24 files changed, 9500 insertions, 0 deletions

diff --git a/fs/xfs/scrub/agheader.c b/fs/xfs/scrub/agheader.c
new file mode 100644
index 000000000000..018aabbd9394
--- /dev/null
+++ b/fs/xfs/scrub/agheader.c
@@ -0,0 +1,967 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_alloc.h"
+#include "xfs_ialloc.h"
+#include "xfs_rmap.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+
+/*
+ * Walk all the blocks in the AGFL.  The fn function can return any negative
+ * error code or XFS_BTREE_QUERY_RANGE_ABORT.
+ */
+int
+xfs_scrub_walk_agfl(
+	struct xfs_scrub_context	*sc,
+	int				(*fn)(struct xfs_scrub_context *,
+					      xfs_agblock_t bno, void *),
+	void				*priv)
+{
+	struct xfs_agf			*agf;
+	__be32				*agfl_bno;
+	struct xfs_mount		*mp = sc->mp;
+	unsigned int			flfirst;
+	unsigned int			fllast;
+	int				i;
+	int				error;
+
+	agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, sc->sa.agfl_bp);
+	flfirst = be32_to_cpu(agf->agf_flfirst);
+	fllast = be32_to_cpu(agf->agf_fllast);
+
+	/* Nothing to walk in an empty AGFL. */
+	if (agf->agf_flcount == cpu_to_be32(0))
+		return 0;
+
+	/* first to last is a consecutive list. */
+	if (fllast >= flfirst) {
+		for (i = flfirst; i <= fllast; i++) {
+			error = fn(sc, be32_to_cpu(agfl_bno[i]), priv);
+			if (error)
+				return error;
+			if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+				return error;
+		}
+
+		return 0;
+	}
+
+	/* first to the end */
+	for (i = flfirst; i < xfs_agfl_size(mp); i++) {
+		error = fn(sc, be32_to_cpu(agfl_bno[i]), priv);
+		if (error)
+			return error;
+		if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+			return error;
+	}
+
+	/* the start to last. */
+	for (i = 0; i <= fllast; i++) {
+		error = fn(sc, be32_to_cpu(agfl_bno[i]), priv);
+		if (error)
+			return error;
+		if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+			return error;
+	}
+
+	return 0;
+}
+
+/* Superblock */
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_superblock_xref(
+	struct xfs_scrub_context	*sc,
+	struct xfs_buf			*bp)
+{
+	struct xfs_owner_info		oinfo;
+	struct xfs_mount		*mp = sc->mp;
+	xfs_agnumber_t			agno = sc->sm->sm_agno;
+	xfs_agblock_t			agbno;
+	int				error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	agbno = XFS_SB_BLOCK(mp);
+
+	error = xfs_scrub_ag_init(sc, agno, &sc->sa);
+	if (!xfs_scrub_xref_process_error(sc, agno, agbno, &error))
+		return;
+
+	xfs_scrub_xref_is_used_space(sc, agbno, 1);
+	xfs_scrub_xref_is_not_inode_chunk(sc, agbno, 1);
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_FS);
+	xfs_scrub_xref_is_owned_by(sc, agbno, 1, &oinfo);
+	xfs_scrub_xref_is_not_shared(sc, agbno, 1);
+
+	/* scrub teardown will take care of sc->sa for us */
+}
+
+/*
+ * Scrub the filesystem superblock.
+ *
+ * Note: We do /not/ attempt to check AG 0's superblock.  Mount is
+ * responsible for validating all the geometry information in sb 0, so
+ * if the filesystem is capable of initiating online scrub, then clearly
+ * sb 0 is ok and we can use its information to check everything else.
+ */
+int
+xfs_scrub_superblock(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_buf			*bp;
+	struct xfs_dsb			*sb;
+	xfs_agnumber_t			agno;
+	uint32_t			v2_ok;
+	__be32				features_mask;
+	int				error;
+	__be16				vernum_mask;
+
+	agno = sc->sm->sm_agno;
+	if (agno == 0)
+		return 0;
+
+	error = xfs_trans_read_buf(mp, sc->tp, mp->m_ddev_targp,
+		  XFS_AGB_TO_DADDR(mp, agno, XFS_SB_BLOCK(mp)),
+		  XFS_FSS_TO_BB(mp, 1), 0, &bp, &xfs_sb_buf_ops);
+	/*
+	 * The superblock verifier can return several different error codes
+	 * if it thinks the superblock doesn't look right.  For a mount these
+	 * would all get bounced back to userspace, but if we're here then the
+	 * fs mounted successfully, which means that this secondary superblock
+	 * is simply incorrect.  Treat all these codes the same way we treat
+	 * any corruption.
+	 */
+	switch (error) {
+	case -EINVAL:	/* also -EWRONGFS */
+	case -ENOSYS:
+	case -EFBIG:
+		error = -EFSCORRUPTED;
+	default:
+		break;
+	}
+	if (!xfs_scrub_process_error(sc, agno, XFS_SB_BLOCK(mp), &error))
+		return error;
+
+	sb = XFS_BUF_TO_SBP(bp);
+
+	/*
+	 * Verify the geometries match.  Fields that are permanently
+	 * set by mkfs are checked; fields that can be updated later
+	 * (and are not propagated to backup superblocks) are preen
+	 * checked.
+	 */
+	if (sb->sb_blocksize != cpu_to_be32(mp->m_sb.sb_blocksize))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_dblocks != cpu_to_be64(mp->m_sb.sb_dblocks))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_rblocks != cpu_to_be64(mp->m_sb.sb_rblocks))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_rextents != cpu_to_be64(mp->m_sb.sb_rextents))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (!uuid_equal(&sb->sb_uuid, &mp->m_sb.sb_uuid))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	if (sb->sb_logstart != cpu_to_be64(mp->m_sb.sb_logstart))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_rootino != cpu_to_be64(mp->m_sb.sb_rootino))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	if (sb->sb_rbmino != cpu_to_be64(mp->m_sb.sb_rbmino))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	if (sb->sb_rsumino != cpu_to_be64(mp->m_sb.sb_rsumino))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	if (sb->sb_rextsize != cpu_to_be32(mp->m_sb.sb_rextsize))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_agblocks != cpu_to_be32(mp->m_sb.sb_agblocks))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_agcount != cpu_to_be32(mp->m_sb.sb_agcount))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_rbmblocks != cpu_to_be32(mp->m_sb.sb_rbmblocks))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_logblocks != cpu_to_be32(mp->m_sb.sb_logblocks))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	/* Check sb_versionnum bits that are set at mkfs time. */
+	vernum_mask = cpu_to_be16(~XFS_SB_VERSION_OKBITS |
+				  XFS_SB_VERSION_NUMBITS |
+				  XFS_SB_VERSION_ALIGNBIT |
+				  XFS_SB_VERSION_DALIGNBIT |
+				  XFS_SB_VERSION_SHAREDBIT |
+				  XFS_SB_VERSION_LOGV2BIT |
+				  XFS_SB_VERSION_SECTORBIT |
+				  XFS_SB_VERSION_EXTFLGBIT |
+				  XFS_SB_VERSION_DIRV2BIT);
+	if ((sb->sb_versionnum & vernum_mask) !=
+	    (cpu_to_be16(mp->m_sb.sb_versionnum) & vernum_mask))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	/* Check sb_versionnum bits that can be set after mkfs time. */
+	vernum_mask = cpu_to_be16(XFS_SB_VERSION_ATTRBIT |
+				  XFS_SB_VERSION_NLINKBIT |
+				  XFS_SB_VERSION_QUOTABIT);
+	if ((sb->sb_versionnum & vernum_mask) !=
+	    (cpu_to_be16(mp->m_sb.sb_versionnum) & vernum_mask))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	if (sb->sb_sectsize != cpu_to_be16(mp->m_sb.sb_sectsize))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_inodesize != cpu_to_be16(mp->m_sb.sb_inodesize))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_inopblock != cpu_to_be16(mp->m_sb.sb_inopblock))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (memcmp(sb->sb_fname, mp->m_sb.sb_fname, sizeof(sb->sb_fname)))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	if (sb->sb_blocklog != mp->m_sb.sb_blocklog)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_sectlog != mp->m_sb.sb_sectlog)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_inodelog != mp->m_sb.sb_inodelog)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_inopblog != mp->m_sb.sb_inopblog)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_agblklog != mp->m_sb.sb_agblklog)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_rextslog != mp->m_sb.sb_rextslog)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_imax_pct != mp->m_sb.sb_imax_pct)
+		xfs_scrub_block_set_preen(sc, bp);
+
+	/*
+	 * Skip the summary counters since we track them in memory anyway.
+	 * sb_icount, sb_ifree, sb_fdblocks, sb_frexents
+	 */
+
+	if (sb->sb_uquotino != cpu_to_be64(mp->m_sb.sb_uquotino))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	if (sb->sb_gquotino != cpu_to_be64(mp->m_sb.sb_gquotino))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	/*
+	 * Skip the quota flags since repair will force quotacheck.
+	 * sb_qflags
+	 */
+
+	if (sb->sb_flags != mp->m_sb.sb_flags)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_shared_vn != mp->m_sb.sb_shared_vn)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_inoalignmt != cpu_to_be32(mp->m_sb.sb_inoalignmt))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_unit != cpu_to_be32(mp->m_sb.sb_unit))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	if (sb->sb_width != cpu_to_be32(mp->m_sb.sb_width))
+		xfs_scrub_block_set_preen(sc, bp);
+
+	if (sb->sb_dirblklog != mp->m_sb.sb_dirblklog)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_logsectlog != mp->m_sb.sb_logsectlog)
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_logsectsize != cpu_to_be16(mp->m_sb.sb_logsectsize))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (sb->sb_logsunit != cpu_to_be32(mp->m_sb.sb_logsunit))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	/* Do we see any invalid bits in sb_features2? */
+	if (!xfs_sb_version_hasmorebits(&mp->m_sb)) {
+		if (sb->sb_features2 != 0)
+			xfs_scrub_block_set_corrupt(sc, bp);
+	} else {
+		v2_ok = XFS_SB_VERSION2_OKBITS;
+		if (XFS_SB_VERSION_NUM(&mp->m_sb) >= XFS_SB_VERSION_5)
+			v2_ok |= XFS_SB_VERSION2_CRCBIT;
+
+		if (!!(sb->sb_features2 & cpu_to_be32(~v2_ok)))
+			xfs_scrub_block_set_corrupt(sc, bp);
+
+		if (sb->sb_features2 != sb->sb_bad_features2)
+			xfs_scrub_block_set_preen(sc, bp);
+	}
+
+	/* Check sb_features2 flags that are set at mkfs time. */
+	features_mask = cpu_to_be32(XFS_SB_VERSION2_LAZYSBCOUNTBIT |
+				    XFS_SB_VERSION2_PROJID32BIT |
+				    XFS_SB_VERSION2_CRCBIT |
+				    XFS_SB_VERSION2_FTYPE);
+	if ((sb->sb_features2 & features_mask) !=
+	    (cpu_to_be32(mp->m_sb.sb_features2) & features_mask))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	/* Check sb_features2 flags that can be set after mkfs time. */
+	features_mask = cpu_to_be32(XFS_SB_VERSION2_ATTR2BIT);
+	if ((sb->sb_features2 & features_mask) !=
+	    (cpu_to_be32(mp->m_sb.sb_features2) & features_mask))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	if (!xfs_sb_version_hascrc(&mp->m_sb)) {
+		/* all v5 fields must be zero */
+		if (memchr_inv(&sb->sb_features_compat, 0,
+				sizeof(struct xfs_dsb) -
+				offsetof(struct xfs_dsb, sb_features_compat)))
+			xfs_scrub_block_set_corrupt(sc, bp);
+	} else {
+		/* Check compat flags; all are set at mkfs time. */
+		features_mask = cpu_to_be32(XFS_SB_FEAT_COMPAT_UNKNOWN);
+		if ((sb->sb_features_compat & features_mask) !=
+		    (cpu_to_be32(mp->m_sb.sb_features_compat) & features_mask))
+			xfs_scrub_block_set_corrupt(sc, bp);
+
+		/* Check ro compat flags; all are set at mkfs time. */
+		features_mask = cpu_to_be32(XFS_SB_FEAT_RO_COMPAT_UNKNOWN |
+					    XFS_SB_FEAT_RO_COMPAT_FINOBT |
+					    XFS_SB_FEAT_RO_COMPAT_RMAPBT |
+					    XFS_SB_FEAT_RO_COMPAT_REFLINK);
+		if ((sb->sb_features_ro_compat & features_mask) !=
+		    (cpu_to_be32(mp->m_sb.sb_features_ro_compat) &
+		     features_mask))
+			xfs_scrub_block_set_corrupt(sc, bp);
+
+		/* Check incompat flags; all are set at mkfs time. */
+		features_mask = cpu_to_be32(XFS_SB_FEAT_INCOMPAT_UNKNOWN |
+					    XFS_SB_FEAT_INCOMPAT_FTYPE |
+					    XFS_SB_FEAT_INCOMPAT_SPINODES |
+					    XFS_SB_FEAT_INCOMPAT_META_UUID);
+		if ((sb->sb_features_incompat & features_mask) !=
+		    (cpu_to_be32(mp->m_sb.sb_features_incompat) &
+		     features_mask))
+			xfs_scrub_block_set_corrupt(sc, bp);
+
+		/* Check log incompat flags; all are set at mkfs time. */
+		features_mask = cpu_to_be32(XFS_SB_FEAT_INCOMPAT_LOG_UNKNOWN);
+		if ((sb->sb_features_log_incompat & features_mask) !=
+		    (cpu_to_be32(mp->m_sb.sb_features_log_incompat) &
+		     features_mask))
+			xfs_scrub_block_set_corrupt(sc, bp);
+
+		/* Don't care about sb_crc */
+
+		if (sb->sb_spino_align != cpu_to_be32(mp->m_sb.sb_spino_align))
+			xfs_scrub_block_set_corrupt(sc, bp);
+
+		if (sb->sb_pquotino != cpu_to_be64(mp->m_sb.sb_pquotino))
+			xfs_scrub_block_set_preen(sc, bp);
+
+		/* Don't care about sb_lsn */
+	}
+
+	if (xfs_sb_version_hasmetauuid(&mp->m_sb)) {
+		/* The metadata UUID must be the same for all supers */
+		if (!uuid_equal(&sb->sb_meta_uuid, &mp->m_sb.sb_meta_uuid))
+			xfs_scrub_block_set_corrupt(sc, bp);
+	}
+
+	/* Everything else must be zero. */
+	if (memchr_inv(sb + 1, 0,
+			BBTOB(bp->b_length) - sizeof(struct xfs_dsb)))
+		xfs_scrub_block_set_corrupt(sc, bp);
+
+	xfs_scrub_superblock_xref(sc, bp);
+
+	return error;
+}
+
+/* AGF */
+
+/* Tally freespace record lengths. */
+STATIC int
+xfs_scrub_agf_record_bno_lengths(
+	struct xfs_btree_cur		*cur,
+	struct xfs_alloc_rec_incore	*rec,
+	void				*priv)
+{
+	xfs_extlen_t			*blocks = priv;
+
+	(*blocks) += rec->ar_blockcount;
+	return 0;
+}
+
+/* Check agf_freeblks */
+static inline void
+xfs_scrub_agf_xref_freeblks(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_agf			*agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	xfs_extlen_t			blocks = 0;
+	int				error;
+
+	if (!sc->sa.bno_cur)
+		return;
+
+	error = xfs_alloc_query_all(sc->sa.bno_cur,
+			xfs_scrub_agf_record_bno_lengths, &blocks);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.bno_cur))
+		return;
+	if (blocks != be32_to_cpu(agf->agf_freeblks))
+		xfs_scrub_block_xref_set_corrupt(sc, sc->sa.agf_bp);
+}
+
+/* Cross reference the AGF with the cntbt (freespace by length btree) */
+static inline void
+xfs_scrub_agf_xref_cntbt(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_agf			*agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	xfs_agblock_t			agbno;
+	xfs_extlen_t			blocks;
+	int				have;
+	int				error;
+
+	if (!sc->sa.cnt_cur)
+		return;
+
+	/* Any freespace at all? */
+	error = xfs_alloc_lookup_le(sc->sa.cnt_cur, 0, -1U, &have);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.cnt_cur))
+		return;
+	if (!have) {
+		if (agf->agf_freeblks != be32_to_cpu(0))
+			xfs_scrub_block_xref_set_corrupt(sc, sc->sa.agf_bp);
+		return;
+	}
+
+	/* Check agf_longest */
+	error = xfs_alloc_get_rec(sc->sa.cnt_cur, &agbno, &blocks, &have);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.cnt_cur))
+		return;
+	if (!have || blocks != be32_to_cpu(agf->agf_longest))
+		xfs_scrub_block_xref_set_corrupt(sc, sc->sa.agf_bp);
+}
+
+/* Check the btree block counts in the AGF against the btrees. */
+STATIC void
+xfs_scrub_agf_xref_btreeblks(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_agf			*agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	struct xfs_mount		*mp = sc->mp;
+	xfs_agblock_t			blocks;
+	xfs_agblock_t			btreeblks;
+	int				error;
+
+	/* Check agf_rmap_blocks; set up for agf_btreeblks check */
+	if (sc->sa.rmap_cur) {
+		error = xfs_btree_count_blocks(sc->sa.rmap_cur, &blocks);
+		if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.rmap_cur))
+			return;
+		btreeblks = blocks - 1;
+		if (blocks != be32_to_cpu(agf->agf_rmap_blocks))
+			xfs_scrub_block_xref_set_corrupt(sc, sc->sa.agf_bp);
+	} else {
+		btreeblks = 0;
+	}
+
+	/*
+	 * No rmap cursor; we can't xref if we have the rmapbt feature.
+	 * We also can't do it if we're missing the free space btree cursors.
+	 */
+	if ((xfs_sb_version_hasrmapbt(&mp->m_sb) && !sc->sa.rmap_cur) ||
+	    !sc->sa.bno_cur || !sc->sa.cnt_cur)
+		return;
+
+	/* Check agf_btreeblks */
+	error = xfs_btree_count_blocks(sc->sa.bno_cur, &blocks);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.bno_cur))
+		return;
+	btreeblks += blocks - 1;
+
+	error = xfs_btree_count_blocks(sc->sa.cnt_cur, &blocks);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.cnt_cur))
+		return;
+	btreeblks += blocks - 1;
+
+	if (btreeblks != be32_to_cpu(agf->agf_btreeblks))
+		xfs_scrub_block_xref_set_corrupt(sc, sc->sa.agf_bp);
+}
+
+/* Check agf_refcount_blocks against tree size */
+static inline void
+xfs_scrub_agf_xref_refcblks(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_agf			*agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	xfs_agblock_t			blocks;
+	int				error;
+
+	if (!sc->sa.refc_cur)
+		return;
+
+	error = xfs_btree_count_blocks(sc->sa.refc_cur, &blocks);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.refc_cur))
+		return;
+	if (blocks != be32_to_cpu(agf->agf_refcount_blocks))
+		xfs_scrub_block_xref_set_corrupt(sc, sc->sa.agf_bp);
+}
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_agf_xref(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_owner_info		oinfo;
+	struct xfs_mount		*mp = sc->mp;
+	xfs_agblock_t			agbno;
+	int				error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	agbno = XFS_AGF_BLOCK(mp);
+
+	error = xfs_scrub_ag_btcur_init(sc, &sc->sa);
+	if (error)
+		return;
+
+	xfs_scrub_xref_is_used_space(sc, agbno, 1);
+	xfs_scrub_agf_xref_freeblks(sc);
+	xfs_scrub_agf_xref_cntbt(sc);
+	xfs_scrub_xref_is_not_inode_chunk(sc, agbno, 1);
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_FS);
+	xfs_scrub_xref_is_owned_by(sc, agbno, 1, &oinfo);
+	xfs_scrub_agf_xref_btreeblks(sc);
+	xfs_scrub_xref_is_not_shared(sc, agbno, 1);
+	xfs_scrub_agf_xref_refcblks(sc);
+
+	/* scrub teardown will take care of sc->sa for us */
+}
+
+/* Scrub the AGF. */
+int
+xfs_scrub_agf(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_agf			*agf;
+	xfs_agnumber_t			agno;
+	xfs_agblock_t			agbno;
+	xfs_agblock_t			eoag;
+	xfs_agblock_t			agfl_first;
+	xfs_agblock_t			agfl_last;
+	xfs_agblock_t			agfl_count;
+	xfs_agblock_t			fl_count;
+	int				level;
+	int				error = 0;
+
+	agno = sc->sa.agno = sc->sm->sm_agno;
+	error = xfs_scrub_ag_read_headers(sc, agno, &sc->sa.agi_bp,
+			&sc->sa.agf_bp, &sc->sa.agfl_bp);
+	if (!xfs_scrub_process_error(sc, agno, XFS_AGF_BLOCK(sc->mp), &error))
+		goto out;
+	xfs_scrub_buffer_recheck(sc, sc->sa.agf_bp);
+
+	agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+
+	/* Check the AG length */
+	eoag = be32_to_cpu(agf->agf_length);
+	if (eoag != xfs_ag_block_count(mp, agno))
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+
+	/* Check the AGF btree roots and levels */
+	agbno = be32_to_cpu(agf->agf_roots[XFS_BTNUM_BNO]);
+	if (!xfs_verify_agbno(mp, agno, agbno))
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+
+	agbno = be32_to_cpu(agf->agf_roots[XFS_BTNUM_CNT]);
+	if (!xfs_verify_agbno(mp, agno, agbno))
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+
+	level = be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]);
+	if (level <= 0 || level > XFS_BTREE_MAXLEVELS)
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+
+	level = be32_to_cpu(agf->agf_levels[XFS_BTNUM_CNT]);
+	if (level <= 0 || level > XFS_BTREE_MAXLEVELS)
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+
+	if (xfs_sb_version_hasrmapbt(&mp->m_sb)) {
+		agbno = be32_to_cpu(agf->agf_roots[XFS_BTNUM_RMAP]);
+		if (!xfs_verify_agbno(mp, agno, agbno))
+			xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+
+		level = be32_to_cpu(agf->agf_levels[XFS_BTNUM_RMAP]);
+		if (level <= 0 || level > XFS_BTREE_MAXLEVELS)
+			xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+	}
+
+	if (xfs_sb_version_hasreflink(&mp->m_sb)) {
+		agbno = be32_to_cpu(agf->agf_refcount_root);
+		if (!xfs_verify_agbno(mp, agno, agbno))
+			xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+
+		level = be32_to_cpu(agf->agf_refcount_level);
+		if (level <= 0 || level > XFS_BTREE_MAXLEVELS)
+			xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+	}
+
+	/* Check the AGFL counters */
+	agfl_first = be32_to_cpu(agf->agf_flfirst);
+	agfl_last = be32_to_cpu(agf->agf_fllast);
+	agfl_count = be32_to_cpu(agf->agf_flcount);
+	if (agfl_last > agfl_first)
+		fl_count = agfl_last - agfl_first + 1;
+	else
+		fl_count = xfs_agfl_size(mp) - agfl_first + agfl_last + 1;
+	if (agfl_count != 0 && fl_count != agfl_count)
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+
+	xfs_scrub_agf_xref(sc);
+out:
+	return error;
+}
+
+/* AGFL */
+
+struct xfs_scrub_agfl_info {
+	struct xfs_owner_info		oinfo;
+	unsigned int			sz_entries;
+	unsigned int			nr_entries;
+	xfs_agblock_t			*entries;
+};
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_agfl_block_xref(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	struct xfs_owner_info		*oinfo)
+{
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	xfs_scrub_xref_is_used_space(sc, agbno, 1);
+	xfs_scrub_xref_is_not_inode_chunk(sc, agbno, 1);
+	xfs_scrub_xref_is_owned_by(sc, agbno, 1, oinfo);
+	xfs_scrub_xref_is_not_shared(sc, agbno, 1);
+}
+
+/* Scrub an AGFL block. */
+STATIC int
+xfs_scrub_agfl_block(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	void				*priv)
+{
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_scrub_agfl_info	*sai = priv;
+	xfs_agnumber_t			agno = sc->sa.agno;
+
+	if (xfs_verify_agbno(mp, agno, agbno) &&
+	    sai->nr_entries < sai->sz_entries)
+		sai->entries[sai->nr_entries++] = agbno;
+	else
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agfl_bp);
+
+	xfs_scrub_agfl_block_xref(sc, agbno, priv);
+
+	return 0;
+}
+
+static int
+xfs_scrub_agblock_cmp(
+	const void		*pa,
+	const void		*pb)
+{
+	const xfs_agblock_t	*a = pa;
+	const xfs_agblock_t	*b = pb;
+
+	return (int)*a - (int)*b;
+}
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_agfl_xref(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_owner_info		oinfo;
+	struct xfs_mount		*mp = sc->mp;
+	xfs_agblock_t			agbno;
+	int				error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	agbno = XFS_AGFL_BLOCK(mp);
+
+	error = xfs_scrub_ag_btcur_init(sc, &sc->sa);
+	if (error)
+		return;
+
+	xfs_scrub_xref_is_used_space(sc, agbno, 1);
+	xfs_scrub_xref_is_not_inode_chunk(sc, agbno, 1);
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_FS);
+	xfs_scrub_xref_is_owned_by(sc, agbno, 1, &oinfo);
+	xfs_scrub_xref_is_not_shared(sc, agbno, 1);
+
+	/*
+	 * Scrub teardown will take care of sc->sa for us.  Leave sc->sa
+	 * active so that the agfl block xref can use it too.
+	 */
+}
+
+/* Scrub the AGFL. */
+int
+xfs_scrub_agfl(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_scrub_agfl_info	sai;
+	struct xfs_agf			*agf;
+	xfs_agnumber_t			agno;
+	unsigned int			agflcount;
+	unsigned int			i;
+	int				error;
+
+	agno = sc->sa.agno = sc->sm->sm_agno;
+	error = xfs_scrub_ag_read_headers(sc, agno, &sc->sa.agi_bp,
+			&sc->sa.agf_bp, &sc->sa.agfl_bp);
+	if (!xfs_scrub_process_error(sc, agno, XFS_AGFL_BLOCK(sc->mp), &error))
+		goto out;
+	if (!sc->sa.agf_bp)
+		return -EFSCORRUPTED;
+	xfs_scrub_buffer_recheck(sc, sc->sa.agfl_bp);
+
+	xfs_scrub_agfl_xref(sc);
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	/* Allocate buffer to ensure uniqueness of AGFL entries. */
+	agf = XFS_BUF_TO_AGF(sc->sa.agf_bp);
+	agflcount = be32_to_cpu(agf->agf_flcount);
+	if (agflcount > xfs_agfl_size(sc->mp)) {
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+		goto out;
+	}
+	memset(&sai, 0, sizeof(sai));
+	sai.sz_entries = agflcount;
+	sai.entries = kmem_zalloc(sizeof(xfs_agblock_t) * agflcount, KM_NOFS);
+	if (!sai.entries) {
+		error = -ENOMEM;
+		goto out;
+	}
+
+	/* Check the blocks in the AGFL. */
+	xfs_rmap_ag_owner(&sai.oinfo, XFS_RMAP_OWN_AG);
+	error = xfs_scrub_walk_agfl(sc, xfs_scrub_agfl_block, &sai);
+	if (error)
+		goto out_free;
+
+	if (agflcount != sai.nr_entries) {
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+		goto out_free;
+	}
+
+	/* Sort entries, check for duplicates. */
+	sort(sai.entries, sai.nr_entries, sizeof(sai.entries[0]),
+			xfs_scrub_agblock_cmp, NULL);
+	for (i = 1; i < sai.nr_entries; i++) {
+		if (sai.entries[i] == sai.entries[i - 1]) {
+			xfs_scrub_block_set_corrupt(sc, sc->sa.agf_bp);
+			break;
+		}
+	}
+
+out_free:
+	kmem_free(sai.entries);
+out:
+	return error;
+}
+
+/* AGI */
+
+/* Check agi_count/agi_freecount */
+static inline void
+xfs_scrub_agi_xref_icounts(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_agi			*agi = XFS_BUF_TO_AGI(sc->sa.agi_bp);
+	xfs_agino_t			icount;
+	xfs_agino_t			freecount;
+	int				error;
+
+	if (!sc->sa.ino_cur)
+		return;
+
+	error = xfs_ialloc_count_inodes(sc->sa.ino_cur, &icount, &freecount);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.ino_cur))
+		return;
+	if (be32_to_cpu(agi->agi_count) != icount ||
+	    be32_to_cpu(agi->agi_freecount) != freecount)
+		xfs_scrub_block_xref_set_corrupt(sc, sc->sa.agi_bp);
+}
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_agi_xref(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_owner_info		oinfo;
+	struct xfs_mount		*mp = sc->mp;
+	xfs_agblock_t			agbno;
+	int				error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	agbno = XFS_AGI_BLOCK(mp);
+
+	error = xfs_scrub_ag_btcur_init(sc, &sc->sa);
+	if (error)
+		return;
+
+	xfs_scrub_xref_is_used_space(sc, agbno, 1);
+	xfs_scrub_xref_is_not_inode_chunk(sc, agbno, 1);
+	xfs_scrub_agi_xref_icounts(sc);
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_FS);
+	xfs_scrub_xref_is_owned_by(sc, agbno, 1, &oinfo);
+	xfs_scrub_xref_is_not_shared(sc, agbno, 1);
+
+	/* scrub teardown will take care of sc->sa for us */
+}
+
+/* Scrub the AGI. */
+int
+xfs_scrub_agi(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_agi			*agi;
+	xfs_agnumber_t			agno;
+	xfs_agblock_t			agbno;
+	xfs_agblock_t			eoag;
+	xfs_agino_t			agino;
+	xfs_agino_t			first_agino;
+	xfs_agino_t			last_agino;
+	xfs_agino_t			icount;
+	int				i;
+	int				level;
+	int				error = 0;
+
+	agno = sc->sa.agno = sc->sm->sm_agno;
+	error = xfs_scrub_ag_read_headers(sc, agno, &sc->sa.agi_bp,
+			&sc->sa.agf_bp, &sc->sa.agfl_bp);
+	if (!xfs_scrub_process_error(sc, agno, XFS_AGI_BLOCK(sc->mp), &error))
+		goto out;
+	xfs_scrub_buffer_recheck(sc, sc->sa.agi_bp);
+
+	agi = XFS_BUF_TO_AGI(sc->sa.agi_bp);
+
+	/* Check the AG length */
+	eoag = be32_to_cpu(agi->agi_length);
+	if (eoag != xfs_ag_block_count(mp, agno))
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+
+	/* Check btree roots and levels */
+	agbno = be32_to_cpu(agi->agi_root);
+	if (!xfs_verify_agbno(mp, agno, agbno))
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+
+	level = be32_to_cpu(agi->agi_level);
+	if (level <= 0 || level > XFS_BTREE_MAXLEVELS)
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+
+	if (xfs_sb_version_hasfinobt(&mp->m_sb)) {
+		agbno = be32_to_cpu(agi->agi_free_root);
+		if (!xfs_verify_agbno(mp, agno, agbno))
+			xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+
+		level = be32_to_cpu(agi->agi_free_level);
+		if (level <= 0 || level > XFS_BTREE_MAXLEVELS)
+			xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+	}
+
+	/* Check inode counters */
+	xfs_ialloc_agino_range(mp, agno, &first_agino, &last_agino);
+	icount = be32_to_cpu(agi->agi_count);
+	if (icount > last_agino - first_agino + 1 ||
+	    icount < be32_to_cpu(agi->agi_freecount))
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+
+	/* Check inode pointers */
+	agino = be32_to_cpu(agi->agi_newino);
+	if (agino != NULLAGINO && !xfs_verify_agino(mp, agno, agino))
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+
+	agino = be32_to_cpu(agi->agi_dirino);
+	if (agino != NULLAGINO && !xfs_verify_agino(mp, agno, agino))
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+
+	/* Check unlinked inode buckets */
+	for (i = 0; i < XFS_AGI_UNLINKED_BUCKETS; i++) {
+		agino = be32_to_cpu(agi->agi_unlinked[i]);
+		if (agino == NULLAGINO)
+			continue;
+		if (!xfs_verify_agino(mp, agno, agino))
+			xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+	}
+
+	if (agi->agi_pad32 != cpu_to_be32(0))
+		xfs_scrub_block_set_corrupt(sc, sc->sa.agi_bp);
+
+	xfs_scrub_agi_xref(sc);
+out:
+	return error;
+}
diff --git a/fs/xfs/scrub/alloc.c b/fs/xfs/scrub/alloc.c
new file mode 100644
index 000000000000..517c079d3f68
--- /dev/null
+++ b/fs/xfs/scrub/alloc.c
@@ -0,0 +1,183 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_alloc.h"
+#include "xfs_rmap.h"
+#include "xfs_alloc.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+
+/*
+ * Set us up to scrub free space btrees.
+ */
+int
+xfs_scrub_setup_ag_allocbt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	return xfs_scrub_setup_ag_btree(sc, ip, false);
+}
+
+/* Free space btree scrubber. */
+/*
+ * Ensure there's a corresponding cntbt/bnobt record matching this
+ * bnobt/cntbt record, respectively.
