ip,
+	struct xfs_trans	*tp,
 	struct xfs_icluster	*xic)
 {
-	xfs_mount_t		*mp = free_ip->i_mount;
+	struct xfs_mount	*mp = free_ip->i_mount;
+	struct xfs_ino_geometry	*igeo = M_IGEO(mp);
+	struct xfs_buf		*bp;
+	xfs_daddr_t		blkno;
+	xfs_ino_t		inum = xic->first_ino;
 	int			nbufs;
 	int			i, j;
 	int			ioffset;
-	xfs_daddr_t		blkno;
-	xfs_buf_t		*bp;
-	xfs_inode_t		*ip;
-	xfs_inode_log_item_t	*iip;
-	struct xfs_log_item	*lip;
-	struct xfs_perag	*pag;
-	struct xfs_ino_geometry	*igeo = M_IGEO(mp);
-	xfs_ino_t		inum;
 	int			error;
 
-	inum = xic->first_ino;
-	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, inum));
 	nbufs = igeo->ialloc_blks / igeo->blocks_per_cluster;
 
 	for (j = 0; j < nbufs; j++, inum += igeo->inodes_per_cluster) {
@@ -2593,148 +2655,20 @@ xfs_ifree_cluster(
 		bp->b_ops = &xfs_inode_buf_ops;
 
 		/*
-		 * Walk the inodes already attached to the buffer and mark them
-		 * stale. These will all have the flush locks held, so an
-		 * in-memory inode walk can't lock them. By marking them all
-		 * stale first, we will not attempt to lock them in the loop
-		 * below as the XFS_ISTALE flag will be set.
+		 * Now we need to set all the cached clean inodes as XFS_ISTALE,
+		 * too. This requires lookups, and will skip inodes that we've
+		 * already marked XFS_ISTALE.
 		 */
-		list_for_each_entry(lip, &bp->b_li_list, li_bio_list) {
-			if (lip->li_type == XFS_LI_INODE) {
-				iip = (xfs_inode_log_item_t *)lip;
-				ASSERT(iip->ili_logged == 1);
-				lip->li_cb = xfs_istale_done;
-				xfs_trans_ail_copy_lsn(mp->m_ail,
-							&iip->ili_flush_lsn,
-							&iip->ili_item.li_lsn);
-				xfs_iflags_set(iip->ili_inode, XFS_ISTALE);
-			}
-		}
-
-
-		/*
-		 * For each inode in memory attempt to add it to the inode
-		 * buffer and set it up for being staled on buffer IO
-		 * completion.  This is safe as we've locked out tail pushing
-		 * and flushing by locking the buffer.
-		 *
-		 * We have already marked every inode that was part of a
-		 * transaction stale above, which means there is no point in
-		 * even trying to lock them.
-		 */
-		for (i = 0; i < igeo->inodes_per_cluster; i++) {
-retry:
-			rcu_read_lock();
-			ip = radix_tree_lookup(&pag->pag_ici_root,
-					XFS_INO_TO_AGINO(mp, (inum + i)));
-
-			/* Inode not in memory, nothing to do */
-			if (!ip) {
-				rcu_read_unlock();
-				continue;
-			}
-
-			/*
-			 * because this is an RCU protected lookup, we could
-			 * find a recently freed or even reallocated inode
-			 * during the lookup. We need to check under the
-			 * i_flags_lock for a valid inode here. Skip it if it
-			 * is not valid, the wrong inode or stale.
-			 */
-			spin_lock(&ip->i_flags_lock);
-			if (ip->i_ino != inum + i ||
-			    __xfs_iflags_test(ip, XFS_ISTALE)) {
-				spin_unlock(&ip->i_flags_lock);
-				rcu_read_unlock();
-				continue;
-			}
-			spin_unlock(&ip->i_flags_lock);
-
-			/*
-			 * Don't try to lock/unlock the current inode, but we
-			 * _cannot_ skip the other inodes that we did not find
-			 * in the list attached to the buffer and are not
-			 * already marked stale. If we can't lock it, back off
-			 * and retry.
-			 */
-			if (ip != free_ip) {
-				if (!xfs_ilock_nowait(ip, XFS_ILOCK_EXCL)) {
-					rcu_read_unlock();
-					delay(1);
-					goto retry;
-				}
-
-				/*
-				 * Check the inode number again in case we're
-				 * racing with freeing in xfs_reclaim_inode().
-				 * See the comments in that function for more
-				 * information as to why the initial check is
-				 * not sufficient.
-				 */
-				if (ip->i_ino != inum + i) {
-					xfs_iunlock(ip, XFS_ILOCK_EXCL);
-					rcu_read_unlock();
-					continue;
-				}
-			}
-			rcu_read_unlock();
-
-			xfs_iflock(ip);
-			xfs_iflags_set(ip, XFS_ISTALE);
-
-			/*
-			 * we don't need to attach clean inodes or those only
-			 * with unlogged changes (which we throw away, anyway).
-			 */
-			iip = ip->i_itemp;
-			if (!iip || xfs_inode_clean(ip)) {
-				ASSERT(ip != free_ip);
-				xfs_ifunlock(ip);
-				xfs_iunlock(ip, XFS_ILOCK_EXCL);
-				continue;
-			}
-
-			iip->ili_last_fields = iip->ili_fields;
-			iip->ili_fields = 0;
-			iip->ili_fsync_fields = 0;
-			iip->ili_logged = 1;
-			xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
-						&iip->ili_item.li_lsn);
-
-			xfs_buf_attach_iodone(bp, xfs_istale_done,
-						  &iip->ili_item);
-
-			if (ip != free_ip)
-				xfs_iunlock(ip, XFS_ILOCK_EXCL);
-		}
+		for (i = 0; i < igeo->inodes_per_cluster; i++)
+			xfs_ifree_mark_inode_stale(bp, free_ip, inum + i);
 
 		xfs_trans_stale_inode_buf(tp, bp);
 		xfs_trans_binval(tp, bp);
 	}
-
-	xfs_perag_put(pag);
 	return 0;
 }
 
 /*
- * Free any local-format buffers sitting around before we reset to
- * extents format.
- */
-static inline void
-xfs_ifree_local_data(
-	struct xfs_inode	*ip,
-	int			whichfork)
-{
-	struct xfs_ifork	*ifp;
-
-	if (XFS_IFORK_FORMAT(ip, whichfork) != XFS_DINODE_FMT_LOCAL)
-		return;
-
-	ifp = XFS_IFORK_PTR(ip, whichfork);
-	xfs_idata_realloc(ip, -ifp->if_bytes, whichfork);
-}
-
-/*
  * This is called to return an inode to the inode free list.
  * The inode should already be truncated to 0 length and have
  * no pages associated with it.  This routine also assumes that
@@ -2751,11 +2685,11 @@ xfs_ifree(
 {
 	int			error;
 	struct xfs_icluster	xic = { 0 };
+	struct xfs_inode_log_item *iip = ip->i_itemp;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 	ASSERT(VFS_I(ip)->i_nlink == 0);
-	ASSERT(ip->i_d.di_nextents == 0);
-	ASSERT(ip->i_d.di_anextents == 0);
+	ASSERT(ip->i_df.if_nextents == 0);
 	ASSERT(ip->i_d.di_size == 0 || !S_ISREG(VFS_I(ip)->i_mode));
 	ASSERT(ip->i_d.di_nblocks == 0);
 
@@ -2770,19 +2704,28 @@ xfs_ifree(
 	if (error)
 		return error;
 
-	xfs_ifree_local_data(ip, XFS_DATA_FORK);
-	xfs_ifree_local_data(ip, XFS_ATTR_FORK);
+	/*
+	 * Free any local-format data sitting around before we reset the
+	 * data fork to extents format.  Note that the attr fork data has
+	 * already been freed by xfs_attr_inactive.
+	 */
+	if (ip->i_df.if_format == XFS_DINODE_FMT_LOCAL) {
+		kmem_free(ip->i_df.if_u1.if_data);
+		ip->i_df.if_u1.if_data = NULL;
+		ip->i_df.if_bytes = 0;
+	}
 
 	VFS_I(ip)->i_mode = 0;		/* mark incore inode as free */
 	ip->i_d.di_flags = 0;
 	ip->i_d.di_flags2 = 0;
 	ip->i_d.di_dmevmask = 0;
 	ip->i_d.di_forkoff = 0;		/* mark the attr fork not in use */
-	ip->i_d.di_format = XFS_DINODE_FMT_EXTENTS;
-	ip->i_d.di_aformat = XFS_DINODE_FMT_EXTENTS;
+	ip->i_df.if_format = XFS_DINODE_FMT_EXTENTS;
 
 	/* Don't attempt to replay owner changes for a deleted inode */
-	ip->i_itemp->ili_fields &= ~(XFS_ILOG_AOWNER|XFS_ILOG_DOWNER);
+	spin_lock(&iip->ili_lock);
+	iip->ili_fields &= ~(XFS_ILOG_AOWNER | XFS_ILOG_DOWNER);
+	spin_unlock(&iip->ili_lock);
 
 	/*
 	 * Bump the generation count so no one will be confused
@@ -3162,7 +3105,7 @@ out_trans_abort:
 /*
  * xfs_rename_alloc_whiteout()
  *
- * Return a referenced, unlinked, unlocked inode that that can be used as a
+ * Return a referenced, unlinked, unlocked inode that can be used as a
  * whiteout in a rename transaction. We use a tmpfile inode here so that if we
  * crash between allocating the inode and linking it into the rename transaction
  * recovery will free the inode and we won't leak it.
@@ -3489,374 +3432,76 @@ out_release_wip:
 	return error;
 }
 
-STATIC int
-xfs_iflush_cluster(
-	struct xfs_inode	*ip,
-	struct xfs_buf		*bp)
-{
-	struct xfs_mount	*mp = ip->i_mount;
-	struct xfs_perag	*pag;
-	unsigned long		first_index, mask;
-	int			cilist_size;
-	struct xfs_inode	**cilist;
-	struct xfs_inode	*cip;
-	struct xfs_ino_geometry	*igeo = M_IGEO(mp);
-	int			nr_found;
-	int			clcount = 0;
-	int			i;
-
-	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
-
-	cilist_size = igeo->inodes_per_cluster * sizeof(struct xfs_inode *);
-	cilist = kmem_alloc(cilist_size, KM_MAYFAIL|KM_NOFS);
-	if (!cilist)
-		goto out_put;
-
-	mask = ~(igeo->inodes_per_cluster - 1);
-	first_index = XFS_INO_TO_AGINO(mp, ip->i_ino) & mask;
-	rcu_read_lock();
-	/* really need a gang lookup range call here */
-	nr_found = radix_tree_gang_lookup(&pag->pag_ici_root, (void**)cilist,
-					first_index, igeo->inodes_per_cluster);
-	if (nr_found == 0)
-		goto out_free;
-
-	for (i = 0; i < nr_found; i++) {
-		cip = cilist[i];
-		if (cip == ip)
-			continue;
-
-		/*
-		 * because this is an RCU protected lookup, we could find a
-		 * recently freed or even reallocated inode during the lookup.
-		 * We need to check under the i_flags_lock for a valid inode
-		 * here. Skip it if it is not valid or the wrong inode.
-		 */
-		spin_lock(&cip->i_flags_lock);
-		if (!cip->i_ino ||
-		    __xfs_iflags_test(cip, XFS_ISTALE)) {
-			spin_unlock(&cip->i_flags_lock);
-			continue;
-		}
-
-		/*
-		 * Once we fall off the end of the cluster, no point checking
-		 * any more inodes in the list because they will also all be
-		 * outside the cluster.
-		 */
-		if ((XFS_INO_TO_AGINO(mp, cip->i_ino) & mask) != first_index) {
-			spin_unlock(&cip->i_flags_lock);
-			break;
-		}
-		spin_unlock(&cip->i_flags_lock);
-
-		/*
-		 * Do an un-protected check to see if the inode is dirty and
-		 * is a candidate for flushing.  These checks will be repeated
-		 * later after the appropriate locks are acquired.
-		 */
-		if (xfs_inode_clean(cip) && xfs_ipincount(cip) == 0)
-			continue;
-
-		/*
-		 * Try to get locks.  If any are unavailable or it is pinned,
-		 * then this inode cannot be flushed and is skipped.
-		 */
-
-		if (!xfs_ilock_nowait(cip, XFS_ILOCK_SHARED))
-			continue;
-		if (!xfs_iflock_nowait(cip)) {
-			xfs_iunlock(cip, XFS_ILOCK_SHARED);
-			continue;
-		}
-		if (xfs_ipincount(cip)) {
-			xfs_ifunlock(cip);
-			xfs_iunlock(cip, XFS_ILOCK_SHARED);
-			continue;
-		}
-
-
-		/*
-		 * Check the inode number again, just to be certain we are not
-		 * racing with freeing in xfs_reclaim_inode(). See the comments
-		 * in that function for more information as to why the initial
-		 * check is not sufficient.
-		 */
-		if (!cip->i_ino) {
-			xfs_ifunlock(cip);
-			xfs_iunlock(cip, XFS_ILOCK_SHARED);
-			continue;
-		}
-
-		/*
-		 * arriving here means that this inode can be flushed.  First
-		 * re-check that it's dirty before flushing.
-		 */
-		if (!xfs_inode_clean(cip)) {
-			int	error;
-			error = xfs_iflush_int(cip, bp);
-			if (error) {
-				xfs_iunlock(cip, XFS_ILOCK_SHARED);
-				goto cluster_corrupt_out;
-			}
-			clcount++;
-		} else {
-			xfs_ifunlock(cip);
-		}
-		xfs_iunlock(cip, XFS_ILOCK_SHARED);
-	}
-
-	if (clcount) {
-		XFS_STATS_INC(mp, xs_icluster_flushcnt);
-		XFS_STATS_ADD(mp, xs_icluster_flushinode, clcount);
-	}
-
-out_free:
-	rcu_read_unlock();
-	kmem_free(cilist);
-out_put:
-	xfs_perag_put(pag);
-	return 0;
-
-
-cluster_corrupt_out:
-	/*
-	 * Corruption detected in the clustering loop.  Invalidate the
-	 * inode buffer and shut down the filesystem.
-	 */
-	rcu_read_unlock();
-
-	/*
-	 * We'll always have an inode attached to the buffer for completion
-	 * process by the time we are called from xfs_iflush(). Hence we have
-	 * always need to do IO completion processing to abort the inodes
-	 * attached to the buffer.  handle them just like the shutdown case in
-	 * xfs_buf_submit().
-	 */
-	ASSERT(bp->b_iodone);
-	bp->b_flags |= XBF_ASYNC;
-	bp->b_flags &= ~XBF_DONE;
-	xfs_buf_stale(bp);
-	xfs_buf_ioerror(bp, -EIO);
-	xfs_buf_ioend(bp);
-
-	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
-
-	/* abort the corrupt inode, as it was not attached to the buffer */
-	xfs_iflush_abort(cip, false);
-	kmem_free(cilist);
-	xfs_perag_put(pag);
-	return -EFSCORRUPTED;
-}
-
-/*
- * Flush dirty inode metadata into the backing buffer.
- *
- * The caller must have the inode lock and the inode flush lock held.  The
- * inode lock will still be held upon return to the caller, and the inode
- * flush lock will be released after the inode has reached the disk.
- *
- * The caller must write out the buffer returned in *bpp and release it.
- */
-int
+static int
 xfs_iflush(
 	struct xfs_inode	*ip,
-	struct xfs_buf		**bpp)
-{
-	struct xfs_mount	*mp = ip->i_mount;
-	struct xfs_buf		*bp = NULL;
-	struct xfs_dinode	*dip;
-	int			error;
-
-	XFS_STATS_INC(mp, xs_iflush_count);
-
-	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
-	ASSERT(xfs_isiflocked(ip));
-	ASSERT(ip->i_d.di_format != XFS_DINODE_FMT_BTREE ||
-	       ip->i_d.di_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK));
-
-	*bpp = NULL;
-
-	xfs_iunpin_wait(ip);
-
-	/*
-	 * For stale inodes we cannot rely on the backing buffer remaining
-	 * stale in cache for the remaining life of the stale inode and so
-	 * xfs_imap_to_bp() below may give us a buffer that no longer contains
-	 * inodes below. We have to check this after ensuring the inode is
-	 * unpinned so that it is safe to reclaim the stale inode after the
-	 * flush call.
-	 */
-	if (xfs_iflags_test(ip, XFS_ISTALE)) {
-		xfs_ifunlock(ip);
-		return 0;
-	}
-
-	/*
-	 * This may have been unpinned because the filesystem is shutting
-	 * down forcibly. If that's the case we must not write this inode
-	 * to disk, because the log record didn't make it to disk.
-	 *
-	 * We also have to remove the log item from the AIL in this case,
-	 * as we wait for an empty AIL as part of the unmount process.
-	 */
-	if (XFS_FORCED_SHUTDOWN(mp)) {
-		error = -EIO;
-		goto abort_out;
-	}
-
-	/*
-	 * Get the buffer containing the on-disk inode. We are doing a try-lock
-	 * operation here, so we may get  an EAGAIN error. In that case, we
-	 * simply want to return with the inode still dirty.
-	 *
-	 * If we get any other error, we effectively have a corruption situation
-	 * and we cannot flush the inode, so we treat it the same as failing
-	 * xfs_iflush_int().
-	 */
-	error = xfs_imap_to_bp(mp, NULL, &ip->i_imap, &dip, &bp, XBF_TRYLOCK,
-			       0);
-	if (error == -EAGAIN) {
-		xfs_ifunlock(ip);
-		return error;
-	}
-	if (error)
-		goto corrupt_out;
-
-	/*
-	 * First flush out the inode that xfs_iflush was called with.
-	 */
-	error = xfs_iflush_int(ip, bp);
-	if (error)
-		goto corrupt_out;
-
-	/*
-	 * If the buffer is pinned then push on the log now so we won't
-	 * get stuck waiting in the write for too long.
-	 */
-	if (xfs_buf_ispinned(bp))
-		xfs_log_force(mp, 0);
-
-	/*
-	 * inode clustering: try to gather other inodes into this write
-	 *
-	 * Note: Any error during clustering will result in the filesystem
-	 * being shut down and completion callbacks run on the cluster buffer.
-	 * As we have already flushed and attached this inode to the buffer,
-	 * it has already been aborted and released by xfs_iflush_cluster() and
-	 * so we have no further error handling to do here.
-	 */
-	error = xfs_iflush_cluster(ip, bp);
-	if (error)
-		return error;
-
-	*bpp = bp;
-	return 0;
-
-corrupt_out:
-	if (bp)
-		xfs_buf_relse(bp);
-	xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
-abort_out:
-	/* abort the corrupt inode, as it was not attached to the buffer */
-	xfs_iflush_abort(ip, false);
-	return error;
-}
-
-/*
- * If there are inline format data / attr forks attached to this inode,
- * make sure they're not corrupt.
- */
-bool
-xfs_inode_verify_forks(
-	struct xfs_inode	*ip)
-{
-	struct xfs_ifork	*ifp;
-	xfs_failaddr_t		fa;
-
-	fa = xfs_ifork_verify_data(ip, &xfs_default_ifork_ops);
-	if (fa) {
-		ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
-		xfs_inode_verifier_error(ip, -EFSCORRUPTED, "data fork",
-				ifp->if_u1.if_data, ifp->if_bytes, fa);
-		return false;
-	}
-
-	fa = xfs_ifork_verify_attr(ip, &xfs_default_ifork_ops);
-	if (fa) {
-		ifp = XFS_IFORK_PTR(ip, XFS_ATTR_FORK);
-		xfs_inode_verifier_error(ip, -EFSCORRUPTED, "attr fork",
-				ifp ? ifp->if_u1.if_data : NULL,
-				ifp ? ifp->if_bytes : 0, fa);
-		return false;
-	}
-	return true;
-}
-
-STATIC int
-xfs_iflush_int(
-	struct xfs_inode	*ip,
 	struct xfs_buf		*bp)
 {
 	struct xfs_inode_log_item *iip = ip->i_itemp;
 	struct xfs_dinode	*dip;
 	struct xfs_mount	*mp = ip->i_mount;
+	int			error;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL|XFS_ILOCK_SHARED));
 	ASSERT(xfs_isiflocked(ip));
-	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);
+	ASSERT(ip->i_df.if_format != XFS_DINODE_FMT_BTREE ||
+	       ip->i_df.if_nextents > XFS_IFORK_MAXEXT(ip, XFS_DATA_FORK));
+	ASSERT(iip->ili_item.li_buf == bp);
 
-	/* set *dip = inode's place in the buffer */
 	dip = xfs_buf_offset(bp, ip->i_imap.im_boffset);
 
+	/*
+	 * We don't flush the inode if any of the following checks fail, but we
+	 * do still update the log item and attach to the backing buffer as if
+	 * the flush happened. This is a formality to facilitate predictable
+	 * error handling as the caller will shutdown and fail the buffer.
+	 */
+	error = -EFSCORRUPTED;
 	if (XFS_TEST_ERROR(dip->di_magic != cpu_to_be16(XFS_DINODE_MAGIC),
 			       mp, XFS_ERRTAG_IFLUSH_1)) {
 		xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
 			"%s: Bad inode %Lu magic number 0x%x, ptr "PTR_FMT,
 			__func__, ip->i_ino, be16_to_cpu(dip->di_magic), dip);
-		goto corrupt_out;
+		goto flush_out;
 	}
 	if (S_ISREG(VFS_I(ip)->i_mode)) {
 		if (XFS_TEST_ERROR(
-		    (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS) &&
-		    (ip->i_d.di_format != XFS_DINODE_FMT_BTREE),
+		    ip->i_df.if_format != XFS_DINODE_FMT_EXTENTS &&
+		    ip->i_df.if_format != XFS_DINODE_FMT_BTREE,
 		    mp, XFS_ERRTAG_IFLUSH_3)) {
 			xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
 				"%s: Bad regular inode %Lu, ptr "PTR_FMT,
 				__func__, ip->i_ino, ip);
-			goto corrupt_out;
+			goto flush_out;
 		}
 	} else if (S_ISDIR(VFS_I(ip)->i_mode)) {
 		if (XFS_TEST_ERROR(
-		    (ip->i_d.di_format != XFS_DINODE_FMT_EXTENTS) &&
-		    (ip->i_d.di_format != XFS_DINODE_FMT_BTREE) &&
-		    (ip->i_d.di_format != XFS_DINODE_FMT_LOCAL),
+		    ip->i_df.if_format != XFS_DINODE_FMT_EXTENTS &&
+		    ip->i_df.if_format != XFS_DINODE_FMT_BTREE &&
+		    ip->i_df.if_format != XFS_DINODE_FMT_LOCAL,
 		    mp, XFS_ERRTAG_IFLUSH_4)) {
 			xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
 				"%s: Bad directory inode %Lu, ptr "PTR_FMT,
 				__func__, ip->i_ino, ip);
-			goto corrupt_out;
+			goto flush_out;
 		}
 	}
-	if (XFS_TEST_ERROR(ip->i_d.di_nextents + ip->i_d.di_anextents >
+	if (XFS_TEST_ERROR(ip->i_df.if_nextents + xfs_ifork_nextents(ip->i_afp) >
 				ip->i_d.di_nblocks, mp, XFS_ERRTAG_IFLUSH_5)) {
 		xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
 			"%s: detected corrupt incore inode %Lu, "
 			"total extents = %d, nblocks = %Ld, ptr "PTR_FMT,
 			__func__, ip->i_ino,
-			ip->i_d.di_nextents + ip->i_d.di_anextents,
+			ip->i_df.if_nextents + xfs_ifork_nextents(ip->i_afp),
 			ip->i_d.di_nblocks, ip);
-		goto corrupt_out;
+		goto flush_out;
 	}
 	if (XFS_TEST_ERROR(ip->i_d.di_forkoff > mp->m_sb.sb_inodesize,
 				mp, XFS_ERRTAG_IFLUSH_6)) {
 		xfs_alert_tag(mp, XFS_PTAG_IFLUSH,
 			"%s: bad inode %Lu, forkoff 0x%x, ptr "PTR_FMT,
 			__func__, ip->i_ino, ip->i_d.di_forkoff, ip);
-		goto corrupt_out;
+		goto flush_out;
 	}
 
 	/*
@@ -3868,12 +3513,19 @@ xfs_iflush_int(
 	 * backwards compatibility with old kernels that predate logging all
 	 * inode changes.
 	 */
-	if (ip->i_d.di_version < 3)
+	if (!xfs_sb_version_has_v3inode(&mp->m_sb))
 		ip->i_d.di_flushiter++;
 
-	/* Check the inline fork data before we write out. */
-	if (!xfs_inode_verify_forks(ip))
-		goto corrupt_out;
+	/*
+	 * If there are inline format data / attr forks attached to this inode,
+	 * make sure they are not corrupt.
+	 */
+	if (ip->i_df.if_format == XFS_DINODE_FMT_LOCAL &&
+	    xfs_ifork_verify_local_data(ip))
+		goto flush_out;
+	if (ip->i_afp && ip->i_afp->if_format == XFS_DINODE_FMT_LOCAL &&
+	    xfs_ifork_verify_local_attr(ip))
+		goto flush_out;
 
 	/*
 	 * Copy the dirty parts of the inode into the on-disk inode.  We always
@@ -3889,7 +3541,6 @@ xfs_iflush_int(
 	xfs_iflush_fork(ip, dip, iip, XFS_DATA_FORK);
 	if (XFS_IFORK_Q(ip))
 		xfs_iflush_fork(ip, dip, iip, XFS_ATTR_FORK);
-	xfs_inobp_check(mp, bp);
 
 	/*
 	 * We've recorded everything logged in the inode, so we'd like to clear
@@ -3906,41 +3557,148 @@ xfs_iflush_int(
 	 * know that the information those bits represent is permanently on
 	 * disk.  As long as the flush completes before the inode is logged
 	 * again, then both ili_fields and ili_last_fields will be cleared.
-	 *
-	 * We can play with the ili_fields bits here, because the inode lock
-	 * must be held exclusively in order to set bits there and the flush
-	 * lock protects the ili_last_fields bits.  Set ili_logged so the flush
-	 * done routine can tell whether or not to look in the AIL.  Also, store
-	 * the current LSN of the inode so that we can tell whether the item has
-	 * moved in the AIL from xfs_iflush_done().  In order to read the lsn we
-	 * need the AIL lock, because it is a 64 bit value that cannot be read
-	 * atomically.
 	 */
+	error = 0;
+flush_out:
+	spin_lock(&iip->ili_lock);
 	iip->ili_last_fields = iip->ili_fields;
 	iip->ili_fields = 0;
 	iip->ili_fsync_fields = 0;
-	iip->ili_logged = 1;
+	spin_unlock(&iip->ili_lock);
 
+	/*
+	 * Store the current LSN of the inode so that we can tell whether the
+	 * item has moved in the AIL from xfs_iflush_done().
+	 */
 	xfs_trans_ail_copy_lsn(mp->m_ail, &iip->ili_flush_lsn,
 				&iip->ili_item.li_lsn);
 
+	/* generate the checksum. */
+	xfs_dinode_calc_crc(mp, dip);
+	return error;
+}
+
+/*
+ * Non-blocking flush of dirty inode metadata into the backing buffer.
+ *
+ * The caller must have a reference to the inode and hold the cluster buffer
+ * locked. The function will walk across all the inodes on the cluster buffer it
+ * can find and lock without blocking, and flush them to the cluster buffer.
+ *
+ * On successful flushing of at least one inode, the caller must write out the
+ * buffer and release it. If no inodes are flushed, -EAGAIN will be returned and
+ * the caller needs to release the buffer. On failure, the filesystem will be
+ * shut down, the buffer will have been unlocked and released, and EFSCORRUPTED
+ * will be returned.
+ */
+int
+xfs_iflush_cluster(
+	struct xfs_buf		*bp)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+	struct xfs_log_item	*lip, *n;
+	struct xfs_inode	*ip;
+	struct xfs_inode_log_item *iip;
+	int			clcount = 0;
+	int			error = 0;
+
 	/*
-	 * Attach the function xfs_iflush_done to the inode's
-	 * buffer.  This will remove the inode from the AIL
-	 * and unlock the inode's flush lock when the inode is
-	 * completely written to disk.
+	 * We must use the safe variant here as on shutdown xfs_iflush_abort()
+	 * can remove itself from the list.
 	 */
-	xfs_buf_attach_iodone(bp, xfs_iflush_done, &iip->ili_item);
+	list_for_each_entry_safe(lip, n, &bp->b_li_list, li_bio_list) {
+		iip = (struct xfs_inode_log_item *)lip;
+		ip = iip->ili_inode;
 
-	/* generate the checksum. */
-	xfs_dinode_calc_crc(mp, dip);
+		/*
+		 * Quick and dirty check to avoid locks if possible.
+		 */
+		if (__xfs_iflags_test(ip, XFS_IRECLAIM | XFS_IFLOCK))
+			continue;
+		if (xfs_ipincount(ip))
+			continue;
+
+		/*
+		 * The inode is still attached to the buffer, which means it is
+		 * dirty but reclaim might try to grab it. Check carefully for
+		 * that, and grab the ilock while still holding the i_flags_lock
+		 * to guarantee reclaim will not be able to reclaim this inode
+		 * once we drop the i_flags_lock.
+		 */
+		spin_lock(&ip->i_flags_lock);
+		ASSERT(!__xfs_iflags_test(ip, XFS_ISTALE));
+		if (__xfs_iflags_test(ip, XFS_IRECLAIM | XFS_IFLOCK)) {
+			spin_unlock(&ip->i_flags_lock);
+			continue;
+		}
+
+		/*
+		 * ILOCK will pin the inode against reclaim and prevent
+		 * concurrent transactions modifying the inode while we are
+		 * flushing the inode.
+		 */
+		if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED)) {
+			spin_unlock(&ip->i_flags_lock);
+			continue;
+		}
+		spin_unlock(&ip->i_flags_lock);
 
-	ASSERT(!list_empty(&bp->b_li_list));
-	ASSERT(bp->b_iodone != NULL);
+		/*
+		 * Skip inodes that are already flush locked as they have
+		 * already been written to the buffer.
+		 */
+		if (!xfs_iflock_nowait(ip)) {
+			xfs_iunlock(ip, XFS_ILOCK_SHARED);
+			continue;
+		}
+
+		/*
+		 * Abort flushing this inode if we are shut down because the
+		 * inode may not currently be in the AIL. This can occur when
+		 * log I/O failure unpins the inode without inserting into the
+		 * AIL, leaving a dirty/unpinned inode attached to the buffer
+		 * that otherwise looks like it should be flushed.
+		 */
+		if (XFS_FORCED_SHUTDOWN(mp)) {
+			xfs_iunpin_wait(ip);
+			/* xfs_iflush_abort() drops the flush lock */
+			xfs_iflush_abort(ip);
+			xfs_iunlock(ip, XFS_ILOCK_SHARED);
+			error = -EIO;
+			continue;
+		}
+
+		/* don't block waiting on a log force to unpin dirty inodes */
+		if (xfs_ipincount(ip)) {
+			xfs_ifunlock(ip);
+			xfs_iunlock(ip, XFS_ILOCK_SHARED);
+			continue;
+		}
+
+		if (!xfs_inode_clean(ip))
+			error = xfs_iflush(ip, bp);
+		else
+			xfs_ifunlock(ip);
+		xfs_iunlock(ip, XFS_ILOCK_SHARED);
+		if (error)
+			break;
+		clcount++;
+	}
+
+	if (error) {
+		bp->b_flags |= XBF_ASYNC;
+		xfs_buf_ioend_fail(bp);
+		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+		return error;
+	}
+
+	if (!clcount)
+		return -EAGAIN;
+
+	XFS_STATS_INC(mp, xs_icluster_flushcnt);
+	XFS_STATS_ADD(mp, xs_icluster_flushinode, clcount);
 	return 0;
 
-corrupt_out:
-	return -EFSCORRUPTED;
 }
 
 /* Release an inode. */
@@ -3951,3 +3709,115 @@ xfs_irele(
 	trace_xfs_irele(ip, _RET_IP_);
 	iput(VFS_I(ip));
 }
+
+/*
+ * Ensure all commited transactions touching the inode are written to the log.
+ */
+int
+xfs_log_force_inode(
+	struct xfs_inode	*ip)
+{
+	xfs_lsn_t		lsn = 0;
+
+	xfs_ilock(ip, XFS_ILOCK_SHARED);
+	if (xfs_ipincount(ip))
+		lsn = ip->i_itemp->ili_last_lsn;
+	xfs_iunlock(ip, XFS_ILOCK_SHARED);
+
+	if (!lsn)
+		return 0;
+	return xfs_log_force_lsn(ip->i_mount, lsn, XFS_LOG_SYNC, NULL);
+}
+
+/*
+ * Grab the exclusive iolock for a data copy from src to dest, making sure to
+ * abide vfs locking order (lowest pointer value goes first) and breaking the
+ * layout leases before proceeding.  The loop is needed because we cannot call
+ * the blocking break_layout() with the iolocks held, and therefore have to
+ * back out both locks.
+ */
+static int
+xfs_iolock_two_inodes_and_break_layout(
+	struct inode		*src,
+	struct inode		*dest)
+{
+	int			error;
+
+	if (src > dest)
+		swap(src, dest);
+
+retry:
+	/* Wait to break both inodes' layouts before we start locking. */
+	error = break_layout(src, true);
+	if (error)
+		return error;
+	if (src != dest) {
+		error = break_layout(dest, true);
+		if (error)
+			return error;
+	}
+
+	/* Lock one inode and make sure nobody got in and leased it. */
+	inode_lock(src);
+	error = break_layout(src, false);
+	if (error) {
+		inode_unlock(src);
+		if (error == -EWOULDBLOCK)
+			goto retry;
+		return error;
+	}
+
+	if (src == dest)
+		return 0;
+
+	/* Lock the other inode and make sure nobody got in and leased it. */
+	inode_lock_nested(dest, I_MUTEX_NONDIR2);
+	error = break_layout(dest, false);
+	if (error) {
+		inode_unlock(src);
+		inode_unlock(dest);
+		if (error == -EWOULDBLOCK)
+			goto retry;
+		return error;
+	}
+
+	return 0;
+}
+
+/*
+ * Lock two inodes so that userspace cannot initiate I/O via file syscalls or
+ * mmap activity.
+ */
+int
+xfs_ilock2_io_mmap(
+	struct xfs_inode	*ip1,
+	struct xfs_inode	*ip2)
+{
+	int			ret;
+
+	ret = xfs_iolock_two_inodes_and_break_layout(VFS_I(ip1), VFS_I(ip2));
+	if (ret)
+		return ret;
+	if (ip1 == ip2)
+		xfs_ilock(ip1, XFS_MMAPLOCK_EXCL);
+	else
+		xfs_lock_two_inodes(ip1, XFS_MMAPLOCK_EXCL,
+				    ip2, XFS_MMAPLOCK_EXCL);
+	return 0;
+}
+
+/* Unlock both inodes to allow IO and mmap activity. */
+void
+xfs_iunlock2_io_mmap(
+	struct xfs_inode	*ip1,
+	struct xfs_inode	*ip2)
+{
+	bool			same_inode = (ip1 == ip2);
+
+	xfs_iunlock(ip2, XFS_MMAPLOCK_EXCL);
+	if (!same_inode)
+		xfs_iunlock(ip1, XFS_MMAPLOCK_EXCL);
+	inode_unlock(VFS_I(ip2));
+	if (!same_inode)
+		inode_unlock(VFS_I(ip1));
+}
diff --git a/fs/xfs/xfs_inode.h b/fs/xfs/xfs_inode.h
index 492e53992fa9..e9a8bb184d1f 100644
--- a/fs/xfs/xfs_inode.h
+++ b/fs/xfs/xfs_inode.h
@@ -57,9 +57,6 @@ typedef struct xfs_inode {
 
 	struct xfs_icdinode	i_d;		/* most of ondisk inode */
 
-	xfs_extnum_t		i_cnextents;	/* # of extents in cow fork */
-	unsigned int		i_cformat;	/* format of cow fork */
-
 	/* VFS inode */
 	struct inode		i_vnode;	/* embedded VFS inode */
 
@@ -218,8 +215,7 @@ static inline bool xfs_inode_has_cow_data(struct xfs_inode *ip)
 #define XFS_IFLOCK		(1 << __XFS_IFLOCK_BIT)
 #define __XFS_IPINNED_BIT	8	 /* wakeup key for zero pin count */
 #define XFS_IPINNED		(1 << __XFS_IPINNED_BIT)
-#define XFS_IDONTCACHE		(1 << 9) /* don't cache the inode long term */
-#define XFS_IEOFBLOCKS		(1 << 10)/* has the preallocblocks tag set */
+#define XFS_IEOFBLOCKS		(1 << 9) /* has the preallocblocks tag set */
 /*
  * If this unlinked inode is in the middle of recovery, don't let drop_inode
  * truncate and free the inode.  This can happen if we iget the inode during
@@ -426,10 +422,11 @@ int		xfs_itruncate_extents_flags(struct xfs_trans **,
 				struct xfs_inode *, int, xfs_fsize_t, int);
 void		xfs_iext_realloc(xfs_inode_t *, int, int);
 
+int		xfs_log_force_inode(struct xfs_inode *ip);
 void		xfs_iunpin_wait(xfs_inode_t *);
 #define xfs_ipincount(ip)	((unsigned int) atomic_read(&ip->i_pincount))
 
-int		xfs_iflush(struct xfs_inode *, struct xfs_buf **);
+int		xfs_iflush_cluster(struct xfs_buf *);
 void		xfs_lock_two_inodes(struct xfs_inode *ip0, uint ip0_mode,
 				struct xfs_inode *ip1, uint ip1_mode);
 
@@ -466,6 +463,7 @@ int	xfs_break_layouts(struct inode *inode, uint *iolock,
 /* from xfs_iops.c */
 extern void xfs_setup_inode(struct xfs_inode *ip);
 extern void xfs_setup_iops(struct xfs_inode *ip);
+extern void xfs_diflags_to_iflags(struct xfs_inode *ip, bool init);
 
 /*
  * When setting up a newly allocated inode, we need to call
@@ -496,11 +494,12 @@ extern struct kmem_zone	*xfs_inode_zone;
 /* The default CoW extent size hint. */
 #define XFS_DEFAULT_COWEXTSZ_HINT 32
 
-bool xfs_inode_verify_forks(struct xfs_inode *ip);
-
 int xfs_iunlink_init(struct xfs_perag *pag);
 void xfs_iunlink_destroy(struct xfs_perag *pag);
 
 void xfs_end_io(struct work_struct *work);
 
+int xfs_ilock2_io_mmap(struct xfs_inode *ip1, struct xfs_inode *ip2);
+void xfs_iunlock2_io_mmap(struct xfs_inode *ip1, struct xfs_inode *ip2);
+
 #endif	/* __XFS_INODE_H__ */
diff --git a/fs/xfs/xfs_inode_item.c b/fs/xfs/xfs_inode_item.c
index 8bd5d0de6321..6c65938cee1c 100644
--- a/fs/xfs/xfs_inode_item.c
+++ b/fs/xfs/xfs_inode_item.c
@@ -36,10 +36,10 @@ xfs_inode_item_data_fork_size(
 {
 	struct xfs_inode	*ip = iip->ili_inode;
 
-	switch (ip->i_d.di_format) {
+	switch (ip->i_df.if_format) {
 	case XFS_DINODE_FMT_EXTENTS:
 		if ((iip->ili_fields & XFS_ILOG_DEXT) &&
-		    ip->i_d.di_nextents > 0 &&
+		    ip->i_df.if_nextents > 0 &&
 		    ip->i_df.if_bytes > 0) {
 			/* worst case, doesn't subtract delalloc extents */
 			*nbytes += XFS_IFORK_DSIZE(ip);
@@ -77,10 +77,10 @@ xfs_inode_item_attr_fork_size(
 {
 	struct xfs_inode	*ip = iip->ili_inode;
 
-	switch (ip->i_d.di_aformat) {
+	switch (ip->i_afp->if_format) {
 	case XFS_DINODE_FMT_EXTENTS:
 		if ((iip->ili_fields & XFS_ILOG_AEXT) &&
-		    ip->i_d.di_anextents > 0 &&
+		    ip->i_afp->if_nextents > 0 &&
 		    ip->i_afp->if_bytes > 0) {
 			/* worst case, doesn't subtract unused space */
 			*nbytes += XFS_IFORK_ASIZE(ip);
@@ -125,7 +125,7 @@ xfs_inode_item_size(
 
 	*nvecs += 2;
 	*nbytes += sizeof(struct xfs_inode_log_format) +
-		   xfs_log_dinode_size(ip->i_d.di_version);
+		   xfs_log_dinode_size(ip->i_mount);
 
 	xfs_inode_item_data_fork_size(iip, nvecs, nbytes);
 	if (XFS_IFORK_Q(ip))
@@ -142,13 +142,13 @@ xfs_inode_item_format_data_fork(
 	struct xfs_inode	*ip = iip->ili_inode;
 	size_t			data_bytes;
 
-	switch (ip->i_d.di_format) {
+	switch (ip->i_df.if_format) {
 	case XFS_DINODE_FMT_EXTENTS:
 		iip->ili_fields &=
 			~(XFS_ILOG_DDATA | XFS_ILOG_DBROOT | XFS_ILOG_DEV);
 
 		if ((iip->ili_fields & XFS_ILOG_DEXT) &&
-		    ip->i_d.di_nextents > 0 &&
+		    ip->i_df.if_nextents > 0 &&
 		    ip->i_df.if_bytes > 0) {
 			struct xfs_bmbt_rec *p;
 
@@ -191,7 +191,7 @@ xfs_inode_item_format_data_fork(
 		    ip->i_df.if_bytes > 0) {
 			/*
 			 * Round i_bytes up to a word boundary.
-			 * The underlying memory is guaranteed to
+			 * The underlying memory is guaranteed
 			 * to be there by xfs_idata_realloc().
 			 */
 			data_bytes = roundup(ip->i_df.if_bytes, 4);
@@ -227,18 +227,18 @@ xfs_inode_item_format_attr_fork(
 	struct xfs_inode	*ip = iip->ili_inode;
 	size_t			data_bytes;
 
-	switch (ip->i_d.di_aformat) {
+	switch (ip->i_afp->if_format) {
 	case XFS_DINODE_FMT_EXTENTS:
 		iip->ili_fields &=
 			~(XFS_ILOG_ADATA | XFS_ILOG_ABROOT);
 
 		if ((iip->ili_fields & XFS_ILOG_AEXT) &&
-		    ip->i_d.di_anextents > 0 &&
+		    ip->i_afp->if_nextents > 0 &&
 		    ip->i_afp->if_bytes > 0) {
 			struct xfs_bmbt_rec *p;
 
 			ASSERT(xfs_iext_count(ip->i_afp) ==
-				ip->i_d.di_anextents);
+				ip->i_afp->if_nextents);
 
 			p = xlog_prepare_iovec(lv, vecp, XLOG_REG_TYPE_IATTR_EXT);
 			data_bytes = xfs_iextents_copy(ip, p, XFS_ATTR_FORK);
@@ -275,7 +275,7 @@ xfs_inode_item_format_attr_fork(
 		    ip->i_afp->if_bytes > 0) {
 			/*
 			 * Round i_bytes up to a word boundary.
-			 * The underlying memory is guaranteed to
+			 * The underlying memory is guaranteed
 			 * to be there by xfs_idata_realloc().
 			 */
 			data_bytes = roundup(ip->i_afp->if_bytes, 4);
@@ -305,11 +305,9 @@ xfs_inode_to_log_dinode(
 	struct inode		*inode = VFS_I(ip);
 
 	to->di_magic = XFS_DINODE_MAGIC;
-
-	to->di_version = from->di_version;
-	to->di_format = from->di_format;
-	to->di_uid = from->di_uid;
-	to->di_gid = from->di_gid;
+	to->di_format = xfs_ifork_format(&ip->i_df);
+	to->di_uid = i_uid_read(inode);
+	to->di_gid = i_gid_read(inode);
 	to->di_projid_lo = from->di_projid & 0xffff;
 	to->di_projid_hi = from->di_projid >> 16;
 
@@ -328,10 +326,10 @@ xfs_inode_to_log_dinode(
 	to->di_size = from->di_size;
 	to->di_nblocks = from->di_nblocks;
 	to->di_extsize = from->di_extsize;
-	to->di_nextents = from->di_nextents;
-	to->di_anextents = from->di_anextents;
+	to->di_nextents = xfs_ifork_nextents(&ip->i_df);
+	to->di_anextents = xfs_ifork_nextents(ip->i_afp);
 	to->di_forkoff = from->di_forkoff;
-	to->di_aformat = from->di_aformat;
+	to->di_aformat = xfs_ifork_format(ip->i_afp);
 	to->di_dmevmask = from->di_dmevmask;
 	to->di_dmstate = from->di_dmstate;
 	to->di_flags = from->di_flags;
@@ -339,7 +337,8 @@ xfs_inode_to_log_dinode(
 	/* log a dummy value to ensure log structure is fully initialised */
 	to->di_next_unlinked = NULLAGINO;
 
-	if (from->di_version == 3) {
+	if (xfs_sb_version_has_v3inode(&ip->i_mount->m_sb)) {
+		to->di_version = 3;
 		to->di_changecount = inode_peek_iversion(inode);
 		to->di_crtime.t_sec = from->di_crtime.tv_sec;
 		to->di_crtime.t_nsec = from->di_crtime.tv_nsec;
@@ -351,6 +350,7 @@ xfs_inode_to_log_dinode(
 		uuid_copy(&to->di_uuid, &ip->i_mount->m_sb.sb_meta_uuid);
 		to->di_flushiter = 0;
 	} else {
+		to->di_version = 2;
 		to->di_flushiter = from->di_flushiter;
 	}
 }
@@ -370,7 +370,7 @@ xfs_inode_item_format_core(
 
 	dic = xlog_prepare_iovec(lv, vecp, XLOG_REG_TYPE_ICORE);
 	xfs_inode_to_log_dinode(ip, dic, ip->i_itemp->ili_item.li_lsn);
-	xlog_finish_iovec(lv, *vecp, xfs_log_dinode_size(ip->i_d.di_version));
+	xlog_finish_iovec(lv, *vecp, xfs_log_dinode_size(ip->i_mount));
 }
 
 /*
@@ -395,8 +395,6 @@ xfs_inode_item_format(
 	struct xfs_log_iovec	*vecp = NULL;
 	struct xfs_inode_log_format *ilf;
 
-	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;
@@ -441,6 +439,7 @@ xfs_inode_item_pin(
 	struct xfs_inode	*ip = INODE_ITEM(lip)->ili_inode;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
+	ASSERT(lip->li_buf);
 
 	trace_xfs_inode_pin(ip, _RET_IP_);
 	atomic_inc(&ip->i_pincount);
@@ -452,6 +451,12 @@ xfs_inode_item_pin(
  * item which was previously pinned with a call to xfs_inode_item_pin().
  *
  * Also wake up anyone in xfs_iunpin_wait() if the count goes to 0.
+ *
+ * Note that unpin can race with inode cluster buffer freeing marking the buffer
+ * stale. In that case, flush completions are run from the buffer unpin call,
+ * which may happen before the inode is unpinned. If we lose the race, there
+ * will be no buffer attached to the log item, but the inode will be marked
+ * XFS_ISTALE.
  */
 STATIC void
 xfs_inode_item_unpin(
@@ -461,28 +466,12 @@ xfs_inode_item_unpin(
 	struct xfs_inode	*ip = INODE_ITEM(lip)->ili_inode;
 
 	trace_xfs_inode_unpin(ip, _RET_IP_);
+	ASSERT(lip->li_buf || xfs_iflags_test(ip, XFS_ISTALE));
 	ASSERT(atomic_read(&ip->i_pincount) > 0);
 	if (atomic_dec_and_test(&ip->i_pincount))
 		wake_up_bit(&ip->i_flags, __XFS_IPINNED_BIT);
 }
 
-/*
- * Callback used to mark a buffer with XFS_LI_FAILED when items in the buffer
- * have been failed during writeback
- *
- * This informs the AIL that the inode is already flush locked on the next push,
- * and acquires a hold on the buffer to ensure that it isn't reclaimed before
- * dirty data makes it to disk.
- */
-STATIC void
-xfs_inode_item_error(
-	struct xfs_log_item	*lip,
-	struct xfs_buf		*bp)
-{
-	ASSERT(xfs_isiflocked(INODE_ITEM(lip)->ili_inode));
-	xfs_set_li_failed(lip, bp);
-}
-
 STATIC uint
 xfs_inode_item_push(
 	struct xfs_log_item	*lip,
@@ -496,69 +485,44 @@ xfs_inode_item_push(
 	uint			rval = XFS_ITEM_SUCCESS;
 	int			error;
 
-	if (xfs_ipincount(ip) > 0)
-		return XFS_ITEM_PINNED;
-
-	/*
-	 * The buffer containing this item failed to be written back
-	 * previously. Resubmit the buffer for IO.
-	 */
-	if (test_bit(XFS_LI_FAILED, &lip->li_flags)) {
-		if (!xfs_buf_trylock(bp))
-			return XFS_ITEM_LOCKED;
+	ASSERT(iip->ili_item.li_buf);
 
-		if (!xfs_buf_resubmit_failed_buffers(bp, buffer_list))
-			rval = XFS_ITEM_FLUSHING;
+	if (xfs_ipincount(ip) > 0 || xfs_buf_ispinned(bp) ||
+	    (ip->i_flags & XFS_ISTALE))
+		return XFS_ITEM_PINNED;
 
-		xfs_buf_unlock(bp);
-		return rval;
-	}
+	/* If the inode is already flush locked, we're already flushing. */
+	if (xfs_isiflocked(ip))
+		return XFS_ITEM_FLUSHING;
 
-	if (!xfs_ilock_nowait(ip, XFS_ILOCK_SHARED))
+	if (!xfs_buf_trylock(bp))
 		return XFS_ITEM_LOCKED;
 
-	/*
-	 * Re-check the pincount now that we stabilized the value by
-	 * taking the ilock.
-	 */
-	if (xfs_ipincount(ip) > 0) {
-		rval = XFS_ITEM_PINNED;
-		goto out_unlock;
-	}
-
-	/*
-	 * Stale inode items should force out the iclog.
-	 */
-	if (ip->i_flags & XFS_ISTALE) {
-		rval = XFS_ITEM_PINNED;
-		goto out_unlock;
-	}
+	spin_unlock(&lip->li_ailp->ail_lock);
 
 	/*
-	 * Someone else is already flushing the inode.  Nothing we can do
-	 * here but wait for the flush to finish and remove the item from
-	 * the AIL.
+	 * We need to hold a reference for flushing the cluster buffer as it may
+	 * fail the buffer without IO submission. In which case, we better get a
+	 * reference for that completion because otherwise we don't get a
+	 * reference for IO until we queue the buffer for delwri submission.
 	 */
-	if (!xfs_iflock_nowait(ip)) {
-		rval = XFS_ITEM_FLUSHING;
-		goto out_unlock;
-	}
-
-	ASSERT(iip->ili_fields != 0 || XFS_FORCED_SHUTDOWN(ip->i_mount));
-	ASSERT(iip->ili_logged == 0 || XFS_FORCED_SHUTDOWN(ip->i_mount));
-
-	spin_unlock(&lip->li_ailp->ail_lock);
-
-	error = xfs_iflush(ip, &bp);
+	xfs_buf_hold(bp);
+	error = xfs_iflush_cluster(bp);
 	if (!error) {
 		if (!xfs_buf_delwri_queue(bp, buffer_list))
 			rval = XFS_ITEM_FLUSHING;
 		xfs_buf_relse(bp);
+	} else {
+		/*
+		 * Release the buffer if we were unable to flush anything. On
+		 * any other error, the buffer has already been released.
+		 */
+		if (error == -EAGAIN)
+			xfs_buf_relse(bp);
+		rval = XFS_ITEM_LOCKED;
 	}
 
 	spin_lock(&lip->li_ailp->ail_lock);
-out_unlock:
-	xfs_iunlock(ip, XFS_ILOCK_SHARED);
 	return rval;
 }
 
@@ -637,7 +601,6 @@ static const struct xfs_item_ops xfs_inode_item_ops = {
 	.iop_committed	= xfs_inode_item_committed,
 	.iop_push	= xfs_inode_item_push,
 	.iop_committing	= xfs_inode_item_committing,
-	.iop_error	= xfs_inode_item_error
 };
 
 
@@ -652,9 +615,11 @@ xfs_inode_item_init(
 	struct xfs_inode_log_item *iip;
 
 	ASSERT(ip->i_itemp == NULL);
-	iip = ip->i_itemp = kmem_zone_zalloc(xfs_ili_zone, 0);
+	iip = ip->i_itemp = kmem_cache_zalloc(xfs_ili_zone,
+					      GFP_KERNEL | __GFP_NOFAIL);
 
 	iip->ili_inode = ip;
+	spin_lock_init(&iip->ili_lock);
 	xfs_log_item_init(mp, &iip->ili_item, XFS_LI_INODE,
 						&xfs_inode_item_ops);
 }
@@ -664,112 +629,129 @@ xfs_inode_item_init(
  */
 void
 xfs_inode_item_destroy(
-	xfs_inode_t	*ip)
+	struct xfs_inode	*ip)
 {
-	kmem_free(ip->i_itemp->ili_item.li_lv_shadow);
-	kmem_cache_free(xfs_ili_zone, ip->i_itemp);
+	struct xfs_inode_log_item *iip = ip->i_itemp;
+
+	ASSERT(iip->ili_item.li_buf == NULL);
+
+	ip->i_itemp = NULL;
+	kmem_free(iip->ili_item.li_lv_shadow);
+	kmem_cache_free(xfs_ili_zone, iip);
 }
 
 
 /*
- * This is the inode flushing I/O completion routine.  It is called
- * from interrupt level when the buffer containing the inode is
- * flushed to disk.  It is responsible for removing the inode item
- * from the AIL if it has not been re-logged, and unlocking the inode's
- * flush lock.
- *
- * To reduce AIL lock traffic as much as possible, we scan the buffer log item
- * list for other inodes that will run this function. We remove them from the
- * buffer list so we can process all the inode IO completions in one AIL lock
- * traversal.
+ * We only want to pull the item from the AIL if it is actually there
+ * and its location in the log has not changed since we started the
+ * flush.  Thus, we only bother if the inode's lsn has not changed.
  */
-void
-xfs_iflush_done(
-	struct xfs_buf		*bp,
-	struct xfs_log_item	*lip)
+static void
+xfs_iflush_ail_updates(
+	struct xfs_ail		*ailp,
+	struct list_head	*list)
 {
-	struct xfs_inode_log_item *iip;
-	struct xfs_log_item	*blip, *n;
-	struct xfs_ail		*ailp = lip->li_ailp;
-	int			need_ail = 0;
-	LIST_HEAD(tmp);
+	struct xfs_log_item	*lip;
+	xfs_lsn_t		tail_lsn = 0;
 
-	/*
-	 * Scan the buffer IO completions for other inodes being completed and
-	 * attach them to the current inode log item.
-	 */
-
-	list_add_tail(&lip->li_bio_list, &tmp);
+	/* this is an opencoded batch version of xfs_trans_ail_delete */
+	spin_lock(&ailp->ail_lock);
+	list_for_each_entry(lip, list, li_bio_list) {
+		xfs_lsn_t	lsn;
 
-	list_for_each_entry_safe(blip, n, &bp->b_li_list, li_bio_list) {
-		if (lip->li_cb != xfs_iflush_done)
+		clear_bit(XFS_LI_FAILED, &lip->li_flags);
+		if (INODE_ITEM(lip)->ili_flush_lsn != lip->li_lsn)
 			continue;
 
-		list_move_tail(&blip->li_bio_list, &tmp);
+		lsn = xfs_ail_delete_one(ailp, lip);
+		if (!tail_lsn && lsn)
+			tail_lsn = lsn;
+	}
+	xfs_ail_update_finish(ailp, tail_lsn);
+}
+
+/*
+ * Walk the list of inodes that have completed their IOs. If they are clean
+ * remove them from the list and dissociate them from the buffer. Buffers that
+ * are still dirty remain linked to the buffer and on the list. Caller must
+ * handle them appropriately.
+ */
+static void
+xfs_iflush_finish(
+	struct xfs_buf		*bp,
+	struct list_head	*list)
+{
+	struct xfs_log_item	*lip, *n;
+
+	list_for_each_entry_safe(lip, n, list, li_bio_list) {
+		struct xfs_inode_log_item *iip = INODE_ITEM(lip);
+		bool	drop_buffer = false;
+
+		spin_lock(&iip->ili_lock);
+
 		/*
-		 * while we have the item, do the unlocked check for needing
-		 * the AIL lock.
+		 * Remove the reference to the cluster buffer if the inode is
+		 * clean in memory and drop the buffer reference once we've
+		 * dropped the locks we hold.
 		 */
-		iip = INODE_ITEM(blip);
-		if ((iip->ili_logged && blip->li_lsn == iip->ili_flush_lsn) ||
-		    test_bit(XFS_LI_FAILED, &blip->li_flags))
-			need_ail++;
+		ASSERT(iip->ili_item.li_buf == bp);
+		if (!iip->ili_fields) {
+			iip->ili_item.li_buf = NULL;
+			list_del_init(&lip->li_bio_list);
+			drop_buffer = true;
+		}
+		iip->ili_last_fields = 0;
+		iip->ili_flush_lsn = 0;
+		spin_unlock(&iip->ili_lock);
+		xfs_ifunlock(iip->ili_inode);
+		if (drop_buffer)
+			xfs_buf_rele(bp);
 	}
+}
 
-	/* make sure we capture the state of the initial inode. */
-	iip = INODE_ITEM(lip);
-	if ((iip->ili_logged && lip->li_lsn == iip->ili_flush_lsn) ||
-	    test_bit(XFS_LI_FAILED, &lip->li_flags))
-		need_ail++;
+/*
+ * Inode buffer IO completion routine.  It is responsible for removing inodes
+ * attached to the buffer from the AIL if they have not been re-logged, as well
+ * as completing the flush and unlocking the inode.
+ */
+void
+xfs_iflush_done(
+	struct xfs_buf		*bp)
+{
+	struct xfs_log_item	*lip, *n;
+	LIST_HEAD(flushed_inodes);
+	LIST_HEAD(ail_updates);
 
 	/*
-	 * We only want to pull the item from the AIL if it is
-	 * actually there and its location in the log has not
-	 * changed since we started the flush.  Thus, we only bother
-	 * if the ili_logged flag is set and the inode's lsn has not
-	 * changed.  First we check the lsn outside
-	 * the lock since it's cheaper, and then we recheck while
-	 * holding the lock before removing the inode from the AIL.
+	 * Pull the attached inodes from the buffer one at a time and take the
+	 * appropriate action on them.
 	 */
-	if (need_ail) {
-		bool			mlip_changed = false;
-
-		/* this is an opencoded batch version of xfs_trans_ail_delete */
-		spin_lock(&ailp->ail_lock);
-		list_for_each_entry(blip, &tmp, li_bio_list) {
-			if (INODE_ITEM(blip)->ili_logged &&
-			    blip->li_lsn == INODE_ITEM(blip)->ili_flush_lsn)
-				mlip_changed |= xfs_ail_delete_one(ailp, blip);
-			else {
-				xfs_clear_li_failed(blip);
-			}
-		}
+	list_for_each_entry_safe(lip, n, &bp->b_li_list, li_bio_list) {
+		struct xfs_inode_log_item *iip = INODE_ITEM(lip);
 
-		if (mlip_changed) {
-			if (!XFS_FORCED_SHUTDOWN(ailp->ail_mount))
-				xlog_assign_tail_lsn_locked(ailp->ail_mount);
-			if (list_empty(&ailp->ail_head))
-				wake_up_all(&ailp->ail_empty);
+		if (xfs_iflags_test(iip->ili_inode, XFS_ISTALE)) {
+			xfs_iflush_abort(iip->ili_inode);
+			continue;
 		}
-		spin_unlock(&ailp->ail_lock);
+		if (!iip->ili_last_fields)
+			continue;
 
-		if (mlip_changed)
-			xfs_log_space_wake(ailp->ail_mount);
+		/* Do an unlocked check for needing the AIL lock. */
+		if (iip->ili_flush_lsn == lip->li_lsn ||
+		    test_bit(XFS_LI_FAILED, &lip->li_flags))
+			list_move_tail(&lip->li_bio_list, &ail_updates);
+		else
+			list_move_tail(&lip->li_bio_list, &flushed_inodes);
 	}
 
-	/*
-	 * clean up and unlock the flush lock now we are done. We can clear the
-	 * ili_last_fields bits now that we know that the data corresponding to
-	 * them is safely on disk.
-	 */
-	list_for_each_entry_safe(blip, n, &tmp, li_bio_list) {
-		list_del_init(&blip->li_bio_list);
-		iip = INODE_ITEM(blip);
-		iip->ili_logged = 0;
-		iip->ili_last_fields = 0;
-		xfs_ifunlock(iip->ili_inode);
+	if (!list_empty(&ail_updates)) {
+		xfs_iflush_ail_updates(bp->b_mount->m_ail, &ail_updates);
+		list_splice_tail(&ail_updates, &flushed_inodes);
 	}
-	list_del(&tmp);
+
+	xfs_iflush_finish(bp, &flushed_inodes);
+	if (!list_empty(&flushed_inodes))
+		list_splice_tail(&flushed_inodes, &bp->b_li_list);
 }
 
 /*
@@ -780,42 +762,37 @@ xfs_iflush_done(
  */
 void
 xfs_iflush_abort(
-	xfs_inode_t		*ip,
-	bool			stale)
+	struct xfs_inode	*ip)
 {
-	xfs_inode_log_item_t	*iip = ip->i_itemp;
+	struct xfs_inode_log_item *iip = ip->i_itemp;
+	struct xfs_buf		*bp = NULL;
 
 	if (iip) {
-		if (test_bit(XFS_LI_IN_AIL, &iip->ili_item.li_flags)) {
-			xfs_trans_ail_remove(&iip->ili_item,
-					     stale ? SHUTDOWN_LOG_IO_ERROR :
-						     SHUTDOWN_CORRUPT_INCORE);
-		}
-		iip->ili_logged = 0;
 		/*
-		 * Clear the ili_last_fields bits now that we know that the
-		 * data corresponding to them is safely on disk.
+		 * Clear the failed bit before removing the item from the AIL so
+		 * xfs_trans_ail_delete() doesn't try to clear and release the
+		 * buffer attached to the log item before we are done with it.
 		 */
-		iip->ili_last_fields = 0;
+		clear_bit(XFS_LI_FAILED, &iip->ili_item.li_flags);
+		xfs_trans_ail_delete(&iip->ili_item, 0);
+
 		/*
 		 * Clear the inode logging fields so no more flushes are
 		 * attempted.
 		 */
+		spin_lock(&iip->ili_lock);
+		iip->ili_last_fields = 0;
 		iip->ili_fields = 0;
 		iip->ili_fsync_fields = 0;
+		iip->ili_flush_lsn = 0;
+		bp = iip->ili_item.li_buf;
+		iip->ili_item.li_buf = NULL;
+		list_del_init(&iip->ili_item.li_bio_list);
+		spin_unlock(&iip->ili_lock);
 	}
-	/*
-	 * Release the inode's flush lock since we're done with it.
-	 */
 	xfs_ifunlock(ip);
-}
-
-void
-xfs_istale_done(
-	struct xfs_buf		*bp,
-	struct xfs_log_item	*lip)
-{
-	xfs_iflush_abort(INODE_ITEM(lip)->ili_inode, true);
+	if (bp)
+		xfs_buf_rele(bp);
 }
 
 /*
diff --git a/fs/xfs/xfs_inode_item.h b/fs/xfs/xfs_inode_item.h
index 07a60e74c39c..048b5e7dee90 100644
--- a/fs/xfs/xfs_inode_item.h
+++ b/fs/xfs/xfs_inode_item.h
@@ -13,28 +13,38 @@ struct xfs_bmbt_rec;
 struct xfs_inode;
 struct xfs_mount;
 
-typedef struct xfs_inode_log_item {
+struct xfs_inode_log_item {
 	struct xfs_log_item	ili_item;	   /* common portion */
 	struct xfs_inode	*ili_inode;	   /* inode ptr */
-	xfs_lsn_t		ili_flush_lsn;	   /* lsn at last flush */
-	xfs_lsn_t		ili_last_lsn;	   /* lsn at last transaction */
-	unsigned short		ili_lock_flags;	   /* lock flags */
-	unsigned short		ili_logged;	   /* flushed logged data */
+	unsigned short		ili_lock_flags;	   /* inode lock flags */
+	/*
+	 * The ili_lock protects the interactions between the dirty state and
+	 * the flush state of the inode log item. This allows us to do atomic
+	 * modifications of multiple state fields without having to hold a
+	 * specific inode lock to serialise them.
+	 *
+	 * We need atomic changes between inode dirtying, inode flushing and
+	 * inode completion, but these all hold different combinations of
+	 * ILOCK and iflock and hence we need some other method of serialising
+	 * updates to the flush state.
+	 */
+	spinlock_t		ili_lock;	   /* flush state lock */
 	unsigned int		ili_last_fields;   /* fields when flushed */
 	unsigned int		ili_fields;	   /* fields to be logged */
 	unsigned int		ili_fsync_fields;  /* logged since last fsync */
-} xfs_inode_log_item_t;
+	xfs_lsn_t		ili_flush_lsn;	   /* lsn at last flush */
+	xfs_lsn_t		ili_last_lsn;	   /* lsn at last transaction */
+};
 
-static inline int xfs_inode_clean(xfs_inode_t *ip)
+static inline int xfs_inode_clean(struct xfs_inode *ip)
 {
 	return !ip->i_itemp || !(ip->i_itemp->ili_fields & XFS_ILOG_ALL);
 }
 
 extern void xfs_inode_item_init(struct xfs_inode *, struct xfs_mount *);
 extern void xfs_inode_item_destroy(struct xfs_inode *);
-extern void xfs_iflush_done(struct xfs_buf *, struct xfs_log_item *);
-extern void xfs_istale_done(struct xfs_buf *, struct xfs_log_item *);
-extern void xfs_iflush_abort(struct xfs_inode *, bool);
+extern void xfs_iflush_done(struct xfs_buf *);
+extern void xfs_iflush_abort(struct xfs_inode *);
 extern int xfs_inode_item_format_convert(xfs_log_iovec_t *,
 					 struct xfs_inode_log_format *);
 
diff --git a/fs/xfs/xfs_inode_item_recover.c b/fs/xfs/xfs_inode_item_recover.c
new file mode 100644
index 000000000000..5e0d291835b3
--- /dev/null
+++ b/fs/xfs/xfs_inode_item_recover.c
@@ -0,0 +1,394 @@
+// SPDX-License-Identifier: GPL-2.0
+/*
+ * Copyright (c) 2000-2006 Silicon Graphics, Inc.
+ * All Rights Reserved.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_log_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_inode.h"
+#include "xfs_trans.h"
+#include "xfs_inode_item.h"
+#include "xfs_trace.h"
+#include "xfs_trans_priv.h"
+#include "xfs_buf_item.h"
+#include "xfs_log.h"
+#include "xfs_error.h"
+#include "xfs_log_priv.h"
+#include "xfs_log_recover.h"
+#include "xfs_icache.h"
+#include "xfs_bmap_btree.h"
+
+STATIC void
+xlog_recover_inode_ra_pass2(
+	struct xlog                     *log,
+	struct xlog_recover_item        *item)
+{
+	if (item->ri_buf[0].i_len == sizeof(struct xfs_inode_log_format)) {
+		struct xfs_inode_log_format	*ilfp = item->ri_buf[0].i_addr;
+
+		xlog_buf_readahead(log, ilfp->ilf_blkno, ilfp->ilf_len,
+				   &xfs_inode_buf_ra_ops);
+	} else {
+		struct xfs_inode_log_format_32	*ilfp = item->ri_buf[0].i_addr;
+
+		xlog_buf_readahead(log, ilfp->ilf_blkno, ilfp->ilf_len,
+				   &xfs_inode_buf_ra_ops);
+	}
+}
+
+/*
+ * Inode fork owner changes
+ *
+ * If we have been told that we have to reparent the inode fork, it's because an
+ * extent swap operation on a CRC enabled filesystem has been done and we are
+ * replaying it. We need to walk the BMBT of the appropriate fork and change the
+ * owners of it.
+ *
+ * The complexity here is that we don't have an inode context to work with, so
+ * after we've replayed the inode we need to instantiate one.  This is where the
+ * fun begins.
+ *
+ * We are in the middle of log recovery, so we can't run transactions. That
+ * means we cannot use cache coherent inode instantiation via xfs_iget(), as
+ * that will result in the corresponding iput() running the inode through
+ * xfs_inactive(). If we've just replayed an inode core that changes the link
+ * count to zero (i.e. it's been unlinked), then xfs_inactive() will run
+ * transactions (bad!).
+ *
+ * So, to avoid this, we instantiate an inode directly from the inode core we've
+ * just recovered. We have the buffer still locked, and all we really need to
+ * instantiate is the inode core and the forks being modified. We can do this
+ * manually, then run the inode btree owner change, and then tear down the
+ * xfs_inode without having to run any transactions at all.
+ *
+ * Also, because we don't have a transaction context available here but need to
+ * gather all the buffers we modify for writeback so we pass the buffer_list
+ * instead for the operation to use.
+ */
+
+STATIC int
+xfs_recover_inode_owner_change(
+	struct xfs_mount	*mp,
+	struct xfs_dinode	*dip,
+	struct xfs_inode_log_format *in_f,
+	struct list_head	*buffer_list)
+{
+	struct xfs_inode	*ip;
+	int			error;
+
+	ASSERT(in_f->ilf_fields & (XFS_ILOG_DOWNER|XFS_ILOG_AOWNER));
+
+	ip = xfs_inode_alloc(mp, in_f->ilf_ino);
+	if (!ip)
+		return -ENOMEM;
+
+	/* instantiate the inode */
+	ASSERT(dip->di_version >= 3);
+
+	error = xfs_inode_from_disk(ip, dip);
+	if (error)
+		goto out_free_ip;
+
+	if (in_f->ilf_fields & XFS_ILOG_DOWNER) {
+		ASSERT(in_f->ilf_fields & XFS_ILOG_DBROOT);
+		error = xfs_bmbt_change_owner(NULL, ip, XFS_DATA_FORK,
+					      ip->i_ino, buffer_list);
+		if (error)
+			goto out_free_ip;
+	}
+
+	if (in_f->ilf_fields & XFS_ILOG_AOWNER) {
+		ASSERT(in_f->ilf_fields & XFS_ILOG_ABROOT);
+		error = xfs_bmbt_change_owner(NULL, ip, XFS_ATTR_FORK,
+					      ip->i_ino, buffer_list);
+		if (error)
+			goto out_free_ip;
+	}
+
+out_free_ip:
+	xfs_inode_free(ip);
+	return error;
+}
+
+STATIC int
+xlog_recover_inode_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			current_lsn)
+{
+	struct xfs_inode_log_format	*in_f;
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_buf			*bp;
+	struct xfs_dinode		*dip;
+	int				len;
+	char				*src;
+	char				*dest;
+	int				error;
+	int				attr_index;
+	uint				fields;
+	struct xfs_log_dinode		*ldip;
+	uint				isize;
+	int				need_free = 0;
+
+	if (item->ri_buf[0].i_len == sizeof(struct xfs_inode_log_format)) {
+		in_f = item->ri_buf[0].i_addr;
+	} else {
+		in_f = kmem_alloc(sizeof(struct xfs_inode_log_format), 0);
+		need_free = 1;
+		error = xfs_inode_item_format_convert(&item->ri_buf[0], in_f);
+		if (error)
+			goto error;
+	}
+
+	/*
+	 * Inode buffers can be freed, look out for it,
+	 * and do not replay the inode.
+	 */
+	if (xlog_is_buffer_cancelled(log, in_f->ilf_blkno, in_f->ilf_len)) {
+		error = 0;
+		trace_xfs_log_recover_inode_cancel(log, in_f);
+		goto error;
+	}
+	trace_xfs_log_recover_inode_recover(log, in_f);
+
+	error = xfs_buf_read(mp->m_ddev_targp, in_f->ilf_blkno, in_f->ilf_len,
+			0, &bp, &xfs_inode_buf_ops);
+	if (error)
+		goto error;
+	ASSERT(in_f->ilf_fields & XFS_ILOG_CORE);
+	dip = xfs_buf_offset(bp, in_f->ilf_boffset);
+
+	/*
+	 * Make sure the place we're flushing out to really looks
+	 * like an inode!
+	 */
+	if (XFS_IS_CORRUPT(mp, !xfs_verify_magic16(bp, dip->di_magic))) {
+		xfs_alert(mp,
+	"%s: Bad inode magic number, dip = "PTR_FMT", dino bp = "PTR_FMT", ino = %Ld",
+			__func__, dip, bp, in_f->ilf_ino);
+		error = -EFSCORRUPTED;
+		goto out_release;
+	}
+	ldip = item->ri_buf[1].i_addr;
+	if (XFS_IS_CORRUPT(mp, ldip->di_magic != XFS_DINODE_MAGIC)) {
+		xfs_alert(mp,
+			"%s: Bad inode log record, rec ptr "PTR_FMT", ino %Ld",
+			__func__, item, in_f->ilf_ino);
+		error = -EFSCORRUPTED;
+		goto out_release;
+	}
+
+	/*
+	 * If the inode has an LSN in it, recover the inode only if it's less
+	 * than the lsn of the transaction we are replaying. Note: we still
+	 * need to replay an owner change even though the inode is more recent
+	 * than the transaction as there is no guarantee that all the btree
+	 * blocks are more recent than this transaction, too.
+	 */
+	if (dip->di_version >= 3) {
+		xfs_lsn_t	lsn = be64_to_cpu(dip->di_lsn);
+
+		if (lsn && lsn != -1 && XFS_LSN_CMP(lsn, current_lsn) >= 0) {
+			trace_xfs_log_recover_inode_skip(log, in_f);
+			error = 0;
+			goto out_owner_change;
+		}
+	}
+
+	/*
+	 * di_flushiter is only valid for v1/2 inodes. All changes for v3 inodes
+	 * are transactional and if ordering is necessary we can determine that
+	 * more accurately by the LSN field in the V3 inode core. Don't trust
+	 * the inode versions we might be changing them here - use the
+	 * superblock flag to determine whether we need to look at di_flushiter
+	 * to skip replay when the on disk inode is newer than the log one
+	 */
+	if (!xfs_sb_version_has_v3inode(&mp->m_sb) &&
+	    ldip->di_flushiter < be16_to_cpu(dip->di_flushiter)) {
+		/*
+		 * Deal with the wrap case, DI_MAX_FLUSH is less
+		 * than smaller numbers
+		 */
+		if (be16_to_cpu(dip->di_flushiter) == DI_MAX_FLUSH &&
+		    ldip->di_flushiter < (DI_MAX_FLUSH >> 1)) {
+			/* do nothing */
+		} else {
+			trace_xfs_log_recover_inode_skip(log, in_f);
+			error = 0;
+			goto out_release;
+		}
+	}
+
+	/* Take the opportunity to reset the flush iteration count */
+	ldip->di_flushiter = 0;
+
+	if (unlikely(S_ISREG(ldip->di_mode))) {
+		if ((ldip->di_format != XFS_DINODE_FMT_EXTENTS) &&
+		    (ldip->di_format != XFS_DINODE_FMT_BTREE)) {
+			XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(3)",
+					 XFS_ERRLEVEL_LOW, mp, ldip,
+					 sizeof(*ldip));
+			xfs_alert(mp,
+		"%s: Bad regular inode log record, rec ptr "PTR_FMT", "
+		"ino ptr = "PTR_FMT", ino bp = "PTR_FMT", ino %Ld",
+				__func__, item, dip, bp, in_f->ilf_ino);
+			error = -EFSCORRUPTED;
+			goto out_release;
+		}
+	} else if (unlikely(S_ISDIR(ldip->di_mode))) {
+		if ((ldip->di_format != XFS_DINODE_FMT_EXTENTS) &&
+		    (ldip->di_format != XFS_DINODE_FMT_BTREE) &&
+		    (ldip->di_format != XFS_DINODE_FMT_LOCAL)) {
+			XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(4)",
+					     XFS_ERRLEVEL_LOW, mp, ldip,
+					     sizeof(*ldip));
+			xfs_alert(mp,
+		"%s: Bad dir inode log record, rec ptr "PTR_FMT", "
+		"ino ptr = "PTR_FMT", ino bp = "PTR_FMT", ino %Ld",
+				__func__, item, dip, bp, in_f->ilf_ino);
+			error = -EFSCORRUPTED;
+			goto out_release;
+		}
+	}
+	if (unlikely(ldip->di_nextents + ldip->di_anextents > ldip->di_nblocks)){
+		XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(5)",
+				     XFS_ERRLEVEL_LOW, mp, ldip,
+				     sizeof(*ldip));
+		xfs_alert(mp,
+	"%s: Bad inode log record, rec ptr "PTR_FMT", dino ptr "PTR_FMT", "
+	"dino bp "PTR_FMT", ino %Ld, total extents = %d, nblocks = %Ld",
+			__func__, item, dip, bp, in_f->ilf_ino,
+			ldip->di_nextents + ldip->di_anextents,
+			ldip->di_nblocks);
+		error = -EFSCORRUPTED;
+		goto out_release;
+	}
+	if (unlikely(ldip->di_forkoff > mp->m_sb.sb_inodesize)) {
+		XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(6)",
+				     XFS_ERRLEVEL_LOW, mp, ldip,
+				     sizeof(*ldip));
+		xfs_alert(mp,
+	"%s: Bad inode log record, rec ptr "PTR_FMT", dino ptr "PTR_FMT", "
+	"dino bp "PTR_FMT", ino %Ld, forkoff 0x%x", __func__,
+			item, dip, bp, in_f->ilf_ino, ldip->di_forkoff);
+		error = -EFSCORRUPTED;
+		goto out_release;
+	}
+	isize = xfs_log_dinode_size(mp);
+	if (unlikely(item->ri_buf[1].i_len > isize)) {
+		XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(7)",
+				     XFS_ERRLEVEL_LOW, mp, ldip,
+				     sizeof(*ldip));
+		xfs_alert(mp,
+			"%s: Bad inode log record length %d, rec ptr "PTR_FMT,
+			__func__, item->ri_buf[1].i_len, item);
+		error = -EFSCORRUPTED;
+		goto out_release;
+	}
+
+	/* recover the log dinode inode into the on disk inode */
+	xfs_log_dinode_to_disk(ldip, dip);
+
+	fields = in_f->ilf_fields;
+	if (fields & XFS_ILOG_DEV)
+		xfs_dinode_put_rdev(dip, in_f->ilf_u.ilfu_rdev);
+
+	if (in_f->ilf_size == 2)
+		goto out_owner_change;
+	len = item->ri_buf[2].i_len;
+	src = item->ri_buf[2].i_addr;
+	ASSERT(in_f->ilf_size <= 4);
+	ASSERT((in_f->ilf_size == 3) || (fields & XFS_ILOG_AFORK));
+	ASSERT(!(fields & XFS_ILOG_DFORK) ||
+	       (len == in_f->ilf_dsize));
+
+	switch (fields & XFS_ILOG_DFORK) {
+	case XFS_ILOG_DDATA:
+	case XFS_ILOG_DEXT:
+		memcpy(XFS_DFORK_DPTR(dip), src, len);
+		break;
+
+	case XFS_ILOG_DBROOT:
+		xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src, len,
+				 (struct xfs_bmdr_block *)XFS_DFORK_DPTR(dip),
+				 XFS_DFORK_DSIZE(dip, mp));
+		break;
+
+	default:
+		/*
+		 * There are no data fork flags set.
+		 */
+		ASSERT((fields & XFS_ILOG_DFORK) == 0);
+		break;
+	}
+
+	/*
+	 * If we logged any attribute data, recover it.  There may or
+	 * may not have been any other non-core data logged in this
+	 * transaction.
+	 */
+	if (in_f->ilf_fields & XFS_ILOG_AFORK) {
+		if (in_f->ilf_fields & XFS_ILOG_DFORK) {
+			attr_index = 3;
+		} else {
+			attr_index = 2;
+		}
+		len = item->ri_buf[attr_index].i_len;
+		src = item->ri_buf[attr_index].i_addr;
+		ASSERT(len == in_f->ilf_asize);
+
+		switch (in_f->ilf_fields & XFS_ILOG_AFORK) {
+		case XFS_ILOG_ADATA:
+		case XFS_ILOG_AEXT:
+			dest = XFS_DFORK_APTR(dip);
+			ASSERT(len <= XFS_DFORK_ASIZE(dip, mp));
+			memcpy(dest, src, len);
+			break;
+
+		case XFS_ILOG_ABROOT:
+			dest = XFS_DFORK_APTR(dip);
+			xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src,
+					 len, (struct xfs_bmdr_block *)dest,
+					 XFS_DFORK_ASIZE(dip, mp));
+			break;
+
+		default:
+			xfs_warn(log->l_mp, "%s: Invalid flag", __func__);
+			ASSERT(0);
+			error = -EFSCORRUPTED;
+			goto out_release;
+		}
+	}
+
+out_owner_change:
+	/* Recover the swapext owner change unless inode has been deleted */
+	if ((in_f->ilf_fields & (XFS_ILOG_DOWNER|XFS_ILOG_AOWNER)) &&
+	    (dip->di_mode != 0))
+		error = xfs_recover_inode_owner_change(mp, dip, in_f,
+						       buffer_list);
+	/* re-generate the checksum. */
+	xfs_dinode_calc_crc(log->l_mp, dip);
+
+	ASSERT(bp->b_mount == mp);
+	bp->b_flags |= _XBF_LOGRECOVERY;
+	xfs_buf_delwri_queue(bp, buffer_list);
+
+out_release:
+	xfs_buf_relse(bp);
+error:
+	if (need_free)
+		kmem_free(in_f);
+	return error;
+}
+
+const struct xlog_recover_item_ops xlog_inode_item_ops = {
+	.item_type		= XFS_LI_INODE,
+	.ra_pass2		= xlog_recover_inode_ra_pass2,
+	.commit_pass2		= xlog_recover_inode_commit_pass2,
+};
diff --git a/fs/xfs/xfs_ioctl.c b/fs/xfs/xfs_ioctl.c
index d42de92cb283..6f22a66777cd 100644
--- a/fs/xfs/xfs_ioctl.c
+++ b/fs/xfs/xfs_ioctl.c
@@ -35,6 +35,8 @@
 #include "xfs_health.h"
 #include "xfs_reflink.h"
 #include "xfs_ioctl.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
 