+ */
+STATIC void
+xfs_scrub_allocbt_xref_other(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len)
+{
+	struct xfs_btree_cur		**pcur;
+	xfs_agblock_t			fbno;
+	xfs_extlen_t			flen;
+	int				has_otherrec;
+	int				error;
+
+	if (sc->sm->sm_type == XFS_SCRUB_TYPE_BNOBT)
+		pcur = &sc->sa.cnt_cur;
+	else
+		pcur = &sc->sa.bno_cur;
+	if (!*pcur)
+		return;
+
+	error = xfs_alloc_lookup_le(*pcur, agbno, len, &has_otherrec);
+	if (!xfs_scrub_should_check_xref(sc, &error, pcur))
+		return;
+	if (!has_otherrec) {
+		xfs_scrub_btree_xref_set_corrupt(sc, *pcur, 0);
+		return;
+	}
+
+	error = xfs_alloc_get_rec(*pcur, &fbno, &flen, &has_otherrec);
+	if (!xfs_scrub_should_check_xref(sc, &error, pcur))
+		return;
+	if (!has_otherrec) {
+		xfs_scrub_btree_xref_set_corrupt(sc, *pcur, 0);
+		return;
+	}
+
+	if (fbno != agbno || flen != len)
+		xfs_scrub_btree_xref_set_corrupt(sc, *pcur, 0);
+}
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_allocbt_xref(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len)
+{
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	xfs_scrub_allocbt_xref_other(sc, agbno, len);
+	xfs_scrub_xref_is_not_inode_chunk(sc, agbno, len);
+	xfs_scrub_xref_has_no_owner(sc, agbno, len);
+	xfs_scrub_xref_is_not_shared(sc, agbno, len);
+}
+
+/* Scrub a bnobt/cntbt record. */
+STATIC int
+xfs_scrub_allocbt_rec(
+	struct xfs_scrub_btree		*bs,
+	union xfs_btree_rec		*rec)
+{
+	struct xfs_mount		*mp = bs->cur->bc_mp;
+	xfs_agnumber_t			agno = bs->cur->bc_private.a.agno;
+	xfs_agblock_t			bno;
+	xfs_extlen_t			len;
+	int				error = 0;
+
+	bno = be32_to_cpu(rec->alloc.ar_startblock);
+	len = be32_to_cpu(rec->alloc.ar_blockcount);
+
+	if (bno + len <= bno ||
+	    !xfs_verify_agbno(mp, agno, bno) ||
+	    !xfs_verify_agbno(mp, agno, bno + len - 1))
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	xfs_scrub_allocbt_xref(bs->sc, bno, len);
+
+	return error;
+}
+
+/* Scrub the freespace btrees for some AG. */
+STATIC int
+xfs_scrub_allocbt(
+	struct xfs_scrub_context	*sc,
+	xfs_btnum_t			which)
+{
+	struct xfs_owner_info		oinfo;
+	struct xfs_btree_cur		*cur;
+
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_AG);
+	cur = which == XFS_BTNUM_BNO ? sc->sa.bno_cur : sc->sa.cnt_cur;
+	return xfs_scrub_btree(sc, cur, xfs_scrub_allocbt_rec, &oinfo, NULL);
+}
+
+int
+xfs_scrub_bnobt(
+	struct xfs_scrub_context	*sc)
+{
+	return xfs_scrub_allocbt(sc, XFS_BTNUM_BNO);
+}
+
+int
+xfs_scrub_cntbt(
+	struct xfs_scrub_context	*sc)
+{
+	return xfs_scrub_allocbt(sc, XFS_BTNUM_CNT);
+}
+
+/* xref check that the extent is not free */
+void
+xfs_scrub_xref_is_used_space(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len)
+{
+	bool				is_freesp;
+	int				error;
+
+	if (!sc->sa.bno_cur)
+		return;
+
+	error = xfs_alloc_has_record(sc->sa.bno_cur, agbno, len, &is_freesp);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.bno_cur))
+		return;
+	if (is_freesp)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.bno_cur, 0);
+}
diff --git a/fs/xfs/scrub/attr.c b/fs/xfs/scrub/attr.c
new file mode 100644
index 000000000000..127575f0abfb
--- /dev/null
+++ b/fs/xfs/scrub/attr.c
@@ -0,0 +1,471 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_dir2.h"
+#include "xfs_attr.h"
+#include "xfs_attr_leaf.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/dabtree.h"
+#include "scrub/trace.h"
+
+#include <linux/posix_acl_xattr.h>
+#include <linux/xattr.h>
+
+/* Set us up to scrub an inode's extended attributes. */
+int
+xfs_scrub_setup_xattr(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	size_t				sz;
+
+	/*
+	 * Allocate the buffer without the inode lock held.  We need enough
+	 * space to read every xattr value in the file or enough space to
+	 * hold three copies of the xattr free space bitmap.  (Not both at
+	 * the same time.)
+	 */
+	sz = max_t(size_t, XATTR_SIZE_MAX, 3 * sizeof(long) *
+			BITS_TO_LONGS(sc->mp->m_attr_geo->blksize));
+	sc->buf = kmem_zalloc_large(sz, KM_SLEEP);
+	if (!sc->buf)
+		return -ENOMEM;
+
+	return xfs_scrub_setup_inode_contents(sc, ip, 0);
+}
+
+/* Extended Attributes */
+
+struct xfs_scrub_xattr {
+	struct xfs_attr_list_context	context;
+	struct xfs_scrub_context	*sc;
+};
+
+/*
+ * 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)
+ * 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.
+ */
+static void
+xfs_scrub_xattr_listent(
+	struct xfs_attr_list_context	*context,
+	int				flags,
+	unsigned char			*name,
+	int				namelen,
+	int				valuelen)
+{
+	struct xfs_scrub_xattr		*sx;
+	struct xfs_da_args		args = { NULL };
+	int				error = 0;
+
+	sx = container_of(context, struct xfs_scrub_xattr, context);
+
+	if (flags & XFS_ATTR_INCOMPLETE) {
+		/* Incomplete attr key, just mark the inode for preening. */
+		xfs_scrub_ino_set_preen(sx->sc, context->dp->i_ino);
+		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.geo = context->dp->i_mount->m_attr_geo;
+	args.whichfork = XFS_ATTR_FORK;
+	args.dp = context->dp;
+	args.name = name;
+	args.namelen = namelen;
+	args.hashval = xfs_da_hashname(args.name, args.namelen);
+	args.trans = context->tp;
+	args.value = sx->sc->buf;
+	args.valuelen = XATTR_SIZE_MAX;
+
+	error = xfs_attr_get_ilocked(context->dp, &args);
+	if (error == -EEXIST)
+		error = 0;
+	if (!xfs_scrub_fblock_process_error(sx->sc, XFS_ATTR_FORK, args.blkno,
+			&error))
+		goto fail_xref;
+	if (args.valuelen != valuelen)
+		xfs_scrub_fblock_set_corrupt(sx->sc, XFS_ATTR_FORK,
+					     args.blkno);
+
+fail_xref:
+	return;
+}
+
+/*
+ * Mark a range [start, start+len) in this map.  Returns true if the
+ * region was free, and false if there's a conflict or a problem.
+ *
+ * Within a char, the lowest bit of the char represents the byte with
+ * the smallest address
+ */
+STATIC bool
+xfs_scrub_xattr_set_map(
+	struct xfs_scrub_context	*sc,
+	unsigned long			*map,
+	unsigned int			start,
+	unsigned int			len)
+{
+	unsigned int			mapsize = sc->mp->m_attr_geo->blksize;
+	bool				ret = true;
+
+	if (start >= mapsize)
+		return false;
+	if (start + len > mapsize) {
+		len = mapsize - start;
+		ret = false;
+	}
+
+	if (find_next_bit(map, mapsize, start) < start + len)
+		ret = false;
+	bitmap_set(map, start, len);
+
+	return ret;
+}
+
+/*
+ * Check the leaf freemap from the usage bitmap.  Returns false if the
+ * attr freemap has problems or points to used space.
+ */
+STATIC bool
+xfs_scrub_xattr_check_freemap(
+	struct xfs_scrub_context	*sc,
+	unsigned long			*map,
+	struct xfs_attr3_icleaf_hdr	*leafhdr)
+{
+	unsigned long			*freemap;
+	unsigned long			*dstmap;
+	unsigned int			mapsize = sc->mp->m_attr_geo->blksize;
+	int				i;
+
+	/* Construct bitmap of freemap contents. */
+	freemap = (unsigned long *)sc->buf + BITS_TO_LONGS(mapsize);
+	bitmap_zero(freemap, mapsize);
+	for (i = 0; i < XFS_ATTR_LEAF_MAPSIZE; i++) {
+		if (!xfs_scrub_xattr_set_map(sc, freemap,
+				leafhdr->freemap[i].base,
+				leafhdr->freemap[i].size))
+			return false;
+	}
+
+	/* Look for bits that are set in freemap and are marked in use. */
+	dstmap = freemap + BITS_TO_LONGS(mapsize);
+	return bitmap_and(dstmap, freemap, map, mapsize) == 0;
+}
+
+/*
+ * Check this leaf entry's relations to everything else.
+ * Returns the number of bytes used for the name/value data.
+ */
+STATIC void
+xfs_scrub_xattr_entry(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	char				*buf_end,
+	struct xfs_attr_leafblock	*leaf,
+	struct xfs_attr3_icleaf_hdr	*leafhdr,
+	unsigned long			*usedmap,
+	struct xfs_attr_leaf_entry	*ent,
+	int				idx,
+	unsigned int			*usedbytes,
+	__u32				*last_hashval)
+{
+	struct xfs_mount		*mp = ds->state->mp;
+	char				*name_end;
+	struct xfs_attr_leaf_name_local	*lentry;
+	struct xfs_attr_leaf_name_remote *rentry;
+	unsigned int			nameidx;
+	unsigned int			namesize;
+
+	if (ent->pad2 != 0)
+		xfs_scrub_da_set_corrupt(ds, level);
+
+	/* Hash values in order? */
+	if (be32_to_cpu(ent->hashval) < *last_hashval)
+		xfs_scrub_da_set_corrupt(ds, level);
+	*last_hashval = be32_to_cpu(ent->hashval);
+
+	nameidx = be16_to_cpu(ent->nameidx);
+	if (nameidx < leafhdr->firstused ||
+	    nameidx >= mp->m_attr_geo->blksize) {
+		xfs_scrub_da_set_corrupt(ds, level);
+		return;
+	}
+
+	/* Check the name information. */
+	if (ent->flags & XFS_ATTR_LOCAL) {
+		lentry = xfs_attr3_leaf_name_local(leaf, idx);
+		namesize = xfs_attr_leaf_entsize_local(lentry->namelen,
+				be16_to_cpu(lentry->valuelen));
+		name_end = (char *)lentry + namesize;
+		if (lentry->namelen == 0)
+			xfs_scrub_da_set_corrupt(ds, level);
+	} else {
+		rentry = xfs_attr3_leaf_name_remote(leaf, idx);
+		namesize = xfs_attr_leaf_entsize_remote(rentry->namelen);
+		name_end = (char *)rentry + namesize;
+		if (rentry->namelen == 0 || rentry->valueblk == 0)
+			xfs_scrub_da_set_corrupt(ds, level);
+	}
+	if (name_end > buf_end)
+		xfs_scrub_da_set_corrupt(ds, level);
+
+	if (!xfs_scrub_xattr_set_map(ds->sc, usedmap, nameidx, namesize))
+		xfs_scrub_da_set_corrupt(ds, level);
+	if (!(ds->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		*usedbytes += namesize;
+}
+
+/* Scrub an attribute leaf. */
+STATIC int
+xfs_scrub_xattr_block(
+	struct xfs_scrub_da_btree	*ds,
+	int				level)
+{
+	struct xfs_attr3_icleaf_hdr	leafhdr;
+	struct xfs_mount		*mp = ds->state->mp;
+	struct xfs_da_state_blk		*blk = &ds->state->path.blk[level];
+	struct xfs_buf			*bp = blk->bp;
+	xfs_dablk_t			*last_checked = ds->private;
+	struct xfs_attr_leafblock	*leaf = bp->b_addr;
+	struct xfs_attr_leaf_entry	*ent;
+	struct xfs_attr_leaf_entry	*entries;
+	unsigned long			*usedmap = ds->sc->buf;
+	char				*buf_end;
+	size_t				off;
+	__u32				last_hashval = 0;
+	unsigned int			usedbytes = 0;
+	unsigned int			hdrsize;
+	int				i;
+
+	if (*last_checked == blk->blkno)
+		return 0;
+	*last_checked = blk->blkno;
+	bitmap_zero(usedmap, mp->m_attr_geo->blksize);
+
+	/* Check all the padding. */
+	if (xfs_sb_version_hascrc(&ds->sc->mp->m_sb)) {
+		struct xfs_attr3_leafblock	*leaf = bp->b_addr;
+
+		if (leaf->hdr.pad1 != 0 || leaf->hdr.pad2 != 0 ||
+		    leaf->hdr.info.hdr.pad != 0)
+			xfs_scrub_da_set_corrupt(ds, level);
+	} else {
+		if (leaf->hdr.pad1 != 0 || leaf->hdr.info.pad != 0)
+			xfs_scrub_da_set_corrupt(ds, level);
+	}
+
+	/* Check the leaf header */
+	xfs_attr3_leaf_hdr_from_disk(mp->m_attr_geo, &leafhdr, leaf);
+	hdrsize = xfs_attr3_leaf_hdr_size(leaf);
+
+	if (leafhdr.usedbytes > mp->m_attr_geo->blksize)
+		xfs_scrub_da_set_corrupt(ds, level);
+	if (leafhdr.firstused > mp->m_attr_geo->blksize)
+		xfs_scrub_da_set_corrupt(ds, level);
+	if (leafhdr.firstused < hdrsize)
+		xfs_scrub_da_set_corrupt(ds, level);
+	if (!xfs_scrub_xattr_set_map(ds->sc, usedmap, 0, hdrsize))
+		xfs_scrub_da_set_corrupt(ds, level);
+
+	if (ds->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	entries = xfs_attr3_leaf_entryp(leaf);
+	if ((char *)&entries[leafhdr.count] > (char *)leaf + leafhdr.firstused)
+		xfs_scrub_da_set_corrupt(ds, level);
+
+	buf_end = (char *)bp->b_addr + mp->m_attr_geo->blksize;
+	for (i = 0, ent = entries; i < leafhdr.count; ent++, i++) {
+		/* Mark the leaf entry itself. */
+		off = (char *)ent - (char *)leaf;
+		if (!xfs_scrub_xattr_set_map(ds->sc, usedmap, off,
+				sizeof(xfs_attr_leaf_entry_t))) {
+			xfs_scrub_da_set_corrupt(ds, level);
+			goto out;
+		}
+
+		/* Check the entry and nameval. */
+		xfs_scrub_xattr_entry(ds, level, buf_end, leaf, &leafhdr,
+				usedmap, ent, i, &usedbytes, &last_hashval);
+
+		if (ds->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+			goto out;
+	}
+
+	if (!xfs_scrub_xattr_check_freemap(ds->sc, usedmap, &leafhdr))
+		xfs_scrub_da_set_corrupt(ds, level);
+
+	if (leafhdr.usedbytes != usedbytes)
+		xfs_scrub_da_set_corrupt(ds, level);
+
+out:
+	return 0;
+}
+
+/* Scrub a attribute btree record. */
+STATIC int
+xfs_scrub_xattr_rec(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	void				*rec)
+{
+	struct xfs_mount		*mp = ds->state->mp;
+	struct xfs_attr_leaf_entry	*ent = rec;
+	struct xfs_da_state_blk		*blk;
+	struct xfs_attr_leaf_name_local	*lentry;
+	struct xfs_attr_leaf_name_remote	*rentry;
+	struct xfs_buf			*bp;
+	xfs_dahash_t			calc_hash;
+	xfs_dahash_t			hash;
+	int				nameidx;
+	int				hdrsize;
+	unsigned int			badflags;
+	int				error;
+
+	blk = &ds->state->path.blk[level];
+
+	/* Check the whole block, if necessary. */
+	error = xfs_scrub_xattr_block(ds, level);
+	if (error)
+		goto out;
+	if (ds->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	/* Check the hash of the entry. */
+	error = xfs_scrub_da_btree_hash(ds, level, &ent->hashval);
+	if (error)
+		goto out;
+
+	/* Find the attr entry's location. */
+	bp = blk->bp;
+	hdrsize = xfs_attr3_leaf_hdr_size(bp->b_addr);
+	nameidx = be16_to_cpu(ent->nameidx);
+	if (nameidx < hdrsize || nameidx >= mp->m_attr_geo->blksize) {
+		xfs_scrub_da_set_corrupt(ds, level);
+		goto out;
+	}
+
+	/* Retrieve the entry and check it. */
+	hash = be32_to_cpu(ent->hashval);
+	badflags = ~(XFS_ATTR_LOCAL | XFS_ATTR_ROOT | XFS_ATTR_SECURE |
+			XFS_ATTR_INCOMPLETE);
+	if ((ent->flags & badflags) != 0)
+		xfs_scrub_da_set_corrupt(ds, level);
+	if (ent->flags & XFS_ATTR_LOCAL) {
+		lentry = (struct xfs_attr_leaf_name_local *)
+				(((char *)bp->b_addr) + nameidx);
+		if (lentry->namelen <= 0) {
+			xfs_scrub_da_set_corrupt(ds, level);
+			goto out;
+		}
+		calc_hash = xfs_da_hashname(lentry->nameval, lentry->namelen);
+	} else {
+		rentry = (struct xfs_attr_leaf_name_remote *)
+				(((char *)bp->b_addr) + nameidx);
+		if (rentry->namelen <= 0) {
+			xfs_scrub_da_set_corrupt(ds, level);
+			goto out;
+		}
+		calc_hash = xfs_da_hashname(rentry->name, rentry->namelen);
+	}
+	if (calc_hash != hash)
+		xfs_scrub_da_set_corrupt(ds, level);
+
+out:
+	return error;
+}
+
+/* Scrub the extended attribute metadata. */
+int
+xfs_scrub_xattr(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_scrub_xattr		sx;
+	struct attrlist_cursor_kern	cursor = { 0 };
+	xfs_dablk_t			last_checked = -1U;
+	int				error = 0;
+
+	if (!xfs_inode_hasattr(sc->ip))
+		return -ENOENT;
+
+	memset(&sx, 0, sizeof(sx));
+	/* Check attribute tree structure */
+	error = xfs_scrub_da_btree(sc, XFS_ATTR_FORK, xfs_scrub_xattr_rec,
+			&last_checked);
+	if (error)
+		goto out;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	/* 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 = xfs_scrub_xattr_listent;
+	sx.context.tp = sc->tp;
+	sx.context.flags = ATTR_INCOMPLETE;
+	sx.sc = sc;
+
+	/*
+	 * Look up every xattr in this file by name.
+	 *
+	 * Use the backend implementation of xfs_attr_list to call
+	 * xfs_scrub_xattr_listent on every attribute key in this inode.
+	 * In other words, we use the same iterator/callback mechanism
+	 * that listattr uses to scrub extended attributes, though in our
+	 * _listent function, we check the value of the attribute.
+	 *
+	 * The VFS only locks i_rwsem when modifying attrs, so keep all
+	 * three locks held because that's the only way to ensure we're
+	 * the only thread poking into the da btree.  We traverse the da
+	 * btree while holding a leaf buffer locked for the xattr name
+	 * iteration, which doesn't really follow the usual buffer
+	 * locking order.
+	 */
+	error = xfs_attr_list_int_ilocked(&sx.context);
+	if (!xfs_scrub_fblock_process_error(sc, XFS_ATTR_FORK, 0, &error))
+		goto out;
+out:
+	return error;
+}
diff --git a/fs/xfs/scrub/bmap.c b/fs/xfs/scrub/bmap.c
new file mode 100644
index 000000000000..639d14b51e90
--- /dev/null
+++ b/fs/xfs/scrub/bmap.c
@@ -0,0 +1,734 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_inode_fork.h"
+#include "xfs_alloc.h"
+#include "xfs_rtalloc.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_util.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_rmap.h"
+#include "xfs_rmap_btree.h"
+#include "xfs_refcount.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+
+/* Set us up with an inode's bmap. */
+int
+xfs_scrub_setup_inode_bmap(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	struct xfs_mount		*mp = sc->mp;
+	int				error;
+
+	error = xfs_scrub_get_inode(sc, ip);
+	if (error)
+		goto out;
+
+	sc->ilock_flags = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL;
+	xfs_ilock(sc->ip, sc->ilock_flags);
+
+	/*
+	 * We don't want any ephemeral data fork updates sitting around
+	 * while we inspect block mappings, so wait for directio to finish
+	 * and flush dirty data if we have delalloc reservations.
+	 */
+	if (S_ISREG(VFS_I(sc->ip)->i_mode) &&
+	    sc->sm->sm_type == XFS_SCRUB_TYPE_BMBTD) {
+		inode_dio_wait(VFS_I(sc->ip));
+		error = filemap_write_and_wait(VFS_I(sc->ip)->i_mapping);
+		if (error)
+			goto out;
+	}
+
+	/* Got the inode, lock it and we're ready to go. */
+	error = xfs_scrub_trans_alloc(sc->sm, mp, &sc->tp);
+	if (error)
+		goto out;
+	sc->ilock_flags |= XFS_ILOCK_EXCL;
+	xfs_ilock(sc->ip, XFS_ILOCK_EXCL);
+
+out:
+	/* scrub teardown will unlock and release the inode */
+	return error;
+}
+
+/*
+ * Inode fork block mapping (BMBT) scrubber.
+ * More complex than the others because we have to scrub
+ * all the extents regardless of whether or not the fork
+ * is in btree format.
+ */
+
+struct xfs_scrub_bmap_info {
+	struct xfs_scrub_context	*sc;
+	xfs_fileoff_t			lastoff;
+	bool				is_rt;
+	bool				is_shared;
+	int				whichfork;
+};
+
+/* Look for a corresponding rmap for this irec. */
+static inline bool
+xfs_scrub_bmap_get_rmap(
+	struct xfs_scrub_bmap_info	*info,
+	struct xfs_bmbt_irec		*irec,
+	xfs_agblock_t			agbno,
+	uint64_t			owner,
+	struct xfs_rmap_irec		*rmap)
+{
+	xfs_fileoff_t			offset;
+	unsigned int			rflags = 0;
+	int				has_rmap;
+	int				error;
+
+	if (info->whichfork == XFS_ATTR_FORK)
+		rflags |= XFS_RMAP_ATTR_FORK;
+
+	/*
+	 * CoW staging extents are owned (on disk) by the refcountbt, so
+	 * their rmaps do not have offsets.
+	 */
+	if (info->whichfork == XFS_COW_FORK)
+		offset = 0;
+	else
+		offset = irec->br_startoff;
+
+	/*
+	 * If the caller thinks this could be a shared bmbt extent (IOWs,
+	 * any data fork extent of a reflink inode) then we have to use the
+	 * range rmap lookup to make sure we get the correct owner/offset.
+	 */
+	if (info->is_shared) {
+		error = xfs_rmap_lookup_le_range(info->sc->sa.rmap_cur, agbno,
+				owner, offset, rflags, rmap, &has_rmap);
+		if (!xfs_scrub_should_check_xref(info->sc, &error,
+				&info->sc->sa.rmap_cur))
+			return false;
+		goto out;
+	}
+
+	/*
+	 * Otherwise, use the (faster) regular lookup.
+	 */
+	error = xfs_rmap_lookup_le(info->sc->sa.rmap_cur, agbno, 0, owner,
+			offset, rflags, &has_rmap);
+	if (!xfs_scrub_should_check_xref(info->sc, &error,
+			&info->sc->sa.rmap_cur))
+		return false;
+	if (!has_rmap)
+		goto out;
+
+	error = xfs_rmap_get_rec(info->sc->sa.rmap_cur, rmap, &has_rmap);
+	if (!xfs_scrub_should_check_xref(info->sc, &error,
+			&info->sc->sa.rmap_cur))
+		return false;
+
+out:
+	if (!has_rmap)
+		xfs_scrub_fblock_xref_set_corrupt(info->sc, info->whichfork,
+			irec->br_startoff);
+	return has_rmap;
+}
+
+/* Make sure that we have rmapbt records for this extent. */
+STATIC void
+xfs_scrub_bmap_xref_rmap(
+	struct xfs_scrub_bmap_info	*info,
+	struct xfs_bmbt_irec		*irec,
+	xfs_agblock_t			agbno)
+{
+	struct xfs_rmap_irec		rmap;
+	unsigned long long		rmap_end;
+	uint64_t			owner;
+
+	if (!info->sc->sa.rmap_cur)
+		return;
+
+	if (info->whichfork == XFS_COW_FORK)
+		owner = XFS_RMAP_OWN_COW;
+	else
+		owner = info->sc->ip->i_ino;
+
+	/* Find the rmap record for this irec. */
+	if (!xfs_scrub_bmap_get_rmap(info, irec, agbno, owner, &rmap))
+		return;
+
+	/* Check the rmap. */
+	rmap_end = (unsigned long long)rmap.rm_startblock + rmap.rm_blockcount;
+	if (rmap.rm_startblock > agbno ||
+	    agbno + irec->br_blockcount > rmap_end)
+		xfs_scrub_fblock_xref_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+
+	/*
+	 * Check the logical offsets if applicable.  CoW staging extents
+	 * don't track logical offsets since the mappings only exist in
+	 * memory.
+	 */
+	if (info->whichfork != XFS_COW_FORK) {
+		rmap_end = (unsigned long long)rmap.rm_offset +
+				rmap.rm_blockcount;
+		if (rmap.rm_offset > irec->br_startoff ||
+		    irec->br_startoff + irec->br_blockcount > rmap_end)
+			xfs_scrub_fblock_xref_set_corrupt(info->sc,
+					info->whichfork, irec->br_startoff);
+	}
+
+	if (rmap.rm_owner != owner)
+		xfs_scrub_fblock_xref_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+
+	/*
+	 * Check for discrepancies between the unwritten flag in the irec and
+	 * the rmap.  Note that the (in-memory) CoW fork distinguishes between
+	 * unwritten and written extents, but we don't track that in the rmap
+	 * records because the blocks are owned (on-disk) by the refcountbt,
+	 * which doesn't track unwritten state.
+	 */
+	if (owner != XFS_RMAP_OWN_COW &&
+	    irec->br_state == XFS_EXT_UNWRITTEN &&
+	    !(rmap.rm_flags & XFS_RMAP_UNWRITTEN))
+		xfs_scrub_fblock_xref_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+
+	if (info->whichfork == XFS_ATTR_FORK &&
+	    !(rmap.rm_flags & XFS_RMAP_ATTR_FORK))
+		xfs_scrub_fblock_xref_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+	if (rmap.rm_flags & XFS_RMAP_BMBT_BLOCK)
+		xfs_scrub_fblock_xref_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+}
+
+/* Cross-reference a single rtdev extent record. */
+STATIC void
+xfs_scrub_bmap_rt_extent_xref(
+	struct xfs_scrub_bmap_info	*info,
+	struct xfs_inode		*ip,
+	struct xfs_btree_cur		*cur,
+	struct xfs_bmbt_irec		*irec)
+{
+	if (info->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	xfs_scrub_xref_is_used_rt_space(info->sc, irec->br_startblock,
+			irec->br_blockcount);
+}
+
+/* Cross-reference a single datadev extent record. */
+STATIC void
+xfs_scrub_bmap_extent_xref(
+	struct xfs_scrub_bmap_info	*info,
+	struct xfs_inode		*ip,
+	struct xfs_btree_cur		*cur,
+	struct xfs_bmbt_irec		*irec)
+{
+	struct xfs_mount		*mp = info->sc->mp;
+	xfs_agnumber_t			agno;
+	xfs_agblock_t			agbno;
+	xfs_extlen_t			len;
+	int				error;
+
+	if (info->sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	agno = XFS_FSB_TO_AGNO(mp, irec->br_startblock);
+	agbno = XFS_FSB_TO_AGBNO(mp, irec->br_startblock);
+	len = irec->br_blockcount;
+
+	error = xfs_scrub_ag_init(info->sc, agno, &info->sc->sa);
+	if (!xfs_scrub_fblock_process_error(info->sc, info->whichfork,
+			irec->br_startoff, &error))
+		return;
+
+	xfs_scrub_xref_is_used_space(info->sc, agbno, len);
+	xfs_scrub_xref_is_not_inode_chunk(info->sc, agbno, len);
+	xfs_scrub_bmap_xref_rmap(info, irec, agbno);
+	switch (info->whichfork) {
+	case XFS_DATA_FORK:
+		if (xfs_is_reflink_inode(info->sc->ip))
+			break;
+		/* fall through */
+	case XFS_ATTR_FORK:
+		xfs_scrub_xref_is_not_shared(info->sc, agbno,
+				irec->br_blockcount);
+		break;
+	case XFS_COW_FORK:
+		xfs_scrub_xref_is_cow_staging(info->sc, agbno,
+				irec->br_blockcount);
+		break;
+	}
+
+	xfs_scrub_ag_free(info->sc, &info->sc->sa);
+}
+
+/* Scrub a single extent record. */
+STATIC int
+xfs_scrub_bmap_extent(
+	struct xfs_inode		*ip,
+	struct xfs_btree_cur		*cur,
+	struct xfs_scrub_bmap_info	*info,
+	struct xfs_bmbt_irec		*irec)
+{
+	struct xfs_mount		*mp = info->sc->mp;
+	struct xfs_buf			*bp = NULL;
+	xfs_filblks_t			end;
+	int				error = 0;
+
+	if (cur)
+		xfs_btree_get_block(cur, 0, &bp);
+
+	/*
+	 * Check for out-of-order extents.  This record could have come
+	 * from the incore list, for which there is no ordering check.
+	 */
+	if (irec->br_startoff < info->lastoff)
+		xfs_scrub_fblock_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+
+	/* There should never be a "hole" extent in either extent list. */
+	if (irec->br_startblock == HOLESTARTBLOCK)
+		xfs_scrub_fblock_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+
+	/*
+	 * Check for delalloc extents.  We never iterate the ones in the
+	 * in-core extent scan, and we should never see these in the bmbt.
+	 */
+	if (isnullstartblock(irec->br_startblock))
+		xfs_scrub_fblock_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+
+	/* Make sure the extent points to a valid place. */
+	if (irec->br_blockcount > MAXEXTLEN)
+		xfs_scrub_fblock_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+	if (irec->br_startblock + irec->br_blockcount <= irec->br_startblock)
+		xfs_scrub_fblock_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+	end = irec->br_startblock + irec->br_blockcount - 1;
+	if (info->is_rt &&
+	    (!xfs_verify_rtbno(mp, irec->br_startblock) ||
+	     !xfs_verify_rtbno(mp, end)))
+		xfs_scrub_fblock_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+	if (!info->is_rt &&
+	    (!xfs_verify_fsbno(mp, irec->br_startblock) ||
+	     !xfs_verify_fsbno(mp, end) ||
+	     XFS_FSB_TO_AGNO(mp, irec->br_startblock) !=
+				XFS_FSB_TO_AGNO(mp, end)))
+		xfs_scrub_fblock_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+
+	/* We don't allow unwritten extents on attr forks. */
+	if (irec->br_state == XFS_EXT_UNWRITTEN &&
+	    info->whichfork == XFS_ATTR_FORK)
+		xfs_scrub_fblock_set_corrupt(info->sc, info->whichfork,
+				irec->br_startoff);
+
+	if (info->is_rt)
+		xfs_scrub_bmap_rt_extent_xref(info, ip, cur, irec);
+	else
+		xfs_scrub_bmap_extent_xref(info, ip, cur, irec);
+
+	info->lastoff = irec->br_startoff + irec->br_blockcount;
+	return error;
+}
+
+/* Scrub a bmbt record. */
+STATIC int
+xfs_scrub_bmapbt_rec(
+	struct xfs_scrub_btree		*bs,
+	union xfs_btree_rec		*rec)
+{
+	struct xfs_bmbt_irec		irec;
+	struct xfs_scrub_bmap_info	*info = bs->private;
+	struct xfs_inode		*ip = bs->cur->bc_private.b.ip;
+	struct xfs_buf			*bp = NULL;
+	struct xfs_btree_block		*block;
+	uint64_t			owner;
+	int				i;
+
+	/*
+	 * Check the owners of the btree blocks up to the level below
+	 * the root since the verifiers don't do that.
+	 */
+	if (xfs_sb_version_hascrc(&bs->cur->bc_mp->m_sb) &&
+	    bs->cur->bc_ptrs[0] == 1) {
+		for (i = 0; i < bs->cur->bc_nlevels - 1; i++) {
+			block = xfs_btree_get_block(bs->cur, i, &bp);
+			owner = be64_to_cpu(block->bb_u.l.bb_owner);
+			if (owner != ip->i_ino)
+				xfs_scrub_fblock_set_corrupt(bs->sc,
+						info->whichfork, 0);
+		}
+	}
+
+	/* Set up the in-core record and scrub it. */
+	xfs_bmbt_disk_get_all(&rec->bmbt, &irec);
+	return xfs_scrub_bmap_extent(ip, bs->cur, info, &irec);
+}
+
+/* Scan the btree records. */
+STATIC int
+xfs_scrub_bmap_btree(
+	struct xfs_scrub_context	*sc,
+	int				whichfork,
+	struct xfs_scrub_bmap_info	*info)
+{
+	struct xfs_owner_info		oinfo;
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_inode		*ip = sc->ip;
+	struct xfs_btree_cur		*cur;
+	int				error;
+
+	cur = xfs_bmbt_init_cursor(mp, sc->tp, ip, whichfork);
+	xfs_rmap_ino_bmbt_owner(&oinfo, ip->i_ino, whichfork);
+	error = xfs_scrub_btree(sc, cur, xfs_scrub_bmapbt_rec, &oinfo, info);
+	xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR :
+					  XFS_BTREE_NOERROR);
+	return error;
+}
+
+struct xfs_scrub_bmap_check_rmap_info {
+	struct xfs_scrub_context	*sc;
+	int				whichfork;
+	struct xfs_iext_cursor		icur;
+};
+
+/* Can we find bmaps that fit this rmap? */
+STATIC int
+xfs_scrub_bmap_check_rmap(
+	struct xfs_btree_cur		*cur,
+	struct xfs_rmap_irec		*rec,
+	void				*priv)
+{
+	struct xfs_bmbt_irec		irec;
+	struct xfs_scrub_bmap_check_rmap_info	*sbcri = priv;
+	struct xfs_ifork		*ifp;
+	struct xfs_scrub_context	*sc = sbcri->sc;
+	bool				have_map;
+
+	/* Is this even the right fork? */
+	if (rec->rm_owner != sc->ip->i_ino)
+		return 0;
+	if ((sbcri->whichfork == XFS_ATTR_FORK) ^
+	    !!(rec->rm_flags & XFS_RMAP_ATTR_FORK))
+		return 0;
+	if (rec->rm_flags & XFS_RMAP_BMBT_BLOCK)
+		return 0;
+
+	/* Now look up the bmbt record. */
+	ifp = XFS_IFORK_PTR(sc->ip, sbcri->whichfork);
+	if (!ifp) {
+		xfs_scrub_fblock_set_corrupt(sc, sbcri->whichfork,
+				rec->rm_offset);
+		goto out;
+	}
+	have_map = xfs_iext_lookup_extent(sc->ip, ifp, rec->rm_offset,
+			&sbcri->icur, &irec);
+	if (!have_map)
+		xfs_scrub_fblock_set_corrupt(sc, sbcri->whichfork,
+				rec->rm_offset);
+	/*
+	 * bmap extent record lengths are constrained to 2^21 blocks in length
+	 * because of space constraints in the on-disk metadata structure.
+	 * However, rmap extent record lengths are constrained only by AG
+	 * length, so we have to loop through the bmbt to make sure that the
+	 * entire rmap is covered by bmbt records.
+	 */
+	while (have_map) {
+		if (irec.br_startoff != rec->rm_offset)
+			xfs_scrub_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))
+			xfs_scrub_fblock_set_corrupt(sc, sbcri->whichfork,
+					rec->rm_offset);
+		if (irec.br_blockcount > rec->rm_blockcount)
+			xfs_scrub_fblock_set_corrupt(sc, sbcri->whichfork,
+					rec->rm_offset);
+		if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+			break;
+		rec->rm_startblock += irec.br_blockcount;
+		rec->rm_offset += irec.br_blockcount;
+		rec->rm_blockcount -= irec.br_blockcount;
+		if (rec->rm_blockcount == 0)
+			break;
+		have_map = xfs_iext_next_extent(ifp, &sbcri->icur, &irec);
+		if (!have_map)
+			xfs_scrub_fblock_set_corrupt(sc, sbcri->whichfork,
+					rec->rm_offset);
+	}
+
+out:
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return XFS_BTREE_QUERY_RANGE_ABORT;
+	return 0;
+}
+
+/* Make sure each rmap has a corresponding bmbt entry. */
+STATIC int
+xfs_scrub_bmap_check_ag_rmaps(
+	struct xfs_scrub_context	*sc,
+	int				whichfork,
+	xfs_agnumber_t			agno)
+{
+	struct xfs_scrub_bmap_check_rmap_info	sbcri;
+	struct xfs_btree_cur		*cur;
+	struct xfs_buf			*agf;
+	int				error;
+
+	error = xfs_alloc_read_agf(sc->mp, sc->tp, agno, 0, &agf);
+	if (error)
+		return error;
+
+	cur = xfs_rmapbt_init_cursor(sc->mp, sc->tp, agf, agno);
+	if (!cur) {
+		error = -ENOMEM;
+		goto out_agf;
+	}
+
+	sbcri.sc = sc;
+	sbcri.whichfork = whichfork;
+	error = xfs_rmap_query_all(cur, xfs_scrub_bmap_check_rmap, &sbcri);
+	if (error == XFS_BTREE_QUERY_RANGE_ABORT)
+		error = 0;
+
+	xfs_btree_del_cursor(cur, error ? XFS_BTREE_ERROR : XFS_BTREE_NOERROR);
+out_agf:
+	xfs_trans_brelse(sc->tp, agf);
+	return error;
+}
+
+/* Make sure each rmap has a corresponding bmbt entry. */
+STATIC int
+xfs_scrub_bmap_check_rmaps(
+	struct xfs_scrub_context	*sc,
+	int				whichfork)
+{
+	loff_t				size;
+	xfs_agnumber_t			agno;
+	int				error;
+
+	if (!xfs_sb_version_hasrmapbt(&sc->mp->m_sb) ||
+	    whichfork == XFS_COW_FORK ||
+	    (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+		return 0;
+
+	/* Don't support realtime rmap checks yet. */
+	if (XFS_IS_REALTIME_INODE(sc->ip) && whichfork == XFS_DATA_FORK)
+		return 0;
+
+	/*
+	 * 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.
+	 * The inode repair code can zap broken iforks, which means we have
+	 * 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))
+		return 0;
+
+	for (agno = 0; agno < sc->mp->m_sb.sb_agcount; agno++) {
+		error = xfs_scrub_bmap_check_ag_rmaps(sc, whichfork, agno);
+		if (error)
+			return error;
+		if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+			break;
+	}
+
+	return 0;
+}
+
+/*
+ * Scrub an inode fork's block mappings.
+ *
+ * First we scan every record in every btree block, if applicable.
+ * Then we unconditionally scan the incore extent cache.
+ */
+STATIC int
+xfs_scrub_bmap(
+	struct xfs_scrub_context	*sc,
+	int				whichfork)
+{
+	struct xfs_bmbt_irec		irec;
+	struct xfs_scrub_bmap_info	info = { NULL };
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_inode		*ip = sc->ip;
+	struct xfs_ifork		*ifp;
+	xfs_fileoff_t			endoff;
+	struct xfs_iext_cursor		icur;
+	int				error = 0;
+
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+
+	info.is_rt = whichfork == XFS_DATA_FORK && XFS_IS_REALTIME_INODE(ip);
+	info.whichfork = whichfork;
+	info.is_shared = whichfork == XFS_DATA_FORK && xfs_is_reflink_inode(ip);
+	info.sc = sc;
+
+	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)) {
+			xfs_scrub_ino_set_corrupt(sc, sc->ip->i_ino);
+			goto out;
+		}
+		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))
+			xfs_scrub_ino_set_corrupt(sc, sc->ip->i_ino);
+		break;
+	default:
+		ASSERT(whichfork == XFS_DATA_FORK);
+		break;
+	}
+
+	/* Check the fork values */
+	switch (XFS_IFORK_FORMAT(ip, whichfork)) {
+	case XFS_DINODE_FMT_UUID:
+	case XFS_DINODE_FMT_DEV:
+	case XFS_DINODE_FMT_LOCAL:
+		/* No mappings to check. */
+		goto out;
+	case XFS_DINODE_FMT_EXTENTS:
+		if (!(ifp->if_flags & XFS_IFEXTENTS)) {
+			xfs_scrub_fblock_set_corrupt(sc, whichfork, 0);
+			goto out;
+		}
+		break;
+	case XFS_DINODE_FMT_BTREE:
+		if (whichfork == XFS_COW_FORK) {
+			xfs_scrub_fblock_set_corrupt(sc, whichfork, 0);
+			goto out;
+		}
+
+		error = xfs_scrub_bmap_btree(sc, whichfork, &info);
+		if (error)
+			goto out;
+		break;
+	default:
+		xfs_scrub_fblock_set_corrupt(sc, whichfork, 0);
+		goto out;
+	}
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	/* Now try to scrub the in-memory extent list. */
+        if (!(ifp->if_flags & XFS_IFEXTENTS)) {
+		error = xfs_iread_extents(sc->tp, ip, whichfork);
+		if (!xfs_scrub_fblock_process_error(sc, whichfork, 0, &error))
+			goto out;
+	}
+
+	/* Find the offset of the last extent in the mapping. */
+	error = xfs_bmap_last_offset(ip, &endoff, whichfork);
+	if (!xfs_scrub_fblock_process_error(sc, whichfork, 0, &error))
+		goto out;
+
+	/* Scrub extent records. */
+	info.lastoff = 0;
+	ifp = XFS_IFORK_PTR(ip, whichfork);
+	for_each_xfs_iext(ifp, &icur, &irec) {
+		if (xfs_scrub_should_terminate(sc, &error))
+			break;
+		if (isnullstartblock(irec.br_startblock))
+			continue;
+		if (irec.br_startoff >= endoff) {
+			xfs_scrub_fblock_set_corrupt(sc, whichfork,
+					irec.br_startoff);
+			goto out;
+		}
+		error = xfs_scrub_bmap_extent(ip, NULL, &info, &irec);
+		if (error)
+			goto out;
+	}
+
+out_check_rmap:
+	error = xfs_scrub_bmap_check_rmaps(sc, whichfork);
+	if (!xfs_scrub_fblock_xref_process_error(sc, whichfork, 0, &error))
+		goto out;
+out:
+	return error;
+}
+
+/* Scrub an inode's data fork. */
+int
+xfs_scrub_bmap_data(
+	struct xfs_scrub_context	*sc)
+{
+	return xfs_scrub_bmap(sc, XFS_DATA_FORK);
+}
+
+/* Scrub an inode's attr fork. */
+int
+xfs_scrub_bmap_attr(
+	struct xfs_scrub_context	*sc)
+{
+	return xfs_scrub_bmap(sc, XFS_ATTR_FORK);
+}
+
+/* Scrub an inode's CoW fork. */
+int
+xfs_scrub_bmap_cow(
+	struct xfs_scrub_context	*sc)
+{
+	if (!xfs_is_reflink_inode(sc->ip))
+		return -ENOENT;
+
+	return xfs_scrub_bmap(sc, XFS_COW_FORK);
+}
diff --git a/fs/xfs/scrub/btree.c b/fs/xfs/scrub/btree.c
new file mode 100644
index 000000000000..54218168c8f9
--- /dev/null
+++ b/fs/xfs/scrub/btree.c
@@ -0,0 +1,676 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_alloc.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+
+/* btree scrubbing */
+
+/*
+ * Check for btree operation errors.  See the section about handling
+ * operational errors in common.c.
+ */
+static bool
+__xfs_scrub_btree_process_error(
+	struct xfs_scrub_context	*sc,
+	struct xfs_btree_cur		*cur,
+	int				level,
+	int				*error,
+	__u32				errflag,
+	void				*ret_ip)
+{
+	if (*error == 0)
+		return true;
+
+	switch (*error) {
+	case -EDEADLOCK:
+		/* Used to restart an op with deadlock avoidance. */
+		trace_xfs_scrub_deadlock_retry(sc->ip, sc->sm, *error);
+		break;
+	case -EFSBADCRC:
+	case -EFSCORRUPTED:
+		/* Note the badness but don't abort. */
+		sc->sm->sm_flags |= errflag;
+		*error = 0;
+		/* fall through */
+	default:
+		if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+			trace_xfs_scrub_ifork_btree_op_error(sc, cur, level,
+					*error, ret_ip);
+		else
+			trace_xfs_scrub_btree_op_error(sc, cur, level,
+					*error, ret_ip);
+		break;
+	}
+	return false;
+}
+
+bool
+xfs_scrub_btree_process_error(
+	struct xfs_scrub_context	*sc,
+	struct xfs_btree_cur		*cur,
+	int				level,
+	int				*error)
+{
+	return __xfs_scrub_btree_process_error(sc, cur, level, error,
+			XFS_SCRUB_OFLAG_CORRUPT, __return_address);
+}
+
+bool
+xfs_scrub_btree_xref_process_error(
+	struct xfs_scrub_context	*sc,
+	struct xfs_btree_cur		*cur,
+	int				level,
+	int				*error)
+{
+	return __xfs_scrub_btree_process_error(sc, cur, level, error,
+			XFS_SCRUB_OFLAG_XFAIL, __return_address);
+}
+
+/* Record btree block corruption. */
+static void
+__xfs_scrub_btree_set_corrupt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_btree_cur		*cur,
+	int				level,
+	__u32				errflag,
+	void				*ret_ip)
+{
+	sc->sm->sm_flags |= errflag;
+
+	if (cur->bc_flags & XFS_BTREE_ROOT_IN_INODE)
+		trace_xfs_scrub_ifork_btree_error(sc, cur, level,
+				ret_ip);
+	else
+		trace_xfs_scrub_btree_error(sc, cur, level,
+				ret_ip);
+}
+
+void
+xfs_scrub_btree_set_corrupt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_btree_cur		*cur,
+	int				level)
+{
+	__xfs_scrub_btree_set_corrupt(sc, cur, level, XFS_SCRUB_OFLAG_CORRUPT,
+			__return_address);
+}
+
+void
+xfs_scrub_btree_xref_set_corrupt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_btree_cur		*cur,
+	int				level)
+{
+	__xfs_scrub_btree_set_corrupt(sc, cur, level, XFS_SCRUB_OFLAG_XCORRUPT,
+			__return_address);
+}
+
+/*
+ * Make sure this record is in order and doesn't stray outside of the parent
+ * keys.
+ */
+STATIC void
+xfs_scrub_btree_rec(
+	struct xfs_scrub_btree	*bs)
+{
+	struct xfs_btree_cur	*cur = bs->cur;
+	union xfs_btree_rec	*rec;
+	union xfs_btree_key	key;
+	union xfs_btree_key	hkey;
+	union xfs_btree_key	*keyp;
+	struct xfs_btree_block	*block;
+	struct xfs_btree_block	*keyblock;
+	struct xfs_buf		*bp;
+
+	block = xfs_btree_get_block(cur, 0, &bp);
+	rec = xfs_btree_rec_addr(cur, cur->bc_ptrs[0], block);
+
+	trace_xfs_scrub_btree_rec(bs->sc, cur, 0);
+
+	/* If this isn't the first record, are they in order? */
+	if (!bs->firstrec && !cur->bc_ops->recs_inorder(cur, &bs->lastrec, rec))
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, 0);
+	bs->firstrec = false;
+	memcpy(&bs->lastrec, rec, cur->bc_ops->rec_len);
+
+	if (cur->bc_nlevels == 1)
+		return;
+
+	/* Is this at least as large as the parent low key? */
+	cur->bc_ops->init_key_from_rec(&key, rec);
+	keyblock = xfs_btree_get_block(cur, 1, &bp);
+	keyp = xfs_btree_key_addr(cur, cur->bc_ptrs[1], keyblock);
+	if (cur->bc_ops->diff_two_keys(cur, &key, keyp) < 0)
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, 1);
+
+	if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING))
+		return;
+
+	/* Is this no larger than the parent high key? */
+	cur->bc_ops->init_high_key_from_rec(&hkey, rec);
+	keyp = xfs_btree_high_key_addr(cur, cur->bc_ptrs[1], keyblock);
+	if (cur->bc_ops->diff_two_keys(cur, keyp, &hkey) < 0)
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, 1);
+}
+
+/*
+ * Make sure this key is in order and doesn't stray outside of the parent
+ * keys.
+ */
+STATIC void
+xfs_scrub_btree_key(
+	struct xfs_scrub_btree	*bs,
+	int			level)
+{
+	struct xfs_btree_cur	*cur = bs->cur;
+	union xfs_btree_key	*key;
+	union xfs_btree_key	*keyp;
+	struct xfs_btree_block	*block;
+	struct xfs_btree_block	*keyblock;
+	struct xfs_buf		*bp;
+
+	block = xfs_btree_get_block(cur, level, &bp);
+	key = xfs_btree_key_addr(cur, cur->bc_ptrs[level], block);
+
+	trace_xfs_scrub_btree_key(bs->sc, cur, level);
+
+	/* If this isn't the first key, are they in order? */
+	if (!bs->firstkey[level] &&
+	    !cur->bc_ops->keys_inorder(cur, &bs->lastkey[level], key))
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, level);
+	bs->firstkey[level] = false;
+	memcpy(&bs->lastkey[level], key, cur->bc_ops->key_len);
+
+	if (level + 1 >= cur->bc_nlevels)
+		return;
+
+	/* Is this at least as large as the parent low key? */
+	keyblock = xfs_btree_get_block(cur, level + 1, &bp);
+	keyp = xfs_btree_key_addr(cur, cur->bc_ptrs[level + 1], keyblock);
+	if (cur->bc_ops->diff_two_keys(cur, key, keyp) < 0)
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, level);
+
+	if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING))
+		return;
+
+	/* Is this no larger than the parent high key? */
+	key = xfs_btree_high_key_addr(cur, cur->bc_ptrs[level], block);
+	keyp = xfs_btree_high_key_addr(cur, cur->bc_ptrs[level + 1], keyblock);
+	if (cur->bc_ops->diff_two_keys(cur, keyp, key) < 0)
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, level);
+}
+
+/*
+ * Check a btree pointer.  Returns true if it's ok to use this pointer.
+ * Callers do not need to set the corrupt flag.
+ */
+static bool
+xfs_scrub_btree_ptr_ok(
+	struct xfs_scrub_btree		*bs,
+	int				level,
+	union xfs_btree_ptr		*ptr)
+{
+	bool				res;
+
+	/* A btree rooted in an inode has no block pointer to the root. */
+	if ((bs->cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) &&
+	    level == bs->cur->bc_nlevels)
+		return true;
+
+	/* Otherwise, check the pointers. */
+	if (bs->cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		res = xfs_btree_check_lptr(bs->cur, be64_to_cpu(ptr->l), level);
+	else
+		res = xfs_btree_check_sptr(bs->cur, be32_to_cpu(ptr->s), level);
+	if (!res)
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, level);
+
+	return res;
+}
+
+/* Check that a btree block's sibling matches what we expect it. */
+STATIC int
+xfs_scrub_btree_block_check_sibling(
+	struct xfs_scrub_btree		*bs,
+	int				level,
+	int				direction,
+	union xfs_btree_ptr		*sibling)
+{
+	struct xfs_btree_cur		*cur = bs->cur;
+	struct xfs_btree_block		*pblock;
+	struct xfs_buf			*pbp;
+	struct xfs_btree_cur		*ncur = NULL;
+	union xfs_btree_ptr		*pp;
+	int				success;
+	int				error;
+
+	error = xfs_btree_dup_cursor(cur, &ncur);
+	if (!xfs_scrub_btree_process_error(bs->sc, cur, level + 1, &error) ||
+	    !ncur)
+		return error;
+
+	/*
+	 * If the pointer is null, we shouldn't be able to move the upper
+	 * level pointer anywhere.
+	 */
+	if (xfs_btree_ptr_is_null(cur, sibling)) {
+		if (direction > 0)
+			error = xfs_btree_increment(ncur, level + 1, &success);
+		else
+			error = xfs_btree_decrement(ncur, level + 1, &success);
+		if (error == 0 && success)
+			xfs_scrub_btree_set_corrupt(bs->sc, cur, level);
+		error = 0;
+		goto out;
+	}
+
+	/* Increment upper level pointer. */
+	if (direction > 0)
+		error = xfs_btree_increment(ncur, level + 1, &success);
+	else
+		error = xfs_btree_decrement(ncur, level + 1, &success);
+	if (!xfs_scrub_btree_process_error(bs->sc, cur, level + 1, &error))
+		goto out;
+	if (!success) {
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, level + 1);
+		goto out;
+	}
+
+	/* Compare upper level pointer to sibling pointer. */
+	pblock = xfs_btree_get_block(ncur, level + 1, &pbp);
+	pp = xfs_btree_ptr_addr(ncur, ncur->bc_ptrs[level + 1], pblock);
+	if (!xfs_scrub_btree_ptr_ok(bs, level + 1, pp))
+		goto out;
+	if (pbp)
+		xfs_scrub_buffer_recheck(bs->sc, pbp);
+
+	if (xfs_btree_diff_two_ptrs(cur, pp, sibling))
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, level);
+out:
+	xfs_btree_del_cursor(ncur, XFS_BTREE_ERROR);
+	return error;
+}
+
+/* Check the siblings of a btree block. */
+STATIC int
+xfs_scrub_btree_block_check_siblings(
+	struct xfs_scrub_btree		*bs,
+	struct xfs_btree_block		*block)
+{
+	struct xfs_btree_cur		*cur = bs->cur;
+	union xfs_btree_ptr		leftsib;
+	union xfs_btree_ptr		rightsib;
+	int				level;
+	int				error = 0;
+
+	xfs_btree_get_sibling(cur, block, &leftsib, XFS_BB_LEFTSIB);
+	xfs_btree_get_sibling(cur, block, &rightsib, XFS_BB_RIGHTSIB);
+	level = xfs_btree_get_level(block);
+
+	/* Root block should never have siblings. */
+	if (level == cur->bc_nlevels - 1) {
+		if (!xfs_btree_ptr_is_null(cur, &leftsib) ||
+		    !xfs_btree_ptr_is_null(cur, &rightsib))
+			xfs_scrub_btree_set_corrupt(bs->sc, cur, level);
+		goto out;
+	}
+
+	/*
+	 * Does the left & right sibling pointers match the adjacent
+	 * parent level pointers?
+	 * (These function absorbs error codes for us.)
+	 */
+	error = xfs_scrub_btree_block_check_sibling(bs, level, -1, &leftsib);
+	if (error)
+		return error;
+	error = xfs_scrub_btree_block_check_sibling(bs, level, 1, &rightsib);
+	if (error)
+		return error;
+out:
+	return error;
+}
+
+struct check_owner {
+	struct list_head	list;
+	xfs_daddr_t		daddr;
+	int			level;
+};
+
+/*
+ * Make sure this btree block isn't in the free list and that there's
+ * an rmap record for it.
+ */
+STATIC int
+xfs_scrub_btree_check_block_owner(
+	struct xfs_scrub_btree		*bs,
+	int				level,
+	xfs_daddr_t			daddr)
+{
+	xfs_agnumber_t			agno;
+	xfs_agblock_t			agbno;
+	xfs_btnum_t			btnum;
+	bool				init_sa;
+	int				error = 0;
+
+	if (!bs->cur)
+		return 0;
+
+	btnum = bs->cur->bc_btnum;
+	agno = xfs_daddr_to_agno(bs->cur->bc_mp, daddr);
+	agbno = xfs_daddr_to_agbno(bs->cur->bc_mp, daddr);
+
+	init_sa = bs->cur->bc_flags & XFS_BTREE_LONG_PTRS;
+	if (init_sa) {
+		error = xfs_scrub_ag_init(bs->sc, agno, &bs->sc->sa);
+		if (!xfs_scrub_btree_xref_process_error(bs->sc, bs->cur,
+				level, &error))
+			return error;
+	}
+
+	xfs_scrub_xref_is_used_space(bs->sc, agbno, 1);
+	/*
+	 * The bnobt scrubber aliases bs->cur to bs->sc->sa.bno_cur, so we
+	 * have to nullify it (to shut down further block owner checks) if
+	 * self-xref encounters problems.
+	 */
+	if (!bs->sc->sa.bno_cur && btnum == XFS_BTNUM_BNO)
+		bs->cur = NULL;
+
+	xfs_scrub_xref_is_owned_by(bs->sc, agbno, 1, bs->oinfo);
+	if (!bs->sc->sa.rmap_cur && btnum == XFS_BTNUM_RMAP)
+		bs->cur = NULL;
+
+	if (init_sa)
+		xfs_scrub_ag_free(bs->sc, &bs->sc->sa);
+
+	return error;
+}
+
+/* Check the owner of a btree block. */
+STATIC int
+xfs_scrub_btree_check_owner(
+	struct xfs_scrub_btree		*bs,
+	int				level,
+	struct xfs_buf			*bp)
+{
+	struct xfs_btree_cur		*cur = bs->cur;
+	struct check_owner		*co;
+
+	if ((cur->bc_flags & XFS_BTREE_ROOT_IN_INODE) && bp == NULL)
+		return 0;
+
+	/*
+	 * We want to cross-reference each btree block with the bnobt
+	 * and the rmapbt.  We cannot cross-reference the bnobt or
+	 * rmapbt while scanning the bnobt or rmapbt, respectively,
+	 * because we cannot alter the cursor and we'd prefer not to
+	 * duplicate cursors.  Therefore, save the buffer daddr for
+	 * later scanning.
+	 */
+	if (cur->bc_btnum == XFS_BTNUM_BNO || cur->bc_btnum == XFS_BTNUM_RMAP) {
+		co = kmem_alloc(sizeof(struct check_owner),
+				KM_MAYFAIL | KM_NOFS);
+		if (!co)
+			return -ENOMEM;
+		co->level = level;
+		co->daddr = XFS_BUF_ADDR(bp);
+		list_add_tail(&co->list, &bs->to_check);
+		return 0;
+	}
+
+	return xfs_scrub_btree_check_block_owner(bs, level, XFS_BUF_ADDR(bp));
+}
+
+/*
+ * Grab and scrub a btree block given a btree pointer.  Returns block
+ * and buffer pointers (if applicable) if they're ok to use.
+ */
+STATIC int
+xfs_scrub_btree_get_block(
+	struct xfs_scrub_btree		*bs,
+	int				level,
+	union xfs_btree_ptr		*pp,
+	struct xfs_btree_block		**pblock,
+	struct xfs_buf			**pbp)
+{
+	void				*failed_at;
+	int				error;
+
+	*pblock = NULL;
+	*pbp = NULL;
+
+	error = xfs_btree_lookup_get_block(bs->cur, level, pp, pblock);
+	if (!xfs_scrub_btree_process_error(bs->sc, bs->cur, level, &error) ||
+	    !*pblock)
+		return error;
+
+	xfs_btree_get_block(bs->cur, level, pbp);
+	if (bs->cur->bc_flags & XFS_BTREE_LONG_PTRS)
+		failed_at = __xfs_btree_check_lblock(bs->cur, *pblock,
+				level, *pbp);
+	else
+		failed_at = __xfs_btree_check_sblock(bs->cur, *pblock,
+				 level, *pbp);
+	if (failed_at) {
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, level);
+		return 0;
+	}
+	if (*pbp)
+		xfs_scrub_buffer_recheck(bs->sc, *pbp);
+
+	/*
+	 * Check the block's owner; this function absorbs error codes
+	 * for us.
+	 */
+	error = xfs_scrub_btree_check_owner(bs, level, *pbp);
+	if (error)
+		return error;
+
+	/*
+	 * Check the block's siblings; this function absorbs error codes
+	 * for us.
+	 */
+	return xfs_scrub_btree_block_check_siblings(bs, *pblock);
+}
+
+/*
+ * Check that the low and high keys of this block match the keys stored
+ * in the parent block.
+ */
+STATIC void
+xfs_scrub_btree_block_keys(
+	struct xfs_scrub_btree		*bs,
+	int				level,
+	struct xfs_btree_block		*block)
+{
+	union xfs_btree_key		block_keys;
+	struct xfs_btree_cur		*cur = bs->cur;
+	union xfs_btree_key		*high_bk;
+	union xfs_btree_key		*parent_keys;
+	union xfs_btree_key		*high_pk;
+	struct xfs_btree_block		*parent_block;
+	struct xfs_buf			*bp;
+
+	if (level >= cur->bc_nlevels - 1)
+		return;
+
+	/* Calculate the keys for this block. */
+	xfs_btree_get_keys(cur, block, &block_keys);
+
+	/* Obtain the parent's copy of the keys for this block. */
+	parent_block = xfs_btree_get_block(cur, level + 1, &bp);
+	parent_keys = xfs_btree_key_addr(cur, cur->bc_ptrs[level + 1],
+			parent_block);
+
+	if (cur->bc_ops->diff_two_keys(cur, &block_keys, parent_keys) != 0)
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, 1);
+
+	if (!(cur->bc_flags & XFS_BTREE_OVERLAPPING))
+		return;
+
+	/* Get high keys */
+	high_bk = xfs_btree_high_key_from_key(cur, &block_keys);
+	high_pk = xfs_btree_high_key_addr(cur, cur->bc_ptrs[level + 1],
+			parent_block);
+
+	if (cur->bc_ops->diff_two_keys(cur, high_bk, high_pk) != 0)
+		xfs_scrub_btree_set_corrupt(bs->sc, cur, 1);
+}
+
+/*
+ * Visit all nodes and leaves of a btree.  Check that all pointers and
+ * records are in order, that the keys reflect the records, and use a callback
+ * so that the caller can verify individual records.
+ */
+int
+xfs_scrub_btree(
+	struct xfs_scrub_context	*sc,
+	struct xfs_btree_cur		*cur,
+	xfs_scrub_btree_rec_fn		scrub_fn,
+	struct xfs_owner_info		*oinfo,
+	void				*private)
+{
+	struct xfs_scrub_btree		bs = { NULL };
+	union xfs_btree_ptr		ptr;
+	union xfs_btree_ptr		*pp;
+	union xfs_btree_rec		*recp;
+	struct xfs_btree_block		*block;
+	int				level;
+	struct xfs_buf			*bp;
+	struct check_owner		*co;
+	struct check_owner		*n;
+	int				i;
+	int				error = 0;
+
+	/* Initialize scrub state */
+	bs.cur = cur;
+	bs.scrub_rec = scrub_fn;
+	bs.oinfo = oinfo;
+	bs.firstrec = true;
+	bs.private = private;
+	bs.sc = sc;
+	for (i = 0; i < XFS_BTREE_MAXLEVELS; i++)
+		bs.firstkey[i] = true;
+	INIT_LIST_HEAD(&bs.to_check);
+
+	/* Don't try to check a tree with a height we can't handle. */
+	if (cur->bc_nlevels > XFS_BTREE_MAXLEVELS) {
+		xfs_scrub_btree_set_corrupt(sc, cur, 0);
+		goto out;
+	}
+
+	/*
+	 * Load the root of the btree.  The helper function absorbs
+	 * error codes for us.
+	 */
+	level = cur->bc_nlevels - 1;
+	cur->bc_ops->init_ptr_from_cur(cur, &ptr);
+	if (!xfs_scrub_btree_ptr_ok(&bs, cur->bc_nlevels, &ptr))
+		goto out;
+	error = xfs_scrub_btree_get_block(&bs, level, &ptr, &block, &bp);
+	if (error || !block)
+		goto out;
+
+	cur->bc_ptrs[level] = 1;
+
+	while (level < cur->bc_nlevels) {
+		block = xfs_btree_get_block(cur, level, &bp);
+
+		if (level == 0) {
+			/* End of leaf, pop back towards the root. */
+			if (cur->bc_ptrs[level] >
+			    be16_to_cpu(block->bb_numrecs)) {
+				xfs_scrub_btree_block_keys(&bs, level, block);
+				if (level < cur->bc_nlevels - 1)
+					cur->bc_ptrs[level + 1]++;
+				level++;
+				continue;
+			}
+
+			/* Records in order for scrub? */
+			xfs_scrub_btree_rec(&bs);
+
+			/* Call out to the record checker. */
+			recp = xfs_btree_rec_addr(cur, cur->bc_ptrs[0], block);
+			error = bs.scrub_rec(&bs, recp);
+			if (error)
+				break;
+			if (xfs_scrub_should_terminate(sc, &error) ||
+			    (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+				break;
+
+			cur->bc_ptrs[level]++;
+			continue;
+		}
+
+		/* End of node, pop back towards the root. */
+		if (cur->bc_ptrs[level] > be16_to_cpu(block->bb_numrecs)) {
+			xfs_scrub_btree_block_keys(&bs, level, block);
+			if (level < cur->bc_nlevels - 1)
+				cur->bc_ptrs[level + 1]++;
+			level++;
+			continue;
+		}
+
+		/* Keys in order for scrub? */
+		xfs_scrub_btree_key(&bs, level);
+
+		/* Drill another level deeper. */
+		pp = xfs_btree_ptr_addr(cur, cur->bc_ptrs[level], block);
+		if (!xfs_scrub_btree_ptr_ok(&bs, level, pp)) {
+			cur->bc_ptrs[level]++;
+			continue;
+		}
+		level--;
+		error = xfs_scrub_btree_get_block(&bs, level, pp, &block, &bp);
+		if (error || !block)
+			goto out;
+
+		cur->bc_ptrs[level] = 1;
+	}
+
+out:
+	/* Process deferred owner checks on btree blocks. */
+	list_for_each_entry_safe(co, n, &bs.to_check, list) {
+		if (!error && bs.cur)
+			error = xfs_scrub_btree_check_block_owner(&bs,
+					co->level, co->daddr);
+		list_del(&co->list);
+		kmem_free(co);
+	}
+
+	return error;
+}
diff --git a/fs/xfs/scrub/btree.h b/fs/xfs/scrub/btree.h
new file mode 100644
index 000000000000..e2b868ede70b
--- /dev/null
+++ b/fs/xfs/scrub/btree.h
@@ -0,0 +1,66 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef __XFS_SCRUB_BTREE_H__
+#define __XFS_SCRUB_BTREE_H__
+
+/* btree scrub */
+
+/* Check for btree operation errors. */
+bool xfs_scrub_btree_process_error(struct xfs_scrub_context *sc,
+		struct xfs_btree_cur *cur, int level, int *error);
+
+/* Check for btree xref operation errors. */
+bool xfs_scrub_btree_xref_process_error(struct xfs_scrub_context *sc,
+				struct xfs_btree_cur *cur, int level,
+				int *error);
+
+/* Check for btree corruption. */
+void xfs_scrub_btree_set_corrupt(struct xfs_scrub_context *sc,
+		struct xfs_btree_cur *cur, int level);
+
+/* Check for btree xref discrepancies. */
+void xfs_scrub_btree_xref_set_corrupt(struct xfs_scrub_context *sc,
+		struct xfs_btree_cur *cur, int level);
+
+struct xfs_scrub_btree;
+typedef int (*xfs_scrub_btree_rec_fn)(
+	struct xfs_scrub_btree	*bs,
+	union xfs_btree_rec	*rec);
+
+struct xfs_scrub_btree {
+	/* caller-provided scrub state */
+	struct xfs_scrub_context	*sc;
+	struct xfs_btree_cur		*cur;
+	xfs_scrub_btree_rec_fn		scrub_rec;
+	struct xfs_owner_info		*oinfo;
+	void				*private;
+
+	/* internal scrub state */
+	union xfs_btree_rec		lastrec;
+	bool				firstrec;
+	union xfs_btree_key		lastkey[XFS_BTREE_MAXLEVELS];
+	bool				firstkey[XFS_BTREE_MAXLEVELS];
+	struct list_head		to_check;
+};
+int xfs_scrub_btree(struct xfs_scrub_context *sc, struct xfs_btree_cur *cur,
+		    xfs_scrub_btree_rec_fn scrub_fn,
+		    struct xfs_owner_info *oinfo, void *private);
+
+#endif /* __XFS_SCRUB_BTREE_H__ */
diff --git a/fs/xfs/scrub/common.c b/fs/xfs/scrub/common.c
new file mode 100644
index 000000000000..8ed91d5c868d
--- /dev/null
+++ b/fs/xfs/scrub/common.c
@@ -0,0 +1,775 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_itable.h"
+#include "xfs_alloc.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_ialloc.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_refcount.h"
+#include "xfs_refcount_btree.h"
+#include "xfs_rmap.h"
+#include "xfs_rmap_btree.h"
+#include "xfs_log.h"
+#include "xfs_trans_priv.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/btree.h"
+
+/* Common code for the metadata scrubbers. */
+
+/*
+ * Handling operational errors.