 #include <linux/mount.h>
 #include <linux/namei.h>
@@ -292,62 +294,173 @@ xfs_readlink_by_handle(
 	return error;
 }
 
-STATIC int
-xfs_attrlist_by_handle(
-	struct file		*parfilp,
-	void			__user *arg)
+/*
+ * Format an attribute and copy it out to the user's buffer.
+ * Take care to check values and protect against them changing later,
+ * we may be reading them directly out of a user buffer.
+ */
+static void
+xfs_ioc_attr_put_listent(
+	struct xfs_attr_list_context *context,
+	int			flags,
+	unsigned char		*name,
+	int			namelen,
+	int			valuelen)
 {
-	int			error = -ENOMEM;
-	attrlist_cursor_kern_t	*cursor;
-	struct xfs_fsop_attrlist_handlereq __user	*p = arg;
-	xfs_fsop_attrlist_handlereq_t al_hreq;
-	struct dentry		*dentry;
-	char			*kbuf;
+	struct xfs_attrlist	*alist = context->buffer;
+	struct xfs_attrlist_ent	*aep;
+	int			arraytop;
 
-	if (!capable(CAP_SYS_ADMIN))
-		return -EPERM;
-	if (copy_from_user(&al_hreq, arg, sizeof(xfs_fsop_attrlist_handlereq_t)))
-		return -EFAULT;
-	if (al_hreq.buflen < sizeof(struct attrlist) ||
-	    al_hreq.buflen > XFS_XATTR_LIST_MAX)
+	ASSERT(!context->seen_enough);
+	ASSERT(context->count >= 0);
+	ASSERT(context->count < (ATTR_MAX_VALUELEN/8));
+	ASSERT(context->firstu >= sizeof(*alist));
+	ASSERT(context->firstu <= context->bufsize);
+
+	/*
+	 * Only list entries in the right namespace.
+	 */
+	if (context->attr_filter != (flags & XFS_ATTR_NSP_ONDISK_MASK))
+		return;
+
+	arraytop = sizeof(*alist) +
+			context->count * sizeof(alist->al_offset[0]);
+
+	/* decrement by the actual bytes used by the attr */
+	context->firstu -= round_up(offsetof(struct xfs_attrlist_ent, a_name) +
+			namelen + 1, sizeof(uint32_t));
+	if (context->firstu < arraytop) {
+		trace_xfs_attr_list_full(context);
+		alist->al_more = 1;
+		context->seen_enough = 1;
+		return;
+	}
+
+	aep = context->buffer + context->firstu;
+	aep->a_valuelen = valuelen;
+	memcpy(aep->a_name, name, namelen);
+	aep->a_name[namelen] = 0;
+	alist->al_offset[context->count++] = context->firstu;
+	alist->al_count = context->count;
+	trace_xfs_attr_list_add(context);
+}
+
+static unsigned int
+xfs_attr_filter(
+	u32			ioc_flags)
+{
+	if (ioc_flags & XFS_IOC_ATTR_ROOT)
+		return XFS_ATTR_ROOT;
+	if (ioc_flags & XFS_IOC_ATTR_SECURE)
+		return XFS_ATTR_SECURE;
+	return 0;
+}
+
+static unsigned int
+xfs_attr_flags(
+	u32			ioc_flags)
+{
+	if (ioc_flags & XFS_IOC_ATTR_CREATE)
+		return XATTR_CREATE;
+	if (ioc_flags & XFS_IOC_ATTR_REPLACE)
+		return XATTR_REPLACE;
+	return 0;
+}
+
+int
+xfs_ioc_attr_list(
+	struct xfs_inode		*dp,
+	void __user			*ubuf,
+	int				bufsize,
+	int				flags,
+	struct xfs_attrlist_cursor __user *ucursor)
+{
+	struct xfs_attr_list_context	context = { };
+	struct xfs_attrlist		*alist;
+	void				*buffer;
+	int				error;
+
+	if (bufsize < sizeof(struct xfs_attrlist) ||
+	    bufsize > XFS_XATTR_LIST_MAX)
 		return -EINVAL;
 
 	/*
 	 * Reject flags, only allow namespaces.
 	 */
-	if (al_hreq.flags & ~(ATTR_ROOT | ATTR_SECURE))
+	if (flags & ~(XFS_IOC_ATTR_ROOT | XFS_IOC_ATTR_SECURE))
+		return -EINVAL;
+	if (flags == (XFS_IOC_ATTR_ROOT | XFS_IOC_ATTR_SECURE))
 		return -EINVAL;
 
-	dentry = xfs_handlereq_to_dentry(parfilp, &al_hreq.hreq);
-	if (IS_ERR(dentry))
-		return PTR_ERR(dentry);
+	/*
+	 * Validate the cursor.
+	 */
+	if (copy_from_user(&context.cursor, ucursor, sizeof(context.cursor)))
+		return -EFAULT;
+	if (context.cursor.pad1 || context.cursor.pad2)
+		return -EINVAL;
+	if (!context.cursor.initted &&
+	    (context.cursor.hashval || context.cursor.blkno ||
+	     context.cursor.offset))
+		return -EINVAL;
 
-	kbuf = kmem_zalloc_large(al_hreq.buflen, 0);
-	if (!kbuf)
-		goto out_dput;
+	buffer = kmem_zalloc_large(bufsize, 0);
+	if (!buffer)
+		return -ENOMEM;
 
-	cursor = (attrlist_cursor_kern_t *)&al_hreq.pos;
-	error = xfs_attr_list(XFS_I(d_inode(dentry)), kbuf, al_hreq.buflen,
-					al_hreq.flags, cursor);
+	/*
+	 * Initialize the output buffer.
+	 */
+	context.dp = dp;
+	context.resynch = 1;
+	context.attr_filter = xfs_attr_filter(flags);
+	context.buffer = buffer;
+	context.bufsize = round_down(bufsize, sizeof(uint32_t));
+	context.firstu = context.bufsize;
+	context.put_listent = xfs_ioc_attr_put_listent;
+
+	alist = context.buffer;
+	alist->al_count = 0;
+	alist->al_more = 0;
+	alist->al_offset[0] = context.bufsize;
+
+	error = xfs_attr_list(&context);
 	if (error)
-		goto out_kfree;
+		goto out_free;
 
-	if (copy_to_user(&p->pos, cursor, sizeof(attrlist_cursor_kern_t))) {
+	if (copy_to_user(ubuf, buffer, bufsize) ||
+	    copy_to_user(ucursor, &context.cursor, sizeof(context.cursor)))
 		error = -EFAULT;
-		goto out_kfree;
-	}
+out_free:
+	kmem_free(buffer);
+	return error;
+}
 
-	if (copy_to_user(al_hreq.buffer, kbuf, al_hreq.buflen))
-		error = -EFAULT;
+STATIC int
+xfs_attrlist_by_handle(
+	struct file		*parfilp,
+	struct xfs_fsop_attrlist_handlereq __user *p)
+{
+	struct xfs_fsop_attrlist_handlereq al_hreq;
+	struct dentry		*dentry;
+	int			error = -ENOMEM;
 
-out_kfree:
-	kmem_free(kbuf);
-out_dput:
+	if (!capable(CAP_SYS_ADMIN))
+		return -EPERM;
+	if (copy_from_user(&al_hreq, p, sizeof(al_hreq)))
+		return -EFAULT;
+
+	dentry = xfs_handlereq_to_dentry(parfilp, &al_hreq.hreq);
+	if (IS_ERR(dentry))
+		return PTR_ERR(dentry);
+
+	error = xfs_ioc_attr_list(XFS_I(d_inode(dentry)), al_hreq.buffer,
+				  al_hreq.buflen, al_hreq.flags, &p->pos);
 	dput(dentry);
 	return error;
 }
 
-int
+static int
 xfs_attrmulti_attr_get(
 	struct inode		*inode,
 	unsigned char		*name,
@@ -355,31 +468,33 @@ xfs_attrmulti_attr_get(
 	uint32_t		*len,
 	uint32_t		flags)
 {
-	unsigned char		*kbuf;
-	int			error = -EFAULT;
-	size_t			namelen;
+	struct xfs_da_args	args = {
+		.dp		= XFS_I(inode),
+		.attr_filter	= xfs_attr_filter(flags),
+		.attr_flags	= xfs_attr_flags(flags),
+		.name		= name,
+		.namelen	= strlen(name),
+		.valuelen	= *len,
+	};
+	int			error;
 
 	if (*len > XFS_XATTR_SIZE_MAX)
 		return -EINVAL;
-	kbuf = kmem_zalloc_large(*len, 0);
-	if (!kbuf)
-		return -ENOMEM;
 
-	namelen = strlen(name);
-	error = xfs_attr_get(XFS_I(inode), name, namelen, &kbuf, (int *)len,
-			     flags);
+	error = xfs_attr_get(&args);
 	if (error)
 		goto out_kfree;
 
-	if (copy_to_user(ubuf, kbuf, *len))
+	*len = args.valuelen;
+	if (copy_to_user(ubuf, args.value, args.valuelen))
 		error = -EFAULT;
 
 out_kfree:
-	kmem_free(kbuf);
+	kmem_free(args.value);
 	return error;
 }
 
-int
+static int
 xfs_attrmulti_attr_set(
 	struct inode		*inode,
 	unsigned char		*name,
@@ -387,42 +502,75 @@ xfs_attrmulti_attr_set(
 	uint32_t		len,
 	uint32_t		flags)
 {
-	unsigned char		*kbuf;
+	struct xfs_da_args	args = {
+		.dp		= XFS_I(inode),
+		.attr_filter	= xfs_attr_filter(flags),
+		.attr_flags	= xfs_attr_flags(flags),
+		.name		= name,
+		.namelen	= strlen(name),
+	};
 	int			error;
-	size_t			namelen;
 
 	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
 		return -EPERM;
-	if (len > XFS_XATTR_SIZE_MAX)
-		return -EINVAL;
 
-	kbuf = memdup_user(ubuf, len);
-	if (IS_ERR(kbuf))
-		return PTR_ERR(kbuf);
+	if (ubuf) {
+		if (len > XFS_XATTR_SIZE_MAX)
+			return -EINVAL;
+		args.value = memdup_user(ubuf, len);
+		if (IS_ERR(args.value))
+			return PTR_ERR(args.value);
+		args.valuelen = len;
+	}
 
-	namelen = strlen(name);
-	error = xfs_attr_set(XFS_I(inode), name, namelen, kbuf, len, flags);
-	if (!error)
-		xfs_forget_acl(inode, name, flags);
-	kfree(kbuf);
+	error = xfs_attr_set(&args);
+	if (!error && (flags & XFS_IOC_ATTR_ROOT))
+		xfs_forget_acl(inode, name);
+	kfree(args.value);
 	return error;
 }
 
 int
-xfs_attrmulti_attr_remove(
+xfs_ioc_attrmulti_one(
+	struct file		*parfilp,
 	struct inode		*inode,
-	unsigned char		*name,
+	uint32_t		opcode,
+	void __user		*uname,
+	void __user		*value,
+	uint32_t		*len,
 	uint32_t		flags)
 {
+	unsigned char		*name;
 	int			error;
-	size_t			namelen;
 
-	if (IS_IMMUTABLE(inode) || IS_APPEND(inode))
-		return -EPERM;
-	namelen = strlen(name);
-	error = xfs_attr_remove(XFS_I(inode), name, namelen, flags);
-	if (!error)
-		xfs_forget_acl(inode, name, flags);
+	if ((flags & XFS_IOC_ATTR_ROOT) && (flags & XFS_IOC_ATTR_SECURE))
+		return -EINVAL;
+
+	name = strndup_user(uname, MAXNAMELEN);
+	if (IS_ERR(name))
+		return PTR_ERR(name);
+
+	switch (opcode) {
+	case ATTR_OP_GET:
+		error = xfs_attrmulti_attr_get(inode, name, value, len, flags);
+		break;
+	case ATTR_OP_REMOVE:
+		value = NULL;
+		*len = 0;
+		/* fall through */
+	case ATTR_OP_SET:
+		error = mnt_want_write_file(parfilp);
+		if (error)
+			break;
+		error = xfs_attrmulti_attr_set(inode, name, value, *len, flags);
+		mnt_drop_write_file(parfilp);
+		break;
+	default:
+		error = -EINVAL;
+		break;
+	}
+
+	kfree(name);
 	return error;
 }
 
@@ -436,7 +584,6 @@ xfs_attrmulti_by_handle(
 	xfs_fsop_attrmulti_handlereq_t am_hreq;
 	struct dentry		*dentry;
 	unsigned int		i, size;
-	unsigned char		*attr_name;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
@@ -462,63 +609,17 @@ xfs_attrmulti_by_handle(
 		goto out_dput;
 	}
 
-	error = -ENOMEM;
-	attr_name = kmalloc(MAXNAMELEN, GFP_KERNEL);
-	if (!attr_name)
-		goto out_kfree_ops;
-
 	error = 0;
 	for (i = 0; i < am_hreq.opcount; i++) {
-		if ((ops[i].am_flags & ATTR_ROOT) &&
-		    (ops[i].am_flags & ATTR_SECURE)) {
-			ops[i].am_error = -EINVAL;
-			continue;
-		}
-		ops[i].am_flags &= ~ATTR_KERNEL_FLAGS;
-
-		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;
-		if (ops[i].am_error < 0)
-			break;
-
-		switch (ops[i].am_opcode) {
-		case ATTR_OP_GET:
-			ops[i].am_error = xfs_attrmulti_attr_get(
-					d_inode(dentry), attr_name,
-					ops[i].am_attrvalue, &ops[i].am_length,
-					ops[i].am_flags);
-			break;
-		case ATTR_OP_SET:
-			ops[i].am_error = mnt_want_write_file(parfilp);
-			if (ops[i].am_error)
-				break;
-			ops[i].am_error = xfs_attrmulti_attr_set(
-					d_inode(dentry), attr_name,
-					ops[i].am_attrvalue, ops[i].am_length,
-					ops[i].am_flags);
-			mnt_drop_write_file(parfilp);
-			break;
-		case ATTR_OP_REMOVE:
-			ops[i].am_error = mnt_want_write_file(parfilp);
-			if (ops[i].am_error)
-				break;
-			ops[i].am_error = xfs_attrmulti_attr_remove(
-					d_inode(dentry), attr_name,
-					ops[i].am_flags);
-			mnt_drop_write_file(parfilp);
-			break;
-		default:
-			ops[i].am_error = -EINVAL;
-		}
+		ops[i].am_error = xfs_ioc_attrmulti_one(parfilp,
+				d_inode(dentry), ops[i].am_opcode,
+				ops[i].am_attrname, ops[i].am_attrvalue,
+				&ops[i].am_length, ops[i].am_flags);
 	}
 
 	if (copy_to_user(am_hreq.ops, ops, size))
 		error = -EFAULT;
 
-	kfree(attr_name);
- out_kfree_ops:
 	kfree(ops);
  out_dput:
 	dput(dentry);
@@ -974,13 +1075,18 @@ xfs_merge_ioc_xflags(
 		xflags |= FS_XFLAG_NODUMP;
 	else
 		xflags &= ~FS_XFLAG_NODUMP;
+	if (flags & FS_DAX_FL)
+		xflags |= FS_XFLAG_DAX;
+	else
+		xflags &= ~FS_XFLAG_DAX;
 
 	return xflags;
 }
 
 STATIC unsigned int
 xfs_di2lxflags(
-	uint16_t	di_flags)
+	uint16_t	di_flags,
+	uint64_t	di_flags2)
 {
 	unsigned int	flags = 0;
 
@@ -994,6 +1100,9 @@ xfs_di2lxflags(
 		flags |= FS_NOATIME_FL;
 	if (di_flags & XFS_DIFLAG_NODUMP)
 		flags |= FS_NODUMP_FL;
+	if (di_flags2 & XFS_DIFLAG2_DAX) {
+		flags |= FS_DAX_FL;
+	}
 	return flags;
 }
 
@@ -1003,26 +1112,17 @@ xfs_fill_fsxattr(
 	bool			attr,
 	struct fsxattr		*fa)
 {
+	struct xfs_ifork	*ifp = attr ? ip->i_afp : &ip->i_df;
+
 	simple_fill_fsxattr(fa, xfs_ip2xflags(ip));
 	fa->fsx_extsize = ip->i_d.di_extsize << ip->i_mount->m_sb.sb_blocklog;
 	fa->fsx_cowextsize = ip->i_d.di_cowextsize <<
 			ip->i_mount->m_sb.sb_blocklog;
 	fa->fsx_projid = ip->i_d.di_projid;
-
-	if (attr) {
-		if (ip->i_afp) {
-			if (ip->i_afp->if_flags & XFS_IFEXTENTS)
-				fa->fsx_nextents = xfs_iext_count(ip->i_afp);
-			else
-				fa->fsx_nextents = ip->i_d.di_anextents;
-		} else
-			fa->fsx_nextents = 0;
-	} else {
-		if (ip->i_df.if_flags & XFS_IFEXTENTS)
-			fa->fsx_nextents = xfs_iext_count(&ip->i_df);
-		else
-			fa->fsx_nextents = ip->i_d.di_nextents;
-	}
+	if (ifp && (ifp->if_flags & XFS_IFEXTENTS))
+		fa->fsx_nextents = xfs_iext_count(ifp);
+	else
+		fa->fsx_nextents = xfs_ifork_nextents(ifp);
 }
 
 STATIC int
@@ -1100,37 +1200,6 @@ xfs_flags2diflags2(
 	return di_flags2;
 }
 
-STATIC void
-xfs_diflags_to_linux(
-	struct xfs_inode	*ip)
-{
-	struct inode		*inode = VFS_I(ip);
-	unsigned int		xflags = xfs_ip2xflags(ip);
-
-	if (xflags & FS_XFLAG_IMMUTABLE)
-		inode->i_flags |= S_IMMUTABLE;
-	else
-		inode->i_flags &= ~S_IMMUTABLE;
-	if (xflags & FS_XFLAG_APPEND)
-		inode->i_flags |= S_APPEND;
-	else
-		inode->i_flags &= ~S_APPEND;
-	if (xflags & FS_XFLAG_SYNC)
-		inode->i_flags |= S_SYNC;
-	else
-		inode->i_flags &= ~S_SYNC;
-	if (xflags & FS_XFLAG_NOATIME)
-		inode->i_flags |= S_NOATIME;
-	else
-		inode->i_flags &= ~S_NOATIME;
-#if 0	/* disabled until the flag switching races are sorted out */
-	if (xflags & FS_XFLAG_DAX)
-		inode->i_flags |= S_DAX;
-	else
-		inode->i_flags &= ~S_DAX;
-#endif
-}
-
 static int
 xfs_ioctl_setattr_xflags(
 	struct xfs_trans	*tp,
@@ -1141,7 +1210,7 @@ xfs_ioctl_setattr_xflags(
 	uint64_t		di_flags2;
 
 	/* Can't change realtime flag if any extents are allocated. */
-	if ((ip->i_d.di_nextents || ip->i_delayed_blks) &&
+	if ((ip->i_df.if_nextents || ip->i_delayed_blks) &&
 	    XFS_IS_REALTIME_INODE(ip) != (fa->fsx_xflags & FS_XFLAG_REALTIME))
 		return -EINVAL;
 
@@ -1162,77 +1231,39 @@ xfs_ioctl_setattr_xflags(
 
 	/* diflags2 only valid for v3 inodes. */
 	di_flags2 = xfs_flags2diflags2(ip, fa->fsx_xflags);
-	if (di_flags2 && ip->i_d.di_version < 3)
+	if (di_flags2 && !xfs_sb_version_has_v3inode(&mp->m_sb))
 		return -EINVAL;
 
 	ip->i_d.di_flags = xfs_flags2diflags(ip, fa->fsx_xflags);
 	ip->i_d.di_flags2 = di_flags2;
 
-	xfs_diflags_to_linux(ip);
+	xfs_diflags_to_iflags(ip, false);
 	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_CHG);
 	xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 	XFS_STATS_INC(mp, xs_ig_attrchg);
 	return 0;
 }
 
-/*
- * If we are changing DAX flags, we have to ensure the file is clean and any
- * cached objects in the address space are invalidated and removed. This
- * requires us to lock out other IO and page faults similar to a truncate
- * operation. The locks need to be held until the transaction has been committed
- * so that the cache invalidation is atomic with respect to the DAX flag
- * manipulation.
- */
-static int
-xfs_ioctl_setattr_dax_invalidate(
+static void
+xfs_ioctl_setattr_prepare_dax(
 	struct xfs_inode	*ip,
-	struct fsxattr		*fa,
-	int			*join_flags)
+	struct fsxattr		*fa)
 {
-	struct inode		*inode = VFS_I(ip);
-	struct super_block	*sb = inode->i_sb;
-	int			error;
-
-	*join_flags = 0;
-
-	/*
-	 * It is only valid to set the DAX flag on regular files and
-	 * directories on filesystems where the block size is equal to the page
-	 * size. On directories it serves as an inherited hint so we don't
-	 * have to check the device for dax support or flush pagecache.
-	 */
-	if (fa->fsx_xflags & FS_XFLAG_DAX) {
-		struct xfs_buftarg	*target = xfs_inode_buftarg(ip);
-
-		if (!bdev_dax_supported(target->bt_bdev, sb->s_blocksize))
-			return -EINVAL;
-	}
-
-	/* If the DAX state is not changing, we have nothing to do here. */
-	if ((fa->fsx_xflags & FS_XFLAG_DAX) && IS_DAX(inode))
-		return 0;
-	if (!(fa->fsx_xflags & FS_XFLAG_DAX) && !IS_DAX(inode))
-		return 0;
+	struct xfs_mount	*mp = ip->i_mount;
+	struct inode            *inode = VFS_I(ip);
 
 	if (S_ISDIR(inode->i_mode))
-		return 0;
+		return;
 
-	/* lock, flush and invalidate mapping in preparation for flag change */
-	xfs_ilock(ip, XFS_MMAPLOCK_EXCL | XFS_IOLOCK_EXCL);
-	error = filemap_write_and_wait(inode->i_mapping);
-	if (error)
-		goto out_unlock;
-	error = invalidate_inode_pages2(inode->i_mapping);
-	if (error)
-		goto out_unlock;
-
-	*join_flags = XFS_MMAPLOCK_EXCL | XFS_IOLOCK_EXCL;
-	return 0;
-
-out_unlock:
-	xfs_iunlock(ip, XFS_MMAPLOCK_EXCL | XFS_IOLOCK_EXCL);
-	return error;
+	if ((mp->m_flags & XFS_MOUNT_DAX_ALWAYS) ||
+	    (mp->m_flags & XFS_MOUNT_DAX_NEVER))
+		return;
 
+	if (((fa->fsx_xflags & FS_XFLAG_DAX) &&
+	    !(ip->i_d.di_flags2 & XFS_DIFLAG2_DAX)) ||
+	    (!(fa->fsx_xflags & FS_XFLAG_DAX) &&
+	     (ip->i_d.di_flags2 & XFS_DIFLAG2_DAX)))
+		d_mark_dontcache(inode);
 }
 
 /*
@@ -1240,17 +1271,10 @@ out_unlock:
  * have permission to do so. On success, return a clean transaction and the
  * inode locked exclusively ready for further operation specific checks. On
  * failure, return an error without modifying or locking the inode.
- *
- * The inode might already be IO locked on call. If this is the case, it is
- * indicated in @join_flags and we take full responsibility for ensuring they
- * are unlocked from now on. Hence if we have an error here, we still have to
- * unlock them. Otherwise, once they are joined to the transaction, they will
- * be unlocked on commit/cancel.
  */
 static struct xfs_trans *
 xfs_ioctl_setattr_get_trans(
-	struct xfs_inode	*ip,
-	int			join_flags)
+	struct xfs_inode	*ip)
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_trans	*tp;
@@ -1267,8 +1291,7 @@ xfs_ioctl_setattr_get_trans(
 		goto out_unlock;
 
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL | join_flags);
-	join_flags = 0;
+	xfs_trans_ijoin(tp, ip, XFS_ILOCK_EXCL);
 
 	/*
 	 * CAP_FOWNER overrides the following restrictions:
@@ -1289,8 +1312,6 @@ xfs_ioctl_setattr_get_trans(
 out_cancel:
 	xfs_trans_cancel(tp);
 out_unlock:
-	if (join_flags)
-		xfs_iunlock(ip, join_flags);
 	return ERR_PTR(error);
 }
 
@@ -1319,7 +1340,7 @@ xfs_ioctl_setattr_check_extsize(
 	xfs_extlen_t		size;
 	xfs_fsblock_t		extsize_fsb;
 
-	if (S_ISREG(VFS_I(ip)->i_mode) && ip->i_d.di_nextents &&
+	if (S_ISREG(VFS_I(ip)->i_mode) && ip->i_df.if_nextents &&
 	    ((ip->i_d.di_extsize << mp->m_sb.sb_blocklog) != fa->fsx_extsize))
 		return -EINVAL;
 
@@ -1372,8 +1393,7 @@ xfs_ioctl_setattr_check_cowextsize(
 	if (!(fa->fsx_xflags & FS_XFLAG_COWEXTSIZE))
 		return 0;
 
-	if (!xfs_sb_version_hasreflink(&ip->i_mount->m_sb) ||
-	    ip->i_d.di_version != 3)
+	if (!xfs_sb_version_hasreflink(&ip->i_mount->m_sb))
 		return -EINVAL;
 
 	if (fa->fsx_cowextsize == 0)
@@ -1413,11 +1433,9 @@ xfs_ioctl_setattr(
 	struct fsxattr		old_fa;
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_trans	*tp;
-	struct xfs_dquot	*udqp = NULL;
 	struct xfs_dquot	*pdqp = NULL;
 	struct xfs_dquot	*olddquot = NULL;
 	int			code;
-	int			join_flags = 0;
 
 	trace_xfs_ioctl_setattr(ip);
 
@@ -1434,25 +1452,16 @@ xfs_ioctl_setattr(
 	 * because the i_*dquot fields will get updated anyway.
 	 */
 	if (XFS_IS_QUOTA_ON(mp)) {
-		code = xfs_qm_vop_dqalloc(ip, ip->i_d.di_uid,
-					 ip->i_d.di_gid, fa->fsx_projid,
-					 XFS_QMOPT_PQUOTA, &udqp, NULL, &pdqp);
+		code = xfs_qm_vop_dqalloc(ip, VFS_I(ip)->i_uid,
+				VFS_I(ip)->i_gid, fa->fsx_projid,
+				XFS_QMOPT_PQUOTA, NULL, NULL, &pdqp);
 		if (code)
 			return code;
 	}
 
-	/*
-	 * Changing DAX config may require inode locking for mapping
-	 * invalidation. These need to be held all the way to transaction commit
-	 * or cancel time, so need to be passed through to
-	 * xfs_ioctl_setattr_get_trans() so it can apply them to the join call
-	 * appropriately.
-	 */
-	code = xfs_ioctl_setattr_dax_invalidate(ip, fa, &join_flags);
-	if (code)
-		goto error_free_dquots;
+	xfs_ioctl_setattr_prepare_dax(ip, fa);
 
-	tp = xfs_ioctl_setattr_get_trans(ip, join_flags);
+	tp = xfs_ioctl_setattr_get_trans(ip);
 	if (IS_ERR(tp)) {
 		code = PTR_ERR(tp);
 		goto error_free_dquots;
@@ -1460,7 +1469,7 @@ xfs_ioctl_setattr(
 
 	if (XFS_IS_QUOTA_RUNNING(mp) && XFS_IS_PQUOTA_ON(mp) &&
 	    ip->i_d.di_projid != fa->fsx_projid) {
-		code = xfs_qm_vop_chown_reserve(tp, ip, udqp, NULL, pdqp,
+		code = xfs_qm_vop_chown_reserve(tp, ip, NULL, NULL, pdqp,
 				capable(CAP_FOWNER) ?  XFS_QMOPT_FORCE_RES : 0);
 		if (code)	/* out of quota */
 			goto error_trans_cancel;
@@ -1501,7 +1510,6 @@ xfs_ioctl_setattr(
 			olddquot = xfs_qm_vop_chown(tp, ip,
 						&ip->i_pdquot, pdqp);
 		}
-		ASSERT(ip->i_d.di_version > 1);
 		ip->i_d.di_projid = fa->fsx_projid;
 	}
 
@@ -1514,7 +1522,7 @@ xfs_ioctl_setattr(
 		ip->i_d.di_extsize = fa->fsx_extsize >> mp->m_sb.sb_blocklog;
 	else
 		ip->i_d.di_extsize = 0;
-	if (ip->i_d.di_version == 3 &&
+	if (xfs_sb_version_has_v3inode(&mp->m_sb) &&
 	    (ip->i_d.di_flags2 & XFS_DIFLAG2_COWEXTSIZE))
 		ip->i_d.di_cowextsize = fa->fsx_cowextsize >>
 				mp->m_sb.sb_blocklog;
@@ -1527,7 +1535,6 @@ xfs_ioctl_setattr(
 	 * Release any dquot(s) the inode had kept before chown.
 	 */
 	xfs_qm_dqrele(olddquot);
-	xfs_qm_dqrele(udqp);
 	xfs_qm_dqrele(pdqp);
 
 	return code;
@@ -1535,7 +1542,6 @@ xfs_ioctl_setattr(
 error_trans_cancel:
 	xfs_trans_cancel(tp);
 error_free_dquots:
-	xfs_qm_dqrele(udqp);
 	xfs_qm_dqrele(pdqp);
 	return code;
 }
@@ -1567,7 +1573,7 @@ xfs_ioc_getxflags(
 {
 	unsigned int		flags;
 
-	flags = xfs_di2lxflags(ip->i_d.di_flags);
+	flags = xfs_di2lxflags(ip->i_d.di_flags, ip->i_d.di_flags2);
 	if (copy_to_user(arg, &flags, sizeof(flags)))
 		return -EFAULT;
 	return 0;
@@ -1583,7 +1589,6 @@ xfs_ioc_setxflags(
 	struct fsxattr		fa;
 	struct fsxattr		old_fa;
 	unsigned int		flags;
-	int			join_flags = 0;
 	int			error;
 
 	if (copy_from_user(&flags, arg, sizeof(flags)))
@@ -1591,7 +1596,7 @@ xfs_ioc_setxflags(
 
 	if (flags & ~(FS_IMMUTABLE_FL | FS_APPEND_FL | \
 		      FS_NOATIME_FL | FS_NODUMP_FL | \
-		      FS_SYNC_FL))
+		      FS_SYNC_FL | FS_DAX_FL))
 		return -EOPNOTSUPP;
 
 	fa.fsx_xflags = xfs_merge_ioc_xflags(flags, xfs_ip2xflags(ip));
@@ -1600,18 +1605,9 @@ xfs_ioc_setxflags(
 	if (error)
 		return error;
 
-	/*
-	 * Changing DAX config may require inode locking for mapping
-	 * invalidation. These need to be held all the way to transaction commit
-	 * or cancel time, so need to be passed through to
-	 * xfs_ioctl_setattr_get_trans() so it can apply them to the join call
-	 * appropriately.
-	 */
-	error = xfs_ioctl_setattr_dax_invalidate(ip, &fa, &join_flags);
-	if (error)
-		goto out_drop_write;
+	xfs_ioctl_setattr_prepare_dax(ip, &fa);
 
-	tp = xfs_ioctl_setattr_get_trans(ip, join_flags);
+	tp = xfs_ioctl_setattr_get_trans(ip);
 	if (IS_ERR(tp)) {
 		error = PTR_ERR(tp);
 		goto out_drop_write;
@@ -1983,6 +1979,41 @@ out:
 	return error;
 }
 
+static inline int
+xfs_fs_eofblocks_from_user(
+	struct xfs_fs_eofblocks		*src,
+	struct xfs_eofblocks		*dst)
+{
+	if (src->eof_version != XFS_EOFBLOCKS_VERSION)
+		return -EINVAL;
+
+	if (src->eof_flags & ~XFS_EOF_FLAGS_VALID)
+		return -EINVAL;
+
+	if (memchr_inv(&src->pad32, 0, sizeof(src->pad32)) ||
+	    memchr_inv(src->pad64, 0, sizeof(src->pad64)))
+		return -EINVAL;
+
+	dst->eof_flags = src->eof_flags;
+	dst->eof_prid = src->eof_prid;
+	dst->eof_min_file_size = src->eof_min_file_size;
+
+	dst->eof_uid = INVALID_UID;
+	if (src->eof_flags & XFS_EOF_FLAGS_UID) {
+		dst->eof_uid = make_kuid(current_user_ns(), src->eof_uid);
+		if (!uid_valid(dst->eof_uid))
+			return -EINVAL;
+	}
+
+	dst->eof_gid = INVALID_GID;
+	if (src->eof_flags & XFS_EOF_FLAGS_GID) {
+		dst->eof_gid = make_kgid(current_user_ns(), src->eof_gid);
+		if (!gid_valid(dst->eof_gid))
+			return -EINVAL;
+	}
+	return 0;
+}
+
 /*
  * Note: some of the ioctl's return positive numbers as a
  * byte count indicating success, such as readlink_by_handle.
@@ -2264,7 +2295,10 @@ xfs_file_ioctl(
 		if (error)
 			return error;
 
-		return xfs_icache_free_eofblocks(mp, &keofb);
+		sb_start_write(mp->m_super);
+		error = xfs_icache_free_eofblocks(mp, &keofb);
+		sb_end_write(mp->m_super);
+		return error;
 	}
 
 	default:
diff --git a/fs/xfs/xfs_ioctl.h b/fs/xfs/xfs_ioctl.h
index 420bd95dc326..bab6a5a92407 100644
--- a/fs/xfs/xfs_ioctl.h
+++ b/fs/xfs/xfs_ioctl.h
@@ -6,6 +6,11 @@
 #ifndef __XFS_IOCTL_H__
 #define __XFS_IOCTL_H__
 
+struct xfs_bstat;
+struct xfs_ibulk;
+struct xfs_inogrp;
+
+
 extern int
 xfs_ioc_space(
 	struct file		*filp,
@@ -30,27 +35,11 @@ xfs_readlink_by_handle(
 	struct file		*parfilp,
 	xfs_fsop_handlereq_t	*hreq);
 
-extern int
-xfs_attrmulti_attr_get(
-	struct inode		*inode,
-	unsigned char		*name,
-	unsigned char		__user *ubuf,
-	uint32_t		*len,
-	uint32_t		flags);
-
-extern int
-xfs_attrmulti_attr_set(
-	struct inode		*inode,
-	unsigned char		*name,
-	const unsigned char	__user *ubuf,
-	uint32_t		len,
-	uint32_t		flags);
-
-extern int
-xfs_attrmulti_attr_remove(
-	struct inode		*inode,
-	unsigned char		*name,
-	uint32_t		flags);
+int xfs_ioc_attrmulti_one(struct file *parfilp, struct inode *inode,
+		uint32_t opcode, void __user *uname, void __user *value,
+		uint32_t *len, uint32_t flags);
+int xfs_ioc_attr_list(struct xfs_inode *dp, void __user *ubuf, int bufsize,
+	int flags, struct xfs_attrlist_cursor __user *ucursor);
 
 extern struct dentry *
 xfs_handle_to_dentry(
@@ -70,10 +59,6 @@ xfs_file_compat_ioctl(
 	unsigned int		cmd,
 	unsigned long		arg);
 
-struct xfs_ibulk;
-struct xfs_bstat;
-struct xfs_inogrp;
-
 int xfs_fsbulkstat_one_fmt(struct xfs_ibulk *breq,
 			   const struct xfs_bulkstat *bstat);
 int xfs_fsinumbers_fmt(struct xfs_ibulk *breq, const struct xfs_inumbers *igrp);
diff --git a/fs/xfs/xfs_ioctl32.c b/fs/xfs/xfs_ioctl32.c
index 769581a79c58..c1771e728117 100644
--- a/fs/xfs/xfs_ioctl32.c
+++ b/fs/xfs/xfs_ioctl32.c
@@ -352,56 +352,24 @@ xfs_compat_handlereq_to_dentry(
 STATIC int
 xfs_compat_attrlist_by_handle(
 	struct file		*parfilp,
-	void			__user *arg)
+	compat_xfs_fsop_attrlist_handlereq_t __user *p)
 {
-	int			error;
-	attrlist_cursor_kern_t	*cursor;
-	compat_xfs_fsop_attrlist_handlereq_t __user *p = arg;
 	compat_xfs_fsop_attrlist_handlereq_t al_hreq;
 	struct dentry		*dentry;
-	char			*kbuf;
+	int			error;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
-	if (copy_from_user(&al_hreq, arg,
-			   sizeof(compat_xfs_fsop_attrlist_handlereq_t)))
+	if (copy_from_user(&al_hreq, p, sizeof(al_hreq)))
 		return -EFAULT;
-	if (al_hreq.buflen < sizeof(struct attrlist) ||
-	    al_hreq.buflen > XFS_XATTR_LIST_MAX)
-		return -EINVAL;
-
-	/*
-	 * Reject flags, only allow namespaces.
-	 */
-	if (al_hreq.flags & ~(ATTR_ROOT | ATTR_SECURE))
-		return -EINVAL;
 
 	dentry = xfs_compat_handlereq_to_dentry(parfilp, &al_hreq.hreq);
 	if (IS_ERR(dentry))
 		return PTR_ERR(dentry);
 
-	error = -ENOMEM;
-	kbuf = kmem_zalloc_large(al_hreq.buflen, 0);
-	if (!kbuf)
-		goto out_dput;
-
-	cursor = (attrlist_cursor_kern_t *)&al_hreq.pos;
-	error = xfs_attr_list(XFS_I(d_inode(dentry)), kbuf, al_hreq.buflen,
-					al_hreq.flags, cursor);
-	if (error)
-		goto out_kfree;
-
-	if (copy_to_user(&p->pos, cursor, sizeof(attrlist_cursor_kern_t))) {
-		error = -EFAULT;
-		goto out_kfree;
-	}
-
-	if (copy_to_user(compat_ptr(al_hreq.buffer), kbuf, al_hreq.buflen))
-		error = -EFAULT;
-
-out_kfree:
-	kmem_free(kbuf);
-out_dput:
+	error = xfs_ioc_attr_list(XFS_I(d_inode(dentry)),
+			compat_ptr(al_hreq.buffer), al_hreq.buflen,
+			al_hreq.flags, &p->pos);
 	dput(dentry);
 	return error;
 }
@@ -416,7 +384,6 @@ xfs_compat_attrmulti_by_handle(
 	compat_xfs_fsop_attrmulti_handlereq_t	am_hreq;
 	struct dentry				*dentry;
 	unsigned int				i, size;
-	unsigned char				*attr_name;
 
 	if (!capable(CAP_SYS_ADMIN))
 		return -EPERM;
@@ -443,64 +410,18 @@ xfs_compat_attrmulti_by_handle(
 		goto out_dput;
 	}
 
-	error = -ENOMEM;
-	attr_name = kmalloc(MAXNAMELEN, GFP_KERNEL);
-	if (!attr_name)
-		goto out_kfree_ops;
-
 	error = 0;
 	for (i = 0; i < am_hreq.opcount; i++) {
-		if ((ops[i].am_flags & ATTR_ROOT) &&
-		    (ops[i].am_flags & ATTR_SECURE)) {
-			ops[i].am_error = -EINVAL;
-			continue;
-		}
-		ops[i].am_flags &= ~ATTR_KERNEL_FLAGS;
-
-		ops[i].am_error = strncpy_from_user((char *)attr_name,
+		ops[i].am_error = xfs_ioc_attrmulti_one(parfilp,
+				d_inode(dentry), ops[i].am_opcode,
 				compat_ptr(ops[i].am_attrname),
-				MAXNAMELEN);
-		if (ops[i].am_error == 0 || ops[i].am_error == MAXNAMELEN)
-			error = -ERANGE;
-		if (ops[i].am_error < 0)
-			break;
-
-		switch (ops[i].am_opcode) {
-		case ATTR_OP_GET:
-			ops[i].am_error = xfs_attrmulti_attr_get(
-					d_inode(dentry), attr_name,
-					compat_ptr(ops[i].am_attrvalue),
-					&ops[i].am_length, ops[i].am_flags);
-			break;
-		case ATTR_OP_SET:
-			ops[i].am_error = mnt_want_write_file(parfilp);
-			if (ops[i].am_error)
-				break;
-			ops[i].am_error = xfs_attrmulti_attr_set(
-					d_inode(dentry), attr_name,
-					compat_ptr(ops[i].am_attrvalue),
-					ops[i].am_length, ops[i].am_flags);
-			mnt_drop_write_file(parfilp);
-			break;
-		case ATTR_OP_REMOVE:
-			ops[i].am_error = mnt_want_write_file(parfilp);
-			if (ops[i].am_error)
-				break;
-			ops[i].am_error = xfs_attrmulti_attr_remove(
-					d_inode(dentry), attr_name,
-					ops[i].am_flags);
-			mnt_drop_write_file(parfilp);
-			break;
-		default:
-			ops[i].am_error = -EINVAL;
-		}
+				compat_ptr(ops[i].am_attrvalue),
+				&ops[i].am_length, ops[i].am_flags);
 	}
 
 	if (copy_to_user(compat_ptr(am_hreq.ops), ops, size))
 		error = -EFAULT;
 
-	kfree(attr_name);
- out_kfree_ops:
 	kfree(ops);
  out_dput:
 	dput(dentry);
diff --git a/fs/xfs/xfs_iomap.c b/fs/xfs/xfs_iomap.c
index bb590a267a7f..3abb8b9d6f4c 100644
--- a/fs/xfs/xfs_iomap.c
+++ b/fs/xfs/xfs_iomap.c
@@ -293,11 +293,11 @@ out_trans_cancel:
 
 STATIC bool
 xfs_quota_need_throttle(
-	struct xfs_inode *ip,
-	int type,
-	xfs_fsblock_t alloc_blocks)
+	struct xfs_inode	*ip,
+	xfs_dqtype_t		type,
+	xfs_fsblock_t		alloc_blocks)
 {
-	struct xfs_dquot *dq = xfs_inode_dquot(ip, type);
+	struct xfs_dquot	*dq = xfs_inode_dquot(ip, type);
 
 	if (!dq || !xfs_this_quota_on(ip->i_mount, type))
 		return false;
@@ -307,7 +307,7 @@ xfs_quota_need_throttle(
 		return false;
 
 	/* under the lo watermark, no throttle */
-	if (dq->q_res_bcount + alloc_blocks < dq->q_prealloc_lo_wmark)
+	if (dq->q_blk.reserved + alloc_blocks < dq->q_prealloc_lo_wmark)
 		return false;
 
 	return true;
@@ -315,24 +315,24 @@ xfs_quota_need_throttle(
 
 STATIC void
 xfs_quota_calc_throttle(
-	struct xfs_inode *ip,
-	int type,
-	xfs_fsblock_t *qblocks,
-	int *qshift,
-	int64_t	*qfreesp)
+	struct xfs_inode	*ip,
+	xfs_dqtype_t		type,
+	xfs_fsblock_t		*qblocks,
+	int			*qshift,
+	int64_t			*qfreesp)
 {
-	int64_t freesp;
-	int shift = 0;
-	struct xfs_dquot *dq = xfs_inode_dquot(ip, type);
+	struct xfs_dquot	*dq = xfs_inode_dquot(ip, type);
+	int64_t			freesp;
+	int			shift = 0;
 
 	/* no dq, or over hi wmark, squash the prealloc completely */
-	if (!dq || dq->q_res_bcount >= dq->q_prealloc_hi_wmark) {
+	if (!dq || dq->q_blk.reserved >= dq->q_prealloc_hi_wmark) {
 		*qblocks = 0;
 		*qfreesp = 0;
 		return;
 	}
 
-	freesp = dq->q_prealloc_hi_wmark - dq->q_res_bcount;
+	freesp = dq->q_prealloc_hi_wmark - dq->q_blk.reserved;
 	if (freesp < dq->q_low_space[XFS_QLOWSP_5_PCNT]) {
 		shift = 2;
 		if (freesp < dq->q_low_space[XFS_QLOWSP_3_PCNT])
@@ -352,22 +352,10 @@ xfs_quota_calc_throttle(
 }
 
 /*
- * If we are doing a write at the end of the file and there are no allocations
- * past this one, then extend the allocation out to the file system's write
- * iosize.
- *
  * If we don't have a user specified preallocation size, dynamically increase
  * the preallocation size as the size of the file grows.  Cap the maximum size
  * at a single extent or less if the filesystem is near full. The closer the
- * filesystem is to full, the smaller the maximum prealocation.
- *
- * As an exception we don't do any preallocation at all if the file is smaller
- * than the minimum preallocation and we are using the default dynamic
- * preallocation scheme, as it is likely this is the only write to the file that
- * is going to be done.
- *
- * We clean up any extra space left over when the file is closed in
- * xfs_inactive().
+ * filesystem is to being full, the smaller the maximum preallocation.
  */
 STATIC xfs_fsblock_t
 xfs_iomap_prealloc_size(
@@ -377,63 +365,70 @@ xfs_iomap_prealloc_size(
 	loff_t			count,
 	struct xfs_iext_cursor	*icur)
 {
+	struct xfs_iext_cursor	ncur = *icur;
+	struct xfs_bmbt_irec	prev, got;
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_ifork	*ifp = XFS_IFORK_PTR(ip, whichfork);
 	xfs_fileoff_t		offset_fsb = XFS_B_TO_FSBT(mp, offset);
-	struct xfs_bmbt_irec	prev;
-	int			shift = 0;
 	int64_t			freesp;
 	xfs_fsblock_t		qblocks;
-	int			qshift = 0;
 	xfs_fsblock_t		alloc_blocks = 0;
+	xfs_extlen_t		plen;
+	int			shift = 0;
+	int			qshift = 0;
 
-	if (offset + count <= XFS_ISIZE(ip))
-		return 0;
-
-	if (!(mp->m_flags & XFS_MOUNT_ALLOCSIZE) &&
-	    (XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_allocsize_blocks)))
+	/*
+	 * As an exception we don't do any preallocation at all if the file is
+	 * smaller than the minimum preallocation and we are using the default
+	 * dynamic preallocation scheme, as it is likely this is the only write
+	 * to the file that is going to be done.
+	 */
+	if (XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_allocsize_blocks))
 		return 0;
 
 	/*
-	 * If an explicit allocsize is set, the file is small, or we
-	 * are writing behind a hole, then use the minimum prealloc:
+	 * Use the minimum preallocation size for small files or if we are
+	 * writing right after a hole.
 	 */
-	if ((mp->m_flags & XFS_MOUNT_ALLOCSIZE) ||
-	    XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_dalign) ||
-	    !xfs_iext_peek_prev_extent(ifp, icur, &prev) ||
+	if (XFS_ISIZE(ip) < XFS_FSB_TO_B(mp, mp->m_dalign) ||
+	    !xfs_iext_prev_extent(ifp, &ncur, &prev) ||
 	    prev.br_startoff + prev.br_blockcount < offset_fsb)
 		return mp->m_allocsize_blocks;
 
 	/*
-	 * Determine the initial size of the preallocation. We are beyond the
-	 * current EOF here, but we need to take into account whether this is
-	 * a sparse write or an extending write when determining the
-	 * preallocation size.  Hence we need to look up the extent that ends
-	 * at the current write offset and use the result to determine the
-	 * preallocation size.
-	 *
-	 * If the extent is a hole, then preallocation is essentially disabled.
-	 * Otherwise we take the size of the preceding data extent as the basis
-	 * for the preallocation size. If the size of the extent is greater than
-	 * half the maximum extent length, then use the current offset as the
-	 * basis. This ensures that for large files the preallocation size
-	 * always extends to MAXEXTLEN rather than falling short due to things
-	 * like stripe unit/width alignment of real extents.
+	 * Take the size of the preceding data extents as the basis for the
+	 * preallocation size. Note that we don't care if the previous extents
+	 * are written or not.
 	 */
-	if (prev.br_blockcount <= (MAXEXTLEN >> 1))
-		alloc_blocks = prev.br_blockcount << 1;
-	else
+	plen = prev.br_blockcount;
+	while (xfs_iext_prev_extent(ifp, &ncur, &got)) {
+		if (plen > MAXEXTLEN / 2 ||
+		    isnullstartblock(got.br_startblock) ||
+		    got.br_startoff + got.br_blockcount != prev.br_startoff ||
+		    got.br_startblock + got.br_blockcount != prev.br_startblock)
+			break;
+		plen += got.br_blockcount;
+		prev = got;
+	}
+
+	/*
+	 * If the size of the extents is greater than half the maximum extent
+	 * length, then use the current offset as the basis.  This ensures that
+	 * for large files the preallocation size always extends to MAXEXTLEN
+	 * rather than falling short due to things like stripe unit/width
+	 * alignment of real extents.
+	 */
+	alloc_blocks = plen * 2;
+	if (alloc_blocks > MAXEXTLEN)
 		alloc_blocks = XFS_B_TO_FSB(mp, offset);
-	if (!alloc_blocks)
-		goto check_writeio;
 	qblocks = alloc_blocks;
 
 	/*
 	 * MAXEXTLEN is not a power of two value but we round the prealloc down
 	 * to the nearest power of two value after throttling. To prevent the
-	 * round down from unconditionally reducing the maximum supported prealloc
-	 * size, we round up first, apply appropriate throttling, round down and
-	 * cap the value to MAXEXTLEN.
+	 * round down from unconditionally reducing the maximum supported
+	 * prealloc size, we round up first, apply appropriate throttling,
+	 * round down and cap the value to MAXEXTLEN.
 	 */
 	alloc_blocks = XFS_FILEOFF_MIN(roundup_pow_of_two(MAXEXTLEN),
 				       alloc_blocks);
@@ -455,14 +450,14 @@ xfs_iomap_prealloc_size(
 	 * Check each quota to cap the prealloc size, provide a shift value to
 	 * throttle with and adjust amount of available space.
 	 */
-	if (xfs_quota_need_throttle(ip, XFS_DQ_USER, alloc_blocks))
-		xfs_quota_calc_throttle(ip, XFS_DQ_USER, &qblocks, &qshift,
+	if (xfs_quota_need_throttle(ip, XFS_DQTYPE_USER, alloc_blocks))
+		xfs_quota_calc_throttle(ip, XFS_DQTYPE_USER, &qblocks, &qshift,
 					&freesp);
-	if (xfs_quota_need_throttle(ip, XFS_DQ_GROUP, alloc_blocks))
-		xfs_quota_calc_throttle(ip, XFS_DQ_GROUP, &qblocks, &qshift,
+	if (xfs_quota_need_throttle(ip, XFS_DQTYPE_GROUP, alloc_blocks))
+		xfs_quota_calc_throttle(ip, XFS_DQTYPE_GROUP, &qblocks, &qshift,
 					&freesp);
-	if (xfs_quota_need_throttle(ip, XFS_DQ_PROJ, alloc_blocks))
-		xfs_quota_calc_throttle(ip, XFS_DQ_PROJ, &qblocks, &qshift,
+	if (xfs_quota_need_throttle(ip, XFS_DQTYPE_PROJ, alloc_blocks))
+		xfs_quota_calc_throttle(ip, XFS_DQTYPE_PROJ, &qblocks, &qshift,
 					&freesp);
 
 	/*
@@ -494,7 +489,6 @@ xfs_iomap_prealloc_size(
 	 */
 	while (alloc_blocks && alloc_blocks >= freesp)
 		alloc_blocks >>= 4;
-check_writeio:
 	if (alloc_blocks < mp->m_allocsize_blocks)
 		alloc_blocks = mp->m_allocsize_blocks;
 	trace_xfs_iomap_prealloc_size(ip, alloc_blocks, shift,
@@ -563,7 +557,7 @@ xfs_iomap_write_unwritten(
 		xfs_trans_ijoin(tp, ip, 0);
 
 		error = xfs_trans_reserve_quota_nblks(tp, ip, resblks, 0,
-				XFS_QMOPT_RES_REGBLKS);
+				XFS_QMOPT_RES_REGBLKS | XFS_QMOPT_FORCE_RES);
 		if (error)
 			goto error_on_bmapi_transaction;
 
@@ -856,7 +850,7 @@ xfs_buffered_write_iomap_begin(
 
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 
-	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(ip, XFS_DATA_FORK)) ||
+	if (XFS_IS_CORRUPT(mp, !xfs_ifork_has_extents(&ip->i_df)) ||
 	    XFS_TEST_ERROR(false, mp, XFS_ERRTAG_BMAPIFORMAT)) {
 		error = -EFSCORRUPTED;
 		goto out_unlock;
@@ -871,7 +865,7 @@ xfs_buffered_write_iomap_begin(
 	}
 
 	/*
-	 * Search the data fork fork first to look up our source mapping.  We
+	 * Search the data fork first to look up our source mapping.  We
 	 * always need the data fork map, as we have to return it to the
 	 * iomap code so that the higher level write code can read data in to
 	 * perform read-modify-write cycles for unaligned writes.
@@ -961,9 +955,16 @@ xfs_buffered_write_iomap_begin(
 	if (error)
 		goto out_unlock;
 
-	if (eof) {
-		prealloc_blocks = xfs_iomap_prealloc_size(ip, allocfork, offset,
-				count, &icur);
+	if (eof && offset + count > XFS_ISIZE(ip)) {
+		/*
+		 * Determine the initial size of the preallocation.
+		 * We clean up any extra preallocation when the file is closed.
+		 */
+		if (mp->m_flags & XFS_MOUNT_ALLOCSIZE)
+			prealloc_blocks = mp->m_allocsize_blocks;
+		else
+			prealloc_blocks = xfs_iomap_prealloc_size(ip, allocfork,
+						offset, count, &icur);
 		if (prealloc_blocks) {
 			xfs_extlen_t	align;
 			xfs_off_t	end_offset;
@@ -1258,12 +1259,12 @@ xfs_xattr_iomap_begin(
 	lockmode = xfs_ilock_attr_map_shared(ip);
 
 	/* if there are no attribute fork or extents, return ENOENT */
-	if (!XFS_IFORK_Q(ip) || !ip->i_d.di_anextents) {
+	if (!XFS_IFORK_Q(ip) || !ip->i_afp->if_nextents) {
 		error = -ENOENT;
 		goto out_unlock;
 	}
 
-	ASSERT(ip->i_d.di_aformat != XFS_DINODE_FMT_LOCAL);
+	ASSERT(ip->i_afp->if_format != XFS_DINODE_FMT_LOCAL);
 	error = xfs_bmapi_read(ip, offset_fsb, end_fsb - offset_fsb, &imap,
 			       &nimaps, XFS_BMAPI_ATTRFORK);
 out_unlock:
diff --git a/fs/xfs/xfs_iops.c b/fs/xfs/xfs_iops.c
index 81f2f93caec0..80a13c8561d8 100644
--- a/fs/xfs/xfs_iops.c
+++ b/fs/xfs/xfs_iops.c
@@ -22,17 +22,17 @@
 #include "xfs_iomap.h"
 #include "xfs_error.h"
 
-#include <linux/xattr.h>
 #include <linux/posix_acl.h>
 #include <linux/security.h>
 #include <linux/iversion.h>
+#include <linux/fiemap.h>
 
 /*
- * Directories have different lock order w.r.t. mmap_sem compared to regular
+ * Directories have different lock order w.r.t. mmap_lock compared to regular
  * files. This is due to readdir potentially triggering page faults on a user
  * buffer inside filldir(), and this happens with the ilock on the directory
  * held. For regular files, the lock order is the other way around - the
- * mmap_sem is taken during the page fault, and then we lock the ilock to do
+ * mmap_lock is taken during the page fault, and then we lock the ilock to do
  * block mapping. Hence we need a different class for the directory ilock so
  * that lockdep can tell them apart.
  */
@@ -50,10 +50,15 @@ xfs_initxattrs(
 	int			error = 0;
 
 	for (xattr = xattr_array; xattr->name != NULL; xattr++) {
-		error = xfs_attr_set(ip, xattr->name,
-				     strlen(xattr->name),
-				     xattr->value, xattr->value_len,
-				     ATTR_SECURE);
+		struct xfs_da_args	args = {
+			.dp		= ip,
+			.attr_filter	= XFS_ATTR_SECURE,
+			.name		= xattr->name,
+			.namelen	= strlen(xattr->name),
+			.value		= xattr->value,
+			.valuelen	= xattr->value_len,
+		};
+		error = xfs_attr_set(&args);
 		if (error < 0)
 			break;
 	}
@@ -553,7 +558,7 @@ xfs_vn_getattr(
 	stat->blocks =
 		XFS_FSB_TO_BB(mp, ip->i_d.di_nblocks + ip->i_delayed_blks);
 
-	if (ip->i_d.di_version == 3) {
+	if (xfs_sb_version_has_v3inode(&mp->m_sb)) {
 		if (request_mask & STATX_BTIME) {
 			stat->result_mask |= STATX_BTIME;
 			stat->btime = ip->i_d.di_crtime;
@@ -692,9 +697,7 @@ xfs_setattr_nonsize(
 		 */
 		ASSERT(udqp == NULL);
 		ASSERT(gdqp == NULL);
-		error = xfs_qm_vop_dqalloc(ip, xfs_kuid_to_uid(uid),
-					   xfs_kgid_to_gid(gid),
-					   ip->i_d.di_projid,
+		error = xfs_qm_vop_dqalloc(ip, uid, gid, ip->i_d.di_projid,
 					   qflags, &udqp, &gdqp, NULL);
 		if (error)
 			return error;
@@ -736,12 +739,7 @@ xfs_setattr_nonsize(
 			if (error)	/* out of quota */
 				goto out_cancel;
 		}
-	}
 
-	/*
-	 * Change file ownership.  Must be the owner or privileged.
-	 */
-	if (mask & (ATTR_UID|ATTR_GID)) {
 		/*
 		 * CAP_FSETID overrides the following restrictions:
 		 *
@@ -763,7 +761,6 @@ xfs_setattr_nonsize(
 				olddquot1 = xfs_qm_vop_chown(tp, ip,
 							&ip->i_udquot, udqp);
 			}
-			ip->i_d.di_uid = xfs_kuid_to_uid(uid);
 			inode->i_uid = uid;
 		}
 		if (!gid_eq(igid, gid)) {
@@ -775,7 +772,6 @@ xfs_setattr_nonsize(
 				olddquot2 = xfs_qm_vop_chown(tp, ip,
 							&ip->i_gdquot, gdqp);
 			}
-			ip->i_d.di_gid = xfs_kgid_to_gid(gid);
 			inode->i_gid = gid;
 		}
 	}
@@ -877,7 +873,7 @@ xfs_setattr_size(
 	/*
 	 * Short circuit the truncate case for zero length files.
 	 */
-	if (newsize == 0 && oldsize == 0 && ip->i_d.di_nextents == 0) {
+	if (newsize == 0 && oldsize == 0 && ip->i_df.if_nextents == 0) {
 		if (!(iattr->ia_valid & (ATTR_CTIME|ATTR_MTIME)))
 			return 0;
 
@@ -1243,13 +1239,12 @@ xfs_inode_supports_dax(
 {
 	struct xfs_mount	*mp = ip->i_mount;
 
-	/* Only supported on non-reflinked files. */
-	if (!S_ISREG(VFS_I(ip)->i_mode) || xfs_is_reflink_inode(ip))
+	/* Only supported on regular files. */
+	if (!S_ISREG(VFS_I(ip)->i_mode))
 		return false;
 
-	/* DAX mount option or DAX iflag must be set. */
-	if (!(mp->m_flags & XFS_MOUNT_DAX) &&
-	    !(ip->i_d.di_flags2 & XFS_DIFLAG2_DAX))
+	/* Only supported on non-reflinked files. */
+	if (xfs_is_reflink_inode(ip))
 		return false;
 
 	/* Block size must match page size */
@@ -1260,26 +1255,51 @@ xfs_inode_supports_dax(
 	return xfs_inode_buftarg(ip)->bt_daxdev != NULL;
 }
 
-STATIC void
+static bool
+xfs_inode_should_enable_dax(
+	struct xfs_inode *ip)
+{
+	if (!IS_ENABLED(CONFIG_FS_DAX))
+		return false;
+	if (ip->i_mount->m_flags & XFS_MOUNT_DAX_NEVER)
+		return false;
+	if (!xfs_inode_supports_dax(ip))
+		return false;
+	if (ip->i_mount->m_flags & XFS_MOUNT_DAX_ALWAYS)
+		return true;
+	if (ip->i_d.di_flags2 & XFS_DIFLAG2_DAX)
+		return true;
+	return false;
+}
+
+void
 xfs_diflags_to_iflags(
-	struct inode		*inode,
-	struct xfs_inode	*ip)
+	struct xfs_inode	*ip,
+	bool init)
 {
-	uint16_t		flags = ip->i_d.di_flags;
-
-	inode->i_flags &= ~(S_IMMUTABLE | S_APPEND | S_SYNC |
-			    S_NOATIME | S_DAX);
-
-	if (flags & XFS_DIFLAG_IMMUTABLE)
-		inode->i_flags |= S_IMMUTABLE;
-	if (flags & XFS_DIFLAG_APPEND)
-		inode->i_flags |= S_APPEND;
-	if (flags & XFS_DIFLAG_SYNC)
-		inode->i_flags |= S_SYNC;
-	if (flags & XFS_DIFLAG_NOATIME)
-		inode->i_flags |= S_NOATIME;
-	if (xfs_inode_supports_dax(ip))
-		inode->i_flags |= S_DAX;
+	struct inode            *inode = VFS_I(ip);
+	unsigned int            xflags = xfs_ip2xflags(ip);
+	unsigned int            flags = 0;
+
+	ASSERT(!(IS_DAX(inode) && init));
+
+	if (xflags & FS_XFLAG_IMMUTABLE)
+		flags |= S_IMMUTABLE;
+	if (xflags & FS_XFLAG_APPEND)
+		flags |= S_APPEND;
+	if (xflags & FS_XFLAG_SYNC)
+		flags |= S_SYNC;
+	if (xflags & FS_XFLAG_NOATIME)
+		flags |= S_NOATIME;
+	if (init && xfs_inode_should_enable_dax(ip))
+		flags |= S_DAX;
+
+	/*
+	 * S_DAX can only be set during inode initialization and is never set by
+	 * the VFS, so we cannot mask off S_DAX in i_flags.
+	 */
+	inode->i_flags &= ~(S_IMMUTABLE | S_APPEND | S_SYNC | S_NOATIME);
+	inode->i_flags |= flags;
 }
 
 /*
@@ -1304,11 +1324,8 @@ xfs_setup_inode(
 	/* make the inode look hashed for the writeback code */
 	inode_fake_hash(inode);
 
-	inode->i_uid    = xfs_uid_to_kuid(ip->i_d.di_uid);
-	inode->i_gid    = xfs_gid_to_kgid(ip->i_d.di_gid);
-
 	i_size_write(inode, ip->i_d.di_size);
-	xfs_diflags_to_iflags(inode, ip);
+	xfs_diflags_to_iflags(ip, true);
 
 	if (S_ISDIR(inode->i_mode)) {
 		/*
diff --git a/fs/xfs/xfs_itable.c b/fs/xfs/xfs_itable.c
index 4b31c29b7e6b..16ca97a7ff00 100644
--- a/fs/xfs/xfs_itable.c
+++ b/fs/xfs/xfs_itable.c
@@ -86,8 +86,8 @@ xfs_bulkstat_one_int(
 	 */
 	buf->bs_projectid = ip->i_d.di_projid;
 	buf->bs_ino = ino;
-	buf->bs_uid = dic->di_uid;
-	buf->bs_gid = dic->di_gid;
+	buf->bs_uid = i_uid_read(inode);
+	buf->bs_gid = i_gid_read(inode);
 	buf->bs_size = dic->di_size;
 
 	buf->bs_nlink = inode->i_nlink;
@@ -104,18 +104,18 @@ xfs_bulkstat_one_int(
 
 	buf->bs_xflags = xfs_ip2xflags(ip);
 	buf->bs_extsize_blks = dic->di_extsize;
-	buf->bs_extents = dic->di_nextents;
+	buf->bs_extents = xfs_ifork_nextents(&ip->i_df);
 	xfs_bulkstat_health(ip, buf);
-	buf->bs_aextents = dic->di_anextents;
+	buf->bs_aextents = xfs_ifork_nextents(ip->i_afp);
 	buf->bs_forkoff = XFS_IFORK_BOFF(ip);
 	buf->bs_version = XFS_BULKSTAT_VERSION_V5;
 
-	if (dic->di_version == 3) {
+	if (xfs_sb_version_has_v3inode(&mp->m_sb)) {
 		if (dic->di_flags2 & XFS_DIFLAG2_COWEXTSIZE)
 			buf->bs_cowextsize_blks = dic->di_cowextsize;
 	}
 
-	switch (dic->di_format) {
+	switch (ip->i_df.if_format) {
 	case XFS_DINODE_FMT_DEV:
 		buf->bs_rdev = sysv_encode_dev(inode->i_rdev);
 		buf->bs_blksize = BLKDEV_IOSIZE;
diff --git a/fs/xfs/xfs_linux.h b/fs/xfs/xfs_linux.h
index 8738bb03f253..ab737fed7b12 100644
--- a/fs/xfs/xfs_linux.h
+++ b/fs/xfs/xfs_linux.h
@@ -60,6 +60,7 @@ typedef __u32			xfs_nlink_t;
 #include <linux/list_sort.h>
 #include <linux/ratelimit.h>
 #include <linux/rhashtable.h>
+#include <linux/xattr.h>
 
 #include <asm/page.h>
 #include <asm/div64.h>
@@ -101,12 +102,8 @@ typedef __u32			xfs_nlink_t;
 #define xfs_cowb_secs		xfs_params.cowb_timer.val
 
 #define current_cpu()		(raw_smp_processor_id())
-#define current_pid()		(current->pid)
-#define current_test_flags(f)	(current->flags & (f))
 #define current_set_flags_nested(sp, f)		\
 		(*(sp) = current->flags, current->flags |= (f))
-#define current_clear_flags_nested(sp, f)	\
-		(*(sp) = current->flags, current->flags &= ~(f))
 #define current_restore_flags_nested(sp, f)	\
 		(current->flags = ((current->flags & ~(f)) | (*(sp) & (f))))
 
@@ -163,32 +160,6 @@ struct xstats {
 
 extern struct xstats xfsstats;
 
-/* Kernel uid/gid conversion. These are used to convert to/from the on disk
- * uid_t/gid_t types to the kuid_t/kgid_t types that the kernel uses internally.
- * The conversion here is type only, the value will remain the same since we
- * are converting to the init_user_ns. The uid is later mapped to a particular
- * user namespace value when crossing the kernel/user boundary.
- */
-static inline uint32_t xfs_kuid_to_uid(kuid_t uid)
-{
-	return from_kuid(&init_user_ns, uid);
-}
-
-static inline kuid_t xfs_uid_to_kuid(uint32_t uid)
-{
-	return make_kuid(&init_user_ns, uid);
-}
-
-static inline uint32_t xfs_kgid_to_gid(kgid_t gid)
-{
-	return from_kgid(&init_user_ns, gid);
-}
-
-static inline kgid_t xfs_gid_to_kgid(uint32_t gid)
-{
-	return make_kgid(&init_user_ns, gid);
-}
-
 static inline dev_t xfs_to_linux_dev_t(xfs_dev_t dev)
 {
 	return MKDEV(sysv_major(dev) & 0x1ff, sysv_minor(dev));
diff --git a/fs/xfs/xfs_log.c b/fs/xfs/xfs_log.c
index f6006d94a581..ad0c69ee8947 100644
--- a/fs/xfs/xfs_log.c
+++ b/fs/xfs/xfs_log.c
@@ -24,13 +24,6 @@
 kmem_zone_t	*xfs_log_ticket_zone;
 