+ *
+ * The *_process_error() family of functions are used to process error return
+ * codes from functions called as part of a scrub operation.
+ *
+ * If there's no error, we return true to tell the caller that it's ok
+ * to move on to the next check in its list.
+ *
+ * For non-verifier errors (e.g. ENOMEM) we return false to tell the
+ * caller that something bad happened, and we preserve *error so that
+ * the caller can return the *error up the stack to userspace.
+ *
+ * Verifier errors (EFSBADCRC/EFSCORRUPTED) are recorded by setting
+ * OFLAG_CORRUPT in sm_flags and the *error is cleared.  In other words,
+ * we track verifier errors (and failed scrub checks) via OFLAG_CORRUPT,
+ * not via return codes.  We return false to tell the caller that
+ * something bad happened.  Since the error has been cleared, the caller
+ * will (presumably) return that zero and scrubbing will move on to
+ * whatever's next.
+ *
+ * ftrace can be used to record the precise metadata location and the
+ * approximate code location of the failed operation.
+ */
+
+/* Check for operational errors. */
+static bool
+__xfs_scrub_process_error(
+	struct xfs_scrub_context	*sc,
+	xfs_agnumber_t			agno,
+	xfs_agblock_t			bno,
+	int				*error,
+	__u32				errflag,
+	void				*ret_ip)
+{
+	switch (*error) {
+	case 0:
+		return true;
+	case -EDEADLOCK:
+		/* Used to restart an op with deadlock avoidance. */
+		trace_xfs_scrub_deadlock_retry(sc->ip, sc->sm, *error);
+		break;
+	case -EFSBADCRC:
+	case -EFSCORRUPTED:
+		/* Note the badness but don't abort. */
+		sc->sm->sm_flags |= errflag;
+		*error = 0;
+		/* fall through */
+	default:
+		trace_xfs_scrub_op_error(sc, agno, bno, *error,
+				ret_ip);
+		break;
+	}
+	return false;
+}
+
+bool
+xfs_scrub_process_error(
+	struct xfs_scrub_context	*sc,
+	xfs_agnumber_t			agno,
+	xfs_agblock_t			bno,
+	int				*error)
+{
+	return __xfs_scrub_process_error(sc, agno, bno, error,
+			XFS_SCRUB_OFLAG_CORRUPT, __return_address);
+}
+
+bool
+xfs_scrub_xref_process_error(
+	struct xfs_scrub_context	*sc,
+	xfs_agnumber_t			agno,
+	xfs_agblock_t			bno,
+	int				*error)
+{
+	return __xfs_scrub_process_error(sc, agno, bno, error,
+			XFS_SCRUB_OFLAG_XFAIL, __return_address);
+}
+
+/* Check for operational errors for a file offset. */
+static bool
+__xfs_scrub_fblock_process_error(
+	struct xfs_scrub_context	*sc,
+	int				whichfork,
+	xfs_fileoff_t			offset,
+	int				*error,
+	__u32				errflag,
+	void				*ret_ip)
+{
+	switch (*error) {
+	case 0:
+		return true;
+	case -EDEADLOCK:
+		/* Used to restart an op with deadlock avoidance. */
+		trace_xfs_scrub_deadlock_retry(sc->ip, sc->sm, *error);
+		break;
+	case -EFSBADCRC:
+	case -EFSCORRUPTED:
+		/* Note the badness but don't abort. */
+		sc->sm->sm_flags |= errflag;
+		*error = 0;
+		/* fall through */
+	default:
+		trace_xfs_scrub_file_op_error(sc, whichfork, offset, *error,
+				ret_ip);
+		break;
+	}
+	return false;
+}
+
+bool
+xfs_scrub_fblock_process_error(
+	struct xfs_scrub_context	*sc,
+	int				whichfork,
+	xfs_fileoff_t			offset,
+	int				*error)
+{
+	return __xfs_scrub_fblock_process_error(sc, whichfork, offset, error,
+			XFS_SCRUB_OFLAG_CORRUPT, __return_address);
+}
+
+bool
+xfs_scrub_fblock_xref_process_error(
+	struct xfs_scrub_context	*sc,
+	int				whichfork,
+	xfs_fileoff_t			offset,
+	int				*error)
+{
+	return __xfs_scrub_fblock_process_error(sc, whichfork, offset, error,
+			XFS_SCRUB_OFLAG_XFAIL, __return_address);
+}
+
+/*
+ * Handling scrub corruption/optimization/warning checks.
+ *
+ * The *_set_{corrupt,preen,warning}() family of functions are used to
+ * record the presence of metadata that is incorrect (corrupt), could be
+ * optimized somehow (preen), or should be flagged for administrative
+ * review but is not incorrect (warn).
+ *
+ * ftrace can be used to record the precise metadata location and
+ * approximate code location of the failed check.
+ */
+
+/* Record a block which could be optimized. */
+void
+xfs_scrub_block_set_preen(
+	struct xfs_scrub_context	*sc,
+	struct xfs_buf			*bp)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_PREEN;
+	trace_xfs_scrub_block_preen(sc, bp->b_bn, __return_address);
+}
+
+/*
+ * Record an inode which could be optimized.  The trace data will
+ * include the block given by bp if bp is given; otherwise it will use
+ * the block location of the inode record itself.
+ */
+void
+xfs_scrub_ino_set_preen(
+	struct xfs_scrub_context	*sc,
+	xfs_ino_t			ino)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_PREEN;
+	trace_xfs_scrub_ino_preen(sc, ino, __return_address);
+}
+
+/* Record a corrupt block. */
+void
+xfs_scrub_block_set_corrupt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_buf			*bp)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
+	trace_xfs_scrub_block_error(sc, bp->b_bn, __return_address);
+}
+
+/* Record a corruption while cross-referencing. */
+void
+xfs_scrub_block_xref_set_corrupt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_buf			*bp)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_XCORRUPT;
+	trace_xfs_scrub_block_error(sc, bp->b_bn, __return_address);
+}
+
+/*
+ * Record a corrupt inode.  The trace data will include the block given
+ * by bp if bp is given; otherwise it will use the block location of the
+ * inode record itself.
+ */
+void
+xfs_scrub_ino_set_corrupt(
+	struct xfs_scrub_context	*sc,
+	xfs_ino_t			ino)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
+	trace_xfs_scrub_ino_error(sc, ino, __return_address);
+}
+
+/* Record a corruption while cross-referencing with an inode. */
+void
+xfs_scrub_ino_xref_set_corrupt(
+	struct xfs_scrub_context	*sc,
+	xfs_ino_t			ino)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_XCORRUPT;
+	trace_xfs_scrub_ino_error(sc, ino, __return_address);
+}
+
+/* Record corruption in a block indexed by a file fork. */
+void
+xfs_scrub_fblock_set_corrupt(
+	struct xfs_scrub_context	*sc,
+	int				whichfork,
+	xfs_fileoff_t			offset)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
+	trace_xfs_scrub_fblock_error(sc, whichfork, offset, __return_address);
+}
+
+/* Record a corruption while cross-referencing a fork block. */
+void
+xfs_scrub_fblock_xref_set_corrupt(
+	struct xfs_scrub_context	*sc,
+	int				whichfork,
+	xfs_fileoff_t			offset)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_XCORRUPT;
+	trace_xfs_scrub_fblock_error(sc, whichfork, offset, __return_address);
+}
+
+/*
+ * Warn about inodes that need administrative review but is not
+ * incorrect.
+ */
+void
+xfs_scrub_ino_set_warning(
+	struct xfs_scrub_context	*sc,
+	xfs_ino_t			ino)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_WARNING;
+	trace_xfs_scrub_ino_warning(sc, ino, __return_address);
+}
+
+/* Warn about a block indexed by a file fork that needs review. */
+void
+xfs_scrub_fblock_set_warning(
+	struct xfs_scrub_context	*sc,
+	int				whichfork,
+	xfs_fileoff_t			offset)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_WARNING;
+	trace_xfs_scrub_fblock_warning(sc, whichfork, offset, __return_address);
+}
+
+/* Signal an incomplete scrub. */
+void
+xfs_scrub_set_incomplete(
+	struct xfs_scrub_context	*sc)
+{
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_INCOMPLETE;
+	trace_xfs_scrub_incomplete(sc, __return_address);
+}
+
+/*
+ * rmap scrubbing -- compute the number of blocks with a given owner,
+ * at least according to the reverse mapping data.
+ */
+
+struct xfs_scrub_rmap_ownedby_info {
+	struct xfs_owner_info	*oinfo;
+	xfs_filblks_t		*blocks;
+};
+
+STATIC int
+xfs_scrub_count_rmap_ownedby_irec(
+	struct xfs_btree_cur			*cur,
+	struct xfs_rmap_irec			*rec,
+	void					*priv)
+{
+	struct xfs_scrub_rmap_ownedby_info	*sroi = priv;
+	bool					irec_attr;
+	bool					oinfo_attr;
+
+	irec_attr = rec->rm_flags & XFS_RMAP_ATTR_FORK;
+	oinfo_attr = sroi->oinfo->oi_flags & XFS_OWNER_INFO_ATTR_FORK;
+
+	if (rec->rm_owner != sroi->oinfo->oi_owner)
+		return 0;
+
+	if (XFS_RMAP_NON_INODE_OWNER(rec->rm_owner) || irec_attr == oinfo_attr)
+		(*sroi->blocks) += rec->rm_blockcount;
+
+	return 0;
+}
+
+/*
+ * Calculate the number of blocks the rmap thinks are owned by something.
+ * The caller should pass us an rmapbt cursor.
+ */
+int
+xfs_scrub_count_rmap_ownedby_ag(
+	struct xfs_scrub_context		*sc,
+	struct xfs_btree_cur			*cur,
+	struct xfs_owner_info			*oinfo,
+	xfs_filblks_t				*blocks)
+{
+	struct xfs_scrub_rmap_ownedby_info	sroi;
+
+	sroi.oinfo = oinfo;
+	*blocks = 0;
+	sroi.blocks = blocks;
+
+	return xfs_rmap_query_all(cur, xfs_scrub_count_rmap_ownedby_irec,
+			&sroi);
+}
+
+/*
+ * AG scrubbing
+ *
+ * These helpers facilitate locking an allocation group's header
+ * buffers, setting up cursors for all btrees that are present, and
+ * cleaning everything up once we're through.
+ */
+
+/* Decide if we want to return an AG header read failure. */
+static inline bool
+want_ag_read_header_failure(
+	struct xfs_scrub_context	*sc,
+	unsigned int			type)
+{
+	/* Return all AG header read failures when scanning btrees. */
+	if (sc->sm->sm_type != XFS_SCRUB_TYPE_AGF &&
+	    sc->sm->sm_type != XFS_SCRUB_TYPE_AGFL &&
+	    sc->sm->sm_type != XFS_SCRUB_TYPE_AGI)
+		return true;
+	/*
+	 * If we're scanning a given type of AG header, we only want to
+	 * see read failures from that specific header.  We'd like the
+	 * other headers to cross-check them, but this isn't required.
+	 */
+	if (sc->sm->sm_type == type)
+		return true;
+	return false;
+}
+
+/*
+ * Grab all the headers for an AG.
+ *
+ * The headers should be released by xfs_scrub_ag_free, but as a fail
+ * safe we attach all the buffers we grab to the scrub transaction so
+ * they'll all be freed when we cancel it.
+ */
+int
+xfs_scrub_ag_read_headers(
+	struct xfs_scrub_context	*sc,
+	xfs_agnumber_t			agno,
+	struct xfs_buf			**agi,
+	struct xfs_buf			**agf,
+	struct xfs_buf			**agfl)
+{
+	struct xfs_mount		*mp = sc->mp;
+	int				error;
+
+	error = xfs_ialloc_read_agi(mp, sc->tp, agno, agi);
+	if (error && want_ag_read_header_failure(sc, XFS_SCRUB_TYPE_AGI))
+		goto out;
+
+	error = xfs_alloc_read_agf(mp, sc->tp, agno, 0, agf);
+	if (error && want_ag_read_header_failure(sc, XFS_SCRUB_TYPE_AGF))
+		goto out;
+
+	error = xfs_alloc_read_agfl(mp, sc->tp, agno, agfl);
+	if (error && want_ag_read_header_failure(sc, XFS_SCRUB_TYPE_AGFL))
+		goto out;
+	error = 0;
+out:
+	return error;
+}
+
+/* Release all the AG btree cursors. */
+void
+xfs_scrub_ag_btcur_free(
+	struct xfs_scrub_ag		*sa)
+{
+	if (sa->refc_cur)
+		xfs_btree_del_cursor(sa->refc_cur, XFS_BTREE_ERROR);
+	if (sa->rmap_cur)
+		xfs_btree_del_cursor(sa->rmap_cur, XFS_BTREE_ERROR);
+	if (sa->fino_cur)
+		xfs_btree_del_cursor(sa->fino_cur, XFS_BTREE_ERROR);
+	if (sa->ino_cur)
+		xfs_btree_del_cursor(sa->ino_cur, XFS_BTREE_ERROR);
+	if (sa->cnt_cur)
+		xfs_btree_del_cursor(sa->cnt_cur, XFS_BTREE_ERROR);
+	if (sa->bno_cur)
+		xfs_btree_del_cursor(sa->bno_cur, XFS_BTREE_ERROR);
+
+	sa->refc_cur = NULL;
+	sa->rmap_cur = NULL;
+	sa->fino_cur = NULL;
+	sa->ino_cur = NULL;
+	sa->bno_cur = NULL;
+	sa->cnt_cur = NULL;
+}
+
+/* Initialize all the btree cursors for an AG. */
+int
+xfs_scrub_ag_btcur_init(
+	struct xfs_scrub_context	*sc,
+	struct xfs_scrub_ag		*sa)
+{
+	struct xfs_mount		*mp = sc->mp;
+	xfs_agnumber_t			agno = sa->agno;
+
+	if (sa->agf_bp) {
+		/* Set up a bnobt cursor for cross-referencing. */
+		sa->bno_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp,
+				agno, XFS_BTNUM_BNO);
+		if (!sa->bno_cur)
+			goto err;
+
+		/* Set up a cntbt cursor for cross-referencing. */
+		sa->cnt_cur = xfs_allocbt_init_cursor(mp, sc->tp, sa->agf_bp,
+				agno, XFS_BTNUM_CNT);
+		if (!sa->cnt_cur)
+			goto err;
+	}
+
+	/* Set up a inobt cursor for cross-referencing. */
+	if (sa->agi_bp) {
+		sa->ino_cur = xfs_inobt_init_cursor(mp, sc->tp, sa->agi_bp,
+					agno, XFS_BTNUM_INO);
+		if (!sa->ino_cur)
+			goto err;
+	}
+
+	/* Set up a finobt cursor for cross-referencing. */
+	if (sa->agi_bp && xfs_sb_version_hasfinobt(&mp->m_sb)) {
+		sa->fino_cur = xfs_inobt_init_cursor(mp, sc->tp, sa->agi_bp,
+				agno, XFS_BTNUM_FINO);
+		if (!sa->fino_cur)
+			goto err;
+	}
+
+	/* Set up a rmapbt cursor for cross-referencing. */
+	if (sa->agf_bp && xfs_sb_version_hasrmapbt(&mp->m_sb)) {
+		sa->rmap_cur = xfs_rmapbt_init_cursor(mp, sc->tp, sa->agf_bp,
+				agno);
+		if (!sa->rmap_cur)
+			goto err;
+	}
+
+	/* Set up a refcountbt cursor for cross-referencing. */
+	if (sa->agf_bp && xfs_sb_version_hasreflink(&mp->m_sb)) {
+		sa->refc_cur = xfs_refcountbt_init_cursor(mp, sc->tp,
+				sa->agf_bp, agno, NULL);
+		if (!sa->refc_cur)
+			goto err;
+	}
+
+	return 0;
+err:
+	return -ENOMEM;
+}
+
+/* Release the AG header context and btree cursors. */
+void
+xfs_scrub_ag_free(
+	struct xfs_scrub_context	*sc,
+	struct xfs_scrub_ag		*sa)
+{
+	xfs_scrub_ag_btcur_free(sa);
+	if (sa->agfl_bp) {
+		xfs_trans_brelse(sc->tp, sa->agfl_bp);
+		sa->agfl_bp = NULL;
+	}
+	if (sa->agf_bp) {
+		xfs_trans_brelse(sc->tp, sa->agf_bp);
+		sa->agf_bp = NULL;
+	}
+	if (sa->agi_bp) {
+		xfs_trans_brelse(sc->tp, sa->agi_bp);
+		sa->agi_bp = NULL;
+	}
+	sa->agno = NULLAGNUMBER;
+}
+
+/*
+ * For scrub, grab the AGI and the AGF headers, in that order.  Locking
+ * order requires us to get the AGI before the AGF.  We use the
+ * transaction to avoid deadlocking on crosslinked metadata buffers;
+ * either the caller passes one in (bmap scrub) or we have to create a
+ * transaction ourselves.
+ */
+int
+xfs_scrub_ag_init(
+	struct xfs_scrub_context	*sc,
+	xfs_agnumber_t			agno,
+	struct xfs_scrub_ag		*sa)
+{
+	int				error;
+
+	sa->agno = agno;
+	error = xfs_scrub_ag_read_headers(sc, agno, &sa->agi_bp,
+			&sa->agf_bp, &sa->agfl_bp);
+	if (error)
+		return error;
+
+	return xfs_scrub_ag_btcur_init(sc, sa);
+}
+
+/* Per-scrubber setup functions */
+
+/* Set us up with a transaction and an empty context. */
+int
+xfs_scrub_setup_fs(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	return xfs_scrub_trans_alloc(sc->sm, sc->mp, &sc->tp);
+}
+
+/* Set us up with AG headers and btree cursors. */
+int
+xfs_scrub_setup_ag_btree(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip,
+	bool				force_log)
+{
+	struct xfs_mount		*mp = sc->mp;
+	int				error;
+
+	/*
+	 * If the caller asks us to checkpont the log, do so.  This
+	 * expensive operation should be performed infrequently and only
+	 * as a last resort.  Any caller that sets force_log should
+	 * document why they need to do so.
+	 */
+	if (force_log) {
+		error = xfs_scrub_checkpoint_log(mp);
+		if (error)
+			return error;
+	}
+
+	error = xfs_scrub_setup_fs(sc, ip);
+	if (error)
+		return error;
+
+	return xfs_scrub_ag_init(sc, sc->sm->sm_agno, &sc->sa);
+}
+
+/* Push everything out of the log onto disk. */
+int
+xfs_scrub_checkpoint_log(
+	struct xfs_mount	*mp)
+{
+	int			error;
+
+	error = xfs_log_force(mp, XFS_LOG_SYNC);
+	if (error)
+		return error;
+	xfs_ail_push_all_sync(mp->m_ail);
+	return 0;
+}
+
+/*
+ * Given an inode and the scrub control structure, grab either the
+ * inode referenced in the control structure or the inode passed in.
+ * The inode is not locked.
+ */
+int
+xfs_scrub_get_inode(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip_in)
+{
+	struct xfs_imap			imap;
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_inode		*ip = NULL;
+	int				error;
+
+	/* We want to scan the inode we already had opened. */
+	if (sc->sm->sm_ino == 0 || sc->sm->sm_ino == ip_in->i_ino) {
+		sc->ip = ip_in;
+		return 0;
+	}
+
+	/* Look up the inode, see if the generation number matches. */
+	if (xfs_internal_inum(mp, sc->sm->sm_ino))
+		return -ENOENT;
+	error = xfs_iget(mp, NULL, sc->sm->sm_ino,
+			XFS_IGET_UNTRUSTED | XFS_IGET_DONTCACHE, 0, &ip);
+	switch (error) {
+	case -ENOENT:
+		/* Inode doesn't exist, just bail out. */
+		return error;
+	case 0:
+		/* Got an inode, continue. */
+		break;
+	case -EINVAL:
+		/*
+		 * -EINVAL with IGET_UNTRUSTED could mean one of several
+		 * things: userspace gave us an inode number that doesn't
+		 * correspond to fs space, or doesn't have an inobt entry;
+		 * or it could simply mean that the inode buffer failed the
+		 * read verifiers.
+		 *
+		 * Try just the inode mapping lookup -- if it succeeds, then
+		 * the inode buffer verifier failed and something needs fixing.
+		 * Otherwise, we really couldn't find it so tell userspace
+		 * that it no longer exists.
+		 */
+		error = xfs_imap(sc->mp, sc->tp, sc->sm->sm_ino, &imap,
+				XFS_IGET_UNTRUSTED | XFS_IGET_DONTCACHE);
+		if (error)
+			return -ENOENT;
+		error = -EFSCORRUPTED;
+		/* fall through */
+	default:
+		trace_xfs_scrub_op_error(sc,
+				XFS_INO_TO_AGNO(mp, sc->sm->sm_ino),
+				XFS_INO_TO_AGBNO(mp, sc->sm->sm_ino),
+				error, __return_address);
+		return error;
+	}
+	if (VFS_I(ip)->i_generation != sc->sm->sm_gen) {
+		iput(VFS_I(ip));
+		return -ENOENT;
+	}
+
+	sc->ip = ip;
+	return 0;
+}
+
+/* Set us up to scrub a file's contents. */
+int
+xfs_scrub_setup_inode_contents(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip,
+	unsigned int			resblks)
+{
+	struct xfs_mount		*mp = sc->mp;
+	int				error;
+
+	error = xfs_scrub_get_inode(sc, ip);
+	if (error)
+		return error;
+
+	/* Got the inode, lock it and we're ready to go. */
+	sc->ilock_flags = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL;
+	xfs_ilock(sc->ip, sc->ilock_flags);
+	error = xfs_scrub_trans_alloc(sc->sm, mp, &sc->tp);
+	if (error)
+		goto out;
+	sc->ilock_flags |= XFS_ILOCK_EXCL;
+	xfs_ilock(sc->ip, XFS_ILOCK_EXCL);
+
+out:
+	/* scrub teardown will unlock and release the inode for us */
+	return error;
+}
+
+/*
+ * Predicate that decides if we need to evaluate the cross-reference check.
+ * If there was an error accessing the cross-reference btree, just delete
+ * the cursor and skip the check.
+ */
+bool
+xfs_scrub_should_check_xref(
+	struct xfs_scrub_context	*sc,
+	int				*error,
+	struct xfs_btree_cur		**curpp)
+{
+	if (*error == 0)
+		return true;
+
+	if (curpp) {
+		/* If we've already given up on xref, just bail out. */
+		if (!*curpp)
+			return false;
+
+		/* xref error, delete cursor and bail out. */
+		xfs_btree_del_cursor(*curpp, XFS_BTREE_ERROR);
+		*curpp = NULL;
+	}
+
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_XFAIL;
+	trace_xfs_scrub_xref_error(sc, *error, __return_address);
+
+	/*
+	 * Errors encountered during cross-referencing with another
+	 * data structure should not cause this scrubber to abort.
+	 */
+	*error = 0;
+	return false;
+}
+
+/* Run the structure verifiers on in-memory buffers to detect bad memory. */
+void
+xfs_scrub_buffer_recheck(
+	struct xfs_scrub_context	*sc,
+	struct xfs_buf			*bp)
+{
+	xfs_failaddr_t			fa;
+
+	if (bp->b_ops == NULL) {
+		xfs_scrub_block_set_corrupt(sc, bp);
+		return;
+	}
+	if (bp->b_ops->verify_struct == NULL) {
+		xfs_scrub_set_incomplete(sc);
+		return;
+	}
+	fa = bp->b_ops->verify_struct(bp);
+	if (!fa)
+		return;
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
+	trace_xfs_scrub_block_error(sc, bp->b_bn, fa);
+}
diff --git a/fs/xfs/scrub/common.h b/fs/xfs/scrub/common.h
new file mode 100644
index 000000000000..deaf60400981
--- /dev/null
+++ b/fs/xfs/scrub/common.h
@@ -0,0 +1,160 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef __XFS_SCRUB_COMMON_H__
+#define __XFS_SCRUB_COMMON_H__
+
+/*
+ * We /could/ terminate a scrub/repair operation early.  If we're not
+ * in a good place to continue (fatal signal, etc.) then bail out.
+ * Note that we're careful not to make any judgements about *error.
+ */
+static inline bool
+xfs_scrub_should_terminate(
+	struct xfs_scrub_context	*sc,
+	int				*error)
+{
+	if (fatal_signal_pending(current)) {
+		if (*error == 0)
+			*error = -EAGAIN;
+		return true;
+	}
+	return false;
+}
+
+/*
+ * Grab an empty transaction so that we can re-grab locked buffers if
+ * one of our btrees turns out to be cyclic.
+ */
+static inline int
+xfs_scrub_trans_alloc(
+	struct xfs_scrub_metadata	*sm,
+	struct xfs_mount		*mp,
+	struct xfs_trans		**tpp)
+{
+	return xfs_trans_alloc_empty(mp, tpp);
+}
+
+bool xfs_scrub_process_error(struct xfs_scrub_context *sc, xfs_agnumber_t agno,
+		xfs_agblock_t bno, int *error);
+bool xfs_scrub_fblock_process_error(struct xfs_scrub_context *sc, int whichfork,
+		xfs_fileoff_t offset, int *error);
+
+bool xfs_scrub_xref_process_error(struct xfs_scrub_context *sc,
+		xfs_agnumber_t agno, xfs_agblock_t bno, int *error);
+bool xfs_scrub_fblock_xref_process_error(struct xfs_scrub_context *sc,
+		int whichfork, xfs_fileoff_t offset, int *error);
+
+void xfs_scrub_block_set_preen(struct xfs_scrub_context *sc,
+		struct xfs_buf *bp);
+void xfs_scrub_ino_set_preen(struct xfs_scrub_context *sc, xfs_ino_t ino);
+
+void xfs_scrub_block_set_corrupt(struct xfs_scrub_context *sc,
+		struct xfs_buf *bp);
+void xfs_scrub_ino_set_corrupt(struct xfs_scrub_context *sc, xfs_ino_t ino);
+void xfs_scrub_fblock_set_corrupt(struct xfs_scrub_context *sc, int whichfork,
+		xfs_fileoff_t offset);
+
+void xfs_scrub_block_xref_set_corrupt(struct xfs_scrub_context *sc,
+		struct xfs_buf *bp);
+void xfs_scrub_ino_xref_set_corrupt(struct xfs_scrub_context *sc,
+		xfs_ino_t ino);
+void xfs_scrub_fblock_xref_set_corrupt(struct xfs_scrub_context *sc,
+		int whichfork, xfs_fileoff_t offset);
+
+void xfs_scrub_ino_set_warning(struct xfs_scrub_context *sc, xfs_ino_t ino);
+void xfs_scrub_fblock_set_warning(struct xfs_scrub_context *sc, int whichfork,
+		xfs_fileoff_t offset);
+
+void xfs_scrub_set_incomplete(struct xfs_scrub_context *sc);
+int xfs_scrub_checkpoint_log(struct xfs_mount *mp);
+
+/* Are we set up for a cross-referencing check? */
+bool xfs_scrub_should_check_xref(struct xfs_scrub_context *sc, int *error,
+			   struct xfs_btree_cur **curpp);
+
+/* Setup functions */
+int xfs_scrub_setup_fs(struct xfs_scrub_context *sc, struct xfs_inode *ip);
+int xfs_scrub_setup_ag_allocbt(struct xfs_scrub_context *sc,
+			       struct xfs_inode *ip);
+int xfs_scrub_setup_ag_iallocbt(struct xfs_scrub_context *sc,
+				struct xfs_inode *ip);
+int xfs_scrub_setup_ag_rmapbt(struct xfs_scrub_context *sc,
+			      struct xfs_inode *ip);
+int xfs_scrub_setup_ag_refcountbt(struct xfs_scrub_context *sc,
+				  struct xfs_inode *ip);
+int xfs_scrub_setup_inode(struct xfs_scrub_context *sc,
+			  struct xfs_inode *ip);
+int xfs_scrub_setup_inode_bmap(struct xfs_scrub_context *sc,
+			       struct xfs_inode *ip);
+int xfs_scrub_setup_inode_bmap_data(struct xfs_scrub_context *sc,
+				    struct xfs_inode *ip);
+int xfs_scrub_setup_directory(struct xfs_scrub_context *sc,
+			      struct xfs_inode *ip);
+int xfs_scrub_setup_xattr(struct xfs_scrub_context *sc,
+			  struct xfs_inode *ip);
+int xfs_scrub_setup_symlink(struct xfs_scrub_context *sc,
+			    struct xfs_inode *ip);
+int xfs_scrub_setup_parent(struct xfs_scrub_context *sc,
+			   struct xfs_inode *ip);
+#ifdef CONFIG_XFS_RT
+int xfs_scrub_setup_rt(struct xfs_scrub_context *sc, struct xfs_inode *ip);
+#else
+static inline int
+xfs_scrub_setup_rt(struct xfs_scrub_context *sc, struct xfs_inode *ip)
+{
+	return -ENOENT;
+}
+#endif
+#ifdef CONFIG_XFS_QUOTA
+int xfs_scrub_setup_quota(struct xfs_scrub_context *sc, struct xfs_inode *ip);
+#else
+static inline int
+xfs_scrub_setup_quota(struct xfs_scrub_context *sc, struct xfs_inode *ip)
+{
+	return -ENOENT;
+}
+#endif
+
+void xfs_scrub_ag_free(struct xfs_scrub_context *sc, struct xfs_scrub_ag *sa);
+int xfs_scrub_ag_init(struct xfs_scrub_context *sc, xfs_agnumber_t agno,
+		      struct xfs_scrub_ag *sa);
+int xfs_scrub_ag_read_headers(struct xfs_scrub_context *sc, xfs_agnumber_t agno,
+			      struct xfs_buf **agi, struct xfs_buf **agf,
+			      struct xfs_buf **agfl);
+void xfs_scrub_ag_btcur_free(struct xfs_scrub_ag *sa);
+int xfs_scrub_ag_btcur_init(struct xfs_scrub_context *sc,
+			    struct xfs_scrub_ag *sa);
+int xfs_scrub_walk_agfl(struct xfs_scrub_context *sc,
+			int (*fn)(struct xfs_scrub_context *, xfs_agblock_t bno,
+				  void *),
+			void *priv);
+int xfs_scrub_count_rmap_ownedby_ag(struct xfs_scrub_context *sc,
+				    struct xfs_btree_cur *cur,
+				    struct xfs_owner_info *oinfo,
+				    xfs_filblks_t *blocks);
+
+int xfs_scrub_setup_ag_btree(struct xfs_scrub_context *sc,
+			     struct xfs_inode *ip, bool force_log);
+int xfs_scrub_get_inode(struct xfs_scrub_context *sc, struct xfs_inode *ip_in);
+int xfs_scrub_setup_inode_contents(struct xfs_scrub_context *sc,
+				   struct xfs_inode *ip, unsigned int resblks);
+void xfs_scrub_buffer_recheck(struct xfs_scrub_context *sc, struct xfs_buf *bp);
+
+#endif	/* __XFS_SCRUB_COMMON_H__ */
diff --git a/fs/xfs/scrub/dabtree.c b/fs/xfs/scrub/dabtree.c
new file mode 100644
index 000000000000..bffdb7dc09bf
--- /dev/null
+++ b/fs/xfs/scrub/dabtree.c
@@ -0,0 +1,613 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_inode_fork.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_dir2.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_attr_leaf.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/dabtree.h"
+
+/* Directory/Attribute Btree */
+
+/*
+ * Check for da btree operation errors.  See the section about handling
+ * operational errors in common.c.