 /* Local miscellaneous function prototypes */
-STATIC int
-xlog_commit_record(
-	struct xlog		*log,
-	struct xlog_ticket	*ticket,
-	struct xlog_in_core	**iclog,
-	xfs_lsn_t		*commitlsnp);
-
 STATIC struct xlog *
 xlog_alloc_log(
 	struct xfs_mount	*mp,
@@ -47,8 +40,7 @@ xlog_dealloc_log(
 
 /* local state machine functions */
 STATIC void xlog_state_done_syncing(
-	struct xlog_in_core	*iclog,
-	bool			aborted);
+	struct xlog_in_core	*iclog);
 STATIC int
 xlog_state_get_iclog_space(
 	struct xlog		*log,
@@ -63,23 +55,10 @@ xlog_state_switch_iclogs(
 	struct xlog_in_core	*iclog,
 	int			eventual_size);
 STATIC void
-xlog_state_want_sync(
-	struct xlog		*log,
-	struct xlog_in_core	*iclog);
-
-STATIC void
 xlog_grant_push_ail(
 	struct xlog		*log,
 	int			need_bytes);
 STATIC void
-xlog_regrant_reserve_log_space(
-	struct xlog		*log,
-	struct xlog_ticket	*ticket);
-STATIC void
-xlog_ungrant_log_space(
-	struct xlog		*log,
-	struct xlog_ticket	*ticket);
-STATIC void
 xlog_sync(
 	struct xlog		*log,
 	struct xlog_in_core	*iclog);
@@ -454,7 +433,7 @@ xfs_log_reserve(
 	XFS_STATS_INC(mp, xs_try_logspace);
 
 	ASSERT(*ticp == NULL);
-	tic = xlog_ticket_alloc(log, unit_bytes, cnt, client, permanent, 0);
+	tic = xlog_ticket_alloc(log, unit_bytes, cnt, client, permanent);
 	*ticp = tic;
 
 	xlog_grant_push_ail(log, tic->t_cnt ? tic->t_unit_res * tic->t_cnt
@@ -484,73 +463,6 @@ out_error:
 	return error;
 }
 
-
-/*
- * NOTES:
- *
- *	1. currblock field gets updated at startup and after in-core logs
- *		marked as with WANT_SYNC.
- */
-
-/*
- * This routine is called when a user of a log manager ticket is done with
- * the reservation.  If the ticket was ever used, then a commit record for
- * the associated transaction is written out as a log operation header with
- * no data.  The flag XLOG_TIC_INITED is set when the first write occurs with
- * a given ticket.  If the ticket was one with a permanent reservation, then
- * a few operations are done differently.  Permanent reservation tickets by
- * default don't release the reservation.  They just commit the current
- * transaction with the belief that the reservation is still needed.  A flag
- * must be passed in before permanent reservations are actually released.
- * When these type of tickets are not released, they need to be set into
- * the inited state again.  By doing this, a start record will be written
- * out when the next write occurs.
- */
-xfs_lsn_t
-xfs_log_done(
-	struct xfs_mount	*mp,
-	struct xlog_ticket	*ticket,
-	struct xlog_in_core	**iclog,
-	bool			regrant)
-{
-	struct xlog		*log = mp->m_log;
-	xfs_lsn_t		lsn = 0;
-
-	if (XLOG_FORCED_SHUTDOWN(log) ||
-	    /*
-	     * If nothing was ever written, don't write out commit record.
-	     * If we get an error, just continue and give back the log ticket.
-	     */
-	    (((ticket->t_flags & XLOG_TIC_INITED) == 0) &&
-	     (xlog_commit_record(log, ticket, iclog, &lsn)))) {
-		lsn = (xfs_lsn_t) -1;
-		regrant = false;
-	}
-
-
-	if (!regrant) {
-		trace_xfs_log_done_nonperm(log, ticket);
-
-		/*
-		 * Release ticket if not permanent reservation or a specific
-		 * request has been made to release a permanent reservation.
-		 */
-		xlog_ungrant_log_space(log, ticket);
-	} else {
-		trace_xfs_log_done_perm(log, ticket);
-
-		xlog_regrant_reserve_log_space(log, ticket);
-		/* If this ticket was a permanent reservation and we aren't
-		 * trying to release it, reset the inited flags; so next time
-		 * we write, a start record will be written out.
-		 */
-		ticket->t_flags |= XLOG_TIC_INITED;
-	}
-
-	xfs_log_ticket_put(ticket);
-	return lsn;
-}
-
 static bool
 __xlog_state_release_iclog(
 	struct xlog		*log,
@@ -597,26 +509,21 @@ xlog_state_release_iclog(
 	return 0;
 }
 
-int
+void
 xfs_log_release_iclog(
-	struct xfs_mount        *mp,
 	struct xlog_in_core	*iclog)
 {
-	struct xlog		*log = mp->m_log;
-	bool			sync;
-
-	if (iclog->ic_state == XLOG_STATE_IOERROR) {
-		xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR);
-		return -EIO;
-	}
+	struct xlog		*log = iclog->ic_log;
+	bool			sync = false;
 
 	if (atomic_dec_and_lock(&iclog->ic_refcnt, &log->l_icloglock)) {
-		sync = __xlog_state_release_iclog(log, iclog);
+		if (iclog->ic_state != XLOG_STATE_IOERROR)
+			sync = __xlog_state_release_iclog(log, iclog);
 		spin_unlock(&log->l_icloglock);
-		if (sync)
-			xlog_sync(log, iclog);
 	}
-	return 0;
+
+	if (sync)
+		xlog_sync(log, iclog);
 }
 
 /*
@@ -855,32 +762,69 @@ xfs_log_mount_cancel(
 }
 
 /*
- * Final log writes as part of unmount.
- *
- * Mark the filesystem clean as unmount happens.  Note that during relocation
- * this routine needs to be executed as part of source-bag while the
- * deallocation must not be done until source-end.
+ * Wait for the iclog to be written disk, or return an error if the log has been
+ * shut down.
  */
+static int
+xlog_wait_on_iclog(
+	struct xlog_in_core	*iclog)
+		__releases(iclog->ic_log->l_icloglock)
+{
+	struct xlog		*log = iclog->ic_log;
 
-/* Actually write the unmount record to disk. */
-static void
-xfs_log_write_unmount_record(
-	struct xfs_mount	*mp)
+	if (!XLOG_FORCED_SHUTDOWN(log) &&
+	    iclog->ic_state != XLOG_STATE_ACTIVE &&
+	    iclog->ic_state != XLOG_STATE_DIRTY) {
+		XFS_STATS_INC(log->l_mp, xs_log_force_sleep);
+		xlog_wait(&iclog->ic_force_wait, &log->l_icloglock);
+	} else {
+		spin_unlock(&log->l_icloglock);
+	}
+
+	if (XLOG_FORCED_SHUTDOWN(log))
+		return -EIO;
+	return 0;
+}
+
+/*
+ * Write out an unmount record using the ticket provided. We have to account for
+ * the data space used in the unmount ticket as this write is not done from a
+ * transaction context that has already done the accounting for us.
+ */
+static int
+xlog_write_unmount_record(
+	struct xlog		*log,
+	struct xlog_ticket	*ticket,
+	xfs_lsn_t		*lsn,
+	uint			flags)
 {
-	/* the data section must be 32 bit size aligned */
-	struct xfs_unmount_log_format magic = {
+	struct xfs_unmount_log_format ulf = {
 		.magic = XLOG_UNMOUNT_TYPE,
 	};
 	struct xfs_log_iovec reg = {
-		.i_addr = &magic,
-		.i_len = sizeof(magic),
+		.i_addr = &ulf,
+		.i_len = sizeof(ulf),
 		.i_type = XLOG_REG_TYPE_UNMOUNT,
 	};
 	struct xfs_log_vec vec = {
 		.lv_niovecs = 1,
 		.lv_iovecp = &reg,
 	};
-	struct xlog		*log = mp->m_log;
+
+	/* account for space used by record data */
+	ticket->t_curr_res -= sizeof(ulf);
+	return xlog_write(log, &vec, ticket, lsn, NULL, flags, false);
+}
+
+/*
+ * Mark the filesystem clean by writing an unmount record to the head of the
+ * log.
+ */
+static void
+xlog_unmount_write(
+	struct xlog		*log)
+{
+	struct xfs_mount	*mp = log->l_mp;
 	struct xlog_in_core	*iclog;
 	struct xlog_ticket	*tic = NULL;
 	xfs_lsn_t		lsn;
@@ -891,23 +835,7 @@ xfs_log_write_unmount_record(
 	if (error)
 		goto out_err;
 
-	/*
-	 * If we think the summary counters are bad, clear the unmount header
-	 * flag in the unmount record so that the summary counters will be
-	 * recalculated during log recovery at next mount.  Refer to
-	 * xlog_check_unmount_rec for more details.
-	 */
-	if (XFS_TEST_ERROR(xfs_fs_has_sickness(mp, XFS_SICK_FS_COUNTERS), mp,
-			XFS_ERRTAG_FORCE_SUMMARY_RECALC)) {
-		xfs_alert(mp, "%s: will fix summary counters at next mount",
-				__func__);
-		flags &= ~XLOG_UNMOUNT_TRANS;
-	}
-
-	/* remove inited flag, and account for space used */
-	tic->t_flags = 0;
-	tic->t_curr_res -= sizeof(magic);
-	error = xlog_write(log, &vec, tic, &lsn, NULL, flags);
+	error = xlog_write_unmount_record(log, tic, &lsn, flags);
 	/*
 	 * At this point, we're umounting anyway, so there's no point in
 	 * transitioning log state to IOERROR. Just continue...
@@ -919,28 +847,32 @@ out_err:
 	spin_lock(&log->l_icloglock);
 	iclog = log->l_iclog;
 	atomic_inc(&iclog->ic_refcnt);
-	xlog_state_want_sync(log, iclog);
+	if (iclog->ic_state == XLOG_STATE_ACTIVE)
+		xlog_state_switch_iclogs(log, iclog, 0);
+	else
+		ASSERT(iclog->ic_state == XLOG_STATE_WANT_SYNC ||
+		       iclog->ic_state == XLOG_STATE_IOERROR);
 	error = xlog_state_release_iclog(log, iclog);
-	switch (iclog->ic_state) {
-	default:
-		if (!XLOG_FORCED_SHUTDOWN(log)) {
-			xlog_wait(&iclog->ic_force_wait, &log->l_icloglock);
-			break;
-		}
-		/* fall through */
-	case XLOG_STATE_ACTIVE:
-	case XLOG_STATE_DIRTY:
-		spin_unlock(&log->l_icloglock);
-		break;
-	}
+	xlog_wait_on_iclog(iclog);
 
 	if (tic) {
 		trace_xfs_log_umount_write(log, tic);
-		xlog_ungrant_log_space(log, tic);
-		xfs_log_ticket_put(tic);
+		xfs_log_ticket_ungrant(log, tic);
 	}
 }
 
+static void
+xfs_log_unmount_verify_iclog(
+	struct xlog		*log)
+{
+	struct xlog_in_core	*iclog = log->l_iclog;
+
+	do {
+		ASSERT(iclog->ic_state == XLOG_STATE_ACTIVE);
+		ASSERT(iclog->ic_offset == 0);
+	} while ((iclog = iclog->ic_next) != log->l_iclog);
+}
+
 /*
  * Unmount record used to have a string "Unmount filesystem--" in the
  * data section where the "Un" was really a magic number (XLOG_UNMOUNT_TYPE).
@@ -948,16 +880,11 @@ out_err:
  * currently architecture converted and "Unmount" is a bit foo.
  * As far as I know, there weren't any dependencies on the old behaviour.
  */
-
-static int
-xfs_log_unmount_write(xfs_mount_t *mp)
+static void
+xfs_log_unmount_write(
+	struct xfs_mount	*mp)
 {
-	struct xlog	 *log = mp->m_log;
-	xlog_in_core_t	 *iclog;
-#ifdef DEBUG
-	xlog_in_core_t	 *first_iclog;
-#endif
-	int		 error;
+	struct xlog		*log = mp->m_log;
 
 	/*
 	 * Don't write out unmount record on norecovery mounts or ro devices.
@@ -966,57 +893,30 @@ xfs_log_unmount_write(xfs_mount_t *mp)
 	if (mp->m_flags & XFS_MOUNT_NORECOVERY ||
 	    xfs_readonly_buftarg(log->l_targ)) {
 		ASSERT(mp->m_flags & XFS_MOUNT_RDONLY);
-		return 0;
+		return;
 	}
 
-	error = xfs_log_force(mp, XFS_LOG_SYNC);
-	ASSERT(error || !(XLOG_FORCED_SHUTDOWN(log)));
+	xfs_log_force(mp, XFS_LOG_SYNC);
 
-#ifdef DEBUG
-	first_iclog = iclog = log->l_iclog;
-	do {
-		if (iclog->ic_state != XLOG_STATE_IOERROR) {
-			ASSERT(iclog->ic_state == XLOG_STATE_ACTIVE);
-			ASSERT(iclog->ic_offset == 0);
-		}
-		iclog = iclog->ic_next;
-	} while (iclog != first_iclog);
-#endif
-	if (! (XLOG_FORCED_SHUTDOWN(log))) {
-		xfs_log_write_unmount_record(mp);
-	} else {
-		/*
-		 * We're already in forced_shutdown mode, couldn't
-		 * even attempt to write out the unmount transaction.
-		 *
-		 * Go through the motions of sync'ing and releasing
-		 * the iclog, even though no I/O will actually happen,
-		 * we need to wait for other log I/Os that may already
-		 * be in progress.  Do this as a separate section of
-		 * code so we'll know if we ever get stuck here that
-		 * we're in this odd situation of trying to unmount
-		 * a file system that went into forced_shutdown as
-		 * the result of an unmount..
-		 */
-		spin_lock(&log->l_icloglock);
-		iclog = log->l_iclog;
-		atomic_inc(&iclog->ic_refcnt);
-		xlog_state_want_sync(log, iclog);
-		error =  xlog_state_release_iclog(log, iclog);
-		switch (iclog->ic_state) {
-		case XLOG_STATE_ACTIVE:
-		case XLOG_STATE_DIRTY:
-		case XLOG_STATE_IOERROR:
-			spin_unlock(&log->l_icloglock);
-			break;
-		default:
-			xlog_wait(&iclog->ic_force_wait, &log->l_icloglock);
-			break;
-		}
+	if (XLOG_FORCED_SHUTDOWN(log))
+		return;
+
+	/*
+	 * If we think the summary counters are bad, avoid writing the unmount
+	 * record to force log recovery at next mount, after which the summary
+	 * counters will be recalculated.  Refer to xlog_check_unmount_rec for
+	 * more details.
+	 */
+	if (XFS_TEST_ERROR(xfs_fs_has_sickness(mp, XFS_SICK_FS_COUNTERS), mp,
+			XFS_ERRTAG_FORCE_SUMMARY_RECALC)) {
+		xfs_alert(mp, "%s: will fix summary counters at next mount",
+				__func__);
+		return;
 	}
 
-	return error;
-}	/* xfs_log_unmount_write */
+	xfs_log_unmount_verify_iclog(log);
+	xlog_unmount_write(log);
+}
 
 /*
  * Empty the log for unmount/freeze.
@@ -1279,7 +1179,6 @@ xlog_ioend_work(
 	struct xlog_in_core     *iclog =
 		container_of(work, struct xlog_in_core, ic_end_io_work);
 	struct xlog		*log = iclog->ic_log;
-	bool			aborted = false;
 	int			error;
 
 	error = blk_status_to_errno(iclog->ic_bio.bi_status);
@@ -1295,17 +1194,9 @@ xlog_ioend_work(
 	if (XFS_TEST_ERROR(error, log->l_mp, XFS_ERRTAG_IODONE_IOERR)) {
 		xfs_alert(log->l_mp, "log I/O error %d", error);
 		xfs_force_shutdown(log->l_mp, SHUTDOWN_LOG_IO_ERROR);
-		/*
-		 * This flag will be propagated to the trans-committed
-		 * callback routines to let them know that the log-commit
-		 * didn't succeed.
-		 */
-		aborted = true;
-	} else if (iclog->ic_state == XLOG_STATE_IOERROR) {
-		aborted = true;
 	}
 
-	xlog_state_done_syncing(iclog, aborted);
+	xlog_state_done_syncing(iclog);
 	bio_uninit(&iclog->ic_bio);
 
 	/*
@@ -1551,20 +1442,17 @@ out:
 	return ERR_PTR(error);
 }	/* xlog_alloc_log */
 
-
 /*
  * Write out the commit record of a transaction associated with the given
- * ticket.  Return the lsn of the commit record.
+ * ticket to close off a running log write. Return the lsn of the commit record.
  */
-STATIC int
+int
 xlog_commit_record(
 	struct xlog		*log,
 	struct xlog_ticket	*ticket,
 	struct xlog_in_core	**iclog,
-	xfs_lsn_t		*commitlsnp)
+	xfs_lsn_t		*lsn)
 {
-	struct xfs_mount *mp = log->l_mp;
-	int	error;
 	struct xfs_log_iovec reg = {
 		.i_addr = NULL,
 		.i_len = 0,
@@ -1574,12 +1462,15 @@ xlog_commit_record(
 		.lv_niovecs = 1,
 		.lv_iovecp = &reg,
 	};
+	int	error;
 
-	ASSERT_ALWAYS(iclog);
-	error = xlog_write(log, &vec, ticket, commitlsnp, iclog,
-					XLOG_COMMIT_TRANS);
+	if (XLOG_FORCED_SHUTDOWN(log))
+		return -EIO;
+
+	error = xlog_write(log, &vec, ticket, lsn, iclog, XLOG_COMMIT_TRANS,
+			   false);
 	if (error)
-		xfs_force_shutdown(mp, SHUTDOWN_LOG_IO_ERROR);
+		xfs_force_shutdown(log->l_mp, SHUTDOWN_LOG_IO_ERROR);
 	return error;
 }
 
@@ -1739,7 +1630,7 @@ xlog_bio_end_io(
 		   &iclog->ic_end_io_work);
 }
 
-static void
+static int
 xlog_map_iclog_data(
 	struct bio		*bio,
 	void			*data,
@@ -1750,11 +1641,14 @@ xlog_map_iclog_data(
 		unsigned int	off = offset_in_page(data);
 		size_t		len = min_t(size_t, count, PAGE_SIZE - off);
 
-		WARN_ON_ONCE(bio_add_page(bio, page, len, off) != len);
+		if (bio_add_page(bio, page, len, off) != len)
+			return -EIO;
 
 		data += len;
 		count -= len;
 	} while (count);
+
+	return 0;
 }
 
 STATIC void
@@ -1784,7 +1678,7 @@ xlog_write_iclog(
 		 * the buffer manually, the code needs to be kept in sync
 		 * with the I/O completion path.
 		 */
-		xlog_state_done_syncing(iclog, true);
+		xlog_state_done_syncing(iclog);
 		up(&iclog->ic_sema);
 		return;
 	}
@@ -1794,11 +1688,22 @@ xlog_write_iclog(
 	iclog->ic_bio.bi_iter.bi_sector = log->l_logBBstart + bno;
 	iclog->ic_bio.bi_end_io = xlog_bio_end_io;
 	iclog->ic_bio.bi_private = iclog;
-	iclog->ic_bio.bi_opf = REQ_OP_WRITE | REQ_META | REQ_SYNC | REQ_FUA;
+
+	/*
+	 * We use REQ_SYNC | REQ_IDLE here to tell the block layer the are more
+	 * IOs coming immediately after this one. This prevents the block layer
+	 * writeback throttle from throttling log writes behind background
+	 * metadata writeback and causing priority inversions.
+	 */
+	iclog->ic_bio.bi_opf = REQ_OP_WRITE | REQ_META | REQ_SYNC |
+				REQ_IDLE | REQ_FUA;
 	if (need_flush)
 		iclog->ic_bio.bi_opf |= REQ_PREFLUSH;
 
-	xlog_map_iclog_data(&iclog->ic_bio, iclog->ic_data, count);
+	if (xlog_map_iclog_data(&iclog->ic_bio, iclog->ic_data, count)) {
+		xfs_force_shutdown(log->l_mp, SHUTDOWN_LOG_IO_ERROR);
+		return;
+	}
 	if (is_vmalloc_addr(iclog->ic_data))
 		flush_kernel_vmap_range(iclog->ic_data, count);
 
@@ -2011,7 +1916,7 @@ xlog_dealloc_log(
 	log->l_mp->m_log = NULL;
 	destroy_workqueue(log->l_ioend_workqueue);
 	kmem_free(log);
-}	/* xlog_dealloc_log */
+}
 
 /*
  * Update counters atomically now that memcpy is done.
@@ -2148,23 +2053,21 @@ xlog_print_trans(
 }
 
 /*
- * Calculate the potential space needed by the log vector.  Each region gets
- * its own xlog_op_header_t and may need to be double word aligned.
+ * Calculate the potential space needed by the log vector.  We may need a start
+ * record, and each region gets its own struct xlog_op_header and may need to be
+ * double word aligned.
  */
 static int
 xlog_write_calc_vec_length(
 	struct xlog_ticket	*ticket,
-	struct xfs_log_vec	*log_vector)
+	struct xfs_log_vec	*log_vector,
+	bool			need_start_rec)
 {
 	struct xfs_log_vec	*lv;
-	int			headers = 0;
+	int			headers = need_start_rec ? 1 : 0;
 	int			len = 0;
 	int			i;
 
-	/* acct for start rec of xact */
-	if (ticket->t_flags & XLOG_TIC_INITED)
-		headers++;
-
 	for (lv = log_vector; lv; lv = lv->lv_next) {
 		/* we don't write ordered log vectors */
 		if (lv->lv_buf_len == XFS_LOG_VEC_ORDERED)
@@ -2186,27 +2089,16 @@ xlog_write_calc_vec_length(
 	return len;
 }
 
-/*
- * If first write for transaction, insert start record  We can't be trying to
- * commit if we are inited.  We can't have any "partial_copy" if we are inited.
- */
-static int
+static void
 xlog_write_start_rec(
 	struct xlog_op_header	*ophdr,
 	struct xlog_ticket	*ticket)
 {
-	if (!(ticket->t_flags & XLOG_TIC_INITED))
-		return 0;
-
 	ophdr->oh_tid	= cpu_to_be32(ticket->t_tid);
 	ophdr->oh_clientid = ticket->t_clientid;
 	ophdr->oh_len = 0;
 	ophdr->oh_flags = XLOG_START_TRANS;
 	ophdr->oh_res2 = 0;
-
-	ticket->t_flags &= ~XLOG_TIC_INITED;
-
-	return sizeof(struct xlog_op_header);
 }
 
 static xlog_op_header_t *
@@ -2328,7 +2220,11 @@ xlog_write_copy_finish(
 		*record_cnt = 0;
 		*data_cnt = 0;
 
-		xlog_state_want_sync(log, iclog);
+		if (iclog->ic_state == XLOG_STATE_ACTIVE)
+			xlog_state_switch_iclogs(log, iclog, 0);
+		else
+			ASSERT(iclog->ic_state == XLOG_STATE_WANT_SYNC ||
+			       iclog->ic_state == XLOG_STATE_IOERROR);
 		if (!commit_iclog)
 			goto release_iclog;
 		spin_unlock(&log->l_icloglock);
@@ -2391,13 +2287,14 @@ xlog_write(
 	struct xlog_ticket	*ticket,
 	xfs_lsn_t		*start_lsn,
 	struct xlog_in_core	**commit_iclog,
-	uint			flags)
+	uint			flags,
+	bool			need_start_rec)
 {
 	struct xlog_in_core	*iclog = NULL;
-	struct xfs_log_iovec	*vecp;
-	struct xfs_log_vec	*lv;
+	struct xfs_log_vec	*lv = log_vector;
+	struct xfs_log_iovec	*vecp = lv->lv_iovecp;
+	int			index = 0;
 	int			len;
-	int			index;
 	int			partial_copy = 0;
 	int			partial_copy_len = 0;
 	int			contwr = 0;
@@ -2405,25 +2302,13 @@ xlog_write(
 	int			data_cnt = 0;
 	int			error = 0;
 
-	*start_lsn = 0;
-
-	len = xlog_write_calc_vec_length(ticket, log_vector);
-
 	/*
-	 * Region headers and bytes are already accounted for.
-	 * We only need to take into account start records and
-	 * split regions in this function.
+	 * If this is a commit or unmount transaction, we don't need a start
+	 * record to be written.  We do, however, have to account for the
+	 * commit or unmount header that gets written. Hence we always have
+	 * to account for an extra xlog_op_header here.
 	 */
-	if (ticket->t_flags & XLOG_TIC_INITED)
-		ticket->t_curr_res -= sizeof(xlog_op_header_t);
-
-	/*
-	 * Commit record headers need to be accounted for. These
-	 * come in as separate writes so are easy to detect.
-	 */
-	if (flags & (XLOG_COMMIT_TRANS | XLOG_UNMOUNT_TRANS))
-		ticket->t_curr_res -= sizeof(xlog_op_header_t);
-
+	ticket->t_curr_res -= sizeof(struct xlog_op_header);
 	if (ticket->t_curr_res < 0) {
 		xfs_alert_tag(log->l_mp, XFS_PTAG_LOGRES,
 		     "ctx ticket reservation ran out. Need to up reservation");
@@ -2431,9 +2316,8 @@ xlog_write(
 		xfs_force_shutdown(log->l_mp, SHUTDOWN_LOG_IO_ERROR);
 	}
 
-	index = 0;
-	lv = log_vector;
-	vecp = lv->lv_iovecp;
+	len = xlog_write_calc_vec_length(ticket, log_vector, need_start_rec);
+	*start_lsn = 0;
 	while (lv && (!lv->lv_niovecs || index < lv->lv_niovecs)) {
 		void		*ptr;
 		int		log_offset;
@@ -2457,7 +2341,6 @@ xlog_write(
 		while (lv && (!lv->lv_niovecs || index < lv->lv_niovecs)) {
 			struct xfs_log_iovec	*reg;
 			struct xlog_op_header	*ophdr;
-			int			start_rec_copy;
 			int			copy_len;
 			int			copy_off;
 			bool			ordered = false;
@@ -2473,11 +2356,15 @@ xlog_write(
 			ASSERT(reg->i_len % sizeof(int32_t) == 0);
 			ASSERT((unsigned long)ptr % sizeof(int32_t) == 0);
 
-			start_rec_copy = xlog_write_start_rec(ptr, ticket);
-			if (start_rec_copy) {
-				record_cnt++;
+			/*
+			 * Before we start formatting log vectors, we need to
+			 * write a start record. Only do this for the first
+			 * iclog we write to.
+			 */
+			if (need_start_rec) {
+				xlog_write_start_rec(ptr, ticket);
 				xlog_write_adv_cnt(&ptr, &len, &log_offset,
-						   start_rec_copy);
+						sizeof(struct xlog_op_header));
 			}
 
 			ophdr = xlog_write_setup_ophdr(log, ptr, ticket, flags);
@@ -2509,8 +2396,13 @@ xlog_write(
 				xlog_write_adv_cnt(&ptr, &len, &log_offset,
 						   copy_len);
 			}
-			copy_len += start_rec_copy + sizeof(xlog_op_header_t);
+			copy_len += sizeof(struct xlog_op_header);
 			record_cnt++;
+			if (need_start_rec) {
+				copy_len += sizeof(struct xlog_op_header);
+				record_cnt++;
+				need_start_rec = false;
+			}
 			data_cnt += contwr ? copy_len : 0;
 
 			error = xlog_write_copy_finish(log, iclog, flags,
@@ -2567,119 +2459,106 @@ next_lv:
 	return error;
 }
 
+static void
+xlog_state_activate_iclog(
+	struct xlog_in_core	*iclog,
+	int			*iclogs_changed)
+{
+	ASSERT(list_empty_careful(&iclog->ic_callbacks));
 
-/*****************************************************************************
- *
- *		State Machine functions
- *
- *****************************************************************************
- */
+	/*
+	 * If the number of ops in this iclog indicate it just contains the
+	 * dummy transaction, we can change state into IDLE (the second time
+	 * around). Otherwise we should change the state into NEED a dummy.
+	 * We don't need to cover the dummy.
+	 */
+	if (*iclogs_changed == 0 &&
+	    iclog->ic_header.h_num_logops == cpu_to_be32(XLOG_COVER_OPS)) {
+		*iclogs_changed = 1;
+	} else {
+		/*
+		 * We have two dirty iclogs so start over.  This could also be
+		 * num of ops indicating this is not the dummy going out.
+		 */
+		*iclogs_changed = 2;
+	}
+
+	iclog->ic_state	= XLOG_STATE_ACTIVE;
+	iclog->ic_offset = 0;
+	iclog->ic_header.h_num_logops = 0;
+	memset(iclog->ic_header.h_cycle_data, 0,
+		sizeof(iclog->ic_header.h_cycle_data));
+	iclog->ic_header.h_lsn = 0;
+}
 
 /*
- * An iclog has just finished IO completion processing, so we need to update
- * the iclog state and propagate that up into the overall log state. Hence we
- * prepare the iclog for cleaning, and then clean all the pending dirty iclogs
- * starting from the head, and then wake up any threads that are waiting for the
- * iclog to be marked clean.
- *
- * The ordering of marking iclogs ACTIVE must be maintained, so an iclog
- * doesn't become ACTIVE beyond one that is SYNCING.  This is also required to
- * maintain the notion that we use a ordered wait queue to hold off would be
- * writers to the log when every iclog is trying to sync to disk.
- *
- * Caller must hold the icloglock before calling us.
- *
- * State Change: !IOERROR -> DIRTY -> ACTIVE
+ * Loop through all iclogs and mark all iclogs currently marked DIRTY as
+ * ACTIVE after iclog I/O has completed.
  */
-STATIC void
-xlog_state_clean_iclog(
+static void
+xlog_state_activate_iclogs(
 	struct xlog		*log,
-	struct xlog_in_core	*dirty_iclog)
+	int			*iclogs_changed)
 {
-	struct xlog_in_core	*iclog;
-	int			changed = 0;
-
-	/* Prepare the completed iclog. */
-	if (dirty_iclog->ic_state != XLOG_STATE_IOERROR)
-		dirty_iclog->ic_state = XLOG_STATE_DIRTY;
+	struct xlog_in_core	*iclog = log->l_iclog;
 
-	/* Walk all the iclogs to update the ordered active state. */
-	iclog = log->l_iclog;
 	do {
-		if (iclog->ic_state == XLOG_STATE_DIRTY) {
-			iclog->ic_state	= XLOG_STATE_ACTIVE;
-			iclog->ic_offset       = 0;
-			ASSERT(list_empty_careful(&iclog->ic_callbacks));
-			/*
-			 * If the number of ops in this iclog indicate it just
-			 * contains the dummy transaction, we can
-			 * change state into IDLE (the second time around).
-			 * Otherwise we should change the state into
-			 * NEED a dummy.
-			 * We don't need to cover the dummy.
-			 */
-			if (!changed &&
-			   (be32_to_cpu(iclog->ic_header.h_num_logops) ==
-			   		XLOG_COVER_OPS)) {
-				changed = 1;
-			} else {
-				/*
-				 * We have two dirty iclogs so start over
-				 * This could also be num of ops indicates
-				 * this is not the dummy going out.
-				 */
-				changed = 2;
-			}
-			iclog->ic_header.h_num_logops = 0;
-			memset(iclog->ic_header.h_cycle_data, 0,
-			      sizeof(iclog->ic_header.h_cycle_data));
-			iclog->ic_header.h_lsn = 0;
-		} else if (iclog->ic_state == XLOG_STATE_ACTIVE)
-			/* do nothing */;
-		else
-			break;	/* stop cleaning */
-		iclog = iclog->ic_next;
-	} while (iclog != log->l_iclog);
-
+		if (iclog->ic_state == XLOG_STATE_DIRTY)
+			xlog_state_activate_iclog(iclog, iclogs_changed);
+		/*
+		 * The ordering of marking iclogs ACTIVE must be maintained, so
+		 * an iclog doesn't become ACTIVE beyond one that is SYNCING.
+		 */
+		else if (iclog->ic_state != XLOG_STATE_ACTIVE)
+			break;
+	} while ((iclog = iclog->ic_next) != log->l_iclog);
+}
 
+static int
+xlog_covered_state(
+	int			prev_state,
+	int			iclogs_changed)
+{
 	/*
-	 * Wake up threads waiting in xfs_log_force() for the dirty iclog
-	 * to be cleaned.
+	 * We usually go to NEED. But we go to NEED2 if the changed indicates we
+	 * are done writing the dummy record.  If we are done with the second
+	 * dummy recored (DONE2), then we go to IDLE.
 	 */
-	wake_up_all(&dirty_iclog->ic_force_wait);
+	switch (prev_state) {
+	case XLOG_STATE_COVER_IDLE:
+	case XLOG_STATE_COVER_NEED:
+	case XLOG_STATE_COVER_NEED2:
+		break;
+	case XLOG_STATE_COVER_DONE:
+		if (iclogs_changed == 1)
+			return XLOG_STATE_COVER_NEED2;
+		break;
+	case XLOG_STATE_COVER_DONE2:
+		if (iclogs_changed == 1)
+			return XLOG_STATE_COVER_IDLE;
+		break;
+	default:
+		ASSERT(0);
+	}
 
-	/*
-	 * Change state for the dummy log recording.
-	 * We usually go to NEED. But we go to NEED2 if the changed indicates
-	 * we are done writing the dummy record.
-	 * If we are done with the second dummy recored (DONE2), then
-	 * we go to IDLE.
-	 */
-	if (changed) {
-		switch (log->l_covered_state) {
-		case XLOG_STATE_COVER_IDLE:
-		case XLOG_STATE_COVER_NEED:
-		case XLOG_STATE_COVER_NEED2:
-			log->l_covered_state = XLOG_STATE_COVER_NEED;
-			break;
+	return XLOG_STATE_COVER_NEED;
+}
 
-		case XLOG_STATE_COVER_DONE:
-			if (changed == 1)
-				log->l_covered_state = XLOG_STATE_COVER_NEED2;
-			else
-				log->l_covered_state = XLOG_STATE_COVER_NEED;
-			break;
+STATIC void
+xlog_state_clean_iclog(
+	struct xlog		*log,
+	struct xlog_in_core	*dirty_iclog)
+{
+	int			iclogs_changed = 0;
 
-		case XLOG_STATE_COVER_DONE2:
-			if (changed == 1)
-				log->l_covered_state = XLOG_STATE_COVER_IDLE;
-			else
-				log->l_covered_state = XLOG_STATE_COVER_NEED;
-			break;
+	dirty_iclog->ic_state = XLOG_STATE_DIRTY;
 
-		default:
-			ASSERT(0);
-		}
+	xlog_state_activate_iclogs(log, &iclogs_changed);
+	wake_up_all(&dirty_iclog->ic_force_wait);
+
+	if (iclogs_changed) {
+		log->l_covered_state = xlog_covered_state(log->l_covered_state,
+				iclogs_changed);
 	}
 }
 
@@ -2808,8 +2687,7 @@ xlog_state_iodone_process_iclog(
 static void
 xlog_state_do_iclog_callbacks(
 	struct xlog		*log,
-	struct xlog_in_core	*iclog,
-	bool			aborted)
+	struct xlog_in_core	*iclog)
 		__releases(&log->l_icloglock)
 		__acquires(&log->l_icloglock)
 {
@@ -2821,7 +2699,7 @@ xlog_state_do_iclog_callbacks(
 		list_splice_init(&iclog->ic_callbacks, &tmp);
 
 		spin_unlock(&iclog->ic_callback_lock);
-		xlog_cil_process_committed(&tmp, aborted);
+		xlog_cil_process_committed(&tmp);
 		spin_lock(&iclog->ic_callback_lock);
 	}
 
@@ -2836,8 +2714,7 @@ xlog_state_do_iclog_callbacks(
 
 STATIC void
 xlog_state_do_callback(
-	struct xlog		*log,
-	bool			aborted)
+	struct xlog		*log)
 {
 	struct xlog_in_core	*iclog;
 	struct xlog_in_core	*first_iclog;
@@ -2878,9 +2755,11 @@ xlog_state_do_callback(
 			 * we'll have to run at least one more complete loop.
 			 */
 			cycled_icloglock = true;
-			xlog_state_do_iclog_callbacks(log, iclog, aborted);
-
-			xlog_state_clean_iclog(log, iclog);
+			xlog_state_do_iclog_callbacks(log, iclog);
+			if (XLOG_FORCED_SHUTDOWN(log))
+				wake_up_all(&iclog->ic_force_wait);
+			else
+				xlog_state_clean_iclog(log, iclog);
 			iclog = iclog->ic_next;
 		} while (first_iclog != iclog);
 
@@ -2916,25 +2795,22 @@ xlog_state_do_callback(
  */
 STATIC void
 xlog_state_done_syncing(
-	struct xlog_in_core	*iclog,
-	bool			aborted)
+	struct xlog_in_core	*iclog)
 {
 	struct xlog		*log = iclog->ic_log;
 
 	spin_lock(&log->l_icloglock);
-
 	ASSERT(atomic_read(&iclog->ic_refcnt) == 0);
 
 	/*
 	 * If we got an error, either on the first buffer, or in the case of
-	 * split log writes, on the second, we mark ALL iclogs STATE_IOERROR,
-	 * and none should ever be attempted to be written to disk
-	 * again.
+	 * split log writes, on the second, we shut down the file system and
+	 * no iclogs should ever be attempted to be written to disk again.
 	 */
-	if (iclog->ic_state == XLOG_STATE_SYNCING)
+	if (!XLOG_FORCED_SHUTDOWN(log)) {
+		ASSERT(iclog->ic_state == XLOG_STATE_SYNCING);
 		iclog->ic_state = XLOG_STATE_DONE_SYNC;
-	else
-		ASSERT(iclog->ic_state == XLOG_STATE_IOERROR);
+	}
 
 	/*
 	 * Someone could be sleeping prior to writing out the next
@@ -2943,9 +2819,8 @@ xlog_state_done_syncing(
 	 */
 	wake_up_all(&iclog->ic_write_wait);
 	spin_unlock(&log->l_icloglock);
-	xlog_state_do_callback(log, aborted);	/* also cleans log */
-}	/* xlog_state_done_syncing */
-
+	xlog_state_do_callback(log);
+}
 
 /*
  * If the head of the in-core log ring is not (ACTIVE or DIRTY), then we must
@@ -3064,21 +2939,21 @@ restart:
 
 	*logoffsetp = log_offset;
 	return 0;
-}	/* xlog_state_get_iclog_space */
-
-/* The first cnt-1 times through here we don't need to
- * move the grant write head because the permanent
- * reservation has reserved cnt times the unit amount.
- * Release part of current permanent unit reservation and
- * reset current reservation to be one units worth.  Also
- * move grant reservation head forward.
+}
+
+/*
+ * The first cnt-1 times a ticket goes through here we don't need to move the
+ * grant write head because the permanent reservation has reserved cnt times the
+ * unit amount.  Release part of current permanent unit reservation and reset
+ * current reservation to be one units worth.  Also move grant reservation head
+ * forward.
  */
-STATIC void
-xlog_regrant_reserve_log_space(
+void
+xfs_log_ticket_regrant(
 	struct xlog		*log,
 	struct xlog_ticket	*ticket)
 {
-	trace_xfs_log_regrant_reserve_enter(log, ticket);
+	trace_xfs_log_ticket_regrant(log, ticket);
 
 	if (ticket->t_cnt > 0)
 		ticket->t_cnt--;
@@ -3090,21 +2965,20 @@ xlog_regrant_reserve_log_space(
 	ticket->t_curr_res = ticket->t_unit_res;
 	xlog_tic_reset_res(ticket);
 
-	trace_xfs_log_regrant_reserve_sub(log, ticket);
+	trace_xfs_log_ticket_regrant_sub(log, ticket);
 
 	/* just return if we still have some of the pre-reserved space */
-	if (ticket->t_cnt > 0)
-		return;
-
-	xlog_grant_add_space(log, &log->l_reserve_head.grant,
-					ticket->t_unit_res);
-
-	trace_xfs_log_regrant_reserve_exit(log, ticket);
+	if (!ticket->t_cnt) {
+		xlog_grant_add_space(log, &log->l_reserve_head.grant,
+				     ticket->t_unit_res);
+		trace_xfs_log_ticket_regrant_exit(log, ticket);
 
-	ticket->t_curr_res = ticket->t_unit_res;
-	xlog_tic_reset_res(ticket);
-}	/* xlog_regrant_reserve_log_space */
+		ticket->t_curr_res = ticket->t_unit_res;
+		xlog_tic_reset_res(ticket);
+	}
 
+	xfs_log_ticket_put(ticket);
+}
 
 /*
  * Give back the space left from a reservation.
@@ -3120,18 +2994,19 @@ xlog_regrant_reserve_log_space(
  * space, the count will stay at zero and the only space remaining will be
  * in the current reservation field.
  */
-STATIC void
-xlog_ungrant_log_space(
+void
+xfs_log_ticket_ungrant(
 	struct xlog		*log,
 	struct xlog_ticket	*ticket)
 {
-	int	bytes;
+	int			bytes;
+
+	trace_xfs_log_ticket_ungrant(log, ticket);
 
 	if (ticket->t_cnt > 0)
 		ticket->t_cnt--;
 
-	trace_xfs_log_ungrant_enter(log, ticket);
-	trace_xfs_log_ungrant_sub(log, ticket);
+	trace_xfs_log_ticket_ungrant_sub(log, ticket);
 
 	/*
 	 * If this is a permanent reservation ticket, we may be able to free
@@ -3146,17 +3021,15 @@ xlog_ungrant_log_space(
 	xlog_grant_sub_space(log, &log->l_reserve_head.grant, bytes);
 	xlog_grant_sub_space(log, &log->l_write_head.grant, bytes);
 
-	trace_xfs_log_ungrant_exit(log, ticket);
+	trace_xfs_log_ticket_ungrant_exit(log, ticket);
 
 	xfs_log_space_wake(log->l_mp);
+	xfs_log_ticket_put(ticket);
 }
 
 /*
- * This routine will mark the current iclog in the ring as WANT_SYNC
- * and move the current iclog pointer to the next iclog in the ring.
- * When this routine is called from xlog_state_get_iclog_space(), the
- * exact size of the iclog has not yet been determined.  All we know is
- * that every data block.  We have run out of space in this log record.
+ * This routine will mark the current iclog in the ring as WANT_SYNC and move
+ * the current iclog pointer to the next iclog in the ring.
  */
 STATIC void
 xlog_state_switch_iclogs(
@@ -3165,6 +3038,8 @@ xlog_state_switch_iclogs(
 	int			eventual_size)
 {
 	ASSERT(iclog->ic_state == XLOG_STATE_ACTIVE);
+	assert_spin_locked(&log->l_icloglock);
+
 	if (!eventual_size)
 		eventual_size = iclog->ic_offset;
 	iclog->ic_state = XLOG_STATE_WANT_SYNC;
@@ -3199,7 +3074,7 @@ xlog_state_switch_iclogs(
 	}
 	ASSERT(iclog == log->l_iclog);
 	log->l_iclog = iclog->ic_next;
-}	/* xlog_state_switch_iclogs */
+}
 
 /*
  * Write out all data in the in-core log as of this exact moment in time.
@@ -3259,9 +3134,6 @@ xfs_log_force(
 		 * previous iclog and go to sleep.
 		 */
 		iclog = iclog->ic_prev;
-		if (iclog->ic_state == XLOG_STATE_ACTIVE ||
-		    iclog->ic_state == XLOG_STATE_DIRTY)
-			goto out_unlock;
 	} else if (iclog->ic_state == XLOG_STATE_ACTIVE) {
 		if (atomic_read(&iclog->ic_refcnt) == 0) {
 			/*
@@ -3277,8 +3149,7 @@ xfs_log_force(
 			if (xlog_state_release_iclog(log, iclog))
 				goto out_error;
 
-			if (be64_to_cpu(iclog->ic_header.h_lsn) != lsn ||
-			    iclog->ic_state == XLOG_STATE_DIRTY)
+			if (be64_to_cpu(iclog->ic_header.h_lsn) != lsn)
 				goto out_unlock;
 		} else {
 			/*
@@ -3298,17 +3169,8 @@ xfs_log_force(
 		;
 	}
 
-	if (!(flags & XFS_LOG_SYNC))
-		goto out_unlock;
-
-	if (iclog->ic_state == XLOG_STATE_IOERROR)
-		goto out_error;
-	XFS_STATS_INC(mp, xs_log_force_sleep);
-	xlog_wait(&iclog->ic_force_wait, &log->l_icloglock);
-	if (iclog->ic_state == XLOG_STATE_IOERROR)
-		return -EIO;
-	return 0;
-
+	if (flags & XFS_LOG_SYNC)
+		return xlog_wait_on_iclog(iclog);
 out_unlock:
 	spin_unlock(&log->l_icloglock);
 	return 0;
@@ -3339,9 +3201,6 @@ __xfs_log_force_lsn(
 			goto out_unlock;
 	}
 
-	if (iclog->ic_state == XLOG_STATE_DIRTY)
-		goto out_unlock;
-
 	if (iclog->ic_state == XLOG_STATE_ACTIVE) {
 		/*
 		 * We sleep here if we haven't already slept (e.g. this is the
@@ -3375,20 +3234,8 @@ __xfs_log_force_lsn(
 			*log_flushed = 1;
 	}
 
-	if (!(flags & XFS_LOG_SYNC) ||
-	    (iclog->ic_state == XLOG_STATE_ACTIVE ||
-	     iclog->ic_state == XLOG_STATE_DIRTY))
-		goto out_unlock;
-
-	if (iclog->ic_state == XLOG_STATE_IOERROR)
-		goto out_error;
-
-	XFS_STATS_INC(mp, xs_log_force_sleep);
-	xlog_wait(&iclog->ic_force_wait, &log->l_icloglock);
-	if (iclog->ic_state == XLOG_STATE_IOERROR)
-		return -EIO;
-	return 0;
-
+	if (flags & XFS_LOG_SYNC)
+		return xlog_wait_on_iclog(iclog);
 out_unlock:
 	spin_unlock(&log->l_icloglock);
 	return 0;
@@ -3435,33 +3282,6 @@ xfs_log_force_lsn(
 }
 
 /*
- * Called when we want to mark the current iclog as being ready to sync to
- * disk.
- */
-STATIC void
-xlog_state_want_sync(
-	struct xlog		*log,
-	struct xlog_in_core	*iclog)
-{
-	assert_spin_locked(&log->l_icloglock);
-
-	if (iclog->ic_state == XLOG_STATE_ACTIVE) {
-		xlog_state_switch_iclogs(log, iclog, 0);
-	} else {
-		ASSERT(iclog->ic_state == XLOG_STATE_WANT_SYNC ||
-		       iclog->ic_state == XLOG_STATE_IOERROR);
-	}
-}
-
-
-/*****************************************************************************
- *
- *		TICKET functions
- *
- *****************************************************************************
- */
-
-/*
  * Free a used ticket when its refcount falls to zero.
  */
 void
@@ -3588,15 +3408,12 @@ xlog_ticket_alloc(
 	int			unit_bytes,
 	int			cnt,
 	char			client,
-	bool			permanent,
-	xfs_km_flags_t		alloc_flags)
+	bool			permanent)
 {
 	struct xlog_ticket	*tic;
 	int			unit_res;
 
-	tic = kmem_zone_zalloc(xfs_log_ticket_zone, alloc_flags);
-	if (!tic)
-		return NULL;
+	tic = kmem_cache_zalloc(xfs_log_ticket_zone, GFP_NOFS | __GFP_NOFAIL);
 
 	unit_res = xfs_log_calc_unit_res(log->l_mp, unit_bytes);
 
@@ -3609,7 +3426,6 @@ xlog_ticket_alloc(
 	tic->t_ocnt		= cnt;
 	tic->t_tid		= prandom_u32();
 	tic->t_clientid		= client;
-	tic->t_flags		= XLOG_TIC_INITED;
 	if (permanent)
 		tic->t_flags |= XLOG_TIC_PERM_RESERV;
 
@@ -3618,13 +3434,6 @@ xlog_ticket_alloc(
 	return tic;
 }
 
-
-/******************************************************************************
- *
- *		Log debug routines
- *
- ******************************************************************************
- */
 #if defined(DEBUG)
 /*
  * Make sure that the destination ptr is within the valid data region of
@@ -3710,7 +3519,7 @@ xlog_verify_tail_lsn(
 	if (blocks < BTOBB(iclog->ic_offset) + 1)
 		xfs_emerg(log->l_mp, "%s: ran out of log space", __func__);
     }
-}	/* xlog_verify_tail_lsn */
+}
 
 /*
  * Perform a number of checks on the iclog before writing to disk.
@@ -3813,7 +3622,7 @@ xlog_verify_iclog(
 		}
 		ptr += sizeof(xlog_op_header_t) + op_len;
 	}
-}	/* xlog_verify_iclog */
+}
 #endif
 
 /*
@@ -3937,7 +3746,7 @@ xfs_log_force_umount(
 	spin_lock(&log->l_cilp->xc_push_lock);
 	wake_up_all(&log->l_cilp->xc_commit_wait);
 	spin_unlock(&log->l_cilp->xc_push_lock);
-	xlog_state_do_callback(log, true);
+	xlog_state_do_callback(log);
 
 	/* return non-zero if log IOERROR transition had already happened */
 	return retval;
diff --git a/fs/xfs/xfs_log.h b/fs/xfs/xfs_log.h
index 84e06805160f..1412d6993f1e 100644
--- a/fs/xfs/xfs_log.h
+++ b/fs/xfs/xfs_log.h
@@ -105,10 +105,6 @@ struct xfs_log_item;
 struct xfs_item_ops;
 struct xfs_trans;
 
-xfs_lsn_t xfs_log_done(struct xfs_mount *mp,
-		       struct xlog_ticket *ticket,
-		       struct xlog_in_core **iclog,
-		       bool regrant);
 int	  xfs_log_force(struct xfs_mount *mp, uint flags);
 int	  xfs_log_force_lsn(struct xfs_mount *mp, xfs_lsn_t lsn, uint flags,
 		int *log_forced);
@@ -121,8 +117,7 @@ void	xfs_log_mount_cancel(struct xfs_mount *);
 xfs_lsn_t xlog_assign_tail_lsn(struct xfs_mount *mp);
 xfs_lsn_t xlog_assign_tail_lsn_locked(struct xfs_mount *mp);
 void	  xfs_log_space_wake(struct xfs_mount *mp);
-int	  xfs_log_release_iclog(struct xfs_mount *mp,
-			 struct xlog_in_core	 *iclog);
+void	  xfs_log_release_iclog(struct xlog_in_core *iclog);
 int	  xfs_log_reserve(struct xfs_mount *mp,
 			  int		   length,
 			  int		   count,
@@ -138,7 +133,7 @@ void	  xfs_log_ticket_put(struct xlog_ticket *ticket);
 
 void	xfs_log_commit_cil(struct xfs_mount *mp, struct xfs_trans *tp,
 				xfs_lsn_t *commit_lsn, bool regrant);
-void	xlog_cil_process_committed(struct list_head *list, bool aborted);
+void	xlog_cil_process_committed(struct list_head *list);
 bool	xfs_log_item_in_current_chkpt(struct xfs_log_item *lip);
 
 void	xfs_log_work_queue(struct xfs_mount *mp);
diff --git a/fs/xfs/xfs_log_cil.c b/fs/xfs/xfs_log_cil.c
index 48435cf2aa16..b0ef071b3cb5 100644
--- a/fs/xfs/xfs_log_cil.c
+++ b/fs/xfs/xfs_log_cil.c
@@ -37,8 +37,7 @@ xlog_cil_ticket_alloc(
 {
 	struct xlog_ticket *tic;
 
-	tic = xlog_ticket_alloc(log, 0, 1, XFS_TRANSACTION, 0,
-				KM_NOFS);
+	tic = xlog_ticket_alloc(log, 0, 1, XFS_TRANSACTION, 0);
 
 	/*
 	 * set the current reservation to zero so we know to steal the basic
@@ -240,7 +239,7 @@ xfs_cil_prepare_item(
 	 * this CIL context and so we need to pin it. If we are replacing the
 	 * old_lv, then remove the space it accounts for and make it the shadow
 	 * buffer for later freeing. In both cases we are now switching to the
-	 * shadow buffer, so update the the pointer to it appropriately.
+	 * shadow buffer, so update the pointer to it appropriately.
 	 */
 	if (!old_lv) {
 		if (lv->lv_item->li_ops->iop_pin)
@@ -574,10 +573,10 @@ xlog_discard_busy_extents(
  */
 static void
 xlog_cil_committed(
-	struct xfs_cil_ctx	*ctx,
-	bool			abort)
+	struct xfs_cil_ctx	*ctx)
 {
 	struct xfs_mount	*mp = ctx->cil->xc_log->l_mp;
+	bool			abort = XLOG_FORCED_SHUTDOWN(ctx->cil->xc_log);
 
 	/*
 	 * If the I/O failed, we're aborting the commit and already shutdown.
@@ -613,37 +612,38 @@ xlog_cil_committed(
 
 void
 xlog_cil_process_committed(
-	struct list_head	*list,
-	bool			aborted)
+	struct list_head	*list)
 {
 	struct xfs_cil_ctx	*ctx;
 
 	while ((ctx = list_first_entry_or_null(list,
 			struct xfs_cil_ctx, iclog_entry))) {
 		list_del(&ctx->iclog_entry);
-		xlog_cil_committed(ctx, aborted);
+		xlog_cil_committed(ctx);
 	}
 }
 
 /*
- * Push the Committed Item List to the log. If @push_seq flag is zero, then it
- * is a background flush and so we can chose to ignore it. Otherwise, if the
- * current sequence is the same as @push_seq we need to do a flush. If
- * @push_seq is less than the current sequence, then it has already been
+ * Push the Committed Item List to the log.
+ *
+ * If the current sequence is the same as xc_push_seq we need to do a flush. If
+ * xc_push_seq is less than the current sequence, then it has already been
  * flushed and we don't need to do anything - the caller will wait for it to
  * complete if necessary.
  *
- * @push_seq is a value rather than a flag because that allows us to do an
- * unlocked check of the sequence number for a match. Hence we can allows log
- * forces to run racily and not issue pushes for the same sequence twice. If we
- * get a race between multiple pushes for the same sequence they will block on
- * the first one and then abort, hence avoiding needless pushes.
+ * xc_push_seq is checked unlocked against the sequence number for a match.
+ * Hence we can allow log forces to run racily and not issue pushes for the
+ * same sequence twice.  If we get a race between multiple pushes for the same
+ * sequence they will block on the first one and then abort, hence avoiding
+ * needless pushes.
  */
-STATIC int
-xlog_cil_push(
-	struct xlog		*log)
+static void
+xlog_cil_push_work(
+	struct work_struct	*work)
 {
-	struct xfs_cil		*cil = log->l_cilp;
+	struct xfs_cil		*cil =
+		container_of(work, struct xfs_cil, xc_push_work);
+	struct xlog		*log = cil->xc_log;
 	struct xfs_log_vec	*lv;
 	struct xfs_cil_ctx	*ctx;
 	struct xfs_cil_ctx	*new_ctx;
@@ -657,9 +657,6 @@ xlog_cil_push(
 	xfs_lsn_t		commit_lsn;
 	xfs_lsn_t		push_seq;
 
-	if (!cil)
-		return 0;
-
 	new_ctx = kmem_zalloc(sizeof(*new_ctx), KM_NOFS);
 	new_ctx->ticket = xlog_cil_ticket_alloc(log);
 
@@ -671,6 +668,12 @@ xlog_cil_push(
 	ASSERT(push_seq <= ctx->sequence);
 
 	/*
+	 * Wake up any background push waiters now this context is being pushed.
+	 */
+	if (ctx->space_used >= XLOG_CIL_BLOCKING_SPACE_LIMIT(log))
+		wake_up_all(&cil->xc_push_wait);
+
+	/*
 	 * Check if we've anything to push. If there is nothing, then we don't
 	 * move on to a new sequence number and so we have to be able to push
 	 * this sequence again later.
@@ -740,7 +743,7 @@ xlog_cil_push(
 
 	/*
 	 * initialise the new context and attach it to the CIL. Then attach
-	 * the current context to the CIL committing lsit so it can be found
+	 * the current context to the CIL committing list so it can be found
 	 * during log forces to extract the commit lsn of the sequence that
 	 * needs to be forced.
 	 */
@@ -803,7 +806,7 @@ xlog_cil_push(
 	lvhdr.lv_iovecp = &lhdr;
 	lvhdr.lv_next = ctx->lv_chain;
 
-	error = xlog_write(log, &lvhdr, tic, &ctx->start_lsn, NULL, 0);
+	error = xlog_write(log, &lvhdr, tic, &ctx->start_lsn, NULL, 0, true);
 	if (error)
 		goto out_abort_free_ticket;
 
@@ -841,10 +844,11 @@ restart:
 	}
 	spin_unlock(&cil->xc_push_lock);
 
-	/* xfs_log_done always frees the ticket on error. */
-	commit_lsn = xfs_log_done(log->l_mp, tic, &commit_iclog, false);
-	if (commit_lsn == -1)
-		goto out_abort;
+	error = xlog_commit_record(log, tic, &commit_iclog, &commit_lsn);
+	if (error)
+		goto out_abort_free_ticket;
+
+	xfs_log_ticket_ungrant(log, tic);
 
 	spin_lock(&commit_iclog->ic_callback_lock);
 	if (commit_iclog->ic_state == XLOG_STATE_IOERROR) {
@@ -867,28 +871,20 @@ restart:
 	spin_unlock(&cil->xc_push_lock);
 
 	/* release the hounds! */
-	return xfs_log_release_iclog(log->l_mp, commit_iclog);
+	xfs_log_release_iclog(commit_iclog);
+	return;
 
 out_skip:
 	up_write(&cil->xc_ctx_lock);
 	xfs_log_ticket_put(new_ctx->ticket);
 	kmem_free(new_ctx);
-	return 0;
+	return;
 
 out_abort_free_ticket:
-	xfs_log_ticket_put(tic);
+	xfs_log_ticket_ungrant(log, tic);
 out_abort:
-	xlog_cil_committed(ctx, true);
-	return -EIO;
-}
-
-static void
-xlog_cil_push_work(
-	struct work_struct	*work)
-{
-	struct xfs_cil		*cil = container_of(work, struct xfs_cil,
-							xc_push_work);
-	xlog_cil_push(cil->xc_log);
+	ASSERT(XLOG_FORCED_SHUTDOWN(log));
+	xlog_cil_committed(ctx);
 }
 
 /*
@@ -900,7 +896,7 @@ xlog_cil_push_work(
  */
 static void
 xlog_cil_push_background(
-	struct xlog	*log)
+	struct xlog	*log) __releases(cil->xc_ctx_lock)
 {
 	struct xfs_cil	*cil = log->l_cilp;
 
@@ -914,14 +910,36 @@ xlog_cil_push_background(
 	 * don't do a background push if we haven't used up all the
 	 * space available yet.
 	 */
-	if (cil->xc_ctx->space_used < XLOG_CIL_SPACE_LIMIT(log))
+	if (cil->xc_ctx->space_used < XLOG_CIL_SPACE_LIMIT(log)) {
+		up_read(&cil->xc_ctx_lock);
 		return;
+	}
 
 	spin_lock(&cil->xc_push_lock);
 	if (cil->xc_push_seq < cil->xc_current_sequence) {
 		cil->xc_push_seq = cil->xc_current_sequence;
 		queue_work(log->l_mp->m_cil_workqueue, &cil->xc_push_work);
 	}
+
+	/*
+	 * Drop the context lock now, we can't hold that if we need to sleep
+	 * because we are over the blocking threshold. The push_lock is still
+	 * held, so blocking threshold sleep/wakeup is still correctly
+	 * serialised here.
+	 */
+	up_read(&cil->xc_ctx_lock);
+
+	/*
+	 * If we are well over the space limit, throttle the work that is being
+	 * done until the push work on this context has begun.
+	 */
+	if (cil->xc_ctx->space_used >= XLOG_CIL_BLOCKING_SPACE_LIMIT(log)) {
+		trace_xfs_log_cil_wait(log, cil->xc_ctx->ticket);
+		ASSERT(cil->xc_ctx->space_used < log->l_logsize);
+		xlog_wait(&cil->xc_push_wait, &cil->xc_push_lock);
+		return;
+	}
+
 	spin_unlock(&cil->xc_push_lock);
 
 }
@@ -1017,7 +1035,10 @@ xfs_log_commit_cil(
 	if (commit_lsn)
 		*commit_lsn = xc_commit_lsn;
 
-	xfs_log_done(mp, tp->t_ticket, NULL, regrant);
+	if (regrant && !XLOG_FORCED_SHUTDOWN(log))
+		xfs_log_ticket_regrant(log, tp->t_ticket);
+	else
+		xfs_log_ticket_ungrant(log, tp->t_ticket);
 	tp->t_ticket = NULL;
 	xfs_trans_unreserve_and_mod_sb(tp);
 
@@ -1038,9 +1059,9 @@ xfs_log_commit_cil(
 		if (lip->li_ops->iop_committing)
 			lip->li_ops->iop_committing(lip, xc_commit_lsn);
 	}
-	xlog_cil_push_background(log);
 
-	up_read(&cil->xc_ctx_lock);
+	/* xlog_cil_push_background() releases cil->xc_ctx_lock */
+	xlog_cil_push_background(log);
 }
 
 /*
@@ -1194,6 +1215,7 @@ xlog_cil_init(
 	INIT_LIST_HEAD(&cil->xc_committing);
 	spin_lock_init(&cil->xc_cil_lock);
 	spin_lock_init(&cil->xc_push_lock);
+	init_waitqueue_head(&cil->xc_push_wait);
 	init_rwsem(&cil->xc_ctx_lock);
 	init_waitqueue_head(&cil->xc_commit_wait);
 
diff --git a/fs/xfs/xfs_log_priv.h b/fs/xfs/xfs_log_priv.h
index b192c5a9f9fd..1c6fdbf3d506 100644
--- a/fs/xfs/xfs_log_priv.h
+++ b/fs/xfs/xfs_log_priv.h
@@ -51,13 +51,11 @@ enum xlog_iclog_state {
 };
 
 /*
- * Flags to log ticket
+ * Log ticket flags
  */
-#define XLOG_TIC_INITED		0x1	/* has been initialized */
-#define XLOG_TIC_PERM_RESERV	0x2	/* permanent reservation */
+#define XLOG_TIC_PERM_RESERV	0x1	/* permanent reservation */
 
 #define XLOG_TIC_FLAGS \
-	{ XLOG_TIC_INITED,	"XLOG_TIC_INITED" }, \
 	{ XLOG_TIC_PERM_RESERV,	"XLOG_TIC_PERM_RESERV" }
 
 /*
@@ -275,6 +273,7 @@ struct xfs_cil {
 	wait_queue_head_t	xc_commit_wait;
 	xfs_lsn_t		xc_current_sequence;
 	struct work_struct	xc_push_work;
+	wait_queue_head_t	xc_push_wait;	/* background push throttle */
 } ____cacheline_aligned_in_smp;
 
 /*
@@ -318,13 +317,53 @@ struct xfs_cil {
  * tries to keep 25% of the log free, so we need to keep below that limit or we
  * risk running out of free log space to start any new transactions.
  *
- * In order to keep background CIL push efficient, we will set a lower
- * threshold at which background pushing is attempted without blocking current
- * transaction commits.  A separate, higher bound defines when CIL pushes are
- * enforced to ensure we stay within our maximum checkpoint size bounds.
- * threshold, yet give us plenty of space for aggregation on large logs.
+ * In order to keep background CIL push efficient, we only need to ensure the
+ * CIL is large enough to maintain sufficient in-memory relogging to avoid
+ * repeated physical writes of frequently modified metadata. If we allow the CIL
+ * to grow to a substantial fraction of the log, then we may be pinning hundreds
+ * of megabytes of metadata in memory until the CIL flushes. This can cause
+ * issues when we are running low on memory - pinned memory cannot be reclaimed,
+ * and the CIL consumes a lot of memory. Hence we need to set an upper physical
+ * size limit for the CIL that limits the maximum amount of memory pinned by the
+ * CIL but does not limit performance by reducing relogging efficiency
+ * significantly.
+ *
+ * As such, the CIL push threshold ends up being the smaller of two thresholds:
+ * - a threshold large enough that it allows CIL to be pushed and progress to be
+ *   made without excessive blocking of incoming transaction commits. This is
+ *   defined to be 12.5% of the log space - half the 25% push threshold of the
+ *   AIL.
+ * - small enough that it doesn't pin excessive amounts of memory but maintains
+ *   close to peak relogging efficiency. This is defined to be 16x the iclog
+ *   buffer window (32MB) as measurements have shown this to be roughly the
+ *   point of diminishing performance increases under highly concurrent
+ *   modification workloads.
+ *
+ * To prevent the CIL from overflowing upper commit size bounds, we introduce a
+ * new threshold at which we block committing transactions until the background
+ * CIL commit commences and switches to a new context. While this is not a hard
+ * limit, it forces the process committing a transaction to the CIL to block and
+ * yeild the CPU, giving the CIL push work a chance to be scheduled and start
+ * work. This prevents a process running lots of transactions from overfilling
+ * the CIL because it is not yielding the CPU. We set the blocking limit at
+ * twice the background push space threshold so we keep in line with the AIL
+ * push thresholds.
+ *
+ * Note: this is not a -hard- limit as blocking is applied after the transaction
+ * is inserted into the CIL and the push has been triggered. It is largely a
+ * throttling mechanism that allows the CIL push to be scheduled and run. A hard
+ * limit will be difficult to implement without introducing global serialisation
+ * in the CIL commit fast path, and it's not at all clear that we actually need
+ * such hard limits given the ~7 years we've run without a hard limit before
+ * finding the first situation where a checkpoint size overflow actually
+ * occurred. Hence the simple throttle, and an ASSERT check to tell us that
+ * we've overrun the max size.
  */
-#define XLOG_CIL_SPACE_LIMIT(log)	(log->l_logsize >> 3)
+#define XLOG_CIL_SPACE_LIMIT(log)	\
+	min_t(int, (log)->l_logsize >> 3, BBTOB(XLOG_TOTAL_REC_SHIFT(log)) << 4)
+
+#define XLOG_CIL_BLOCKING_SPACE_LIMIT(log)	\
+	(XLOG_CIL_SPACE_LIMIT(log) * 2)
 
 /*
  * ticket grant locks, queues and accounting have their own cachlines
@@ -402,7 +441,8 @@ struct xlog {
 #define XLOG_BUF_CANCEL_BUCKET(log, blkno) \
 	((log)->l_buf_cancel_table + ((uint64_t)blkno % XLOG_BC_TABLE_SIZE))
 
-#define XLOG_FORCED_SHUTDOWN(log)	((log)->l_flags & XLOG_IO_ERROR)
+#define XLOG_FORCED_SHUTDOWN(log) \
+	(unlikely((log)->l_flags & XLOG_IO_ERROR))
 
 /* common routines */
 extern int
@@ -424,9 +464,7 @@ xlog_ticket_alloc(
 	int		unit_bytes,
 	int		count,
 	char		client,
-	bool		permanent,
-	xfs_km_flags_t	alloc_flags);
-
+	bool		permanent);
 
 static inline void
 xlog_write_adv_cnt(void **ptr, int *len, int *off, size_t bytes)
@@ -438,14 +476,14 @@ xlog_write_adv_cnt(void **ptr, int *len, int *off, size_t bytes)
 
 void	xlog_print_tic_res(struct xfs_mount *mp, struct xlog_ticket *ticket);
 void	xlog_print_trans(struct xfs_trans *);
-int
-xlog_write(
-	struct xlog		*log,
-	struct xfs_log_vec	*log_vector,
-	struct xlog_ticket	*tic,
-	xfs_lsn_t		*start_lsn,
-	struct xlog_in_core	**commit_iclog,
-	uint			flags);
+int	xlog_write(struct xlog *log, struct xfs_log_vec *log_vector,
+		struct xlog_ticket *tic, xfs_lsn_t *start_lsn,
+		struct xlog_in_core **commit_iclog, uint flags,
+		bool need_start_rec);
+int	xlog_commit_record(struct xlog *log, struct xlog_ticket *ticket,
+		struct xlog_in_core **iclog, xfs_lsn_t *lsn);
+void	xfs_log_ticket_ungrant(struct xlog *log, struct xlog_ticket *ticket);
+void	xfs_log_ticket_regrant(struct xlog *log, struct xlog_ticket *ticket);
 
 /*
  * When we crack an atomic LSN, we sample it first so that the value will not
@@ -525,12 +563,6 @@ xlog_cil_force(struct xlog *log)
 }
 
 /*
- * Unmount record type is used as a pseudo transaction type for the ticket.
- * It's value must be outside the range of XFS_TRANS_* values.
- */
-#define XLOG_UNMOUNT_REC_TYPE	(-1U)
-
-/*
  * Wrapper function for waiting on a wait queue serialised against wakeups
  * by a spinlock. This matches the semantics of all the wait queues used in the
  * log code.
diff --git a/fs/xfs/xfs_log_recover.c b/fs/xfs/xfs_log_recover.c
index 25cfc85dbaa7..e2ec91b2d0f4 100644
--- a/fs/xfs/xfs_log_recover.c
+++ b/fs/xfs/xfs_log_recover.c
@@ -18,21 +18,13 @@
 #include "xfs_log.h"
 #include "xfs_log_priv.h"
 #include "xfs_log_recover.h"
-#include "xfs_inode_item.h"
-#include "xfs_extfree_item.h"
 #include "xfs_trans_priv.h"
 #include "xfs_alloc.h"
 #include "xfs_ialloc.h"
-#include "xfs_quota.h"
 #include "xfs_trace.h"
 #include "xfs_icache.h"
-#include "xfs_bmap_btree.h"
 #include "xfs_error.h"
-#include "xfs_dir2.h"
-#include "xfs_rmap_item.h"
 #include "xfs_buf_item.h"
-#include "xfs_refcount_item.h"
-#include "xfs_bmap_item.h"
 
 #define BLK_AVG(blk1, blk2)	((blk1+blk2) >> 1)
 
@@ -56,17 +48,6 @@ xlog_do_recovery_pass(
         struct xlog *, xfs_daddr_t, xfs_daddr_t, int, xfs_daddr_t *);
 
 /*
- * This structure is used during recovery to record the buf log items which
- * have been canceled and should not be replayed.
- */
-struct xfs_buf_cancel {
-	xfs_daddr_t		bc_blkno;
-	uint			bc_len;
-	int			bc_refcount;
-	struct list_head	bc_list;
-};
-
-/*
  * Sector aligned buffer routines for buffer create/read/write/access
  */
 
@@ -284,7 +265,7 @@ xlog_header_check_mount(
 	return 0;
 }
 
-STATIC void
+void
 xlog_recover_iodone(
 	struct xfs_buf	*bp)
 {
@@ -306,9 +287,8 @@ xlog_recover_iodone(
 	if (bp->b_log_item)
 		xfs_buf_item_relse(bp);
 	ASSERT(bp->b_log_item == NULL);
-
-	bp->b_iodone = NULL;
-	xfs_buf_ioend(bp);
+	bp->b_flags &= ~_XBF_LOGRECOVERY;
+	xfs_buf_ioend_finish(bp);
 }
 
 /*
@@ -1120,7 +1100,7 @@ xlog_verify_head(
 		 *
 		 * Note that xlog_find_tail() clears the blocks at the new head
 		 * (i.e., the records with invalid CRC) if the cycle number
-		 * matches the the current cycle.
+		 * matches the current cycle.
 		 */
 		found = xlog_rseek_logrec_hdr(log, first_bad, *tail_blk, 1,
 				buffer, rhead_blk, rhead, wrapped);
@@ -1779,12 +1759,72 @@ xlog_clear_stale_blocks(
 	return 0;
 }
 
+/*
+ * Release the recovered intent item in the AIL that matches the given intent
+ * type and intent id.
+ */
+void
+xlog_recover_release_intent(
+	struct xlog		*log,
+	unsigned short		intent_type,
+	uint64_t		intent_id)
+{
+	struct xfs_ail_cursor	cur;
+	struct xfs_log_item	*lip;
+	struct xfs_ail		*ailp = log->l_ailp;
+
+	spin_lock(&ailp->ail_lock);
+	for (lip = xfs_trans_ail_cursor_first(ailp, &cur, 0); lip != NULL;
+	     lip = xfs_trans_ail_cursor_next(ailp, &cur)) {
+		if (lip->li_type != intent_type)
+			continue;
+		if (!lip->li_ops->iop_match(lip, intent_id))
+			continue;
+
+		spin_unlock(&ailp->ail_lock);
+		lip->li_ops->iop_release(lip);
+		spin_lock(&ailp->ail_lock);
+		break;
+	}
+
+	xfs_trans_ail_cursor_done(&cur);
+	spin_unlock(&ailp->ail_lock);
+}
+
 /******************************************************************************
  *
  *		Log recover routines
  *
  ******************************************************************************
  */
+static const struct xlog_recover_item_ops *xlog_recover_item_ops[] = {
+	&xlog_buf_item_ops,
+	&xlog_inode_item_ops,
+	&xlog_dquot_item_ops,
+	&xlog_quotaoff_item_ops,
+	&xlog_icreate_item_ops,
+	&xlog_efi_item_ops,
+	&xlog_efd_item_ops,
+	&xlog_rui_item_ops,
+	&xlog_rud_item_ops,
+	&xlog_cui_item_ops,
+	&xlog_cud_item_ops,
+	&xlog_bui_item_ops,
+	&xlog_bud_item_ops,
+};
+
+static const struct xlog_recover_item_ops *
+xlog_find_item_ops(
+	struct xlog_recover_item		*item)
+{
+	unsigned int				i;
+
+	for (i = 0; i < ARRAY_SIZE(xlog_recover_item_ops); i++)
+		if (ITEM_TYPE(item) == xlog_recover_item_ops[i]->item_type)
+			return xlog_recover_item_ops[i];
+
+	return NULL;
+}
 