+ */
+bool
+xfs_scrub_da_process_error(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	int				*error)
+{
+	struct xfs_scrub_context	*sc = ds->sc;
+
+	if (*error == 0)
+		return true;
+
+	switch (*error) {
+	case -EDEADLOCK:
+		/* Used to restart an op with deadlock avoidance. */
+		trace_xfs_scrub_deadlock_retry(sc->ip, sc->sm, *error);
+		break;
+	case -EFSBADCRC:
+	case -EFSCORRUPTED:
+		/* Note the badness but don't abort. */
+		sc->sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
+		*error = 0;
+		/* fall through */
+	default:
+		trace_xfs_scrub_file_op_error(sc, ds->dargs.whichfork,
+				xfs_dir2_da_to_db(ds->dargs.geo,
+					ds->state->path.blk[level].blkno),
+				*error, __return_address);
+		break;
+	}
+	return false;
+}
+
+/*
+ * Check for da btree corruption.  See the section about handling
+ * operational errors in common.c.
+ */
+void
+xfs_scrub_da_set_corrupt(
+	struct xfs_scrub_da_btree	*ds,
+	int				level)
+{
+	struct xfs_scrub_context	*sc = ds->sc;
+
+	sc->sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
+
+	trace_xfs_scrub_fblock_error(sc, ds->dargs.whichfork,
+			xfs_dir2_da_to_db(ds->dargs.geo,
+				ds->state->path.blk[level].blkno),
+			__return_address);
+}
+
+/* Find an entry at a certain level in a da btree. */
+STATIC void *
+xfs_scrub_da_btree_entry(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	int				rec)
+{
+	char				*ents;
+	struct xfs_da_state_blk		*blk;
+	void				*baddr;
+
+	/* Dispatch the entry finding function. */
+	blk = &ds->state->path.blk[level];
+	baddr = blk->bp->b_addr;
+	switch (blk->magic) {
+	case XFS_ATTR_LEAF_MAGIC:
+	case XFS_ATTR3_LEAF_MAGIC:
+		ents = (char *)xfs_attr3_leaf_entryp(baddr);
+		return ents + (rec * sizeof(struct xfs_attr_leaf_entry));
+	case XFS_DIR2_LEAFN_MAGIC:
+	case XFS_DIR3_LEAFN_MAGIC:
+		ents = (char *)ds->dargs.dp->d_ops->leaf_ents_p(baddr);
+		return ents + (rec * sizeof(struct xfs_dir2_leaf_entry));
+	case XFS_DIR2_LEAF1_MAGIC:
+	case XFS_DIR3_LEAF1_MAGIC:
+		ents = (char *)ds->dargs.dp->d_ops->leaf_ents_p(baddr);
+		return ents + (rec * sizeof(struct xfs_dir2_leaf_entry));
+	case XFS_DA_NODE_MAGIC:
+	case XFS_DA3_NODE_MAGIC:
+		ents = (char *)ds->dargs.dp->d_ops->node_tree_p(baddr);
+		return ents + (rec * sizeof(struct xfs_da_node_entry));
+	}
+
+	return NULL;
+}
+
+/* Scrub a da btree hash (key). */
+int
+xfs_scrub_da_btree_hash(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	__be32				*hashp)
+{
+	struct xfs_da_state_blk		*blks;
+	struct xfs_da_node_entry	*entry;
+	xfs_dahash_t			hash;
+	xfs_dahash_t			parent_hash;
+
+	/* Is this hash in order? */
+	hash = be32_to_cpu(*hashp);
+	if (hash < ds->hashes[level])
+		xfs_scrub_da_set_corrupt(ds, level);
+	ds->hashes[level] = hash;
+
+	if (level == 0)
+		return 0;
+
+	/* Is this hash no larger than the parent hash? */
+	blks = ds->state->path.blk;
+	entry = xfs_scrub_da_btree_entry(ds, level - 1, blks[level - 1].index);
+	parent_hash = be32_to_cpu(entry->hashval);
+	if (parent_hash < hash)
+		xfs_scrub_da_set_corrupt(ds, level);
+
+	return 0;
+}
+
+/*
+ * Check a da btree pointer.  Returns true if it's ok to use this
+ * pointer.
+ */
+STATIC bool
+xfs_scrub_da_btree_ptr_ok(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	xfs_dablk_t			blkno)
+{
+	if (blkno < ds->lowest || (ds->highest != 0 && blkno >= ds->highest)) {
+		xfs_scrub_da_set_corrupt(ds, level);
+		return false;
+	}
+
+	return true;
+}
+
+/*
+ * The da btree scrubber can handle leaf1 blocks as a degenerate
+ * form of leafn blocks.  Since the regular da code doesn't handle
+ * leaf1, we must multiplex the verifiers.
+ */
+static void
+xfs_scrub_da_btree_read_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_da_blkinfo	*info = bp->b_addr;
+
+	switch (be16_to_cpu(info->magic)) {
+	case XFS_DIR2_LEAF1_MAGIC:
+	case XFS_DIR3_LEAF1_MAGIC:
+		bp->b_ops = &xfs_dir3_leaf1_buf_ops;
+		bp->b_ops->verify_read(bp);
+		return;
+	default:
+		/*
+		 * xfs_da3_node_buf_ops already know how to handle
+		 * DA*_NODE, ATTR*_LEAF, and DIR*_LEAFN blocks.
+		 */
+		bp->b_ops = &xfs_da3_node_buf_ops;
+		bp->b_ops->verify_read(bp);
+		return;
+	}
+}
+static void
+xfs_scrub_da_btree_write_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_da_blkinfo	*info = bp->b_addr;
+
+	switch (be16_to_cpu(info->magic)) {
+	case XFS_DIR2_LEAF1_MAGIC:
+	case XFS_DIR3_LEAF1_MAGIC:
+		bp->b_ops = &xfs_dir3_leaf1_buf_ops;
+		bp->b_ops->verify_write(bp);
+		return;
+	default:
+		/*
+		 * xfs_da3_node_buf_ops already know how to handle
+		 * DA*_NODE, ATTR*_LEAF, and DIR*_LEAFN blocks.
+		 */
+		bp->b_ops = &xfs_da3_node_buf_ops;
+		bp->b_ops->verify_write(bp);
+		return;
+	}
+}
+static void *
+xfs_scrub_da_btree_verify(
+	struct xfs_buf		*bp)
+{
+	struct xfs_da_blkinfo	*info = bp->b_addr;
+
+	switch (be16_to_cpu(info->magic)) {
+	case XFS_DIR2_LEAF1_MAGIC:
+	case XFS_DIR3_LEAF1_MAGIC:
+		bp->b_ops = &xfs_dir3_leaf1_buf_ops;
+		return bp->b_ops->verify_struct(bp);
+	default:
+		bp->b_ops = &xfs_da3_node_buf_ops;
+		return bp->b_ops->verify_struct(bp);
+	}
+}
+
+static const struct xfs_buf_ops xfs_scrub_da_btree_buf_ops = {
+	.name = "xfs_scrub_da_btree",
+	.verify_read = xfs_scrub_da_btree_read_verify,
+	.verify_write = xfs_scrub_da_btree_write_verify,
+	.verify_struct = xfs_scrub_da_btree_verify,
+};
+
+/* Check a block's sibling. */
+STATIC int
+xfs_scrub_da_btree_block_check_sibling(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	int				direction,
+	xfs_dablk_t			sibling)
+{
+	int				retval;
+	int				error;
+
+	memcpy(&ds->state->altpath, &ds->state->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);
+		if (error == 0 && retval == 0)
+			xfs_scrub_da_set_corrupt(ds, level);
+		error = 0;
+		goto out;
+	}
+
+	/* Move the alternate cursor one block in the direction given. */
+	error = xfs_da3_path_shift(ds->state, &ds->state->altpath,
+			direction, false, &retval);
+	if (!xfs_scrub_da_process_error(ds, level, &error))
+		return error;
+	if (retval) {
+		xfs_scrub_da_set_corrupt(ds, level);
+		return error;
+	}
+	if (ds->state->altpath.blk[level].bp)
+		xfs_scrub_buffer_recheck(ds->sc,
+				ds->state->altpath.blk[level].bp);
+
+	/* Compare upper level pointer to sibling pointer. */
+	if (ds->state->altpath.blk[level].blkno != sibling)
+		xfs_scrub_da_set_corrupt(ds, level);
+	xfs_trans_brelse(ds->dargs.trans, ds->state->altpath.blk[level].bp);
+out:
+	return error;
+}
+
+/* Check a block's sibling pointers. */
+STATIC int
+xfs_scrub_da_btree_block_check_siblings(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	struct xfs_da_blkinfo		*hdr)
+{
+	xfs_dablk_t			forw;
+	xfs_dablk_t			back;
+	int				error = 0;
+
+	forw = be32_to_cpu(hdr->forw);
+	back = be32_to_cpu(hdr->back);
+
+	/* Top level blocks should not have sibling pointers. */
+	if (level == 0) {
+		if (forw != 0 || back != 0)
+			xfs_scrub_da_set_corrupt(ds, level);
+		return 0;
+	}
+
+	/*
+	 * Check back (left) and forw (right) pointers.  These functions
+	 * absorb error codes for us.
+	 */
+	error = xfs_scrub_da_btree_block_check_sibling(ds, level, 0, back);
+	if (error)
+		goto out;
+	error = xfs_scrub_da_btree_block_check_sibling(ds, level, 1, forw);
+
+out:
+	memset(&ds->state->altpath, 0, sizeof(ds->state->altpath));
+	return error;
+}
+
+/* Load a dir/attribute block from a btree. */
+STATIC int
+xfs_scrub_da_btree_block(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	xfs_dablk_t			blkno)
+{
+	struct xfs_da_state_blk		*blk;
+	struct xfs_da_intnode		*node;
+	struct xfs_da_node_entry	*btree;
+	struct xfs_da3_blkinfo		*hdr3;
+	struct xfs_da_args		*dargs = &ds->dargs;
+	struct xfs_inode		*ip = ds->dargs.dp;
+	xfs_ino_t			owner;
+	int				*pmaxrecs;
+	struct xfs_da3_icnode_hdr	nodehdr;
+	int				error = 0;
+
+	blk = &ds->state->path.blk[level];
+	ds->state->path.active = level + 1;
+
+	/* Release old block. */
+	if (blk->bp) {
+		xfs_trans_brelse(dargs->trans, blk->bp);
+		blk->bp = NULL;
+	}
+
+	/* Check the pointer. */
+	blk->blkno = blkno;
+	if (!xfs_scrub_da_btree_ptr_ok(ds, level, blkno))
+		goto out_nobuf;
+
+	/* Read the buffer. */
+	error = xfs_da_read_buf(dargs->trans, dargs->dp, blk->blkno, -2,
+			&blk->bp, dargs->whichfork,
+			&xfs_scrub_da_btree_buf_ops);
+	if (!xfs_scrub_da_process_error(ds, level, &error))
+		goto out_nobuf;
+	if (blk->bp)
+		xfs_scrub_buffer_recheck(ds->sc, blk->bp);
+
+	/*
+	 * We didn't find a dir btree root block, which means that
+	 * there's no LEAF1/LEAFN tree (at least not where it's supposed
+	 * to be), so jump out now.
+	 */
+	if (ds->dargs.whichfork == XFS_DATA_FORK && level == 0 &&
+			blk->bp == NULL)
+		goto out_nobuf;
+
+	/* It's /not/ ok for attr trees not to have a da btree. */
+	if (blk->bp == NULL) {
+		xfs_scrub_da_set_corrupt(ds, level);
+		goto out_nobuf;
+	}
+
+	hdr3 = blk->bp->b_addr;
+	blk->magic = be16_to_cpu(hdr3->hdr.magic);
+	pmaxrecs = &ds->maxrecs[level];
+
+	/* We only started zeroing the header on v5 filesystems. */
+	if (xfs_sb_version_hascrc(&ds->sc->mp->m_sb) && hdr3->hdr.pad)
+		xfs_scrub_da_set_corrupt(ds, level);
+
+	/* Check the owner. */
+	if (xfs_sb_version_hascrc(&ip->i_mount->m_sb)) {
+		owner = be64_to_cpu(hdr3->owner);
+		if (owner != ip->i_ino)
+			xfs_scrub_da_set_corrupt(ds, level);
+	}
+
+	/* Check the siblings. */
+	error = xfs_scrub_da_btree_block_check_siblings(ds, level, &hdr3->hdr);
+	if (error)
+		goto out;
+
+	/* Interpret the buffer. */
+	switch (blk->magic) {
+	case XFS_ATTR_LEAF_MAGIC:
+	case XFS_ATTR3_LEAF_MAGIC:
+		xfs_trans_buf_set_type(dargs->trans, blk->bp,
+				XFS_BLFT_ATTR_LEAF_BUF);
+		blk->magic = XFS_ATTR_LEAF_MAGIC;
+		blk->hashval = xfs_attr_leaf_lasthash(blk->bp, pmaxrecs);
+		if (ds->tree_level != 0)
+			xfs_scrub_da_set_corrupt(ds, level);
+		break;
+	case XFS_DIR2_LEAFN_MAGIC:
+	case XFS_DIR3_LEAFN_MAGIC:
+		xfs_trans_buf_set_type(dargs->trans, blk->bp,
+				XFS_BLFT_DIR_LEAFN_BUF);
+		blk->magic = XFS_DIR2_LEAFN_MAGIC;
+		blk->hashval = xfs_dir2_leaf_lasthash(ip, blk->bp, pmaxrecs);
+		if (ds->tree_level != 0)
+			xfs_scrub_da_set_corrupt(ds, level);
+		break;
+	case XFS_DIR2_LEAF1_MAGIC:
+	case XFS_DIR3_LEAF1_MAGIC:
+		xfs_trans_buf_set_type(dargs->trans, blk->bp,
+				XFS_BLFT_DIR_LEAF1_BUF);
+		blk->magic = XFS_DIR2_LEAF1_MAGIC;
+		blk->hashval = xfs_dir2_leaf_lasthash(ip, blk->bp, pmaxrecs);
+		if (ds->tree_level != 0)
+			xfs_scrub_da_set_corrupt(ds, level);
+		break;
+	case XFS_DA_NODE_MAGIC:
+	case XFS_DA3_NODE_MAGIC:
+		xfs_trans_buf_set_type(dargs->trans, blk->bp,
+				XFS_BLFT_DA_NODE_BUF);
+		blk->magic = XFS_DA_NODE_MAGIC;
+		node = blk->bp->b_addr;
+		ip->d_ops->node_hdr_from_disk(&nodehdr, node);
+		btree = ip->d_ops->node_tree_p(node);
+		*pmaxrecs = nodehdr.count;
+		blk->hashval = be32_to_cpu(btree[*pmaxrecs - 1].hashval);
+		if (level == 0) {
+			if (nodehdr.level >= XFS_DA_NODE_MAXDEPTH) {
+				xfs_scrub_da_set_corrupt(ds, level);
+				goto out_freebp;
+			}
+			ds->tree_level = nodehdr.level;
+		} else {
+			if (ds->tree_level != nodehdr.level) {
+				xfs_scrub_da_set_corrupt(ds, level);
+				goto out_freebp;
+			}
+		}
+
+		/* XXX: Check hdr3.pad32 once we know how to fix it. */
+		break;
+	default:
+		xfs_scrub_da_set_corrupt(ds, level);
+		goto out_freebp;
+	}
+
+out:
+	return error;
+out_freebp:
+	xfs_trans_brelse(dargs->trans, blk->bp);
+	blk->bp = NULL;
+out_nobuf:
+	blk->blkno = 0;
+	return error;
+}
+
+/* Visit all nodes and leaves of a da btree. */
+int
+xfs_scrub_da_btree(
+	struct xfs_scrub_context	*sc,
+	int				whichfork,
+	xfs_scrub_da_btree_rec_fn	scrub_fn,
+	void				*private)
+{
+	struct xfs_scrub_da_btree	ds = {};
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_da_state_blk		*blks;
+	struct xfs_da_node_entry	*key;
+	void				*rec;
+	xfs_dablk_t			blkno;
+	int				level;
+	int				error;
+
+	/* Skip short format data structures; no btree to scan. */
+	if (XFS_IFORK_FORMAT(sc->ip, whichfork) != XFS_DINODE_FMT_EXTENTS &&
+	    XFS_IFORK_FORMAT(sc->ip, whichfork) != XFS_DINODE_FMT_BTREE)
+		return 0;
+
+	/* Set up initial da state. */
+	ds.dargs.dp = sc->ip;
+	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.sc = sc;
+	ds.private = private;
+	if (whichfork == XFS_ATTR_FORK) {
+		ds.dargs.geo = mp->m_attr_geo;
+		ds.lowest = 0;
+		ds.highest = 0;
+	} else {
+		ds.dargs.geo = mp->m_dir_geo;
+		ds.lowest = ds.dargs.geo->leafblk;
+		ds.highest = ds.dargs.geo->freeblk;
+	}
+	blkno = ds.lowest;
+	level = 0;
+
+	/* Find the root of the da tree, if present. */
+	blks = ds.state->path.blk;
+	error = xfs_scrub_da_btree_block(&ds, level, blkno);
+	if (error)
+		goto out_state;
+	/*
+	 * We didn't find a block at ds.lowest, which means that there's
+	 * no LEAF1/LEAFN tree (at least not where it's supposed to be),
+	 * so jump out now.
+	 */
+	if (blks[level].bp == NULL)
+		goto out_state;
+
+	blks[level].index = 0;
+	while (level >= 0 && level < XFS_DA_NODE_MAXDEPTH) {
+		/* Handle leaf block. */
+		if (blks[level].magic != XFS_DA_NODE_MAGIC) {
+			/* End of leaf, pop back towards the root. */
+			if (blks[level].index >= ds.maxrecs[level]) {
+				if (level > 0)
+					blks[level - 1].index++;
+				ds.tree_level++;
+				level--;
+				continue;
+			}
+
+			/* Dispatch record scrubbing. */
+			rec = xfs_scrub_da_btree_entry(&ds, level,
+					blks[level].index);
+			error = scrub_fn(&ds, level, rec);
+			if (error)
+				break;
+			if (xfs_scrub_should_terminate(sc, &error) ||
+			    (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT))
+				break;
+
+			blks[level].index++;
+			continue;
+		}
+
+
+		/* End of node, pop back towards the root. */
+		if (blks[level].index >= ds.maxrecs[level]) {
+			if (level > 0)
+				blks[level - 1].index++;
+			ds.tree_level++;
+			level--;
+			continue;
+		}
+
+		/* Hashes in order for scrub? */
+		key = xfs_scrub_da_btree_entry(&ds, level, blks[level].index);
+		error = xfs_scrub_da_btree_hash(&ds, level, &key->hashval);
+		if (error)
+			goto out;
+
+		/* Drill another level deeper. */
+		blkno = be32_to_cpu(key->before);
+		level++;
+		ds.tree_level--;
+		error = xfs_scrub_da_btree_block(&ds, level, blkno);
+		if (error)
+			goto out;
+		if (blks[level].bp == NULL)
+			goto out;
+
+		blks[level].index = 0;
+	}
+
+out:
+	/* Release all the buffers we're tracking. */
+	for (level = 0; level < XFS_DA_NODE_MAXDEPTH; level++) {
+		if (blks[level].bp == NULL)
+			continue;
+		xfs_trans_brelse(sc->tp, blks[level].bp);
+		blks[level].bp = NULL;
+	}
+
+out_state:
+	xfs_da_state_free(ds.state);
+	return error;
+}
diff --git a/fs/xfs/scrub/dabtree.h b/fs/xfs/scrub/dabtree.h
new file mode 100644
index 000000000000..d31468d68cef
--- /dev/null
+++ b/fs/xfs/scrub/dabtree.h
@@ -0,0 +1,59 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef __XFS_SCRUB_DABTREE_H__
+#define __XFS_SCRUB_DABTREE_H__
+
+/* dir/attr btree */
+
+struct xfs_scrub_da_btree {
+	struct xfs_da_args		dargs;
+	xfs_dahash_t			hashes[XFS_DA_NODE_MAXDEPTH];
+	int				maxrecs[XFS_DA_NODE_MAXDEPTH];
+	struct xfs_da_state		*state;
+	struct xfs_scrub_context	*sc;
+	void				*private;
+
+	/*
+	 * Lowest and highest directory block address in which we expect
+	 * to find dir/attr btree node blocks.  For a directory this
+	 * (presumably) means between LEAF_OFFSET and FREE_OFFSET; for
+	 * attributes there is no limit.
+	 */
+	xfs_dablk_t			lowest;
+	xfs_dablk_t			highest;
+
+	int				tree_level;
+};
+
+typedef int (*xfs_scrub_da_btree_rec_fn)(struct xfs_scrub_da_btree *ds,
+		int level, void *rec);
+
+/* Check for da btree operation errors. */
+bool xfs_scrub_da_process_error(struct xfs_scrub_da_btree *ds, int level, int *error);
+
+/* Check for da btree corruption. */
+void xfs_scrub_da_set_corrupt(struct xfs_scrub_da_btree *ds, int level);
+
+int xfs_scrub_da_btree_hash(struct xfs_scrub_da_btree *ds, int level,
+			    __be32 *hashp);
+int xfs_scrub_da_btree(struct xfs_scrub_context *sc, int whichfork,
+		       xfs_scrub_da_btree_rec_fn scrub_fn, void *private);
+
+#endif /* __XFS_SCRUB_DABTREE_H__ */
diff --git a/fs/xfs/scrub/dir.c b/fs/xfs/scrub/dir.c
new file mode 100644
index 000000000000..38f29806eb54
--- /dev/null
+++ b/fs/xfs/scrub/dir.c
@@ -0,0 +1,842 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_itable.h"
+#include "xfs_da_format.h"
+#include "xfs_da_btree.h"
+#include "xfs_dir2.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_ialloc.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/dabtree.h"
+
+/* Set us up to scrub directories. */
+int
+xfs_scrub_setup_directory(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	return xfs_scrub_setup_inode_contents(sc, ip, 0);
+}
+
+/* Directories */
+
+/* Scrub a directory entry. */
+
+struct xfs_scrub_dir_ctx {
+	/* VFS fill-directory iterator */
+	struct dir_context		dir_iter;
+
+	struct xfs_scrub_context	*sc;
+};
+
+/* Check that an inode's mode matches a given DT_ type. */
+STATIC int
+xfs_scrub_dir_check_ftype(
+	struct xfs_scrub_dir_ctx	*sdc,
+	xfs_fileoff_t			offset,
+	xfs_ino_t			inum,
+	int				dtype)
+{
+	struct xfs_mount		*mp = sdc->sc->mp;
+	struct xfs_inode		*ip;
+	int				ino_dtype;
+	int				error = 0;
+
+	if (!xfs_sb_version_hasftype(&mp->m_sb)) {
+		if (dtype != DT_UNKNOWN && dtype != DT_DIR)
+			xfs_scrub_fblock_set_corrupt(sdc->sc, XFS_DATA_FORK,
+					offset);
+		goto out;
+	}
+
+	/*
+	 * Grab the inode pointed to by the dirent.  We release the
+	 * inode before we cancel the scrub transaction.  Since we're
+	 * don't know a priori that releasing the inode won't trigger
+	 * eofblocks cleanup (which allocates what would be a nested
+	 * transaction), we can't use DONTCACHE here because DONTCACHE
+	 * inodes can trigger immediate inactive cleanup of the inode.
+	 */
+	error = xfs_iget(mp, sdc->sc->tp, inum, 0, 0, &ip);
+	if (!xfs_scrub_fblock_xref_process_error(sdc->sc, XFS_DATA_FORK, offset,
+			&error))
+		goto out;
+
+	/* Convert mode to the DT_* values that dir_emit uses. */
+	ino_dtype = xfs_dir3_get_dtype(mp,
+			xfs_mode_to_ftype(VFS_I(ip)->i_mode));
+	if (ino_dtype != dtype)
+		xfs_scrub_fblock_set_corrupt(sdc->sc, XFS_DATA_FORK, offset);
+	iput(VFS_I(ip));
+out:
+	return error;
+}
+
+/*
+ * Scrub a single directory entry.
+ *
+ * We use the VFS directory iterator (i.e. readdir) to call this
+ * function for every directory entry in a directory.  Once we're here,
+ * we check the inode number to make sure it's sane, then we check that
+ * we can look up this filename.  Finally, we check the ftype.
+ */
+STATIC int
+xfs_scrub_dir_actor(
+	struct dir_context		*dir_iter,
+	const char			*name,
+	int				namelen,
+	loff_t				pos,
+	u64				ino,
+	unsigned			type)
+{
+	struct xfs_mount		*mp;
+	struct xfs_inode		*ip;
+	struct xfs_scrub_dir_ctx	*sdc;
+	struct xfs_name			xname;
+	xfs_ino_t			lookup_ino;
+	xfs_dablk_t			offset;
+	int				error = 0;
+
+	sdc = container_of(dir_iter, struct xfs_scrub_dir_ctx, dir_iter);
+	ip = sdc->sc->ip;
+	mp = ip->i_mount;
+	offset = xfs_dir2_db_to_da(mp->m_dir_geo,
+			xfs_dir2_dataptr_to_db(mp->m_dir_geo, pos));
+
+	/* Does this inode number make sense? */
+	if (!xfs_verify_dir_ino(mp, ino)) {
+		xfs_scrub_fblock_set_corrupt(sdc->sc, XFS_DATA_FORK, offset);
+		goto out;
+	}
+
+	if (!strncmp(".", name, namelen)) {
+		/* If this is "." then check that the inum matches the dir. */
+		if (xfs_sb_version_hasftype(&mp->m_sb) && type != DT_DIR)
+			xfs_scrub_fblock_set_corrupt(sdc->sc, XFS_DATA_FORK,
+					offset);
+		if (ino != ip->i_ino)
+			xfs_scrub_fblock_set_corrupt(sdc->sc, XFS_DATA_FORK,
+					offset);
+	} else if (!strncmp("..", name, namelen)) {
+		/*
+		 * If this is ".." in the root inode, check that the inum
+		 * matches this dir.
+		 */
+		if (xfs_sb_version_hasftype(&mp->m_sb) && type != DT_DIR)
+			xfs_scrub_fblock_set_corrupt(sdc->sc, XFS_DATA_FORK,
+					offset);
+		if (ip->i_ino == mp->m_sb.sb_rootino && ino != ip->i_ino)
+			xfs_scrub_fblock_set_corrupt(sdc->sc, XFS_DATA_FORK,
+					offset);
+	}
+
+	/* Verify that we can look up this name by hash. */
+	xname.name = name;
+	xname.len = namelen;
+	xname.type = XFS_DIR3_FT_UNKNOWN;
+
+	error = xfs_dir_lookup(sdc->sc->tp, ip, &xname, &lookup_ino, NULL);
+	if (!xfs_scrub_fblock_process_error(sdc->sc, XFS_DATA_FORK, offset,
+			&error))
+		goto fail_xref;
+	if (lookup_ino != ino) {
+		xfs_scrub_fblock_set_corrupt(sdc->sc, XFS_DATA_FORK, offset);
+		goto out;
+	}
+
+	/* Verify the file type.  This function absorbs error codes. */
+	error = xfs_scrub_dir_check_ftype(sdc, offset, lookup_ino, type);
+	if (error)
+		goto out;
+out:
+	return error;
+fail_xref:
+	return error;
+}
+
+/* Scrub a directory btree record. */
+STATIC int
+xfs_scrub_dir_rec(
+	struct xfs_scrub_da_btree	*ds,
+	int				level,
+	void				*rec)
+{
+	struct xfs_mount		*mp = ds->state->mp;
+	struct xfs_dir2_leaf_entry	*ent = rec;
+	struct xfs_inode		*dp = ds->dargs.dp;
+	struct xfs_dir2_data_entry	*dent;
+	struct xfs_buf			*bp;
+	char				*p, *endp;
+	xfs_ino_t			ino;
+	xfs_dablk_t			rec_bno;
+	xfs_dir2_db_t			db;
+	xfs_dir2_data_aoff_t		off;
+	xfs_dir2_dataptr_t		ptr;
+	xfs_dahash_t			calc_hash;
+	xfs_dahash_t			hash;
+	unsigned int			tag;
+	int				error;
+
+	/* Check the hash of the entry. */
+	error = xfs_scrub_da_btree_hash(ds, level, &ent->hashval);
+	if (error)
+		goto out;
+
+	/* Valid hash pointer? */
+	ptr = be32_to_cpu(ent->address);
+	if (ptr == 0)
+		return 0;
+
+	/* Find the directory entry's location. */
+	db = xfs_dir2_dataptr_to_db(mp->m_dir_geo, ptr);
+	off = xfs_dir2_dataptr_to_off(mp->m_dir_geo, ptr);
+	rec_bno = xfs_dir2_db_to_da(mp->m_dir_geo, db);
+
+	if (rec_bno >= mp->m_dir_geo->leafblk) {
+		xfs_scrub_da_set_corrupt(ds, level);
+		goto out;
+	}
+	error = xfs_dir3_data_read(ds->dargs.trans, dp, rec_bno, -2, &bp);
+	if (!xfs_scrub_fblock_process_error(ds->sc, XFS_DATA_FORK, rec_bno,
+			&error))
+		goto out;
+	if (!bp) {
+		xfs_scrub_fblock_set_corrupt(ds->sc, XFS_DATA_FORK, rec_bno);
+		goto out;
+	}
+	xfs_scrub_buffer_recheck(ds->sc, bp);
+
+	dent = (struct xfs_dir2_data_entry *)(((char *)bp->b_addr) + off);
+
+	/* Make sure we got a real directory entry. */
+	p = (char *)mp->m_dir_inode_ops->data_entry_p(bp->b_addr);
+	endp = xfs_dir3_data_endp(mp->m_dir_geo, bp->b_addr);
+	if (!endp) {
+		xfs_scrub_fblock_set_corrupt(ds->sc, XFS_DATA_FORK, rec_bno);
+		goto out_relse;
+	}
+	while (p < endp) {
+		struct xfs_dir2_data_entry	*dep;
+		struct xfs_dir2_data_unused	*dup;
+
+		dup = (struct xfs_dir2_data_unused *)p;
+		if (be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG) {
+			p += be16_to_cpu(dup->length);
+			continue;
+		}
+		dep = (struct xfs_dir2_data_entry *)p;
+		if (dep == dent)
+			break;
+		p += mp->m_dir_inode_ops->data_entsize(dep->namelen);
+	}
+	if (p >= endp) {
+		xfs_scrub_fblock_set_corrupt(ds->sc, XFS_DATA_FORK, rec_bno);
+		goto out_relse;
+	}
+
+	/* Retrieve the entry, sanity check it, and compare hashes. */
+	ino = be64_to_cpu(dent->inumber);
+	hash = be32_to_cpu(ent->hashval);
+	tag = be16_to_cpup(dp->d_ops->data_entry_tag_p(dent));
+	if (!xfs_verify_dir_ino(mp, ino) || tag != off)
+		xfs_scrub_fblock_set_corrupt(ds->sc, XFS_DATA_FORK, rec_bno);
+	if (dent->namelen == 0) {
+		xfs_scrub_fblock_set_corrupt(ds->sc, XFS_DATA_FORK, rec_bno);
+		goto out_relse;
+	}
+	calc_hash = xfs_da_hashname(dent->name, dent->namelen);
+	if (calc_hash != hash)
+		xfs_scrub_fblock_set_corrupt(ds->sc, XFS_DATA_FORK, rec_bno);
+
+out_relse:
+	xfs_trans_brelse(ds->dargs.trans, bp);
+out:
+	return error;
+}
+
+/*
+ * Is this unused entry either in the bestfree or smaller than all of
+ * them?  We've already checked that the bestfrees are sorted longest to
+ * shortest, and that there aren't any bogus entries.