 /*
  * Sort the log items in the transaction.
@@ -1841,54 +1881,23 @@ xlog_recover_reorder_trans(
 	struct xlog_recover	*trans,
 	int			pass)
 {
-	xlog_recover_item_t	*item, *n;
+	struct xlog_recover_item *item, *n;
 	int			error = 0;
 	LIST_HEAD(sort_list);
 	LIST_HEAD(cancel_list);
 	LIST_HEAD(buffer_list);
 	LIST_HEAD(inode_buffer_list);
-	LIST_HEAD(inode_list);
+	LIST_HEAD(item_list);
 
 	list_splice_init(&trans->r_itemq, &sort_list);
 	list_for_each_entry_safe(item, n, &sort_list, ri_list) {
-		xfs_buf_log_format_t	*buf_f = item->ri_buf[0].i_addr;
+		enum xlog_recover_reorder	fate = XLOG_REORDER_ITEM_LIST;
 
-		switch (ITEM_TYPE(item)) {
-		case XFS_LI_ICREATE:
-			list_move_tail(&item->ri_list, &buffer_list);
-			break;
-		case XFS_LI_BUF:
-			if (buf_f->blf_flags & XFS_BLF_CANCEL) {
-				trace_xfs_log_recover_item_reorder_head(log,
-							trans, item, pass);
-				list_move(&item->ri_list, &cancel_list);
-				break;
-			}
-			if (buf_f->blf_flags & XFS_BLF_INODE_BUF) {
-				list_move(&item->ri_list, &inode_buffer_list);
-				break;
-			}
-			list_move_tail(&item->ri_list, &buffer_list);
-			break;
-		case XFS_LI_INODE:
-		case XFS_LI_DQUOT:
-		case XFS_LI_QUOTAOFF:
-		case XFS_LI_EFD:
-		case XFS_LI_EFI:
-		case XFS_LI_RUI:
-		case XFS_LI_RUD:
-		case XFS_LI_CUI:
-		case XFS_LI_CUD:
-		case XFS_LI_BUI:
-		case XFS_LI_BUD:
-			trace_xfs_log_recover_item_reorder_tail(log,
-							trans, item, pass);
-			list_move_tail(&item->ri_list, &inode_list);
-			break;
-		default:
+		item->ri_ops = xlog_find_item_ops(item);
+		if (!item->ri_ops) {
 			xfs_warn(log->l_mp,
-				"%s: unrecognized type of log operation",
-				__func__);
+				"%s: unrecognized type of log operation (%d)",
+				__func__, ITEM_TYPE(item));
 			ASSERT(0);
 			/*
 			 * return the remaining items back to the transaction
@@ -1896,16 +1905,38 @@ xlog_recover_reorder_trans(
 			 */
 			if (!list_empty(&sort_list))
 				list_splice_init(&sort_list, &trans->r_itemq);
-			error = -EIO;
-			goto out;
+			error = -EFSCORRUPTED;
+			break;
+		}
+
+		if (item->ri_ops->reorder)
+			fate = item->ri_ops->reorder(item);
+
+		switch (fate) {
+		case XLOG_REORDER_BUFFER_LIST:
+			list_move_tail(&item->ri_list, &buffer_list);
+			break;
+		case XLOG_REORDER_CANCEL_LIST:
+			trace_xfs_log_recover_item_reorder_head(log,
+					trans, item, pass);
+			list_move(&item->ri_list, &cancel_list);
+			break;
+		case XLOG_REORDER_INODE_BUFFER_LIST:
+			list_move(&item->ri_list, &inode_buffer_list);
+			break;
+		case XLOG_REORDER_ITEM_LIST:
+			trace_xfs_log_recover_item_reorder_tail(log,
+							trans, item, pass);
+			list_move_tail(&item->ri_list, &item_list);
+			break;
 		}
 	}
-out:
+
 	ASSERT(list_empty(&sort_list));
 	if (!list_empty(&buffer_list))
 		list_splice(&buffer_list, &trans->r_itemq);
-	if (!list_empty(&inode_list))
-		list_splice_tail(&inode_list, &trans->r_itemq);
+	if (!list_empty(&item_list))
+		list_splice_tail(&item_list, &trans->r_itemq);
 	if (!list_empty(&inode_buffer_list))
 		list_splice_tail(&inode_buffer_list, &trans->r_itemq);
 	if (!list_empty(&cancel_list))
@@ -1913,2152 +1944,15 @@ out:
 	return error;
 }
 
-/*
- * Build up the table of buf cancel records so that we don't replay
- * cancelled data in the second pass.  For buffer records that are
- * not cancel records, there is nothing to do here so we just return.
- *
- * If we get a cancel record which is already in the table, this indicates
- * that the buffer was cancelled multiple times.  In order to ensure
- * that during pass 2 we keep the record in the table until we reach its
- * last occurrence in the log, we keep a reference count in the cancel
- * record in the table to tell us how many times we expect to see this
- * record during the second pass.
- */
-STATIC int
-xlog_recover_buffer_pass1(
-	struct xlog			*log,
-	struct xlog_recover_item	*item)
-{
-	xfs_buf_log_format_t	*buf_f = item->ri_buf[0].i_addr;
-	struct list_head	*bucket;
-	struct xfs_buf_cancel	*bcp;
-
-	if (!xfs_buf_log_check_iovec(&item->ri_buf[0])) {
-		xfs_err(log->l_mp, "bad buffer log item size (%d)",
-				item->ri_buf[0].i_len);
-		return -EFSCORRUPTED;
-	}
-
-	/*
-	 * If this isn't a cancel buffer item, then just return.
-	 */
-	if (!(buf_f->blf_flags & XFS_BLF_CANCEL)) {
-		trace_xfs_log_recover_buf_not_cancel(log, buf_f);
-		return 0;
-	}
-
-	/*
-	 * Insert an xfs_buf_cancel record into the hash table of them.
-	 * If there is already an identical record, bump its reference count.
-	 */
-	bucket = XLOG_BUF_CANCEL_BUCKET(log, buf_f->blf_blkno);
-	list_for_each_entry(bcp, bucket, bc_list) {
-		if (bcp->bc_blkno == buf_f->blf_blkno &&
-		    bcp->bc_len == buf_f->blf_len) {
-			bcp->bc_refcount++;
-			trace_xfs_log_recover_buf_cancel_ref_inc(log, buf_f);
-			return 0;
-		}
-	}
-
-	bcp = kmem_alloc(sizeof(struct xfs_buf_cancel), 0);
-	bcp->bc_blkno = buf_f->blf_blkno;
-	bcp->bc_len = buf_f->blf_len;
-	bcp->bc_refcount = 1;
-	list_add_tail(&bcp->bc_list, bucket);
-
-	trace_xfs_log_recover_buf_cancel_add(log, buf_f);
-	return 0;
-}
-
-/*
- * Check to see whether the buffer being recovered has a corresponding
- * entry in the buffer cancel record table. If it is, return the cancel
- * buffer structure to the caller.
- */
-STATIC struct xfs_buf_cancel *
-xlog_peek_buffer_cancelled(
+void
+xlog_buf_readahead(
 	struct xlog		*log,
 	xfs_daddr_t		blkno,
 	uint			len,
-	unsigned short			flags)
+	const struct xfs_buf_ops *ops)
 {
-	struct list_head	*bucket;
-	struct xfs_buf_cancel	*bcp;
-
-	if (!log->l_buf_cancel_table) {
-		/* empty table means no cancelled buffers in the log */
-		ASSERT(!(flags & XFS_BLF_CANCEL));
-		return NULL;
-	}
-
-	bucket = XLOG_BUF_CANCEL_BUCKET(log, blkno);
-	list_for_each_entry(bcp, bucket, bc_list) {
-		if (bcp->bc_blkno == blkno && bcp->bc_len == len)
-			return bcp;
-	}
-
-	/*
-	 * We didn't find a corresponding entry in the table, so return 0 so
-	 * that the buffer is NOT cancelled.
-	 */
-	ASSERT(!(flags & XFS_BLF_CANCEL));
-	return NULL;
-}
-
-/*
- * If the buffer is being cancelled then return 1 so that it will be cancelled,
- * otherwise return 0.  If the buffer is actually a buffer cancel item
- * (XFS_BLF_CANCEL is set), then decrement the refcount on the entry in the
- * table and remove it from the table if this is the last reference.
- *
- * We remove the cancel record from the table when we encounter its last
- * occurrence in the log so that if the same buffer is re-used again after its
- * last cancellation we actually replay the changes made at that point.
- */
-STATIC int
-xlog_check_buffer_cancelled(
-	struct xlog		*log,
-	xfs_daddr_t		blkno,
-	uint			len,
-	unsigned short			flags)
-{
-	struct xfs_buf_cancel	*bcp;
-
-	bcp = xlog_peek_buffer_cancelled(log, blkno, len, flags);
-	if (!bcp)
-		return 0;
-
-	/*
-	 * We've go a match, so return 1 so that the recovery of this buffer
-	 * is cancelled.  If this buffer is actually a buffer cancel log
-	 * item, then decrement the refcount on the one in the table and
-	 * remove it if this is the last reference.
-	 */
-	if (flags & XFS_BLF_CANCEL) {
-		if (--bcp->bc_refcount == 0) {
-			list_del(&bcp->bc_list);
-			kmem_free(bcp);
-		}
-	}
-	return 1;
-}
-
-/*
- * Perform recovery for a buffer full of inodes.  In these buffers, the only
- * data which should be recovered is that which corresponds to the
- * di_next_unlinked pointers in the on disk inode structures.  The rest of the
- * data for the inodes is always logged through the inodes themselves rather
- * than the inode buffer and is recovered in xlog_recover_inode_pass2().
- *
- * The only time when buffers full of inodes are fully recovered is when the
- * buffer is full of newly allocated inodes.  In this case the buffer will
- * not be marked as an inode buffer and so will be sent to
- * xlog_recover_do_reg_buffer() below during recovery.
- */
-STATIC int
-xlog_recover_do_inode_buffer(
-	struct xfs_mount	*mp,
-	xlog_recover_item_t	*item,
-	struct xfs_buf		*bp,
-	xfs_buf_log_format_t	*buf_f)
-{
-	int			i;
-	int			item_index = 0;
-	int			bit = 0;
-	int			nbits = 0;
-	int			reg_buf_offset = 0;
-	int			reg_buf_bytes = 0;
-	int			next_unlinked_offset;
-	int			inodes_per_buf;
-	xfs_agino_t		*logged_nextp;
-	xfs_agino_t		*buffer_nextp;
-
-	trace_xfs_log_recover_buf_inode_buf(mp->m_log, buf_f);
-
-	/*
-	 * Post recovery validation only works properly on CRC enabled
-	 * filesystems.
-	 */
-	if (xfs_sb_version_hascrc(&mp->m_sb))
-		bp->b_ops = &xfs_inode_buf_ops;
-
-	inodes_per_buf = BBTOB(bp->b_length) >> mp->m_sb.sb_inodelog;
-	for (i = 0; i < inodes_per_buf; i++) {
-		next_unlinked_offset = (i * mp->m_sb.sb_inodesize) +
-			offsetof(xfs_dinode_t, di_next_unlinked);
-
-		while (next_unlinked_offset >=
-		       (reg_buf_offset + reg_buf_bytes)) {
-			/*
-			 * The next di_next_unlinked field is beyond
-			 * the current logged region.  Find the next
-			 * logged region that contains or is beyond
-			 * the current di_next_unlinked field.
-			 */
-			bit += nbits;
-			bit = xfs_next_bit(buf_f->blf_data_map,
-					   buf_f->blf_map_size, bit);
-
-			/*
-			 * If there are no more logged regions in the
-			 * buffer, then we're done.
-			 */
-			if (bit == -1)
-				return 0;
-
-			nbits = xfs_contig_bits(buf_f->blf_data_map,
-						buf_f->blf_map_size, bit);
-			ASSERT(nbits > 0);
-			reg_buf_offset = bit << XFS_BLF_SHIFT;
-			reg_buf_bytes = nbits << XFS_BLF_SHIFT;
-			item_index++;
-		}
-
-		/*
-		 * If the current logged region starts after the current
-		 * di_next_unlinked field, then move on to the next
-		 * di_next_unlinked field.
-		 */
-		if (next_unlinked_offset < reg_buf_offset)
-			continue;
-
-		ASSERT(item->ri_buf[item_index].i_addr != NULL);
-		ASSERT((item->ri_buf[item_index].i_len % XFS_BLF_CHUNK) == 0);
-		ASSERT((reg_buf_offset + reg_buf_bytes) <= BBTOB(bp->b_length));
-
-		/*
-		 * The current logged region contains a copy of the
-		 * current di_next_unlinked field.  Extract its value
-		 * and copy it to the buffer copy.
-		 */
-		logged_nextp = item->ri_buf[item_index].i_addr +
-				next_unlinked_offset - reg_buf_offset;
-		if (XFS_IS_CORRUPT(mp, *logged_nextp == 0)) {
-			xfs_alert(mp,
-		"Bad inode buffer log record (ptr = "PTR_FMT", bp = "PTR_FMT"). "
-		"Trying to replay bad (0) inode di_next_unlinked field.",
-				item, bp);
-			return -EFSCORRUPTED;
-		}
-
-		buffer_nextp = xfs_buf_offset(bp, next_unlinked_offset);
-		*buffer_nextp = *logged_nextp;
-
-		/*
-		 * If necessary, recalculate the CRC in the on-disk inode. We
-		 * have to leave the inode in a consistent state for whoever
-		 * reads it next....
-		 */
-		xfs_dinode_calc_crc(mp,
-				xfs_buf_offset(bp, i * mp->m_sb.sb_inodesize));
-
-	}
-
-	return 0;
-}
-
-/*
- * V5 filesystems know the age of the buffer on disk being recovered. We can
- * have newer objects on disk than we are replaying, and so for these cases we
- * don't want to replay the current change as that will make the buffer contents
- * temporarily invalid on disk.
- *
- * The magic number might not match the buffer type we are going to recover
- * (e.g. reallocated blocks), so we ignore the xfs_buf_log_format flags.  Hence
- * extract the LSN of the existing object in the buffer based on it's current
- * magic number.  If we don't recognise the magic number in the buffer, then
- * return a LSN of -1 so that the caller knows it was an unrecognised block and
- * so can recover the buffer.
- *
- * Note: we cannot rely solely on magic number matches to determine that the
- * buffer has a valid LSN - we also need to verify that it belongs to this
- * filesystem, so we need to extract the object's LSN and compare it to that
- * which we read from the superblock. If the UUIDs don't match, then we've got a
- * stale metadata block from an old filesystem instance that we need to recover
- * over the top of.
- */
-static xfs_lsn_t
-xlog_recover_get_buf_lsn(
-	struct xfs_mount	*mp,
-	struct xfs_buf		*bp)
-{
-	uint32_t		magic32;
-	uint16_t		magic16;
-	uint16_t		magicda;
-	void			*blk = bp->b_addr;
-	uuid_t			*uuid;
-	xfs_lsn_t		lsn = -1;
-
-	/* v4 filesystems always recover immediately */
-	if (!xfs_sb_version_hascrc(&mp->m_sb))
-		goto recover_immediately;
-
-	magic32 = be32_to_cpu(*(__be32 *)blk);
-	switch (magic32) {
-	case XFS_ABTB_CRC_MAGIC:
-	case XFS_ABTC_CRC_MAGIC:
-	case XFS_ABTB_MAGIC:
-	case XFS_ABTC_MAGIC:
-	case XFS_RMAP_CRC_MAGIC:
-	case XFS_REFC_CRC_MAGIC:
-	case XFS_IBT_CRC_MAGIC:
-	case XFS_IBT_MAGIC: {
-		struct xfs_btree_block *btb = blk;
-
-		lsn = be64_to_cpu(btb->bb_u.s.bb_lsn);
-		uuid = &btb->bb_u.s.bb_uuid;
-		break;
-	}
-	case XFS_BMAP_CRC_MAGIC:
-	case XFS_BMAP_MAGIC: {
-		struct xfs_btree_block *btb = blk;
-
-		lsn = be64_to_cpu(btb->bb_u.l.bb_lsn);
-		uuid = &btb->bb_u.l.bb_uuid;
-		break;
-	}
-	case XFS_AGF_MAGIC:
-		lsn = be64_to_cpu(((struct xfs_agf *)blk)->agf_lsn);
-		uuid = &((struct xfs_agf *)blk)->agf_uuid;
-		break;
-	case XFS_AGFL_MAGIC:
-		lsn = be64_to_cpu(((struct xfs_agfl *)blk)->agfl_lsn);
-		uuid = &((struct xfs_agfl *)blk)->agfl_uuid;
-		break;
-	case XFS_AGI_MAGIC:
-		lsn = be64_to_cpu(((struct xfs_agi *)blk)->agi_lsn);
-		uuid = &((struct xfs_agi *)blk)->agi_uuid;
-		break;
-	case XFS_SYMLINK_MAGIC:
-		lsn = be64_to_cpu(((struct xfs_dsymlink_hdr *)blk)->sl_lsn);
-		uuid = &((struct xfs_dsymlink_hdr *)blk)->sl_uuid;
-		break;
-	case XFS_DIR3_BLOCK_MAGIC:
-	case XFS_DIR3_DATA_MAGIC:
-	case XFS_DIR3_FREE_MAGIC:
-		lsn = be64_to_cpu(((struct xfs_dir3_blk_hdr *)blk)->lsn);
-		uuid = &((struct xfs_dir3_blk_hdr *)blk)->uuid;
-		break;
-	case XFS_ATTR3_RMT_MAGIC:
-		/*
-		 * Remote attr blocks are written synchronously, rather than
-		 * being logged. That means they do not contain a valid LSN
-		 * (i.e. transactionally ordered) in them, and hence any time we
-		 * see a buffer to replay over the top of a remote attribute
-		 * block we should simply do so.
-		 */
-		goto recover_immediately;
-	case XFS_SB_MAGIC:
-		/*
-		 * superblock uuids are magic. We may or may not have a
-		 * sb_meta_uuid on disk, but it will be set in the in-core
-		 * superblock. We set the uuid pointer for verification
-		 * according to the superblock feature mask to ensure we check
-		 * the relevant UUID in the superblock.
-		 */
-		lsn = be64_to_cpu(((struct xfs_dsb *)blk)->sb_lsn);
-		if (xfs_sb_version_hasmetauuid(&mp->m_sb))
-			uuid = &((struct xfs_dsb *)blk)->sb_meta_uuid;
-		else
-			uuid = &((struct xfs_dsb *)blk)->sb_uuid;
-		break;
-	default:
-		break;
-	}
-
-	if (lsn != (xfs_lsn_t)-1) {
-		if (!uuid_equal(&mp->m_sb.sb_meta_uuid, uuid))
-			goto recover_immediately;
-		return lsn;
-	}
-
-	magicda = be16_to_cpu(((struct xfs_da_blkinfo *)blk)->magic);
-	switch (magicda) {
-	case XFS_DIR3_LEAF1_MAGIC:
-	case XFS_DIR3_LEAFN_MAGIC:
-	case XFS_DA3_NODE_MAGIC:
-		lsn = be64_to_cpu(((struct xfs_da3_blkinfo *)blk)->lsn);
-		uuid = &((struct xfs_da3_blkinfo *)blk)->uuid;
-		break;
-	default:
-		break;
-	}
-
-	if (lsn != (xfs_lsn_t)-1) {
-		if (!uuid_equal(&mp->m_sb.sb_uuid, uuid))
-			goto recover_immediately;
-		return lsn;
-	}
-
-	/*
-	 * We do individual object checks on dquot and inode buffers as they
-	 * have their own individual LSN records. Also, we could have a stale
-	 * buffer here, so we have to at least recognise these buffer types.
-	 *
-	 * A notd complexity here is inode unlinked list processing - it logs
-	 * the inode directly in the buffer, but we don't know which inodes have
-	 * been modified, and there is no global buffer LSN. Hence we need to
-	 * recover all inode buffer types immediately. This problem will be
-	 * fixed by logical logging of the unlinked list modifications.
-	 */
-	magic16 = be16_to_cpu(*(__be16 *)blk);
-	switch (magic16) {
-	case XFS_DQUOT_MAGIC:
-	case XFS_DINODE_MAGIC:
-		goto recover_immediately;
-	default:
-		break;
-	}
-
-	/* unknown buffer contents, recover immediately */
-
-recover_immediately:
-	return (xfs_lsn_t)-1;
-
-}
-
-/*
- * Validate the recovered buffer is of the correct type and attach the
- * appropriate buffer operations to them for writeback. Magic numbers are in a
- * few places:
- *	the first 16 bits of the buffer (inode buffer, dquot buffer),
- *	the first 32 bits of the buffer (most blocks),
- *	inside a struct xfs_da_blkinfo at the start of the buffer.
- */
-static void
-xlog_recover_validate_buf_type(
-	struct xfs_mount	*mp,
-	struct xfs_buf		*bp,
-	xfs_buf_log_format_t	*buf_f,
-	xfs_lsn_t		current_lsn)
-{
-	struct xfs_da_blkinfo	*info = bp->b_addr;
-	uint32_t		magic32;
-	uint16_t		magic16;
-	uint16_t		magicda;
-	char			*warnmsg = NULL;
-
-	/*
-	 * We can only do post recovery validation on items on CRC enabled
-	 * fielsystems as we need to know when the buffer was written to be able
-	 * to determine if we should have replayed the item. If we replay old
-	 * metadata over a newer buffer, then it will enter a temporarily
-	 * inconsistent state resulting in verification failures. Hence for now
-	 * just avoid the verification stage for non-crc filesystems
-	 */
-	if (!xfs_sb_version_hascrc(&mp->m_sb))
-		return;
-
-	magic32 = be32_to_cpu(*(__be32 *)bp->b_addr);
-	magic16 = be16_to_cpu(*(__be16*)bp->b_addr);
-	magicda = be16_to_cpu(info->magic);
-	switch (xfs_blft_from_flags(buf_f)) {
-	case XFS_BLFT_BTREE_BUF:
-		switch (magic32) {
-		case XFS_ABTB_CRC_MAGIC:
-		case XFS_ABTB_MAGIC:
-			bp->b_ops = &xfs_bnobt_buf_ops;
-			break;
-		case XFS_ABTC_CRC_MAGIC:
-		case XFS_ABTC_MAGIC:
-			bp->b_ops = &xfs_cntbt_buf_ops;
-			break;
-		case XFS_IBT_CRC_MAGIC:
-		case XFS_IBT_MAGIC:
-			bp->b_ops = &xfs_inobt_buf_ops;
-			break;
-		case XFS_FIBT_CRC_MAGIC:
-		case XFS_FIBT_MAGIC:
-			bp->b_ops = &xfs_finobt_buf_ops;
-			break;
-		case XFS_BMAP_CRC_MAGIC:
-		case XFS_BMAP_MAGIC:
-			bp->b_ops = &xfs_bmbt_buf_ops;
-			break;
-		case XFS_RMAP_CRC_MAGIC:
-			bp->b_ops = &xfs_rmapbt_buf_ops;
-			break;
-		case XFS_REFC_CRC_MAGIC:
-			bp->b_ops = &xfs_refcountbt_buf_ops;
-			break;
-		default:
-			warnmsg = "Bad btree block magic!";
-			break;
-		}
-		break;
-	case XFS_BLFT_AGF_BUF:
-		if (magic32 != XFS_AGF_MAGIC) {
-			warnmsg = "Bad AGF block magic!";
-			break;
-		}
-		bp->b_ops = &xfs_agf_buf_ops;
-		break;
-	case XFS_BLFT_AGFL_BUF:
-		if (magic32 != XFS_AGFL_MAGIC) {
-			warnmsg = "Bad AGFL block magic!";
-			break;
-		}
-		bp->b_ops = &xfs_agfl_buf_ops;
-		break;
-	case XFS_BLFT_AGI_BUF:
-		if (magic32 != XFS_AGI_MAGIC) {
-			warnmsg = "Bad AGI block magic!";
-			break;
-		}
-		bp->b_ops = &xfs_agi_buf_ops;
-		break;
-	case XFS_BLFT_UDQUOT_BUF:
-	case XFS_BLFT_PDQUOT_BUF:
-	case XFS_BLFT_GDQUOT_BUF:
-#ifdef CONFIG_XFS_QUOTA
-		if (magic16 != XFS_DQUOT_MAGIC) {
-			warnmsg = "Bad DQUOT block magic!";
-			break;
-		}
-		bp->b_ops = &xfs_dquot_buf_ops;
-#else
-		xfs_alert(mp,
-	"Trying to recover dquots without QUOTA support built in!");
-		ASSERT(0);
-#endif
-		break;
-	case XFS_BLFT_DINO_BUF:
-		if (magic16 != XFS_DINODE_MAGIC) {
-			warnmsg = "Bad INODE block magic!";
-			break;
-		}
-		bp->b_ops = &xfs_inode_buf_ops;
-		break;
-	case XFS_BLFT_SYMLINK_BUF:
-		if (magic32 != XFS_SYMLINK_MAGIC) {
-			warnmsg = "Bad symlink block magic!";
-			break;
-		}
-		bp->b_ops = &xfs_symlink_buf_ops;
-		break;
-	case XFS_BLFT_DIR_BLOCK_BUF:
-		if (magic32 != XFS_DIR2_BLOCK_MAGIC &&
-		    magic32 != XFS_DIR3_BLOCK_MAGIC) {
-			warnmsg = "Bad dir block magic!";
-			break;
-		}
-		bp->b_ops = &xfs_dir3_block_buf_ops;
-		break;
-	case XFS_BLFT_DIR_DATA_BUF:
-		if (magic32 != XFS_DIR2_DATA_MAGIC &&
-		    magic32 != XFS_DIR3_DATA_MAGIC) {
-			warnmsg = "Bad dir data magic!";
-			break;
-		}
-		bp->b_ops = &xfs_dir3_data_buf_ops;
-		break;
-	case XFS_BLFT_DIR_FREE_BUF:
-		if (magic32 != XFS_DIR2_FREE_MAGIC &&
-		    magic32 != XFS_DIR3_FREE_MAGIC) {
-			warnmsg = "Bad dir3 free magic!";
-			break;
-		}
-		bp->b_ops = &xfs_dir3_free_buf_ops;
-		break;
-	case XFS_BLFT_DIR_LEAF1_BUF:
-		if (magicda != XFS_DIR2_LEAF1_MAGIC &&
-		    magicda != XFS_DIR3_LEAF1_MAGIC) {
-			warnmsg = "Bad dir leaf1 magic!";
-			break;
-		}
-		bp->b_ops = &xfs_dir3_leaf1_buf_ops;
-		break;
-	case XFS_BLFT_DIR_LEAFN_BUF:
-		if (magicda != XFS_DIR2_LEAFN_MAGIC &&
-		    magicda != XFS_DIR3_LEAFN_MAGIC) {
-			warnmsg = "Bad dir leafn magic!";
-			break;
-		}
-		bp->b_ops = &xfs_dir3_leafn_buf_ops;
-		break;
-	case XFS_BLFT_DA_NODE_BUF:
-		if (magicda != XFS_DA_NODE_MAGIC &&
-		    magicda != XFS_DA3_NODE_MAGIC) {
-			warnmsg = "Bad da node magic!";
-			break;
-		}
-		bp->b_ops = &xfs_da3_node_buf_ops;
-		break;
-	case XFS_BLFT_ATTR_LEAF_BUF:
-		if (magicda != XFS_ATTR_LEAF_MAGIC &&
-		    magicda != XFS_ATTR3_LEAF_MAGIC) {
-			warnmsg = "Bad attr leaf magic!";
-			break;
-		}
-		bp->b_ops = &xfs_attr3_leaf_buf_ops;
-		break;
-	case XFS_BLFT_ATTR_RMT_BUF:
-		if (magic32 != XFS_ATTR3_RMT_MAGIC) {
-			warnmsg = "Bad attr remote magic!";
-			break;
-		}
-		bp->b_ops = &xfs_attr3_rmt_buf_ops;
-		break;
-	case XFS_BLFT_SB_BUF:
-		if (magic32 != XFS_SB_MAGIC) {
-			warnmsg = "Bad SB block magic!";
-			break;
-		}
-		bp->b_ops = &xfs_sb_buf_ops;
-		break;
-#ifdef CONFIG_XFS_RT
-	case XFS_BLFT_RTBITMAP_BUF:
-	case XFS_BLFT_RTSUMMARY_BUF:
-		/* no magic numbers for verification of RT buffers */
-		bp->b_ops = &xfs_rtbuf_ops;
-		break;
-#endif /* CONFIG_XFS_RT */
-	default:
-		xfs_warn(mp, "Unknown buffer type %d!",
-			 xfs_blft_from_flags(buf_f));
-		break;
-	}
-
-	/*
-	 * Nothing else to do in the case of a NULL current LSN as this means
-	 * the buffer is more recent than the change in the log and will be
-	 * skipped.
-	 */
-	if (current_lsn == NULLCOMMITLSN)
-		return;
-
-	if (warnmsg) {
-		xfs_warn(mp, warnmsg);
-		ASSERT(0);
-	}
-
-	/*
-	 * We must update the metadata LSN of the buffer as it is written out to
-	 * ensure that older transactions never replay over this one and corrupt
-	 * the buffer. This can occur if log recovery is interrupted at some
-	 * point after the current transaction completes, at which point a
-	 * subsequent mount starts recovery from the beginning.
-	 *
-	 * Write verifiers update the metadata LSN from log items attached to
-	 * the buffer. Therefore, initialize a bli purely to carry the LSN to
-	 * the verifier. We'll clean it up in our ->iodone() callback.
-	 */
-	if (bp->b_ops) {
-		struct xfs_buf_log_item	*bip;
-
-		ASSERT(!bp->b_iodone || bp->b_iodone == xlog_recover_iodone);
-		bp->b_iodone = xlog_recover_iodone;
-		xfs_buf_item_init(bp, mp);
-		bip = bp->b_log_item;
-		bip->bli_item.li_lsn = current_lsn;
-	}
-}
-
-/*
- * Perform a 'normal' buffer recovery.  Each logged region of the
- * buffer should be copied over the corresponding region in the
- * given buffer.  The bitmap in the buf log format structure indicates
- * where to place the logged data.
- */
-STATIC void
-xlog_recover_do_reg_buffer(
-	struct xfs_mount	*mp,
-	xlog_recover_item_t	*item,
-	struct xfs_buf		*bp,
-	xfs_buf_log_format_t	*buf_f,
-	xfs_lsn_t		current_lsn)
-{
-	int			i;
-	int			bit;
-	int			nbits;
-	xfs_failaddr_t		fa;
-	const size_t		size_disk_dquot = sizeof(struct xfs_disk_dquot);
-
-	trace_xfs_log_recover_buf_reg_buf(mp->m_log, buf_f);
-
-	bit = 0;
-	i = 1;  /* 0 is the buf format structure */
-	while (1) {
-		bit = xfs_next_bit(buf_f->blf_data_map,
-				   buf_f->blf_map_size, bit);
-		if (bit == -1)
-			break;
-		nbits = xfs_contig_bits(buf_f->blf_data_map,
-					buf_f->blf_map_size, bit);
-		ASSERT(nbits > 0);
-		ASSERT(item->ri_buf[i].i_addr != NULL);
-		ASSERT(item->ri_buf[i].i_len % XFS_BLF_CHUNK == 0);
-		ASSERT(BBTOB(bp->b_length) >=
-		       ((uint)bit << XFS_BLF_SHIFT) + (nbits << XFS_BLF_SHIFT));
-
-		/*
-		 * The dirty regions logged in the buffer, even though
-		 * contiguous, may span multiple chunks. This is because the
-		 * dirty region may span a physical page boundary in a buffer
-		 * and hence be split into two separate vectors for writing into
-		 * the log. Hence we need to trim nbits back to the length of
-		 * the current region being copied out of the log.
-		 */
-		if (item->ri_buf[i].i_len < (nbits << XFS_BLF_SHIFT))
-			nbits = item->ri_buf[i].i_len >> XFS_BLF_SHIFT;
-
-		/*
-		 * Do a sanity check if this is a dquot buffer. Just checking
-		 * the first dquot in the buffer should do. XXXThis is
-		 * probably a good thing to do for other buf types also.
-		 */
-		fa = NULL;
-		if (buf_f->blf_flags &
-		   (XFS_BLF_UDQUOT_BUF|XFS_BLF_PDQUOT_BUF|XFS_BLF_GDQUOT_BUF)) {
-			if (item->ri_buf[i].i_addr == NULL) {
-				xfs_alert(mp,
-					"XFS: NULL dquot in %s.", __func__);
-				goto next;
-			}
-			if (item->ri_buf[i].i_len < size_disk_dquot) {
-				xfs_alert(mp,
-					"XFS: dquot too small (%d) in %s.",
-					item->ri_buf[i].i_len, __func__);
-				goto next;
-			}
-			fa = xfs_dquot_verify(mp, item->ri_buf[i].i_addr,
-					       -1, 0);
-			if (fa) {
-				xfs_alert(mp,
-	"dquot corrupt at %pS trying to replay into block 0x%llx",
-					fa, bp->b_bn);
-				goto next;
-			}
-		}
-
-		memcpy(xfs_buf_offset(bp,
-			(uint)bit << XFS_BLF_SHIFT),	/* dest */
-			item->ri_buf[i].i_addr,		/* source */
-			nbits<<XFS_BLF_SHIFT);		/* length */
- next:
-		i++;
-		bit += nbits;
-	}
-
-	/* Shouldn't be any more regions */
-	ASSERT(i == item->ri_total);
-
-	xlog_recover_validate_buf_type(mp, bp, buf_f, current_lsn);
-}
-
-/*
- * Perform a dquot buffer recovery.
- * Simple algorithm: if we have found a QUOTAOFF log item of the same type
- * (ie. USR or GRP), then just toss this buffer away; don't recover it.
- * Else, treat it as a regular buffer and do recovery.
- *
- * Return false if the buffer was tossed and true if we recovered the buffer to
- * indicate to the caller if the buffer needs writing.
- */
-STATIC bool
-xlog_recover_do_dquot_buffer(
-	struct xfs_mount		*mp,
-	struct xlog			*log,
-	struct xlog_recover_item	*item,
-	struct xfs_buf			*bp,
-	struct xfs_buf_log_format	*buf_f)
-{
-	uint			type;
-
-	trace_xfs_log_recover_buf_dquot_buf(log, buf_f);
-
-	/*
-	 * Filesystems are required to send in quota flags at mount time.
-	 */
-	if (!mp->m_qflags)
-		return false;
-
-	type = 0;
-	if (buf_f->blf_flags & XFS_BLF_UDQUOT_BUF)
-		type |= XFS_DQ_USER;
-	if (buf_f->blf_flags & XFS_BLF_PDQUOT_BUF)
-		type |= XFS_DQ_PROJ;
-	if (buf_f->blf_flags & XFS_BLF_GDQUOT_BUF)
-		type |= XFS_DQ_GROUP;
-	/*
-	 * This type of quotas was turned off, so ignore this buffer
-	 */
-	if (log->l_quotaoffs_flag & type)
-		return false;
-
-	xlog_recover_do_reg_buffer(mp, item, bp, buf_f, NULLCOMMITLSN);
-	return true;
-}
-
-/*
- * This routine replays a modification made to a buffer at runtime.
- * There are actually two types of buffer, regular and inode, which
- * are handled differently.  Inode buffers are handled differently
- * in that we only recover a specific set of data from them, namely
- * the inode di_next_unlinked fields.  This is because all other inode
- * data is actually logged via inode records and any data we replay
- * here which overlaps that may be stale.
- *
- * When meta-data buffers are freed at run time we log a buffer item
- * with the XFS_BLF_CANCEL bit set to indicate that previous copies
- * of the buffer in the log should not be replayed at recovery time.
- * This is so that if the blocks covered by the buffer are reused for
- * file data before we crash we don't end up replaying old, freed
- * meta-data into a user's file.
- *
- * To handle the cancellation of buffer log items, we make two passes
- * over the log during recovery.  During the first we build a table of
- * those buffers which have been cancelled, and during the second we
- * only replay those buffers which do not have corresponding cancel
- * records in the table.  See xlog_recover_buffer_pass[1,2] above
- * for more details on the implementation of the table of cancel records.
- */
-STATIC int
-xlog_recover_buffer_pass2(
-	struct xlog			*log,
-	struct list_head		*buffer_list,
-	struct xlog_recover_item	*item,
-	xfs_lsn_t			current_lsn)
-{
-	xfs_buf_log_format_t	*buf_f = item->ri_buf[0].i_addr;
-	xfs_mount_t		*mp = log->l_mp;
-	xfs_buf_t		*bp;
-	int			error;
-	uint			buf_flags;
-	xfs_lsn_t		lsn;
-
-	/*
-	 * In this pass we only want to recover all the buffers which have
-	 * not been cancelled and are not cancellation buffers themselves.
-	 */
-	if (xlog_check_buffer_cancelled(log, buf_f->blf_blkno,
-			buf_f->blf_len, buf_f->blf_flags)) {
-		trace_xfs_log_recover_buf_cancel(log, buf_f);
-		return 0;
-	}
-
-	trace_xfs_log_recover_buf_recover(log, buf_f);
-
-	buf_flags = 0;
-	if (buf_f->blf_flags & XFS_BLF_INODE_BUF)
-		buf_flags |= XBF_UNMAPPED;
-
-	error = xfs_buf_read(mp->m_ddev_targp, buf_f->blf_blkno, buf_f->blf_len,
-			  buf_flags, &bp, NULL);
-	if (error)
-		return error;
-
-	/*
-	 * Recover the buffer only if we get an LSN from it and it's less than
-	 * the lsn of the transaction we are replaying.
-	 *
-	 * Note that we have to be extremely careful of readahead here.
-	 * Readahead does not attach verfiers to the buffers so if we don't
-	 * actually do any replay after readahead because of the LSN we found
-	 * in the buffer if more recent than that current transaction then we
-	 * need to attach the verifier directly. Failure to do so can lead to
-	 * future recovery actions (e.g. EFI and unlinked list recovery) can
-	 * operate on the buffers and they won't get the verifier attached. This
-	 * can lead to blocks on disk having the correct content but a stale
-	 * CRC.
-	 *
-	 * It is safe to assume these clean buffers are currently up to date.
-	 * If the buffer is dirtied by a later transaction being replayed, then
-	 * the verifier will be reset to match whatever recover turns that
-	 * buffer into.
-	 */
-	lsn = xlog_recover_get_buf_lsn(mp, bp);
-	if (lsn && lsn != -1 && XFS_LSN_CMP(lsn, current_lsn) >= 0) {
-		trace_xfs_log_recover_buf_skip(log, buf_f);
-		xlog_recover_validate_buf_type(mp, bp, buf_f, NULLCOMMITLSN);
-		goto out_release;
-	}
-
-	if (buf_f->blf_flags & XFS_BLF_INODE_BUF) {
-		error = xlog_recover_do_inode_buffer(mp, item, bp, buf_f);
-		if (error)
-			goto out_release;
-	} else if (buf_f->blf_flags &
-		  (XFS_BLF_UDQUOT_BUF|XFS_BLF_PDQUOT_BUF|XFS_BLF_GDQUOT_BUF)) {
-		bool	dirty;
-
-		dirty = xlog_recover_do_dquot_buffer(mp, log, item, bp, buf_f);
-		if (!dirty)
-			goto out_release;
-	} else {
-		xlog_recover_do_reg_buffer(mp, item, bp, buf_f, current_lsn);
-	}
-
-	/*
-	 * Perform delayed write on the buffer.  Asynchronous writes will be
-	 * slower when taking into account all the buffers to be flushed.
-	 *
-	 * Also make sure that only inode buffers with good sizes stay in
-	 * the buffer cache.  The kernel moves inodes in buffers of 1 block
-	 * or inode_cluster_size bytes, whichever is bigger.  The inode
-	 * buffers in the log can be a different size if the log was generated
-	 * by an older kernel using unclustered inode buffers or a newer kernel
-	 * running with a different inode cluster size.  Regardless, if the
-	 * the inode buffer size isn't max(blocksize, inode_cluster_size)
-	 * for *our* value of inode_cluster_size, then we need to keep
-	 * the buffer out of the buffer cache so that the buffer won't
-	 * overlap with future reads of those inodes.
-	 */
-	if (XFS_DINODE_MAGIC ==
-	    be16_to_cpu(*((__be16 *)xfs_buf_offset(bp, 0))) &&
-	    (BBTOB(bp->b_length) != M_IGEO(log->l_mp)->inode_cluster_size)) {
-		xfs_buf_stale(bp);
-		error = xfs_bwrite(bp);
-	} else {
-		ASSERT(bp->b_mount == mp);
-		bp->b_iodone = xlog_recover_iodone;
-		xfs_buf_delwri_queue(bp, buffer_list);
-	}
-
-out_release:
-	xfs_buf_relse(bp);
-	return error;
-}
-
-/*
- * Inode fork owner changes
- *
- * If we have been told that we have to reparent the inode fork, it's because an
- * extent swap operation on a CRC enabled filesystem has been done and we are
- * replaying it. We need to walk the BMBT of the appropriate fork and change the
- * owners of it.
- *
- * The complexity here is that we don't have an inode context to work with, so
- * after we've replayed the inode we need to instantiate one.  This is where the
- * fun begins.
- *
- * We are in the middle of log recovery, so we can't run transactions. That
- * means we cannot use cache coherent inode instantiation via xfs_iget(), as
- * that will result in the corresponding iput() running the inode through
- * xfs_inactive(). If we've just replayed an inode core that changes the link
- * count to zero (i.e. it's been unlinked), then xfs_inactive() will run
- * transactions (bad!).
- *
- * So, to avoid this, we instantiate an inode directly from the inode core we've
- * just recovered. We have the buffer still locked, and all we really need to
- * instantiate is the inode core and the forks being modified. We can do this
- * manually, then run the inode btree owner change, and then tear down the
- * xfs_inode without having to run any transactions at all.
- *
- * Also, because we don't have a transaction context available here but need to
- * gather all the buffers we modify for writeback so we pass the buffer_list
- * instead for the operation to use.
- */
-
-STATIC int
-xfs_recover_inode_owner_change(
-	struct xfs_mount	*mp,
-	struct xfs_dinode	*dip,
-	struct xfs_inode_log_format *in_f,
-	struct list_head	*buffer_list)
-{
-	struct xfs_inode	*ip;
-	int			error;
-
-	ASSERT(in_f->ilf_fields & (XFS_ILOG_DOWNER|XFS_ILOG_AOWNER));
-
-	ip = xfs_inode_alloc(mp, in_f->ilf_ino);
-	if (!ip)
-		return -ENOMEM;
-
-	/* instantiate the inode */
-	xfs_inode_from_disk(ip, dip);
-	ASSERT(ip->i_d.di_version >= 3);
-
-	error = xfs_iformat_fork(ip, dip);
-	if (error)
-		goto out_free_ip;
-
-	if (!xfs_inode_verify_forks(ip)) {
-		error = -EFSCORRUPTED;
-		goto out_free_ip;
-	}
-
-	if (in_f->ilf_fields & XFS_ILOG_DOWNER) {
-		ASSERT(in_f->ilf_fields & XFS_ILOG_DBROOT);
-		error = xfs_bmbt_change_owner(NULL, ip, XFS_DATA_FORK,
-					      ip->i_ino, buffer_list);
-		if (error)
-			goto out_free_ip;
-	}
-
-	if (in_f->ilf_fields & XFS_ILOG_AOWNER) {
-		ASSERT(in_f->ilf_fields & XFS_ILOG_ABROOT);
-		error = xfs_bmbt_change_owner(NULL, ip, XFS_ATTR_FORK,
-					      ip->i_ino, buffer_list);
-		if (error)
-			goto out_free_ip;
-	}
-
-out_free_ip:
-	xfs_inode_free(ip);
-	return error;
-}
-
-STATIC int
-xlog_recover_inode_pass2(
-	struct xlog			*log,
-	struct list_head		*buffer_list,
-	struct xlog_recover_item	*item,
-	xfs_lsn_t			current_lsn)
-{
-	struct xfs_inode_log_format	*in_f;
-	xfs_mount_t		*mp = log->l_mp;
-	xfs_buf_t		*bp;
-	xfs_dinode_t		*dip;
-	int			len;
-	char			*src;
-	char			*dest;
-	int			error;
-	int			attr_index;
-	uint			fields;
-	struct xfs_log_dinode	*ldip;
-	uint			isize;
-	int			need_free = 0;
-
-	if (item->ri_buf[0].i_len == sizeof(struct xfs_inode_log_format)) {
-		in_f = item->ri_buf[0].i_addr;
-	} else {
-		in_f = kmem_alloc(sizeof(struct xfs_inode_log_format), 0);
-		need_free = 1;
-		error = xfs_inode_item_format_convert(&item->ri_buf[0], in_f);
-		if (error)
-			goto error;
-	}
-
-	/*
-	 * Inode buffers can be freed, look out for it,
-	 * and do not replay the inode.
-	 */
-	if (xlog_check_buffer_cancelled(log, in_f->ilf_blkno,
-					in_f->ilf_len, 0)) {
-		error = 0;
-		trace_xfs_log_recover_inode_cancel(log, in_f);
-		goto error;
-	}
-	trace_xfs_log_recover_inode_recover(log, in_f);
-
-	error = xfs_buf_read(mp->m_ddev_targp, in_f->ilf_blkno, in_f->ilf_len,
-			0, &bp, &xfs_inode_buf_ops);
-	if (error)
-		goto error;
-	ASSERT(in_f->ilf_fields & XFS_ILOG_CORE);
-	dip = xfs_buf_offset(bp, in_f->ilf_boffset);
-
-	/*
-	 * Make sure the place we're flushing out to really looks
-	 * like an inode!
-	 */
-	if (XFS_IS_CORRUPT(mp, !xfs_verify_magic16(bp, dip->di_magic))) {
-		xfs_alert(mp,
-	"%s: Bad inode magic number, dip = "PTR_FMT", dino bp = "PTR_FMT", ino = %Ld",
-			__func__, dip, bp, in_f->ilf_ino);
-		error = -EFSCORRUPTED;
-		goto out_release;
-	}
-	ldip = item->ri_buf[1].i_addr;
-	if (XFS_IS_CORRUPT(mp, ldip->di_magic != XFS_DINODE_MAGIC)) {
-		xfs_alert(mp,
-			"%s: Bad inode log record, rec ptr "PTR_FMT", ino %Ld",
-			__func__, item, in_f->ilf_ino);
-		error = -EFSCORRUPTED;
-		goto out_release;
-	}
-
-	/*
-	 * If the inode has an LSN in it, recover the inode only if it's less
-	 * than the lsn of the transaction we are replaying. Note: we still
-	 * need to replay an owner change even though the inode is more recent
-	 * than the transaction as there is no guarantee that all the btree
-	 * blocks are more recent than this transaction, too.
-	 */
-	if (dip->di_version >= 3) {
-		xfs_lsn_t	lsn = be64_to_cpu(dip->di_lsn);
-
-		if (lsn && lsn != -1 && XFS_LSN_CMP(lsn, current_lsn) >= 0) {
-			trace_xfs_log_recover_inode_skip(log, in_f);
-			error = 0;
-			goto out_owner_change;
-		}
-	}
-
-	/*
-	 * di_flushiter is only valid for v1/2 inodes. All changes for v3 inodes
-	 * are transactional and if ordering is necessary we can determine that
-	 * more accurately by the LSN field in the V3 inode core. Don't trust
-	 * the inode versions we might be changing them here - use the
-	 * superblock flag to determine whether we need to look at di_flushiter
-	 * to skip replay when the on disk inode is newer than the log one
-	 */
-	if (!xfs_sb_version_hascrc(&mp->m_sb) &&
-	    ldip->di_flushiter < be16_to_cpu(dip->di_flushiter)) {
-		/*
-		 * Deal with the wrap case, DI_MAX_FLUSH is less
-		 * than smaller numbers
-		 */
-		if (be16_to_cpu(dip->di_flushiter) == DI_MAX_FLUSH &&
-		    ldip->di_flushiter < (DI_MAX_FLUSH >> 1)) {
-			/* do nothing */
-		} else {
-			trace_xfs_log_recover_inode_skip(log, in_f);
-			error = 0;
-			goto out_release;
-		}
-	}
-
-	/* Take the opportunity to reset the flush iteration count */
-	ldip->di_flushiter = 0;
-
-	if (unlikely(S_ISREG(ldip->di_mode))) {
-		if ((ldip->di_format != XFS_DINODE_FMT_EXTENTS) &&
-		    (ldip->di_format != XFS_DINODE_FMT_BTREE)) {
-			XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(3)",
-					 XFS_ERRLEVEL_LOW, mp, ldip,
-					 sizeof(*ldip));
-			xfs_alert(mp,
-		"%s: Bad regular inode log record, rec ptr "PTR_FMT", "
-		"ino ptr = "PTR_FMT", ino bp = "PTR_FMT", ino %Ld",
-				__func__, item, dip, bp, in_f->ilf_ino);
-			error = -EFSCORRUPTED;
-			goto out_release;
-		}
-	} else if (unlikely(S_ISDIR(ldip->di_mode))) {
-		if ((ldip->di_format != XFS_DINODE_FMT_EXTENTS) &&
-		    (ldip->di_format != XFS_DINODE_FMT_BTREE) &&
-		    (ldip->di_format != XFS_DINODE_FMT_LOCAL)) {
-			XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(4)",
-					     XFS_ERRLEVEL_LOW, mp, ldip,
-					     sizeof(*ldip));
-			xfs_alert(mp,
-		"%s: Bad dir inode log record, rec ptr "PTR_FMT", "
-		"ino ptr = "PTR_FMT", ino bp = "PTR_FMT", ino %Ld",
-				__func__, item, dip, bp, in_f->ilf_ino);
-			error = -EFSCORRUPTED;
-			goto out_release;
-		}
-	}
-	if (unlikely(ldip->di_nextents + ldip->di_anextents > ldip->di_nblocks)){
-		XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(5)",
-				     XFS_ERRLEVEL_LOW, mp, ldip,
-				     sizeof(*ldip));
-		xfs_alert(mp,
-	"%s: Bad inode log record, rec ptr "PTR_FMT", dino ptr "PTR_FMT", "
-	"dino bp "PTR_FMT", ino %Ld, total extents = %d, nblocks = %Ld",
-			__func__, item, dip, bp, in_f->ilf_ino,
-			ldip->di_nextents + ldip->di_anextents,
-			ldip->di_nblocks);
-		error = -EFSCORRUPTED;
-		goto out_release;
-	}
-	if (unlikely(ldip->di_forkoff > mp->m_sb.sb_inodesize)) {
-		XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(6)",
-				     XFS_ERRLEVEL_LOW, mp, ldip,
-				     sizeof(*ldip));
-		xfs_alert(mp,
-	"%s: Bad inode log record, rec ptr "PTR_FMT", dino ptr "PTR_FMT", "
-	"dino bp "PTR_FMT", ino %Ld, forkoff 0x%x", __func__,
-			item, dip, bp, in_f->ilf_ino, ldip->di_forkoff);
-		error = -EFSCORRUPTED;
-		goto out_release;
-	}
-	isize = xfs_log_dinode_size(ldip->di_version);
-	if (unlikely(item->ri_buf[1].i_len > isize)) {
-		XFS_CORRUPTION_ERROR("xlog_recover_inode_pass2(7)",
-				     XFS_ERRLEVEL_LOW, mp, ldip,
-				     sizeof(*ldip));
-		xfs_alert(mp,
-			"%s: Bad inode log record length %d, rec ptr "PTR_FMT,
-			__func__, item->ri_buf[1].i_len, item);
-		error = -EFSCORRUPTED;
-		goto out_release;
-	}
-
-	/* recover the log dinode inode into the on disk inode */
-	xfs_log_dinode_to_disk(ldip, dip);
-
-	fields = in_f->ilf_fields;
-	if (fields & XFS_ILOG_DEV)
-		xfs_dinode_put_rdev(dip, in_f->ilf_u.ilfu_rdev);
-
-	if (in_f->ilf_size == 2)
-		goto out_owner_change;
-	len = item->ri_buf[2].i_len;
-	src = item->ri_buf[2].i_addr;
-	ASSERT(in_f->ilf_size <= 4);
-	ASSERT((in_f->ilf_size == 3) || (fields & XFS_ILOG_AFORK));
-	ASSERT(!(fields & XFS_ILOG_DFORK) ||
-	       (len == in_f->ilf_dsize));
-
-	switch (fields & XFS_ILOG_DFORK) {
-	case XFS_ILOG_DDATA:
-	case XFS_ILOG_DEXT:
-		memcpy(XFS_DFORK_DPTR(dip), src, len);
-		break;
-
-	case XFS_ILOG_DBROOT:
-		xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src, len,
-				 (xfs_bmdr_block_t *)XFS_DFORK_DPTR(dip),
-				 XFS_DFORK_DSIZE(dip, mp));
-		break;
-
-	default:
-		/*
-		 * There are no data fork flags set.
-		 */
-		ASSERT((fields & XFS_ILOG_DFORK) == 0);
-		break;
-	}
-
-	/*
-	 * If we logged any attribute data, recover it.  There may or
-	 * may not have been any other non-core data logged in this
-	 * transaction.
-	 */
-	if (in_f->ilf_fields & XFS_ILOG_AFORK) {
-		if (in_f->ilf_fields & XFS_ILOG_DFORK) {
-			attr_index = 3;
-		} else {
-			attr_index = 2;
-		}
-		len = item->ri_buf[attr_index].i_len;
-		src = item->ri_buf[attr_index].i_addr;
-		ASSERT(len == in_f->ilf_asize);
-
-		switch (in_f->ilf_fields & XFS_ILOG_AFORK) {
-		case XFS_ILOG_ADATA:
-		case XFS_ILOG_AEXT:
-			dest = XFS_DFORK_APTR(dip);
-			ASSERT(len <= XFS_DFORK_ASIZE(dip, mp));
-			memcpy(dest, src, len);
-			break;
-
-		case XFS_ILOG_ABROOT:
-			dest = XFS_DFORK_APTR(dip);
-			xfs_bmbt_to_bmdr(mp, (struct xfs_btree_block *)src,
-					 len, (xfs_bmdr_block_t*)dest,
-					 XFS_DFORK_ASIZE(dip, mp));
-			break;
-
-		default:
-			xfs_warn(log->l_mp, "%s: Invalid flag", __func__);
-			ASSERT(0);
-			error = -EFSCORRUPTED;
-			goto out_release;
-		}
-	}
-
-out_owner_change:
-	/* Recover the swapext owner change unless inode has been deleted */
-	if ((in_f->ilf_fields & (XFS_ILOG_DOWNER|XFS_ILOG_AOWNER)) &&
-	    (dip->di_mode != 0))
-		error = xfs_recover_inode_owner_change(mp, dip, in_f,
-						       buffer_list);
-	/* re-generate the checksum. */
-	xfs_dinode_calc_crc(log->l_mp, dip);
-
-	ASSERT(bp->b_mount == mp);
-	bp->b_iodone = xlog_recover_iodone;
-	xfs_buf_delwri_queue(bp, buffer_list);
-
-out_release:
-	xfs_buf_relse(bp);
-error:
-	if (need_free)
-		kmem_free(in_f);
-	return error;
-}
-
-/*
- * Recover QUOTAOFF records. We simply make a note of it in the xlog
- * structure, so that we know not to do any dquot item or dquot buffer recovery,
- * of that type.
- */
-STATIC int
-xlog_recover_quotaoff_pass1(
-	struct xlog			*log,
-	struct xlog_recover_item	*item)
-{
-	xfs_qoff_logformat_t	*qoff_f = item->ri_buf[0].i_addr;
-	ASSERT(qoff_f);
-
-	/*
-	 * The logitem format's flag tells us if this was user quotaoff,
-	 * group/project quotaoff or both.
-	 */
-	if (qoff_f->qf_flags & XFS_UQUOTA_ACCT)
-		log->l_quotaoffs_flag |= XFS_DQ_USER;
-	if (qoff_f->qf_flags & XFS_PQUOTA_ACCT)
-		log->l_quotaoffs_flag |= XFS_DQ_PROJ;
-	if (qoff_f->qf_flags & XFS_GQUOTA_ACCT)
-		log->l_quotaoffs_flag |= XFS_DQ_GROUP;
-
-	return 0;
-}
-
-/*
- * Recover a dquot record
- */
-STATIC int
-xlog_recover_dquot_pass2(
-	struct xlog			*log,
-	struct list_head		*buffer_list,
-	struct xlog_recover_item	*item,
-	xfs_lsn_t			current_lsn)
-{
-	xfs_mount_t		*mp = log->l_mp;
-	xfs_buf_t		*bp;
-	struct xfs_disk_dquot	*ddq, *recddq;
-	xfs_failaddr_t		fa;
-	int			error;
-	xfs_dq_logformat_t	*dq_f;
-	uint			type;
-
-
-	/*
-	 * Filesystems are required to send in quota flags at mount time.
-	 */
-	if (mp->m_qflags == 0)
-		return 0;
-
-	recddq = item->ri_buf[1].i_addr;
-	if (recddq == NULL) {
-		xfs_alert(log->l_mp, "NULL dquot in %s.", __func__);
-		return -EFSCORRUPTED;
-	}
-	if (item->ri_buf[1].i_len < sizeof(struct xfs_disk_dquot)) {
-		xfs_alert(log->l_mp, "dquot too small (%d) in %s.",
-			item->ri_buf[1].i_len, __func__);
-		return -EFSCORRUPTED;
-	}
-
-	/*
-	 * This type of quotas was turned off, so ignore this record.
-	 */
-	type = recddq->d_flags & (XFS_DQ_USER | XFS_DQ_PROJ | XFS_DQ_GROUP);
-	ASSERT(type);
-	if (log->l_quotaoffs_flag & type)
-		return 0;
-
-	/*
-	 * At this point we know that quota was _not_ turned off.
-	 * Since the mount flags are not indicating to us otherwise, this
-	 * must mean that quota is on, and the dquot needs to be replayed.
-	 * Remember that we may not have fully recovered the superblock yet,
-	 * so we can't do the usual trick of looking at the SB quota bits.
-	 *
-	 * The other possibility, of course, is that the quota subsystem was
-	 * removed since the last mount - ENOSYS.
-	 */
-	dq_f = item->ri_buf[0].i_addr;
-	ASSERT(dq_f);
-	fa = xfs_dquot_verify(mp, recddq, dq_f->qlf_id, 0);
-	if (fa) {
-		xfs_alert(mp, "corrupt dquot ID 0x%x in log at %pS",
-				dq_f->qlf_id, fa);
-		return -EFSCORRUPTED;
-	}
-	ASSERT(dq_f->qlf_len == 1);
-
-	/*
-	 * At this point we are assuming that the dquots have been allocated
-	 * and hence the buffer has valid dquots stamped in it. It should,
-	 * therefore, pass verifier validation. If the dquot is bad, then the
-	 * we'll return an error here, so we don't need to specifically check
-	 * the dquot in the buffer after the verifier has run.
-	 */
-	error = xfs_trans_read_buf(mp, NULL, mp->m_ddev_targp, dq_f->qlf_blkno,
-				   XFS_FSB_TO_BB(mp, dq_f->qlf_len), 0, &bp,
-				   &xfs_dquot_buf_ops);
-	if (error)
-		return error;
-
-	ASSERT(bp);
-	ddq = xfs_buf_offset(bp, dq_f->qlf_boffset);
-
-	/*
-	 * If the dquot has an LSN in it, recover the dquot only if it's less
-	 * than the lsn of the transaction we are replaying.
-	 */
-	if (xfs_sb_version_hascrc(&mp->m_sb)) {
-		struct xfs_dqblk *dqb = (struct xfs_dqblk *)ddq;
-		xfs_lsn_t	lsn = be64_to_cpu(dqb->dd_lsn);
-
-		if (lsn && lsn != -1 && XFS_LSN_CMP(lsn, current_lsn) >= 0) {
-			goto out_release;
-		}
-	}
-
-	memcpy(ddq, recddq, item->ri_buf[1].i_len);
-	if (xfs_sb_version_hascrc(&mp->m_sb)) {
-		xfs_update_cksum((char *)ddq, sizeof(struct xfs_dqblk),
-				 XFS_DQUOT_CRC_OFF);
-	}
-
-	ASSERT(dq_f->qlf_size == 2);
-	ASSERT(bp->b_mount == mp);
-	bp->b_iodone = xlog_recover_iodone;
-	xfs_buf_delwri_queue(bp, buffer_list);
-
-out_release:
-	xfs_buf_relse(bp);
-	return 0;
-}
-
-/*
- * This routine is called to create an in-core extent free intent
- * item from the efi format structure which was logged on disk.
- * It allocates an in-core efi, copies the extents from the format
- * structure into it, and adds the efi to the AIL with the given
- * LSN.
- */
-STATIC int
-xlog_recover_efi_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item,
-	xfs_lsn_t			lsn)
-{
-	int				error;
-	struct xfs_mount		*mp = log->l_mp;
-	struct xfs_efi_log_item		*efip;
-	struct xfs_efi_log_format	*efi_formatp;
-
-	efi_formatp = item->ri_buf[0].i_addr;
-
-	efip = xfs_efi_init(mp, efi_formatp->efi_nextents);
-	error = xfs_efi_copy_format(&item->ri_buf[0], &efip->efi_format);
-	if (error) {
-		xfs_efi_item_free(efip);
-		return error;
-	}
-	atomic_set(&efip->efi_next_extent, efi_formatp->efi_nextents);
-
-	spin_lock(&log->l_ailp->ail_lock);
-	/*
-	 * The EFI has two references. One for the EFD and one for EFI to ensure
-	 * it makes it into the AIL. Insert the EFI into the AIL directly and
-	 * drop the EFI reference. Note that xfs_trans_ail_update() drops the
-	 * AIL lock.
-	 */
-	xfs_trans_ail_update(log->l_ailp, &efip->efi_item, lsn);
-	xfs_efi_release(efip);
-	return 0;
-}
-
-
-/*
- * This routine is called when an EFD format structure is found in a committed
- * transaction in the log. Its purpose is to cancel the corresponding EFI if it
- * was still in the log. To do this it searches the AIL for the EFI with an id
- * equal to that in the EFD format structure. If we find it we drop the EFD
- * reference, which removes the EFI from the AIL and frees it.
- */
-STATIC int
-xlog_recover_efd_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item)
-{
-	xfs_efd_log_format_t	*efd_formatp;
-	xfs_efi_log_item_t	*efip = NULL;
-	struct xfs_log_item	*lip;
-	uint64_t		efi_id;
-	struct xfs_ail_cursor	cur;
-	struct xfs_ail		*ailp = log->l_ailp;
-
-	efd_formatp = item->ri_buf[0].i_addr;
-	ASSERT((item->ri_buf[0].i_len == (sizeof(xfs_efd_log_format_32_t) +
-		((efd_formatp->efd_nextents - 1) * sizeof(xfs_extent_32_t)))) ||
-	       (item->ri_buf[0].i_len == (sizeof(xfs_efd_log_format_64_t) +
-		((efd_formatp->efd_nextents - 1) * sizeof(xfs_extent_64_t)))));
-	efi_id = efd_formatp->efd_efi_id;
-
-	/*
-	 * Search for the EFI with the id in the EFD format structure in the
-	 * AIL.
-	 */
-	spin_lock(&ailp->ail_lock);
-	lip = xfs_trans_ail_cursor_first(ailp, &cur, 0);
-	while (lip != NULL) {
-		if (lip->li_type == XFS_LI_EFI) {
-			efip = (xfs_efi_log_item_t *)lip;
-			if (efip->efi_format.efi_id == efi_id) {
-				/*
-				 * Drop the EFD reference to the EFI. This
-				 * removes the EFI from the AIL and frees it.
-				 */
-				spin_unlock(&ailp->ail_lock);
-				xfs_efi_release(efip);
-				spin_lock(&ailp->ail_lock);
-				break;
-			}
-		}
-		lip = xfs_trans_ail_cursor_next(ailp, &cur);
-	}
-
-	xfs_trans_ail_cursor_done(&cur);
-	spin_unlock(&ailp->ail_lock);
-
-	return 0;
-}
-
-/*
- * This routine is called to create an in-core extent rmap update
- * item from the rui format structure which was logged on disk.
- * It allocates an in-core rui, copies the extents from the format
- * structure into it, and adds the rui to the AIL with the given
- * LSN.
- */
-STATIC int
-xlog_recover_rui_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item,
-	xfs_lsn_t			lsn)
-{
-	int				error;
-	struct xfs_mount		*mp = log->l_mp;
-	struct xfs_rui_log_item		*ruip;
-	struct xfs_rui_log_format	*rui_formatp;
-
-	rui_formatp = item->ri_buf[0].i_addr;
-
-	ruip = xfs_rui_init(mp, rui_formatp->rui_nextents);
-	error = xfs_rui_copy_format(&item->ri_buf[0], &ruip->rui_format);
-	if (error) {
-		xfs_rui_item_free(ruip);
-		return error;
-	}
-	atomic_set(&ruip->rui_next_extent, rui_formatp->rui_nextents);
-
-	spin_lock(&log->l_ailp->ail_lock);
-	/*
-	 * The RUI has two references. One for the RUD and one for RUI to ensure
-	 * it makes it into the AIL. Insert the RUI into the AIL directly and
-	 * drop the RUI reference. Note that xfs_trans_ail_update() drops the
-	 * AIL lock.
-	 */
-	xfs_trans_ail_update(log->l_ailp, &ruip->rui_item, lsn);
-	xfs_rui_release(ruip);
-	return 0;
-}
-
-
-/*
- * This routine is called when an RUD format structure is found in a committed
- * transaction in the log. Its purpose is to cancel the corresponding RUI if it
- * was still in the log. To do this it searches the AIL for the RUI with an id
- * equal to that in the RUD format structure. If we find it we drop the RUD
- * reference, which removes the RUI from the AIL and frees it.
- */
-STATIC int
-xlog_recover_rud_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item)
-{
-	struct xfs_rud_log_format	*rud_formatp;
-	struct xfs_rui_log_item		*ruip = NULL;
-	struct xfs_log_item		*lip;
-	uint64_t			rui_id;
-	struct xfs_ail_cursor		cur;
-	struct xfs_ail			*ailp = log->l_ailp;
-
-	rud_formatp = item->ri_buf[0].i_addr;
-	ASSERT(item->ri_buf[0].i_len == sizeof(struct xfs_rud_log_format));
-	rui_id = rud_formatp->rud_rui_id;
-
-	/*
-	 * Search for the RUI with the id in the RUD format structure in the
-	 * AIL.
-	 */
-	spin_lock(&ailp->ail_lock);
-	lip = xfs_trans_ail_cursor_first(ailp, &cur, 0);
-	while (lip != NULL) {
-		if (lip->li_type == XFS_LI_RUI) {
-			ruip = (struct xfs_rui_log_item *)lip;
-			if (ruip->rui_format.rui_id == rui_id) {
-				/*
-				 * Drop the RUD reference to the RUI. This
-				 * removes the RUI from the AIL and frees it.
-				 */
-				spin_unlock(&ailp->ail_lock);
-				xfs_rui_release(ruip);
-				spin_lock(&ailp->ail_lock);
-				break;
-			}
-		}
-		lip = xfs_trans_ail_cursor_next(ailp, &cur);
-	}
-
-	xfs_trans_ail_cursor_done(&cur);
-	spin_unlock(&ailp->ail_lock);
-
-	return 0;
-}
-
-/*
- * Copy an CUI format buffer from the given buf, and into the destination
- * CUI format structure.  The CUI/CUD items were designed not to need any
- * special alignment handling.
- */
-static int
-xfs_cui_copy_format(
-	struct xfs_log_iovec		*buf,
-	struct xfs_cui_log_format	*dst_cui_fmt)
-{
-	struct xfs_cui_log_format	*src_cui_fmt;
-	uint				len;
-
-	src_cui_fmt = buf->i_addr;
-	len = xfs_cui_log_format_sizeof(src_cui_fmt->cui_nextents);
-
-	if (buf->i_len == len) {
-		memcpy(dst_cui_fmt, src_cui_fmt, len);
-		return 0;
-	}
-	XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, NULL);
-	return -EFSCORRUPTED;
-}
-
-/*
- * This routine is called to create an in-core extent refcount update
- * item from the cui format structure which was logged on disk.
- * It allocates an in-core cui, copies the extents from the format
- * structure into it, and adds the cui to the AIL with the given
- * LSN.
- */
-STATIC int
-xlog_recover_cui_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item,
-	xfs_lsn_t			lsn)
-{
-	int				error;
-	struct xfs_mount		*mp = log->l_mp;
-	struct xfs_cui_log_item		*cuip;
-	struct xfs_cui_log_format	*cui_formatp;
-
-	cui_formatp = item->ri_buf[0].i_addr;
-
-	cuip = xfs_cui_init(mp, cui_formatp->cui_nextents);
-	error = xfs_cui_copy_format(&item->ri_buf[0], &cuip->cui_format);
-	if (error) {
-		xfs_cui_item_free(cuip);
-		return error;
-	}
-	atomic_set(&cuip->cui_next_extent, cui_formatp->cui_nextents);
-
-	spin_lock(&log->l_ailp->ail_lock);
-	/*
-	 * The CUI has two references. One for the CUD and one for CUI to ensure
-	 * it makes it into the AIL. Insert the CUI into the AIL directly and
-	 * drop the CUI reference. Note that xfs_trans_ail_update() drops the
-	 * AIL lock.
-	 */
-	xfs_trans_ail_update(log->l_ailp, &cuip->cui_item, lsn);
-	xfs_cui_release(cuip);
-	return 0;
-}
-
-
-/*
- * This routine is called when an CUD format structure is found in a committed
- * transaction in the log. Its purpose is to cancel the corresponding CUI if it
- * was still in the log. To do this it searches the AIL for the CUI with an id
- * equal to that in the CUD format structure. If we find it we drop the CUD
- * reference, which removes the CUI from the AIL and frees it.
- */
-STATIC int
-xlog_recover_cud_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item)
-{
-	struct xfs_cud_log_format	*cud_formatp;
-	struct xfs_cui_log_item		*cuip = NULL;
-	struct xfs_log_item		*lip;
-	uint64_t			cui_id;
-	struct xfs_ail_cursor		cur;
-	struct xfs_ail			*ailp = log->l_ailp;
-
-	cud_formatp = item->ri_buf[0].i_addr;
-	if (item->ri_buf[0].i_len != sizeof(struct xfs_cud_log_format)) {
-		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, log->l_mp);
-		return -EFSCORRUPTED;
-	}
-	cui_id = cud_formatp->cud_cui_id;
-
-	/*
-	 * Search for the CUI with the id in the CUD format structure in the
-	 * AIL.
-	 */
-	spin_lock(&ailp->ail_lock);
-	lip = xfs_trans_ail_cursor_first(ailp, &cur, 0);
-	while (lip != NULL) {
-		if (lip->li_type == XFS_LI_CUI) {
-			cuip = (struct xfs_cui_log_item *)lip;
-			if (cuip->cui_format.cui_id == cui_id) {
-				/*
-				 * Drop the CUD reference to the CUI. This
-				 * removes the CUI from the AIL and frees it.
-				 */
-				spin_unlock(&ailp->ail_lock);
-				xfs_cui_release(cuip);
-				spin_lock(&ailp->ail_lock);
-				break;
-			}
-		}
-		lip = xfs_trans_ail_cursor_next(ailp, &cur);
-	}
-
-	xfs_trans_ail_cursor_done(&cur);
-	spin_unlock(&ailp->ail_lock);
-
-	return 0;
-}
-
-/*
- * Copy an BUI format buffer from the given buf, and into the destination
- * BUI format structure.  The BUI/BUD items were designed not to need any
- * special alignment handling.
- */
-static int
-xfs_bui_copy_format(
-	struct xfs_log_iovec		*buf,
-	struct xfs_bui_log_format	*dst_bui_fmt)
-{
-	struct xfs_bui_log_format	*src_bui_fmt;
-	uint				len;
-
-	src_bui_fmt = buf->i_addr;
-	len = xfs_bui_log_format_sizeof(src_bui_fmt->bui_nextents);
-
-	if (buf->i_len == len) {
-		memcpy(dst_bui_fmt, src_bui_fmt, len);
-		return 0;
-	}
-	XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, NULL);
-	return -EFSCORRUPTED;
-}
-
-/*
- * This routine is called to create an in-core extent bmap update
- * item from the bui format structure which was logged on disk.
- * It allocates an in-core bui, copies the extents from the format
- * structure into it, and adds the bui to the AIL with the given
- * LSN.
- */
-STATIC int
-xlog_recover_bui_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item,
-	xfs_lsn_t			lsn)
-{
-	int				error;
-	struct xfs_mount		*mp = log->l_mp;
-	struct xfs_bui_log_item		*buip;
-	struct xfs_bui_log_format	*bui_formatp;
-
-	bui_formatp = item->ri_buf[0].i_addr;
-
-	if (bui_formatp->bui_nextents != XFS_BUI_MAX_FAST_EXTENTS) {
-		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, log->l_mp);
-		return -EFSCORRUPTED;
-	}
-	buip = xfs_bui_init(mp);
-	error = xfs_bui_copy_format(&item->ri_buf[0], &buip->bui_format);
-	if (error) {
-		xfs_bui_item_free(buip);
-		return error;
-	}
-	atomic_set(&buip->bui_next_extent, bui_formatp->bui_nextents);
-
-	spin_lock(&log->l_ailp->ail_lock);
-	/*
-	 * The RUI has two references. One for the RUD and one for RUI to ensure
-	 * it makes it into the AIL. Insert the RUI into the AIL directly and
-	 * drop the RUI reference. Note that xfs_trans_ail_update() drops the
-	 * AIL lock.
-	 */
-	xfs_trans_ail_update(log->l_ailp, &buip->bui_item, lsn);
-	xfs_bui_release(buip);
-	return 0;
-}
-
-
-/*
- * This routine is called when an BUD format structure is found in a committed
- * transaction in the log. Its purpose is to cancel the corresponding BUI if it
- * was still in the log. To do this it searches the AIL for the BUI with an id
- * equal to that in the BUD format structure. If we find it we drop the BUD
- * reference, which removes the BUI from the AIL and frees it.
- */
-STATIC int
-xlog_recover_bud_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item)
-{
-	struct xfs_bud_log_format	*bud_formatp;
-	struct xfs_bui_log_item		*buip = NULL;
-	struct xfs_log_item		*lip;
-	uint64_t			bui_id;
-	struct xfs_ail_cursor		cur;
-	struct xfs_ail			*ailp = log->l_ailp;
-
-	bud_formatp = item->ri_buf[0].i_addr;
-	if (item->ri_buf[0].i_len != sizeof(struct xfs_bud_log_format)) {
-		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, log->l_mp);
-		return -EFSCORRUPTED;
-	}
-	bui_id = bud_formatp->bud_bui_id;
-
-	/*
-	 * Search for the BUI with the id in the BUD format structure in the
-	 * AIL.
-	 */
-	spin_lock(&ailp->ail_lock);
-	lip = xfs_trans_ail_cursor_first(ailp, &cur, 0);
-	while (lip != NULL) {
-		if (lip->li_type == XFS_LI_BUI) {
-			buip = (struct xfs_bui_log_item *)lip;
-			if (buip->bui_format.bui_id == bui_id) {
-				/*
-				 * Drop the BUD reference to the BUI. This
-				 * removes the BUI from the AIL and frees it.
-				 */
-				spin_unlock(&ailp->ail_lock);
-				xfs_bui_release(buip);
-				spin_lock(&ailp->ail_lock);
-				break;
-			}
-		}
-		lip = xfs_trans_ail_cursor_next(ailp, &cur);
-	}
-
-	xfs_trans_ail_cursor_done(&cur);
-	spin_unlock(&ailp->ail_lock);
-
-	return 0;
-}
-
-/*
- * This routine is called when an inode create format structure is found in a
- * committed transaction in the log.  It's purpose is to initialise the inodes
- * being allocated on disk. This requires us to get inode cluster buffers that
- * match the range to be initialised, stamped with inode templates and written
- * by delayed write so that subsequent modifications will hit the cached buffer
- * and only need writing out at the end of recovery.
- */
-STATIC int
-xlog_recover_do_icreate_pass2(
-	struct xlog		*log,
-	struct list_head	*buffer_list,
-	xlog_recover_item_t	*item)
-{
-	struct xfs_mount	*mp = log->l_mp;
-	struct xfs_icreate_log	*icl;
-	struct xfs_ino_geometry	*igeo = M_IGEO(mp);
-	xfs_agnumber_t		agno;
-	xfs_agblock_t		agbno;
-	unsigned int		count;
-	unsigned int		isize;
-	xfs_agblock_t		length;
-	int			bb_per_cluster;
-	int			cancel_count;
-	int			nbufs;
-	int			i;
-
-	icl = (struct xfs_icreate_log *)item->ri_buf[0].i_addr;
-	if (icl->icl_type != XFS_LI_ICREATE) {
-		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad type");
-		return -EINVAL;
-	}
-
-	if (icl->icl_size != 1) {
-		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad icl size");
-		return -EINVAL;
-	}
-
-	agno = be32_to_cpu(icl->icl_ag);
-	if (agno >= mp->m_sb.sb_agcount) {
-		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad agno");
-		return -EINVAL;
-	}
-	agbno = be32_to_cpu(icl->icl_agbno);
-	if (!agbno || agbno == NULLAGBLOCK || agbno >= mp->m_sb.sb_agblocks) {
-		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad agbno");
-		return -EINVAL;
-	}
-	isize = be32_to_cpu(icl->icl_isize);
-	if (isize != mp->m_sb.sb_inodesize) {
-		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad isize");
-		return -EINVAL;
-	}
-	count = be32_to_cpu(icl->icl_count);
-	if (!count) {
-		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad count");
-		return -EINVAL;
-	}
-	length = be32_to_cpu(icl->icl_length);
-	if (!length || length >= mp->m_sb.sb_agblocks) {
-		xfs_warn(log->l_mp, "xlog_recover_do_icreate_trans: bad length");
-		return -EINVAL;
-	}
-
-	/*
-	 * The inode chunk is either full or sparse and we only support
-	 * m_ino_geo.ialloc_min_blks sized sparse allocations at this time.
-	 */
-	if (length != igeo->ialloc_blks &&
-	    length != igeo->ialloc_min_blks) {
-		xfs_warn(log->l_mp,
-			 "%s: unsupported chunk length", __FUNCTION__);
-		return -EINVAL;
-	}
-
-	/* verify inode count is consistent with extent length */
-	if ((count >> mp->m_sb.sb_inopblog) != length) {
-		xfs_warn(log->l_mp,
-			 "%s: inconsistent inode count and chunk length",
-			 __FUNCTION__);
-		return -EINVAL;
-	}
-
-	/*
-	 * The icreate transaction can cover multiple cluster buffers and these
-	 * buffers could have been freed and reused. Check the individual
-	 * buffers for cancellation so we don't overwrite anything written after
-	 * a cancellation.
-	 */
-	bb_per_cluster = XFS_FSB_TO_BB(mp, igeo->blocks_per_cluster);
-	nbufs = length / igeo->blocks_per_cluster;
-	for (i = 0, cancel_count = 0; i < nbufs; i++) {
-		xfs_daddr_t	daddr;
-
-		daddr = XFS_AGB_TO_DADDR(mp, agno,
-				agbno + i * igeo->blocks_per_cluster);
-		if (xlog_check_buffer_cancelled(log, daddr, bb_per_cluster, 0))
-			cancel_count++;
-	}
-
-	/*
-	 * We currently only use icreate for a single allocation at a time. This
-	 * means we should expect either all or none of the buffers to be
-	 * cancelled. Be conservative and skip replay if at least one buffer is
-	 * cancelled, but warn the user that something is awry if the buffers
-	 * are not consistent.
-	 *
-	 * XXX: This must be refined to only skip cancelled clusters once we use
-	 * icreate for multiple chunk allocations.
-	 */
-	ASSERT(!cancel_count || cancel_count == nbufs);
-	if (cancel_count) {
-		if (cancel_count != nbufs)
-			xfs_warn(mp,
-	"WARNING: partial inode chunk cancellation, skipped icreate.");
-		trace_xfs_log_recover_icreate_cancel(log, icl);
-		return 0;
-	}
-
-	trace_xfs_log_recover_icreate_recover(log, icl);
-	return xfs_ialloc_inode_init(mp, NULL, buffer_list, count, agno, agbno,
-				     length, be32_to_cpu(icl->icl_gen));
-}
-
-STATIC void
-xlog_recover_buffer_ra_pass2(
-	struct xlog                     *log,
-	struct xlog_recover_item        *item)
-{
-	struct xfs_buf_log_format	*buf_f = item->ri_buf[0].i_addr;
-	struct xfs_mount		*mp = log->l_mp;
-
-	if (xlog_peek_buffer_cancelled(log, buf_f->blf_blkno,
-			buf_f->blf_len, buf_f->blf_flags)) {
-		return;
-	}
-
-	xfs_buf_readahead(mp->m_ddev_targp, buf_f->blf_blkno,
-				buf_f->blf_len, NULL);
-}
-
-STATIC void
-xlog_recover_inode_ra_pass2(
-	struct xlog                     *log,
-	struct xlog_recover_item        *item)
-{
-	struct xfs_inode_log_format	ilf_buf;
-	struct xfs_inode_log_format	*ilfp;
-	struct xfs_mount		*mp = log->l_mp;
-	int			error;
-
-	if (item->ri_buf[0].i_len == sizeof(struct xfs_inode_log_format)) {
-		ilfp = item->ri_buf[0].i_addr;
-	} else {
-		ilfp = &ilf_buf;
-		memset(ilfp, 0, sizeof(*ilfp));
-		error = xfs_inode_item_format_convert(&item->ri_buf[0], ilfp);
-		if (error)
-			return;
-	}
-
-	if (xlog_peek_buffer_cancelled(log, ilfp->ilf_blkno, ilfp->ilf_len, 0))
-		return;
-
-	xfs_buf_readahead(mp->m_ddev_targp, ilfp->ilf_blkno,
-				ilfp->ilf_len, &xfs_inode_buf_ra_ops);
-}
-
-STATIC void
-xlog_recover_dquot_ra_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item)
-{
-	struct xfs_mount	*mp = log->l_mp;
-	struct xfs_disk_dquot	*recddq;
-	struct xfs_dq_logformat	*dq_f;
-	uint			type;
-	int			len;
-
-
-	if (mp->m_qflags == 0)
-		return;
-
-	recddq = item->ri_buf[1].i_addr;
-	if (recddq == NULL)
-		return;
-	if (item->ri_buf[1].i_len < sizeof(struct xfs_disk_dquot))
-		return;
-
-	type = recddq->d_flags & (XFS_DQ_USER | XFS_DQ_PROJ | XFS_DQ_GROUP);
-	ASSERT(type);
-	if (log->l_quotaoffs_flag & type)
-		return;
-
-	dq_f = item->ri_buf[0].i_addr;
-	ASSERT(dq_f);
-	ASSERT(dq_f->qlf_len == 1);
-
-	len = XFS_FSB_TO_BB(mp, dq_f->qlf_len);
-	if (xlog_peek_buffer_cancelled(log, dq_f->qlf_blkno, len, 0))
-		return;
-
-	xfs_buf_readahead(mp->m_ddev_targp, dq_f->qlf_blkno, len,
-			  &xfs_dquot_buf_ra_ops);
-}
-
-STATIC void
-xlog_recover_ra_pass2(
-	struct xlog			*log,
-	struct xlog_recover_item	*item)
-{
-	switch (ITEM_TYPE(item)) {
-	case XFS_LI_BUF:
-		xlog_recover_buffer_ra_pass2(log, item);
-		break;
-	case XFS_LI_INODE:
-		xlog_recover_inode_ra_pass2(log, item);
-		break;
-	case XFS_LI_DQUOT:
-		xlog_recover_dquot_ra_pass2(log, item);
-		break;
-	case XFS_LI_EFI:
-	case XFS_LI_EFD:
-	case XFS_LI_QUOTAOFF:
-	case XFS_LI_RUI:
-	case XFS_LI_RUD:
-	case XFS_LI_CUI:
-	case XFS_LI_CUD:
-	case XFS_LI_BUI:
-	case XFS_LI_BUD:
-	default:
-		break;
-	}
-}
-
-STATIC int
-xlog_recover_commit_pass1(
-	struct xlog			*log,
-	struct xlog_recover		*trans,
-	struct xlog_recover_item	*item)
-{
-	trace_xfs_log_recover_item_recover(log, trans, item, XLOG_RECOVER_PASS1);
-
-	switch (ITEM_TYPE(item)) {
-	case XFS_LI_BUF:
-		return xlog_recover_buffer_pass1(log, item);
-	case XFS_LI_QUOTAOFF:
-		return xlog_recover_quotaoff_pass1(log, item);
-	case XFS_LI_INODE:
-	case XFS_LI_EFI:
-	case XFS_LI_EFD:
-	case XFS_LI_DQUOT:
-	case XFS_LI_ICREATE:
-	case XFS_LI_RUI:
-	case XFS_LI_RUD:
-	case XFS_LI_CUI:
-	case XFS_LI_CUD:
-	case XFS_LI_BUI:
-	case XFS_LI_BUD:
-		/* nothing to do in pass 1 */
-		return 0;
-	default:
-		xfs_warn(log->l_mp, "%s: invalid item type (%d)",
-			__func__, ITEM_TYPE(item));
-		ASSERT(0);
-		return -EFSCORRUPTED;
-	}
-}
-
-STATIC int
-xlog_recover_commit_pass2(
-	struct xlog			*log,
-	struct xlog_recover		*trans,
-	struct list_head		*buffer_list,
-	struct xlog_recover_item	*item)
-{
-	trace_xfs_log_recover_item_recover(log, trans, item, XLOG_RECOVER_PASS2);
-
-	switch (ITEM_TYPE(item)) {
-	case XFS_LI_BUF:
-		return xlog_recover_buffer_pass2(log, buffer_list, item,
-						 trans->r_lsn);
-	case XFS_LI_INODE:
-		return xlog_recover_inode_pass2(log, buffer_list, item,
-						 trans->r_lsn);
-	case XFS_LI_EFI:
-		return xlog_recover_efi_pass2(log, item, trans->r_lsn);
-	case XFS_LI_EFD:
-		return xlog_recover_efd_pass2(log, item);
-	case XFS_LI_RUI:
-		return xlog_recover_rui_pass2(log, item, trans->r_lsn);
-	case XFS_LI_RUD:
-		return xlog_recover_rud_pass2(log, item);
-	case XFS_LI_CUI:
-		return xlog_recover_cui_pass2(log, item, trans->r_lsn);
-	case XFS_LI_CUD:
-		return xlog_recover_cud_pass2(log, item);
-	case XFS_LI_BUI:
-		return xlog_recover_bui_pass2(log, item, trans->r_lsn);
-	case XFS_LI_BUD:
-		return xlog_recover_bud_pass2(log, item);
-	case XFS_LI_DQUOT:
-		return xlog_recover_dquot_pass2(log, buffer_list, item,
-						trans->r_lsn);
-	case XFS_LI_ICREATE:
-		return xlog_recover_do_icreate_pass2(log, buffer_list, item);
-	case XFS_LI_QUOTAOFF:
-		/* nothing to do in pass2 */
-		return 0;
-	default:
-		xfs_warn(log->l_mp, "%s: invalid item type (%d)",
-			__func__, ITEM_TYPE(item));
-		ASSERT(0);
-		return -EFSCORRUPTED;
-	}
+	if (!xlog_is_buffer_cancelled(log, blkno, len))
+		xfs_buf_readahead(log->l_mp->m_ddev_targp, blkno, len, ops);
 }
 
 STATIC int
@@ -4072,8 +1966,12 @@ xlog_recover_items_pass2(
 	int				error = 0;
 
 	list_for_each_entry(item, item_list, ri_list) {
-		error = xlog_recover_commit_pass2(log, trans,
-					  buffer_list, item);
+		trace_xfs_log_recover_item_recover(log, trans, item,
+				XLOG_RECOVER_PASS2);
+
+		if (item->ri_ops->commit_pass2)
+			error = item->ri_ops->commit_pass2(log, buffer_list,
+					item, trans->r_lsn);
 		if (error)
 			return error;
 	}
@@ -4110,12 +2008,16 @@ xlog_recover_commit_trans(
 		return error;
 
 	list_for_each_entry_safe(item, next, &trans->r_itemq, ri_list) {
+		trace_xfs_log_recover_item_recover(log, trans, item, pass);
+
 		switch (pass) {
 		case XLOG_RECOVER_PASS1:
-			error = xlog_recover_commit_pass1(log, trans, item);
+			if (item->ri_ops->commit_pass1)
+				error = item->ri_ops->commit_pass1(log, item);
 			break;
 		case XLOG_RECOVER_PASS2:
-			xlog_recover_ra_pass2(log, item);
+			if (item->ri_ops->ra_pass2)
+				item->ri_ops->ra_pass2(log, item);
 			list_move_tail(&item->ri_list, &ra_list);
 			items_queued++;
 			if (items_queued >= XLOG_RECOVER_COMMIT_QUEUE_MAX) {
@@ -4152,9 +2054,9 @@ STATIC void
 xlog_recover_add_item(
 	struct list_head	*head)
 {
-	xlog_recover_item_t	*item;
+	struct xlog_recover_item *item;
 
-	item = kmem_zalloc(sizeof(xlog_recover_item_t), 0);
+	item = kmem_zalloc(sizeof(struct xlog_recover_item), 0);
 	INIT_LIST_HEAD(&item->ri_list);
 	list_add_tail(&item->ri_list, head);
 }
@@ -4166,7 +2068,7 @@ xlog_recover_add_to_cont_trans(
 	char			*dp,
 	int			len)
 {
-	xlog_recover_item_t	*item;
+	struct xlog_recover_item *item;
 	char			*ptr, *old_ptr;
 	int			old_len;
 
@@ -4189,7 +2091,8 @@ xlog_recover_add_to_cont_trans(
 	}
 
 	/* take the tail entry */
-	item = list_entry(trans->r_itemq.prev, xlog_recover_item_t, ri_list);
+	item = list_entry(trans->r_itemq.prev, struct xlog_recover_item,
+			  ri_list);
 
 	old_ptr = item->ri_buf[item->ri_cnt-1].i_addr;
 	old_len = item->ri_buf[item->ri_cnt-1].i_len;
@@ -4223,7 +2126,7 @@ xlog_recover_add_to_trans(
 	int			len)
 {
 	struct xfs_inode_log_format	*in_f;			/* any will do */
-	xlog_recover_item_t	*item;
+	struct xlog_recover_item *item;
 	char			*ptr;
 
 	if (!len)
@@ -4259,13 +2162,14 @@ xlog_recover_add_to_trans(
 	in_f = (struct xfs_inode_log_format *)ptr;
 
 	/* take the tail entry */
-	item = list_entry(trans->r_itemq.prev, xlog_recover_item_t, ri_list);
+	item = list_entry(trans->r_itemq.prev, struct xlog_recover_item,
+			  ri_list);
 	if (item->ri_total != 0 &&
 	     item->ri_total == item->ri_cnt) {
 		/* tail item is in use, get a new one */
 		xlog_recover_add_item(&trans->r_itemq);
 		item = list_entry(trans->r_itemq.prev,
-					xlog_recover_item_t, ri_list);
+					struct xlog_recover_item, ri_list);
 	}
 
 	if (item->ri_total == 0) {		/* first region to be added */
@@ -4311,7 +2215,7 @@ STATIC void
 xlog_recover_free_trans(
 	struct xlog_recover	*trans)
 {
-	xlog_recover_item_t	*item, *n;
+	struct xlog_recover_item *item, *n;
 	int			i;
 
 	hlist_del_init(&trans->r_list);
@@ -4563,180 +2467,6 @@ xlog_recover_process_data(
 	return 0;
 }
 
-/* Recover the EFI if necessary. */
-STATIC int
-xlog_recover_process_efi(
-	struct xfs_mount		*mp,
-	struct xfs_ail			*ailp,
-	struct xfs_log_item		*lip)
-{
-	struct xfs_efi_log_item		*efip;
-	int				error;
-
-	/*
-	 * Skip EFIs that we've already processed.
-	 */
-	efip = container_of(lip, struct xfs_efi_log_item, efi_item);
-	if (test_bit(XFS_EFI_RECOVERED, &efip->efi_flags))
-		return 0;
-
-	spin_unlock(&ailp->ail_lock);
-	error = xfs_efi_recover(mp, efip);
-	spin_lock(&ailp->ail_lock);
-
-	return error;
-}
-
-/* Release the EFI since we're cancelling everything. */
-STATIC void
-xlog_recover_cancel_efi(
-	struct xfs_mount		*mp,
-	struct xfs_ail			*ailp,
-	struct xfs_log_item		*lip)
-{
-	struct xfs_efi_log_item		*efip;
-
-	efip = container_of(lip, struct xfs_efi_log_item, efi_item);
-
-	spin_unlock(&ailp->ail_lock);
-	xfs_efi_release(efip);
-	spin_lock(&ailp->ail_lock);
-}
-
-/* Recover the RUI if necessary. */
-STATIC int
-xlog_recover_process_rui(
-	struct xfs_mount		*mp,
-	struct xfs_ail			*ailp,
-	struct xfs_log_item		*lip)
-{
-	struct xfs_rui_log_item		*ruip;
-	int				error;
-
-	/*
-	 * Skip RUIs that we've already processed.
-	 */
-	ruip = container_of(lip, struct xfs_rui_log_item, rui_item);
-	if (test_bit(XFS_RUI_RECOVERED, &ruip->rui_flags))
-		return 0;
-
-	spin_unlock(&ailp->ail_lock);
-	error = xfs_rui_recover(mp, ruip);
-	spin_lock(&ailp->ail_lock);
-
-	return error;
-}
-
-/* Release the RUI since we're cancelling everything. */
-STATIC void
-xlog_recover_cancel_rui(
-	struct xfs_mount		*mp,
-	struct xfs_ail			*ailp,
-	struct xfs_log_item		*lip)
-{
-	struct xfs_rui_log_item		*ruip;
-
-	ruip = container_of(lip, struct xfs_rui_log_item, rui_item);
-
-	spin_unlock(&ailp->ail_lock);
-	xfs_rui_release(ruip);
-	spin_lock(&ailp->ail_lock);
-}
-
-/* Recover the CUI if necessary. */
-STATIC int
-xlog_recover_process_cui(
-	struct xfs_trans		*parent_tp,
-	struct xfs_ail			*ailp,
-	struct xfs_log_item		*lip)
-{
-	struct xfs_cui_log_item		*cuip;
-	int				error;
-
-	/*
-	 * Skip CUIs that we've already processed.
-	 */
-	cuip = container_of(lip, struct xfs_cui_log_item, cui_item);
-	if (test_bit(XFS_CUI_RECOVERED, &cuip->cui_flags))
-		return 0;
-
-	spin_unlock(&ailp->ail_lock);
-	error = xfs_cui_recover(parent_tp, cuip);
-	spin_lock(&ailp->ail_lock);
-
-	return error;
-}
-
-/* Release the CUI since we're cancelling everything. */
-STATIC void
-xlog_recover_cancel_cui(
-	struct xfs_mount		*mp,
-	struct xfs_ail			*ailp,
-	struct xfs_log_item		*lip)
-{
-	struct xfs_cui_log_item		*cuip;
-
-	cuip = container_of(lip, struct xfs_cui_log_item, cui_item);
-
-	spin_unlock(&ailp->ail_lock);
-	xfs_cui_release(cuip);
-	spin_lock(&ailp->ail_lock);
-}
-
-/* Recover the BUI if necessary. */
-STATIC int
-xlog_recover_process_bui(
-	struct xfs_trans		*parent_tp,
-	struct xfs_ail			*ailp,
-	struct xfs_log_item		*lip)
-{
-	struct xfs_bui_log_item		*buip;
-	int				error;
-
-	/*
-	 * Skip BUIs that we've already processed.
-	 */
-	buip = container_of(lip, struct xfs_bui_log_item, bui_item);
-	if (test_bit(XFS_BUI_RECOVERED, &buip->bui_flags))
-		return 0;
-
-	spin_unlock(&ailp->ail_lock);
-	error = xfs_bui_recover(parent_tp, buip);
-	spin_lock(&ailp->ail_lock);
-
-	return error;
-}
-
-/* Release the BUI since we're cancelling everything. */
-STATIC void
-xlog_recover_cancel_bui(
-	struct xfs_mount		*mp,
-	struct xfs_ail			*ailp,
-	struct xfs_log_item		*lip)
-{
-	struct xfs_bui_log_item		*buip;
-
-	buip = container_of(lip, struct xfs_bui_log_item, bui_item);
-
-	spin_unlock(&ailp->ail_lock);
-	xfs_bui_release(buip);
-	spin_lock(&ailp->ail_lock);
-}
-
-/* Is this log item a deferred action intent? */
-static inline bool xlog_item_is_intent(struct xfs_log_item *lip)
-{
-	switch (lip->li_type) {
-	case XFS_LI_EFI:
-	case XFS_LI_RUI:
-	case XFS_LI_CUI:
-	case XFS_LI_BUI:
-		return true;
-	default:
-		return false;
-	}
-}
-
 /* Take all the collected deferred ops and finish them in order. */
 static int
 xlog_finish_defer_ops(
@@ -4771,6 +2501,13 @@ xlog_finish_defer_ops(
 	return xfs_trans_commit(tp);
 }
 
+/* Is this log item a deferred action intent? */
+static inline bool xlog_item_is_intent(struct xfs_log_item *lip)
+{
+	return lip->li_ops->iop_recover != NULL &&
+	       lip->li_ops->iop_match != NULL;
+}
+
 /*
  * When this is called, all of the log intent items which did not have
  * corresponding log done items should be in the AIL.  What we do now
@@ -4841,23 +2578,14 @@ xlog_recover_process_intents(
 
 		/*
 		 * NOTE: If your intent processing routine can create more
-		 * deferred ops, you /must/ attach them to the dfops in this
-		 * routine or else those subsequent intents will get
+		 * deferred ops, you /must/ attach them to the transaction in
+		 * this routine or else those subsequent intents will get
 		 * replayed in the wrong order!
 		 */
-		switch (lip->li_type) {
-		case XFS_LI_EFI:
-			error = xlog_recover_process_efi(log->l_mp, ailp, lip);
-			break;
-		case XFS_LI_RUI:
-			error = xlog_recover_process_rui(log->l_mp, ailp, lip);
-			break;
-		case XFS_LI_CUI:
-			error = xlog_recover_process_cui(parent_tp, ailp, lip);
-			break;
-		case XFS_LI_BUI:
-			error = xlog_recover_process_bui(parent_tp, ailp, lip);
-			break;
+		if (!test_and_set_bit(XFS_LI_RECOVERED, &lip->li_flags)) {
+			spin_unlock(&ailp->ail_lock);
+			error = lip->li_ops->iop_recover(lip, parent_tp);
+			spin_lock(&ailp->ail_lock);
 		}
 		if (error)
 			goto out;
@@ -4901,21 +2629,9 @@ xlog_recover_cancel_intents(
 			break;
 		}
 
-		switch (lip->li_type) {
-		case XFS_LI_EFI:
-			xlog_recover_cancel_efi(log->l_mp, ailp, lip);
-			break;
-		case XFS_LI_RUI:
-			xlog_recover_cancel_rui(log->l_mp, ailp, lip);
-			break;
-		case XFS_LI_CUI:
-			xlog_recover_cancel_cui(log->l_mp, ailp, lip);
-			break;
-		case XFS_LI_BUI:
-			xlog_recover_cancel_bui(log->l_mp, ailp, lip);
-			break;
-		}
-
+		spin_unlock(&ailp->ail_lock);
+		lip->li_ops->iop_release(lip);
+		spin_lock(&ailp->ail_lock);
 		lip = xfs_trans_ail_cursor_next(ailp, &cur);
 	}
 
@@ -4947,7 +2663,7 @@ xlog_recover_clear_agi_bucket(
 	if (error)
 		goto out_abort;
 
-	agi = XFS_BUF_TO_AGI(agibp);
+	agi = agibp->b_addr;
 	agi->agi_unlinked[bucket] = cpu_to_be32(NULLAGINO);
 	offset = offsetof(xfs_agi_t, agi_unlinked) +
 		 (sizeof(xfs_agino_t) * bucket);
@@ -4987,7 +2703,7 @@ xlog_recover_process_one_iunlink(
 	/*
 	 * Get the on disk inode to find the next inode in the bucket.
 	 */
-	error = xfs_imap_to_bp(mp, NULL, &ip->i_imap, &dip, &ibp, 0, 0);
+	error = xfs_imap_to_bp(mp, NULL, &ip->i_imap, &dip, &ibp, 0);
 	if (error)
 		goto fail_iput;
 
@@ -5083,7 +2799,7 @@ xlog_recover_process_iunlinks(
 		 * buffer reference though, so that it stays pinned in memory
 		 * while we need the buffer.
 		 */
-		agi = XFS_BUF_TO_AGI(agibp);
+		agi = agibp->b_addr;
 		xfs_buf_unlock(agibp);
 
 		for (bucket = 0; bucket < XFS_AGI_UNLINKED_BUCKETS; bucket++) {
@@ -5636,7 +3352,7 @@ xlog_do_recover(
 
 	/* Convert superblock from on-disk format */
 	sbp = &mp->m_sb;
-	xfs_sb_from_disk(sbp, XFS_BUF_TO_SBP(bp));
+	xfs_sb_from_disk(sbp, bp->b_addr);
 	xfs_buf_relse(bp);
 
 	/* re-initialise in-core superblock and geometry structures */
@@ -5809,7 +3525,6 @@ xlog_recover_check_summary(
 	struct xlog	*log)
 {
 	xfs_mount_t	*mp;
-	xfs_agf_t	*agfp;
 	xfs_buf_t	*agfbp;
 	xfs_buf_t	*agibp;
 	xfs_agnumber_t	agno;
@@ -5829,7 +3544,8 @@ xlog_recover_check_summary(
 			xfs_alert(mp, "%s agf read failed agno %d error %d",
 						__func__, agno, error);
 		} else {
-			agfp = XFS_BUF_TO_AGF(agfbp);
+			struct xfs_agf	*agfp = agfbp->b_addr;
+
 			freeblks += be32_to_cpu(agfp->agf_freeblks) +
 				    be32_to_cpu(agfp->agf_flcount);
 			xfs_buf_relse(agfbp);
@@ -5840,7 +3556,7 @@ xlog_recover_check_summary(
 			xfs_alert(mp, "%s agi read failed agno %d error %d",
 						__func__, agno, error);
 		} else {
-			struct xfs_agi	*agi = XFS_BUF_TO_AGI(agibp);
+			struct xfs_agi	*agi = agibp->b_addr;
 
 			itotal += be32_to_cpu(agi->agi_count);
 			ifree += be32_to_cpu(agi->agi_freecount);
diff --git a/fs/xfs/xfs_message.c b/fs/xfs/xfs_message.c
index e0f9d3b6abe9..bc66d95c8d4c 100644
--- a/fs/xfs/xfs_message.c
+++ b/fs/xfs/xfs_message.c
@@ -117,3 +117,25 @@ xfs_hex_dump(const void *p, int length)
 {
 	print_hex_dump(KERN_ALERT, "", DUMP_PREFIX_OFFSET, 16, 1, p, length, 1);
 }
+
+void
+xfs_buf_alert_ratelimited(
+	struct xfs_buf		*bp,
+	const char		*rlmsg,
+	const char		*fmt,
+	...)
+{
+	struct xfs_mount	*mp = bp->b_mount;
+	struct va_format	vaf;
+	va_list			args;
+
+	/* use the more aggressive per-target rate limit for buffers */
+	if (!___ratelimit(&bp->b_target->bt_ioerror_rl, rlmsg))
+		return;
+
+	va_start(args, fmt);
+	vaf.fmt = fmt;
+	vaf.va = &args;
+	__xfs_printk(KERN_ALERT, mp, &vaf);
+	va_end(args);
+}
diff --git a/fs/xfs/xfs_message.h b/fs/xfs/xfs_message.h
index 0b05e10995a0..4d9bd6bb63ca 100644
--- a/fs/xfs/xfs_message.h
+++ b/fs/xfs/xfs_message.h
@@ -31,15 +31,27 @@ void xfs_debug(const struct xfs_mount *mp, const char *fmt, ...)
 }
 #endif
 
-#define xfs_printk_ratelimited(func, dev, fmt, ...)		\
+#define xfs_printk_ratelimited(func, dev, fmt, ...)			\
 do {									\
 	static DEFINE_RATELIMIT_STATE(_rs,				\
 				      DEFAULT_RATELIMIT_INTERVAL,	\
 				      DEFAULT_RATELIMIT_BURST);		\
 	if (__ratelimit(&_rs))						\
-		func(dev, fmt, ##__VA_ARGS__);			\
+		func(dev, fmt, ##__VA_ARGS__);				\
 } while (0)
 
+#define xfs_printk_once(func, dev, fmt, ...)			\
+({								\
+	static bool __section(.data.once) __print_once;		\
+	bool __ret_print_once = !__print_once; 			\
+								\
+	if (!__print_once) {					\
+		__print_once = true;				\
+		func(dev, fmt, ##__VA_ARGS__);			\
+	}							\
+	unlikely(__ret_print_once);				\
+})
+
 #define xfs_emerg_ratelimited(dev, fmt, ...)				\
 	xfs_printk_ratelimited(xfs_emerg, dev, fmt, ##__VA_ARGS__)
 #define xfs_alert_ratelimited(dev, fmt, ...)				\
@@ -57,9 +69,17 @@ do {									\
 #define xfs_debug_ratelimited(dev, fmt, ...)				\
 	xfs_printk_ratelimited(xfs_debug, dev, fmt, ##__VA_ARGS__)
 
+#define xfs_warn_once(dev, fmt, ...)				\
+	xfs_printk_once(xfs_warn, dev, fmt, ##__VA_ARGS__)
+#define xfs_notice_once(dev, fmt, ...)				\
+	xfs_printk_once(xfs_notice, dev, fmt, ##__VA_ARGS__)
+
 void assfail(struct xfs_mount *mp, char *expr, char *f, int l);
 void asswarn(struct xfs_mount *mp, char *expr, char *f, int l);
 
 extern void xfs_hex_dump(const void *p, int length);
 
+void xfs_buf_alert_ratelimited(struct xfs_buf *bp, const char *rlmsg,
+			       const char *fmt, ...);
+
 #endif	/* __XFS_MESSAGE_H */
diff --git a/fs/xfs/xfs_mount.c b/fs/xfs/xfs_mount.c
index 56efe140c923..c8ae49a1e99c 100644
--- a/fs/xfs/xfs_mount.c
+++ b/fs/xfs/xfs_mount.c
@@ -148,7 +148,6 @@ xfs_free_perag(
 		ASSERT(atomic_read(&pag->pag_ref) == 0);
 		xfs_iunlink_destroy(pag);
 		xfs_buf_hash_destroy(pag);
-		mutex_destroy(&pag->pag_ici_reclaim_lock);
 		call_rcu(&pag->rcu_head, __xfs_free_perag);
 	}
 }
@@ -200,7 +199,6 @@ xfs_initialize_perag(
 		pag->pag_agno = index;
 		pag->pag_mount = mp;
 		spin_lock_init(&pag->pag_ici_lock);
-		mutex_init(&pag->pag_ici_reclaim_lock);
 		INIT_RADIX_TREE(&pag->pag_ici_root, GFP_ATOMIC);
 		if (xfs_buf_hash_init(pag))
 			goto out_free_pag;
@@ -242,7 +240,6 @@ xfs_initialize_perag(
 out_hash_destroy:
 	xfs_buf_hash_destroy(pag);
 out_free_pag:
-	mutex_destroy(&pag->pag_ici_reclaim_lock);
 	kmem_free(pag);
 out_unwind_new_pags:
 	/* unwind any prior newly initialized pags */
@@ -252,7 +249,6 @@ out_unwind_new_pags:
 			break;
 		xfs_buf_hash_destroy(pag);
 		xfs_iunlink_destroy(pag);
-		mutex_destroy(&pag->pag_ici_reclaim_lock);
 		kmem_free(pag);
 	}
 	return error;
@@ -310,7 +306,7 @@ reread:
 	/*
 	 * Initialize the mount structure from the superblock.
 	 */
-	xfs_sb_from_disk(sbp, XFS_BUF_TO_SBP(bp));
+	xfs_sb_from_disk(sbp, bp->b_addr);
 
 	/*
 	 * If we haven't validated the superblock, do so now before we try
@@ -1015,7 +1011,7 @@ xfs_mountfs(
 	 * quota inodes.
 	 */
 	cancel_delayed_work_sync(&mp->m_reclaim_work);
-	xfs_reclaim_inodes(mp, SYNC_WAIT);
+	xfs_reclaim_inodes(mp);
 	xfs_health_unmount(mp);
  out_log_dealloc:
 	mp->m_flags |= XFS_MOUNT_UNMOUNTING;
@@ -1092,13 +1088,12 @@ xfs_unmountfs(
 	xfs_ail_push_all_sync(mp->m_ail);
 
 	/*
-	 * And reclaim all inodes.  At this point there should be no dirty
-	 * inodes and none should be pinned or locked, but use synchronous
-	 * reclaim just to be sure. We can stop background inode reclaim
-	 * here as well if it is still running.
+	 * Reclaim all inodes. At this point there should be no dirty inodes and
+	 * none should be pinned or locked. Stop background inode reclaim here
+	 * if it is still running.
 	 */
 	cancel_delayed_work_sync(&mp->m_reclaim_work);
-	xfs_reclaim_inodes(mp, SYNC_WAIT);
+	xfs_reclaim_inodes(mp);
 	xfs_health_unmount(mp);
 
 	xfs_qm_unmount(mp);
@@ -1190,39 +1185,6 @@ xfs_log_sbcount(xfs_mount_t *mp)
 }
 
 /*
- * Deltas for the inode count are +/-64, hence we use a large batch size
- * of 128 so we don't need to take the counter lock on every update.
- */
-#define XFS_ICOUNT_BATCH	128
-int
-xfs_mod_icount(
-	struct xfs_mount	*mp,
-	int64_t			delta)
-{
-	percpu_counter_add_batch(&mp->m_icount, delta, XFS_ICOUNT_BATCH);
-	if (__percpu_counter_compare(&mp->m_icount, 0, XFS_ICOUNT_BATCH) < 0) {
-		ASSERT(0);
-		percpu_counter_add(&mp->m_icount, -delta);
-		return -EINVAL;
-	}
-	return 0;
-}
-
-int
-xfs_mod_ifree(
-	struct xfs_mount	*mp,
-	int64_t			delta)
-{
-	percpu_counter_add(&mp->m_ifree, delta);
-	if (percpu_counter_compare(&mp->m_ifree, 0) < 0) {
-		ASSERT(0);
-		percpu_counter_add(&mp->m_ifree, -delta);
-		return -EINVAL;
-	}
-	return 0;
-}
-
-/*
  * Deltas for the block count can vary from 1 to very large, but lock contention
  * only occurs on frequent small block count updates such as in the delayed
  * allocation path for buffered writes (page a time updates). Hence we set
@@ -1300,10 +1262,9 @@ xfs_mod_fdblocks(
 		spin_unlock(&mp->m_sb_lock);
 		return 0;
 	}
-	printk_once(KERN_WARNING
-		"Filesystem \"%s\": reserve blocks depleted! "
-		"Consider increasing reserve pool size.",
-		mp->m_super->s_id);
+	xfs_warn_once(mp,
+"Reserve blocks depleted! Consider increasing reserve pool size.");
+
 fdblocks_enospc:
 	spin_unlock(&mp->m_sb_lock);
 	return -ENOSPC;
diff --git a/fs/xfs/xfs_mount.h b/fs/xfs/xfs_mount.h
index 88ab09ed29e7..a72cfcaa4ad1 100644
--- a/fs/xfs/xfs_mount.h
+++ b/fs/xfs/xfs_mount.h
@@ -55,61 +55,25 @@ struct xfs_error_cfg {
 	long		retry_timeout;	/* in jiffies, -1 = infinite */
 };
 
+/*
+ * The struct xfsmount layout is optimised to separate read-mostly variables
+ * from variables that are frequently modified. We put the read-mostly variables
+ * first, then place all the other variables at the end.
+ *
+ * Typically, read-mostly variables are those that are set at mount time and
+ * never changed again, or only change rarely as a result of things like sysfs
+ * knobs being tweaked.
+ */
 typedef struct xfs_mount {
+	struct xfs_sb		m_sb;		/* copy of fs superblock */
 	struct super_block	*m_super;
-
-	/*
-	 * Bitsets of per-fs metadata that have been checked and/or are sick.
-	 * Callers must hold m_sb_lock to access these two fields.
-	 */
-	uint8_t			m_fs_checked;
-	uint8_t			m_fs_sick;
-	/*
-	 * Bitsets of rt metadata that have been checked and/or are sick.
-	 * Callers must hold m_sb_lock to access this field.
-	 */
-	uint8_t			m_rt_checked;
-	uint8_t			m_rt_sick;
-
 	struct xfs_ail		*m_ail;		/* fs active log item list */
-
-	struct xfs_sb		m_sb;		/* copy of fs superblock */
-	spinlock_t		m_sb_lock;	/* sb counter lock */
-	struct percpu_counter	m_icount;	/* allocated inodes counter */
-	struct percpu_counter	m_ifree;	/* free inodes counter */
-	struct percpu_counter	m_fdblocks;	/* free block counter */
-	/*
-	 * Count of data device blocks reserved for delayed allocations,
-	 * including indlen blocks.  Does not include allocated CoW staging
-	 * extents or anything related to the rt device.
-	 */
-	struct percpu_counter	m_delalloc_blks;
-
 	struct xfs_buf		*m_sb_bp;	/* buffer for superblock */
 	char			*m_rtname;	/* realtime device name */
 	char			*m_logname;	/* external log device name */
-	int			m_bsize;	/* fs logical block size */
-	xfs_agnumber_t		m_agfrotor;	/* last ag where space found */
-	xfs_agnumber_t		m_agirotor;	/* last ag dir inode alloced */
-	spinlock_t		m_agirotor_lock;/* .. and lock protecting it */
-	xfs_agnumber_t		m_maxagi;	/* highest inode alloc group */
-	uint			m_allocsize_log;/* min write size log bytes */
-	uint			m_allocsize_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 */
-	struct xfs_ino_geometry	m_ino_geo;	/* inode geometry */
-	int			m_logbufs;	/* number of log buffers */
-	int			m_logbsize;	/* size of each log buffer */
-	uint			m_rsumlevels;	/* rt summary levels */
-	uint			m_rsumsize;	/* size of rt summary, bytes */
-	/*
-	 * Optional cache of rt summary level per bitmap block with the
-	 * invariant that m_rsum_cache[bbno] <= the minimum i for which
-	 * rsum[i][bbno] != 0. Reads and writes are serialized by the rsumip
-	 * inode lock.
-	 */
-	uint8_t			*m_rsum_cache;
 	struct xfs_inode	*m_rbmip;	/* pointer to bitmap inode */
 	struct xfs_inode	*m_rsumip;	/* pointer to summary inode */
 	struct xfs_inode	*m_rootip;	/* pointer to root directory */
@@ -117,9 +81,26 @@ typedef struct xfs_mount {
 	xfs_buftarg_t		*m_ddev_targp;	/* saves taking the address */
 	xfs_buftarg_t		*m_logdev_targp;/* ptr to log device */
 	xfs_buftarg_t		*m_rtdev_targp;	/* ptr to rt device */
+	/*
+	 * Optional cache of rt summary level per bitmap block with the
+	 * invariant that m_rsum_cache[bbno] <= the minimum i for which
+	 * rsum[i][bbno] != 0. Reads and writes are serialized by the rsumip
+	 * inode lock.
+	 */
+	uint8_t			*m_rsum_cache;
+	struct xfs_mru_cache	*m_filestream;  /* per-mount filestream data */
+	struct workqueue_struct *m_buf_workqueue;
+	struct workqueue_struct	*m_unwritten_workqueue;
+	struct workqueue_struct	*m_cil_workqueue;
+	struct workqueue_struct	*m_reclaim_workqueue;
+	struct workqueue_struct *m_eofblocks_workqueue;
+	struct workqueue_struct	*m_sync_workqueue;
+
+	int			m_bsize;	/* fs logical block size */
 	uint8_t			m_blkbit_log;	/* blocklog + NBBY */
 	uint8_t			m_blkbb_log;	/* blocklog - BBSHIFT */
 	uint8_t			m_agno_log;	/* log #ag's */
+	uint8_t			m_sectbb_log;	/* sectlog - BBSHIFT */
 	uint			m_blockmask;	/* sb_blocksize-1 */
 	uint			m_blockwsize;	/* sb_blocksize in words */
 	uint			m_blockwmask;	/* blockwsize-1 */
@@ -138,42 +119,82 @@ typedef struct xfs_mount {
 	xfs_extlen_t		m_ag_prealloc_blocks; /* reserved ag blocks */
 	uint			m_alloc_set_aside; /* space we can't use */
 	uint			m_ag_max_usable; /* max space per AG */
-	struct radix_tree_root	m_perag_tree;	/* per-ag accounting info */
-	spinlock_t		m_perag_lock;	/* lock for m_perag_tree */
-	struct mutex		m_growlock;	/* growfs mutex */
+	int			m_dalign;	/* stripe unit */
+	int			m_swidth;	/* stripe width */
+	xfs_agnumber_t		m_maxagi;	/* highest inode alloc group */
+	uint			m_allocsize_log;/* min write size log bytes */
+	uint			m_allocsize_blocks; /* min write size blocks */
+	int			m_logbufs;	/* number of log buffers */
+	int			m_logbsize;	/* size of each log buffer */
+	uint			m_rsumlevels;	/* rt summary levels */
+	uint			m_rsumsize;	/* size of rt summary, bytes */
 	int			m_fixedfsid[2];	/* unchanged for life of FS */
-	uint64_t		m_flags;	/* global mount flags */
-	bool			m_finobt_nores; /* no per-AG finobt resv. */
 	uint			m_qflags;	/* quota status flags */
+	uint64_t		m_flags;	/* global mount flags */
+	int64_t			m_low_space[XFS_LOWSP_MAX];
+	struct xfs_ino_geometry	m_ino_geo;	/* inode geometry */
 	struct xfs_trans_resv	m_resv;		/* precomputed res values */
+						/* low free space thresholds */
+	bool			m_always_cow;
+	bool			m_fail_unmount;
+	bool			m_finobt_nores; /* no per-AG finobt resv. */
+	bool			m_update_sb;	/* sb needs update in mount */
+
+	/*
+	 * Bitsets of per-fs metadata that have been checked and/or are sick.
+	 * Callers must hold m_sb_lock to access these two fields.
+	 */
+	uint8_t			m_fs_checked;
+	uint8_t			m_fs_sick;
+	/*
+	 * Bitsets of rt metadata that have been checked and/or are sick.
+	 * Callers must hold m_sb_lock to access this field.
+	 */
+	uint8_t			m_rt_checked;
+	uint8_t			m_rt_sick;
+
+	/*
+	 * End of read-mostly variables. Frequently written variables and locks
+	 * should be placed below this comment from now on. The first variable
+	 * here is marked as cacheline aligned so they it is separated from
+	 * the read-mostly variables.
+	 */
+
+	spinlock_t ____cacheline_aligned m_sb_lock; /* sb counter lock */
+	struct percpu_counter	m_icount;	/* allocated inodes counter */
+	struct percpu_counter	m_ifree;	/* free inodes counter */
+	struct percpu_counter	m_fdblocks;	/* free block counter */
+	/*
+	 * Count of data device blocks reserved for delayed allocations,
+	 * including indlen blocks.  Does not include allocated CoW staging
+	 * extents or anything related to the rt device.
+	 */
+	struct percpu_counter	m_delalloc_blks;
+
+	struct radix_tree_root	m_perag_tree;	/* per-ag accounting info */
+	spinlock_t		m_perag_lock;	/* lock for m_perag_tree */
 	uint64_t		m_resblks;	/* total reserved blocks */
 	uint64_t		m_resblks_avail;/* available reserved blocks */
 	uint64_t		m_resblks_save;	/* reserved blks @ remount,ro */
-	int			m_dalign;	/* stripe unit */
-	int			m_swidth;	/* stripe width */
-	uint8_t			m_sectbb_log;	/* sectlog - BBSHIFT */
-	atomic_t		m_active_trans;	/* number trans frozen */
-	struct xfs_mru_cache	*m_filestream;  /* per-mount filestream data */
 	struct delayed_work	m_reclaim_work;	/* background inode reclaim */
 	struct delayed_work	m_eofblocks_work; /* background eof blocks
 						     trimming */
 	struct delayed_work	m_cowblocks_work; /* background cow blocks
 						     trimming */
-	bool			m_update_sb;	/* sb needs update in mount */
-	int64_t			m_low_space[XFS_LOWSP_MAX];
-						/* low free space thresholds */
 	struct xfs_kobj		m_kobj;
 	struct xfs_kobj		m_error_kobj;
 	struct xfs_kobj		m_error_meta_kobj;
 	struct xfs_error_cfg	m_error_cfg[XFS_ERR_CLASS_MAX][XFS_ERR_ERRNO_MAX];
 	struct xstats		m_stats;	/* per-fs stats */
+	xfs_agnumber_t		m_agfrotor;	/* last ag where space found */
+	xfs_agnumber_t		m_agirotor;	/* last ag dir inode alloced */
+	spinlock_t		m_agirotor_lock;/* .. and lock protecting it */
 
-	struct workqueue_struct *m_buf_workqueue;
-	struct workqueue_struct	*m_unwritten_workqueue;
-	struct workqueue_struct	*m_cil_workqueue;
-	struct workqueue_struct	*m_reclaim_workqueue;
-	struct workqueue_struct *m_eofblocks_workqueue;
-	struct workqueue_struct	*m_sync_workqueue;
+	/*
+	 * Workqueue item so that we can coalesce multiple inode flush attempts
+	 * into a single flush.
+	 */
+	struct work_struct	m_flush_inodes_work;
 
 	/*
 	 * Generation of the filesysyem layout.  This is incremented by each
@@ -185,9 +206,8 @@ typedef struct xfs_mount {
 	 * to various other kinds of pain inflicted on the pNFS server.
 	 */
 	uint32_t		m_generation;
+	struct mutex		m_growlock;	/* growfs mutex */
 
-	bool			m_always_cow;
-	bool			m_fail_unmount;
 #ifdef DEBUG
 	/*
 	 * Frequency with which errors are injected.  Replaces xfs_etest; the
@@ -232,8 +252,8 @@ typedef struct xfs_mount {
 #define XFS_MOUNT_FILESTREAMS	(1ULL << 24)	/* enable the filestreams
 						   allocator */
 #define XFS_MOUNT_NOATTR2	(1ULL << 25)	/* disable use of attr2 format */
-
-#define XFS_MOUNT_DAX		(1ULL << 62)	/* TEST ONLY! */
+#define XFS_MOUNT_DAX_ALWAYS	(1ULL << 26)
+#define XFS_MOUNT_DAX_NEVER	(1ULL << 27)
 
 /*
  * Max and min values for mount-option defined I/O
@@ -254,8 +274,6 @@ void xfs_do_force_shutdown(struct xfs_mount *mp, int flags, char *fname,
 #define SHUTDOWN_LOG_IO_ERROR	0x0002	/* write attempt to the log failed */
 #define SHUTDOWN_FORCE_UMOUNT	0x0004	/* shutdown from a forced unmount */
 #define SHUTDOWN_CORRUPT_INCORE	0x0008	/* corrupt in-memory data structures */
-#define SHUTDOWN_REMOTE_REQ	0x0010	/* shutdown came from remote cell */
-#define SHUTDOWN_DEVICE_REQ	0x0020	/* failed all paths to the device */
 
 /*
  * Flags for xfs_mountfs
@@ -336,7 +354,6 @@ typedef struct xfs_perag {
 	spinlock_t	pag_ici_lock;	/* incore inode cache lock */
 	struct radix_tree_root pag_ici_root;	/* incore inode cache root */
 	int		pag_ici_reclaimable;	/* reclaimable inodes */
-	struct mutex	pag_ici_reclaim_lock;	/* serialisation point */
 	unsigned long	pag_ici_reclaim_cursor;	/* reclaim restart point */
 
 	/* buffer cache index */
@@ -389,8 +406,6 @@ extern int	xfs_initialize_perag(xfs_mount_t *mp, xfs_agnumber_t agcount,
 				     xfs_agnumber_t *maxagi);
 extern void	xfs_unmountfs(xfs_mount_t *);
 
-extern int	xfs_mod_icount(struct xfs_mount *mp, int64_t delta);
-extern int	xfs_mod_ifree(struct xfs_mount *mp, int64_t delta);
 extern int	xfs_mod_fdblocks(struct xfs_mount *mp, int64_t delta,
 				 bool reserved);
 extern int	xfs_mod_frextents(struct xfs_mount *mp, int64_t delta);
diff --git a/fs/xfs/xfs_pnfs.c b/fs/xfs/xfs_pnfs.c
index bb3008d390aa..b101feb2aab4 100644
--- a/fs/xfs/xfs_pnfs.c
+++ b/fs/xfs/xfs_pnfs.c
@@ -58,9 +58,8 @@ xfs_fs_get_uuid(
 {
 	struct xfs_mount	*mp = XFS_M(sb);
 
-	printk_once(KERN_NOTICE
-"XFS (%s): using experimental pNFS feature, use at your own risk!\n",
-		mp->m_super->s_id);
+	xfs_notice_once(mp,
+"Using experimental pNFS feature, use at your own risk!");
 
 	if (*len < sizeof(uuid_t))
 		return -EINVAL;
diff --git a/fs/xfs/xfs_pwork.c b/fs/xfs/xfs_pwork.c
index 4bcc3e61056c..b03333f1c84a 100644
--- a/fs/xfs/xfs_pwork.c
+++ b/fs/xfs/xfs_pwork.c
@@ -132,5 +132,5 @@ xfs_pwork_guess_datadev_parallelism(
 	 * For now we'll go with the most conservative setting possible,
 	 * which is two threads for an SSD and 1 thread everywhere else.
 	 */
-	return blk_queue_nonrot(btp->bt_bdev->bd_queue) ? 2 : 1;
+	return blk_queue_nonrot(btp->bt_bdev->bd_disk->queue) ? 2 : 1;
 }
diff --git a/fs/xfs/xfs_qm.c b/fs/xfs/xfs_qm.c
index 0b0909657bad..be67570badf8 100644
--- a/fs/xfs/xfs_qm.c
+++ b/fs/xfs/xfs_qm.c
@@ -47,7 +47,7 @@ STATIC void	xfs_qm_dqfree_one(struct xfs_dquot *dqp);
 STATIC int
 xfs_qm_dquot_walk(
 	struct xfs_mount	*mp,
-	int			type,
+	xfs_dqtype_t		type,
 	int			(*execute)(struct xfs_dquot *dqp, void *data),
 	void			*data)
 {
@@ -79,7 +79,7 @@ restart:
 		for (i = 0; i < nr_found; i++) {
 			struct xfs_dquot *dqp = batch[i];
 
-			next_index = be32_to_cpu(dqp->q_core.d_id) + 1;
+			next_index = dqp->q_id + 1;
 
 			error = execute(batch[i], data);
 			if (error == -EAGAIN) {
@@ -121,14 +121,13 @@ xfs_qm_dqpurge(
 {
 	struct xfs_mount	*mp = dqp->q_mount;
 	struct xfs_quotainfo	*qi = mp->m_quotainfo;
+	int			error = -EAGAIN;
 
 	xfs_dqlock(dqp);
-	if ((dqp->dq_flags & XFS_DQ_FREEING) || dqp->q_nrefs != 0) {
-		xfs_dqunlock(dqp);
-		return -EAGAIN;
-	}
+	if ((dqp->q_flags & XFS_DQFLAG_FREEING) || dqp->q_nrefs != 0)
+		goto out_unlock;
 
-	dqp->dq_flags |= XFS_DQ_FREEING;
+	dqp->q_flags |= XFS_DQFLAG_FREEING;
 
 	xfs_dqflock(dqp);
 
@@ -139,7 +138,6 @@ xfs_qm_dqpurge(
 	 */
 	if (XFS_DQ_IS_DIRTY(dqp)) {
 		struct xfs_buf	*bp = NULL;
-		int		error;
 
 		/*
 		 * We don't care about getting disk errors here. We need
@@ -149,6 +147,9 @@ xfs_qm_dqpurge(
 		if (!error) {
 			error = xfs_bwrite(bp);
 			xfs_buf_relse(bp);
+		} else if (error == -EAGAIN) {
+			dqp->q_flags &= ~XFS_DQFLAG_FREEING;
+			goto out_unlock;
 		}
 		xfs_dqflock(dqp);
 	}
@@ -160,8 +161,7 @@ xfs_qm_dqpurge(
 	xfs_dqfunlock(dqp);
 	xfs_dqunlock(dqp);
 
-	radix_tree_delete(xfs_dquot_tree(qi, dqp->q_core.d_flags),
-			  be32_to_cpu(dqp->q_core.d_id));
+	radix_tree_delete(xfs_dquot_tree(qi, xfs_dquot_type(dqp)), dqp->q_id);
 	qi->qi_dquots--;
 
 	/*
@@ -174,6 +174,10 @@ xfs_qm_dqpurge(
 
 	xfs_qm_dqdestroy(dqp);
 	return 0;
+
+out_unlock:
+	xfs_dqunlock(dqp);
+	return error;
 }
 
 /*
@@ -185,11 +189,11 @@ xfs_qm_dqpurge_all(
 	uint			flags)
 {
 	if (flags & XFS_QMOPT_UQUOTA)
-		xfs_qm_dquot_walk(mp, XFS_DQ_USER, xfs_qm_dqpurge, NULL);
+		xfs_qm_dquot_walk(mp, XFS_DQTYPE_USER, xfs_qm_dqpurge, NULL);
 	if (flags & XFS_QMOPT_GQUOTA)
-		xfs_qm_dquot_walk(mp, XFS_DQ_GROUP, xfs_qm_dqpurge, NULL);
+		xfs_qm_dquot_walk(mp, XFS_DQTYPE_GROUP, xfs_qm_dqpurge, NULL);
 	if (flags & XFS_QMOPT_PQUOTA)
-		xfs_qm_dquot_walk(mp, XFS_DQ_PROJ, xfs_qm_dqpurge, NULL);
+		xfs_qm_dquot_walk(mp, XFS_DQTYPE_PROJ, xfs_qm_dqpurge, NULL);
 }
 
 /*
@@ -246,7 +250,7 @@ STATIC int
 xfs_qm_dqattach_one(
 	struct xfs_inode	*ip,
 	xfs_dqid_t		id,
-	uint			type,
+	xfs_dqtype_t		type,
 	bool			doalloc,
 	struct xfs_dquot	**IO_idqpp)
 {
@@ -326,23 +330,23 @@ xfs_qm_dqattach_locked(
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
 
 	if (XFS_IS_UQUOTA_ON(mp) && !ip->i_udquot) {
-		error = xfs_qm_dqattach_one(ip, ip->i_d.di_uid, XFS_DQ_USER,
-				doalloc, &ip->i_udquot);
+		error = xfs_qm_dqattach_one(ip, i_uid_read(VFS_I(ip)),
+				XFS_DQTYPE_USER, doalloc, &ip->i_udquot);
 		if (error)
 			goto done;
 		ASSERT(ip->i_udquot);
 	}
 
 	if (XFS_IS_GQUOTA_ON(mp) && !ip->i_gdquot) {
-		error = xfs_qm_dqattach_one(ip, ip->i_d.di_gid, XFS_DQ_GROUP,
-				doalloc, &ip->i_gdquot);
+		error = xfs_qm_dqattach_one(ip, i_gid_read(VFS_I(ip)),
+				XFS_DQTYPE_GROUP, doalloc, &ip->i_gdquot);
 		if (error)
 			goto done;
 		ASSERT(ip->i_gdquot);
 	}
 
 	if (XFS_IS_PQUOTA_ON(mp) && !ip->i_pdquot) {
-		error = xfs_qm_dqattach_one(ip, ip->i_d.di_projid, XFS_DQ_PROJ,
+		error = xfs_qm_dqattach_one(ip, ip->i_d.di_projid, XFS_DQTYPE_PROJ,
 				doalloc, &ip->i_pdquot);
 		if (error)
 			goto done;
@@ -469,7 +473,7 @@ xfs_qm_dquot_isolate(
 	/*
 	 * Prevent lookups now that we are past the point of no return.
 	 */
-	dqp->dq_flags |= XFS_DQ_FREEING;
+	dqp->q_flags |= XFS_DQFLAG_FREEING;
 	xfs_dqunlock(dqp);
 
 	ASSERT(dqp->q_nrefs == 0);
@@ -541,31 +545,29 @@ xfs_qm_shrink_count(
 STATIC void
 xfs_qm_set_defquota(
 	struct xfs_mount	*mp,
-	uint			type,
+	xfs_dqtype_t		type,
 	struct xfs_quotainfo	*qinf)
 {
 	struct xfs_dquot	*dqp;
 	struct xfs_def_quota	*defq;
-	struct xfs_disk_dquot	*ddqp;
 	int			error;
 
 	error = xfs_qm_dqget_uncached(mp, 0, type, &dqp);
 	if (error)
 		return;
 
-	ddqp = &dqp->q_core;
-	defq = xfs_get_defquota(dqp, qinf);
+	defq = xfs_get_defquota(qinf, xfs_dquot_type(dqp));
 
 	/*
 	 * Timers and warnings have been already set, let's just set the
 	 * default limits for this quota type
 	 */
-	defq->bhardlimit = be64_to_cpu(ddqp->d_blk_hardlimit);
-	defq->bsoftlimit = be64_to_cpu(ddqp->d_blk_softlimit);
-	defq->ihardlimit = be64_to_cpu(ddqp->d_ino_hardlimit);
-	defq->isoftlimit = be64_to_cpu(ddqp->d_ino_softlimit);
-	defq->rtbhardlimit = be64_to_cpu(ddqp->d_rtb_hardlimit);
-	defq->rtbsoftlimit = be64_to_cpu(ddqp->d_rtb_softlimit);
+	defq->blk.hard = dqp->q_blk.hardlimit;
+	defq->blk.soft = dqp->q_blk.softlimit;
+	defq->ino.hard = dqp->q_ino.hardlimit;
+	defq->ino.soft = dqp->q_ino.softlimit;
+	defq->rtb.hard = dqp->q_rtb.hardlimit;
+	defq->rtb.soft = dqp->q_rtb.softlimit;
 	xfs_qm_dqdestroy(dqp);
 }
 
@@ -573,19 +575,21 @@ xfs_qm_set_defquota(
 static void
 xfs_qm_init_timelimits(
 	struct xfs_mount	*mp,
-	struct xfs_quotainfo	*qinf)
+	xfs_dqtype_t		type)
 {
-	struct xfs_disk_dquot	*ddqp;
+	struct xfs_quotainfo	*qinf = mp->m_quotainfo;
+	struct xfs_def_quota	*defq;
 	struct xfs_dquot	*dqp;
-	uint			type;
 	int			error;
 
-	qinf->qi_btimelimit = XFS_QM_BTIMELIMIT;
-	qinf->qi_itimelimit = XFS_QM_ITIMELIMIT;
-	qinf->qi_rtbtimelimit = XFS_QM_RTBTIMELIMIT;
-	qinf->qi_bwarnlimit = XFS_QM_BWARNLIMIT;
-	qinf->qi_iwarnlimit = XFS_QM_IWARNLIMIT;
-	qinf->qi_rtbwarnlimit = XFS_QM_RTBWARNLIMIT;
+	defq = xfs_get_defquota(qinf, type);
+
+	defq->blk.time = XFS_QM_BTIMELIMIT;
+	defq->ino.time = XFS_QM_ITIMELIMIT;
+	defq->rtb.time = XFS_QM_RTBTIMELIMIT;
+	defq->blk.warn = XFS_QM_BWARNLIMIT;
+	defq->ino.warn = XFS_QM_IWARNLIMIT;
+	defq->rtb.warn = XFS_QM_RTBWARNLIMIT;
 
 	/*
 	 * We try to get the limits from the superuser's limits fields.
@@ -593,39 +597,28 @@ xfs_qm_init_timelimits(
 	 *
 	 * Since we may not have done a quotacheck by this point, just read
 	 * the dquot without attaching it to any hashtables or lists.
-	 *
-	 * Timers and warnings are globally set by the first timer found in
-	 * user/group/proj quota types, otherwise a default value is used.
-	 * This should be split into different fields per quota type.
 	 */
-	if (XFS_IS_UQUOTA_RUNNING(mp))
-		type = XFS_DQ_USER;
-	else if (XFS_IS_GQUOTA_RUNNING(mp))
-		type = XFS_DQ_GROUP;
-	else
-		type = XFS_DQ_PROJ;
 	error = xfs_qm_dqget_uncached(mp, 0, type, &dqp);
 	if (error)
 		return;
 
-	ddqp = &dqp->q_core;
 	/*
 	 * The warnings and timers set the grace period given to
 	 * a user or group before he or she can not perform any
 	 * more writing. If it is zero, a default is used.
 	 */
-	if (ddqp->d_btimer)
-		qinf->qi_btimelimit = be32_to_cpu(ddqp->d_btimer);
-	if (ddqp->d_itimer)
-		qinf->qi_itimelimit = be32_to_cpu(ddqp->d_itimer);
-	if (ddqp->d_rtbtimer)
-		qinf->qi_rtbtimelimit = be32_to_cpu(ddqp->d_rtbtimer);
-	if (ddqp->d_bwarns)
-		qinf->qi_bwarnlimit = be16_to_cpu(ddqp->d_bwarns);
-	if (ddqp->d_iwarns)
-		qinf->qi_iwarnlimit = be16_to_cpu(ddqp->d_iwarns);
-	if (ddqp->d_rtbwarns)
-		qinf->qi_rtbwarnlimit = be16_to_cpu(ddqp->d_rtbwarns);
+	if (dqp->q_blk.timer)
+		defq->blk.time = dqp->q_blk.timer;
+	if (dqp->q_ino.timer)
+		defq->ino.time = dqp->q_ino.timer;
+	if (dqp->q_rtb.timer)
+		defq->rtb.time = dqp->q_rtb.timer;
+	if (dqp->q_blk.warnings)
+		defq->blk.warn = dqp->q_blk.warnings;
+	if (dqp->q_ino.warnings)
+		defq->ino.warn = dqp->q_ino.warnings;
+	if (dqp->q_rtb.warnings)
+		defq->rtb.warn = dqp->q_rtb.warnings;
 
 	xfs_qm_dqdestroy(dqp);
 }
@@ -671,14 +664,16 @@ xfs_qm_init_quotainfo(
 
 	mp->m_qflags |= (mp->m_sb.sb_qflags & XFS_ALL_QUOTA_CHKD);
 
-	xfs_qm_init_timelimits(mp, qinf);
+	xfs_qm_init_timelimits(mp, XFS_DQTYPE_USER);
+	xfs_qm_init_timelimits(mp, XFS_DQTYPE_GROUP);
+	xfs_qm_init_timelimits(mp, XFS_DQTYPE_PROJ);
 
 	if (XFS_IS_UQUOTA_RUNNING(mp))
-		xfs_qm_set_defquota(mp, XFS_DQ_USER, qinf);
+		xfs_qm_set_defquota(mp, XFS_DQTYPE_USER, qinf);
 	if (XFS_IS_GQUOTA_RUNNING(mp))
-		xfs_qm_set_defquota(mp, XFS_DQ_GROUP, qinf);
+		xfs_qm_set_defquota(mp, XFS_DQTYPE_GROUP, qinf);
 	if (XFS_IS_PQUOTA_RUNNING(mp))
-		xfs_qm_set_defquota(mp, XFS_DQ_PROJ, qinf);
+		xfs_qm_set_defquota(mp, XFS_DQTYPE_PROJ, qinf);
 
 	qinf->qi_shrinker.count_objects = xfs_qm_shrink_count;
 	qinf->qi_shrinker.scan_objects = xfs_qm_shrink_scan;
@@ -776,7 +771,8 @@ xfs_qm_qino_alloc(
 	}
 
 	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_create,
-			XFS_QM_QINOCREATE_SPACE_RES(mp), 0, 0, &tp);
+			need_alloc ? XFS_QM_QINOCREATE_SPACE_RES(mp) : 0,
+			0, 0, &tp);
 	if (error)
 		return error;
 
@@ -827,14 +823,13 @@ xfs_qm_qino_alloc(
 
 STATIC void
 xfs_qm_reset_dqcounts(
-	xfs_mount_t	*mp,
-	xfs_buf_t	*bp,
-	xfs_dqid_t	id,
-	uint		type)
+	struct xfs_mount	*mp,
+	struct xfs_buf		*bp,
+	xfs_dqid_t		id,
+	xfs_dqtype_t		type)
 {
 	struct xfs_dqblk	*dqb;
 	int			j;
-	xfs_failaddr_t		fa;
 
 	trace_xfs_reset_dqcounts(bp, _RET_IP_);
 
@@ -859,24 +854,32 @@ xfs_qm_reset_dqcounts(
 		 * find uninitialised dquot blks. See comment in
 		 * xfs_dquot_verify.
 		 */
-		fa = xfs_dqblk_verify(mp, &dqb[j], id + j, type);
-		if (fa)
+		if (xfs_dqblk_verify(mp, &dqb[j], id + j) ||
+		    (dqb[j].dd_diskdq.d_type & XFS_DQTYPE_REC_MASK) != type)
 			xfs_dqblk_repair(mp, &dqb[j], id + j, type);
 
 		/*
 		 * Reset type in case we are reusing group quota file for
 		 * project quotas or vice versa
 		 */
-		ddq->d_flags = type;
+		ddq->d_type = type;
 		ddq->d_bcount = 0;
 		ddq->d_icount = 0;
 		ddq->d_rtbcount = 0;
-		ddq->d_btimer = 0;
-		ddq->d_itimer = 0;
-		ddq->d_rtbtimer = 0;
-		ddq->d_bwarns = 0;
-		ddq->d_iwarns = 0;
-		ddq->d_rtbwarns = 0;
+
+		/*
+		 * dquot id 0 stores the default grace period and the maximum
+		 * warning limit that were set by the administrator, so we
+		 * should not reset them.
+		 */
+		if (ddq->d_id != 0) {
+			ddq->d_btimer = 0;
+			ddq->d_itimer = 0;
+			ddq->d_rtbtimer = 0;
+			ddq->d_bwarns = 0;
+			ddq->d_iwarns = 0;
+			ddq->d_rtbwarns = 0;
+		}
 
 		if (xfs_sb_version_hascrc(&mp->m_sb)) {
 			xfs_update_cksum((char *)&dqb[j],
@@ -892,17 +895,13 @@ xfs_qm_reset_dqcounts_all(
 	xfs_dqid_t		firstid,
 	xfs_fsblock_t		bno,
 	xfs_filblks_t		blkcnt,
-	uint			flags,
+	xfs_dqtype_t		type,
 	struct list_head	*buffer_list)
 {
 	struct xfs_buf		*bp;
-	int			error;
-	int			type;
+	int			error = 0;
 
 	ASSERT(blkcnt > 0);
-	type = flags & XFS_QMOPT_UQUOTA ? XFS_DQ_USER :
-		(flags & XFS_QMOPT_PQUOTA ? XFS_DQ_PROJ : XFS_DQ_GROUP);
-	error = 0;
 
 	/*
 	 * Blkcnt arg can be a very big number, and might even be
@@ -962,7 +961,7 @@ STATIC int
 xfs_qm_reset_dqcounts_buf(
 	struct xfs_mount	*mp,
 	struct xfs_inode	*qip,
-	uint			flags,
+	xfs_dqtype_t		type,
 	struct list_head	*buffer_list)
 {
 	struct xfs_bmbt_irec	*map;
@@ -1038,7 +1037,7 @@ xfs_qm_reset_dqcounts_buf(
 			error = xfs_qm_reset_dqcounts_all(mp, firstid,
 						   map[i].br_startblock,
 						   map[i].br_blockcount,
-						   flags, buffer_list);
+						   type, buffer_list);
 			if (error)
 				goto out;
 		}
@@ -1060,7 +1059,7 @@ out:
 STATIC int
 xfs_qm_quotacheck_dqadjust(
 	struct xfs_inode	*ip,
-	uint			type,
+	xfs_dqtype_t		type,
 	xfs_qcnt_t		nblks,
 	xfs_qcnt_t		rtblks)
 {
@@ -1086,15 +1085,15 @@ xfs_qm_quotacheck_dqadjust(
 	 * Adjust the inode count and the block count to reflect this inode's
 	 * resource usage.
 	 */
-	be64_add_cpu(&dqp->q_core.d_icount, 1);
-	dqp->q_res_icount++;
+	dqp->q_ino.count++;
+	dqp->q_ino.reserved++;
 	if (nblks) {
-		be64_add_cpu(&dqp->q_core.d_bcount, nblks);
-		dqp->q_res_bcount += nblks;
+		dqp->q_blk.count += nblks;
+		dqp->q_blk.reserved += nblks;
 	}
 	if (rtblks) {
-		be64_add_cpu(&dqp->q_core.d_rtbcount, rtblks);
-		dqp->q_res_rtbcount += rtblks;
+		dqp->q_rtb.count += rtblks;
+		dqp->q_rtb.reserved += rtblks;
 	}
 
 	/*
@@ -1102,12 +1101,12 @@ xfs_qm_quotacheck_dqadjust(
 	 *
 	 * There are no timers for the default values set in the root dquot.
 	 */
-	if (dqp->q_core.d_id) {
-		xfs_qm_adjust_dqlimits(mp, dqp);
-		xfs_qm_adjust_dqtimers(mp, &dqp->q_core);
+	if (dqp->q_id) {
+		xfs_qm_adjust_dqlimits(dqp);
+		xfs_qm_adjust_dqtimers(dqp);
 	}
 
-	dqp->dq_flags |= XFS_DQ_DIRTY;
+	dqp->q_flags |= XFS_DQFLAG_DIRTY;
 	xfs_qm_dqput(dqp);
 	return 0;
 }
@@ -1177,21 +1176,21 @@ xfs_qm_dqusage_adjust(
 	 * and quotaoffs don't race. (Quotachecks happen at mount time only).
 	 */
 	if (XFS_IS_UQUOTA_ON(mp)) {
-		error = xfs_qm_quotacheck_dqadjust(ip, XFS_DQ_USER, nblks,
+		error = xfs_qm_quotacheck_dqadjust(ip, XFS_DQTYPE_USER, nblks,
 				rtblks);
 		if (error)
 			goto error0;
 	}
 
 	if (XFS_IS_GQUOTA_ON(mp)) {
-		error = xfs_qm_quotacheck_dqadjust(ip, XFS_DQ_GROUP, nblks,
+		error = xfs_qm_quotacheck_dqadjust(ip, XFS_DQTYPE_GROUP, nblks,
 				rtblks);
 		if (error)
 			goto error0;
 	}
 
 	if (XFS_IS_PQUOTA_ON(mp)) {
-		error = xfs_qm_quotacheck_dqadjust(ip, XFS_DQ_PROJ, nblks,
+		error = xfs_qm_quotacheck_dqadjust(ip, XFS_DQTYPE_PROJ, nblks,
 				rtblks);
 		if (error)
 			goto error0;
@@ -1213,7 +1212,7 @@ xfs_qm_flush_one(
 	int			error = 0;
 
 	xfs_dqlock(dqp);
-	if (dqp->dq_flags & XFS_DQ_FREEING)
+	if (dqp->q_flags & XFS_DQFLAG_FREEING)
 		goto out_unlock;
 	if (!XFS_DQ_IS_DIRTY(dqp))
 		goto out_unlock;
@@ -1282,7 +1281,7 @@ xfs_qm_quotacheck(
 	 * We don't log our changes till later.
 	 */
 	if (uip) {
-		error = xfs_qm_reset_dqcounts_buf(mp, uip, XFS_QMOPT_UQUOTA,
+		error = xfs_qm_reset_dqcounts_buf(mp, uip, XFS_DQTYPE_USER,
 					 &buffer_list);
 		if (error)
 			goto error_return;
@@ -1290,7 +1289,7 @@ xfs_qm_quotacheck(
 	}
 
 	if (gip) {
-		error = xfs_qm_reset_dqcounts_buf(mp, gip, XFS_QMOPT_GQUOTA,
+		error = xfs_qm_reset_dqcounts_buf(mp, gip, XFS_DQTYPE_GROUP,
 					 &buffer_list);
 		if (error)
 			goto error_return;
@@ -1298,7 +1297,7 @@ xfs_qm_quotacheck(
 	}
 
 	if (pip) {
-		error = xfs_qm_reset_dqcounts_buf(mp, pip, XFS_QMOPT_PQUOTA,
+		error = xfs_qm_reset_dqcounts_buf(mp, pip, XFS_DQTYPE_PROJ,
 					 &buffer_list);
 		if (error)
 			goto error_return;
@@ -1315,17 +1314,17 @@ xfs_qm_quotacheck(
 	 * down to disk buffers if everything was updated successfully.
 	 */
 	if (XFS_IS_UQUOTA_ON(mp)) {
-		error = xfs_qm_dquot_walk(mp, XFS_DQ_USER, xfs_qm_flush_one,
+		error = xfs_qm_dquot_walk(mp, XFS_DQTYPE_USER, xfs_qm_flush_one,
 					  &buffer_list);
 	}
 	if (XFS_IS_GQUOTA_ON(mp)) {
-		error2 = xfs_qm_dquot_walk(mp, XFS_DQ_GROUP, xfs_qm_flush_one,
+		error2 = xfs_qm_dquot_walk(mp, XFS_DQTYPE_GROUP, xfs_qm_flush_one,
 					   &buffer_list);
 		if (!error)
 			error = error2;
 	}
 	if (XFS_IS_PQUOTA_ON(mp)) {
-		error2 = xfs_qm_dquot_walk(mp, XFS_DQ_PROJ, xfs_qm_flush_one,
+		error2 = xfs_qm_dquot_walk(mp, XFS_DQTYPE_PROJ, xfs_qm_flush_one,
 					   &buffer_list);
 		if (!error)
 			error = error2;
@@ -1588,8 +1587,7 @@ xfs_qm_dqfree_one(
 	struct xfs_quotainfo	*qi = mp->m_quotainfo;
 
 	mutex_lock(&qi->qi_tree_lock);
-	radix_tree_delete(xfs_dquot_tree(qi, dqp->q_core.d_flags),
-			  be32_to_cpu(dqp->q_core.d_id));
+	radix_tree_delete(xfs_dquot_tree(qi, xfs_dquot_type(dqp)), dqp->q_id);
 
 	qi->qi_dquots--;
 	mutex_unlock(&qi->qi_tree_lock);
@@ -1613,8 +1611,8 @@ xfs_qm_dqfree_one(
 int
 xfs_qm_vop_dqalloc(
 	struct xfs_inode	*ip,
-	xfs_dqid_t		uid,
-	xfs_dqid_t		gid,
+	kuid_t			uid,
+	kgid_t			gid,
 	prid_t			prid,
 	uint			flags,
 	struct xfs_dquot	**O_udqpp,
@@ -1622,6 +1620,8 @@ xfs_qm_vop_dqalloc(
 	struct xfs_dquot	**O_pdqpp)
 {
 	struct xfs_mount	*mp = ip->i_mount;
+	struct inode		*inode = VFS_I(ip);
+	struct user_namespace	*user_ns = inode->i_sb->s_user_ns;
 	struct xfs_dquot	*uq = NULL;
 	struct xfs_dquot	*gq = NULL;
 	struct xfs_dquot	*pq = NULL;
@@ -1635,7 +1635,7 @@ xfs_qm_vop_dqalloc(
 	xfs_ilock(ip, lockflags);
 
 	if ((flags & XFS_QMOPT_INHERIT) && XFS_INHERIT_GID(ip))
-		gid = ip->i_d.di_gid;
+		gid = inode->i_gid;
 