+ */
+STATIC void
+xfs_scrub_directory_check_free_entry(
+	struct xfs_scrub_context	*sc,
+	xfs_dablk_t			lblk,
+	struct xfs_dir2_data_free	*bf,
+	struct xfs_dir2_data_unused	*dup)
+{
+	struct xfs_dir2_data_free	*dfp;
+	unsigned int			dup_length;
+
+	dup_length = be16_to_cpu(dup->length);
+
+	/* Unused entry is shorter than any of the bestfrees */
+	if (dup_length < be16_to_cpu(bf[XFS_DIR2_DATA_FD_COUNT - 1].length))
+		return;
+
+	for (dfp = &bf[XFS_DIR2_DATA_FD_COUNT - 1]; dfp >= bf; dfp--)
+		if (dup_length == be16_to_cpu(dfp->length))
+			return;
+
+	/* Unused entry should be in the bestfrees but wasn't found. */
+	xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+}
+
+/* Check free space info in a directory data block. */
+STATIC int
+xfs_scrub_directory_data_bestfree(
+	struct xfs_scrub_context	*sc,
+	xfs_dablk_t			lblk,
+	bool				is_block)
+{
+	struct xfs_dir2_data_unused	*dup;
+	struct xfs_dir2_data_free	*dfp;
+	struct xfs_buf			*bp;
+	struct xfs_dir2_data_free	*bf;
+	struct xfs_mount		*mp = sc->mp;
+	const struct xfs_dir_ops	*d_ops;
+	char				*ptr;
+	char				*endptr;
+	u16				tag;
+	unsigned int			nr_bestfrees = 0;
+	unsigned int			nr_frees = 0;
+	unsigned int			smallest_bestfree;
+	int				newlen;
+	int				offset;
+	int				error;
+
+	d_ops = sc->ip->d_ops;
+
+	if (is_block) {
+		/* dir block format */
+		if (lblk != XFS_B_TO_FSBT(mp, XFS_DIR2_DATA_OFFSET))
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+		error = xfs_dir3_block_read(sc->tp, sc->ip, &bp);
+	} else {
+		/* dir data format */
+		error = xfs_dir3_data_read(sc->tp, sc->ip, lblk, -1, &bp);
+	}
+	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
+		goto out;
+	xfs_scrub_buffer_recheck(sc, bp);
+
+	/* XXX: Check xfs_dir3_data_hdr.pad is zero once we start setting it. */
+
+	/* Do the bestfrees correspond to actual free space? */
+	bf = d_ops->data_bestfree_p(bp->b_addr);
+	smallest_bestfree = UINT_MAX;
+	for (dfp = &bf[0]; dfp < &bf[XFS_DIR2_DATA_FD_COUNT]; dfp++) {
+		offset = be16_to_cpu(dfp->offset);
+		if (offset == 0)
+			continue;
+		if (offset >= mp->m_dir_geo->blksize) {
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+			goto out_buf;
+		}
+		dup = (struct xfs_dir2_data_unused *)(bp->b_addr + offset);
+		tag = be16_to_cpu(*xfs_dir2_data_unused_tag_p(dup));
+
+		/* bestfree doesn't match the entry it points at? */
+		if (dup->freetag != cpu_to_be16(XFS_DIR2_DATA_FREE_TAG) ||
+		    be16_to_cpu(dup->length) != be16_to_cpu(dfp->length) ||
+		    tag != ((char *)dup - (char *)bp->b_addr)) {
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+			goto out_buf;
+		}
+
+		/* bestfree records should be ordered largest to smallest */
+		if (smallest_bestfree < be16_to_cpu(dfp->length)) {
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+			goto out_buf;
+		}
+
+		smallest_bestfree = be16_to_cpu(dfp->length);
+		nr_bestfrees++;
+	}
+
+	/* Make sure the bestfrees are actually the best free spaces. */
+	ptr = (char *)d_ops->data_entry_p(bp->b_addr);
+	endptr = xfs_dir3_data_endp(mp->m_dir_geo, bp->b_addr);
+
+	/* Iterate the entries, stopping when we hit or go past the end. */
+	while (ptr < endptr) {
+		dup = (struct xfs_dir2_data_unused *)ptr;
+		/* Skip real entries */
+		if (dup->freetag != cpu_to_be16(XFS_DIR2_DATA_FREE_TAG)) {
+			struct xfs_dir2_data_entry	*dep;
+
+			dep = (struct xfs_dir2_data_entry *)ptr;
+			newlen = d_ops->data_entsize(dep->namelen);
+			if (newlen <= 0) {
+				xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK,
+						lblk);
+				goto out_buf;
+			}
+			ptr += newlen;
+			continue;
+		}
+
+		/* Spot check this free entry */
+		tag = be16_to_cpu(*xfs_dir2_data_unused_tag_p(dup));
+		if (tag != ((char *)dup - (char *)bp->b_addr))
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+
+		/*
+		 * Either this entry is a bestfree or it's smaller than
+		 * any of the bestfrees.
+		 */
+		xfs_scrub_directory_check_free_entry(sc, lblk, bf, dup);
+
+		/* Move on. */
+		newlen = be16_to_cpu(dup->length);
+		if (newlen <= 0) {
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+			goto out_buf;
+		}
+		ptr += newlen;
+		if (ptr <= endptr)
+			nr_frees++;
+	}
+
+	/* We're required to fill all the space. */
+	if (ptr != endptr)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+
+	/* Did we see at least as many free slots as there are bestfrees? */
+	if (nr_frees < nr_bestfrees)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+out_buf:
+	xfs_trans_brelse(sc->tp, bp);
+out:
+	return error;
+}
+
+/*
+ * Does the free space length in the free space index block ($len) match
+ * the longest length in the directory data block's bestfree array?
+ * Assume that we've already checked that the data block's bestfree
+ * array is in order.
+ */
+STATIC void
+xfs_scrub_directory_check_freesp(
+	struct xfs_scrub_context	*sc,
+	xfs_dablk_t			lblk,
+	struct xfs_buf			*dbp,
+	unsigned int			len)
+{
+	struct xfs_dir2_data_free	*dfp;
+
+	dfp = sc->ip->d_ops->data_bestfree_p(dbp->b_addr);
+
+	if (len != be16_to_cpu(dfp->length))
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+
+	if (len > 0 && be16_to_cpu(dfp->offset) == 0)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+}
+
+/* Check free space info in a directory leaf1 block. */
+STATIC int
+xfs_scrub_directory_leaf1_bestfree(
+	struct xfs_scrub_context	*sc,
+	struct xfs_da_args		*args,
+	xfs_dablk_t			lblk)
+{
+	struct xfs_dir3_icleaf_hdr	leafhdr;
+	struct xfs_dir2_leaf_entry	*ents;
+	struct xfs_dir2_leaf_tail	*ltp;
+	struct xfs_dir2_leaf		*leaf;
+	struct xfs_buf			*dbp;
+	struct xfs_buf			*bp;
+	const struct xfs_dir_ops	*d_ops = sc->ip->d_ops;
+	struct xfs_da_geometry		*geo = sc->mp->m_dir_geo;
+	__be16				*bestp;
+	__u16				best;
+	__u32				hash;
+	__u32				lasthash = 0;
+	__u32				bestcount;
+	unsigned int			stale = 0;
+	int				i;
+	int				error;
+
+	/* Read the free space block. */
+	error = xfs_dir3_leaf_read(sc->tp, sc->ip, lblk, -1, &bp);
+	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
+		goto out;
+	xfs_scrub_buffer_recheck(sc, bp);
+
+	leaf = bp->b_addr;
+	d_ops->leaf_hdr_from_disk(&leafhdr, leaf);
+	ents = d_ops->leaf_ents_p(leaf);
+	ltp = xfs_dir2_leaf_tail_p(geo, leaf);
+	bestcount = be32_to_cpu(ltp->bestcount);
+	bestp = xfs_dir2_leaf_bests_p(ltp);
+
+	if (xfs_sb_version_hascrc(&sc->mp->m_sb)) {
+		struct xfs_dir3_leaf_hdr	*hdr3 = bp->b_addr;
+
+		if (hdr3->pad != cpu_to_be32(0))
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+	}
+
+	/*
+	 * There should be as many bestfree slots as there are dir data
+	 * blocks that can fit under i_size.
+	 */
+	if (bestcount != xfs_dir2_byte_to_db(geo, sc->ip->i_d.di_size)) {
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+		goto out;
+	}
+
+	/* Is the leaf count even remotely sane? */
+	if (leafhdr.count > d_ops->leaf_max_ents(geo)) {
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+		goto out;
+	}
+
+	/* Leaves and bests don't overlap in leaf format. */
+	if ((char *)&ents[leafhdr.count] > (char *)bestp) {
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+		goto out;
+	}
+
+	/* Check hash value order, count stale entries.  */
+	for (i = 0; i < leafhdr.count; i++) {
+		hash = be32_to_cpu(ents[i].hashval);
+		if (i > 0 && lasthash > hash)
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+		lasthash = hash;
+		if (ents[i].address == cpu_to_be32(XFS_DIR2_NULL_DATAPTR))
+			stale++;
+	}
+	if (leafhdr.stale != stale)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+
+	/* Check all the bestfree entries. */
+	for (i = 0; i < bestcount; i++, bestp++) {
+		best = be16_to_cpu(*bestp);
+		if (best == NULLDATAOFF)
+			continue;
+		error = xfs_dir3_data_read(sc->tp, sc->ip,
+				i * args->geo->fsbcount, -1, &dbp);
+		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, lblk,
+				&error))
+			continue;
+		xfs_scrub_directory_check_freesp(sc, lblk, dbp, best);
+		xfs_trans_brelse(sc->tp, dbp);
+	}
+out:
+	return error;
+}
+
+/* Check free space info in a directory freespace block. */
+STATIC int
+xfs_scrub_directory_free_bestfree(
+	struct xfs_scrub_context	*sc,
+	struct xfs_da_args		*args,
+	xfs_dablk_t			lblk)
+{
+	struct xfs_dir3_icfree_hdr	freehdr;
+	struct xfs_buf			*dbp;
+	struct xfs_buf			*bp;
+	__be16				*bestp;
+	__u16				best;
+	unsigned int			stale = 0;
+	int				i;
+	int				error;
+
+	/* Read the free space block */
+	error = xfs_dir2_free_read(sc->tp, sc->ip, lblk, &bp);
+	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
+		goto out;
+	xfs_scrub_buffer_recheck(sc, bp);
+
+	if (xfs_sb_version_hascrc(&sc->mp->m_sb)) {
+		struct xfs_dir3_free_hdr	*hdr3 = bp->b_addr;
+
+		if (hdr3->pad != cpu_to_be32(0))
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+	}
+
+	/* Check all the entries. */
+	sc->ip->d_ops->free_hdr_from_disk(&freehdr, bp->b_addr);
+	bestp = sc->ip->d_ops->free_bests_p(bp->b_addr);
+	for (i = 0; i < freehdr.nvalid; i++, bestp++) {
+		best = be16_to_cpu(*bestp);
+		if (best == NULLDATAOFF) {
+			stale++;
+			continue;
+		}
+		error = xfs_dir3_data_read(sc->tp, sc->ip,
+				(freehdr.firstdb + i) * args->geo->fsbcount,
+				-1, &dbp);
+		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, lblk,
+				&error))
+			continue;
+		xfs_scrub_directory_check_freesp(sc, lblk, dbp, best);
+		xfs_trans_brelse(sc->tp, dbp);
+	}
+
+	if (freehdr.nused + stale != freehdr.nvalid)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+out:
+	return error;
+}
+
+/* Check free space information in directories. */
+STATIC int
+xfs_scrub_directory_blocks(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_bmbt_irec		got;
+	struct xfs_da_args		args;
+	struct xfs_ifork		*ifp;
+	struct xfs_mount		*mp = sc->mp;
+	xfs_fileoff_t			leaf_lblk;
+	xfs_fileoff_t			free_lblk;
+	xfs_fileoff_t			lblk;
+	struct xfs_iext_cursor		icur;
+	xfs_dablk_t			dabno;
+	bool				found;
+	int				is_block = 0;
+	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)
+		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);
+
+	/* Is this a block dir? */
+	args.dp = sc->ip;
+	args.geo = mp->m_dir_geo;
+	args.trans = sc->tp;
+	error = xfs_dir2_isblock(&args, &is_block);
+	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, lblk, &error))
+		goto out;
+
+	/* Iterate all the data extents in the directory... */
+	found = xfs_iext_lookup_extent(sc->ip, ifp, lblk, &icur, &got);
+	while (found) {
+		/* Block directories only have a single block at offset 0. */
+		if (is_block &&
+		    (got.br_startoff > 0 ||
+		     got.br_blockcount != args.geo->fsbcount)) {
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK,
+					got.br_startoff);
+			break;
+		}
+
+		/* No more data blocks... */
+		if (got.br_startoff >= leaf_lblk)
+			break;
+
+		/*
+		 * Check each data block's bestfree data.
+		 *
+		 * Iterate all the fsbcount-aligned block offsets in
+		 * this directory.  The directory block reading code is
+		 * smart enough to do its own bmap lookups to handle
+		 * discontiguous directory blocks.  When we're done
+		 * with the extent record, re-query the bmap at the
+		 * next fsbcount-aligned offset to avoid redundant
+		 * block checks.
+		 */
+		for (lblk = roundup((xfs_dablk_t)got.br_startoff,
+				args.geo->fsbcount);
+		     lblk < got.br_startoff + got.br_blockcount;
+		     lblk += args.geo->fsbcount) {
+			error = xfs_scrub_directory_data_bestfree(sc, lblk,
+					is_block);
+			if (error)
+				goto out;
+		}
+		dabno = got.br_startoff + got.br_blockcount;
+		lblk = roundup(dabno, args.geo->fsbcount);
+		found = xfs_iext_lookup_extent(sc->ip, ifp, lblk, &icur, &got);
+	}
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	/* Look for a leaf1 block, which has free info. */
+	if (xfs_iext_lookup_extent(sc->ip, ifp, leaf_lblk, &icur, &got) &&
+	    got.br_startoff == leaf_lblk &&
+	    got.br_blockcount == args.geo->fsbcount &&
+	    !xfs_iext_next_extent(ifp, &icur, &got)) {
+		if (is_block) {
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+			goto out;
+		}
+		error = xfs_scrub_directory_leaf1_bestfree(sc, &args,
+				leaf_lblk);
+		if (error)
+			goto out;
+	}
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	/* Scan for free blocks */
+	lblk = free_lblk;
+	found = xfs_iext_lookup_extent(sc->ip, ifp, lblk, &icur, &got);
+	while (found) {
+		/*
+		 * Dirs can't have blocks mapped above 2^32.
+		 * Single-block dirs shouldn't even be here.
+		 */
+		lblk = got.br_startoff;
+		if (lblk & ~0xFFFFFFFFULL) {
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+			goto out;
+		}
+		if (is_block) {
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, lblk);
+			goto out;
+		}
+
+		/*
+		 * Check each dir free block's bestfree data.
+		 *
+		 * Iterate all the fsbcount-aligned block offsets in
+		 * this directory.  The directory block reading code is
+		 * smart enough to do its own bmap lookups to handle
+		 * discontiguous directory blocks.  When we're done
+		 * with the extent record, re-query the bmap at the
+		 * next fsbcount-aligned offset to avoid redundant
+		 * block checks.
+		 */
+		for (lblk = roundup((xfs_dablk_t)got.br_startoff,
+				args.geo->fsbcount);
+		     lblk < got.br_startoff + got.br_blockcount;
+		     lblk += args.geo->fsbcount) {
+			error = xfs_scrub_directory_free_bestfree(sc, &args,
+					lblk);
+			if (error)
+				goto out;
+		}
+		dabno = got.br_startoff + got.br_blockcount;
+		lblk = roundup(dabno, args.geo->fsbcount);
+		found = xfs_iext_lookup_extent(sc->ip, ifp, lblk, &icur, &got);
+	}
+out:
+	return error;
+}
+
+/* Scrub a whole directory. */
+int
+xfs_scrub_directory(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_scrub_dir_ctx	sdc = {
+		.dir_iter.actor = xfs_scrub_dir_actor,
+		.dir_iter.pos = 0,
+		.sc = sc,
+	};
+	size_t				bufsize;
+	loff_t				oldpos;
+	int				error = 0;
+
+	if (!S_ISDIR(VFS_I(sc->ip)->i_mode))
+		return -ENOENT;
+
+	/* Plausible size? */
+	if (sc->ip->i_d.di_size < xfs_dir2_sf_hdr_size(0)) {
+		xfs_scrub_ino_set_corrupt(sc, sc->ip->i_ino);
+		goto out;
+	}
+
+	/* Check directory tree structure */
+	error = xfs_scrub_da_btree(sc, XFS_DATA_FORK, xfs_scrub_dir_rec, NULL);
+	if (error)
+		return error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return error;
+
+	/* Check the freespace. */
+	error = xfs_scrub_directory_blocks(sc);
+	if (error)
+		return error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return error;
+
+	/*
+	 * Check that every dirent we see can also be looked up by hash.
+	 * Userspace usually asks for a 32k buffer, so we will too.
+	 */
+	bufsize = (size_t)min_t(loff_t, XFS_READDIR_BUFSIZE,
+			sc->ip->i_d.di_size);
+
+	/*
+	 * Look up every name in this directory by hash.
+	 *
+	 * Use the xfs_readdir function to call xfs_scrub_dir_actor on
+	 * every directory entry in this directory.  In _actor, we check
+	 * the name, inode number, and ftype (if applicable) of the
+	 * entry.  xfs_readdir uses the VFS filldir functions to provide
+	 * iteration context.
+	 *
+	 * The VFS grabs a read or write lock via i_rwsem before it reads
+	 * or writes to a directory.  If we've gotten this far we've
+	 * already obtained IOLOCK_EXCL, which (since 4.10) is the same as
+	 * getting a write lock on i_rwsem.  Therefore, it is safe for us
+	 * to drop the ILOCK here in order to reuse the _readdir and
+	 * _dir_lookup routines, which do their own ILOCK locking.
+	 */
+	oldpos = 0;
+	sc->ilock_flags &= ~XFS_ILOCK_EXCL;
+	xfs_iunlock(sc->ip, XFS_ILOCK_EXCL);
+	while (true) {
+		error = xfs_readdir(sc->tp, sc->ip, &sdc.dir_iter, bufsize);
+		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, 0,
+				&error))
+			goto out;
+		if (oldpos == sdc.dir_iter.pos)
+			break;
+		oldpos = sdc.dir_iter.pos;
+	}
+
+out:
+	return error;
+}
diff --git a/fs/xfs/scrub/ialloc.c b/fs/xfs/scrub/ialloc.c
new file mode 100644
index 000000000000..106ca4bd753f
--- /dev/null
+++ b/fs/xfs/scrub/ialloc.c
@@ -0,0 +1,528 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_alloc.h"
+#include "xfs_ialloc.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_icache.h"
+#include "xfs_rmap.h"
+#include "xfs_log.h"
+#include "xfs_trans_priv.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+
+/*
+ * Set us up to scrub inode btrees.
+ * If we detect a discrepancy between the inobt and the inode,
+ * try again after forcing logged inode cores out to disk.
+ */
+int
+xfs_scrub_setup_ag_iallocbt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	return xfs_scrub_setup_ag_btree(sc, ip, sc->try_harder);
+}
+
+/* Inode btree scrubber. */
+
+/*
+ * If we're checking the finobt, cross-reference with the inobt.
+ * Otherwise we're checking the inobt; if there is an finobt, make sure
+ * we have a record or not depending on freecount.
+ */
+static inline void
+xfs_scrub_iallocbt_chunk_xref_other(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inobt_rec_incore	*irec,
+	xfs_agino_t			agino)
+{
+	struct xfs_btree_cur		**pcur;
+	bool				has_irec;
+	int				error;
+
+	if (sc->sm->sm_type == XFS_SCRUB_TYPE_FINOBT)
+		pcur = &sc->sa.ino_cur;
+	else
+		pcur = &sc->sa.fino_cur;
+	if (!(*pcur))
+		return;
+	error = xfs_ialloc_has_inode_record(*pcur, agino, agino, &has_irec);
+	if (!xfs_scrub_should_check_xref(sc, &error, pcur))
+		return;
+	if (((irec->ir_freecount > 0 && !has_irec) ||
+	     (irec->ir_freecount == 0 && has_irec)))
+		xfs_scrub_btree_xref_set_corrupt(sc, *pcur, 0);
+}
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_iallocbt_chunk_xref(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inobt_rec_incore	*irec,
+	xfs_agino_t			agino,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len)
+{
+	struct xfs_owner_info		oinfo;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	xfs_scrub_xref_is_used_space(sc, agbno, len);
+	xfs_scrub_iallocbt_chunk_xref_other(sc, irec, agino);
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INODES);
+	xfs_scrub_xref_is_owned_by(sc, agbno, len, &oinfo);
+	xfs_scrub_xref_is_not_shared(sc, agbno, len);
+}
+
+/* Is this chunk worth checking? */
+STATIC bool
+xfs_scrub_iallocbt_chunk(
+	struct xfs_scrub_btree		*bs,
+	struct xfs_inobt_rec_incore	*irec,
+	xfs_agino_t			agino,
+	xfs_extlen_t			len)
+{
+	struct xfs_mount		*mp = bs->cur->bc_mp;
+	xfs_agnumber_t			agno = bs->cur->bc_private.a.agno;
+	xfs_agblock_t			bno;
+
+	bno = XFS_AGINO_TO_AGBNO(mp, agino);
+	if (bno + len <= bno ||
+	    !xfs_verify_agbno(mp, agno, bno) ||
+	    !xfs_verify_agbno(mp, agno, bno + len - 1))
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	xfs_scrub_iallocbt_chunk_xref(bs->sc, irec, agino, bno, len);
+
+	return true;
+}
+
+/* Count the number of free inodes. */
+static unsigned int
+xfs_scrub_iallocbt_freecount(
+	xfs_inofree_t			freemask)
+{
+	BUILD_BUG_ON(sizeof(freemask) != sizeof(__u64));
+	return hweight64(freemask);
+}
+
+/* Check a particular inode with ir_free. */
+STATIC int
+xfs_scrub_iallocbt_check_cluster_freemask(
+	struct xfs_scrub_btree		*bs,
+	xfs_ino_t			fsino,
+	xfs_agino_t			chunkino,
+	xfs_agino_t			clusterino,
+	struct xfs_inobt_rec_incore	*irec,
+	struct xfs_buf			*bp)
+{
+	struct xfs_dinode		*dip;
+	struct xfs_mount		*mp = bs->cur->bc_mp;
+	bool				inode_is_free = false;
+	bool				freemask_ok;
+	bool				inuse;
+	int				error = 0;
+
+	if (xfs_scrub_should_terminate(bs->sc, &error))
+		return error;
+
+	dip = xfs_buf_offset(bp, clusterino * mp->m_sb.sb_inodesize);
+	if (be16_to_cpu(dip->di_magic) != XFS_DINODE_MAGIC ||
+	    (dip->di_version >= 3 &&
+	     be64_to_cpu(dip->di_ino) != fsino + clusterino)) {
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+		goto out;
+	}
+
+	if (irec->ir_free & XFS_INOBT_MASK(chunkino + clusterino))
+		inode_is_free = true;
+	error = xfs_icache_inode_is_allocated(mp, bs->cur->bc_tp,
+			fsino + clusterino, &inuse);
+	if (error == -ENODATA) {
+		/* Not cached, just read the disk buffer */
+		freemask_ok = inode_is_free ^ !!(dip->di_mode);
+		if (!bs->sc->try_harder && !freemask_ok)
+			return -EDEADLOCK;
+	} else if (error < 0) {
+		/*
+		 * Inode is only half assembled, or there was an IO error,
+		 * or the verifier failed, so don't bother trying to check.
+		 * The inode scrubber can deal with this.
+		 */
+		goto out;
+	} else {
+		/* Inode is all there. */
+		freemask_ok = inode_is_free ^ inuse;
+	}
+	if (!freemask_ok)
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+out:
+	return 0;
+}
+
+/* Make sure the free mask is consistent with what the inodes think. */
+STATIC int
+xfs_scrub_iallocbt_check_freemask(
+	struct xfs_scrub_btree		*bs,
+	struct xfs_inobt_rec_incore	*irec)
+{
+	struct xfs_owner_info		oinfo;
+	struct xfs_imap			imap;
+	struct xfs_mount		*mp = bs->cur->bc_mp;
+	struct xfs_dinode		*dip;
+	struct xfs_buf			*bp;
+	xfs_ino_t			fsino;
+	xfs_agino_t			nr_inodes;
+	xfs_agino_t			agino;
+	xfs_agino_t			chunkino;
+	xfs_agino_t			clusterino;
+	xfs_agblock_t			agbno;
+	int				blks_per_cluster;
+	uint16_t			holemask;
+	uint16_t			ir_holemask;
+	int				error = 0;
+
+	/* Make sure the freemask matches the inode records. */
+	blks_per_cluster = xfs_icluster_size_fsb(mp);
+	nr_inodes = XFS_OFFBNO_TO_AGINO(mp, blks_per_cluster, 0);
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INODES);
+
+	for (agino = irec->ir_startino;
+	     agino < irec->ir_startino + XFS_INODES_PER_CHUNK;
+	     agino += blks_per_cluster * mp->m_sb.sb_inopblock) {
+		fsino = XFS_AGINO_TO_INO(mp, bs->cur->bc_private.a.agno, agino);
+		chunkino = agino - irec->ir_startino;
+		agbno = XFS_AGINO_TO_AGBNO(mp, agino);
+
+		/* Compute the holemask mask for this cluster. */
+		for (clusterino = 0, holemask = 0; clusterino < nr_inodes;
+		     clusterino += XFS_INODES_PER_HOLEMASK_BIT)
+			holemask |= XFS_INOBT_MASK((chunkino + clusterino) /
+					XFS_INODES_PER_HOLEMASK_BIT);
+
+		/* The whole cluster must be a hole or not a hole. */
+		ir_holemask = (irec->ir_holemask & holemask);
+		if (ir_holemask != holemask && ir_holemask != 0) {
+			xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+			continue;
+		}
+
+		/* If any part of this is a hole, skip it. */
+		if (ir_holemask) {
+			xfs_scrub_xref_is_not_owned_by(bs->sc, agbno,
+					blks_per_cluster, &oinfo);
+			continue;
+		}
+
+		xfs_scrub_xref_is_owned_by(bs->sc, agbno, blks_per_cluster,
+				&oinfo);
+
+		/* Grab the inode cluster buffer. */
+		imap.im_blkno = XFS_AGB_TO_DADDR(mp, bs->cur->bc_private.a.agno,
+				agbno);
+		imap.im_len = XFS_FSB_TO_BB(mp, blks_per_cluster);
+		imap.im_boffset = 0;
+
+		error = xfs_imap_to_bp(mp, bs->cur->bc_tp, &imap,
+				&dip, &bp, 0, 0);
+		if (!xfs_scrub_btree_xref_process_error(bs->sc, bs->cur, 0,
+				&error))
+			continue;
+
+		/* Which inodes are free? */
+		for (clusterino = 0; clusterino < nr_inodes; clusterino++) {
+			error = xfs_scrub_iallocbt_check_cluster_freemask(bs,
+					fsino, chunkino, clusterino, irec, bp);
+			if (error) {
+				xfs_trans_brelse(bs->cur->bc_tp, bp);
+				return error;
+			}
+		}
+
+		xfs_trans_brelse(bs->cur->bc_tp, bp);
+	}
+
+	return error;
+}
+
+/* Scrub an inobt/finobt record. */
+STATIC int
+xfs_scrub_iallocbt_rec(
+	struct xfs_scrub_btree		*bs,
+	union xfs_btree_rec		*rec)
+{
+	struct xfs_mount		*mp = bs->cur->bc_mp;
+	xfs_filblks_t			*inode_blocks = bs->private;
+	struct xfs_inobt_rec_incore	irec;
+	uint64_t			holes;
+	xfs_agnumber_t			agno = bs->cur->bc_private.a.agno;
+	xfs_agino_t			agino;
+	xfs_agblock_t			agbno;
+	xfs_extlen_t			len;
+	int				holecount;
+	int				i;
+	int				error = 0;
+	unsigned int			real_freecount;
+	uint16_t			holemask;
+
+	xfs_inobt_btrec_to_irec(mp, rec, &irec);
+
+	if (irec.ir_count > XFS_INODES_PER_CHUNK ||
+	    irec.ir_freecount > XFS_INODES_PER_CHUNK)
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	real_freecount = irec.ir_freecount +
+			(XFS_INODES_PER_CHUNK - irec.ir_count);
+	if (real_freecount != xfs_scrub_iallocbt_freecount(irec.ir_free))
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	agino = irec.ir_startino;
+	/* Record has to be properly aligned within the AG. */
+	if (!xfs_verify_agino(mp, agno, agino) ||
+	    !xfs_verify_agino(mp, agno, agino + XFS_INODES_PER_CHUNK - 1)) {
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+		goto out;
+	}
+
+	/* Make sure this record is aligned to cluster and inoalignmnt size. */
+	agbno = XFS_AGINO_TO_AGBNO(mp, irec.ir_startino);
+	if ((agbno & (xfs_ialloc_cluster_alignment(mp) - 1)) ||
+	    (agbno & (xfs_icluster_size_fsb(mp) - 1)))
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	*inode_blocks += XFS_B_TO_FSB(mp,
+			irec.ir_count * mp->m_sb.sb_inodesize);
+
+	/* Handle non-sparse inodes */
+	if (!xfs_inobt_issparse(irec.ir_holemask)) {
+		len = XFS_B_TO_FSB(mp,
+				XFS_INODES_PER_CHUNK * mp->m_sb.sb_inodesize);
+		if (irec.ir_count != XFS_INODES_PER_CHUNK)
+			xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+		if (!xfs_scrub_iallocbt_chunk(bs, &irec, agino, len))
+			goto out;
+		goto check_freemask;
+	}
+
+	/* Check each chunk of a sparse inode cluster. */
+	holemask = irec.ir_holemask;
+	holecount = 0;
+	len = XFS_B_TO_FSB(mp,
+			XFS_INODES_PER_HOLEMASK_BIT * mp->m_sb.sb_inodesize);
+	holes = ~xfs_inobt_irec_to_allocmask(&irec);
+	if ((holes & irec.ir_free) != holes ||
+	    irec.ir_freecount > irec.ir_count)
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	for (i = 0; i < XFS_INOBT_HOLEMASK_BITS; i++) {
+		if (holemask & 1)
+			holecount += XFS_INODES_PER_HOLEMASK_BIT;
+		else if (!xfs_scrub_iallocbt_chunk(bs, &irec, agino, len))
+			break;
+		holemask >>= 1;
+		agino += XFS_INODES_PER_HOLEMASK_BIT;
+	}
+
+	if (holecount > XFS_INODES_PER_CHUNK ||
+	    holecount + irec.ir_count != XFS_INODES_PER_CHUNK)
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+check_freemask:
+	error = xfs_scrub_iallocbt_check_freemask(bs, &irec);
+	if (error)
+		goto out;
+
+out:
+	return error;
+}
+
+/*
+ * Make sure the inode btrees are as large as the rmap thinks they are.
+ * Don't bother if we're missing btree cursors, as we're already corrupt.
+ */
+STATIC void
+xfs_scrub_iallocbt_xref_rmap_btreeblks(
+	struct xfs_scrub_context	*sc,
+	int				which)
+{
+	struct xfs_owner_info		oinfo;
+	xfs_filblks_t			blocks;
+	xfs_extlen_t			inobt_blocks = 0;
+	xfs_extlen_t			finobt_blocks = 0;
+	int				error;
+
+	if (!sc->sa.ino_cur || !sc->sa.rmap_cur ||
+	    (xfs_sb_version_hasfinobt(&sc->mp->m_sb) && !sc->sa.fino_cur))
+		return;
+
+	/* Check that we saw as many inobt blocks as the rmap says. */
+	error = xfs_btree_count_blocks(sc->sa.ino_cur, &inobt_blocks);
+	if (!xfs_scrub_process_error(sc, 0, 0, &error))
+		return;
+
+	if (sc->sa.fino_cur) {
+		error = xfs_btree_count_blocks(sc->sa.fino_cur, &finobt_blocks);
+		if (!xfs_scrub_process_error(sc, 0, 0, &error))
+			return;
+	}
+
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INOBT);
+	error = xfs_scrub_count_rmap_ownedby_ag(sc, sc->sa.rmap_cur, &oinfo,
+			&blocks);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.rmap_cur))
+		return;
+	if (blocks != inobt_blocks + finobt_blocks)
+		xfs_scrub_btree_set_corrupt(sc, sc->sa.ino_cur, 0);
+}
+
+/*
+ * Make sure that the inobt records point to the same number of blocks as
+ * the rmap says are owned by inodes.
+ */
+STATIC void
+xfs_scrub_iallocbt_xref_rmap_inodes(
+	struct xfs_scrub_context	*sc,
+	int				which,
+	xfs_filblks_t			inode_blocks)
+{
+	struct xfs_owner_info		oinfo;
+	xfs_filblks_t			blocks;
+	int				error;
+
+	if (!sc->sa.rmap_cur)
+		return;
+
+	/* Check that we saw as many inode blocks as the rmap knows about. */
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INODES);
+	error = xfs_scrub_count_rmap_ownedby_ag(sc, sc->sa.rmap_cur, &oinfo,
+			&blocks);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.rmap_cur))
+		return;
+	if (blocks != inode_blocks)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.rmap_cur, 0);
+}
+
+/* Scrub the inode btrees for some AG. */
+STATIC int
+xfs_scrub_iallocbt(
+	struct xfs_scrub_context	*sc,
+	xfs_btnum_t			which)
+{
+	struct xfs_btree_cur		*cur;
+	struct xfs_owner_info		oinfo;
+	xfs_filblks_t			inode_blocks = 0;
+	int				error;
+
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INOBT);
+	cur = which == XFS_BTNUM_INO ? sc->sa.ino_cur : sc->sa.fino_cur;
+	error = xfs_scrub_btree(sc, cur, xfs_scrub_iallocbt_rec, &oinfo,
+			&inode_blocks);
+	if (error)
+		return error;
+
+	xfs_scrub_iallocbt_xref_rmap_btreeblks(sc, which);
+
+	/*
+	 * If we're scrubbing the inode btree, inode_blocks is the number of
+	 * blocks pointed to by all the inode chunk records.  Therefore, we
+	 * should compare to the number of inode chunk blocks that the rmap
+	 * knows about.  We can't do this for the finobt since it only points
+	 * to inode chunks with free inodes.