 	/*
 	 * Attach the dquot(s) to this inode, doing a dquot allocation
@@ -1650,7 +1650,7 @@ xfs_qm_vop_dqalloc(
 	}
 
 	if ((flags & XFS_QMOPT_UQUOTA) && XFS_IS_UQUOTA_ON(mp)) {
-		if (ip->i_d.di_uid != uid) {
+		if (!uid_eq(inode->i_uid, uid)) {
 			/*
 			 * What we need is the dquot that has this uid, and
 			 * if we send the inode to dqget, the uid of the inode
@@ -1661,7 +1661,8 @@ xfs_qm_vop_dqalloc(
 			 * holding ilock.
 			 */
 			xfs_iunlock(ip, lockflags);
-			error = xfs_qm_dqget(mp, uid, XFS_DQ_USER, true, &uq);
+			error = xfs_qm_dqget(mp, from_kuid(user_ns, uid),
+					XFS_DQTYPE_USER, true, &uq);
 			if (error) {
 				ASSERT(error != -ENOENT);
 				return error;
@@ -1682,9 +1683,10 @@ xfs_qm_vop_dqalloc(
 		}
 	}
 	if ((flags & XFS_QMOPT_GQUOTA) && XFS_IS_GQUOTA_ON(mp)) {
-		if (ip->i_d.di_gid != gid) {
+		if (!gid_eq(inode->i_gid, gid)) {
 			xfs_iunlock(ip, lockflags);
-			error = xfs_qm_dqget(mp, gid, XFS_DQ_GROUP, true, &gq);
+			error = xfs_qm_dqget(mp, from_kgid(user_ns, gid),
+					XFS_DQTYPE_GROUP, true, &gq);
 			if (error) {
 				ASSERT(error != -ENOENT);
 				goto error_rele;
@@ -1700,8 +1702,8 @@ xfs_qm_vop_dqalloc(
 	if ((flags & XFS_QMOPT_PQUOTA) && XFS_IS_PQUOTA_ON(mp)) {
 		if (ip->i_d.di_projid != prid) {
 			xfs_iunlock(ip, lockflags);
-			error = xfs_qm_dqget(mp, (xfs_dqid_t)prid, XFS_DQ_PROJ,
-					true, &pq);
+			error = xfs_qm_dqget(mp, (xfs_dqid_t)prid,
+					XFS_DQTYPE_PROJ, true, &pq);
 			if (error) {
 				ASSERT(error != -ENOENT);
 				goto error_rele;
@@ -1714,8 +1716,7 @@ xfs_qm_vop_dqalloc(
 			pq = xfs_qm_dqhold(ip->i_pdquot);
 		}
 	}
-	if (uq)
-		trace_xfs_dquot_dqalloc(ip);
+	trace_xfs_dquot_dqalloc(ip);
 
 	xfs_iunlock(ip, lockflags);
 	if (O_udqpp)
@@ -1792,7 +1793,7 @@ xfs_qm_vop_chown_reserve(
 {
 	struct xfs_mount	*mp = ip->i_mount;
 	uint64_t		delblks;
-	unsigned int		blkflags, prjflags = 0;
+	unsigned int		blkflags;
 	struct xfs_dquot	*udq_unres = NULL;
 	struct xfs_dquot	*gdq_unres = NULL;
 	struct xfs_dquot	*pdq_unres = NULL;
@@ -1810,7 +1811,7 @@ xfs_qm_vop_chown_reserve(
 			XFS_QMOPT_RES_RTBLKS : XFS_QMOPT_RES_REGBLKS;
 
 	if (XFS_IS_UQUOTA_ON(mp) && udqp &&
-	    ip->i_d.di_uid != be32_to_cpu(udqp->q_core.d_id)) {
+	    i_uid_read(VFS_I(ip)) != udqp->q_id) {
 		udq_delblks = udqp;
 		/*
 		 * If there are delayed allocation blocks, then we have to
@@ -1823,7 +1824,7 @@ xfs_qm_vop_chown_reserve(
 		}
 	}
 	if (XFS_IS_GQUOTA_ON(ip->i_mount) && gdqp &&
-	    ip->i_d.di_gid != be32_to_cpu(gdqp->q_core.d_id)) {
+	    i_gid_read(VFS_I(ip)) != gdqp->q_id) {
 		gdq_delblks = gdqp;
 		if (delblks) {
 			ASSERT(ip->i_gdquot);
@@ -1832,8 +1833,7 @@ xfs_qm_vop_chown_reserve(
 	}
 
 	if (XFS_IS_PQUOTA_ON(ip->i_mount) && pdqp &&
-	    ip->i_d.di_projid != be32_to_cpu(pdqp->q_core.d_id)) {
-		prjflags = XFS_QMOPT_ENOSPC;
+	    ip->i_d.di_projid != pdqp->q_id) {
 		pdq_delblks = pdqp;
 		if (delblks) {
 			ASSERT(ip->i_pdquot);
@@ -1843,8 +1843,7 @@ xfs_qm_vop_chown_reserve(
 
 	error = xfs_trans_reserve_quota_bydquots(tp, ip->i_mount,
 				udq_delblks, gdq_delblks, pdq_delblks,
-				ip->i_d.di_nblocks, 1,
-				flags | blkflags | prjflags);
+				ip->i_d.di_nblocks, 1, flags | blkflags);
 	if (error)
 		return error;
 
@@ -1862,8 +1861,7 @@ xfs_qm_vop_chown_reserve(
 		ASSERT(udq_unres || gdq_unres || pdq_unres);
 		error = xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount,
 			    udq_delblks, gdq_delblks, pdq_delblks,
-			    (xfs_qcnt_t)delblks, 0,
-			    flags | blkflags | prjflags);
+			    (xfs_qcnt_t)delblks, 0, flags | blkflags);
 		if (error)
 			return error;
 		xfs_trans_reserve_quota_bydquots(NULL, ip->i_mount,
@@ -1916,24 +1914,24 @@ xfs_qm_vop_create_dqattach(
 		return;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	ASSERT(XFS_IS_QUOTA_RUNNING(mp));
 
 	if (udqp && XFS_IS_UQUOTA_ON(mp)) {
 		ASSERT(ip->i_udquot == NULL);
-		ASSERT(ip->i_d.di_uid == be32_to_cpu(udqp->q_core.d_id));
+		ASSERT(i_uid_read(VFS_I(ip)) == udqp->q_id);
 
 		ip->i_udquot = xfs_qm_dqhold(udqp);
 		xfs_trans_mod_dquot(tp, udqp, XFS_TRANS_DQ_ICOUNT, 1);
 	}
 	if (gdqp && XFS_IS_GQUOTA_ON(mp)) {
 		ASSERT(ip->i_gdquot == NULL);
-		ASSERT(ip->i_d.di_gid == be32_to_cpu(gdqp->q_core.d_id));
+		ASSERT(i_gid_read(VFS_I(ip)) == gdqp->q_id);
+
 		ip->i_gdquot = xfs_qm_dqhold(gdqp);
 		xfs_trans_mod_dquot(tp, gdqp, XFS_TRANS_DQ_ICOUNT, 1);
 	}
 	if (pdqp && XFS_IS_PQUOTA_ON(mp)) {
 		ASSERT(ip->i_pdquot == NULL);
-		ASSERT(ip->i_d.di_projid == be32_to_cpu(pdqp->q_core.d_id));
+		ASSERT(ip->i_d.di_projid == pdqp->q_id);
 
 		ip->i_pdquot = xfs_qm_dqhold(pdqp);
 		xfs_trans_mod_dquot(tp, pdqp, XFS_TRANS_DQ_ICOUNT, 1);
diff --git a/fs/xfs/xfs_qm.h b/fs/xfs/xfs_qm.h
index 4e57edca8bce..9c078c35d924 100644
--- a/fs/xfs/xfs_qm.h
+++ b/fs/xfs/xfs_qm.h
@@ -20,34 +20,28 @@ extern struct kmem_zone	*xfs_qm_dqtrxzone;
 #define XFS_DQITER_MAP_SIZE	10
 
 #define XFS_IS_DQUOT_UNINITIALIZED(dqp) ( \
-	!dqp->q_core.d_blk_hardlimit && \
-	!dqp->q_core.d_blk_softlimit && \
-	!dqp->q_core.d_rtb_hardlimit && \
-	!dqp->q_core.d_rtb_softlimit && \
-	!dqp->q_core.d_ino_hardlimit && \
-	!dqp->q_core.d_ino_softlimit && \
-	!dqp->q_core.d_bcount && \
-	!dqp->q_core.d_rtbcount && \
-	!dqp->q_core.d_icount)
-
-/*
- * This defines the unit of allocation of dquots.
- * Currently, it is just one file system block, and a 4K blk contains 30
- * (136 * 30 = 4080) dquots. It's probably not worth trying to make
- * this more dynamic.
- * XXXsup However, if this number is changed, we have to make sure that we don't
- * implicitly assume that we do allocations in chunks of a single filesystem
- * block in the dquot/xqm code.
- */
-#define XFS_DQUOT_CLUSTER_SIZE_FSB	(xfs_filblks_t)1
+	!dqp->q_blk.hardlimit && \
+	!dqp->q_blk.softlimit && \
+	!dqp->q_rtb.hardlimit && \
+	!dqp->q_rtb.softlimit && \
+	!dqp->q_ino.hardlimit && \
+	!dqp->q_ino.softlimit && \
+	!dqp->q_blk.count && \
+	!dqp->q_rtb.count && \
+	!dqp->q_ino.count)
+
+struct xfs_quota_limits {
+	xfs_qcnt_t		hard;	/* default hard limit */
+	xfs_qcnt_t		soft;	/* default soft limit */
+	time64_t		time;	/* limit for timers */
+	xfs_qwarncnt_t		warn;	/* limit for warnings */
+};
 
+/* Defaults for each quota type: time limits, warn limits, usage limits */
 struct xfs_def_quota {
-	xfs_qcnt_t       bhardlimit;     /* default data blk hard limit */
-	xfs_qcnt_t       bsoftlimit;	 /* default data blk soft limit */
-	xfs_qcnt_t       ihardlimit;	 /* default inode count hard limit */
-	xfs_qcnt_t       isoftlimit;	 /* default inode count soft limit */
-	xfs_qcnt_t	 rtbhardlimit;   /* default realtime blk hard limit */
-	xfs_qcnt_t	 rtbsoftlimit;   /* default realtime blk soft limit */
+	struct xfs_quota_limits	blk;
+	struct xfs_quota_limits	ino;
+	struct xfs_quota_limits	rtb;
 };
 
 /*
@@ -55,41 +49,35 @@ struct xfs_def_quota {
  * The mount structure keeps a pointer to this.
  */
 struct xfs_quotainfo {
-	struct radix_tree_root qi_uquota_tree;
-	struct radix_tree_root qi_gquota_tree;
-	struct radix_tree_root qi_pquota_tree;
-	struct mutex qi_tree_lock;
+	struct radix_tree_root	qi_uquota_tree;
+	struct radix_tree_root	qi_gquota_tree;
+	struct radix_tree_root	qi_pquota_tree;
+	struct mutex		qi_tree_lock;
 	struct xfs_inode	*qi_uquotaip;	/* user quota inode */
 	struct xfs_inode	*qi_gquotaip;	/* group quota inode */
 	struct xfs_inode	*qi_pquotaip;	/* project quota inode */
-	struct list_lru	 qi_lru;
-	int		 qi_dquots;
-	time64_t	 qi_btimelimit;	 /* limit for blks timer */
-	time64_t	 qi_itimelimit;	 /* limit for inodes timer */
-	time64_t	 qi_rtbtimelimit;/* limit for rt blks timer */
-	xfs_qwarncnt_t	 qi_bwarnlimit;	 /* limit for blks warnings */
-	xfs_qwarncnt_t	 qi_iwarnlimit;	 /* limit for inodes warnings */
-	xfs_qwarncnt_t	 qi_rtbwarnlimit;/* limit for rt blks warnings */
-	struct mutex	 qi_quotaofflock;/* to serialize quotaoff */
-	xfs_filblks_t	 qi_dqchunklen;	 /* # BBs in a chunk of dqs */
-	uint		 qi_dqperchunk;	 /* # ondisk dqs in above chunk */
+	struct list_lru		qi_lru;
+	int			qi_dquots;
+	struct mutex		qi_quotaofflock;/* to serialize quotaoff */
+	xfs_filblks_t		qi_dqchunklen;	/* # BBs in a chunk of dqs */
+	uint			qi_dqperchunk;	/* # ondisk dq in above chunk */
 	struct xfs_def_quota	qi_usr_default;
 	struct xfs_def_quota	qi_grp_default;
 	struct xfs_def_quota	qi_prj_default;
-	struct shrinker	qi_shrinker;
+	struct shrinker		qi_shrinker;
 };
 
 static inline struct radix_tree_root *
 xfs_dquot_tree(
 	struct xfs_quotainfo	*qi,
-	int			type)
+	xfs_dqtype_t		type)
 {
 	switch (type) {
-	case XFS_DQ_USER:
+	case XFS_DQTYPE_USER:
 		return &qi->qi_uquota_tree;
-	case XFS_DQ_GROUP:
+	case XFS_DQTYPE_GROUP:
 		return &qi->qi_gquota_tree;
-	case XFS_DQ_PROJ:
+	case XFS_DQTYPE_PROJ:
 		return &qi->qi_pquota_tree;
 	default:
 		ASSERT(0);
@@ -98,14 +86,14 @@ xfs_dquot_tree(
 }
 
 static inline struct xfs_inode *
-xfs_quota_inode(xfs_mount_t *mp, uint dq_flags)
+xfs_quota_inode(struct xfs_mount *mp, xfs_dqtype_t type)
 {
-	switch (dq_flags & XFS_DQ_ALLTYPES) {
-	case XFS_DQ_USER:
+	switch (type) {
+	case XFS_DQTYPE_USER:
 		return mp->m_quotainfo->qi_uquotaip;
-	case XFS_DQ_GROUP:
+	case XFS_DQTYPE_GROUP:
 		return mp->m_quotainfo->qi_gquotaip;
-	case XFS_DQ_PROJ:
+	case XFS_DQTYPE_PROJ:
 		return mp->m_quotainfo->qi_pquotaip;
 	default:
 		ASSERT(0);
@@ -154,29 +142,35 @@ extern void		xfs_qm_dqrele_all_inodes(struct xfs_mount *, uint);
 
 /* quota ops */
 extern int		xfs_qm_scall_trunc_qfiles(struct xfs_mount *, uint);
-extern int		xfs_qm_scall_getquota(struct xfs_mount *, xfs_dqid_t,
-					uint, struct qc_dqblk *);
-extern int		xfs_qm_scall_getquota_next(struct xfs_mount *,
-					xfs_dqid_t *, uint, struct qc_dqblk *);
-extern int		xfs_qm_scall_setqlim(struct xfs_mount *, xfs_dqid_t, uint,
-					struct qc_dqblk *);
+extern int		xfs_qm_scall_getquota(struct xfs_mount *mp,
+					xfs_dqid_t id,
+					xfs_dqtype_t type,
+					struct qc_dqblk *dst);
+extern int		xfs_qm_scall_getquota_next(struct xfs_mount *mp,
+					xfs_dqid_t *id,
+					xfs_dqtype_t type,
+					struct qc_dqblk *dst);
+extern int		xfs_qm_scall_setqlim(struct xfs_mount *mp,
+					xfs_dqid_t id,
+					xfs_dqtype_t type,
+					struct qc_dqblk *newlim);
 extern int		xfs_qm_scall_quotaon(struct xfs_mount *, uint);
 extern int		xfs_qm_scall_quotaoff(struct xfs_mount *, uint);
 
 static inline struct xfs_def_quota *
-xfs_get_defquota(struct xfs_dquot *dqp, struct xfs_quotainfo *qi)
+xfs_get_defquota(struct xfs_quotainfo *qi, xfs_dqtype_t type)
 {
-	struct xfs_def_quota *defq;
-
-	if (XFS_QM_ISUDQ(dqp))
-		defq = &qi->qi_usr_default;
-	else if (XFS_QM_ISGDQ(dqp))
-		defq = &qi->qi_grp_default;
-	else {
-		ASSERT(XFS_QM_ISPDQ(dqp));
-		defq = &qi->qi_prj_default;
+	switch (type) {
+	case XFS_DQTYPE_USER:
+		return &qi->qi_usr_default;
+	case XFS_DQTYPE_GROUP:
+		return &qi->qi_grp_default;
+	case XFS_DQTYPE_PROJ:
+		return &qi->qi_prj_default;
+	default:
+		ASSERT(0);
+		return NULL;
 	}
-	return defq;
 }
 
 #endif /* __XFS_QM_H__ */
diff --git a/fs/xfs/xfs_qm_bhv.c b/fs/xfs/xfs_qm_bhv.c
index fc2fa418919f..639398091ad6 100644
--- a/fs/xfs/xfs_qm_bhv.c
+++ b/fs/xfs/xfs_qm_bhv.c
@@ -23,24 +23,24 @@ xfs_fill_statvfs_from_dquot(
 {
 	uint64_t		limit;
 
-	limit = dqp->q_core.d_blk_softlimit ?
-		be64_to_cpu(dqp->q_core.d_blk_softlimit) :
-		be64_to_cpu(dqp->q_core.d_blk_hardlimit);
+	limit = dqp->q_blk.softlimit ?
+		dqp->q_blk.softlimit :
+		dqp->q_blk.hardlimit;
 	if (limit && statp->f_blocks > limit) {
 		statp->f_blocks = limit;
 		statp->f_bfree = statp->f_bavail =
-			(statp->f_blocks > dqp->q_res_bcount) ?
-			 (statp->f_blocks - dqp->q_res_bcount) : 0;
+			(statp->f_blocks > dqp->q_blk.reserved) ?
+			 (statp->f_blocks - dqp->q_blk.reserved) : 0;
 	}
 
-	limit = dqp->q_core.d_ino_softlimit ?
-		be64_to_cpu(dqp->q_core.d_ino_softlimit) :
-		be64_to_cpu(dqp->q_core.d_ino_hardlimit);
+	limit = dqp->q_ino.softlimit ?
+		dqp->q_ino.softlimit :
+		dqp->q_ino.hardlimit;
 	if (limit && statp->f_files > limit) {
 		statp->f_files = limit;
 		statp->f_ffree =
-			(statp->f_files > dqp->q_res_icount) ?
-			 (statp->f_files - dqp->q_res_icount) : 0;
+			(statp->f_files > dqp->q_ino.reserved) ?
+			 (statp->f_files - dqp->q_ino.reserved) : 0;
 	}
 }
 
@@ -60,7 +60,7 @@ xfs_qm_statvfs(
 	struct xfs_mount	*mp = ip->i_mount;
 	struct xfs_dquot	*dqp;
 
-	if (!xfs_qm_dqget(mp, ip->i_d.di_projid, XFS_DQ_PROJ, false, &dqp)) {
+	if (!xfs_qm_dqget(mp, ip->i_d.di_projid, XFS_DQTYPE_PROJ, false, &dqp)) {
 		xfs_fill_statvfs_from_dquot(statp, dqp);
 		xfs_qm_dqput(dqp);
 	}
diff --git a/fs/xfs/xfs_qm_syscalls.c b/fs/xfs/xfs_qm_syscalls.c
index 1ea82764bf89..1c542b4a5220 100644
--- a/fs/xfs/xfs_qm_syscalls.c
+++ b/fs/xfs/xfs_qm_syscalls.c
@@ -29,8 +29,6 @@ xfs_qm_log_quotaoff(
 	int			error;
 	struct xfs_qoff_logitem	*qoffi;
 
-	*qoffstartp = NULL;
-
 	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_qm_quotaoff, 0, 0, 0, &tp);
 	if (error)
 		goto out;
@@ -62,7 +60,7 @@ out:
 STATIC int
 xfs_qm_log_quotaoff_end(
 	struct xfs_mount	*mp,
-	struct xfs_qoff_logitem	*startqoff,
+	struct xfs_qoff_logitem	**startqoff,
 	uint			flags)
 {
 	struct xfs_trans	*tp;
@@ -73,9 +71,10 @@ xfs_qm_log_quotaoff_end(
 	if (error)
 		return error;
 
-	qoffi = xfs_trans_get_qoff_item(tp, startqoff,
+	qoffi = xfs_trans_get_qoff_item(tp, *startqoff,
 					flags & XFS_ALL_QUOTA_ACCT);
 	xfs_trans_log_quotaoff_item(tp, qoffi);
+	*startqoff = NULL;
 
 	/*
 	 * We have to make sure that the transaction is secure on disk before we
@@ -103,7 +102,7 @@ xfs_qm_scall_quotaoff(
 	uint			dqtype;
 	int			error;
 	uint			inactivate_flags;
-	struct xfs_qoff_logitem	*qoffstart;
+	struct xfs_qoff_logitem	*qoffstart = NULL;
 
 	/*
 	 * No file system can have quotas enabled on disk but not in core.
@@ -228,7 +227,7 @@ xfs_qm_scall_quotaoff(
 	 * So, we have QUOTAOFF start and end logitems; the start
 	 * logitem won't get overwritten until the end logitem appears...
 	 */
-	error = xfs_qm_log_quotaoff_end(mp, qoffstart, flags);
+	error = xfs_qm_log_quotaoff_end(mp, &qoffstart, flags);
 	if (error) {
 		/* We're screwed now. Shutdown is the only option. */
 		xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
@@ -261,6 +260,8 @@ xfs_qm_scall_quotaoff(
 	}
 
 out_unlock:
+	if (error && qoffstart)
+		xfs_qm_qoff_logitem_relse(qoffstart);
 	mutex_unlock(&q->qi_quotaofflock);
 	return error;
 }
@@ -301,7 +302,7 @@ xfs_qm_scall_trunc_qfile(
 		goto out_unlock;
 	}
 
-	ASSERT(ip->i_d.di_nextents == 0);
+	ASSERT(ip->i_df.if_nextents == 0);
 
 	xfs_trans_ichgtime(tp, ip, XFS_ICHGTIME_MOD | XFS_ICHGTIME_CHG);
 	error = xfs_trans_commit(tp);
@@ -321,23 +322,23 @@ xfs_qm_scall_trunc_qfiles(
 	int		error = -EINVAL;
 
 	if (!xfs_sb_version_hasquota(&mp->m_sb) || flags == 0 ||
-	    (flags & ~XFS_DQ_ALLTYPES)) {
+	    (flags & ~XFS_QMOPT_QUOTALL)) {
 		xfs_debug(mp, "%s: flags=%x m_qflags=%x",
 			__func__, flags, mp->m_qflags);
 		return -EINVAL;
 	}
 
-	if (flags & XFS_DQ_USER) {
+	if (flags & XFS_QMOPT_UQUOTA) {
 		error = xfs_qm_scall_trunc_qfile(mp, mp->m_sb.sb_uquotino);
 		if (error)
 			return error;
 	}
-	if (flags & XFS_DQ_GROUP) {
+	if (flags & XFS_QMOPT_GQUOTA) {
 		error = xfs_qm_scall_trunc_qfile(mp, mp->m_sb.sb_gquotino);
 		if (error)
 			return error;
 	}
-	if (flags & XFS_DQ_PROJ)
+	if (flags & XFS_QMOPT_PQUOTA)
 		error = xfs_qm_scall_trunc_qfile(mp, mp->m_sb.sb_pquotino);
 
 	return error;
@@ -356,11 +357,11 @@ xfs_qm_scall_quotaon(
 	int		error;
 	uint		qf;
 
-	flags &= (XFS_ALL_QUOTA_ACCT | XFS_ALL_QUOTA_ENFD);
 	/*
-	 * Switching on quota accounting must be done at mount time.
+	 * Switching on quota accounting must be done at mount time,
+	 * only consider quota enforcement stuff here.
 	 */
-	flags &= ~(XFS_ALL_QUOTA_ACCT);
+	flags &= XFS_ALL_QUOTA_ENFD;
 
 	if (flags == 0) {
 		xfs_debug(mp, "%s: zero flags, m_qflags=%x",
@@ -436,20 +437,73 @@ xfs_qm_scall_quotaon(
 	(QC_LIMIT_MASK | QC_TIMER_MASK | QC_WARNS_MASK)
 
 /*
+ * Adjust limits of this quota, and the defaults if passed in.  Returns true
+ * if the new limits made sense and were applied, false otherwise.
+ */
+static inline bool
+xfs_setqlim_limits(
+	struct xfs_mount	*mp,
+	struct xfs_dquot_res	*res,
+	struct xfs_quota_limits	*qlim,
+	xfs_qcnt_t		hard,
+	xfs_qcnt_t		soft,
+	const char		*tag)
+{
+	/* The hard limit can't be less than the soft limit. */
+	if (hard != 0 && hard < soft) {
+		xfs_debug(mp, "%shard %lld < %ssoft %lld", tag, hard, tag,
+				soft);
+		return false;
+	}
+
+	res->hardlimit = hard;
+	res->softlimit = soft;
+	if (qlim) {
+		qlim->hard = hard;
+		qlim->soft = soft;
+	}
+
+	return true;
+}
+
+static inline void
+xfs_setqlim_warns(
+	struct xfs_dquot_res	*res,
+	struct xfs_quota_limits	*qlim,
+	int			warns)
+{
+	res->warnings = warns;
+	if (qlim)
+		qlim->warn = warns;
+}
+
+static inline void
+xfs_setqlim_timer(
+	struct xfs_dquot_res	*res,
+	struct xfs_quota_limits	*qlim,
+	s64			timer)
+{
+	res->timer = timer;
+	if (qlim)
+		qlim->time = timer;
+}
+
+/*
  * Adjust quota limits, and start/stop timers accordingly.
  */
 int
 xfs_qm_scall_setqlim(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		id,
-	uint			type,
+	xfs_dqtype_t		type,
 	struct qc_dqblk		*newlim)
 {
 	struct xfs_quotainfo	*q = mp->m_quotainfo;
-	struct xfs_disk_dquot	*ddq;
 	struct xfs_dquot	*dqp;
 	struct xfs_trans	*tp;
 	struct xfs_def_quota	*defq;
+	struct xfs_dquot_res	*res;
+	struct xfs_quota_limits	*qlim;
 	int			error;
 	xfs_qcnt_t		hard, soft;
 
@@ -478,7 +532,7 @@ xfs_qm_scall_setqlim(
 		goto out_unlock;
 	}
 
-	defq = xfs_get_defquota(dqp, q);
+	defq = xfs_get_defquota(q, xfs_dquot_type(dqp));
 	xfs_dqunlock(dqp);
 
 	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_qm_setqlim, 0, 0, 0, &tp);
@@ -487,99 +541,74 @@ xfs_qm_scall_setqlim(
 
 	xfs_dqlock(dqp);
 	xfs_trans_dqjoin(tp, dqp);
-	ddq = &dqp->q_core;
 
 	/*
+	 * Update quota limits, warnings, and timers, and the defaults
+	 * if we're touching id == 0.
+	 *
 	 * Make sure that hardlimits are >= soft limits before changing.
+	 *
+	 * Update warnings counter(s) if requested.
+	 *
+	 * Timelimits for the super user set the relative time the other users
+	 * can be over quota for this file system. If it is zero a default is
+	 * used.  Ditto for the default soft and hard limit values (already
+	 * done, above), and for warnings.
+	 *
+	 * For other IDs, userspace can bump out the grace period if over
+	 * the soft limit.
 	 */
+
+	/* Blocks on the data device. */
 	hard = (newlim->d_fieldmask & QC_SPC_HARD) ?
 		(xfs_qcnt_t) XFS_B_TO_FSB(mp, newlim->d_spc_hardlimit) :
-			be64_to_cpu(ddq->d_blk_hardlimit);
+			dqp->q_blk.hardlimit;
 	soft = (newlim->d_fieldmask & QC_SPC_SOFT) ?
 		(xfs_qcnt_t) XFS_B_TO_FSB(mp, newlim->d_spc_softlimit) :
-			be64_to_cpu(ddq->d_blk_softlimit);
-	if (hard == 0 || hard >= soft) {
-		ddq->d_blk_hardlimit = cpu_to_be64(hard);
-		ddq->d_blk_softlimit = cpu_to_be64(soft);
+			dqp->q_blk.softlimit;
+	res = &dqp->q_blk;
+	qlim = id == 0 ? &defq->blk : NULL;
+
+	if (xfs_setqlim_limits(mp, res, qlim, hard, soft, "blk"))
 		xfs_dquot_set_prealloc_limits(dqp);
-		if (id == 0) {
-			defq->bhardlimit = hard;
-			defq->bsoftlimit = soft;
-		}
-	} else {
-		xfs_debug(mp, "blkhard %Ld < blksoft %Ld", hard, soft);
-	}
+	if (newlim->d_fieldmask & QC_SPC_WARNS)
+		xfs_setqlim_warns(res, qlim, newlim->d_spc_warns);
+	if (newlim->d_fieldmask & QC_SPC_TIMER)
+		xfs_setqlim_timer(res, qlim, newlim->d_spc_timer);
+
+	/* Blocks on the realtime device. */
 	hard = (newlim->d_fieldmask & QC_RT_SPC_HARD) ?
 		(xfs_qcnt_t) XFS_B_TO_FSB(mp, newlim->d_rt_spc_hardlimit) :
-			be64_to_cpu(ddq->d_rtb_hardlimit);
+			dqp->q_rtb.hardlimit;
 	soft = (newlim->d_fieldmask & QC_RT_SPC_SOFT) ?
 		(xfs_qcnt_t) XFS_B_TO_FSB(mp, newlim->d_rt_spc_softlimit) :
-			be64_to_cpu(ddq->d_rtb_softlimit);
-	if (hard == 0 || hard >= soft) {
-		ddq->d_rtb_hardlimit = cpu_to_be64(hard);
-		ddq->d_rtb_softlimit = cpu_to_be64(soft);
-		if (id == 0) {
-			defq->rtbhardlimit = hard;
-			defq->rtbsoftlimit = soft;
-		}
-	} else {
-		xfs_debug(mp, "rtbhard %Ld < rtbsoft %Ld", hard, soft);
-	}
+			dqp->q_rtb.softlimit;
+	res = &dqp->q_rtb;
+	qlim = id == 0 ? &defq->rtb : NULL;
+
+	xfs_setqlim_limits(mp, res, qlim, hard, soft, "rtb");
+	if (newlim->d_fieldmask & QC_RT_SPC_WARNS)
+		xfs_setqlim_warns(res, qlim, newlim->d_rt_spc_warns);
+	if (newlim->d_fieldmask & QC_RT_SPC_TIMER)
+		xfs_setqlim_timer(res, qlim, newlim->d_rt_spc_timer);
 
+	/* Inodes */
 	hard = (newlim->d_fieldmask & QC_INO_HARD) ?
 		(xfs_qcnt_t) newlim->d_ino_hardlimit :
-			be64_to_cpu(ddq->d_ino_hardlimit);
+			dqp->q_ino.hardlimit;
 	soft = (newlim->d_fieldmask & QC_INO_SOFT) ?
 		(xfs_qcnt_t) newlim->d_ino_softlimit :
-			be64_to_cpu(ddq->d_ino_softlimit);
-	if (hard == 0 || hard >= soft) {
-		ddq->d_ino_hardlimit = cpu_to_be64(hard);
-		ddq->d_ino_softlimit = cpu_to_be64(soft);
-		if (id == 0) {
-			defq->ihardlimit = hard;
-			defq->isoftlimit = soft;
-		}
-	} else {
-		xfs_debug(mp, "ihard %Ld < isoft %Ld", hard, soft);
-	}
+			dqp->q_ino.softlimit;
+	res = &dqp->q_ino;
+	qlim = id == 0 ? &defq->ino : NULL;
 
-	/*
-	 * Update warnings counter(s) if requested
-	 */
-	if (newlim->d_fieldmask & QC_SPC_WARNS)
-		ddq->d_bwarns = cpu_to_be16(newlim->d_spc_warns);
+	xfs_setqlim_limits(mp, res, qlim, hard, soft, "ino");
 	if (newlim->d_fieldmask & QC_INO_WARNS)
-		ddq->d_iwarns = cpu_to_be16(newlim->d_ino_warns);
-	if (newlim->d_fieldmask & QC_RT_SPC_WARNS)
-		ddq->d_rtbwarns = cpu_to_be16(newlim->d_rt_spc_warns);
+		xfs_setqlim_warns(res, qlim, newlim->d_ino_warns);
+	if (newlim->d_fieldmask & QC_INO_TIMER)
+		xfs_setqlim_timer(res, qlim, newlim->d_ino_timer);
 
-	if (id == 0) {
-		/*
-		 * Timelimits for the super user set the relative time
-		 * the other users can be over quota for this file system.
-		 * If it is zero a default is used.  Ditto for the default
-		 * soft and hard limit values (already done, above), and
-		 * for warnings.
-		 */
-		if (newlim->d_fieldmask & QC_SPC_TIMER) {
-			q->qi_btimelimit = newlim->d_spc_timer;
-			ddq->d_btimer = cpu_to_be32(newlim->d_spc_timer);
-		}
-		if (newlim->d_fieldmask & QC_INO_TIMER) {
-			q->qi_itimelimit = newlim->d_ino_timer;
-			ddq->d_itimer = cpu_to_be32(newlim->d_ino_timer);
-		}
-		if (newlim->d_fieldmask & QC_RT_SPC_TIMER) {
-			q->qi_rtbtimelimit = newlim->d_rt_spc_timer;
-			ddq->d_rtbtimer = cpu_to_be32(newlim->d_rt_spc_timer);
-		}
-		if (newlim->d_fieldmask & QC_SPC_WARNS)
-			q->qi_bwarnlimit = newlim->d_spc_warns;
-		if (newlim->d_fieldmask & QC_INO_WARNS)
-			q->qi_iwarnlimit = newlim->d_ino_warns;
-		if (newlim->d_fieldmask & QC_RT_SPC_WARNS)
-			q->qi_rtbwarnlimit = newlim->d_rt_spc_warns;
-	} else {
+	if (id != 0) {
 		/*
 		 * If the user is now over quota, start the timelimit.
 		 * The user will not be 'warned'.
@@ -587,9 +616,9 @@ xfs_qm_scall_setqlim(
 		 * is on or off. We don't really want to bother with iterating
 		 * over all ondisk dquots and turning the timers on/off.
 		 */
-		xfs_qm_adjust_dqtimers(mp, ddq);
+		xfs_qm_adjust_dqtimers(dqp);
 	}
-	dqp->dq_flags |= XFS_DQ_DIRTY;
+	dqp->q_flags |= XFS_DQFLAG_DIRTY;
 	xfs_trans_log_dquot(tp, dqp);
 
 	error = xfs_trans_commit(tp);
@@ -605,58 +634,46 @@ out_unlock:
 static void
 xfs_qm_scall_getquota_fill_qc(
 	struct xfs_mount	*mp,
-	uint			type,
+	xfs_dqtype_t		type,
 	const struct xfs_dquot	*dqp,
 	struct qc_dqblk		*dst)
 {
 	memset(dst, 0, sizeof(*dst));
-	dst->d_spc_hardlimit =
-		XFS_FSB_TO_B(mp, be64_to_cpu(dqp->q_core.d_blk_hardlimit));
-	dst->d_spc_softlimit =
-		XFS_FSB_TO_B(mp, be64_to_cpu(dqp->q_core.d_blk_softlimit));
-	dst->d_ino_hardlimit = be64_to_cpu(dqp->q_core.d_ino_hardlimit);
-	dst->d_ino_softlimit = be64_to_cpu(dqp->q_core.d_ino_softlimit);
-	dst->d_space = XFS_FSB_TO_B(mp, dqp->q_res_bcount);
-	dst->d_ino_count = dqp->q_res_icount;
-	dst->d_spc_timer = be32_to_cpu(dqp->q_core.d_btimer);
-	dst->d_ino_timer = be32_to_cpu(dqp->q_core.d_itimer);
-	dst->d_ino_warns = be16_to_cpu(dqp->q_core.d_iwarns);
-	dst->d_spc_warns = be16_to_cpu(dqp->q_core.d_bwarns);
-	dst->d_rt_spc_hardlimit =
-		XFS_FSB_TO_B(mp, be64_to_cpu(dqp->q_core.d_rtb_hardlimit));
-	dst->d_rt_spc_softlimit =
-		XFS_FSB_TO_B(mp, be64_to_cpu(dqp->q_core.d_rtb_softlimit));
-	dst->d_rt_space = XFS_FSB_TO_B(mp, dqp->q_res_rtbcount);
-	dst->d_rt_spc_timer = be32_to_cpu(dqp->q_core.d_rtbtimer);
-	dst->d_rt_spc_warns = be16_to_cpu(dqp->q_core.d_rtbwarns);
+	dst->d_spc_hardlimit = XFS_FSB_TO_B(mp, dqp->q_blk.hardlimit);
+	dst->d_spc_softlimit = XFS_FSB_TO_B(mp, dqp->q_blk.softlimit);
+	dst->d_ino_hardlimit = dqp->q_ino.hardlimit;
+	dst->d_ino_softlimit = dqp->q_ino.softlimit;
+	dst->d_space = XFS_FSB_TO_B(mp, dqp->q_blk.reserved);
+	dst->d_ino_count = dqp->q_ino.reserved;
+	dst->d_spc_timer = dqp->q_blk.timer;
+	dst->d_ino_timer = dqp->q_ino.timer;
+	dst->d_ino_warns = dqp->q_ino.warnings;
+	dst->d_spc_warns = dqp->q_blk.warnings;
+	dst->d_rt_spc_hardlimit = XFS_FSB_TO_B(mp, dqp->q_rtb.hardlimit);
+	dst->d_rt_spc_softlimit = XFS_FSB_TO_B(mp, dqp->q_rtb.softlimit);
+	dst->d_rt_space = XFS_FSB_TO_B(mp, dqp->q_rtb.reserved);
+	dst->d_rt_spc_timer = dqp->q_rtb.timer;
+	dst->d_rt_spc_warns = dqp->q_rtb.warnings;
 
 	/*
 	 * Internally, we don't reset all the timers when quota enforcement
 	 * gets turned off. No need to confuse the user level code,
 	 * so return zeroes in that case.
 	 */
-	if ((!XFS_IS_UQUOTA_ENFORCED(mp) &&
-	     dqp->q_core.d_flags == XFS_DQ_USER) ||
-	    (!XFS_IS_GQUOTA_ENFORCED(mp) &&
-	     dqp->q_core.d_flags == XFS_DQ_GROUP) ||
-	    (!XFS_IS_PQUOTA_ENFORCED(mp) &&
-	     dqp->q_core.d_flags == XFS_DQ_PROJ)) {
+	if (!xfs_dquot_is_enforced(dqp)) {
 		dst->d_spc_timer = 0;
 		dst->d_ino_timer = 0;
 		dst->d_rt_spc_timer = 0;
 	}
 
 #ifdef DEBUG
-	if (((XFS_IS_UQUOTA_ENFORCED(mp) && type == XFS_DQ_USER) ||
-	     (XFS_IS_GQUOTA_ENFORCED(mp) && type == XFS_DQ_GROUP) ||
-	     (XFS_IS_PQUOTA_ENFORCED(mp) && type == XFS_DQ_PROJ)) &&
-	    dqp->q_core.d_id != 0) {
+	if (xfs_dquot_is_enforced(dqp) && dqp->q_id != 0) {
 		if ((dst->d_space > dst->d_spc_softlimit) &&
 		    (dst->d_spc_softlimit > 0)) {
 			ASSERT(dst->d_spc_timer != 0);
 		}
-		if ((dst->d_ino_count > dst->d_ino_softlimit) &&
-		    (dst->d_ino_softlimit > 0)) {
+		if ((dst->d_ino_count > dqp->q_ino.softlimit) &&
+		    (dqp->q_ino.softlimit > 0)) {
 			ASSERT(dst->d_ino_timer != 0);
 		}
 	}
@@ -668,7 +685,7 @@ int
 xfs_qm_scall_getquota(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		id,
-	uint			type,
+	xfs_dqtype_t		type,
 	struct qc_dqblk		*dst)
 {
 	struct xfs_dquot	*dqp;
@@ -706,7 +723,7 @@ int
 xfs_qm_scall_getquota_next(
 	struct xfs_mount	*mp,
 	xfs_dqid_t		*id,
-	uint			type,
+	xfs_dqtype_t		type,
 	struct qc_dqblk		*dst)
 {
 	struct xfs_dquot	*dqp;
@@ -717,7 +734,7 @@ xfs_qm_scall_getquota_next(
 		return error;
 
 	/* Fill in the ID we actually read from disk */
-	*id = be32_to_cpu(dqp->q_core.d_id);
+	*id = dqp->q_id;
 
 	xfs_qm_scall_getquota_fill_qc(mp, type, dqp, dst);
 
@@ -728,9 +745,10 @@ xfs_qm_scall_getquota_next(
 STATIC int
 xfs_dqrele_inode(
 	struct xfs_inode	*ip,
-	int			flags,
 	void			*args)
 {
+	uint			*flags = args;
+
 	/* skip quota inodes */
 	if (ip == ip->i_mount->m_quotainfo->qi_uquotaip ||
 	    ip == ip->i_mount->m_quotainfo->qi_gquotaip ||
@@ -742,15 +760,15 @@ xfs_dqrele_inode(
 	}
 
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
-	if ((flags & XFS_UQUOTA_ACCT) && ip->i_udquot) {
+	if ((*flags & XFS_UQUOTA_ACCT) && ip->i_udquot) {
 		xfs_qm_dqrele(ip->i_udquot);
 		ip->i_udquot = NULL;
 	}
-	if ((flags & XFS_GQUOTA_ACCT) && ip->i_gdquot) {
+	if ((*flags & XFS_GQUOTA_ACCT) && ip->i_gdquot) {
 		xfs_qm_dqrele(ip->i_gdquot);
 		ip->i_gdquot = NULL;
 	}
-	if ((flags & XFS_PQUOTA_ACCT) && ip->i_pdquot) {
+	if ((*flags & XFS_PQUOTA_ACCT) && ip->i_pdquot) {
 		xfs_qm_dqrele(ip->i_pdquot);
 		ip->i_pdquot = NULL;
 	}
@@ -767,10 +785,10 @@ xfs_dqrele_inode(
  */
 void
 xfs_qm_dqrele_all_inodes(
-	struct xfs_mount *mp,
-	uint		 flags)
+	struct xfs_mount	*mp,
+	uint			flags)
 {
 	ASSERT(mp->m_quotainfo);
-	xfs_inode_ag_iterator_flags(mp, xfs_dqrele_inode, flags, NULL,
-				    XFS_AGITER_INEW_WAIT);
+	xfs_inode_walk(mp, XFS_INODE_WALK_INEW_WAIT, xfs_dqrele_inode,
+			&flags, XFS_ICI_NO_TAG);
 }
diff --git a/fs/xfs/xfs_quota.h b/fs/xfs/xfs_quota.h
index efe42ae7a2f3..06b22e35fc90 100644
--- a/fs/xfs/xfs_quota.h
+++ b/fs/xfs/xfs_quota.h
@@ -13,6 +13,7 @@
  */
 
 struct xfs_trans;
+struct xfs_buf;
 
 /*
  * This check is done typically without holding the inode lock;
@@ -38,14 +39,14 @@ struct xfs_trans;
 
 static inline uint
 xfs_quota_chkd_flag(
-	uint		dqtype)
+	xfs_dqtype_t		type)
 {
-	switch (dqtype) {
-	case XFS_DQ_USER:
+	switch (type) {
+	case XFS_DQTYPE_USER:
 		return XFS_UQUOTA_CHKD;
-	case XFS_DQ_GROUP:
+	case XFS_DQTYPE_GROUP:
 		return XFS_GQUOTA_CHKD;
-	case XFS_DQ_PROJ:
+	case XFS_DQTYPE_PROJ:
 		return XFS_PQUOTA_CHKD;
 	default:
 		return 0;
@@ -86,7 +87,7 @@ extern int xfs_trans_reserve_quota_bydquots(struct xfs_trans *,
 		struct xfs_mount *, struct xfs_dquot *,
 		struct xfs_dquot *, struct xfs_dquot *, int64_t, long, uint);
 
-extern int xfs_qm_vop_dqalloc(struct xfs_inode *, xfs_dqid_t, xfs_dqid_t,
+extern int xfs_qm_vop_dqalloc(struct xfs_inode *, kuid_t, kgid_t,
 		prid_t, uint, struct xfs_dquot **, struct xfs_dquot **,
 		struct xfs_dquot **);
 extern void xfs_qm_vop_create_dqattach(struct xfs_trans *, struct xfs_inode *,
@@ -107,9 +108,11 @@ extern void xfs_qm_mount_quotas(struct xfs_mount *);
 extern void xfs_qm_unmount(struct xfs_mount *);
 extern void xfs_qm_unmount_quotas(struct xfs_mount *);
 
+void		xfs_dquot_done(struct xfs_buf *);
+
 #else
 static inline int
-xfs_qm_vop_dqalloc(struct xfs_inode *ip, xfs_dqid_t uid, xfs_dqid_t gid,
+xfs_qm_vop_dqalloc(struct xfs_inode *ip, kuid_t kuid, kgid_t kgid,
 		prid_t prid, uint flags, struct xfs_dquot **udqp,
 		struct xfs_dquot **gdqp, struct xfs_dquot **pdqp)
 {
@@ -148,6 +151,12 @@ static inline int xfs_trans_reserve_quota_bydquots(struct xfs_trans *tp,
 #define xfs_qm_mount_quotas(mp)
 #define xfs_qm_unmount(mp)
 #define xfs_qm_unmount_quotas(mp)
+
+static inline void xfs_dquot_done(struct xfs_buf *bp)
+{
+	return;
+}
+
 #endif /* CONFIG_XFS_QUOTA */
 
 #define xfs_trans_unreserve_quota_nblks(tp, ip, nblks, ninos, flags) \
diff --git a/fs/xfs/xfs_quotaops.c b/fs/xfs/xfs_quotaops.c
index 38669e827206..d27c0e852c0b 100644
--- a/fs/xfs/xfs_quotaops.c
+++ b/fs/xfs/xfs_quotaops.c
@@ -21,10 +21,10 @@ xfs_qm_fill_state(
 	struct qc_type_state	*tstate,
 	struct xfs_mount	*mp,
 	struct xfs_inode	*ip,
-	xfs_ino_t		ino)
+	xfs_ino_t		ino,
+	struct xfs_def_quota	*defq)
 {
-	struct xfs_quotainfo *q = mp->m_quotainfo;
-	bool tempqip = false;
+	bool			tempqip = false;
 
 	tstate->ino = ino;
 	if (!ip && ino == NULLFSINO)
@@ -36,13 +36,13 @@ xfs_qm_fill_state(
 	}
 	tstate->flags |= QCI_SYSFILE;
 	tstate->blocks = ip->i_d.di_nblocks;
-	tstate->nextents = ip->i_d.di_nextents;
-	tstate->spc_timelimit = (u32)q->qi_btimelimit;
-	tstate->ino_timelimit = (u32)q->qi_itimelimit;
-	tstate->rt_spc_timelimit = (u32)q->qi_rtbtimelimit;
-	tstate->spc_warnlimit = q->qi_bwarnlimit;
-	tstate->ino_warnlimit = q->qi_iwarnlimit;
-	tstate->rt_spc_warnlimit = q->qi_rtbwarnlimit;
+	tstate->nextents = ip->i_df.if_nextents;
+	tstate->spc_timelimit = (u32)defq->blk.time;
+	tstate->ino_timelimit = (u32)defq->ino.time;
+	tstate->rt_spc_timelimit = (u32)defq->rtb.time;
+	tstate->spc_warnlimit = defq->blk.warn;
+	tstate->ino_warnlimit = defq->ino.warn;
+	tstate->rt_spc_warnlimit = defq->rtb.warn;
 	if (tempqip)
 		xfs_irele(ip);
 }
@@ -77,24 +77,24 @@ xfs_fs_get_quota_state(
 		state->s_state[PRJQUOTA].flags |= QCI_LIMITS_ENFORCED;
 
 	xfs_qm_fill_state(&state->s_state[USRQUOTA], mp, q->qi_uquotaip,
-			  mp->m_sb.sb_uquotino);
+			  mp->m_sb.sb_uquotino, &q->qi_usr_default);
 	xfs_qm_fill_state(&state->s_state[GRPQUOTA], mp, q->qi_gquotaip,
-			  mp->m_sb.sb_gquotino);
+			  mp->m_sb.sb_gquotino, &q->qi_grp_default);
 	xfs_qm_fill_state(&state->s_state[PRJQUOTA], mp, q->qi_pquotaip,
-			  mp->m_sb.sb_pquotino);
+			  mp->m_sb.sb_pquotino, &q->qi_prj_default);
 	return 0;
 }
 
-STATIC int
+STATIC xfs_dqtype_t
 xfs_quota_type(int type)
 {
 	switch (type) {
 	case USRQUOTA:
-		return XFS_DQ_USER;
+		return XFS_DQTYPE_USER;
 	case GRPQUOTA:
-		return XFS_DQ_GROUP;
+		return XFS_DQTYPE_GROUP;
 	default:
-		return XFS_DQ_PROJ;
+		return XFS_DQTYPE_PROJ;
 	}
 }
 
@@ -109,8 +109,8 @@ xfs_fs_set_info(
 	int			type,
 	struct qc_info		*info)
 {
-	struct xfs_mount *mp = XFS_M(sb);
-	struct qc_dqblk newlim;
+	struct xfs_mount	*mp = XFS_M(sb);
+	struct qc_dqblk		newlim;
 
 	if (sb_rdonly(sb))
 		return -EROFS;
@@ -205,11 +205,11 @@ xfs_fs_rm_xquota(
 		return -EINVAL;
 
 	if (uflags & FS_USER_QUOTA)
-		flags |= XFS_DQ_USER;
+		flags |= XFS_QMOPT_UQUOTA;
 	if (uflags & FS_GROUP_QUOTA)
-		flags |= XFS_DQ_GROUP;
+		flags |= XFS_QMOPT_GQUOTA;
 	if (uflags & FS_PROJ_QUOTA)
-		flags |= XFS_DQ_PROJ;
+		flags |= XFS_QMOPT_PQUOTA;
 
 	return xfs_qm_scall_trunc_qfiles(mp, flags);
 }
diff --git a/fs/xfs/xfs_refcount_item.c b/fs/xfs/xfs_refcount_item.c
index 8eeed73928cd..ca93b6488377 100644
--- a/fs/xfs/xfs_refcount_item.c
+++ b/fs/xfs/xfs_refcount_item.c
@@ -18,16 +18,20 @@
 #include "xfs_log.h"
 #include "xfs_refcount.h"
 #include "xfs_error.h"
+#include "xfs_log_priv.h"
+#include "xfs_log_recover.h"
 
 kmem_zone_t	*xfs_cui_zone;
 kmem_zone_t	*xfs_cud_zone;
 
+static const struct xfs_item_ops xfs_cui_item_ops;
+
 static inline struct xfs_cui_log_item *CUI_ITEM(struct xfs_log_item *lip)
 {
 	return container_of(lip, struct xfs_cui_log_item, cui_item);
 }
 
-void
+STATIC void
 xfs_cui_item_free(
 	struct xfs_cui_log_item	*cuip)
 {
@@ -44,13 +48,13 @@ xfs_cui_item_free(
  * committed vs unpin operations in bulk insert operations. Hence the reference
  * count to ensure only the last caller frees the CUI.
  */
-void
+STATIC void
 xfs_cui_release(
 	struct xfs_cui_log_item	*cuip)
 {
 	ASSERT(atomic_read(&cuip->cui_refcount) > 0);
 	if (atomic_dec_and_test(&cuip->cui_refcount)) {
-		xfs_trans_ail_remove(&cuip->cui_item, SHUTDOWN_LOG_IO_ERROR);
+		xfs_trans_ail_delete(&cuip->cui_item, SHUTDOWN_LOG_IO_ERROR);
 		xfs_cui_item_free(cuip);
 	}
 }
@@ -123,17 +127,10 @@ xfs_cui_item_release(
 	xfs_cui_release(CUI_ITEM(lip));
 }
 
-static const struct xfs_item_ops xfs_cui_item_ops = {
-	.iop_size	= xfs_cui_item_size,
-	.iop_format	= xfs_cui_item_format,
-	.iop_unpin	= xfs_cui_item_unpin,
-	.iop_release	= xfs_cui_item_release,
-};
-
 /*
  * Allocate and initialize an cui item with the given number of extents.
  */
-struct xfs_cui_log_item *
+STATIC struct xfs_cui_log_item *
 xfs_cui_init(
 	struct xfs_mount		*mp,
 	uint				nextents)
@@ -146,7 +143,8 @@ xfs_cui_init(
 		cuip = kmem_zalloc(xfs_cui_log_item_sizeof(nextents),
 				0);
 	else
-		cuip = kmem_zone_zalloc(xfs_cui_zone, 0);
+		cuip = kmem_cache_zalloc(xfs_cui_zone,
+					 GFP_KERNEL | __GFP_NOFAIL);
 
 	xfs_log_item_init(mp, &cuip->cui_item, XFS_LI_CUI, &xfs_cui_item_ops);
 	cuip->cui_format.cui_nextents = nextents;
@@ -223,7 +221,7 @@ xfs_trans_get_cud(
 {
 	struct xfs_cud_log_item		*cudp;
 
-	cudp = kmem_zone_zalloc(xfs_cud_zone, 0);
+	cudp = kmem_cache_zalloc(xfs_cud_zone, GFP_KERNEL | __GFP_NOFAIL);
 	xfs_log_item_init(tp->t_mountp, &cudp->cud_item, XFS_LI_CUD,
 			  &xfs_cud_item_ops);
 	cudp->cud_cuip = cuip;
@@ -284,27 +282,6 @@ xfs_refcount_update_diff_items(
 		XFS_FSB_TO_AGNO(mp, rb->ri_startblock);
 }
 
-/* Get an CUI. */
-STATIC void *
-xfs_refcount_update_create_intent(
-	struct xfs_trans		*tp,
-	unsigned int			count)
-{
-	struct xfs_cui_log_item		*cuip;
-
-	ASSERT(tp != NULL);
-	ASSERT(count > 0);
-
-	cuip = xfs_cui_init(tp->t_mountp, count);
-	ASSERT(cuip != NULL);
-
-	/*
-	 * Get a log_item_desc to point at the new item.
-	 */
-	xfs_trans_add_item(tp, &cuip->cui_item);
-	return cuip;
-}
-
 /* Set the phys extent flags for this reverse mapping. */
 static void
 xfs_trans_set_refcount_flags(
@@ -328,16 +305,12 @@ xfs_trans_set_refcount_flags(
 STATIC void
 xfs_refcount_update_log_item(
 	struct xfs_trans		*tp,
-	void				*intent,
-	struct list_head		*item)
+	struct xfs_cui_log_item		*cuip,
+	struct xfs_refcount_intent	*refc)
 {
-	struct xfs_cui_log_item		*cuip = intent;
-	struct xfs_refcount_intent	*refc;
 	uint				next_extent;
 	struct xfs_phys_extent		*ext;
 
-	refc = container_of(item, struct xfs_refcount_intent, ri_list);
-
 	tp->t_flags |= XFS_TRANS_DIRTY;
 	set_bit(XFS_LI_DIRTY, &cuip->cui_item.li_flags);
 
@@ -354,23 +327,44 @@ xfs_refcount_update_log_item(
 	xfs_trans_set_refcount_flags(ext, refc->ri_type);
 }
 
+static struct xfs_log_item *
+xfs_refcount_update_create_intent(
+	struct xfs_trans		*tp,
+	struct list_head		*items,
+	unsigned int			count,
+	bool				sort)
+{
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_cui_log_item		*cuip = xfs_cui_init(mp, count);
+	struct xfs_refcount_intent	*refc;
+
+	ASSERT(count > 0);
+
+	xfs_trans_add_item(tp, &cuip->cui_item);
+	if (sort)
+		list_sort(mp, items, xfs_refcount_update_diff_items);
+	list_for_each_entry(refc, items, ri_list)
+		xfs_refcount_update_log_item(tp, cuip, refc);
+	return &cuip->cui_item;
+}
+
 /* Get an CUD so we can process all the deferred refcount updates. */
-STATIC void *
+static struct xfs_log_item *
 xfs_refcount_update_create_done(
 	struct xfs_trans		*tp,
-	void				*intent,
+	struct xfs_log_item		*intent,
 	unsigned int			count)
 {
-	return xfs_trans_get_cud(tp, intent);
+	return &xfs_trans_get_cud(tp, CUI_ITEM(intent))->cud_item;
 }
 
 /* Process a deferred refcount update. */
 STATIC int
 xfs_refcount_update_finish_item(
 	struct xfs_trans		*tp,
+	struct xfs_log_item		*done,
 	struct list_head		*item,
-	void				*done_item,
-	void				**state)
+	struct xfs_btree_cur		**state)
 {
 	struct xfs_refcount_intent	*refc;
 	xfs_fsblock_t			new_fsb;
@@ -378,12 +372,10 @@ xfs_refcount_update_finish_item(
 	int				error;
 
 	refc = container_of(item, struct xfs_refcount_intent, ri_list);
-	error = xfs_trans_log_finish_refcount_update(tp, done_item,
-			refc->ri_type,
-			refc->ri_startblock,
-			refc->ri_blockcount,
-			&new_fsb, &new_aglen,
-			(struct xfs_btree_cur **)state);
+	error = xfs_trans_log_finish_refcount_update(tp, CUD_ITEM(done),
+			refc->ri_type, refc->ri_startblock, refc->ri_blockcount,
+			&new_fsb, &new_aglen, state);
+
 	/* Did we run out of reservation?  Requeue what we didn't finish. */
 	if (!error && new_aglen > 0) {
 		ASSERT(refc->ri_type == XFS_REFCOUNT_INCREASE ||
@@ -396,24 +388,12 @@ xfs_refcount_update_finish_item(
 	return error;
 }
 
-/* Clean up after processing deferred refcounts. */
-STATIC void
-xfs_refcount_update_finish_cleanup(
-	struct xfs_trans	*tp,
-	void			*state,
-	int			error)
-{
-	struct xfs_btree_cur	*rcur = state;
-
-	xfs_refcount_finish_one_cleanup(tp, rcur, error);
-}
-
 /* Abort all pending CUIs. */
 STATIC void
 xfs_refcount_update_abort_intent(
-	void				*intent)
+	struct xfs_log_item		*intent)
 {
-	xfs_cui_release(intent);
+	xfs_cui_release(CUI_ITEM(intent));
 }
 
 /* Cancel a deferred refcount update. */
@@ -429,13 +409,11 @@ xfs_refcount_update_cancel_item(
 
 const struct xfs_defer_op_type xfs_refcount_update_defer_type = {
 	.max_items	= XFS_CUI_MAX_FAST_EXTENTS,
-	.diff_items	= xfs_refcount_update_diff_items,
 	.create_intent	= xfs_refcount_update_create_intent,
 	.abort_intent	= xfs_refcount_update_abort_intent,
-	.log_item	= xfs_refcount_update_log_item,
 	.create_done	= xfs_refcount_update_create_done,
 	.finish_item	= xfs_refcount_update_finish_item,
-	.finish_cleanup = xfs_refcount_update_finish_cleanup,
+	.finish_cleanup = xfs_refcount_finish_one_cleanup,
 	.cancel_item	= xfs_refcount_update_cancel_item,
 };
 
@@ -443,28 +421,27 @@ const struct xfs_defer_op_type xfs_refcount_update_defer_type = {
  * Process a refcount update intent item that was recovered from the log.
  * We need to update the refcountbt.
  */
-int
-xfs_cui_recover(
-	struct xfs_trans		*parent_tp,
-	struct xfs_cui_log_item		*cuip)
+STATIC int
+xfs_cui_item_recover(
+	struct xfs_log_item		*lip,
+	struct xfs_trans		*parent_tp)
 {
-	int				i;
-	int				error = 0;
-	unsigned int			refc_type;
+	struct xfs_bmbt_irec		irec;
+	struct xfs_cui_log_item		*cuip = CUI_ITEM(lip);
 	struct xfs_phys_extent		*refc;
-	xfs_fsblock_t			startblock_fsb;
-	bool				op_ok;
 	struct xfs_cud_log_item		*cudp;
 	struct xfs_trans		*tp;
 	struct xfs_btree_cur		*rcur = NULL;
-	enum xfs_refcount_intent_type	type;
+	struct xfs_mount		*mp = parent_tp->t_mountp;
+	xfs_fsblock_t			startblock_fsb;
 	xfs_fsblock_t			new_fsb;
 	xfs_extlen_t			new_len;
-	struct xfs_bmbt_irec		irec;
+	unsigned int			refc_type;
+	bool				op_ok;
 	bool				requeue_only = false;
-	struct xfs_mount		*mp = parent_tp->t_mountp;
-
-	ASSERT(!test_bit(XFS_CUI_RECOVERED, &cuip->cui_flags));
+	enum xfs_refcount_intent_type	type;
+	int				i;
+	int				error = 0;
 
 	/*
 	 * First check the validity of the extents described by the
@@ -495,7 +472,6 @@ xfs_cui_recover(
 			 * This will pull the CUI from the AIL and
 			 * free the memory associated with it.
 			 */
-			set_bit(XFS_CUI_RECOVERED, &cuip->cui_flags);
 			xfs_cui_release(cuip);
 			return -EFSCORRUPTED;
 		}
@@ -509,7 +485,7 @@ xfs_cui_recover(
 	 * transaction.  Normally, any work that needs to be deferred
 	 * gets attached to the same defer_ops that scheduled the
 	 * refcount update.  However, we're in log recovery here, so we
-	 * we use the passed in defer_ops and to finish up any work that
+	 * use the passed in defer_ops and to finish up any work that
 	 * doesn't fit.  We need to reserve enough blocks to handle a
 	 * full btree split on either end of the refcount range.
 	 */
@@ -579,7 +555,6 @@ xfs_cui_recover(
 	}
 
 	xfs_refcount_finish_one_cleanup(tp, rcur, error);
-	set_bit(XFS_CUI_RECOVERED, &cuip->cui_flags);
 	xfs_defer_move(parent_tp, tp);
 	error = xfs_trans_commit(tp);
 	return error;
@@ -590,3 +565,117 @@ abort_error:
 	xfs_trans_cancel(tp);
 	return error;
 }
+
+STATIC bool
+xfs_cui_item_match(
+	struct xfs_log_item	*lip,
+	uint64_t		intent_id)
+{
+	return CUI_ITEM(lip)->cui_format.cui_id == intent_id;
+}
+
+static const struct xfs_item_ops xfs_cui_item_ops = {
+	.iop_size	= xfs_cui_item_size,
+	.iop_format	= xfs_cui_item_format,
+	.iop_unpin	= xfs_cui_item_unpin,
+	.iop_release	= xfs_cui_item_release,
+	.iop_recover	= xfs_cui_item_recover,
+	.iop_match	= xfs_cui_item_match,
+};
+
+/*
+ * Copy an CUI format buffer from the given buf, and into the destination
+ * CUI format structure.  The CUI/CUD items were designed not to need any
+ * special alignment handling.
+ */
+static int
+xfs_cui_copy_format(
+	struct xfs_log_iovec		*buf,
+	struct xfs_cui_log_format	*dst_cui_fmt)
+{
+	struct xfs_cui_log_format	*src_cui_fmt;
+	uint				len;
+
+	src_cui_fmt = buf->i_addr;
+	len = xfs_cui_log_format_sizeof(src_cui_fmt->cui_nextents);
+
+	if (buf->i_len == len) {
+		memcpy(dst_cui_fmt, src_cui_fmt, len);
+		return 0;
+	}
+	XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, NULL);
+	return -EFSCORRUPTED;
+}
+
+/*
+ * This routine is called to create an in-core extent refcount update
+ * item from the cui format structure which was logged on disk.
+ * It allocates an in-core cui, copies the extents from the format
+ * structure into it, and adds the cui to the AIL with the given
+ * LSN.
+ */
+STATIC int
+xlog_recover_cui_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	int				error;
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_cui_log_item		*cuip;
+	struct xfs_cui_log_format	*cui_formatp;
+
+	cui_formatp = item->ri_buf[0].i_addr;
+
+	cuip = xfs_cui_init(mp, cui_formatp->cui_nextents);
+	error = xfs_cui_copy_format(&item->ri_buf[0], &cuip->cui_format);
+	if (error) {
+		xfs_cui_item_free(cuip);
+		return error;
+	}
+	atomic_set(&cuip->cui_next_extent, cui_formatp->cui_nextents);
+	/*
+	 * Insert the intent into the AIL directly and drop one reference so
+	 * that finishing or canceling the work will drop the other.
+	 */
+	xfs_trans_ail_insert(log->l_ailp, &cuip->cui_item, lsn);
+	xfs_cui_release(cuip);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_cui_item_ops = {
+	.item_type		= XFS_LI_CUI,
+	.commit_pass2		= xlog_recover_cui_commit_pass2,
+};
+
+/*
+ * This routine is called when an CUD format structure is found in a committed
+ * transaction in the log. Its purpose is to cancel the corresponding CUI if it
+ * was still in the log. To do this it searches the AIL for the CUI with an id
+ * equal to that in the CUD format structure. If we find it we drop the CUD
+ * reference, which removes the CUI from the AIL and frees it.
+ */
+STATIC int
+xlog_recover_cud_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_cud_log_format	*cud_formatp;
+
+	cud_formatp = item->ri_buf[0].i_addr;
+	if (item->ri_buf[0].i_len != sizeof(struct xfs_cud_log_format)) {
+		XFS_ERROR_REPORT(__func__, XFS_ERRLEVEL_LOW, log->l_mp);
+		return -EFSCORRUPTED;
+	}
+
+	xlog_recover_release_intent(log, XFS_LI_CUI, cud_formatp->cud_cui_id);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_cud_item_ops = {
+	.item_type		= XFS_LI_CUD,
+	.commit_pass2		= xlog_recover_cud_commit_pass2,
+};
diff --git a/fs/xfs/xfs_refcount_item.h b/fs/xfs/xfs_refcount_item.h
index e47530f30489..f4f2e836540b 100644
--- a/fs/xfs/xfs_refcount_item.h
+++ b/fs/xfs/xfs_refcount_item.h
@@ -33,11 +33,6 @@ struct kmem_zone;
 #define	XFS_CUI_MAX_FAST_EXTENTS	16
 
 /*
- * Define CUI flag bits. Manipulated by set/clear/test_bit operators.
- */
-#define	XFS_CUI_RECOVERED		1
-
-/*
  * This is the "refcount update intent" log item.  It is used to log
  * the fact that some reverse mappings need to change.  It is used in
  * conjunction with the "refcount update done" log item described
@@ -51,7 +46,6 @@ struct xfs_cui_log_item {
 	struct xfs_log_item		cui_item;
 	atomic_t			cui_refcount;
 	atomic_t			cui_next_extent;
-	unsigned long			cui_flags;	/* misc flags */
 	struct xfs_cui_log_format	cui_format;
 };
 
@@ -77,9 +71,4 @@ struct xfs_cud_log_item {
 extern struct kmem_zone	*xfs_cui_zone;
 extern struct kmem_zone	*xfs_cud_zone;
 
-struct xfs_cui_log_item *xfs_cui_init(struct xfs_mount *, uint);
-void xfs_cui_item_free(struct xfs_cui_log_item *);
-void xfs_cui_release(struct xfs_cui_log_item *);
-int xfs_cui_recover(struct xfs_trans *parent_tp, struct xfs_cui_log_item *cuip);
-
 #endif	/* __XFS_REFCOUNT_ITEM_H__ */
diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index b0ce04ffd3cd..16098dc42add 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -179,7 +179,7 @@ xfs_reflink_trim_around_shared(
 	int			error = 0;
 
 	/* Holes, unwritten, and delalloc extents cannot be shared */
-	if (!xfs_is_cow_inode(ip) || !xfs_bmap_is_real_extent(irec)) {
+	if (!xfs_is_cow_inode(ip) || !xfs_bmap_is_written_extent(irec)) {
 		*shared = false;
 		return 0;
 	}
@@ -655,7 +655,7 @@ xfs_reflink_end_cow_extent(
 	 * preallocations can leak into the range we are called upon, and we
 	 * need to skip them.
 	 */
-	if (!xfs_bmap_is_real_extent(&got)) {
+	if (!xfs_bmap_is_written_extent(&got)) {
 		*end_fsb = del.br_startoff;
 		goto out_cancel;
 	}
@@ -721,7 +721,7 @@ xfs_reflink_end_cow(
 	 * repeatedly cycles the ILOCK to allocate one transaction per remapped
 	 * extent.
 	 *
-	 * If we're being called by writeback then the the pages will still
+	 * If we're being called by writeback then the pages will still
 	 * have PageWriteback set, which prevents races with reflink remapping
 	 * and truncate.  Reflink remapping prevents races with writeback by
 	 * taking the iolock and mmaplock before flushing the pages and
@@ -984,40 +984,28 @@ xfs_reflink_ag_has_free_space(
 }
 
 /*
- * Unmap a range of blocks from a file, then map other blocks into the hole.
- * The range to unmap is (destoff : destoff + srcioff + irec->br_blockcount).
- * The extent irec is mapped into dest at irec->br_startoff.
+ * Remap the given extent into the file.  The dmap blockcount will be set to
+ * the number of blocks that were actually remapped.
  */
 STATIC int
 xfs_reflink_remap_extent(
 	struct xfs_inode	*ip,
-	struct xfs_bmbt_irec	*irec,
-	xfs_fileoff_t		destoff,
+	struct xfs_bmbt_irec	*dmap,
 	xfs_off_t		new_isize)
 {
+	struct xfs_bmbt_irec	smap;
 	struct xfs_mount	*mp = ip->i_mount;
-	bool			real_extent = xfs_bmap_is_real_extent(irec);
 	struct xfs_trans	*tp;
-	unsigned int		resblks;
-	struct xfs_bmbt_irec	uirec;
-	xfs_filblks_t		rlen;
-	xfs_filblks_t		unmap_len;
 	xfs_off_t		newlen;
+	int64_t			qres, qdelta;
+	unsigned int		resblks;
+	bool			smap_real;
+	bool			dmap_written = xfs_bmap_is_written_extent(dmap);
+	int			nimaps;
 	int			error;
 
-	unmap_len = irec->br_startoff + irec->br_blockcount - destoff;
-	trace_xfs_reflink_punch_range(ip, destoff, unmap_len);
-
-	/* No reflinking if we're low on space */
-	if (real_extent) {
-		error = xfs_reflink_ag_has_free_space(mp,
-				XFS_FSB_TO_AGNO(mp, irec->br_startblock));
-		if (error)
-			goto out;
-	}
-
 	/* Start a rolling transaction to switch the mappings */
-	resblks = XFS_EXTENTADD_SPACE_RES(ip->i_mount, XFS_DATA_FORK);
+	resblks = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
 	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_write, resblks, 0, 0, &tp);
 	if (error)
 		goto out;
@@ -1025,86 +1013,147 @@ xfs_reflink_remap_extent(
 	xfs_ilock(ip, XFS_ILOCK_EXCL);
 	xfs_trans_ijoin(tp, ip, 0);
 
-	/* If we're not just clearing space, then do we have enough quota? */
-	if (real_extent) {
-		error = xfs_trans_reserve_quota_nblks(tp, ip,
-				irec->br_blockcount, 0, XFS_QMOPT_RES_REGBLKS);
+	/*
+	 * Read what's currently mapped in the destination file into smap.
+	 * If smap isn't a hole, we will have to remove it before we can add
+	 * dmap to the destination file.
+	 */
+	nimaps = 1;
+	error = xfs_bmapi_read(ip, dmap->br_startoff, dmap->br_blockcount,
+			&smap, &nimaps, 0);
+	if (error)
+		goto out_cancel;
+	ASSERT(nimaps == 1 && smap.br_startoff == dmap->br_startoff);
+	smap_real = xfs_bmap_is_real_extent(&smap);
+
+	/*
+	 * We can only remap as many blocks as the smaller of the two extent
+	 * maps, because we can only remap one extent at a time.
+	 */
+	dmap->br_blockcount = min(dmap->br_blockcount, smap.br_blockcount);
+	ASSERT(dmap->br_blockcount == smap.br_blockcount);
+
+	trace_xfs_reflink_remap_extent_dest(ip, &smap);
+
+	/*
+	 * Two extents mapped to the same physical block must not have
+	 * different states; that's filesystem corruption.  Move on to the next
+	 * extent if they're both holes or both the same physical extent.
+	 */
+	if (dmap->br_startblock == smap.br_startblock) {
+		if (dmap->br_state != smap.br_state)
+			error = -EFSCORRUPTED;
+		goto out_cancel;
+	}
+
+	/* If both extents are unwritten, leave them alone. */
+	if (dmap->br_state == XFS_EXT_UNWRITTEN &&
+	    smap.br_state == XFS_EXT_UNWRITTEN)
+		goto out_cancel;
+
+	/* No reflinking if the AG of the dest mapping is low on space. */
+	if (dmap_written) {
+		error = xfs_reflink_ag_has_free_space(mp,
+				XFS_FSB_TO_AGNO(mp, dmap->br_startblock));
 		if (error)
 			goto out_cancel;
 	}
 
-	trace_xfs_reflink_remap(ip, irec->br_startoff,
-				irec->br_blockcount, irec->br_startblock);
-
-	/* Unmap the old blocks in the data fork. */
-	rlen = unmap_len;
-	while (rlen) {
-		ASSERT(tp->t_firstblock == NULLFSBLOCK);
-		error = __xfs_bunmapi(tp, ip, destoff, &rlen, 0, 1);
+	/*
+	 * Compute quota reservation if we think the quota block counter for
+	 * this file could increase.
+	 *
+	 * Adding a written extent to the extent map can cause a bmbt split,
+	 * and removing a mapped extent from the extent can cause a bmbt split.
+	 * The two operations cannot both cause a split since they operate on
+	 * the same index in the bmap btree, so we only need a reservation for
+	 * one bmbt split if either thing is happening.
+	 *
+	 * If we are mapping a written extent into the file, we need to have
+	 * enough quota block count reservation to handle the blocks in that
+	 * extent.  We log only the delta to the quota block counts, so if the
+	 * extent we're unmapping also has blocks allocated to it, we don't
+	 * need a quota reservation for the extent itself.
+	 *
+	 * Note that if we're replacing a delalloc reservation with a written
+	 * extent, we have to take the full quota reservation because removing
+	 * the delalloc reservation gives the block count back to the quota
+	 * count.  This is suboptimal, but the VFS flushed the dest range
+	 * before we started.  That should have removed all the delalloc
+	 * reservations, but we code defensively.
+	 */
+	qres = qdelta = 0;
+	if (smap_real || dmap_written)
+		qres = XFS_EXTENTADD_SPACE_RES(mp, XFS_DATA_FORK);
+	if (!smap_real && dmap_written)
+		qres += dmap->br_blockcount;
+	if (qres > 0) {
+		error = xfs_trans_reserve_quota_nblks(tp, ip, qres, 0,
+				XFS_QMOPT_RES_REGBLKS);
 		if (error)
 			goto out_cancel;
+	}
 
+	if (smap_real) {
 		/*
-		 * Trim the extent to whatever got unmapped.
-		 * Remember, bunmapi works backwards.
+		 * If the extent we're unmapping is backed by storage (written
+		 * or not), unmap the extent and drop its refcount.
 		 */
-		uirec.br_startblock = irec->br_startblock + rlen;
-		uirec.br_startoff = irec->br_startoff + rlen;
-		uirec.br_blockcount = unmap_len - rlen;
-		unmap_len = rlen;
-
-		/* If this isn't a real mapping, we're done. */
-		if (!real_extent || uirec.br_blockcount == 0)
-			goto next_extent;
+		xfs_bmap_unmap_extent(tp, ip, &smap);
+		xfs_refcount_decrease_extent(tp, &smap);
+		qdelta -= smap.br_blockcount;
+	} else if (smap.br_startblock == DELAYSTARTBLOCK) {
+		xfs_filblks_t	len = smap.br_blockcount;
 
-		trace_xfs_reflink_remap(ip, uirec.br_startoff,
-				uirec.br_blockcount, uirec.br_startblock);
-
-		/* Update the refcount tree */
-		xfs_refcount_increase_extent(tp, &uirec);
-
-		/* Map the new blocks into the data fork. */
-		xfs_bmap_map_extent(tp, ip, &uirec);
+		/*
+		 * If the extent we're unmapping is a delalloc reservation,
+		 * we can use the regular bunmapi function to release the
+		 * incore state.  Dropping the delalloc reservation takes care
+		 * of the quota reservation for us.
+		 */
+		error = __xfs_bunmapi(NULL, ip, smap.br_startoff, &len, 0, 1);
+		if (error)
+			goto out_cancel;
+		ASSERT(len == 0);
+	}
 
-		/* Update quota accounting. */
-		xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT,
-				uirec.br_blockcount);
+	/*
+	 * If the extent we're sharing is backed by written storage, increase
+	 * its refcount and map it into the file.
+	 */
+	if (dmap_written) {
+		xfs_refcount_increase_extent(tp, dmap);
+		xfs_bmap_map_extent(tp, ip, dmap);
+		qdelta += dmap->br_blockcount;
+	}
 
-		/* Update dest isize if needed. */
-		newlen = XFS_FSB_TO_B(mp,
-				uirec.br_startoff + uirec.br_blockcount);
-		newlen = min_t(xfs_off_t, newlen, new_isize);
-		if (newlen > i_size_read(VFS_I(ip))) {
-			trace_xfs_reflink_update_inode_size(ip, newlen);
-			i_size_write(VFS_I(ip), newlen);
-			ip->i_d.di_size = newlen;
-			xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
-		}
+	xfs_trans_mod_dquot_byino(tp, ip, XFS_TRANS_DQ_BCOUNT, qdelta);
 
-next_extent:
-		/* Process all the deferred stuff. */
-		error = xfs_defer_finish(&tp);
-		if (error)
-			goto out_cancel;
+	/* Update dest isize if needed. */
+	newlen = XFS_FSB_TO_B(mp, dmap->br_startoff + dmap->br_blockcount);
+	newlen = min_t(xfs_off_t, newlen, new_isize);
+	if (newlen > i_size_read(VFS_I(ip))) {
+		trace_xfs_reflink_update_inode_size(ip, newlen);
+		i_size_write(VFS_I(ip), newlen);
+		ip->i_d.di_size = newlen;
+		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 	}
 
+	/* Commit everything and unlock. */
 	error = xfs_trans_commit(tp);
-	xfs_iunlock(ip, XFS_ILOCK_EXCL);
-	if (error)
-		goto out;
-	return 0;
+	goto out_unlock;
 
 out_cancel:
 	xfs_trans_cancel(tp);
+out_unlock:
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 out:
-	trace_xfs_reflink_remap_extent_error(ip, error, _RET_IP_);
+	if (error)
+		trace_xfs_reflink_remap_extent_error(ip, error, _RET_IP_);
 	return error;
 }
 
-/*
- * Iteratively remap one file's extents (and holes) to another's.
- */
+/* Remap a range of one file to the other. */
 int
 xfs_reflink_remap_blocks(
 	struct xfs_inode	*src,
@@ -1115,25 +1164,22 @@ xfs_reflink_remap_blocks(
 	loff_t			*remapped)
 {
 	struct xfs_bmbt_irec	imap;
-	xfs_fileoff_t		srcoff;
-	xfs_fileoff_t		destoff;
+	struct xfs_mount	*mp = src->i_mount;
+	xfs_fileoff_t		srcoff = XFS_B_TO_FSBT(mp, pos_in);
+	xfs_fileoff_t		destoff = XFS_B_TO_FSBT(mp, pos_out);
 	xfs_filblks_t		len;
-	xfs_filblks_t		range_len;
 	xfs_filblks_t		remapped_len = 0;
 	xfs_off_t		new_isize = pos_out + remap_len;
 	int			nimaps;
 	int			error = 0;
 
-	destoff = XFS_B_TO_FSBT(src->i_mount, pos_out);
-	srcoff = XFS_B_TO_FSBT(src->i_mount, pos_in);
-	len = XFS_B_TO_FSB(src->i_mount, remap_len);
+	len = min_t(xfs_filblks_t, XFS_B_TO_FSB(mp, remap_len),
+			XFS_MAX_FILEOFF);
 
-	/* drange = (destoff, destoff + len); srange = (srcoff, srcoff + len) */
-	while (len) {
-		uint		lock_mode;
+	trace_xfs_reflink_remap_blocks(src, srcoff, len, dest, destoff);
 
-		trace_xfs_reflink_remap_blocks_loop(src, srcoff, len,
-				dest, destoff);
+	while (len > 0) {
+		unsigned int	lock_mode;
 
 		/* Read extent from the source file */
 		nimaps = 1;
@@ -1142,18 +1188,25 @@ xfs_reflink_remap_blocks(
 		xfs_iunlock(src, lock_mode);
 		if (error)
 			break;
-		ASSERT(nimaps == 1);
-
-		trace_xfs_reflink_remap_imap(src, srcoff, len, XFS_DATA_FORK,
-				&imap);
+		/*
+		 * The caller supposedly flushed all dirty pages in the source
+		 * file range, which means that writeback should have allocated
+		 * or deleted all delalloc reservations in that range.  If we
+		 * find one, that's a good sign that something is seriously
+		 * wrong here.
+		 */
+		ASSERT(nimaps == 1 && imap.br_startoff == srcoff);
+		if (imap.br_startblock == DELAYSTARTBLOCK) {
+			ASSERT(imap.br_startblock != DELAYSTARTBLOCK);
+			error = -EFSCORRUPTED;
+			break;
+		}
 
-		/* Translate imap into the destination file. */
-		range_len = imap.br_startoff + imap.br_blockcount - srcoff;
-		imap.br_startoff += destoff - srcoff;
+		trace_xfs_reflink_remap_extent_src(src, &imap);
 
-		/* Clear dest from destoff to the end of imap and map it in. */
-		error = xfs_reflink_remap_extent(dest, &imap, destoff,
-				new_isize);
+		/* Remap into the destination file at the given offset. */
+		imap.br_startoff = destoff;
+		error = xfs_reflink_remap_extent(dest, &imap, new_isize);
 		if (error)
 			break;
 
@@ -1163,10 +1216,10 @@ xfs_reflink_remap_blocks(
 		}
 
 		/* Advance drange/srange */
-		srcoff += range_len;
-		destoff += range_len;
-		len -= range_len;
-		remapped_len += range_len;
+		srcoff += imap.br_blockcount;
+		destoff += imap.br_blockcount;
+		len -= imap.br_blockcount;
+		remapped_len += imap.br_blockcount;
 	}
 
 	if (error)
@@ -1177,81 +1230,6 @@ xfs_reflink_remap_blocks(
 }
 
 /*
- * Grab the exclusive iolock for a data copy from src to dest, making sure to
- * abide vfs locking order (lowest pointer value goes first) and breaking the
- * layout leases before proceeding.  The loop is needed because we cannot call
- * the blocking break_layout() with the iolocks held, and therefore have to
- * back out both locks.
- */
-static int
-xfs_iolock_two_inodes_and_break_layout(
-	struct inode		*src,
-	struct inode		*dest)
-{
-	int			error;
-
-	if (src > dest)
-		swap(src, dest);
-
-retry:
-	/* Wait to break both inodes' layouts before we start locking. */
-	error = break_layout(src, true);
-	if (error)
-		return error;
-	if (src != dest) {
-		error = break_layout(dest, true);
-		if (error)
-			return error;
-	}
-
-	/* Lock one inode and make sure nobody got in and leased it. */
-	inode_lock(src);
-	error = break_layout(src, false);
-	if (error) {
-		inode_unlock(src);
-		if (error == -EWOULDBLOCK)
-			goto retry;
-		return error;
-	}
-
-	if (src == dest)
-		return 0;
-
-	/* Lock the other inode and make sure nobody got in and leased it. */
-	inode_lock_nested(dest, I_MUTEX_NONDIR2);
-	error = break_layout(dest, false);
-	if (error) {
-		inode_unlock(src);
-		inode_unlock(dest);
-		if (error == -EWOULDBLOCK)
-			goto retry;
-		return error;
-	}
-
-	return 0;
-}
-
-/* Unlock both inodes after they've been prepped for a range clone. */
-void
-xfs_reflink_remap_unlock(
-	struct file		*file_in,
-	struct file		*file_out)
-{
-	struct inode		*inode_in = file_inode(file_in);
-	struct xfs_inode	*src = XFS_I(inode_in);
-	struct inode		*inode_out = file_inode(file_out);
-	struct xfs_inode	*dest = XFS_I(inode_out);
-	bool			same_inode = (inode_in == inode_out);
-
-	xfs_iunlock(dest, XFS_MMAPLOCK_EXCL);
-	if (!same_inode)
-		xfs_iunlock(src, XFS_MMAPLOCK_EXCL);
-	inode_unlock(inode_out);
-	if (!same_inode)
-		inode_unlock(inode_in);
-}
-
-/*
  * If we're reflinking to a point past the destination file's EOF, we must
  * zero any speculative post-EOF preallocations that sit between the old EOF
  * and the destination file offset.
@@ -1313,18 +1291,12 @@ xfs_reflink_remap_prep(
 	struct xfs_inode	*src = XFS_I(inode_in);
 	struct inode		*inode_out = file_inode(file_out);
 	struct xfs_inode	*dest = XFS_I(inode_out);
-	bool			same_inode = (inode_in == inode_out);
-	ssize_t			ret;
+	int			ret;
 
 	/* Lock both files against IO */
-	ret = xfs_iolock_two_inodes_and_break_layout(inode_in, inode_out);
+	ret = xfs_ilock2_io_mmap(src, dest);
 	if (ret)
 		return ret;
-	if (same_inode)
-		xfs_ilock(src, XFS_MMAPLOCK_EXCL);
-	else
-		xfs_lock_two_inodes(src, XFS_MMAPLOCK_EXCL, dest,
-				XFS_MMAPLOCK_EXCL);
 
 	/* Check file eligibility and prepare for block sharing. */
 	ret = -EINVAL;
@@ -1338,7 +1310,7 @@ xfs_reflink_remap_prep(
 
 	ret = generic_remap_file_range_prep(file_in, pos_in, file_out, pos_out,
 			len, remap_flags);
-	if (ret < 0 || *len == 0)
+	if (ret || *len == 0)
 		goto out_unlock;
 
 	/* Attach dquots to dest inode before changing block map */
@@ -1373,9 +1345,9 @@ xfs_reflink_remap_prep(
 	if (ret)
 		goto out_unlock;
 
-	return 1;
+	return 0;
 out_unlock:
-	xfs_reflink_remap_unlock(file_in, file_out);
+	xfs_iunlock2_io_mmap(src, dest);
 	return ret;
 }
 
diff --git a/fs/xfs/xfs_reflink.h b/fs/xfs/xfs_reflink.h
index 3e4fd46373ab..487b00434b96 100644
--- a/fs/xfs/xfs_reflink.h
+++ b/fs/xfs/xfs_reflink.h
@@ -56,7 +56,5 @@ extern int xfs_reflink_remap_blocks(struct xfs_inode *src, loff_t pos_in,
 		loff_t *remapped);
 extern int xfs_reflink_update_dest(struct xfs_inode *dest, xfs_off_t newlen,
 		xfs_extlen_t cowextsize, unsigned int remap_flags);
-extern void xfs_reflink_remap_unlock(struct file *file_in,
-		struct file *file_out);
 
 #endif /* __XFS_REFLINK_H */
diff --git a/fs/xfs/xfs_rmap_item.c b/fs/xfs/xfs_rmap_item.c
index 4911b68f95dd..dc5b0753cd51 100644
--- a/fs/xfs/xfs_rmap_item.c
+++ b/fs/xfs/xfs_rmap_item.c
@@ -18,16 +18,20 @@
 #include "xfs_log.h"
 #include "xfs_rmap.h"
 #include "xfs_error.h"
+#include "xfs_log_priv.h"
+#include "xfs_log_recover.h"
 
 kmem_zone_t	*xfs_rui_zone;
 kmem_zone_t	*xfs_rud_zone;
 
+static const struct xfs_item_ops xfs_rui_item_ops;
+
 static inline struct xfs_rui_log_item *RUI_ITEM(struct xfs_log_item *lip)
 {
 	return container_of(lip, struct xfs_rui_log_item, rui_item);
 }
 
-void
+STATIC void
 xfs_rui_item_free(
 	struct xfs_rui_log_item	*ruip)
 {
@@ -44,13 +48,13 @@ xfs_rui_item_free(
  * committed vs unpin operations in bulk insert operations. Hence the reference
  * count to ensure only the last caller frees the RUI.
  */
-void
+STATIC void
 xfs_rui_release(
 	struct xfs_rui_log_item	*ruip)
 {
 	ASSERT(atomic_read(&ruip->rui_refcount) > 0);
 	if (atomic_dec_and_test(&ruip->rui_refcount)) {
-		xfs_trans_ail_remove(&ruip->rui_item, SHUTDOWN_LOG_IO_ERROR);
+		xfs_trans_ail_delete(&ruip->rui_item, SHUTDOWN_LOG_IO_ERROR);
 		xfs_rui_item_free(ruip);
 	}
 }
@@ -122,17 +126,10 @@ xfs_rui_item_release(
 	xfs_rui_release(RUI_ITEM(lip));
 }
 
-static const struct xfs_item_ops xfs_rui_item_ops = {
-	.iop_size	= xfs_rui_item_size,
-	.iop_format	= xfs_rui_item_format,
-	.iop_unpin	= xfs_rui_item_unpin,
-	.iop_release	= xfs_rui_item_release,
-};
-
 /*
  * Allocate and initialize an rui item with the given number of extents.
  */
-struct xfs_rui_log_item *
+STATIC struct xfs_rui_log_item *
 xfs_rui_init(
 	struct xfs_mount		*mp,
 	uint				nextents)
@@ -144,7 +141,8 @@ xfs_rui_init(
 	if (nextents > XFS_RUI_MAX_FAST_EXTENTS)
 		ruip = kmem_zalloc(xfs_rui_log_item_sizeof(nextents), 0);
 	else
-		ruip = kmem_zone_zalloc(xfs_rui_zone, 0);
+		ruip = kmem_cache_zalloc(xfs_rui_zone,
+					 GFP_KERNEL | __GFP_NOFAIL);
 
 	xfs_log_item_init(mp, &ruip->rui_item, XFS_LI_RUI, &xfs_rui_item_ops);
 	ruip->rui_format.rui_nextents = nextents;
@@ -160,7 +158,7 @@ xfs_rui_init(
  * RUI format structure.  The RUI/RUD items were designed not to need any
  * special alignment handling.
  */
-int
+STATIC int
 xfs_rui_copy_format(
 	struct xfs_log_iovec		*buf,
 	struct xfs_rui_log_format	*dst_rui_fmt)
@@ -246,7 +244,7 @@ xfs_trans_get_rud(
 {
 	struct xfs_rud_log_item		*rudp;
 
-	rudp = kmem_zone_zalloc(xfs_rud_zone, 0);
+	rudp = kmem_cache_zalloc(xfs_rud_zone, GFP_KERNEL | __GFP_NOFAIL);
 	xfs_log_item_init(tp->t_mountp, &rudp->rud_item, XFS_LI_RUD,
 			  &xfs_rud_item_ops);
 	rudp->rud_ruip = ruip;
@@ -352,41 +350,16 @@ xfs_rmap_update_diff_items(
 		XFS_FSB_TO_AGNO(mp, rb->ri_bmap.br_startblock);
 }
 
-/* Get an RUI. */
-STATIC void *
-xfs_rmap_update_create_intent(
-	struct xfs_trans		*tp,
-	unsigned int			count)
-{
-	struct xfs_rui_log_item		*ruip;
-
-	ASSERT(tp != NULL);
-	ASSERT(count > 0);
-
-	ruip = xfs_rui_init(tp->t_mountp, count);
-	ASSERT(ruip != NULL);
-
-	/*
-	 * Get a log_item_desc to point at the new item.
-	 */
-	xfs_trans_add_item(tp, &ruip->rui_item);
-	return ruip;
-}
-
 /* Log rmap updates in the intent item. */
 STATIC void
 xfs_rmap_update_log_item(
 	struct xfs_trans		*tp,
-	void				*intent,
-	struct list_head		*item)
+	struct xfs_rui_log_item		*ruip,
+	struct xfs_rmap_intent		*rmap)
 {
-	struct xfs_rui_log_item		*ruip = intent;
-	struct xfs_rmap_intent		*rmap;
 	uint				next_extent;
 	struct xfs_map_extent		*map;
 
-	rmap = container_of(item, struct xfs_rmap_intent, ri_list);
-
 	tp->t_flags |= XFS_TRANS_DIRTY;
 	set_bit(XFS_LI_DIRTY, &ruip->rui_item.li_flags);
 
@@ -406,58 +379,64 @@ xfs_rmap_update_log_item(
 			rmap->ri_bmap.br_state);
 }
 
+static struct xfs_log_item *
+xfs_rmap_update_create_intent(
+	struct xfs_trans		*tp,
+	struct list_head		*items,
+	unsigned int			count,
+	bool				sort)
+{
+	struct xfs_mount		*mp = tp->t_mountp;
+	struct xfs_rui_log_item		*ruip = xfs_rui_init(mp, count);
+	struct xfs_rmap_intent		*rmap;
+
+	ASSERT(count > 0);
+
+	xfs_trans_add_item(tp, &ruip->rui_item);
+	if (sort)
+		list_sort(mp, items, xfs_rmap_update_diff_items);
+	list_for_each_entry(rmap, items, ri_list)
+		xfs_rmap_update_log_item(tp, ruip, rmap);
+	return &ruip->rui_item;
+}
+
 /* Get an RUD so we can process all the deferred rmap updates. */
-STATIC void *
+static struct xfs_log_item *
 xfs_rmap_update_create_done(
 	struct xfs_trans		*tp,
-	void				*intent,
+	struct xfs_log_item		*intent,
 	unsigned int			count)
 {
-	return xfs_trans_get_rud(tp, intent);
+	return &xfs_trans_get_rud(tp, RUI_ITEM(intent))->rud_item;
 }
 
 /* Process a deferred rmap update. */
 STATIC int
 xfs_rmap_update_finish_item(
 	struct xfs_trans		*tp,
+	struct xfs_log_item		*done,
 	struct list_head		*item,
-	void				*done_item,
-	void				**state)
+	struct xfs_btree_cur		**state)
 {
 	struct xfs_rmap_intent		*rmap;
 	int				error;
 
 	rmap = container_of(item, struct xfs_rmap_intent, ri_list);
-	error = xfs_trans_log_finish_rmap_update(tp, done_item,
-			rmap->ri_type,
-			rmap->ri_owner, rmap->ri_whichfork,
-			rmap->ri_bmap.br_startoff,
-			rmap->ri_bmap.br_startblock,
-			rmap->ri_bmap.br_blockcount,
-			rmap->ri_bmap.br_state,
-			(struct xfs_btree_cur **)state);
+	error = xfs_trans_log_finish_rmap_update(tp, RUD_ITEM(done),
+			rmap->ri_type, rmap->ri_owner, rmap->ri_whichfork,
+			rmap->ri_bmap.br_startoff, rmap->ri_bmap.br_startblock,
+			rmap->ri_bmap.br_blockcount, rmap->ri_bmap.br_state,
+			state);
 	kmem_free(rmap);
 	return error;
 }
 
-/* Clean up after processing deferred rmaps. */
-STATIC void
-xfs_rmap_update_finish_cleanup(
-	struct xfs_trans	*tp,
-	void			*state,
-	int			error)
-{
-	struct xfs_btree_cur	*rcur = state;
-
-	xfs_rmap_finish_one_cleanup(tp, rcur, error);
-}
-
 /* Abort all pending RUIs. */
 STATIC void
 xfs_rmap_update_abort_intent(
-	void				*intent)
+	struct xfs_log_item	*intent)
 {
-	xfs_rui_release(intent);
+	xfs_rui_release(RUI_ITEM(intent));
 }
 
 /* Cancel a deferred rmap update. */
@@ -473,13 +452,11 @@ xfs_rmap_update_cancel_item(
 
 const struct xfs_defer_op_type xfs_rmap_update_defer_type = {
 	.max_items	= XFS_RUI_MAX_FAST_EXTENTS,
-	.diff_items	= xfs_rmap_update_diff_items,
 	.create_intent	= xfs_rmap_update_create_intent,
 	.abort_intent	= xfs_rmap_update_abort_intent,
-	.log_item	= xfs_rmap_update_log_item,
 	.create_done	= xfs_rmap_update_create_done,
 	.finish_item	= xfs_rmap_update_finish_item,
-	.finish_cleanup = xfs_rmap_update_finish_cleanup,
+	.finish_cleanup = xfs_rmap_finish_one_cleanup,
 	.cancel_item	= xfs_rmap_update_cancel_item,
 };
 
@@ -487,24 +464,24 @@ const struct xfs_defer_op_type xfs_rmap_update_defer_type = {
  * Process an rmap update intent item that was recovered from the log.
  * We need to update the rmapbt.
  */
-int
-xfs_rui_recover(
-	struct xfs_mount		*mp,
-	struct xfs_rui_log_item		*ruip)
+STATIC int
+xfs_rui_item_recover(
+	struct xfs_log_item		*lip,
+	struct xfs_trans		*parent_tp)
 {
-	int				i;
-	int				error = 0;
+	struct xfs_rui_log_item		*ruip = RUI_ITEM(lip);
 	struct xfs_map_extent		*rmap;
-	xfs_fsblock_t			startblock_fsb;
-	bool				op_ok;
 	struct xfs_rud_log_item		*rudp;
-	enum xfs_rmap_intent_type	type;
-	int				whichfork;
-	xfs_exntst_t			state;
 	struct xfs_trans		*tp;
 	struct xfs_btree_cur		*rcur = NULL;
-
-	ASSERT(!test_bit(XFS_RUI_RECOVERED, &ruip->rui_flags));
+	struct xfs_mount		*mp = parent_tp->t_mountp;
+	xfs_fsblock_t			startblock_fsb;
+	enum xfs_rmap_intent_type	type;
+	xfs_exntst_t			state;
+	bool				op_ok;
+	int				i;
+	int				whichfork;
+	int				error = 0;
 
 	/*
 	 * First check the validity of the extents described by the
@@ -539,7 +516,6 @@ xfs_rui_recover(
 			 * This will pull the RUI from the AIL and
 			 * free the memory associated with it.
 			 */
-			set_bit(XFS_RUI_RECOVERED, &ruip->rui_flags);
 			xfs_rui_release(ruip);
 			return -EFSCORRUPTED;
 		}
@@ -597,7 +573,6 @@ xfs_rui_recover(
 	}
 
 	xfs_rmap_finish_one_cleanup(tp, rcur, error);
-	set_bit(XFS_RUI_RECOVERED, &ruip->rui_flags);
 	error = xfs_trans_commit(tp);
 	return error;
 
@@ -606,3 +581,90 @@ abort_error:
 	xfs_trans_cancel(tp);
 	return error;
 }
+
+STATIC bool
+xfs_rui_item_match(
+	struct xfs_log_item	*lip,
+	uint64_t		intent_id)
+{
+	return RUI_ITEM(lip)->rui_format.rui_id == intent_id;
+}
+
+static const struct xfs_item_ops xfs_rui_item_ops = {
+	.iop_size	= xfs_rui_item_size,
+	.iop_format	= xfs_rui_item_format,
+	.iop_unpin	= xfs_rui_item_unpin,
+	.iop_release	= xfs_rui_item_release,
+	.iop_recover	= xfs_rui_item_recover,
+	.iop_match	= xfs_rui_item_match,
+};
+
+/*
+ * This routine is called to create an in-core extent rmap update
+ * item from the rui format structure which was logged on disk.
+ * It allocates an in-core rui, copies the extents from the format
+ * structure into it, and adds the rui to the AIL with the given
+ * LSN.
+ */
+STATIC int
+xlog_recover_rui_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	int				error;
+	struct xfs_mount		*mp = log->l_mp;
+	struct xfs_rui_log_item		*ruip;
+	struct xfs_rui_log_format	*rui_formatp;
+
+	rui_formatp = item->ri_buf[0].i_addr;
+
+	ruip = xfs_rui_init(mp, rui_formatp->rui_nextents);
+	error = xfs_rui_copy_format(&item->ri_buf[0], &ruip->rui_format);
+	if (error) {
+		xfs_rui_item_free(ruip);
+		return error;
+	}
+	atomic_set(&ruip->rui_next_extent, rui_formatp->rui_nextents);
+	/*
+	 * Insert the intent into the AIL directly and drop one reference so
+	 * that finishing or canceling the work will drop the other.
+	 */
+	xfs_trans_ail_insert(log->l_ailp, &ruip->rui_item, lsn);
+	xfs_rui_release(ruip);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_rui_item_ops = {
+	.item_type		= XFS_LI_RUI,
+	.commit_pass2		= xlog_recover_rui_commit_pass2,
+};
+
+/*
+ * This routine is called when an RUD format structure is found in a committed
+ * transaction in the log. Its purpose is to cancel the corresponding RUI if it
+ * was still in the log. To do this it searches the AIL for the RUI with an id
+ * equal to that in the RUD format structure. If we find it we drop the RUD
+ * reference, which removes the RUI from the AIL and frees it.
+ */
+STATIC int
+xlog_recover_rud_commit_pass2(
+	struct xlog			*log,
+	struct list_head		*buffer_list,
+	struct xlog_recover_item	*item,
+	xfs_lsn_t			lsn)
+{
+	struct xfs_rud_log_format	*rud_formatp;
+
+	rud_formatp = item->ri_buf[0].i_addr;
+	ASSERT(item->ri_buf[0].i_len == sizeof(struct xfs_rud_log_format));
+
+	xlog_recover_release_intent(log, XFS_LI_RUI, rud_formatp->rud_rui_id);
+	return 0;
+}
+
+const struct xlog_recover_item_ops xlog_rud_item_ops = {
+	.item_type		= XFS_LI_RUD,
+	.commit_pass2		= xlog_recover_rud_commit_pass2,
+};
diff --git a/fs/xfs/xfs_rmap_item.h b/fs/xfs/xfs_rmap_item.h
index 8708e4a5aa5c..31e6cdfff71f 100644
--- a/fs/xfs/xfs_rmap_item.h
+++ b/fs/xfs/xfs_rmap_item.h
@@ -36,11 +36,6 @@ struct kmem_zone;
 #define	XFS_RUI_MAX_FAST_EXTENTS	16
 
 /*
- * Define RUI flag bits. Manipulated by set/clear/test_bit operators.
- */
-#define	XFS_RUI_RECOVERED		1
-
-/*
  * This is the "rmap update intent" log item.  It is used to log the fact that
  * some reverse mappings need to change.  It is used in conjunction with the
  * "rmap update done" log item described below.
@@ -52,7 +47,6 @@ struct xfs_rui_log_item {
 	struct xfs_log_item		rui_item;
 	atomic_t			rui_refcount;
 	atomic_t			rui_next_extent;
-	unsigned long			rui_flags;	/* misc flags */
 	struct xfs_rui_log_format	rui_format;
 };
 
@@ -77,11 +71,4 @@ struct xfs_rud_log_item {
 extern struct kmem_zone	*xfs_rui_zone;
 extern struct kmem_zone	*xfs_rud_zone;
 
-struct xfs_rui_log_item *xfs_rui_init(struct xfs_mount *, uint);
-int xfs_rui_copy_format(struct xfs_log_iovec *buf,
-		struct xfs_rui_log_format *dst_rui_fmt);
-void xfs_rui_item_free(struct xfs_rui_log_item *);
-void xfs_rui_release(struct xfs_rui_log_item *);
-int xfs_rui_recover(struct xfs_mount *mp, struct xfs_rui_log_item *ruip);
-
 #endif	/* __XFS_RMAP_ITEM_H__ */
diff --git a/fs/xfs/xfs_stats.c b/fs/xfs/xfs_stats.c
index 113883c4f202..f70f1255220b 100644
--- a/fs/xfs/xfs_stats.c
+++ b/fs/xfs/xfs_stats.c
@@ -57,13 +57,13 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf)
 	/* Loop over all stats groups */
 
 	for (i = j = 0; i < ARRAY_SIZE(xstats); i++) {
-		len += snprintf(buf + len, PATH_MAX - len, "%s",
+		len += scnprintf(buf + len, PATH_MAX - len, "%s",
 				xstats[i].desc);
 		/* inner loop does each group */
 		for (; j < xstats[i].endpoint; j++)
-			len += snprintf(buf + len, PATH_MAX - len, " %u",
+			len += scnprintf(buf + len, PATH_MAX - len, " %u",
 					counter_val(stats, j));
-		len += snprintf(buf + len, PATH_MAX - len, "\n");
+		len += scnprintf(buf + len, PATH_MAX - len, "\n");
 	}
 	/* extra precision counters */
 	for_each_possible_cpu(i) {
@@ -72,9 +72,9 @@ int xfs_stats_format(struct xfsstats __percpu *stats, char *buf)
 		xs_read_bytes += per_cpu_ptr(stats, i)->s.xs_read_bytes;
 	}
 
-	len += snprintf(buf + len, PATH_MAX-len, "xpc %Lu %Lu %Lu\n",
+	len += scnprintf(buf + len, PATH_MAX-len, "xpc %Lu %Lu %Lu\n",
 			xs_xstrat_bytes, xs_write_bytes, xs_read_bytes);
-	len += snprintf(buf + len, PATH_MAX-len, "debug %u\n",
+	len += scnprintf(buf + len, PATH_MAX-len, "debug %u\n",
 #if defined(DEBUG)
 		1);
 #else
diff --git a/fs/xfs/xfs_super.c b/fs/xfs/xfs_super.c
index 2094386af8ac..71ac6c1cdc36 100644
--- a/fs/xfs/xfs_super.c
+++ b/fs/xfs/xfs_super.c
@@ -47,6 +47,39 @@ static struct kset *xfs_kset;		/* top-level xfs sysfs dir */
 static struct xfs_kobj xfs_dbg_kobj;	/* global debug sysfs attrs */
 #endif
 
+enum xfs_dax_mode {
+	XFS_DAX_INODE = 0,
+	XFS_DAX_ALWAYS = 1,
+	XFS_DAX_NEVER = 2,
+};
+
+static void
+xfs_mount_set_dax_mode(
+	struct xfs_mount	*mp,
+	enum xfs_dax_mode	mode)
+{
+	switch (mode) {
+	case XFS_DAX_INODE:
+		mp->m_flags &= ~(XFS_MOUNT_DAX_ALWAYS | XFS_MOUNT_DAX_NEVER);
+		break;
+	case XFS_DAX_ALWAYS:
+		mp->m_flags |= XFS_MOUNT_DAX_ALWAYS;
+		mp->m_flags &= ~XFS_MOUNT_DAX_NEVER;
+		break;
+	case XFS_DAX_NEVER:
+		mp->m_flags |= XFS_MOUNT_DAX_NEVER;
+		mp->m_flags &= ~XFS_MOUNT_DAX_ALWAYS;
+		break;
+	}
+}
+
+static const struct constant_table dax_param_enums[] = {
+	{"inode",	XFS_DAX_INODE },
+	{"always",	XFS_DAX_ALWAYS },
+	{"never",	XFS_DAX_NEVER },
+	{}
+};
+
 /*
  * Table driven mount option parser.
  */
@@ -59,7 +92,7 @@ enum {
 	Opt_filestreams, Opt_quota, Opt_noquota, Opt_usrquota, Opt_grpquota,
 	Opt_prjquota, Opt_uquota, Opt_gquota, Opt_pquota,
 	Opt_uqnoenforce, Opt_gqnoenforce, Opt_pqnoenforce, Opt_qnoenforce,
-	Opt_discard, Opt_nodiscard, Opt_dax,
+	Opt_discard, Opt_nodiscard, Opt_dax, Opt_dax_enum,
 };
 
 static const struct fs_parameter_spec xfs_fs_parameters[] = {
@@ -103,6 +136,7 @@ static const struct fs_parameter_spec xfs_fs_parameters[] = {
 	fsparam_flag("discard",		Opt_discard),
 	fsparam_flag("nodiscard",	Opt_nodiscard),
 	fsparam_flag("dax",		Opt_dax),
+	fsparam_enum("dax",		Opt_dax_enum, dax_param_enums),
 	{}
 };
 
@@ -129,7 +163,8 @@ xfs_fs_show_options(
 		{ XFS_MOUNT_GRPID,		",grpid" },
 		{ XFS_MOUNT_DISCARD,		",discard" },
 		{ XFS_MOUNT_LARGEIO,		",largeio" },
-		{ XFS_MOUNT_DAX,		",dax" },
+		{ XFS_MOUNT_DAX_ALWAYS,		",dax=always" },
+		{ XFS_MOUNT_DAX_NEVER,		",dax=never" },
 		{ 0, NULL }
 	};
 	struct xfs_mount	*mp = XFS_M(root->d_sb);
@@ -305,7 +340,7 @@ void
 xfs_blkdev_issue_flush(
 	xfs_buftarg_t		*buftarg)
 {
-	blkdev_issue_flush(buftarg->bt_bdev, GFP_NOFS, NULL);
+	blkdev_issue_flush(buftarg->bt_bdev, GFP_NOFS);
 }
 
 STATIC void
@@ -516,6 +551,20 @@ xfs_destroy_mount_workqueues(
 	destroy_workqueue(mp->m_buf_workqueue);
 }
 
+static void
+xfs_flush_inodes_worker(
+	struct work_struct	*work)
+{
+	struct xfs_mount	*mp = container_of(work, struct xfs_mount,
+						   m_flush_inodes_work);
+	struct super_block	*sb = mp->m_super;
+
+	if (down_read_trylock(&sb->s_umount)) {
+		sync_inodes_sb(sb);
+		up_read(&sb->s_umount);
+	}
+}
+
 /*
  * Flush all dirty data to disk. Must not be called while holding an XFS_ILOCK
  * or a page lock. We use sync_inodes_sb() here to ensure we block while waiting
@@ -526,12 +575,15 @@ void
 xfs_flush_inodes(
 	struct xfs_mount	*mp)
 {
-	struct super_block	*sb = mp->m_super;
+	/*
+	 * If flush_work() returns true then that means we waited for a flush
+	 * which was already in progress.  Don't bother running another scan.
+	 */
+	if (flush_work(&mp->m_flush_inodes_work))
+		return;
 
-	if (down_read_trylock(&sb->s_umount)) {
-		sync_inodes_sb(sb);
-		up_read(&sb->s_umount);
-	}
+	queue_work(mp->m_sync_workqueue, &mp->m_flush_inodes_work);
+	flush_work(&mp->m_flush_inodes_work);
 }
 
 /* Catch misguided souls that try to use this interface on XFS */
@@ -685,7 +737,7 @@ xfs_fs_drop_inode(
 		return 0;
 	}
 
-	return generic_drop_inode(inode) || (ip->i_flags & XFS_IDONTCACHE);
+	return generic_drop_inode(inode);
 }
 
 static void
@@ -755,7 +807,8 @@ xfs_fs_statfs(
 	statp->f_blocks = sbp->sb_dblocks - lsize;
 	spin_unlock(&mp->m_sb_lock);
 
-	statp->f_bfree = fdblocks - mp->m_alloc_set_aside;
+	/* make sure statp->f_bfree does not underflow */
+	statp->f_bfree = max_t(int64_t, fdblocks - mp->m_alloc_set_aside, 0);
 	statp->f_bavail = statp->f_bfree;
 
 	fakeinos = XFS_FSB_TO_INO(mp, statp->f_bfree);
@@ -821,8 +874,10 @@ xfs_restore_resvblks(struct xfs_mount *mp)
  * there is no log replay required to write the inodes to disk - this is the
  * primary difference between a sync and a quiesce.
  *
- * Note: xfs_log_quiesce() stops background log work - the callers must ensure
- * it is started again when appropriate.
+ * We cancel log work early here to ensure all transactions the log worker may
+ * run have finished before we clean up and log the superblock and write an
+ * unmount record. The unfreeze process is responsible for restarting the log
+ * worker correctly.
  */
 void
 xfs_quiesce_attr(
@@ -830,28 +885,17 @@ xfs_quiesce_attr(
 {
 	int	error = 0;
 
-	/* wait for all modifications to complete */
-	while (atomic_read(&mp->m_active_trans) > 0)
-		delay(100);
+	cancel_delayed_work_sync(&mp->m_log->l_work);
 
 	/* force the log to unpin objects from the now complete transactions */
 	xfs_log_force(mp, XFS_LOG_SYNC);
 
-	/* reclaim inodes to do any IO before the freeze completes */
-	xfs_reclaim_inodes(mp, 0);
-	xfs_reclaim_inodes(mp, SYNC_WAIT);
 
 	/* Push the superblock and write an unmount record */
 	error = xfs_log_sbcount(mp);
 	if (error)
 		xfs_warn(mp, "xfs_attr_quiesce: failed to log sb changes. "
 				"Frozen image may not be consistent.");
-	/*
-	 * Just warn here till VFS can correctly support
-	 * read-only remount without racing.
-	 */
-	WARN_ON(atomic_read(&mp->m_active_trans) != 0);
-
 	xfs_log_quiesce(mp);
 }
 
@@ -866,11 +910,21 @@ xfs_fs_freeze(
 	struct super_block	*sb)
 {
 	struct xfs_mount	*mp = XFS_M(sb);
+	unsigned int		flags;
+	int			ret;
 
+	/*
+	 * The filesystem is now frozen far enough that memory reclaim
+	 * cannot safely operate on the filesystem. Hence we need to
+	 * set a GFP_NOFS context here to avoid recursion deadlocks.
+	 */
+	flags = memalloc_nofs_save();
 	xfs_stop_block_reaping(mp);
 	xfs_save_resvblks(mp);
 	xfs_quiesce_attr(mp);
-	return xfs_sync_sb(mp, true);
+	ret = xfs_sync_sb(mp, true);
+	memalloc_nofs_restore(flags);
+	return ret;
 }
 
 STATIC int
@@ -1244,7 +1298,10 @@ xfs_fc_parse_param(
 		return 0;
 #ifdef CONFIG_FS_DAX
 	case Opt_dax:
-		mp->m_flags |= XFS_MOUNT_DAX;
+		xfs_mount_set_dax_mode(mp, XFS_DAX_ALWAYS);
+		return 0;
+	case Opt_dax_enum:
+		xfs_mount_set_dax_mode(mp, result.uint_32);
 		return 0;
 #endif
 	default:
@@ -1437,7 +1494,7 @@ xfs_fc_fill_super(
 	if (XFS_SB_VERSION_NUM(&mp->m_sb) == XFS_SB_VERSION_5)
 		sb->s_flags |= SB_I_VERSION;
 
-	if (mp->m_flags & XFS_MOUNT_DAX) {
+	if (mp->m_flags & XFS_MOUNT_DAX_ALWAYS) {
 		bool rtdev_is_dax = false, datadev_is_dax;
 
 		xfs_warn(mp,
@@ -1451,7 +1508,7 @@ xfs_fc_fill_super(
 		if (!rtdev_is_dax && !datadev_is_dax) {
 			xfs_alert(mp,
 			"DAX unsupported by block device. Turning off DAX.");
-			mp->m_flags &= ~XFS_MOUNT_DAX;
+			xfs_mount_set_dax_mode(mp, XFS_DAX_NEVER);
 		}
 		if (xfs_sb_version_hasreflink(&mp->m_sb)) {
 			xfs_alert(mp,
@@ -1664,6 +1721,10 @@ xfs_fc_reconfigure(
 	int			flags = fc->sb_flags;
 	int			error;
 
+	/* version 5 superblocks always support version counters. */
+	if (XFS_SB_VERSION_NUM(&mp->m_sb) == XFS_SB_VERSION_5)
+		fc->sb_flags |= SB_I_VERSION;
+
 	error = xfs_fc_validate_params(new_mp);
 	if (error)
 		return error;
@@ -1737,7 +1798,7 @@ static int xfs_init_fs_context(
 	INIT_RADIX_TREE(&mp->m_perag_tree, GFP_ATOMIC);
 	spin_lock_init(&mp->m_perag_lock);
 	mutex_init(&mp->m_growlock);
-	atomic_set(&mp->m_active_trans, 0);
+	INIT_WORK(&mp->m_flush_inodes_work, xfs_flush_inodes_worker);
 	INIT_DELAYED_WORK(&mp->m_reclaim_work, xfs_reclaim_worker);
 	INIT_DELAYED_WORK(&mp->m_eofblocks_work, xfs_eofblocks_worker);
 	INIT_DELAYED_WORK(&mp->m_cowblocks_work, xfs_cowblocks_worker);
@@ -1861,7 +1922,8 @@ xfs_init_zones(void)
 
 	xfs_ili_zone = kmem_cache_create("xfs_ili",
 					 sizeof(struct xfs_inode_log_item), 0,
-					 SLAB_MEM_SPREAD, NULL);
+					 SLAB_RECLAIM_ACCOUNT | SLAB_MEM_SPREAD,
+					 NULL);
 	if (!xfs_ili_zone)
 		goto out_destroy_inode_zone;
 
diff --git a/fs/xfs/xfs_symlink.c b/fs/xfs/xfs_symlink.c
index d762d42ed0ff..8e88a7ca387e 100644
--- a/fs/xfs/xfs_symlink.c
+++ b/fs/xfs/xfs_symlink.c
@@ -176,15 +176,12 @@ xfs_symlink(
 		return -ENAMETOOLONG;
 	ASSERT(pathlen > 0);
 
-	udqp = gdqp = NULL;
 	prid = xfs_get_initial_prid(dp);
 
 	/*
 	 * Make sure that we have allocated dquot(s) on disk.
 	 */
-	error = xfs_qm_vop_dqalloc(dp,
-			xfs_kuid_to_uid(current_fsuid()),
-			xfs_kgid_to_gid(current_fsgid()), prid,
+	error = xfs_qm_vop_dqalloc(dp, current_fsuid(), current_fsgid(), prid,
 			XFS_QMOPT_QUOTALL | XFS_QMOPT_INHERIT,
 			&udqp, &gdqp, &pdqp);
 	if (error)
@@ -194,7 +191,7 @@ xfs_symlink(
 	 * The symlink will fit into the inode data fork?
 	 * There can't be any attributes so we get the whole variable part.
 	 */
-	if (pathlen <= XFS_LITINO(mp, dp->i_d.di_version))
+	if (pathlen <= XFS_LITINO(mp))
 		fs_blocks = 0;
 	else
 		fs_blocks = xfs_symlink_blocks(mp, pathlen);
@@ -246,8 +243,7 @@ xfs_symlink(
 	 */
 	xfs_qm_vop_create_dqattach(tp, ip, udqp, gdqp, pdqp);
 
-	if (resblks)
-		resblks -= XFS_IALLOC_SPACE_RES(mp);
+	resblks -= XFS_IALLOC_SPACE_RES(mp);
 	/*
 	 * If the symlink will fit into the inode, write it inline.
 	 */
@@ -255,7 +251,7 @@ xfs_symlink(
 		xfs_init_local_fork(ip, XFS_DATA_FORK, target_path, pathlen);
 
 		ip->i_d.di_size = pathlen;
-		ip->i_d.di_format = XFS_DINODE_FMT_LOCAL;
+		ip->i_df.if_format = XFS_DINODE_FMT_LOCAL;
 		xfs_trans_log_inode(tp, ip, XFS_ILOG_DDATA | XFS_ILOG_CORE);
 	} else {
 		int	offset;
@@ -268,8 +264,7 @@ xfs_symlink(
 		if (error)
 			goto out_trans_cancel;
 
-		if (resblks)
-			resblks -= fs_blocks;
+		resblks -= fs_blocks;
 		ip->i_d.di_size = pathlen;
 		xfs_trans_log_inode(tp, ip, XFS_ILOG_CORE);
 
@@ -389,7 +384,7 @@ xfs_inactive_symlink_rmt(
 	 * either 1 or 2 extents and that we can
 	 * free them all in one bunmapi call.
 	 */
-	ASSERT(ip->i_d.di_nextents > 0 && ip->i_d.di_nextents <= 2);
+	ASSERT(ip->i_df.if_nextents > 0 && ip->i_df.if_nextents <= 2);
 
 	error = xfs_trans_alloc(mp, &M_RES(mp)->tr_itruncate, 0, 0, 0, &tp);
 	if (error)
diff --git a/fs/xfs/xfs_sysctl.c b/fs/xfs/xfs_sysctl.c
index 31b3bdbd2eba..021ef96d0542 100644
--- a/fs/xfs/xfs_sysctl.c
+++ b/fs/xfs/xfs_sysctl.c
@@ -13,7 +13,7 @@ STATIC int
 xfs_stats_clear_proc_handler(
 	struct ctl_table	*ctl,
 	int			write,
-	void			__user *buffer,
+	void			*buffer,
 	size_t			*lenp,
 	loff_t			*ppos)
 {
@@ -33,7 +33,7 @@ STATIC int
 xfs_panic_mask_proc_handler(
 	struct ctl_table	*ctl,
 	int			write,
-	void			__user *buffer,
+	void			*buffer,
 	size_t			*lenp,
 	loff_t			*ppos)
 {
diff --git a/fs/xfs/xfs_sysfs.h b/fs/xfs/xfs_sysfs.h
index e9f810fc6731..43585850f154 100644
--- a/fs/xfs/xfs_sysfs.h
+++ b/fs/xfs/xfs_sysfs.h
@@ -32,9 +32,11 @@ xfs_sysfs_init(
 	struct xfs_kobj		*parent_kobj,
 	const char		*name)
 {
+	struct kobject		*parent;
+
+	parent = parent_kobj ? &parent_kobj->kobject : NULL;
 	init_completion(&kobj->complete);
-	return kobject_init_and_add(&kobj->kobject, ktype,
-				    &parent_kobj->kobject, "%s", name);
+	return kobject_init_and_add(&kobj->kobject, ktype, parent, "%s", name);
 }
 
 static inline void
diff --git a/fs/xfs/xfs_trace.c b/fs/xfs/xfs_trace.c
index bc85b89f88ca..120398a37c2a 100644
--- a/fs/xfs/xfs_trace.c
+++ b/fs/xfs/xfs_trace.c
@@ -6,6 +6,7 @@
 #include "xfs.h"
 #include "xfs_fs.h"
 #include "xfs_shared.h"
+#include "xfs_bit.h"
 #include "xfs_format.h"
 #include "xfs_log_format.h"
 #include "xfs_trans_resv.h"
@@ -27,6 +28,7 @@
 #include "xfs_log_recover.h"
 #include "xfs_filestream.h"
 #include "xfs_fsmap.h"
+#include "xfs_btree_staging.h"
 
 /*
  * We include this last to have the helpers above available for the trace
diff --git a/fs/xfs/xfs_trace.h b/fs/xfs/xfs_trace.h
index e242988f57fb..abb1d859f226 100644
--- a/fs/xfs/xfs_trace.h
+++ b/fs/xfs/xfs_trace.h
@@ -35,6 +35,13 @@ struct xfs_icreate_log;
 struct xfs_owner_info;
 struct xfs_trans_res;
 struct xfs_inobt_rec_incore;
+union xfs_btree_ptr;
+struct xfs_dqtrx;
+
+#define XFS_ATTR_FILTER_FLAGS \
+	{ XFS_ATTR_ROOT,	"ROOT" }, \
+	{ XFS_ATTR_SECURE,	"SECURE" }, \
+	{ XFS_ATTR_INCOMPLETE,	"INCOMPLETE" }
 
 DECLARE_EVENT_CLASS(xfs_attr_list_class,
 	TP_PROTO(struct xfs_attr_list_context *ctx),
@@ -45,39 +52,39 @@ DECLARE_EVENT_CLASS(xfs_attr_list_class,
 		__field(u32, hashval)
 		__field(u32, blkno)
 		__field(u32, offset)
-		__field(void *, alist)
+		__field(void *, buffer)
 		__field(int, bufsize)
 		__field(int, count)
 		__field(int, firstu)
 		__field(int, dupcnt)
-		__field(int, flags)
+		__field(unsigned int, attr_filter)
 	),
 	TP_fast_assign(
 		__entry->dev = VFS_I(ctx->dp)->i_sb->s_dev;
 		__entry->ino = ctx->dp->i_ino;
-		__entry->hashval = ctx->cursor->hashval;
-		__entry->blkno = ctx->cursor->blkno;
-		__entry->offset = ctx->cursor->offset;
-		__entry->alist = ctx->alist;
+		__entry->hashval = ctx->cursor.hashval;
+		__entry->blkno = ctx->cursor.blkno;
+		__entry->offset = ctx->cursor.offset;
+		__entry->buffer = ctx->buffer;
 		__entry->bufsize = ctx->bufsize;
 		__entry->count = ctx->count;
 		__entry->firstu = ctx->firstu;
-		__entry->flags = ctx->flags;
+		__entry->attr_filter = ctx->attr_filter;
 	),
 	TP_printk("dev %d:%d ino 0x%llx cursor h/b/o 0x%x/0x%x/%u dupcnt %u "
-		  "alist %p size %u count %u firstu %u flags %d %s",
+		  "buffer %p size %u count %u firstu %u filter %s",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		   __entry->ino,
 		   __entry->hashval,
 		   __entry->blkno,
 		   __entry->offset,
 		   __entry->dupcnt,
-		   __entry->alist,
+		   __entry->buffer,
 		   __entry->bufsize,
 		   __entry->count,
 		   __entry->firstu,
-		   __entry->flags,
-		   __print_flags(__entry->flags, "|", XFS_ATTR_FLAGS)
+		   __print_flags(__entry->attr_filter, "|",
+				 XFS_ATTR_FILTER_FLAGS)
 	)
 )
 
@@ -169,31 +176,31 @@ TRACE_EVENT(xfs_attr_list_node_descend,
 		__field(u32, hashval)
 		__field(u32, blkno)
 		__field(u32, offset)
-		__field(void *, alist)
+		__field(void *, buffer)
 		__field(int, bufsize)
 		__field(int, count)
 		__field(int, firstu)
 		__field(int, dupcnt)
-		__field(int, flags)
+		__field(unsigned int, attr_filter)
 		__field(u32, bt_hashval)
 		__field(u32, bt_before)
 	),
 	TP_fast_assign(
 		__entry->dev = VFS_I(ctx->dp)->i_sb->s_dev;
 		__entry->ino = ctx->dp->i_ino;
-		__entry->hashval = ctx->cursor->hashval;
-		__entry->blkno = ctx->cursor->blkno;
-		__entry->offset = ctx->cursor->offset;
-		__entry->alist = ctx->alist;
+		__entry->hashval = ctx->cursor.hashval;
+		__entry->blkno = ctx->cursor.blkno;
+		__entry->offset = ctx->cursor.offset;
+		__entry->buffer = ctx->buffer;
 		__entry->bufsize = ctx->bufsize;
 		__entry->count = ctx->count;
 		__entry->firstu = ctx->firstu;
-		__entry->flags = ctx->flags;
+		__entry->attr_filter = ctx->attr_filter;
 		__entry->bt_hashval = be32_to_cpu(btree->hashval);
 		__entry->bt_before = be32_to_cpu(btree->before);
 	),
 	TP_printk("dev %d:%d ino 0x%llx cursor h/b/o 0x%x/0x%x/%u dupcnt %u "
-		  "alist %p size %u count %u firstu %u flags %d %s "
+		  "buffer %p size %u count %u firstu %u filter %s "
 		  "node hashval %u, node before %u",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		   __entry->ino,
@@ -201,12 +208,12 @@ TRACE_EVENT(xfs_attr_list_node_descend,
 		   __entry->blkno,
 		   __entry->offset,
 		   __entry->dupcnt,
-		   __entry->alist,
+		   __entry->buffer,
 		   __entry->bufsize,
 		   __entry->count,
 		   __entry->firstu,
-		   __entry->flags,
-		   __print_flags(__entry->flags, "|", XFS_ATTR_FLAGS),
+		   __print_flags(__entry->attr_filter, "|",
+				 XFS_ATTR_FILTER_FLAGS),
 		   __entry->bt_hashval,
 		   __entry->bt_before)
 );
@@ -858,44 +865,65 @@ DECLARE_EVENT_CLASS(xfs_dquot_class,
 	TP_STRUCT__entry(
 		__field(dev_t, dev)
 		__field(u32, id)
+		__field(xfs_dqtype_t, type)
 		__field(unsigned, flags)
 		__field(unsigned, nrefs)
 		__field(unsigned long long, res_bcount)
+		__field(unsigned long long, res_rtbcount)
+		__field(unsigned long long, res_icount)
+
 		__field(unsigned long long, bcount)
+		__field(unsigned long long, rtbcount)
 		__field(unsigned long long, icount)
+
 		__field(unsigned long long, blk_hardlimit)
 		__field(unsigned long long, blk_softlimit)
+		__field(unsigned long long, rtb_hardlimit)
+		__field(unsigned long long, rtb_softlimit)
 		__field(unsigned long long, ino_hardlimit)
 		__field(unsigned long long, ino_softlimit)
-	), \
+	),
 	TP_fast_assign(
 		__entry->dev = dqp->q_mount->m_super->s_dev;
-		__entry->id = be32_to_cpu(dqp->q_core.d_id);
-		__entry->flags = dqp->dq_flags;
+		__entry->id = dqp->q_id;
+		__entry->type = dqp->q_type;
+		__entry->flags = dqp->q_flags;
 		__entry->nrefs = dqp->q_nrefs;
-		__entry->res_bcount = dqp->q_res_bcount;
-		__entry->bcount = be64_to_cpu(dqp->q_core.d_bcount);
-		__entry->icount = be64_to_cpu(dqp->q_core.d_icount);
-		__entry->blk_hardlimit =
-			be64_to_cpu(dqp->q_core.d_blk_hardlimit);
-		__entry->blk_softlimit =
-			be64_to_cpu(dqp->q_core.d_blk_softlimit);
-		__entry->ino_hardlimit =
-			be64_to_cpu(dqp->q_core.d_ino_hardlimit);
-		__entry->ino_softlimit =
-			be64_to_cpu(dqp->q_core.d_ino_softlimit);
-	),
-	TP_printk("dev %d:%d id 0x%x flags %s nrefs %u res_bc 0x%llx "
+
+		__entry->res_bcount = dqp->q_blk.reserved;
+		__entry->res_rtbcount = dqp->q_rtb.reserved;
+		__entry->res_icount = dqp->q_ino.reserved;
+
+		__entry->bcount = dqp->q_blk.count;
+		__entry->rtbcount = dqp->q_rtb.count;
+		__entry->icount = dqp->q_ino.count;
+
+		__entry->blk_hardlimit = dqp->q_blk.hardlimit;
+		__entry->blk_softlimit = dqp->q_blk.softlimit;
+		__entry->rtb_hardlimit = dqp->q_rtb.hardlimit;
+		__entry->rtb_softlimit = dqp->q_rtb.softlimit;
+		__entry->ino_hardlimit = dqp->q_ino.hardlimit;
+		__entry->ino_softlimit = dqp->q_ino.softlimit;
+	),
+	TP_printk("dev %d:%d id 0x%x type %s flags %s nrefs %u "
+		  "res_bc 0x%llx res_rtbc 0x%llx res_ic 0x%llx "
 		  "bcnt 0x%llx bhardlimit 0x%llx bsoftlimit 0x%llx "
+		  "rtbcnt 0x%llx rtbhardlimit 0x%llx rtbsoftlimit 0x%llx "
 		  "icnt 0x%llx ihardlimit 0x%llx isoftlimit 0x%llx]",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->id,
-		  __print_flags(__entry->flags, "|", XFS_DQ_FLAGS),
+		  __print_flags(__entry->type, "|", XFS_DQTYPE_STRINGS),
+		  __print_flags(__entry->flags, "|", XFS_DQFLAG_STRINGS),
 		  __entry->nrefs,
 		  __entry->res_bcount,
+		  __entry->res_rtbcount,
+		  __entry->res_icount,
 		  __entry->bcount,
 		  __entry->blk_hardlimit,
 		  __entry->blk_softlimit,
+		  __entry->rtbcount,
+		  __entry->rtb_hardlimit,
+		  __entry->rtb_softlimit,
 		  __entry->icount,
 		  __entry->ino_hardlimit,
 		  __entry->ino_softlimit)
@@ -926,6 +954,125 @@ DEFINE_DQUOT_EVENT(xfs_dqrele);
 DEFINE_DQUOT_EVENT(xfs_dqflush);
 DEFINE_DQUOT_EVENT(xfs_dqflush_force);
 DEFINE_DQUOT_EVENT(xfs_dqflush_done);
+DEFINE_DQUOT_EVENT(xfs_trans_apply_dquot_deltas_before);
+DEFINE_DQUOT_EVENT(xfs_trans_apply_dquot_deltas_after);
+
+#define XFS_QMOPT_FLAGS \
+	{ XFS_QMOPT_UQUOTA,		"UQUOTA" }, \
+	{ XFS_QMOPT_PQUOTA,		"PQUOTA" }, \
+	{ XFS_QMOPT_FORCE_RES,		"FORCE_RES" }, \
+	{ XFS_QMOPT_SBVERSION,		"SBVERSION" }, \
+	{ XFS_QMOPT_GQUOTA,		"GQUOTA" }, \
+	{ XFS_QMOPT_INHERIT,		"INHERIT" }, \
+	{ XFS_QMOPT_RES_REGBLKS,	"RES_REGBLKS" }, \
+	{ XFS_QMOPT_RES_RTBLKS,		"RES_RTBLKS" }, \
+	{ XFS_QMOPT_BCOUNT,		"BCOUNT" }, \
+	{ XFS_QMOPT_ICOUNT,		"ICOUNT" }, \
+	{ XFS_QMOPT_RTBCOUNT,		"RTBCOUNT" }, \
+	{ XFS_QMOPT_DELBCOUNT,		"DELBCOUNT" }, \
+	{ XFS_QMOPT_DELRTBCOUNT,	"DELRTBCOUNT" }, \
+	{ XFS_QMOPT_RES_INOS,		"RES_INOS" }
+
+TRACE_EVENT(xfs_trans_mod_dquot,
+	TP_PROTO(struct xfs_trans *tp, struct xfs_dquot *dqp,
+		 unsigned int field, int64_t delta),
+	TP_ARGS(tp, dqp, field, delta),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_dqtype_t, type)
+		__field(unsigned int, flags)
+		__field(unsigned int, dqid)
+		__field(unsigned int, field)
+		__field(int64_t, delta)
+	),
+	TP_fast_assign(
+		__entry->dev = tp->t_mountp->m_super->s_dev;
+		__entry->type = dqp->q_type;
+		__entry->flags = dqp->q_flags;
+		__entry->dqid = dqp->q_id;
+		__entry->field = field;
+		__entry->delta = delta;
+	),
+	TP_printk("dev %d:%d dquot id 0x%x type %s flags %s field %s delta %lld",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->dqid,
+		  __print_flags(__entry->type, "|", XFS_DQTYPE_STRINGS),
+		  __print_flags(__entry->flags, "|", XFS_DQFLAG_STRINGS),
+		  __print_flags(__entry->field, "|", XFS_QMOPT_FLAGS),
+		  __entry->delta)
+);
+
+DECLARE_EVENT_CLASS(xfs_dqtrx_class,
+	TP_PROTO(struct xfs_dqtrx *qtrx),
+	TP_ARGS(qtrx),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_dqtype_t, type)
+		__field(unsigned int, flags)
+		__field(u32, dqid)
+
+		__field(uint64_t, blk_res)
+		__field(int64_t,  bcount_delta)
+		__field(int64_t,  delbcnt_delta)
+
+		__field(uint64_t, rtblk_res)
+		__field(uint64_t, rtblk_res_used)
+		__field(int64_t,  rtbcount_delta)
+		__field(int64_t,  delrtb_delta)
+
+		__field(uint64_t, ino_res)
+		__field(uint64_t, ino_res_used)
+		__field(int64_t,  icount_delta)
+	),
+	TP_fast_assign(
+		__entry->dev = qtrx->qt_dquot->q_mount->m_super->s_dev;
+		__entry->type = qtrx->qt_dquot->q_type;
+		__entry->flags = qtrx->qt_dquot->q_flags;
+		__entry->dqid = qtrx->qt_dquot->q_id;
+
+		__entry->blk_res = qtrx->qt_blk_res;
+		__entry->bcount_delta = qtrx->qt_bcount_delta;
+		__entry->delbcnt_delta = qtrx->qt_delbcnt_delta;
+
+		__entry->rtblk_res = qtrx->qt_rtblk_res;
+		__entry->rtblk_res_used = qtrx->qt_rtblk_res_used;
+		__entry->rtbcount_delta = qtrx->qt_rtbcount_delta;
+		__entry->delrtb_delta = qtrx->qt_delrtb_delta;
+
+		__entry->ino_res = qtrx->qt_ino_res;
+		__entry->ino_res_used = qtrx->qt_ino_res_used;
+		__entry->icount_delta = qtrx->qt_icount_delta;
+	),
+	TP_printk("dev %d:%d dquot id 0x%x type %s flags %s"
+		  "blk_res %llu bcount_delta %lld delbcnt_delta %lld "
+		  "rtblk_res %llu rtblk_res_used %llu rtbcount_delta %lld delrtb_delta %lld "
+		  "ino_res %llu ino_res_used %llu icount_delta %lld",
+		MAJOR(__entry->dev), MINOR(__entry->dev),
+		__entry->dqid,
+		  __print_flags(__entry->type, "|", XFS_DQTYPE_STRINGS),
+		  __print_flags(__entry->flags, "|", XFS_DQFLAG_STRINGS),
+
+		__entry->blk_res,
+		__entry->bcount_delta,
+		__entry->delbcnt_delta,
+
+		__entry->rtblk_res,
+		__entry->rtblk_res_used,
+		__entry->rtbcount_delta,
+		__entry->delrtb_delta,
+
+		__entry->ino_res,
+		__entry->ino_res_used,
+		__entry->icount_delta)
+)
+
+#define DEFINE_DQTRX_EVENT(name) \
+DEFINE_EVENT(xfs_dqtrx_class, name, \
+	TP_PROTO(struct xfs_dqtrx *qtrx), \
+	TP_ARGS(qtrx))
+DEFINE_DQTRX_EVENT(xfs_trans_apply_dquot_deltas);
+DEFINE_DQTRX_EVENT(xfs_trans_mod_dquot_before);
+DEFINE_DQTRX_EVENT(xfs_trans_mod_dquot_after);
 
 DECLARE_EVENT_CLASS(xfs_loggrant_class,
 	TP_PROTO(struct xlog *log, struct xlog_ticket *tic),
@@ -995,8 +1142,6 @@ DECLARE_EVENT_CLASS(xfs_loggrant_class,
 DEFINE_EVENT(xfs_loggrant_class, name, \
 	TP_PROTO(struct xlog *log, struct xlog_ticket *tic), \
 	TP_ARGS(log, tic))
-DEFINE_LOGGRANT_EVENT(xfs_log_done_nonperm);
-DEFINE_LOGGRANT_EVENT(xfs_log_done_perm);
 DEFINE_LOGGRANT_EVENT(xfs_log_umount_write);
 DEFINE_LOGGRANT_EVENT(xfs_log_grant_sleep);
 DEFINE_LOGGRANT_EVENT(xfs_log_grant_wake);
@@ -1005,12 +1150,13 @@ DEFINE_LOGGRANT_EVENT(xfs_log_reserve);
 DEFINE_LOGGRANT_EVENT(xfs_log_reserve_exit);
 DEFINE_LOGGRANT_EVENT(xfs_log_regrant);
 DEFINE_LOGGRANT_EVENT(xfs_log_regrant_exit);
-DEFINE_LOGGRANT_EVENT(xfs_log_regrant_reserve_enter);
-DEFINE_LOGGRANT_EVENT(xfs_log_regrant_reserve_exit);
-DEFINE_LOGGRANT_EVENT(xfs_log_regrant_reserve_sub);
-DEFINE_LOGGRANT_EVENT(xfs_log_ungrant_enter);
-DEFINE_LOGGRANT_EVENT(xfs_log_ungrant_exit);
-DEFINE_LOGGRANT_EVENT(xfs_log_ungrant_sub);
+DEFINE_LOGGRANT_EVENT(xfs_log_ticket_regrant);
+DEFINE_LOGGRANT_EVENT(xfs_log_ticket_regrant_exit);
+DEFINE_LOGGRANT_EVENT(xfs_log_ticket_regrant_sub);
+DEFINE_LOGGRANT_EVENT(xfs_log_ticket_ungrant);
+DEFINE_LOGGRANT_EVENT(xfs_log_ticket_ungrant_sub);
+DEFINE_LOGGRANT_EVENT(xfs_log_ticket_ungrant_exit);
+DEFINE_LOGGRANT_EVENT(xfs_log_cil_wait);
 
 DECLARE_EVENT_CLASS(xfs_log_item_class,
 	TP_PROTO(struct xfs_log_item *lip),
@@ -1701,7 +1847,8 @@ DECLARE_EVENT_CLASS(xfs_attr_class,
 		__field(int, namelen)
 		__field(int, valuelen)
 		__field(xfs_dahash_t, hashval)
-		__field(int, flags)
+		__field(unsigned int, attr_filter)
+		__field(unsigned int, attr_flags)
 		__field(int, op_flags)
 	),
 	TP_fast_assign(
@@ -1712,11 +1859,12 @@ DECLARE_EVENT_CLASS(xfs_attr_class,
 		__entry->namelen = args->namelen;
 		__entry->valuelen = args->valuelen;
 		__entry->hashval = args->hashval;
-		__entry->flags = args->flags;
+		__entry->attr_filter = args->attr_filter;
+		__entry->attr_flags = args->attr_flags;
 		__entry->op_flags = args->op_flags;
 	),
 	TP_printk("dev %d:%d ino 0x%llx name %.*s namelen %d valuelen %d "
-		  "hashval 0x%x flags %s op_flags %s",
+		  "hashval 0x%x filter %s flags %s op_flags %s",
 		  MAJOR(__entry->dev), MINOR(__entry->dev),
 		  __entry->ino,
 		  __entry->namelen,
@@ -1724,7 +1872,11 @@ DECLARE_EVENT_CLASS(xfs_attr_class,
 		  __entry->namelen,
 		  __entry->valuelen,
 		  __entry->hashval,
-		  __print_flags(__entry->flags, "|", XFS_ATTR_FLAGS),
+		  __print_flags(__entry->attr_filter, "|",
+				XFS_ATTR_FILTER_FLAGS),
+		   __print_flags(__entry->attr_flags, "|",
+				{ XATTR_CREATE,		"CREATE" },
+				{ XATTR_REPLACE,	"REPLACE" }),
 		  __print_flags(__entry->op_flags, "|", XFS_DA_OP_FLAGS))
 )
 
@@ -1886,8 +2038,8 @@ DECLARE_EVENT_CLASS(xfs_swap_extent_class,
 		__entry->dev = VFS_I(ip)->i_sb->s_dev;
 		__entry->which = which;
 		__entry->ino = ip->i_ino;
-		__entry->format = ip->i_d.di_format;
-		__entry->nex = ip->i_d.di_nextents;
+		__entry->format = ip->i_df.if_format;
+		__entry->nex = ip->i_df.if_nextents;
 		__entry->broot_size = ip->i_df.if_broot_bytes;
 		__entry->fork_off = XFS_IFORK_BOFF(ip);
 	),
@@ -3041,8 +3193,7 @@ DEFINE_EVENT(xfs_inode_irec_class, name, \
 DEFINE_INODE_EVENT(xfs_reflink_set_inode_flag);
 DEFINE_INODE_EVENT(xfs_reflink_unset_inode_flag);
 DEFINE_ITRUNC_EVENT(xfs_reflink_update_inode_size);
-DEFINE_IMAP_EVENT(xfs_reflink_remap_imap);
-TRACE_EVENT(xfs_reflink_remap_blocks_loop,
+TRACE_EVENT(xfs_reflink_remap_blocks,
 	TP_PROTO(struct xfs_inode *src, xfs_fileoff_t soffset,
 		 xfs_filblks_t len, struct xfs_inode *dest,
 		 xfs_fileoff_t doffset),
@@ -3073,59 +3224,14 @@ TRACE_EVENT(xfs_reflink_remap_blocks_loop,
 		  __entry->dest_ino,
 		  __entry->dest_lblk)
 );
-TRACE_EVENT(xfs_reflink_punch_range,
-	TP_PROTO(struct xfs_inode *ip, xfs_fileoff_t lblk,
-		 xfs_extlen_t len),
-	TP_ARGS(ip, lblk, len),
-	TP_STRUCT__entry(
-		__field(dev_t, dev)
-		__field(xfs_ino_t, ino)
-		__field(xfs_fileoff_t, lblk)
-		__field(xfs_extlen_t, len)
-	),
-	TP_fast_assign(
-		__entry->dev = VFS_I(ip)->i_sb->s_dev;
-		__entry->ino = ip->i_ino;
-		__entry->lblk = lblk;
-		__entry->len = len;
-	),
-	TP_printk("dev %d:%d ino 0x%llx lblk 0x%llx len 0x%x",
-		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __entry->ino,
-		  __entry->lblk,
-		  __entry->len)
-);
-TRACE_EVENT(xfs_reflink_remap,
-	TP_PROTO(struct xfs_inode *ip, xfs_fileoff_t lblk,
-		 xfs_extlen_t len, xfs_fsblock_t new_pblk),
-	TP_ARGS(ip, lblk, len, new_pblk),
-	TP_STRUCT__entry(
-		__field(dev_t, dev)
-		__field(xfs_ino_t, ino)
-		__field(xfs_fileoff_t, lblk)
-		__field(xfs_extlen_t, len)
-		__field(xfs_fsblock_t, new_pblk)
-	),
-	TP_fast_assign(
-		__entry->dev = VFS_I(ip)->i_sb->s_dev;
-		__entry->ino = ip->i_ino;
-		__entry->lblk = lblk;
-		__entry->len = len;
-		__entry->new_pblk = new_pblk;
-	),
-	TP_printk("dev %d:%d ino 0x%llx lblk 0x%llx len 0x%x new_pblk %llu",
-		  MAJOR(__entry->dev), MINOR(__entry->dev),
-		  __entry->ino,
-		  __entry->lblk,
-		  __entry->len,
-		  __entry->new_pblk)
-);
 DEFINE_DOUBLE_IO_EVENT(xfs_reflink_remap_range);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_remap_range_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_set_inode_flag_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_update_inode_size_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_remap_blocks_error);
 DEFINE_INODE_ERROR_EVENT(xfs_reflink_remap_extent_error);
+DEFINE_INODE_IREC_EVENT(xfs_reflink_remap_extent_src);
+DEFINE_INODE_IREC_EVENT(xfs_reflink_remap_extent_dest);
 
 /* dedupe tracepoints */
 DEFINE_DOUBLE_IO_EVENT(xfs_reflink_compare_extents);
@@ -3571,7 +3677,6 @@ DEFINE_KMEM_EVENT(kmem_alloc);
 DEFINE_KMEM_EVENT(kmem_alloc_io);
 DEFINE_KMEM_EVENT(kmem_alloc_large);
 DEFINE_KMEM_EVENT(kmem_realloc);
-DEFINE_KMEM_EVENT(kmem_zone_alloc);
 
 TRACE_EVENT(xfs_check_new_dalign,
 	TP_PROTO(struct xfs_mount *mp, int new_dalign, xfs_ino_t calc_rootino),
@@ -3594,6 +3699,151 @@ TRACE_EVENT(xfs_check_new_dalign,
 		  __entry->calc_rootino)
 )
 