+	 */
+	if (which == XFS_BTNUM_INO)
+		xfs_scrub_iallocbt_xref_rmap_inodes(sc, which, inode_blocks);
+
+	return error;
+}
+
+int
+xfs_scrub_inobt(
+	struct xfs_scrub_context	*sc)
+{
+	return xfs_scrub_iallocbt(sc, XFS_BTNUM_INO);
+}
+
+int
+xfs_scrub_finobt(
+	struct xfs_scrub_context	*sc)
+{
+	return xfs_scrub_iallocbt(sc, XFS_BTNUM_FINO);
+}
+
+/* See if an inode btree has (or doesn't have) an inode chunk record. */
+static inline void
+xfs_scrub_xref_inode_check(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len,
+	struct xfs_btree_cur		**icur,
+	bool				should_have_inodes)
+{
+	bool				has_inodes;
+	int				error;
+
+	if (!(*icur))
+		return;
+
+	error = xfs_ialloc_has_inodes_at_extent(*icur, agbno, len, &has_inodes);
+	if (!xfs_scrub_should_check_xref(sc, &error, icur))
+		return;
+	if (has_inodes != should_have_inodes)
+		xfs_scrub_btree_xref_set_corrupt(sc, *icur, 0);
+}
+
+/* xref check that the extent is not covered by inodes */
+void
+xfs_scrub_xref_is_not_inode_chunk(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len)
+{
+	xfs_scrub_xref_inode_check(sc, agbno, len, &sc->sa.ino_cur, false);
+	xfs_scrub_xref_inode_check(sc, agbno, len, &sc->sa.fino_cur, false);
+}
+
+/* xref check that the extent is covered by inodes */
+void
+xfs_scrub_xref_is_inode_chunk(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len)
+{
+	xfs_scrub_xref_inode_check(sc, agbno, len, &sc->sa.ino_cur, true);
+}
diff --git a/fs/xfs/scrub/inode.c b/fs/xfs/scrub/inode.c
new file mode 100644
index 000000000000..df14930e4fc5
--- /dev/null
+++ b/fs/xfs/scrub/inode.c
@@ -0,0 +1,611 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_inode_buf.h"
+#include "xfs_inode_fork.h"
+#include "xfs_ialloc.h"
+#include "xfs_da_format.h"
+#include "xfs_reflink.h"
+#include "xfs_rmap.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_util.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+
+/*
+ * Grab total control of the inode metadata.  It doesn't matter here if
+ * the file data is still changing; exclusive access to the metadata is
+ * the goal.
+ */
+int
+xfs_scrub_setup_inode(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	struct xfs_mount		*mp = sc->mp;
+	int				error;
+
+	/*
+	 * Try to get the inode.  If the verifiers fail, we try again
+	 * in raw mode.
+	 */
+	error = xfs_scrub_get_inode(sc, ip);
+	switch (error) {
+	case 0:
+		break;
+	case -EFSCORRUPTED:
+	case -EFSBADCRC:
+		return xfs_scrub_trans_alloc(sc->sm, mp, &sc->tp);
+	default:
+		return error;
+	}
+
+	/* Got the inode, lock it and we're ready to go. */
+	sc->ilock_flags = XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL;
+	xfs_ilock(sc->ip, sc->ilock_flags);
+	error = xfs_scrub_trans_alloc(sc->sm, mp, &sc->tp);
+	if (error)
+		goto out;
+	sc->ilock_flags |= XFS_ILOCK_EXCL;
+	xfs_ilock(sc->ip, XFS_ILOCK_EXCL);
+
+out:
+	/* scrub teardown will unlock and release the inode for us */
+	return error;
+}
+
+/* Inode core */
+
+/* Validate di_extsize hint. */
+STATIC void
+xfs_scrub_inode_extsize(
+	struct xfs_scrub_context	*sc,
+	struct xfs_dinode		*dip,
+	xfs_ino_t			ino,
+	uint16_t			mode,
+	uint16_t			flags)
+{
+	xfs_failaddr_t			fa;
+
+	fa = xfs_inode_validate_extsize(sc->mp, be32_to_cpu(dip->di_extsize),
+			mode, flags);
+	if (fa)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+}
+
+/*
+ * Validate di_cowextsize hint.
+ *
+ * The rules are documented at xfs_ioctl_setattr_check_cowextsize().
+ * These functions must be kept in sync with each other.
+ */
+STATIC void
+xfs_scrub_inode_cowextsize(
+	struct xfs_scrub_context	*sc,
+	struct xfs_dinode		*dip,
+	xfs_ino_t			ino,
+	uint16_t			mode,
+	uint16_t			flags,
+	uint64_t			flags2)
+{
+	xfs_failaddr_t			fa;
+
+	fa = xfs_inode_validate_cowextsize(sc->mp,
+			be32_to_cpu(dip->di_cowextsize), mode, flags,
+			flags2);
+	if (fa)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+}
+
+/* Make sure the di_flags make sense for the inode. */
+STATIC void
+xfs_scrub_inode_flags(
+	struct xfs_scrub_context	*sc,
+	struct xfs_dinode		*dip,
+	xfs_ino_t			ino,
+	uint16_t			mode,
+	uint16_t			flags)
+{
+	struct xfs_mount		*mp = sc->mp;
+
+	if (flags & ~XFS_DIFLAG_ANY)
+		goto bad;
+
+	/* rt flags require rt device */
+	if ((flags & (XFS_DIFLAG_REALTIME | XFS_DIFLAG_RTINHERIT)) &&
+	    !mp->m_rtdev_targp)
+		goto bad;
+
+	/* new rt bitmap flag only valid for rbmino */
+	if ((flags & XFS_DIFLAG_NEWRTBM) && ino != mp->m_sb.sb_rbmino)
+		goto bad;
+
+	/* directory-only flags */
+	if ((flags & (XFS_DIFLAG_RTINHERIT |
+		     XFS_DIFLAG_EXTSZINHERIT |
+		     XFS_DIFLAG_PROJINHERIT |
+		     XFS_DIFLAG_NOSYMLINKS)) &&
+	    !S_ISDIR(mode))
+		goto bad;
+
+	/* file-only flags */
+	if ((flags & (XFS_DIFLAG_REALTIME | FS_XFLAG_EXTSIZE)) &&
+	    !S_ISREG(mode))
+		goto bad;
+
+	/* filestreams and rt make no sense */
+	if ((flags & XFS_DIFLAG_FILESTREAM) && (flags & XFS_DIFLAG_REALTIME))
+		goto bad;
+
+	return;
+bad:
+	xfs_scrub_ino_set_corrupt(sc, ino);
+}
+
+/* Make sure the di_flags2 make sense for the inode. */
+STATIC void
+xfs_scrub_inode_flags2(
+	struct xfs_scrub_context	*sc,
+	struct xfs_dinode		*dip,
+	xfs_ino_t			ino,
+	uint16_t			mode,
+	uint16_t			flags,
+	uint64_t			flags2)
+{
+	struct xfs_mount		*mp = sc->mp;
+
+	if (flags2 & ~XFS_DIFLAG2_ANY)
+		goto bad;
+
+	/* reflink flag requires reflink feature */
+	if ((flags2 & XFS_DIFLAG2_REFLINK) &&
+	    !xfs_sb_version_hasreflink(&mp->m_sb))
+		goto bad;
+
+	/* cowextsize flag is checked w.r.t. mode separately */
+
+	/* file/dir-only flags */
+	if ((flags2 & XFS_DIFLAG2_DAX) && !(S_ISREG(mode) || S_ISDIR(mode)))
+		goto bad;
+
+	/* file-only flags */
+	if ((flags2 & XFS_DIFLAG2_REFLINK) && !S_ISREG(mode))
+		goto bad;
+
+	/* realtime and reflink make no sense, currently */
+	if ((flags & XFS_DIFLAG_REALTIME) && (flags2 & XFS_DIFLAG2_REFLINK))
+		goto bad;
+
+	/* dax and reflink make no sense, currently */
+	if ((flags2 & XFS_DIFLAG2_DAX) && (flags2 & XFS_DIFLAG2_REFLINK))
+		goto bad;
+
+	return;
+bad:
+	xfs_scrub_ino_set_corrupt(sc, ino);
+}
+
+/* Scrub all the ondisk inode fields. */
+STATIC void
+xfs_scrub_dinode(
+	struct xfs_scrub_context	*sc,
+	struct xfs_dinode		*dip,
+	xfs_ino_t			ino)
+{
+	struct xfs_mount		*mp = sc->mp;
+	size_t				fork_recs;
+	unsigned long long		isize;
+	uint64_t			flags2;
+	uint32_t			nextents;
+	uint16_t			flags;
+	uint16_t			mode;
+
+	flags = be16_to_cpu(dip->di_flags);
+	if (dip->di_version >= 3)
+		flags2 = be64_to_cpu(dip->di_flags2);
+	else
+		flags2 = 0;
+
+	/* di_mode */
+	mode = be16_to_cpu(dip->di_mode);
+	switch (mode & S_IFMT) {
+	case S_IFLNK:
+	case S_IFREG:
+	case S_IFDIR:
+	case S_IFCHR:
+	case S_IFBLK:
+	case S_IFIFO:
+	case S_IFSOCK:
+		/* mode is recognized */
+		break;
+	default:
+		xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	}
+
+	/* v1/v2 fields */
+	switch (dip->di_version) {
+	case 1:
+		/*
+		 * We autoconvert v1 inodes into v2 inodes on writeout,
+		 * so just mark this inode for preening.
+		 */
+		xfs_scrub_ino_set_preen(sc, ino);
+		break;
+	case 2:
+	case 3:
+		if (dip->di_onlink != 0)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+
+		if (dip->di_mode == 0 && sc->ip)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+
+		if (dip->di_projid_hi != 0 &&
+		    !xfs_sb_version_hasprojid32bit(&mp->m_sb))
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	default:
+		xfs_scrub_ino_set_corrupt(sc, ino);
+		return;
+	}
+
+	/*
+	 * di_uid/di_gid -- -1 isn't invalid, but there's no way that
+	 * userspace could have created that.
+	 */
+	if (dip->di_uid == cpu_to_be32(-1U) ||
+	    dip->di_gid == cpu_to_be32(-1U))
+		xfs_scrub_ino_set_warning(sc, ino);
+
+	/* di_format */
+	switch (dip->di_format) {
+	case XFS_DINODE_FMT_DEV:
+		if (!S_ISCHR(mode) && !S_ISBLK(mode) &&
+		    !S_ISFIFO(mode) && !S_ISSOCK(mode))
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	case XFS_DINODE_FMT_LOCAL:
+		if (!S_ISDIR(mode) && !S_ISLNK(mode))
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	case XFS_DINODE_FMT_EXTENTS:
+		if (!S_ISREG(mode) && !S_ISDIR(mode) && !S_ISLNK(mode))
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	case XFS_DINODE_FMT_BTREE:
+		if (!S_ISREG(mode) && !S_ISDIR(mode))
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	case XFS_DINODE_FMT_UUID:
+	default:
+		xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	}
+
+	/* di_[amc]time.nsec */
+	if (be32_to_cpu(dip->di_atime.t_nsec) >= NSEC_PER_SEC)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+	if (be32_to_cpu(dip->di_mtime.t_nsec) >= NSEC_PER_SEC)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+	if (be32_to_cpu(dip->di_ctime.t_nsec) >= NSEC_PER_SEC)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+
+	/*
+	 * di_size.  xfs_dinode_verify checks for things that screw up
+	 * the VFS such as the upper bit being set and zero-length
+	 * symlinks/directories, but we can do more here.
+	 */
+	isize = be64_to_cpu(dip->di_size);
+	if (isize & (1ULL << 63))
+		xfs_scrub_ino_set_corrupt(sc, ino);
+
+	/* Devices, fifos, and sockets must have zero size */
+	if (!S_ISDIR(mode) && !S_ISREG(mode) && !S_ISLNK(mode) && isize != 0)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+
+	/* Directories can't be larger than the data section size (32G) */
+	if (S_ISDIR(mode) && (isize == 0 || isize >= XFS_DIR2_SPACE_SIZE))
+		xfs_scrub_ino_set_corrupt(sc, ino);
+
+	/* Symlinks can't be larger than SYMLINK_MAXLEN */
+	if (S_ISLNK(mode) && (isize == 0 || isize >= XFS_SYMLINK_MAXLEN))
+		xfs_scrub_ino_set_corrupt(sc, ino);
+
+	/*
+	 * Warn if the running kernel can't handle the kinds of offsets
+	 * needed to deal with the file size.  In other words, if the
+	 * pagecache can't cache all the blocks in this file due to
+	 * overly large offsets, flag the inode for admin review.
+	 */
+	if (isize >= mp->m_super->s_maxbytes)
+		xfs_scrub_ino_set_warning(sc, ino);
+
+	/* di_nblocks */
+	if (flags2 & XFS_DIFLAG2_REFLINK) {
+		; /* nblocks can exceed dblocks */
+	} else if (flags & XFS_DIFLAG_REALTIME) {
+		/*
+		 * nblocks is the sum of data extents (in the rtdev),
+		 * attr extents (in the datadev), and both forks' bmbt
+		 * blocks (in the datadev).  This clumsy check is the
+		 * best we can do without cross-referencing with the
+		 * inode forks.
+		 */
+		if (be64_to_cpu(dip->di_nblocks) >=
+		    mp->m_sb.sb_dblocks + mp->m_sb.sb_rblocks)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+	} else {
+		if (be64_to_cpu(dip->di_nblocks) >= mp->m_sb.sb_dblocks)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+	}
+
+	xfs_scrub_inode_flags(sc, dip, ino, mode, flags);
+
+	xfs_scrub_inode_extsize(sc, dip, ino, mode, flags);
+
+	/* di_nextents */
+	nextents = be32_to_cpu(dip->di_nextents);
+	fork_recs =  XFS_DFORK_DSIZE(dip, mp) / sizeof(struct xfs_bmbt_rec);
+	switch (dip->di_format) {
+	case XFS_DINODE_FMT_EXTENTS:
+		if (nextents > fork_recs)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	case XFS_DINODE_FMT_BTREE:
+		if (nextents <= fork_recs)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	default:
+		if (nextents != 0)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	}
+
+	/* di_forkoff */
+	if (XFS_DFORK_APTR(dip) >= (char *)dip + mp->m_sb.sb_inodesize)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+	if (dip->di_anextents != 0 && dip->di_forkoff == 0)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+	if (dip->di_forkoff == 0 && dip->di_aformat != XFS_DINODE_FMT_EXTENTS)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+
+	/* di_aformat */
+	if (dip->di_aformat != XFS_DINODE_FMT_LOCAL &&
+	    dip->di_aformat != XFS_DINODE_FMT_EXTENTS &&
+	    dip->di_aformat != XFS_DINODE_FMT_BTREE)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+
+	/* di_anextents */
+	nextents = be16_to_cpu(dip->di_anextents);
+	fork_recs =  XFS_DFORK_ASIZE(dip, mp) / sizeof(struct xfs_bmbt_rec);
+	switch (dip->di_aformat) {
+	case XFS_DINODE_FMT_EXTENTS:
+		if (nextents > fork_recs)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	case XFS_DINODE_FMT_BTREE:
+		if (nextents <= fork_recs)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		break;
+	default:
+		if (nextents != 0)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+	}
+
+	if (dip->di_version >= 3) {
+		if (be32_to_cpu(dip->di_crtime.t_nsec) >= NSEC_PER_SEC)
+			xfs_scrub_ino_set_corrupt(sc, ino);
+		xfs_scrub_inode_flags2(sc, dip, ino, mode, flags, flags2);
+		xfs_scrub_inode_cowextsize(sc, dip, ino, mode, flags,
+				flags2);
+	}
+}
+
+/*
+ * Make sure the finobt doesn't think this inode is free.
+ * We don't have to check the inobt ourselves because we got the inode via
+ * IGET_UNTRUSTED, which checks the inobt for us.
+ */
+static void
+xfs_scrub_inode_xref_finobt(
+	struct xfs_scrub_context	*sc,
+	xfs_ino_t			ino)
+{
+	struct xfs_inobt_rec_incore	rec;
+	xfs_agino_t			agino;
+	int				has_record;
+	int				error;
+
+	if (!sc->sa.fino_cur)
+		return;
+
+	agino = XFS_INO_TO_AGINO(sc->mp, ino);
+
+	/*
+	 * Try to get the finobt record.  If we can't get it, then we're
+	 * in good shape.
+	 */
+	error = xfs_inobt_lookup(sc->sa.fino_cur, agino, XFS_LOOKUP_LE,
+			&has_record);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.fino_cur) ||
+	    !has_record)
+		return;
+
+	error = xfs_inobt_get_rec(sc->sa.fino_cur, &rec, &has_record);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.fino_cur) ||
+	    !has_record)
+		return;
+
+	/*
+	 * Otherwise, make sure this record either doesn't cover this inode,
+	 * or that it does but it's marked present.
+	 */
+	if (rec.ir_startino > agino ||
+	    rec.ir_startino + XFS_INODES_PER_CHUNK <= agino)
+		return;
+
+	if (rec.ir_free & XFS_INOBT_MASK(agino - rec.ir_startino))
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.fino_cur, 0);
+}
+
+/* Cross reference the inode fields with the forks. */
+STATIC void
+xfs_scrub_inode_xref_bmap(
+	struct xfs_scrub_context	*sc,
+	struct xfs_dinode		*dip)
+{
+	xfs_extnum_t			nextents;
+	xfs_filblks_t			count;
+	xfs_filblks_t			acount;
+	int				error;
+
+	/* Walk all the extents to check nextents/naextents/nblocks. */
+	error = xfs_bmap_count_blocks(sc->tp, sc->ip, XFS_DATA_FORK,
+			&nextents, &count);
+	if (!xfs_scrub_should_check_xref(sc, &error, NULL))
+		return;
+	if (nextents < be32_to_cpu(dip->di_nextents))
+		xfs_scrub_ino_xref_set_corrupt(sc, sc->ip->i_ino);
+
+	error = xfs_bmap_count_blocks(sc->tp, sc->ip, XFS_ATTR_FORK,
+			&nextents, &acount);
+	if (!xfs_scrub_should_check_xref(sc, &error, NULL))
+		return;
+	if (nextents != be16_to_cpu(dip->di_anextents))
+		xfs_scrub_ino_xref_set_corrupt(sc, sc->ip->i_ino);
+
+	/* Check nblocks against the inode. */
+	if (count + acount != be64_to_cpu(dip->di_nblocks))
+		xfs_scrub_ino_xref_set_corrupt(sc, sc->ip->i_ino);
+}
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_inode_xref(
+	struct xfs_scrub_context	*sc,
+	xfs_ino_t			ino,
+	struct xfs_dinode		*dip)
+{
+	struct xfs_owner_info		oinfo;
+	xfs_agnumber_t			agno;
+	xfs_agblock_t			agbno;
+	int				error;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	agno = XFS_INO_TO_AGNO(sc->mp, ino);
+	agbno = XFS_INO_TO_AGBNO(sc->mp, ino);
+
+	error = xfs_scrub_ag_init(sc, agno, &sc->sa);
+	if (!xfs_scrub_xref_process_error(sc, agno, agbno, &error))
+		return;
+
+	xfs_scrub_xref_is_used_space(sc, agbno, 1);
+	xfs_scrub_inode_xref_finobt(sc, ino);
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_INODES);
+	xfs_scrub_xref_is_owned_by(sc, agbno, 1, &oinfo);
+	xfs_scrub_xref_is_not_shared(sc, agbno, 1);
+	xfs_scrub_inode_xref_bmap(sc, dip);
+
+	xfs_scrub_ag_free(sc, &sc->sa);
+}
+
+/*
+ * If the reflink iflag disagrees with a scan for shared data fork extents,
+ * either flag an error (shared extents w/ no flag) or a preen (flag set w/o
+ * any shared extents).  We already checked for reflink iflag set on a non
+ * reflink filesystem.
+ */
+static void
+xfs_scrub_inode_check_reflink_iflag(
+	struct xfs_scrub_context	*sc,
+	xfs_ino_t			ino)
+{
+	struct xfs_mount		*mp = sc->mp;
+	bool				has_shared;
+	int				error;
+
+	if (!xfs_sb_version_hasreflink(&mp->m_sb))
+		return;
+
+	error = xfs_reflink_inode_has_shared_extents(sc->tp, sc->ip,
+			&has_shared);
+	if (!xfs_scrub_xref_process_error(sc, XFS_INO_TO_AGNO(mp, ino),
+			XFS_INO_TO_AGBNO(mp, ino), &error))
+		return;
+	if (xfs_is_reflink_inode(sc->ip) && !has_shared)
+		xfs_scrub_ino_set_preen(sc, ino);
+	else if (!xfs_is_reflink_inode(sc->ip) && has_shared)
+		xfs_scrub_ino_set_corrupt(sc, ino);
+}
+
+/* Scrub an inode. */
+int
+xfs_scrub_inode(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_dinode		di;
+	int				error = 0;
+
+	/*
+	 * If sc->ip is NULL, that means that the setup function called
+	 * xfs_iget to look up the inode.  xfs_iget returned a EFSCORRUPTED
+	 * and a NULL inode, so flag the corruption error and return.
+	 */
+	if (!sc->ip) {
+		xfs_scrub_ino_set_corrupt(sc, sc->sm->sm_ino);
+		return 0;
+	}
+
+	/* Scrub the inode core. */
+	xfs_inode_to_disk(sc->ip, &di, 0);
+	xfs_scrub_dinode(sc, &di, sc->ip->i_ino);
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	/*
+	 * Look for discrepancies between file's data blocks and the reflink
+	 * iflag.  We already checked the iflag against the file mode when
+	 * we scrubbed the dinode.
+	 */
+	if (S_ISREG(VFS_I(sc->ip)->i_mode))
+		xfs_scrub_inode_check_reflink_iflag(sc, sc->ip->i_ino);
+
+	xfs_scrub_inode_xref(sc, sc->ip->i_ino, &di);
+out:
+	return error;
+}
diff --git a/fs/xfs/scrub/parent.c b/fs/xfs/scrub/parent.c
new file mode 100644
index 000000000000..1fb88c18d455
--- /dev/null
+++ b/fs/xfs/scrub/parent.c
@@ -0,0 +1,327 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_dir2.h"
+#include "xfs_dir2_priv.h"
+#include "xfs_ialloc.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+
+/* Set us up to scrub parents. */
+int
+xfs_scrub_setup_parent(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	return xfs_scrub_setup_inode_contents(sc, ip, 0);
+}
+
+/* Parent pointers */
+
+/* Look for an entry in a parent pointing to this inode. */
+
+struct xfs_scrub_parent_ctx {
+	struct dir_context		dc;
+	xfs_ino_t			ino;
+	xfs_nlink_t			nlink;
+};
+
+/* Look for a single entry in a directory pointing to an inode. */
+STATIC int
+xfs_scrub_parent_actor(
+	struct dir_context		*dc,
+	const char			*name,
+	int				namelen,
+	loff_t				pos,
+	u64				ino,
+	unsigned			type)
+{
+	struct xfs_scrub_parent_ctx	*spc;
+
+	spc = container_of(dc, struct xfs_scrub_parent_ctx, dc);
+	if (spc->ino == ino)
+		spc->nlink++;
+	return 0;
+}
+
+/* Count the number of dentries in the parent dir that point to this inode. */
+STATIC int
+xfs_scrub_parent_count_parent_dentries(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*parent,
+	xfs_nlink_t			*nlink)
+{
+	struct xfs_scrub_parent_ctx	spc = {
+		.dc.actor = xfs_scrub_parent_actor,
+		.dc.pos = 0,
+		.ino = sc->ip->i_ino,
+		.nlink = 0,
+	};
+	size_t				bufsize;
+	loff_t				oldpos;
+	uint				lock_mode;
+	int				error = 0;
+
+	/*
+	 * If there are any blocks, read-ahead block 0 as we're almost
+	 * certain to have the next operation be a read there.  This is
+	 * how we guarantee that the parent's extent map has been loaded,
+	 * if there is one.
+	 */
+	lock_mode = xfs_ilock_data_map_shared(parent);
+	if (parent->i_d.di_nextents > 0)
+		error = xfs_dir3_data_readahead(parent, 0, -1);
+	xfs_iunlock(parent, lock_mode);
+	if (error)
+		return error;
+
+	/*
+	 * Iterate the parent dir to confirm that there is
+	 * exactly one entry pointing back to the inode being
+	 * scanned.
+	 */
+	bufsize = (size_t)min_t(loff_t, XFS_READDIR_BUFSIZE,
+			parent->i_d.di_size);
+	oldpos = 0;
+	while (true) {
+		error = xfs_readdir(sc->tp, parent, &spc.dc, bufsize);
+		if (error)
+			goto out;
+		if (oldpos == spc.dc.pos)
+			break;
+		oldpos = spc.dc.pos;
+	}
+	*nlink = spc.nlink;
+out:
+	return error;
+}
+
+/*
+ * Given the inode number of the alleged parent of the inode being
+ * scrubbed, try to validate that the parent has exactly one directory
+ * entry pointing back to the inode being scrubbed.
+ */
+STATIC int
+xfs_scrub_parent_validate(
+	struct xfs_scrub_context	*sc,
+	xfs_ino_t			dnum,
+	bool				*try_again)
+{
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_inode		*dp = NULL;
+	xfs_nlink_t			expected_nlink;
+	xfs_nlink_t			nlink;
+	int				error = 0;
+
+	*try_again = false;
+
+	/* '..' must not point to ourselves. */
+	if (sc->ip->i_ino == dnum) {
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		goto out;
+	}
+
+	/*
+	 * If we're an unlinked directory, the parent /won't/ have a link
+	 * to us.  Otherwise, it should have one link.
+	 */
+	expected_nlink = VFS_I(sc->ip)->i_nlink == 0 ? 0 : 1;
+
+	/*
+	 * Grab this parent inode.  We release the inode before we
+	 * cancel the scrub transaction.  Since we're don't know a
+	 * priori that releasing the inode won't trigger eofblocks
+	 * cleanup (which allocates what would be a nested transaction)
+	 * if the parent pointer erroneously points to a file, we
+	 * can't use DONTCACHE here because DONTCACHE inodes can trigger
+	 * immediate inactive cleanup of the inode.
+	 *
+	 * If _iget returns -EINVAL then the parent inode number is garbage
+	 * and the directory is corrupt.  If the _iget returns -EFSCORRUPTED
+	 * or -EFSBADCRC then the parent is corrupt which is a cross
+	 * referencing error.  Any other error is an operational error.
+	 */
+	error = xfs_iget(mp, sc->tp, dnum, XFS_IGET_UNTRUSTED, 0, &dp);
+	if (error == -EINVAL) {
+		error = -EFSCORRUPTED;
+		xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, 0, &error);
+		goto out;
+	}
+	if (!xfs_scrub_fblock_xref_process_error(sc, XFS_DATA_FORK, 0, &error))
+		goto out;
+	if (dp == sc->ip || !S_ISDIR(VFS_I(dp)->i_mode)) {
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		goto out_rele;
+	}
+
+	/*
+	 * We prefer to keep the inode locked while we lock and search
+	 * its alleged parent for a forward reference.  If we can grab
+	 * the iolock, validate the pointers and we're done.  We must
+	 * use nowait here to avoid an ABBA deadlock on the parent and
+	 * the child inodes.
+	 */
+	if (xfs_ilock_nowait(dp, XFS_IOLOCK_SHARED)) {
+		error = xfs_scrub_parent_count_parent_dentries(sc, dp, &nlink);
+		if (!xfs_scrub_fblock_xref_process_error(sc, XFS_DATA_FORK, 0,
+				&error))
+			goto out_unlock;
+		if (nlink != expected_nlink)
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		goto out_unlock;
+	}
+
+	/*
+	 * The game changes if we get here.  We failed to lock the parent,
+	 * so we're going to try to verify both pointers while only holding
+	 * one lock so as to avoid deadlocking with something that's actually
+	 * trying to traverse down the directory tree.
+	 */
+	xfs_iunlock(sc->ip, sc->ilock_flags);
+	sc->ilock_flags = 0;
+	xfs_ilock(dp, XFS_IOLOCK_SHARED);
+
+	/* Go looking for our dentry. */
+	error = xfs_scrub_parent_count_parent_dentries(sc, dp, &nlink);
+	if (!xfs_scrub_fblock_xref_process_error(sc, XFS_DATA_FORK, 0, &error))
+		goto out_unlock;
+
+	/* Drop the parent lock, relock this inode. */
+	xfs_iunlock(dp, XFS_IOLOCK_SHARED);
+	sc->ilock_flags = XFS_IOLOCK_EXCL;
+	xfs_ilock(sc->ip, sc->ilock_flags);
+
+	/*
+	 * If we're an unlinked directory, the parent /won't/ have a link
+	 * to us.  Otherwise, it should have one link.  We have to re-set
+	 * it here because we dropped the lock on sc->ip.
+	 */
+	expected_nlink = VFS_I(sc->ip)->i_nlink == 0 ? 0 : 1;
+
+	/* Look up '..' to see if the inode changed. */
+	error = xfs_dir_lookup(sc->tp, sc->ip, &xfs_name_dotdot, &dnum, NULL);
+	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
+		goto out_rele;
+
+	/* Drat, parent changed.  Try again! */
+	if (dnum != dp->i_ino) {
+		iput(VFS_I(dp));
+		*try_again = true;
+		return 0;
+	}
+	iput(VFS_I(dp));
+
+	/*
+	 * '..' didn't change, so check that there was only one entry
+	 * for us in the parent.
+	 */
+	if (nlink != expected_nlink)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+	return error;
+
+out_unlock:
+	xfs_iunlock(dp, XFS_IOLOCK_SHARED);
+out_rele:
+	iput(VFS_I(dp));
+out:
+	return error;
+}
+
+/* Scrub a parent pointer. */
+int
+xfs_scrub_parent(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_mount		*mp = sc->mp;
+	xfs_ino_t			dnum;
+	bool				try_again;
+	int				tries = 0;
+	int				error = 0;
+
+	/*
+	 * If we're a directory, check that the '..' link points up to
+	 * a directory that has one entry pointing to us.
+	 */
+	if (!S_ISDIR(VFS_I(sc->ip)->i_mode))
+		return -ENOENT;
+
+	/* We're not a special inode, are we? */
+	if (!xfs_verify_dir_ino(mp, sc->ip->i_ino)) {
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		goto out;
+	}
+
+	/*
+	 * The VFS grabs a read or write lock via i_rwsem before it reads
+	 * or writes to a directory.  If we've gotten this far we've
+	 * already obtained IOLOCK_EXCL, which (since 4.10) is the same as
+	 * getting a write lock on i_rwsem.  Therefore, it is safe for us
+	 * to drop the ILOCK here in order to do directory lookups.
+	 */
+	sc->ilock_flags &= ~(XFS_ILOCK_EXCL | XFS_MMAPLOCK_EXCL);
+	xfs_iunlock(sc->ip, XFS_ILOCK_EXCL | XFS_MMAPLOCK_EXCL);
+
+	/* Look up '..' */
+	error = xfs_dir_lookup(sc->tp, sc->ip, &xfs_name_dotdot, &dnum, NULL);
+	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
+		goto out;
+	if (!xfs_verify_dir_ino(mp, dnum)) {
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		goto out;
+	}
+
+	/* Is this the root dir?  Then '..' must point to itself. */
+	if (sc->ip == mp->m_rootip) {
+		if (sc->ip->i_ino != mp->m_sb.sb_rootino ||
+		    sc->ip->i_ino != dnum)
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		goto out;
+	}
+
+	do {
+		error = xfs_scrub_parent_validate(sc, dnum, &try_again);
+		if (error)
+			goto out;
+	} while (try_again && ++tries < 20);
+
+	/*
+	 * We gave it our best shot but failed, so mark this scrub
+	 * incomplete.  Userspace can decide if it wants to try again.
+	 */
+	if (try_again && tries == 20)
+		xfs_scrub_set_incomplete(sc);
+out:
+	return error;
+}
diff --git a/fs/xfs/scrub/quota.c b/fs/xfs/scrub/quota.c
new file mode 100644
index 000000000000..6ba465e6c885
--- /dev/null
+++ b/fs/xfs/scrub/quota.c
@@ -0,0 +1,297 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_inode_fork.h"
+#include "xfs_alloc.h"
+#include "xfs_bmap.h"
+#include "xfs_quota.h"
+#include "xfs_qm.h"
+#include "xfs_dquot.h"
+#include "xfs_dquot_item.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+
+/* Convert a scrub type code to a DQ flag, or return 0 if error. */
+static inline uint
+xfs_scrub_quota_to_dqtype(
+	struct xfs_scrub_context	*sc)
+{
+	switch (sc->sm->sm_type) {
+	case XFS_SCRUB_TYPE_UQUOTA:
+		return XFS_DQ_USER;
+	case XFS_SCRUB_TYPE_GQUOTA:
+		return XFS_DQ_GROUP;
+	case XFS_SCRUB_TYPE_PQUOTA:
+		return XFS_DQ_PROJ;
+	default:
+		return 0;
+	}
+}
+
+/* Set us up to scrub a quota. */
+int
+xfs_scrub_setup_quota(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	uint				dqtype;
+
+	dqtype = xfs_scrub_quota_to_dqtype(sc);
+	if (dqtype == 0)
+		return -EINVAL;
+	if (!xfs_this_quota_on(sc->mp, dqtype))
+		return -ENOENT;
+	return 0;
+}
+
+/* Quotas. */
+
+/* Scrub the fields in an individual quota item. */
+STATIC void
+xfs_scrub_quota_item(
+	struct xfs_scrub_context	*sc,
+	uint				dqtype,
+	struct xfs_dquot		*dq,
+	xfs_dqid_t			id)
+{
+	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;
+
+	offset = id / qi->qi_dqperchunk;
+
+	/*
+	 * We fed $id and DQNEXT into the xfs_qm_dqget call, which means
+	 * that the actual dquot we got must either have the same id or
+	 * the next higher id.
+	 */
+	if (id > be32_to_cpu(d->d_id))
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
+
+	/* Did we get the dquot type we wanted? */
+	if (dqtype != (d->d_flags & XFS_DQ_ALLTYPES))
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
+
+	if (d->d_pad0 != cpu_to_be32(0) || d->d_pad != cpu_to_be16(0))
+		xfs_scrub_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);
+
+	/*
+	 * Warn if the hard limits are larger than the fs.
+	 * Administrators can do this, though in production this seems
+	 * suspect, which is why we flag it for review.
+	 *
+	 * Complain about corruption if the soft limit is greater than
+	 * the hard limit.