+TRACE_EVENT(xfs_btree_commit_afakeroot,
+	TP_PROTO(struct xfs_btree_cur *cur),
+	TP_ARGS(cur),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_btnum_t, btnum)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(unsigned int, levels)
+		__field(unsigned int, blocks)
+	),
+	TP_fast_assign(
+		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->btnum = cur->bc_btnum;
+		__entry->agno = cur->bc_ag.agno;
+		__entry->agbno = cur->bc_ag.afake->af_root;
+		__entry->levels = cur->bc_ag.afake->af_levels;
+		__entry->blocks = cur->bc_ag.afake->af_blocks;
+	),
+	TP_printk("dev %d:%d btree %s ag %u levels %u blocks %u root %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __entry->agno,
+		  __entry->levels,
+		  __entry->blocks,
+		  __entry->agbno)
+)
+
+TRACE_EVENT(xfs_btree_commit_ifakeroot,
+	TP_PROTO(struct xfs_btree_cur *cur),
+	TP_ARGS(cur),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_btnum_t, btnum)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agino_t, agino)
+		__field(unsigned int, levels)
+		__field(unsigned int, blocks)
+		__field(int, whichfork)
+	),
+	TP_fast_assign(
+		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->btnum = cur->bc_btnum;
+		__entry->agno = XFS_INO_TO_AGNO(cur->bc_mp,
+					cur->bc_ino.ip->i_ino);
+		__entry->agino = XFS_INO_TO_AGINO(cur->bc_mp,
+					cur->bc_ino.ip->i_ino);
+		__entry->levels = cur->bc_ino.ifake->if_levels;
+		__entry->blocks = cur->bc_ino.ifake->if_blocks;
+		__entry->whichfork = cur->bc_ino.whichfork;
+	),
+	TP_printk("dev %d:%d btree %s ag %u agino %u whichfork %s levels %u blocks %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __entry->agno,
+		  __entry->agino,
+		  __entry->whichfork == XFS_ATTR_FORK ? "attr" : "data",
+		  __entry->levels,
+		  __entry->blocks)
+)
+
+TRACE_EVENT(xfs_btree_bload_level_geometry,
+	TP_PROTO(struct xfs_btree_cur *cur, unsigned int level,
+		 uint64_t nr_this_level, unsigned int nr_per_block,
+		 unsigned int desired_npb, uint64_t blocks,
+		 uint64_t blocks_with_extra),
+	TP_ARGS(cur, level, nr_this_level, nr_per_block, desired_npb, blocks,
+		blocks_with_extra),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_btnum_t, btnum)
+		__field(unsigned int, level)
+		__field(unsigned int, nlevels)
+		__field(uint64_t, nr_this_level)
+		__field(unsigned int, nr_per_block)
+		__field(unsigned int, desired_npb)
+		__field(unsigned long long, blocks)
+		__field(unsigned long long, blocks_with_extra)
+	),
+	TP_fast_assign(
+		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->btnum = cur->bc_btnum;
+		__entry->level = level;
+		__entry->nlevels = cur->bc_nlevels;
+		__entry->nr_this_level = nr_this_level;
+		__entry->nr_per_block = nr_per_block;
+		__entry->desired_npb = desired_npb;
+		__entry->blocks = blocks;
+		__entry->blocks_with_extra = blocks_with_extra;
+	),
+	TP_printk("dev %d:%d btree %s level %u/%u nr_this_level %llu nr_per_block %u desired_npb %u blocks %llu blocks_with_extra %llu",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __entry->level,
+		  __entry->nlevels,
+		  __entry->nr_this_level,
+		  __entry->nr_per_block,
+		  __entry->desired_npb,
+		  __entry->blocks,
+		  __entry->blocks_with_extra)
+)
+
+TRACE_EVENT(xfs_btree_bload_block,
+	TP_PROTO(struct xfs_btree_cur *cur, unsigned int level,
+		 uint64_t block_idx, uint64_t nr_blocks,
+		 union xfs_btree_ptr *ptr, unsigned int nr_records),
+	TP_ARGS(cur, level, block_idx, nr_blocks, ptr, nr_records),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_btnum_t, btnum)
+		__field(unsigned int, level)
+		__field(unsigned long long, block_idx)
+		__field(unsigned long long, nr_blocks)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, agbno)
+		__field(unsigned int, nr_records)
+	),
+	TP_fast_assign(
+		__entry->dev = cur->bc_mp->m_super->s_dev;
+		__entry->btnum = cur->bc_btnum;
+		__entry->level = level;
+		__entry->block_idx = block_idx;
+		__entry->nr_blocks = nr_blocks;
+		if (cur->bc_flags & XFS_BTREE_LONG_PTRS) {
+			xfs_fsblock_t	fsb = be64_to_cpu(ptr->l);
+
+			__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsb);
+			__entry->agbno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsb);
+		} else {
+			__entry->agno = cur->bc_ag.agno;
+			__entry->agbno = be32_to_cpu(ptr->s);
+		}
+		__entry->nr_records = nr_records;
+	),
+	TP_printk("dev %d:%d btree %s level %u block %llu/%llu fsb (%u/%u) recs %u",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __print_symbolic(__entry->btnum, XFS_BTNUM_STRINGS),
+		  __entry->level,
+		  __entry->block_idx,
+		  __entry->nr_blocks,
+		  __entry->agno,
+		  __entry->agbno,
+		  __entry->nr_records)
+)
+
 #endif /* _TRACE_XFS_H */
 
 #undef TRACE_INCLUDE_PATH
diff --git a/fs/xfs/xfs_trans.c b/fs/xfs/xfs_trans.c
index 3b208f9a865c..ed72867b1a19 100644
--- a/fs/xfs/xfs_trans.c
+++ b/fs/xfs/xfs_trans.c
@@ -9,6 +9,7 @@
 #include "xfs_shared.h"
 #include "xfs_format.h"
 #include "xfs_log_format.h"
+#include "xfs_log_priv.h"
 #include "xfs_trans_resv.h"
 #include "xfs_mount.h"
 #include "xfs_extent_busy.h"
@@ -67,7 +68,6 @@ xfs_trans_free(
 	xfs_extent_busy_clear(tp->t_mountp, &tp->t_busy, false);
 
 	trace_xfs_trans_free(tp, _RET_IP_);
-	atomic_dec(&tp->t_mountp->m_active_trans);
 	if (!(tp->t_flags & XFS_TRANS_NO_WRITECOUNT))
 		sb_end_intwrite(tp->t_mountp->m_super);
 	xfs_trans_free_dqinfo(tp);
@@ -90,7 +90,7 @@ xfs_trans_dup(
 
 	trace_xfs_trans_dup(tp, _RET_IP_);
 
-	ntp = kmem_zone_zalloc(xfs_trans_zone, 0);
+	ntp = kmem_cache_zalloc(xfs_trans_zone, GFP_KERNEL | __GFP_NOFAIL);
 
 	/*
 	 * Initialize the new transaction structure.
@@ -107,7 +107,8 @@ xfs_trans_dup(
 
 	ntp->t_flags = XFS_TRANS_PERM_LOG_RES |
 		       (tp->t_flags & XFS_TRANS_RESERVE) |
-		       (tp->t_flags & XFS_TRANS_NO_WRITECOUNT);
+		       (tp->t_flags & XFS_TRANS_NO_WRITECOUNT) |
+		       (tp->t_flags & XFS_TRANS_RES_FDBLKS);
 	/* We gave our writer reference to the new transaction */
 	tp->t_flags |= XFS_TRANS_NO_WRITECOUNT;
 	ntp->t_ticket = xfs_log_ticket_get(tp->t_ticket);
@@ -124,8 +125,6 @@ xfs_trans_dup(
 	xfs_defer_move(ntp, tp);
 
 	xfs_trans_dup_dqinfo(tp, ntp);
-
-	atomic_inc(&tp->t_mountp->m_active_trans);
 	return ntp;
 }
 
@@ -150,8 +149,9 @@ xfs_trans_reserve(
 	uint			blocks,
 	uint			rtextents)
 {
-	int		error = 0;
-	bool		rsvd = (tp->t_flags & XFS_TRANS_RESERVE) != 0;
+	struct xfs_mount	*mp = tp->t_mountp;
+	int			error = 0;
+	bool			rsvd = (tp->t_flags & XFS_TRANS_RESERVE) != 0;
 
 	/* Mark this thread as being in a transaction */
 	current_set_flags_nested(&tp->t_pflags, PF_MEMALLOC_NOFS);
@@ -162,7 +162,7 @@ xfs_trans_reserve(
 	 * fail if the count would go below zero.
 	 */
 	if (blocks > 0) {
-		error = xfs_mod_fdblocks(tp->t_mountp, -((int64_t)blocks), rsvd);
+		error = xfs_mod_fdblocks(mp, -((int64_t)blocks), rsvd);
 		if (error != 0) {
 			current_restore_flags_nested(&tp->t_pflags, PF_MEMALLOC_NOFS);
 			return -ENOSPC;
@@ -191,9 +191,9 @@ xfs_trans_reserve(
 
 		if (tp->t_ticket != NULL) {
 			ASSERT(resp->tr_logflags & XFS_TRANS_PERM_LOG_RES);
-			error = xfs_log_regrant(tp->t_mountp, tp->t_ticket);
+			error = xfs_log_regrant(mp, tp->t_ticket);
 		} else {
-			error = xfs_log_reserve(tp->t_mountp,
+			error = xfs_log_reserve(mp,
 						resp->tr_logres,
 						resp->tr_logcount,
 						&tp->t_ticket, XFS_TRANSACTION,
@@ -213,7 +213,7 @@ xfs_trans_reserve(
 	 * fail if the count would go below zero.
 	 */
 	if (rtextents > 0) {
-		error = xfs_mod_frextents(tp->t_mountp, -((int64_t)rtextents));
+		error = xfs_mod_frextents(mp, -((int64_t)rtextents));
 		if (error) {
 			error = -ENOSPC;
 			goto undo_log;
@@ -229,7 +229,7 @@ xfs_trans_reserve(
 	 */
 undo_log:
 	if (resp->tr_logres > 0) {
-		xfs_log_done(tp->t_mountp, tp->t_ticket, NULL, false);
+		xfs_log_ticket_ungrant(mp->m_log, tp->t_ticket);
 		tp->t_ticket = NULL;
 		tp->t_log_res = 0;
 		tp->t_flags &= ~XFS_TRANS_PERM_LOG_RES;
@@ -237,7 +237,7 @@ undo_log:
 
 undo_blocks:
 	if (blocks > 0) {
-		xfs_mod_fdblocks(tp->t_mountp, (int64_t)blocks, rsvd);
+		xfs_mod_fdblocks(mp, (int64_t)blocks, rsvd);
 		tp->t_blk_res = 0;
 	}
 
@@ -263,7 +263,7 @@ xfs_trans_alloc(
 	 * GFP_NOFS allocation context so that we avoid lockdep false positives
 	 * by doing GFP_KERNEL allocations inside sb_start_intwrite().
 	 */
-	tp = kmem_zone_zalloc(xfs_trans_zone, 0);
+	tp = kmem_cache_zalloc(xfs_trans_zone, GFP_KERNEL | __GFP_NOFAIL);
 	if (!(flags & XFS_TRANS_NO_WRITECOUNT))
 		sb_start_intwrite(mp->m_super);
 
@@ -273,7 +273,8 @@ xfs_trans_alloc(
 	 */
 	WARN_ON(resp->tr_logres > 0 &&
 		mp->m_super->s_writers.frozen == SB_FREEZE_COMPLETE);
-	atomic_inc(&mp->m_active_trans);
+	ASSERT(!(flags & XFS_TRANS_RES_FDBLKS) ||
+	       xfs_sb_version_haslazysbcount(&mp->m_sb));
 
 	tp->t_magic = XFS_TRANS_HEADER_MAGIC;
 	tp->t_flags = flags;
@@ -297,15 +298,19 @@ xfs_trans_alloc(
 
 /*
  * Create an empty transaction with no reservation.  This is a defensive
- * mechanism for routines that query metadata without actually modifying
- * them -- if the metadata being queried is somehow cross-linked (think a
- * btree block pointer that points higher in the tree), we risk deadlock.
- * However, blocks grabbed as part of a transaction can be re-grabbed.
- * The verifiers will notice the corrupt block and the operation will fail
- * back to userspace without deadlocking.
+ * mechanism for routines that query metadata without actually modifying them --
+ * if the metadata being queried is somehow cross-linked (think a btree block
+ * pointer that points higher in the tree), we risk deadlock.  However, blocks
+ * grabbed as part of a transaction can be re-grabbed.  The verifiers will
+ * notice the corrupt block and the operation will fail back to userspace
+ * without deadlocking.
  *
- * Note the zero-length reservation; this transaction MUST be cancelled
- * without any dirty data.
+ * Note the zero-length reservation; this transaction MUST be cancelled without
+ * any dirty data.
+ *
+ * Callers should obtain freeze protection to avoid a conflict with fs freezing
+ * where we can be grabbing buffers at the same time that freeze is trying to
+ * drain the buffer LRU list.
  */
 int
 xfs_trans_alloc_empty(
@@ -363,6 +368,20 @@ xfs_trans_mod_sb(
 			tp->t_blk_res_used += (uint)-delta;
 			if (tp->t_blk_res_used > tp->t_blk_res)
 				xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+		} else if (delta > 0 && (tp->t_flags & XFS_TRANS_RES_FDBLKS)) {
+			int64_t	blkres_delta;
+
+			/*
+			 * Return freed blocks directly to the reservation
+			 * instead of the global pool, being careful not to
+			 * overflow the trans counter. This is used to preserve
+			 * reservation across chains of transaction rolls that
+			 * repeatedly free and allocate blocks.
+			 */
+			blkres_delta = min_t(int64_t, delta,
+					     UINT_MAX - tp->t_blk_res);
+			tp->t_blk_res += blkres_delta;
+			delta -= blkres_delta;
 		}
 		tp->t_fdblocks_delta += delta;
 		if (xfs_sb_version_haslazysbcount(&mp->m_sb))
@@ -450,7 +469,7 @@ xfs_trans_apply_sb_deltas(
 	int		whole = 0;
 
 	bp = xfs_trans_getsb(tp, tp->t_mountp);
-	sbp = XFS_BUF_TO_SBP(bp);
+	sbp = bp->b_addr;
 
 	/*
 	 * Check that superblock mods match the mods made to AGF counters.
@@ -527,57 +546,9 @@ xfs_trans_apply_sb_deltas(
 				  sizeof(sbp->sb_frextents) - 1);
 }
 
-STATIC int
-xfs_sb_mod8(
-	uint8_t			*field,
-	int8_t			delta)
-{
-	int8_t			counter = *field;
-
-	counter += delta;
-	if (counter < 0) {
-		ASSERT(0);
-		return -EINVAL;
-	}
-	*field = counter;
-	return 0;
-}
-
-STATIC int
-xfs_sb_mod32(
-	uint32_t		*field,
-	int32_t			delta)
-{
-	int32_t			counter = *field;
-
-	counter += delta;
-	if (counter < 0) {
-		ASSERT(0);
-		return -EINVAL;
-	}
-	*field = counter;
-	return 0;
-}
-
-STATIC int
-xfs_sb_mod64(
-	uint64_t		*field,
-	int64_t			delta)
-{
-	int64_t			counter = *field;
-
-	counter += delta;
-	if (counter < 0) {
-		ASSERT(0);
-		return -EINVAL;
-	}
-	*field = counter;
-	return 0;
-}
-
 /*
- * xfs_trans_unreserve_and_mod_sb() is called to release unused reservations
- * and apply superblock counter changes to the in-core superblock.  The
+ * xfs_trans_unreserve_and_mod_sb() is called to release unused reservations and
+ * apply superblock counter changes to the in-core superblock.  The
  * t_res_fdblocks_delta and t_res_frextents_delta fields are explicitly NOT
  * applied to the in-core superblock.  The idea is that that has already been
  * done.
@@ -586,7 +557,12 @@ xfs_sb_mod64(
  * used block counts are not updated in the on disk superblock. In this case,
  * XFS_TRANS_SB_DIRTY will not be set when the transaction is updated but we
  * still need to update the incore superblock with the changes.
+ *
+ * Deltas for the inode count are +/-64, hence we use a large batch size of 128
+ * so we don't need to take the counter lock on every update.
  */
+#define XFS_ICOUNT_BATCH	128
+
 void
 xfs_trans_unreserve_and_mod_sb(
 	struct xfs_trans	*tp)
@@ -622,20 +598,21 @@ xfs_trans_unreserve_and_mod_sb(
 	/* apply the per-cpu counters */
 	if (blkdelta) {
 		error = xfs_mod_fdblocks(mp, blkdelta, rsvd);
-		if (error)
-			goto out;
+		ASSERT(!error);
 	}
 
 	if (idelta) {
-		error = xfs_mod_icount(mp, idelta);
-		if (error)
-			goto out_undo_fdblocks;
+		percpu_counter_add_batch(&mp->m_icount, idelta,
+					 XFS_ICOUNT_BATCH);
+		if (idelta < 0)
+			ASSERT(__percpu_counter_compare(&mp->m_icount, 0,
+							XFS_ICOUNT_BATCH) >= 0);
 	}
 
 	if (ifreedelta) {
-		error = xfs_mod_ifree(mp, ifreedelta);
-		if (error)
-			goto out_undo_icount;
+		percpu_counter_add(&mp->m_ifree, ifreedelta);
+		if (ifreedelta < 0)
+			ASSERT(percpu_counter_compare(&mp->m_ifree, 0) >= 0);
 	}
 
 	if (rtxdelta == 0 && !(tp->t_flags & XFS_TRANS_SB_DIRTY))
@@ -643,95 +620,23 @@ xfs_trans_unreserve_and_mod_sb(
 
 	/* apply remaining deltas */
 	spin_lock(&mp->m_sb_lock);
-	if (rtxdelta) {
-		error = xfs_sb_mod64(&mp->m_sb.sb_frextents, rtxdelta);
-		if (error)
-			goto out_undo_ifree;
-	}
-
-	if (tp->t_dblocks_delta != 0) {
-		error = xfs_sb_mod64(&mp->m_sb.sb_dblocks, tp->t_dblocks_delta);
-		if (error)
-			goto out_undo_frextents;
-	}
-	if (tp->t_agcount_delta != 0) {
-		error = xfs_sb_mod32(&mp->m_sb.sb_agcount, tp->t_agcount_delta);
-		if (error)
-			goto out_undo_dblocks;
-	}
-	if (tp->t_imaxpct_delta != 0) {
-		error = xfs_sb_mod8(&mp->m_sb.sb_imax_pct, tp->t_imaxpct_delta);
-		if (error)
-			goto out_undo_agcount;
-	}
-	if (tp->t_rextsize_delta != 0) {
-		error = xfs_sb_mod32(&mp->m_sb.sb_rextsize,
-				     tp->t_rextsize_delta);
-		if (error)
-			goto out_undo_imaxpct;
-	}
-	if (tp->t_rbmblocks_delta != 0) {
-		error = xfs_sb_mod32(&mp->m_sb.sb_rbmblocks,
-				     tp->t_rbmblocks_delta);
-		if (error)
-			goto out_undo_rextsize;
-	}
-	if (tp->t_rblocks_delta != 0) {
-		error = xfs_sb_mod64(&mp->m_sb.sb_rblocks, tp->t_rblocks_delta);
-		if (error)
-			goto out_undo_rbmblocks;
-	}
-	if (tp->t_rextents_delta != 0) {
-		error = xfs_sb_mod64(&mp->m_sb.sb_rextents,
-				     tp->t_rextents_delta);
-		if (error)
-			goto out_undo_rblocks;
-	}
-	if (tp->t_rextslog_delta != 0) {
-		error = xfs_sb_mod8(&mp->m_sb.sb_rextslog,
-				     tp->t_rextslog_delta);
-		if (error)
-			goto out_undo_rextents;
-	}
+	mp->m_sb.sb_frextents += rtxdelta;
+	mp->m_sb.sb_dblocks += tp->t_dblocks_delta;
+	mp->m_sb.sb_agcount += tp->t_agcount_delta;
+	mp->m_sb.sb_imax_pct += tp->t_imaxpct_delta;
+	mp->m_sb.sb_rextsize += tp->t_rextsize_delta;
+	mp->m_sb.sb_rbmblocks += tp->t_rbmblocks_delta;
+	mp->m_sb.sb_rblocks += tp->t_rblocks_delta;
+	mp->m_sb.sb_rextents += tp->t_rextents_delta;
+	mp->m_sb.sb_rextslog += tp->t_rextslog_delta;
 	spin_unlock(&mp->m_sb_lock);
-	return;
 
-out_undo_rextents:
-	if (tp->t_rextents_delta)
-		xfs_sb_mod64(&mp->m_sb.sb_rextents, -tp->t_rextents_delta);
-out_undo_rblocks:
-	if (tp->t_rblocks_delta)
-		xfs_sb_mod64(&mp->m_sb.sb_rblocks, -tp->t_rblocks_delta);
-out_undo_rbmblocks:
-	if (tp->t_rbmblocks_delta)
-		xfs_sb_mod32(&mp->m_sb.sb_rbmblocks, -tp->t_rbmblocks_delta);
-out_undo_rextsize:
-	if (tp->t_rextsize_delta)
-		xfs_sb_mod32(&mp->m_sb.sb_rextsize, -tp->t_rextsize_delta);
-out_undo_imaxpct:
-	if (tp->t_rextsize_delta)
-		xfs_sb_mod8(&mp->m_sb.sb_imax_pct, -tp->t_imaxpct_delta);
-out_undo_agcount:
-	if (tp->t_agcount_delta)
-		xfs_sb_mod32(&mp->m_sb.sb_agcount, -tp->t_agcount_delta);
-out_undo_dblocks:
-	if (tp->t_dblocks_delta)
-		xfs_sb_mod64(&mp->m_sb.sb_dblocks, -tp->t_dblocks_delta);
-out_undo_frextents:
-	if (rtxdelta)
-		xfs_sb_mod64(&mp->m_sb.sb_frextents, -rtxdelta);
-out_undo_ifree:
-	spin_unlock(&mp->m_sb_lock);
-	if (ifreedelta)
-		xfs_mod_ifree(mp, -ifreedelta);
-out_undo_icount:
-	if (idelta)
-		xfs_mod_icount(mp, -idelta);
-out_undo_fdblocks:
-	if (blkdelta)
-		xfs_mod_fdblocks(mp, -blkdelta, rsvd);
-out:
-	ASSERT(error == 0);
+	/*
+	 * Debug checks outside of the spinlock so they don't lock up the
+	 * machine if they fail.
+	 */
+	ASSERT(mp->m_sb.sb_imax_pct >= 0);
+	ASSERT(mp->m_sb.sb_rextslog >= 0);
 	return;
 }
 
@@ -999,9 +904,10 @@ out_unreserve:
 	 */
 	xfs_trans_unreserve_and_mod_dquots(tp);
 	if (tp->t_ticket) {
-		commit_lsn = xfs_log_done(mp, tp->t_ticket, NULL, regrant);
-		if (commit_lsn == -1 && !error)
-			error = -EIO;
+		if (regrant && !XLOG_FORCED_SHUTDOWN(mp->m_log))
+			xfs_log_ticket_regrant(mp->m_log, tp->t_ticket);
+		else
+			xfs_log_ticket_ungrant(mp->m_log, tp->t_ticket);
 		tp->t_ticket = NULL;
 	}
 	current_restore_flags_nested(&tp->t_pflags, PF_MEMALLOC_NOFS);
@@ -1060,7 +966,7 @@ xfs_trans_cancel(
 	xfs_trans_unreserve_and_mod_dquots(tp);
 
 	if (tp->t_ticket) {
-		xfs_log_done(mp, tp->t_ticket, NULL, false);
+		xfs_log_ticket_ungrant(mp->m_log, tp->t_ticket);
 		tp->t_ticket = NULL;
 	}
 
diff --git a/fs/xfs/xfs_trans.h b/fs/xfs/xfs_trans.h
index 752c7fef9de7..b752501818d2 100644
--- a/fs/xfs/xfs_trans.h
+++ b/fs/xfs/xfs_trans.h
@@ -37,10 +37,6 @@ struct xfs_log_item {
 	unsigned long			li_flags;	/* misc flags */
 	struct xfs_buf			*li_buf;	/* real buffer pointer */
 	struct list_head		li_bio_list;	/* buffer item list */
-	void				(*li_cb)(struct xfs_buf *,
-						 struct xfs_log_item *);
-							/* buffer item iodone */
-							/* callback func */
 	const struct xfs_item_ops	*li_ops;	/* function list */
 
 	/* delayed logging */
@@ -59,12 +55,14 @@ struct xfs_log_item {
 #define	XFS_LI_ABORTED	1
 #define	XFS_LI_FAILED	2
 #define	XFS_LI_DIRTY	3	/* log item dirty in transaction */
+#define	XFS_LI_RECOVERED 4	/* log intent item has been recovered */
 
 #define XFS_LI_FLAGS \
 	{ (1 << XFS_LI_IN_AIL),		"IN_AIL" }, \
 	{ (1 << XFS_LI_ABORTED),	"ABORTED" }, \
 	{ (1 << XFS_LI_FAILED),		"FAILED" }, \
-	{ (1 << XFS_LI_DIRTY),		"DIRTY" }
+	{ (1 << XFS_LI_DIRTY),		"DIRTY" }, \
+	{ (1 << XFS_LI_RECOVERED),	"RECOVERED" }
 
 struct xfs_item_ops {
 	unsigned flags;
@@ -76,7 +74,8 @@ struct xfs_item_ops {
 	void (*iop_committing)(struct xfs_log_item *, xfs_lsn_t commit_lsn);
 	void (*iop_release)(struct xfs_log_item *);
 	xfs_lsn_t (*iop_committed)(struct xfs_log_item *, xfs_lsn_t);
-	void (*iop_error)(struct xfs_log_item *, xfs_buf_t *);
+	int (*iop_recover)(struct xfs_log_item *lip, struct xfs_trans *tp);
+	bool (*iop_match)(struct xfs_log_item *item, uint64_t id);
 };
 
 /*
diff --git a/fs/xfs/xfs_trans_ail.c b/fs/xfs/xfs_trans_ail.c
index 00cc5b8734be..dbb69b4bf3ed 100644
--- a/fs/xfs/xfs_trans_ail.c
+++ b/fs/xfs/xfs_trans_ail.c
@@ -32,6 +32,7 @@ STATIC void
 xfs_ail_check(
 	struct xfs_ail		*ailp,
 	struct xfs_log_item	*lip)
+	__must_hold(&ailp->ail_lock)
 {
 	struct xfs_log_item	*prev_lip;
 	struct xfs_log_item	*next_lip;
@@ -108,17 +109,25 @@ xfs_ail_next(
  * We need the AIL lock in order to get a coherent read of the lsn of the last
  * item in the AIL.
  */
+static xfs_lsn_t
+__xfs_ail_min_lsn(
+	struct xfs_ail		*ailp)
+{
+	struct xfs_log_item	*lip = xfs_ail_min(ailp);
+
+	if (lip)
+		return lip->li_lsn;
+	return 0;
+}
+
 xfs_lsn_t
 xfs_ail_min_lsn(
 	struct xfs_ail		*ailp)
 {
-	xfs_lsn_t		lsn = 0;
-	struct xfs_log_item	*lip;
+	xfs_lsn_t		lsn;
 
 	spin_lock(&ailp->ail_lock);
-	lip = xfs_ail_min(ailp);
-	if (lip)
-		lsn = lip->li_lsn;
+	lsn = __xfs_ail_min_lsn(ailp);
 	spin_unlock(&ailp->ail_lock);
 
 	return lsn;
@@ -336,6 +345,49 @@ xfs_ail_delete(
 	xfs_trans_ail_cursor_clear(ailp, lip);
 }
 
+/*
+ * Requeue a failed buffer for writeback.
+ *
+ * We clear the log item failed state here as well, but we have to be careful
+ * about reference counts because the only active reference counts on the buffer
+ * may be the failed log items. Hence if we clear the log item failed state
+ * before queuing the buffer for IO we can release all active references to
+ * the buffer and free it, leading to use after free problems in
+ * xfs_buf_delwri_queue. It makes no difference to the buffer or log items which
+ * order we process them in - the buffer is locked, and we own the buffer list
+ * so nothing on them is going to change while we are performing this action.
+ *
+ * Hence we can safely queue the buffer for IO before we clear the failed log
+ * item state, therefore  always having an active reference to the buffer and
+ * avoiding the transient zero-reference state that leads to use-after-free.
+ */
+static inline int
+xfsaild_resubmit_item(
+	struct xfs_log_item	*lip,
+	struct list_head	*buffer_list)
+{
+	struct xfs_buf		*bp = lip->li_buf;
+
+	if (!xfs_buf_trylock(bp))
+		return XFS_ITEM_LOCKED;
+
+	if (!xfs_buf_delwri_queue(bp, buffer_list)) {
+		xfs_buf_unlock(bp);
+		return XFS_ITEM_FLUSHING;
+	}
+
+	/* protected by ail_lock */
+	list_for_each_entry(lip, &bp->b_li_list, li_bio_list) {
+		if (bp->b_flags & _XBF_INODES)
+			clear_bit(XFS_LI_FAILED, &lip->li_flags);
+		else
+			xfs_clear_li_failed(lip);
+	}
+
+	xfs_buf_unlock(bp);
+	return XFS_ITEM_SUCCESS;
+}
+
 static inline uint
 xfsaild_push_item(
 	struct xfs_ail		*ailp,
@@ -356,6 +408,8 @@ xfsaild_push_item(
 	 */
 	if (!lip->li_ops->iop_push)
 		return XFS_ITEM_PINNED;
+	if (test_bit(XFS_LI_FAILED, &lip->li_flags))
+		return xfsaild_resubmit_item(lip, &ailp->ail_buf_list);
 	return lip->li_ops->iop_push(lip, &ailp->ail_buf_list);
 }
 
@@ -394,16 +448,10 @@ xfsaild_push(
 	target = ailp->ail_target;
 	ailp->ail_target_prev = target;
 
+	/* we're done if the AIL is empty or our push has reached the end */
 	lip = xfs_trans_ail_cursor_first(ailp, &cur, ailp->ail_last_pushed_lsn);
-	if (!lip) {
-		/*
-		 * If the AIL is empty or our push has reached the end we are
-		 * done now.
-		 */
-		xfs_trans_ail_cursor_done(&cur);
-		spin_unlock(&ailp->ail_lock);
+	if (!lip)
 		goto out_done;
-	}
 
 	XFS_STATS_INC(mp, xs_push_ail);
 
@@ -432,7 +480,7 @@ xfsaild_push(
 			 * inode buffer is locked because we already pushed the
 			 * updates to it as part of inode clustering.
 			 *
-			 * We do not want to to stop flushing just because lots
+			 * We do not want to stop flushing just because lots
 			 * of items are already being flushed, but we need to
 			 * re-try the flushing relatively soon if most of the
 			 * AIL is being flushed.
@@ -467,7 +515,7 @@ xfsaild_push(
 		/*
 		 * Are there too many items we can't do anything with?
 		 *
-		 * If we we are skipping too many items because we can't flush
+		 * If we are skipping too many items because we can't flush
 		 * them or they are already being flushed, we back off and
 		 * given them time to complete whatever operation is being
 		 * done. i.e. remove pressure from the AIL while we can't make
@@ -485,6 +533,8 @@ xfsaild_push(
 			break;
 		lsn = lip->li_lsn;
 	}
+
+out_done:
 	xfs_trans_ail_cursor_done(&cur);
 	spin_unlock(&ailp->ail_lock);
 
@@ -492,7 +542,6 @@ xfsaild_push(
 		ailp->ail_log_flush++;
 
 	if (!count || XFS_LSN_CMP(lsn, target) >= 0) {
-out_done:
 		/*
 		 * We reached the target or the AIL is empty, so wait a bit
 		 * longer for I/O to complete and remove pushed items from the
@@ -529,8 +578,9 @@ xfsaild(
 {
 	struct xfs_ail	*ailp = data;
 	long		tout = 0;	/* milliseconds */
+	unsigned int	noreclaim_flag;
 
-	current->flags |= PF_MEMALLOC;
+	noreclaim_flag = memalloc_noreclaim_save();
 	set_freezable();
 
 	while (1) {
@@ -583,7 +633,8 @@ xfsaild(
 		 */
 		smp_rmb();
 		if (!xfs_ail_min(ailp) &&
-		    ailp->ail_target == ailp->ail_target_prev) {
+		    ailp->ail_target == ailp->ail_target_prev &&
+		    list_empty(&ailp->ail_buf_list)) {
 			spin_unlock(&ailp->ail_lock);
 			freezable_schedule();
 			tout = 0;
@@ -601,6 +652,7 @@ xfsaild(
 		tout = xfsaild_push(ailp);
 	}
 
+	memalloc_noreclaim_restore(noreclaim_flag);
 	return 0;
 }
 
@@ -678,6 +730,28 @@ xfs_ail_push_all_sync(
 	finish_wait(&ailp->ail_empty, &wait);
 }
 
+void
+xfs_ail_update_finish(
+	struct xfs_ail		*ailp,
+	xfs_lsn_t		old_lsn) __releases(ailp->ail_lock)
+{
+	struct xfs_mount	*mp = ailp->ail_mount;
+
+	/* if the tail lsn hasn't changed, don't do updates or wakeups. */
+	if (!old_lsn || old_lsn == __xfs_ail_min_lsn(ailp)) {
+		spin_unlock(&ailp->ail_lock);
+		return;
+	}
+
+	if (!XFS_FORCED_SHUTDOWN(mp))
+		xlog_assign_tail_lsn_locked(mp);
+
+	if (list_empty(&ailp->ail_head))
+		wake_up_all(&ailp->ail_empty);
+	spin_unlock(&ailp->ail_lock);
+	xfs_log_space_wake(mp);
+}
+
 /*
  * xfs_trans_ail_update - bulk AIL insertion operation.
  *
@@ -709,7 +783,7 @@ xfs_trans_ail_update_bulk(
 	xfs_lsn_t		lsn) __releases(ailp->ail_lock)
 {
 	struct xfs_log_item	*mlip;
-	int			mlip_changed = 0;
+	xfs_lsn_t		tail_lsn = 0;
 	int			i;
 	LIST_HEAD(tmp);
 
@@ -724,9 +798,10 @@ xfs_trans_ail_update_bulk(
 				continue;
 
 			trace_xfs_ail_move(lip, lip->li_lsn, lsn);
+			if (mlip == lip && !tail_lsn)
+				tail_lsn = lip->li_lsn;
+
 			xfs_ail_delete(ailp, lip);
-			if (mlip == lip)
-				mlip_changed = 1;
 		} else {
 			trace_xfs_ail_insert(lip, 0, lsn);
 		}
@@ -737,66 +812,58 @@ xfs_trans_ail_update_bulk(
 	if (!list_empty(&tmp))
 		xfs_ail_splice(ailp, cur, &tmp, lsn);
 
-	if (mlip_changed) {
-		if (!XFS_FORCED_SHUTDOWN(ailp->ail_mount))
-			xlog_assign_tail_lsn_locked(ailp->ail_mount);
-		spin_unlock(&ailp->ail_lock);
+	xfs_ail_update_finish(ailp, tail_lsn);
+}
 
-		xfs_log_space_wake(ailp->ail_mount);
-	} else {
-		spin_unlock(&ailp->ail_lock);
-	}
+/* Insert a log item into the AIL. */
+void
+xfs_trans_ail_insert(
+	struct xfs_ail		*ailp,
+	struct xfs_log_item	*lip,
+	xfs_lsn_t		lsn)
+{
+	spin_lock(&ailp->ail_lock);
+	xfs_trans_ail_update_bulk(ailp, NULL, &lip, 1, lsn);
 }
 
-bool
+/*
+ * Delete one log item from the AIL.
+ *
+ * If this item was at the tail of the AIL, return the LSN of the log item so
+ * that we can use it to check if the LSN of the tail of the log has moved
+ * when finishing up the AIL delete process in xfs_ail_update_finish().
+ */
+xfs_lsn_t
 xfs_ail_delete_one(
 	struct xfs_ail		*ailp,
 	struct xfs_log_item	*lip)
 {
 	struct xfs_log_item	*mlip = xfs_ail_min(ailp);
+	xfs_lsn_t		lsn = lip->li_lsn;
 
 	trace_xfs_ail_delete(lip, mlip->li_lsn, lip->li_lsn);
 	xfs_ail_delete(ailp, lip);
-	xfs_clear_li_failed(lip);
 	clear_bit(XFS_LI_IN_AIL, &lip->li_flags);
 	lip->li_lsn = 0;
 
-	return mlip == lip;
+	if (mlip == lip)
+		return lsn;
+	return 0;
 }
 
-/**
- * Remove a log items from the AIL
- *
- * @xfs_trans_ail_delete_bulk takes an array of log items that all need to
- * removed from the AIL. The caller is already holding the AIL lock, and done
- * all the checks necessary to ensure the items passed in via @log_items are
- * ready for deletion. This includes checking that the items are in the AIL.
- *
- * For each log item to be removed, unlink it  from the AIL, clear the IN_AIL
- * flag from the item and reset the item's lsn to 0. If we remove the first
- * item in the AIL, update the log tail to match the new minimum LSN in the
- * AIL.
- *
- * This function will not drop the AIL lock until all items are removed from
- * the AIL to minimise the amount of lock traffic on the AIL. This does not
- * greatly increase the AIL hold time, but does significantly reduce the amount
- * of traffic on the lock, especially during IO completion.
- *
- * This function must be called with the AIL lock held.  The lock is dropped
- * before returning.
- */
 void
 xfs_trans_ail_delete(
-	struct xfs_ail		*ailp,
 	struct xfs_log_item	*lip,
-	int			shutdown_type) __releases(ailp->ail_lock)
+	int			shutdown_type)
 {
+	struct xfs_ail		*ailp = lip->li_ailp;
 	struct xfs_mount	*mp = ailp->ail_mount;
-	bool			mlip_changed;
+	xfs_lsn_t		tail_lsn;
 
+	spin_lock(&ailp->ail_lock);
 	if (!test_bit(XFS_LI_IN_AIL, &lip->li_flags)) {
 		spin_unlock(&ailp->ail_lock);
-		if (!XFS_FORCED_SHUTDOWN(mp)) {
+		if (shutdown_type && !XFS_FORCED_SHUTDOWN(mp)) {
 			xfs_alert_tag(mp, XFS_PTAG_AILDELETE,
 	"%s: attempting to delete a log item that is not in the AIL",
 					__func__);
@@ -805,17 +872,10 @@ xfs_trans_ail_delete(
 		return;
 	}
 
-	mlip_changed = xfs_ail_delete_one(ailp, lip);
-	if (mlip_changed) {
-		if (!XFS_FORCED_SHUTDOWN(mp))
-			xlog_assign_tail_lsn_locked(mp);
-		if (list_empty(&ailp->ail_head))
-			wake_up_all(&ailp->ail_empty);
-	}
-
-	spin_unlock(&ailp->ail_lock);
-	if (mlip_changed)
-		xfs_log_space_wake(ailp->ail_mount);
+	/* xfs_ail_update_finish() drops the AIL lock */
+	xfs_clear_li_failed(lip);
+	tail_lsn = xfs_ail_delete_one(ailp, lip);
+	xfs_ail_update_finish(ailp, tail_lsn);
 }
 
 int
diff --git a/fs/xfs/xfs_trans_buf.c b/fs/xfs/xfs_trans_buf.c
index 08174ffa2118..11cd666cd99a 100644
--- a/fs/xfs/xfs_trans_buf.c
+++ b/fs/xfs/xfs_trans_buf.c
@@ -465,24 +465,16 @@ xfs_trans_dirty_buf(
 
 	ASSERT(bp->b_transp == tp);
 	ASSERT(bip != NULL);
-	ASSERT(bp->b_iodone == NULL ||
-	       bp->b_iodone == xfs_buf_iodone_callbacks);
 
 	/*
 	 * Mark the buffer as needing to be written out eventually,
 	 * and set its iodone function to remove the buffer's buf log
 	 * item from the AIL and free it when the buffer is flushed
-	 * to disk.  See xfs_buf_attach_iodone() for more details
-	 * on li_cb and xfs_buf_iodone_callbacks().
-	 * If we end up aborting this transaction, we trap this buffer
-	 * inside the b_bdstrat callback so that this won't get written to
-	 * disk.
+	 * to disk.
 	 */
 	bp->b_flags |= XBF_DONE;
 
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
-	bp->b_iodone = xfs_buf_iodone_callbacks;
-	bip->bli_item.li_cb = xfs_buf_iodone;
 
 	/*
 	 * If we invalidated the buffer within this transaction, then
@@ -626,6 +618,7 @@ xfs_trans_inode_buf(
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 
 	bip->bli_flags |= XFS_BLI_INODE_BUF;
+	bp->b_flags |= _XBF_INODES;
 	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DINO_BUF);
 }
 
@@ -650,7 +643,7 @@ xfs_trans_stale_inode_buf(
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 
 	bip->bli_flags |= XFS_BLI_STALE_INODE;
-	bip->bli_item.li_cb = xfs_buf_iodone;
+	bp->b_flags |= _XBF_INODES;
 	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DINO_BUF);
 }
 
@@ -675,6 +668,7 @@ xfs_trans_inode_alloc_buf(
 	ASSERT(atomic_read(&bip->bli_refcount) > 0);
 
 	bip->bli_flags |= XFS_BLI_INODE_ALLOC_BUF;
+	bp->b_flags |= _XBF_INODES;
 	xfs_trans_buf_set_type(tp, bp, XFS_BLFT_DINO_BUF);
 }
 
@@ -785,5 +779,6 @@ xfs_trans_dquot_buf(
 		break;
 	}
 
+	bp->b_flags |= _XBF_DQUOTS;
 	xfs_trans_buf_set_type(tp, bp, type);
 }
diff --git a/fs/xfs/xfs_trans_dquot.c b/fs/xfs/xfs_trans_dquot.c
index d1b9869bc5fa..c6ba7ef18e06 100644
--- a/fs/xfs/xfs_trans_dquot.c
+++ b/fs/xfs/xfs_trans_dquot.c
@@ -15,6 +15,7 @@
 #include "xfs_trans_priv.h"
 #include "xfs_quota.h"
 #include "xfs_qm.h"
+#include "xfs_trace.h"
 
 STATIC void	xfs_trans_alloc_dqinfo(xfs_trans_t *);
 
@@ -155,14 +156,19 @@ xfs_trans_get_dqtrx(
 	int			i;
 	struct xfs_dqtrx	*qa;
 
-	if (XFS_QM_ISUDQ(dqp))
+	switch (xfs_dquot_type(dqp)) {
+	case XFS_DQTYPE_USER:
 		qa = tp->t_dqinfo->dqs[XFS_QM_TRANS_USR];
-	else if (XFS_QM_ISGDQ(dqp))
+		break;
+	case XFS_DQTYPE_GROUP:
 		qa = tp->t_dqinfo->dqs[XFS_QM_TRANS_GRP];
-	else if (XFS_QM_ISPDQ(dqp))
+		break;
+	case XFS_DQTYPE_PROJ:
 		qa = tp->t_dqinfo->dqs[XFS_QM_TRANS_PRJ];
-	else
+		break;
+	default:
 		return NULL;
+	}
 
 	for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) {
 		if (qa[i].qt_dquot == NULL ||
@@ -203,6 +209,11 @@ xfs_trans_mod_dquot(
 	if (qtrx->qt_dquot == NULL)
 		qtrx->qt_dquot = dqp;
 
+	if (delta) {
+		trace_xfs_trans_mod_dquot_before(qtrx);
+		trace_xfs_trans_mod_dquot(tp, dqp, field, delta);
+	}
+
 	switch (field) {
 
 		/*
@@ -266,6 +277,10 @@ xfs_trans_mod_dquot(
 	      default:
 		ASSERT(0);
 	}
+
+	if (delta)
+		trace_xfs_trans_mod_dquot_after(qtrx);
+
 	tp->t_flags |= XFS_TRANS_DQ_DIRTY;
 }
 
@@ -293,6 +308,37 @@ xfs_trans_dqlockedjoin(
 	}
 }
 
+/* Apply dqtrx changes to the quota reservation counters. */
+static inline void
+xfs_apply_quota_reservation_deltas(
+	struct xfs_dquot_res	*res,
+	uint64_t		reserved,
+	int64_t			res_used,
+	int64_t			count_delta)
+{
+	if (reserved != 0) {
+		/*
+		 * Subtle math here: If reserved > res_used (the normal case),
+		 * we're simply subtracting the unused transaction quota
+		 * reservation from the dquot reservation.
+		 *
+		 * If, however, res_used > reserved, then we have allocated
+		 * more quota blocks than were reserved for the transaction.
+		 * We must add that excess to the dquot reservation since it
+		 * tracks (usage + resv) and by definition we didn't reserve
+		 * that excess.
+		 */
+		res->reserved -= abs(reserved - res_used);
+	} else if (count_delta != 0) {
+		/*
+		 * These blks were never reserved, either inside a transaction
+		 * or outside one (in a delayed allocation). Also, this isn't
+		 * always a negative number since we sometimes deliberately
+		 * skip quota reservations.
+		 */
+		res->reserved += count_delta;
+	}
+}
 
 /*
  * Called by xfs_trans_commit() and similar in spirit to
@@ -309,7 +355,6 @@ xfs_trans_apply_dquot_deltas(
 	int			i, j;
 	struct xfs_dquot	*dqp;
 	struct xfs_dqtrx	*qtrx, *qa;
-	struct xfs_disk_dquot	*d;
 	int64_t			totalbdelta;
 	int64_t			totalrtbdelta;
 
@@ -328,6 +373,8 @@ xfs_trans_apply_dquot_deltas(
 		xfs_trans_dqlockedjoin(tp, qa);
 
 		for (i = 0; i < XFS_QM_TRANS_MAXDQS; i++) {
+			uint64_t	blk_res_used;
+
 			qtrx = &qa[i];
 			/*
 			 * The array of dquots is filled
@@ -341,7 +388,6 @@ xfs_trans_apply_dquot_deltas(
 			/*
 			 * adjust the actual number of blocks used
 			 */
-			d = &dqp->q_core;
 
 			/*
 			 * The issue here is - sometimes we don't make a blkquota
@@ -360,38 +406,46 @@ xfs_trans_apply_dquot_deltas(
 				qtrx->qt_delbcnt_delta;
 			totalrtbdelta = qtrx->qt_rtbcount_delta +
 				qtrx->qt_delrtb_delta;
+
+			if (totalbdelta != 0 || totalrtbdelta != 0 ||
+			    qtrx->qt_icount_delta != 0) {
+				trace_xfs_trans_apply_dquot_deltas_before(dqp);
+				trace_xfs_trans_apply_dquot_deltas(qtrx);
+			}
+
 #ifdef DEBUG
 			if (totalbdelta < 0)
-				ASSERT(be64_to_cpu(d->d_bcount) >=
-				       -totalbdelta);
+				ASSERT(dqp->q_blk.count >= -totalbdelta);
 
 			if (totalrtbdelta < 0)
-				ASSERT(be64_to_cpu(d->d_rtbcount) >=
-				       -totalrtbdelta);
+				ASSERT(dqp->q_rtb.count >= -totalrtbdelta);
 
 			if (qtrx->qt_icount_delta < 0)
-				ASSERT(be64_to_cpu(d->d_icount) >=
-				       -qtrx->qt_icount_delta);
+				ASSERT(dqp->q_ino.count >= -qtrx->qt_icount_delta);
 #endif
 			if (totalbdelta)
-				be64_add_cpu(&d->d_bcount, (xfs_qcnt_t)totalbdelta);
+				dqp->q_blk.count += totalbdelta;
 
 			if (qtrx->qt_icount_delta)
-				be64_add_cpu(&d->d_icount, (xfs_qcnt_t)qtrx->qt_icount_delta);
+				dqp->q_ino.count += qtrx->qt_icount_delta;
 
 			if (totalrtbdelta)
-				be64_add_cpu(&d->d_rtbcount, (xfs_qcnt_t)totalrtbdelta);
+				dqp->q_rtb.count += totalrtbdelta;
+
+			if (totalbdelta != 0 || totalrtbdelta != 0 ||
+			    qtrx->qt_icount_delta != 0)
+				trace_xfs_trans_apply_dquot_deltas_after(dqp);
 
 			/*
 			 * Get any default limits in use.
 			 * Start/reset the timer(s) if needed.
 			 */
-			if (d->d_id) {
-				xfs_qm_adjust_dqlimits(tp->t_mountp, dqp);
-				xfs_qm_adjust_dqtimers(tp->t_mountp, d);
+			if (dqp->q_id) {
+				xfs_qm_adjust_dqlimits(dqp);
+				xfs_qm_adjust_dqtimers(dqp);
 			}
 
-			dqp->dq_flags |= XFS_DQ_DIRTY;
+			dqp->q_flags |= XFS_DQFLAG_DIRTY;
 			/*
 			 * add this to the list of items to get logged
 			 */
@@ -401,78 +455,31 @@ xfs_trans_apply_dquot_deltas(
 			 * In case of delayed allocations, there's no
 			 * reservation that a transaction structure knows of.
 			 */
-			if (qtrx->qt_blk_res != 0) {
-				uint64_t	blk_res_used = 0;
-
-				if (qtrx->qt_bcount_delta > 0)
-					blk_res_used = qtrx->qt_bcount_delta;
-
-				if (qtrx->qt_blk_res != blk_res_used) {
-					if (qtrx->qt_blk_res > blk_res_used)
-						dqp->q_res_bcount -= (xfs_qcnt_t)
-							(qtrx->qt_blk_res -
-							 blk_res_used);
-					else
-						dqp->q_res_bcount -= (xfs_qcnt_t)
-							(blk_res_used -
-							 qtrx->qt_blk_res);
-				}
-			} else {
-				/*
-				 * These blks were never reserved, either inside
-				 * a transaction or outside one (in a delayed
-				 * allocation). Also, this isn't always a
-				 * negative number since we sometimes
-				 * deliberately skip quota reservations.
-				 */
-				if (qtrx->qt_bcount_delta) {
-					dqp->q_res_bcount +=
-					      (xfs_qcnt_t)qtrx->qt_bcount_delta;
-				}
-			}
+			blk_res_used = max_t(int64_t, 0, qtrx->qt_bcount_delta);
+			xfs_apply_quota_reservation_deltas(&dqp->q_blk,
+					qtrx->qt_blk_res, blk_res_used,
+					qtrx->qt_bcount_delta);
+
 			/*
 			 * Adjust the RT reservation.
 			 */
-			if (qtrx->qt_rtblk_res != 0) {
-				if (qtrx->qt_rtblk_res != qtrx->qt_rtblk_res_used) {
-					if (qtrx->qt_rtblk_res >
-					    qtrx->qt_rtblk_res_used)
-					       dqp->q_res_rtbcount -= (xfs_qcnt_t)
-						       (qtrx->qt_rtblk_res -
-							qtrx->qt_rtblk_res_used);
-					else
-					       dqp->q_res_rtbcount -= (xfs_qcnt_t)
-						       (qtrx->qt_rtblk_res_used -
-							qtrx->qt_rtblk_res);
-				}
-			} else {
-				if (qtrx->qt_rtbcount_delta)
-					dqp->q_res_rtbcount +=
-					    (xfs_qcnt_t)qtrx->qt_rtbcount_delta;
-			}
+			xfs_apply_quota_reservation_deltas(&dqp->q_rtb,
+					qtrx->qt_rtblk_res,
+					qtrx->qt_rtblk_res_used,
+					qtrx->qt_rtbcount_delta);
 
 			/*
 			 * Adjust the inode reservation.
 			 */
-			if (qtrx->qt_ino_res != 0) {
-				ASSERT(qtrx->qt_ino_res >=
-				       qtrx->qt_ino_res_used);
-				if (qtrx->qt_ino_res > qtrx->qt_ino_res_used)
-					dqp->q_res_icount -= (xfs_qcnt_t)
-						(qtrx->qt_ino_res -
-						 qtrx->qt_ino_res_used);
-			} else {
-				if (qtrx->qt_icount_delta)
-					dqp->q_res_icount +=
-					    (xfs_qcnt_t)qtrx->qt_icount_delta;
-			}
-
-			ASSERT(dqp->q_res_bcount >=
-				be64_to_cpu(dqp->q_core.d_bcount));
-			ASSERT(dqp->q_res_icount >=
-				be64_to_cpu(dqp->q_core.d_icount));
-			ASSERT(dqp->q_res_rtbcount >=
-				be64_to_cpu(dqp->q_core.d_rtbcount));
+			ASSERT(qtrx->qt_ino_res >= qtrx->qt_ino_res_used);
+			xfs_apply_quota_reservation_deltas(&dqp->q_ino,
+					qtrx->qt_ino_res,
+					qtrx->qt_ino_res_used,
+					qtrx->qt_icount_delta);
+
+			ASSERT(dqp->q_blk.reserved >= dqp->q_blk.count);
+			ASSERT(dqp->q_ino.reserved >= dqp->q_ino.count);
+			ASSERT(dqp->q_rtb.reserved >= dqp->q_rtb.count);
 		}
 	}
 }
@@ -516,7 +523,7 @@ xfs_trans_unreserve_and_mod_dquots(
 			if (qtrx->qt_blk_res) {
 				xfs_dqlock(dqp);
 				locked = true;
-				dqp->q_res_bcount -=
+				dqp->q_blk.reserved -=
 					(xfs_qcnt_t)qtrx->qt_blk_res;
 			}
 			if (qtrx->qt_ino_res) {
@@ -524,7 +531,7 @@ xfs_trans_unreserve_and_mod_dquots(
 					xfs_dqlock(dqp);
 					locked = true;
 				}
-				dqp->q_res_icount -=
+				dqp->q_ino.reserved -=
 					(xfs_qcnt_t)qtrx->qt_ino_res;
 			}
 
@@ -533,7 +540,7 @@ xfs_trans_unreserve_and_mod_dquots(
 					xfs_dqlock(dqp);
 					locked = true;
 				}
-				dqp->q_res_rtbcount -=
+				dqp->q_rtb.reserved -=
 					(xfs_qcnt_t)qtrx->qt_rtblk_res;
 			}
 			if (locked)
@@ -549,21 +556,80 @@ xfs_quota_warn(
 	struct xfs_dquot	*dqp,
 	int			type)
 {
-	enum quota_type qtype;
+	enum quota_type		qtype;
 
-	if (dqp->dq_flags & XFS_DQ_PROJ)
+	switch (xfs_dquot_type(dqp)) {
+	case XFS_DQTYPE_PROJ:
 		qtype = PRJQUOTA;
-	else if (dqp->dq_flags & XFS_DQ_USER)
+		break;
+	case XFS_DQTYPE_USER:
 		qtype = USRQUOTA;
-	else
+		break;
+	case XFS_DQTYPE_GROUP:
 		qtype = GRPQUOTA;
+		break;
+	default:
+		return;
+	}
 
-	quota_send_warning(make_kqid(&init_user_ns, qtype,
-				     be32_to_cpu(dqp->q_core.d_id)),
+	quota_send_warning(make_kqid(&init_user_ns, qtype, dqp->q_id),
 			   mp->m_super->s_dev, type);
 }
 
 /*
+ * Decide if we can make an additional reservation against a quota resource.
+ * Returns an inode QUOTA_NL_ warning code and whether or not it's fatal.
+ *
+ * Note that we assume that the numeric difference between the inode and block
+ * warning codes will always be 3 since it's userspace ABI now, and will never
+ * decrease the quota reservation, so the *BELOW messages are irrelevant.
+ */
+static inline int
+xfs_dqresv_check(
+	struct xfs_dquot_res	*res,
+	struct xfs_quota_limits	*qlim,
+	int64_t			delta,
+	bool			*fatal)
+{
+	xfs_qcnt_t		hardlimit = res->hardlimit;
+	xfs_qcnt_t		softlimit = res->softlimit;
+	xfs_qcnt_t		total_count = res->reserved + delta;
+
+	BUILD_BUG_ON(QUOTA_NL_BHARDWARN     != QUOTA_NL_IHARDWARN + 3);
+	BUILD_BUG_ON(QUOTA_NL_BSOFTLONGWARN != QUOTA_NL_ISOFTLONGWARN + 3);
+	BUILD_BUG_ON(QUOTA_NL_BSOFTWARN     != QUOTA_NL_ISOFTWARN + 3);
+
+	*fatal = false;
+	if (delta <= 0)
+		return QUOTA_NL_NOWARN;
+
+	if (!hardlimit)
+		hardlimit = qlim->hard;
+	if (!softlimit)
+		softlimit = qlim->soft;
+
+	if (hardlimit && total_count > hardlimit) {
+		*fatal = true;
+		return QUOTA_NL_IHARDWARN;
+	}
+
+	if (softlimit && total_count > softlimit) {
+		time64_t	now = ktime_get_real_seconds();
+
+		if ((res->timer != 0 && now > res->timer) ||
+		    (res->warnings != 0 && res->warnings >= qlim->warn)) {
+			*fatal = true;
+			return QUOTA_NL_ISOFTLONGWARN;
+		}
+
+		res->warnings++;
+		return QUOTA_NL_ISOFTWARN;
+	}
+
+	return QUOTA_NL_NOWARN;
+}
+
+/*
  * This reserves disk blocks and inodes against a dquot.
  * Flags indicate if the dquot is to be locked here and also
  * if the blk reservation is for RT or regular blocks.
@@ -578,110 +644,58 @@ xfs_trans_dqresv(
 	long			ninos,
 	uint			flags)
 {
-	xfs_qcnt_t		hardlimit;
-	xfs_qcnt_t		softlimit;
-	time64_t		timer;
-	xfs_qwarncnt_t		warns;
-	xfs_qwarncnt_t		warnlimit;
-	xfs_qcnt_t		total_count;
-	xfs_qcnt_t		*resbcountp;
 	struct xfs_quotainfo	*q = mp->m_quotainfo;
 	struct xfs_def_quota	*defq;
-
+	struct xfs_dquot_res	*blkres;
+	struct xfs_quota_limits	*qlim;
 
 	xfs_dqlock(dqp);
 
-	defq = xfs_get_defquota(dqp, q);
+	defq = xfs_get_defquota(q, xfs_dquot_type(dqp));
 
 	if (flags & XFS_TRANS_DQ_RES_BLKS) {
-		hardlimit = be64_to_cpu(dqp->q_core.d_blk_hardlimit);
-		if (!hardlimit)
-			hardlimit = defq->bhardlimit;
-		softlimit = be64_to_cpu(dqp->q_core.d_blk_softlimit);
-		if (!softlimit)
-			softlimit = defq->bsoftlimit;
-		timer = be32_to_cpu(dqp->q_core.d_btimer);
-		warns = be16_to_cpu(dqp->q_core.d_bwarns);
-		warnlimit = dqp->q_mount->m_quotainfo->qi_bwarnlimit;
-		resbcountp = &dqp->q_res_bcount;
+		blkres = &dqp->q_blk;
+		qlim = &defq->blk;
 	} else {
-		ASSERT(flags & XFS_TRANS_DQ_RES_RTBLKS);
-		hardlimit = be64_to_cpu(dqp->q_core.d_rtb_hardlimit);
-		if (!hardlimit)
-			hardlimit = defq->rtbhardlimit;
-		softlimit = be64_to_cpu(dqp->q_core.d_rtb_softlimit);
-		if (!softlimit)
-			softlimit = defq->rtbsoftlimit;
-		timer = be32_to_cpu(dqp->q_core.d_rtbtimer);
-		warns = be16_to_cpu(dqp->q_core.d_rtbwarns);
-		warnlimit = dqp->q_mount->m_quotainfo->qi_rtbwarnlimit;
-		resbcountp = &dqp->q_res_rtbcount;
+		blkres = &dqp->q_rtb;
+		qlim = &defq->rtb;
 	}
 
-	if ((flags & XFS_QMOPT_FORCE_RES) == 0 &&
-	    dqp->q_core.d_id &&
-	    ((XFS_IS_UQUOTA_ENFORCED(dqp->q_mount) && XFS_QM_ISUDQ(dqp)) ||
-	     (XFS_IS_GQUOTA_ENFORCED(dqp->q_mount) && XFS_QM_ISGDQ(dqp)) ||
-	     (XFS_IS_PQUOTA_ENFORCED(dqp->q_mount) && XFS_QM_ISPDQ(dqp)))) {
-		if (nblks > 0) {
+	if ((flags & XFS_QMOPT_FORCE_RES) == 0 && dqp->q_id &&
+	    xfs_dquot_is_enforced(dqp)) {
+		int		quota_nl;
+		bool		fatal;
+
+		/*
+		 * dquot is locked already. See if we'd go over the hardlimit
+		 * or exceed the timelimit if we'd reserve resources.
+		 */
+		quota_nl = xfs_dqresv_check(blkres, qlim, nblks, &fatal);
+		if (quota_nl != QUOTA_NL_NOWARN) {
 			/*
-			 * dquot is locked already. See if we'd go over the
-			 * hardlimit or exceed the timelimit if we allocate
-			 * nblks.
+			 * Quota block warning codes are 3 more than the inode
+			 * codes, which we check above.
 			 */
-			total_count = *resbcountp + nblks;
-			if (hardlimit && total_count > hardlimit) {
-				xfs_quota_warn(mp, dqp, QUOTA_NL_BHARDWARN);
+			xfs_quota_warn(mp, dqp, quota_nl + 3);
+			if (fatal)
 				goto error_return;
-			}
-			if (softlimit && total_count > softlimit) {
-				if ((timer != 0 &&
-				     ktime_get_real_seconds() > timer) ||
-				    (warns != 0 && warns >= warnlimit)) {
-					xfs_quota_warn(mp, dqp,
-						       QUOTA_NL_BSOFTLONGWARN);
-					goto error_return;
-				}
-
-				xfs_quota_warn(mp, dqp, QUOTA_NL_BSOFTWARN);
-			}
 		}
-		if (ninos > 0) {
-			total_count = be64_to_cpu(dqp->q_core.d_icount) + ninos;
-			timer = be32_to_cpu(dqp->q_core.d_itimer);
-			warns = be16_to_cpu(dqp->q_core.d_iwarns);
-			warnlimit = dqp->q_mount->m_quotainfo->qi_iwarnlimit;
-			hardlimit = be64_to_cpu(dqp->q_core.d_ino_hardlimit);
-			if (!hardlimit)
-				hardlimit = defq->ihardlimit;
-			softlimit = be64_to_cpu(dqp->q_core.d_ino_softlimit);
-			if (!softlimit)
-				softlimit = defq->isoftlimit;
-
-			if (hardlimit && total_count > hardlimit) {
-				xfs_quota_warn(mp, dqp, QUOTA_NL_IHARDWARN);
+
+		quota_nl = xfs_dqresv_check(&dqp->q_ino, &defq->ino, ninos,
+				&fatal);
+		if (quota_nl != QUOTA_NL_NOWARN) {
+			xfs_quota_warn(mp, dqp, quota_nl);
+			if (fatal)
 				goto error_return;
-			}
-			if (softlimit && total_count > softlimit) {
-				if  ((timer != 0 &&
-				      ktime_get_real_seconds() > timer) ||
-				     (warns != 0 && warns >= warnlimit)) {
-					xfs_quota_warn(mp, dqp,
-						       QUOTA_NL_ISOFTLONGWARN);
-					goto error_return;
-				}
-				xfs_quota_warn(mp, dqp, QUOTA_NL_ISOFTWARN);
-			}
 		}
 	}
 
 	/*
 	 * Change the reservation, but not the actual usage.
-	 * Note that q_res_bcount = q_core.d_bcount + resv
+	 * Note that q_blk.reserved = q_blk.count + resv
 	 */
-	(*resbcountp) += (xfs_qcnt_t)nblks;
-	if (ninos != 0)
-		dqp->q_res_icount += (xfs_qcnt_t)ninos;
+	blkres->reserved += (xfs_qcnt_t)nblks;
+	dqp->q_ino.reserved += (xfs_qcnt_t)ninos;
 
 	/*
 	 * note the reservation amt in the trans struct too,
@@ -702,16 +716,16 @@ xfs_trans_dqresv(
 					    XFS_TRANS_DQ_RES_INOS,
 					    ninos);
 	}
-	ASSERT(dqp->q_res_bcount >= be64_to_cpu(dqp->q_core.d_bcount));
-	ASSERT(dqp->q_res_rtbcount >= be64_to_cpu(dqp->q_core.d_rtbcount));
-	ASSERT(dqp->q_res_icount >= be64_to_cpu(dqp->q_core.d_icount));
+	ASSERT(dqp->q_blk.reserved >= dqp->q_blk.count);
+	ASSERT(dqp->q_rtb.reserved >= dqp->q_rtb.count);
+	ASSERT(dqp->q_ino.reserved >= dqp->q_ino.count);
 
 	xfs_dqunlock(dqp);
 	return 0;
 
 error_return:
 	xfs_dqunlock(dqp);
-	if (flags & XFS_QMOPT_ENOSPC)
+	if (xfs_dquot_type(dqp) == XFS_DQTYPE_PROJ)
 		return -ENOSPC;
 	return -EDQUOT;
 }
@@ -751,8 +765,7 @@ xfs_trans_reserve_quota_bydquots(
 	ASSERT(flags & XFS_QMOPT_RESBLK_MASK);
 
 	if (udqp) {
-		error = xfs_trans_dqresv(tp, mp, udqp, nblks, ninos,
-					(flags & ~XFS_QMOPT_ENOSPC));
+		error = xfs_trans_dqresv(tp, mp, udqp, nblks, ninos, flags);
 		if (error)
 			return error;
 	}
@@ -803,16 +816,12 @@ xfs_trans_reserve_quota_nblks(
 
 	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
 		return 0;
-	if (XFS_IS_PQUOTA_ON(mp))
-		flags |= XFS_QMOPT_ENOSPC;
 
 	ASSERT(!xfs_is_quota_inode(&mp->m_sb, ip->i_ino));
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL));
-	ASSERT((flags & ~(XFS_QMOPT_FORCE_RES | XFS_QMOPT_ENOSPC)) ==
-				XFS_TRANS_DQ_RES_RTBLKS ||
-	       (flags & ~(XFS_QMOPT_FORCE_RES | XFS_QMOPT_ENOSPC)) ==
-				XFS_TRANS_DQ_RES_BLKS);
+	ASSERT((flags & ~(XFS_QMOPT_FORCE_RES)) == XFS_TRANS_DQ_RES_RTBLKS ||
+	       (flags & ~(XFS_QMOPT_FORCE_RES)) == XFS_TRANS_DQ_RES_BLKS);
 
 	/*
 	 * Reserve nblks against these dquots, with trans as the mediator.
@@ -865,7 +874,8 @@ STATIC void
 xfs_trans_alloc_dqinfo(
 	xfs_trans_t	*tp)
 {
-	tp->t_dqinfo = kmem_zone_zalloc(xfs_qm_dqtrxzone, 0);
+	tp->t_dqinfo = kmem_cache_zalloc(xfs_qm_dqtrxzone,
+					 GFP_KERNEL | __GFP_NOFAIL);
 }
 
 void
diff --git a/fs/xfs/xfs_trans_priv.h b/fs/xfs/xfs_trans_priv.h
index 2e073c1c4614..3004aeac9110 100644
--- a/fs/xfs/xfs_trans_priv.h
+++ b/fs/xfs/xfs_trans_priv.h
@@ -91,24 +91,13 @@ xfs_trans_ail_update(
 	xfs_trans_ail_update_bulk(ailp, NULL, &lip, 1, lsn);
 }
 
-bool xfs_ail_delete_one(struct xfs_ail *ailp, struct xfs_log_item *lip);
-void xfs_trans_ail_delete(struct xfs_ail *ailp, struct xfs_log_item *lip,
-		int shutdown_type) __releases(ailp->ail_lock);
+void xfs_trans_ail_insert(struct xfs_ail *ailp, struct xfs_log_item *lip,
+		xfs_lsn_t lsn);
 
-static inline void
-xfs_trans_ail_remove(
-	struct xfs_log_item	*lip,
-	int			shutdown_type)
-{
-	struct xfs_ail		*ailp = lip->li_ailp;
-
-	spin_lock(&ailp->ail_lock);
-	/* xfs_trans_ail_delete() drops the AIL lock */
-	if (test_bit(XFS_LI_IN_AIL, &lip->li_flags))
-		xfs_trans_ail_delete(ailp, lip, shutdown_type);
-	else
-		spin_unlock(&ailp->ail_lock);
-}
+xfs_lsn_t xfs_ail_delete_one(struct xfs_ail *ailp, struct xfs_log_item *lip);
+void xfs_ail_update_finish(struct xfs_ail *ailp, xfs_lsn_t old_lsn)
+			__releases(ailp->ail_lock);
+void xfs_trans_ail_delete(struct xfs_log_item *lip, int shutdown_type);
 
 void			xfs_ail_push(struct xfs_ail *, xfs_lsn_t);
 void			xfs_ail_push_all(struct xfs_ail *);
diff --git a/fs/xfs/xfs_xattr.c b/fs/xfs/xfs_xattr.c
index b0fedb543f97..bca48b308c02 100644
--- a/fs/xfs/xfs_xattr.c
+++ b/fs/xfs/xfs_xattr.c
@@ -12,53 +12,29 @@
 #include "xfs_inode.h"
 #include "xfs_attr.h"
 #include "xfs_acl.h"
+#include "xfs_da_btree.h"
 
 #include <linux/posix_acl_xattr.h>
-#include <linux/xattr.h>
 
 
 static int
 xfs_xattr_get(const struct xattr_handler *handler, struct dentry *unused,
 		struct inode *inode, const char *name, void *value, size_t size)
 {
-	int xflags = handler->flags;
-	struct xfs_inode *ip = XFS_I(inode);
-	int error, asize = size;
-	size_t namelen = strlen(name);
-
-	/* Convert Linux syscall to XFS internal ATTR flags */
-	if (!size) {
-		xflags |= ATTR_KERNOVAL;
-		value = NULL;
-	}
+	struct xfs_da_args	args = {
+		.dp		= XFS_I(inode),
+		.attr_filter	= handler->flags,
+		.name		= name,
+		.namelen	= strlen(name),
+		.value		= value,
+		.valuelen	= size,
+	};
+	int			error;
 
-	error = xfs_attr_get(ip, name, namelen, (unsigned char **)&value,
-			     &asize, xflags);
+	error = xfs_attr_get(&args);
 	if (error)
 		return error;
-	return asize;
-}
-
-void
-xfs_forget_acl(
-	struct inode		*inode,
-	const char		*name,
-	int			xflags)
-{
-	/*
-	 * Invalidate any cached ACLs if the user has bypassed the ACL
-	 * interface. We don't validate the content whatsoever so it is caller
-	 * responsibility to provide data in valid format and ensure i_mode is
-	 * consistent.
-	 */
-	if (xflags & ATTR_ROOT) {
-#ifdef CONFIG_XFS_POSIX_ACL
-		if (!strcmp(name, SGI_ACL_FILE))
-			forget_cached_acl(inode, ACL_TYPE_ACCESS);
-		else if (!strcmp(name, SGI_ACL_DEFAULT))
-			forget_cached_acl(inode, ACL_TYPE_DEFAULT);
-#endif
-	}
+	return args.valuelen;
 }
 
 static int
@@ -66,25 +42,20 @@ xfs_xattr_set(const struct xattr_handler *handler, struct dentry *unused,
 		struct inode *inode, const char *name, const void *value,
 		size_t size, int flags)
 {
-	int			xflags = handler->flags;
-	struct xfs_inode	*ip = XFS_I(inode);
+	struct xfs_da_args	args = {
+		.dp		= XFS_I(inode),
+		.attr_filter	= handler->flags,
+		.attr_flags	= flags,
+		.name		= name,
+		.namelen	= strlen(name),
+		.value		= (void *)value,
+		.valuelen	= size,
+	};
 	int			error;
-	size_t			namelen = strlen(name);
-
-	/* Convert Linux syscall to XFS internal ATTR flags */
-	if (flags & XATTR_CREATE)
-		xflags |= ATTR_CREATE;
-	if (flags & XATTR_REPLACE)
-		xflags |= ATTR_REPLACE;
-
-	if (value)
-		error = xfs_attr_set(ip, name, namelen, (void *)value, size,
-				xflags);
-	else
-		error = xfs_attr_remove(ip, name, namelen, xflags);
-	if (!error)
-		xfs_forget_acl(inode, name, xflags);
 
+	error = xfs_attr_set(&args);
+	if (!error && (handler->flags & XFS_ATTR_ROOT))
+		xfs_forget_acl(inode, name);
 	return error;
 }
 
@@ -97,14 +68,14 @@ static const struct xattr_handler xfs_xattr_user_handler = {
 
 static const struct xattr_handler xfs_xattr_trusted_handler = {
 	.prefix	= XATTR_TRUSTED_PREFIX,
-	.flags	= ATTR_ROOT,
+	.flags	= XFS_ATTR_ROOT,
 	.get	= xfs_xattr_get,
 	.set	= xfs_xattr_set,
 };
 
 static const struct xattr_handler xfs_xattr_security_handler = {
 	.prefix	= XATTR_SECURITY_PREFIX,
-	.flags	= ATTR_SECURE,
+	.flags	= XFS_ATTR_SECURE,
 	.get	= xfs_xattr_get,
 	.set	= xfs_xattr_set,
 };
@@ -134,7 +105,7 @@ __xfs_xattr_put_listent(
 	if (context->count < 0 || context->seen_enough)
 		return;
 
-	if (!context->alist)
+	if (!context->buffer)
 		goto compute_size;
 
 	arraytop = context->count + prefix_len + namelen + 1;
@@ -143,7 +114,7 @@ __xfs_xattr_put_listent(
 		context->seen_enough = 1;
 		return;
 	}
-	offset = (char *)context->alist + context->count;
+	offset = context->buffer + context->count;
 	strncpy(offset, prefix, prefix_len);
 	offset += prefix_len;
 	strncpy(offset, (char *)name, namelen);			/* real name */
@@ -218,7 +189,6 @@ xfs_vn_listxattr(
 	size_t		size)
 {
 	struct xfs_attr_list_context context;
-	struct attrlist_cursor_kern cursor = { 0 };
 	struct inode	*inode = d_inode(dentry);
 	int		error;
 
@@ -227,14 +197,13 @@ xfs_vn_listxattr(
 	 */
 	memset(&context, 0, sizeof(context));
 	context.dp = XFS_I(inode);
-	context.cursor = &cursor;
 	context.resynch = 1;
-	context.alist = size ? data : NULL;
+	context.buffer = size ? data : NULL;
 	context.bufsize = size;
 	context.firstu = context.bufsize;
 	context.put_listent = xfs_xattr_put_listent;
 
-	error = xfs_attr_list_int(&context);
+	error = xfs_attr_list(&context);
 	if (error)
 		return error;
 	if (context.count < 0)