+	 */
+	if (bhard > mp->m_sb.sb_dblocks)
+		xfs_scrub_fblock_set_warning(sc, XFS_DATA_FORK, offset);
+	if (bsoft > bhard)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
+
+	if (ihard > mp->m_maxicount)
+		xfs_scrub_fblock_set_warning(sc, XFS_DATA_FORK, offset);
+	if (isoft > ihard)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
+
+	if (rhard > mp->m_sb.sb_rblocks)
+		xfs_scrub_fblock_set_warning(sc, XFS_DATA_FORK, offset);
+	if (rsoft > rhard)
+		xfs_scrub_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);
+
+	/*
+	 * Check that usage doesn't exceed physical limits.  However, on
+	 * a reflink filesystem we're allowed to exceed physical space
+	 * if there are no quota limits.
+	 */
+	if (xfs_sb_version_hasreflink(&mp->m_sb)) {
+		if (mp->m_sb.sb_dblocks < bcount)
+			xfs_scrub_fblock_set_warning(sc, XFS_DATA_FORK,
+					offset);
+	} else {
+		if (mp->m_sb.sb_dblocks < bcount)
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK,
+					offset);
+	}
+	if (icount > fs_icount || rcount > mp->m_sb.sb_rblocks)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, offset);
+
+	/*
+	 * We can violate the hard limits if the admin suddenly sets a
+	 * lower limit than the actual usage.  However, we flag it for
+	 * admin review.
+	 */
+	if (id != 0 && bhard != 0 && bcount > bhard)
+		xfs_scrub_fblock_set_warning(sc, XFS_DATA_FORK, offset);
+	if (id != 0 && ihard != 0 && icount > ihard)
+		xfs_scrub_fblock_set_warning(sc, XFS_DATA_FORK, offset);
+	if (id != 0 && rhard != 0 && rcount > rhard)
+		xfs_scrub_fblock_set_warning(sc, XFS_DATA_FORK, offset);
+}
+
+/* Scrub all of a quota type's items. */
+int
+xfs_scrub_quota(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_bmbt_irec		irec = { 0 };
+	struct xfs_mount		*mp = sc->mp;
+	struct xfs_inode		*ip;
+	struct xfs_quotainfo		*qi = mp->m_quotainfo;
+	struct xfs_dquot		*dq;
+	xfs_fileoff_t			max_dqid_off;
+	xfs_fileoff_t			off = 0;
+	xfs_dqid_t			id = 0;
+	uint				dqtype;
+	int				nimaps;
+	int				error = 0;
+
+	if (!XFS_IS_QUOTA_RUNNING(mp) || !XFS_IS_QUOTA_ON(mp))
+		return -ENOENT;
+
+	mutex_lock(&qi->qi_quotaofflock);
+	dqtype = xfs_scrub_quota_to_dqtype(sc);
+	if (!xfs_this_quota_on(sc->mp, dqtype)) {
+		error = -ENOENT;
+		goto out_unlock_quota;
+	}
+
+	/* Attach to the quota inode and set sc->ip so that reporting works. */
+	ip = xfs_quota_inode(sc->mp, dqtype);
+	sc->ip = ip;
+
+	/* Look for problem extents. */
+	xfs_ilock(ip, XFS_ILOCK_EXCL);
+	if (ip->i_d.di_flags & XFS_DIFLAG_REALTIME) {
+		xfs_scrub_ino_set_corrupt(sc, sc->ip->i_ino);
+		goto out_unlock_inode;
+	}
+	max_dqid_off = ((xfs_dqid_t)-1) / qi->qi_dqperchunk;
+	while (1) {
+		if (xfs_scrub_should_terminate(sc, &error))
+			break;
+
+		off = irec.br_startoff + irec.br_blockcount;
+		nimaps = 1;
+		error = xfs_bmapi_read(ip, off, -1, &irec, &nimaps,
+				XFS_BMAPI_ENTIRE);
+		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, off,
+				&error))
+			goto out_unlock_inode;
+		if (!nimaps)
+			break;
+		if (irec.br_startblock == HOLESTARTBLOCK)
+			continue;
+
+		/* Check the extent record doesn't point to crap. */
+		if (irec.br_startblock + irec.br_blockcount <=
+		    irec.br_startblock)
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK,
+					irec.br_startoff);
+		if (!xfs_verify_fsbno(mp, irec.br_startblock) ||
+		    !xfs_verify_fsbno(mp, irec.br_startblock +
+					irec.br_blockcount - 1))
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK,
+					irec.br_startoff);
+
+		/*
+		 * Unwritten extents or blocks mapped above the highest
+		 * quota id shouldn't happen.
+		 */
+		if (isnullstartblock(irec.br_startblock) ||
+		    irec.br_startoff > max_dqid_off ||
+		    irec.br_startoff + irec.br_blockcount > max_dqid_off + 1)
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, off);
+	}
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		goto out;
+
+	/* Check all the quota items. */
+	while (id < ((xfs_dqid_t)-1ULL)) {
+		if (xfs_scrub_should_terminate(sc, &error))
+			break;
+
+		error = xfs_qm_dqget(mp, NULL, id, dqtype, XFS_QMOPT_DQNEXT,
+				&dq);
+		if (error == -ENOENT)
+			break;
+		if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK,
+				id * qi->qi_dqperchunk, &error))
+			break;
+
+		xfs_scrub_quota_item(sc, dqtype, dq, id);
+
+		id = be32_to_cpu(dq->q_core.d_id) + 1;
+		xfs_qm_dqput(dq);
+		if (!id)
+			break;
+	}
+
+out:
+	/* We set sc->ip earlier, so make sure we clear it now. */
+	sc->ip = NULL;
+out_unlock_quota:
+	mutex_unlock(&qi->qi_quotaofflock);
+	return error;
+
+out_unlock_inode:
+	xfs_iunlock(ip, XFS_ILOCK_EXCL);
+	goto out;
+}
diff --git a/fs/xfs/scrub/refcount.c b/fs/xfs/scrub/refcount.c
new file mode 100644
index 000000000000..400f1561cd3d
--- /dev/null
+++ b/fs/xfs/scrub/refcount.c
@@ -0,0 +1,515 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_alloc.h"
+#include "xfs_rmap.h"
+#include "xfs_refcount.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+
+/*
+ * Set us up to scrub reference count btrees.
+ */
+int
+xfs_scrub_setup_ag_refcountbt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	return xfs_scrub_setup_ag_btree(sc, ip, false);
+}
+
+/* Reference count btree scrubber. */
+
+/*
+ * Confirming Reference Counts via Reverse Mappings
+ *
+ * We want to count the reverse mappings overlapping a refcount record
+ * (bno, len, refcount), allowing for the possibility that some of the
+ * overlap may come from smaller adjoining reverse mappings, while some
+ * comes from single extents which overlap the range entirely.  The
+ * outer loop is as follows:
+ *
+ * 1. For all reverse mappings overlapping the refcount extent,
+ *    a. If a given rmap completely overlaps, mark it as seen.
+ *    b. Otherwise, record the fragment (in agbno order) for later
+ *       processing.
+ *
+ * Once we've seen all the rmaps, we know that for all blocks in the
+ * refcount record we want to find $refcount owners and we've already
+ * visited $seen extents that overlap all the blocks.  Therefore, we
+ * need to find ($refcount - $seen) owners for every block in the
+ * extent; call that quantity $target_nr.  Proceed as follows:
+ *
+ * 2. Pull the first $target_nr fragments from the list; all of them
+ *    should start at or before the start of the extent.
+ *    Call this subset of fragments the working set.
+ * 3. Until there are no more unprocessed fragments,
+ *    a. Find the shortest fragments in the set and remove them.
+ *    b. Note the block number of the end of these fragments.
+ *    c. Pull the same number of fragments from the list.  All of these
+ *       fragments should start at the block number recorded in the
+ *       previous step.
+ *    d. Put those fragments in the set.
+ * 4. Check that there are $target_nr fragments remaining in the list,
+ *    and that they all end at or beyond the end of the refcount extent.
+ *
+ * If the refcount is correct, all the check conditions in the algorithm
+ * should always hold true.  If not, the refcount is incorrect.
+ */
+struct xfs_scrub_refcnt_frag {
+	struct list_head		list;
+	struct xfs_rmap_irec		rm;
+};
+
+struct xfs_scrub_refcnt_check {
+	struct xfs_scrub_context	*sc;
+	struct list_head		fragments;
+
+	/* refcount extent we're examining */
+	xfs_agblock_t			bno;
+	xfs_extlen_t			len;
+	xfs_nlink_t			refcount;
+
+	/* number of owners seen */
+	xfs_nlink_t			seen;
+};
+
+/*
+ * Decide if the given rmap is large enough that we can redeem it
+ * towards refcount verification now, or if it's a fragment, in
+ * which case we'll hang onto it in the hopes that we'll later
+ * discover that we've collected exactly the correct number of
+ * fragments as the refcountbt says we should have.
+ */
+STATIC int
+xfs_scrub_refcountbt_rmap_check(
+	struct xfs_btree_cur		*cur,
+	struct xfs_rmap_irec		*rec,
+	void				*priv)
+{
+	struct xfs_scrub_refcnt_check	*refchk = priv;
+	struct xfs_scrub_refcnt_frag	*frag;
+	xfs_agblock_t			rm_last;
+	xfs_agblock_t			rc_last;
+	int				error = 0;
+
+	if (xfs_scrub_should_terminate(refchk->sc, &error))
+		return error;
+
+	rm_last = rec->rm_startblock + rec->rm_blockcount - 1;
+	rc_last = refchk->bno + refchk->len - 1;
+
+	/* Confirm that a single-owner refc extent is a CoW stage. */
+	if (refchk->refcount == 1 && rec->rm_owner != XFS_RMAP_OWN_COW) {
+		xfs_scrub_btree_xref_set_corrupt(refchk->sc, cur, 0);
+		return 0;
+	}
+
+	if (rec->rm_startblock <= refchk->bno && rm_last >= rc_last) {
+		/*
+		 * The rmap overlaps the refcount record, so we can confirm
+		 * one refcount owner seen.
+		 */
+		refchk->seen++;
+	} else {
+		/*
+		 * This rmap covers only part of the refcount record, so
+		 * save the fragment for later processing.  If the rmapbt
+		 * is healthy each rmap_irec we see will be in agbno order
+		 * so we don't need insertion sort here.
+		 */
+		frag = kmem_alloc(sizeof(struct xfs_scrub_refcnt_frag),
+				KM_MAYFAIL | KM_NOFS);
+		if (!frag)
+			return -ENOMEM;
+		memcpy(&frag->rm, rec, sizeof(frag->rm));
+		list_add_tail(&frag->list, &refchk->fragments);
+	}
+
+	return 0;
+}
+
+/*
+ * Given a bunch of rmap fragments, iterate through them, keeping
+ * a running tally of the refcount.  If this ever deviates from
+ * what we expect (which is the refcountbt's refcount minus the
+ * number of extents that totally covered the refcountbt extent),
+ * we have a refcountbt error.
+ */
+STATIC void
+xfs_scrub_refcountbt_process_rmap_fragments(
+	struct xfs_scrub_refcnt_check	*refchk)
+{
+	struct list_head		worklist;
+	struct xfs_scrub_refcnt_frag	*frag;
+	struct xfs_scrub_refcnt_frag	*n;
+	xfs_agblock_t			bno;
+	xfs_agblock_t			rbno;
+	xfs_agblock_t			next_rbno;
+	xfs_nlink_t			nr;
+	xfs_nlink_t			target_nr;
+
+	target_nr = refchk->refcount - refchk->seen;
+	if (target_nr == 0)
+		return;
+
+	/*
+	 * There are (refchk->rc.rc_refcount - refchk->nr refcount)
+	 * references we haven't found yet.  Pull that many off the
+	 * fragment list and figure out where the smallest rmap ends
+	 * (and therefore the next rmap should start).  All the rmaps
+	 * we pull off should start at or before the beginning of the
+	 * refcount record's range.
+	 */
+	INIT_LIST_HEAD(&worklist);
+	rbno = NULLAGBLOCK;
+	nr = 1;
+
+	/* Make sure the fragments actually /are/ in agbno order. */
+	bno = 0;
+	list_for_each_entry(frag, &refchk->fragments, list) {
+		if (frag->rm.rm_startblock < bno)
+			goto done;
+		bno = frag->rm.rm_startblock;
+	}
+
+	/*
+	 * Find all the rmaps that start at or before the refc extent,
+	 * and put them on the worklist.
+	 */
+	list_for_each_entry_safe(frag, n, &refchk->fragments, list) {
+		if (frag->rm.rm_startblock > refchk->bno)
+			goto done;
+		bno = frag->rm.rm_startblock + frag->rm.rm_blockcount;
+		if (bno < rbno)
+			rbno = bno;
+		list_move_tail(&frag->list, &worklist);
+		if (nr == target_nr)
+			break;
+		nr++;
+	}
+
+	/*
+	 * We should have found exactly $target_nr rmap fragments starting
+	 * at or before the refcount extent.
+	 */
+	if (nr != target_nr)
+		goto done;
+
+	while (!list_empty(&refchk->fragments)) {
+		/* Discard any fragments ending at rbno from the worklist. */
+		nr = 0;
+		next_rbno = NULLAGBLOCK;
+		list_for_each_entry_safe(frag, n, &worklist, list) {
+			bno = frag->rm.rm_startblock + frag->rm.rm_blockcount;
+			if (bno != rbno) {
+				if (bno < next_rbno)
+					next_rbno = bno;
+				continue;
+			}
+			list_del(&frag->list);
+			kmem_free(frag);
+			nr++;
+		}
+
+		/* Try to add nr rmaps starting at rbno to the worklist. */
+		list_for_each_entry_safe(frag, n, &refchk->fragments, list) {
+			bno = frag->rm.rm_startblock + frag->rm.rm_blockcount;
+			if (frag->rm.rm_startblock != rbno)
+				goto done;
+			list_move_tail(&frag->list, &worklist);
+			if (next_rbno > bno)
+				next_rbno = bno;
+			nr--;
+			if (nr == 0)
+				break;
+		}
+
+		/*
+		 * If we get here and nr > 0, this means that we added fewer
+		 * items to the worklist than we discarded because the fragment
+		 * list ran out of items.  Therefore, we cannot maintain the
+		 * required refcount.  Something is wrong, so we're done.
+		 */
+		if (nr)
+			goto done;
+
+		rbno = next_rbno;
+	}
+
+	/*
+	 * Make sure the last extent we processed ends at or beyond
+	 * the end of the refcount extent.
+	 */
+	if (rbno < refchk->bno + refchk->len)
+		goto done;
+
+	/* Actually record us having seen the remaining refcount. */
+	refchk->seen = refchk->refcount;
+done:
+	/* Delete fragments and work list. */
+	list_for_each_entry_safe(frag, n, &worklist, list) {
+		list_del(&frag->list);
+		kmem_free(frag);
+	}
+	list_for_each_entry_safe(frag, n, &refchk->fragments, list) {
+		list_del(&frag->list);
+		kmem_free(frag);
+	}
+}
+
+/* Use the rmap entries covering this extent to verify the refcount. */
+STATIC void
+xfs_scrub_refcountbt_xref_rmap(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			bno,
+	xfs_extlen_t			len,
+	xfs_nlink_t			refcount)
+{
+	struct xfs_scrub_refcnt_check	refchk = {
+		.sc = sc,
+		.bno = bno,
+		.len = len,
+		.refcount = refcount,
+		.seen = 0,
+	};
+	struct xfs_rmap_irec		low;
+	struct xfs_rmap_irec		high;
+	struct xfs_scrub_refcnt_frag	*frag;
+	struct xfs_scrub_refcnt_frag	*n;
+	int				error;
+
+	if (!sc->sa.rmap_cur)
+		return;
+
+	/* Cross-reference with the rmapbt to confirm the refcount. */
+	memset(&low, 0, sizeof(low));
+	low.rm_startblock = bno;
+	memset(&high, 0xFF, sizeof(high));
+	high.rm_startblock = bno + len - 1;
+
+	INIT_LIST_HEAD(&refchk.fragments);
+	error = xfs_rmap_query_range(sc->sa.rmap_cur, &low, &high,
+			&xfs_scrub_refcountbt_rmap_check, &refchk);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.rmap_cur))
+		goto out_free;
+
+	xfs_scrub_refcountbt_process_rmap_fragments(&refchk);
+	if (refcount != refchk.seen)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.rmap_cur, 0);
+
+out_free:
+	list_for_each_entry_safe(frag, n, &refchk.fragments, list) {
+		list_del(&frag->list);
+		kmem_free(frag);
+	}
+}
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_refcountbt_xref(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len,
+	xfs_nlink_t			refcount)
+{
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	xfs_scrub_xref_is_used_space(sc, agbno, len);
+	xfs_scrub_xref_is_not_inode_chunk(sc, agbno, len);
+	xfs_scrub_refcountbt_xref_rmap(sc, agbno, len, refcount);
+}
+
+/* Scrub a refcountbt record. */
+STATIC int
+xfs_scrub_refcountbt_rec(
+	struct xfs_scrub_btree		*bs,
+	union xfs_btree_rec		*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_agblock_t			bno;
+	xfs_extlen_t			len;
+	xfs_nlink_t			refcount;
+	bool				has_cowflag;
+	int				error = 0;
+
+	bno = be32_to_cpu(rec->refc.rc_startblock);
+	len = be32_to_cpu(rec->refc.rc_blockcount);
+	refcount = be32_to_cpu(rec->refc.rc_refcount);
+
+	/* Only CoW records can have refcount == 1. */
+	has_cowflag = (bno & XFS_REFC_COW_START);
+	if ((refcount == 1 && !has_cowflag) || (refcount != 1 && has_cowflag))
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+	if (has_cowflag)
+		(*cow_blocks) += len;
+
+	/* Check the extent. */
+	bno &= ~XFS_REFC_COW_START;
+	if (bno + len <= bno ||
+	    !xfs_verify_agbno(mp, agno, bno) ||
+	    !xfs_verify_agbno(mp, agno, bno + len - 1))
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	if (refcount == 0)
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	xfs_scrub_refcountbt_xref(bs->sc, bno, len, refcount);
+
+	return error;
+}
+
+/* Make sure we have as many refc blocks as the rmap says. */
+STATIC void
+xfs_scrub_refcount_xref_rmap(
+	struct xfs_scrub_context	*sc,
+	struct xfs_owner_info		*oinfo,
+	xfs_filblks_t			cow_blocks)
+{
+	xfs_extlen_t			refcbt_blocks = 0;
+	xfs_filblks_t			blocks;
+	int				error;
+
+	if (!sc->sa.rmap_cur)
+		return;
+
+	/* Check that we saw as many refcbt blocks as the rmap knows about. */
+	error = xfs_btree_count_blocks(sc->sa.refc_cur, &refcbt_blocks);
+	if (!xfs_scrub_btree_process_error(sc, sc->sa.refc_cur, 0, &error))
+		return;
+	error = xfs_scrub_count_rmap_ownedby_ag(sc, sc->sa.rmap_cur, oinfo,
+			&blocks);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.rmap_cur))
+		return;
+	if (blocks != refcbt_blocks)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.rmap_cur, 0);
+
+	/* Check that we saw as many cow blocks as the rmap knows about. */
+	xfs_rmap_ag_owner(oinfo, XFS_RMAP_OWN_COW);
+	error = xfs_scrub_count_rmap_ownedby_ag(sc, sc->sa.rmap_cur, oinfo,
+			&blocks);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.rmap_cur))
+		return;
+	if (blocks != cow_blocks)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.rmap_cur, 0);
+}
+
+/* Scrub the refcount btree for some AG. */
+int
+xfs_scrub_refcountbt(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_owner_info		oinfo;
+	xfs_agblock_t			cow_blocks = 0;
+	int				error;
+
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_REFC);
+	error = xfs_scrub_btree(sc, sc->sa.refc_cur, xfs_scrub_refcountbt_rec,
+			&oinfo, &cow_blocks);
+	if (error)
+		return error;
+
+	xfs_scrub_refcount_xref_rmap(sc, &oinfo, cow_blocks);
+
+	return 0;
+}
+
+/* xref check that a cow staging extent is marked in the refcountbt. */
+void
+xfs_scrub_xref_is_cow_staging(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len)
+{
+	struct xfs_refcount_irec	rc;
+	bool				has_cowflag;
+	int				has_refcount;
+	int				error;
+
+	if (!sc->sa.refc_cur)
+		return;
+
+	/* Find the CoW staging extent. */
+	error = xfs_refcount_lookup_le(sc->sa.refc_cur,
+			agbno + XFS_REFC_COW_START, &has_refcount);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.refc_cur))
+		return;
+	if (!has_refcount) {
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.refc_cur, 0);
+		return;
+	}
+
+	error = xfs_refcount_get_rec(sc->sa.refc_cur, &rc, &has_refcount);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.refc_cur))
+		return;
+	if (!has_refcount) {
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.refc_cur, 0);
+		return;
+	}
+
+	/* CoW flag must be set, refcount must be 1. */
+	has_cowflag = (rc.rc_startblock & XFS_REFC_COW_START);
+	if (!has_cowflag || rc.rc_refcount != 1)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.refc_cur, 0);
+
+	/* Must be at least as long as what was passed in */
+	if (rc.rc_blockcount < len)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.refc_cur, 0);
+}
+
+/*
+ * xref check that the extent is not shared.  Only file data blocks
+ * can have multiple owners.
+ */
+void
+xfs_scrub_xref_is_not_shared(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			agbno,
+	xfs_extlen_t			len)
+{
+	bool				shared;
+	int				error;
+
+	if (!sc->sa.refc_cur)
+		return;
+
+	error = xfs_refcount_has_record(sc->sa.refc_cur, agbno, len, &shared);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.refc_cur))
+		return;
+	if (shared)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.refc_cur, 0);
+}
diff --git a/fs/xfs/scrub/rmap.c b/fs/xfs/scrub/rmap.c
new file mode 100644
index 000000000000..8f2a7c3ff455
--- /dev/null
+++ b/fs/xfs/scrub/rmap.c
@@ -0,0 +1,261 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_alloc.h"
+#include "xfs_ialloc.h"
+#include "xfs_rmap.h"
+#include "xfs_refcount.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/btree.h"
+#include "scrub/trace.h"
+
+/*
+ * Set us up to scrub reverse mapping btrees.
+ */
+int
+xfs_scrub_setup_ag_rmapbt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	return xfs_scrub_setup_ag_btree(sc, ip, false);
+}
+
+/* Reverse-mapping scrubber. */
+
+/* Cross-reference a rmap against the refcount btree. */
+STATIC void
+xfs_scrub_rmapbt_xref_refc(
+	struct xfs_scrub_context	*sc,
+	struct xfs_rmap_irec		*irec)
+{
+	xfs_agblock_t			fbno;
+	xfs_extlen_t			flen;
+	bool				non_inode;
+	bool				is_bmbt;
+	bool				is_attr;
+	bool				is_unwritten;
+	int				error;
+
+	if (!sc->sa.refc_cur)
+		return;
+
+	non_inode = XFS_RMAP_NON_INODE_OWNER(irec->rm_owner);
+	is_bmbt = irec->rm_flags & XFS_RMAP_BMBT_BLOCK;
+	is_attr = irec->rm_flags & XFS_RMAP_ATTR_FORK;
+	is_unwritten = irec->rm_flags & XFS_RMAP_UNWRITTEN;
+
+	/* If this is shared, must be a data fork extent. */
+	error = xfs_refcount_find_shared(sc->sa.refc_cur, irec->rm_startblock,
+			irec->rm_blockcount, &fbno, &flen, false);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.refc_cur))
+		return;
+	if (flen != 0 && (non_inode || is_attr || is_bmbt || is_unwritten))
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.refc_cur, 0);
+}
+
+/* Cross-reference with the other btrees. */
+STATIC void
+xfs_scrub_rmapbt_xref(
+	struct xfs_scrub_context	*sc,
+	struct xfs_rmap_irec		*irec)
+{
+	xfs_agblock_t			agbno = irec->rm_startblock;
+	xfs_extlen_t			len = irec->rm_blockcount;
+
+	if (sc->sm->sm_flags & XFS_SCRUB_OFLAG_CORRUPT)
+		return;
+
+	xfs_scrub_xref_is_used_space(sc, agbno, len);
+	if (irec->rm_owner == XFS_RMAP_OWN_INODES)
+		xfs_scrub_xref_is_inode_chunk(sc, agbno, len);
+	else
+		xfs_scrub_xref_is_not_inode_chunk(sc, agbno, len);
+	if (irec->rm_owner == XFS_RMAP_OWN_COW)
+		xfs_scrub_xref_is_cow_staging(sc, irec->rm_startblock,
+				irec->rm_blockcount);
+	else
+		xfs_scrub_rmapbt_xref_refc(sc, irec);
+}
+
+/* Scrub an rmapbt record. */
+STATIC int
+xfs_scrub_rmapbt_rec(
+	struct xfs_scrub_btree		*bs,
+	union xfs_btree_rec		*rec)
+{
+	struct xfs_mount		*mp = bs->cur->bc_mp;
+	struct xfs_rmap_irec		irec;
+	xfs_agnumber_t			agno = bs->cur->bc_private.a.agno;
+	bool				non_inode;
+	bool				is_unwritten;
+	bool				is_bmbt;
+	bool				is_attr;
+	int				error;
+
+	error = xfs_rmap_btrec_to_irec(rec, &irec);
+	if (!xfs_scrub_btree_process_error(bs->sc, bs->cur, 0, &error))
+		goto out;
+
+	/* Check extent. */
+	if (irec.rm_startblock + irec.rm_blockcount <= irec.rm_startblock)
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	if (irec.rm_owner == XFS_RMAP_OWN_FS) {
+		/*
+		 * xfs_verify_agbno returns false for static fs metadata.
+		 * Since that only exists at the start of the AG, validate
+		 * that by hand.
+		 */
+		if (irec.rm_startblock != 0 ||
+		    irec.rm_blockcount != XFS_AGFL_BLOCK(mp) + 1)
+			xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+	} else {
+		/*
+		 * Otherwise we must point somewhere past the static metadata
+		 * but before the end of the FS.  Run the regular check.
+		 */
+		if (!xfs_verify_agbno(mp, agno, irec.rm_startblock) ||
+		    !xfs_verify_agbno(mp, agno, irec.rm_startblock +
+				irec.rm_blockcount - 1))
+			xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+	}
+
+	/* Check flags. */
+	non_inode = XFS_RMAP_NON_INODE_OWNER(irec.rm_owner);
+	is_bmbt = irec.rm_flags & XFS_RMAP_BMBT_BLOCK;
+	is_attr = irec.rm_flags & XFS_RMAP_ATTR_FORK;
+	is_unwritten = irec.rm_flags & XFS_RMAP_UNWRITTEN;
+
+	if (is_bmbt && irec.rm_offset != 0)
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	if (non_inode && irec.rm_offset != 0)
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	if (is_unwritten && (is_bmbt || non_inode || is_attr))
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	if (non_inode && (is_bmbt || is_unwritten || is_attr))
+		xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+
+	if (!non_inode) {
+		if (!xfs_verify_ino(mp, irec.rm_owner))
+			xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+	} else {
+		/* Non-inode owner within the magic values? */
+		if (irec.rm_owner <= XFS_RMAP_OWN_MIN ||
+		    irec.rm_owner > XFS_RMAP_OWN_FS)
+			xfs_scrub_btree_set_corrupt(bs->sc, bs->cur, 0);
+	}
+
+	xfs_scrub_rmapbt_xref(bs->sc, &irec);
+out:
+	return error;
+}
+
+/* Scrub the rmap btree for some AG. */
+int
+xfs_scrub_rmapbt(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_owner_info		oinfo;
+
+	xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_AG);
+	return xfs_scrub_btree(sc, sc->sa.rmap_cur, xfs_scrub_rmapbt_rec,
+			&oinfo, NULL);
+}
+
+/* xref check that the extent is owned by a given owner */
+static inline void
+xfs_scrub_xref_check_owner(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			bno,
+	xfs_extlen_t			len,
+	struct xfs_owner_info		*oinfo,
+	bool				should_have_rmap)
+{
+	bool				has_rmap;
+	int				error;
+
+	if (!sc->sa.rmap_cur)
+		return;
+
+	error = xfs_rmap_record_exists(sc->sa.rmap_cur, bno, len, oinfo,
+			&has_rmap);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.rmap_cur))
+		return;
+	if (has_rmap != should_have_rmap)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.rmap_cur, 0);
+}
+
+/* xref check that the extent is owned by a given owner */
+void
+xfs_scrub_xref_is_owned_by(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			bno,
+	xfs_extlen_t			len,
+	struct xfs_owner_info		*oinfo)
+{
+	xfs_scrub_xref_check_owner(sc, bno, len, oinfo, true);
+}
+
+/* xref check that the extent is not owned by a given owner */
+void
+xfs_scrub_xref_is_not_owned_by(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			bno,
+	xfs_extlen_t			len,
+	struct xfs_owner_info		*oinfo)
+{
+	xfs_scrub_xref_check_owner(sc, bno, len, oinfo, false);
+}
+
+/* xref check that the extent has no reverse mapping at all */
+void
+xfs_scrub_xref_has_no_owner(
+	struct xfs_scrub_context	*sc,
+	xfs_agblock_t			bno,
+	xfs_extlen_t			len)
+{
+	bool				has_rmap;
+	int				error;
+
+	if (!sc->sa.rmap_cur)
+		return;
+
+	error = xfs_rmap_has_record(sc->sa.rmap_cur, bno, len, &has_rmap);
+	if (!xfs_scrub_should_check_xref(sc, &error, &sc->sa.rmap_cur))
+		return;
+	if (has_rmap)
+		xfs_scrub_btree_xref_set_corrupt(sc, sc->sa.rmap_cur, 0);
+}
diff --git a/fs/xfs/scrub/rtbitmap.c b/fs/xfs/scrub/rtbitmap.c
new file mode 100644
index 000000000000..39c41dfe08ee
--- /dev/null
+++ b/fs/xfs/scrub/rtbitmap.c
@@ -0,0 +1,122 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_alloc.h"
+#include "xfs_rtalloc.h"
+#include "xfs_inode.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+
+/* Set us up with the realtime metadata locked. */
+int
+xfs_scrub_setup_rt(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	int				error;
+
+	error = xfs_scrub_setup_fs(sc, ip);
+	if (error)
+		return error;
+
+	sc->ilock_flags = XFS_ILOCK_EXCL | XFS_ILOCK_RTBITMAP;
+	sc->ip = sc->mp->m_rbmip;
+	xfs_ilock(sc->ip, sc->ilock_flags);
+
+	return 0;
+}
+
+/* Realtime bitmap. */
+
+/* Scrub a free extent record from the realtime bitmap. */
+STATIC int
+xfs_scrub_rtbitmap_rec(
+	struct xfs_trans		*tp,
+	struct xfs_rtalloc_rec		*rec,
+	void				*priv)
+{
+	struct xfs_scrub_context	*sc = priv;
+
+	if (rec->ar_startblock + rec->ar_blockcount <= rec->ar_startblock ||
+	    !xfs_verify_rtbno(sc->mp, rec->ar_startblock) ||
+	    !xfs_verify_rtbno(sc->mp, rec->ar_startblock +
+			rec->ar_blockcount - 1))
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+	return 0;
+}
+
+/* Scrub the realtime bitmap. */
+int
+xfs_scrub_rtbitmap(
+	struct xfs_scrub_context	*sc)
+{
+	int				error;
+
+	error = xfs_rtalloc_query_all(sc->tp, xfs_scrub_rtbitmap_rec, sc);
+	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
+		goto out;
+
+out:
+	return error;
+}
+
+/* Scrub the realtime summary. */
+int
+xfs_scrub_rtsummary(
+	struct xfs_scrub_context	*sc)
+{
+	/* XXX: implement this some day */
+	return -ENOENT;
+}
+
+
+/* xref check that the extent is not free in the rtbitmap */
+void
+xfs_scrub_xref_is_used_rt_space(
+	struct xfs_scrub_context	*sc,
+	xfs_rtblock_t			fsbno,
+	xfs_extlen_t			len)
+{
+	bool				is_free;
+	int				error;
+
+	xfs_ilock(sc->mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+	error = xfs_rtalloc_extent_is_free(sc->mp, sc->tp, fsbno, len,
+			&is_free);
+	if (!xfs_scrub_should_check_xref(sc, &error, NULL))
+		goto out_unlock;
+	if (is_free)
+		xfs_scrub_ino_xref_set_corrupt(sc, sc->mp->m_rbmip->i_ino);
+out_unlock:
+	xfs_iunlock(sc->mp->m_rbmip, XFS_ILOCK_SHARED | XFS_ILOCK_RTBITMAP);
+}
diff --git a/fs/xfs/scrub/scrub.c b/fs/xfs/scrub/scrub.c
new file mode 100644
index 000000000000..26c75967a072
--- /dev/null
+++ b/fs/xfs/scrub/scrub.c
@@ -0,0 +1,462 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_icache.h"
+#include "xfs_itable.h"
+#include "xfs_alloc.h"
+#include "xfs_alloc_btree.h"
+#include "xfs_bmap.h"
+#include "xfs_bmap_btree.h"
+#include "xfs_ialloc.h"
+#include "xfs_ialloc_btree.h"
+#include "xfs_refcount.h"
+#include "xfs_refcount_btree.h"
+#include "xfs_rmap.h"
+#include "xfs_rmap_btree.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+#include "scrub/btree.h"
+
+/*
+ * Online Scrub and Repair
+ *
+ * Traditionally, XFS (the kernel driver) did not know how to check or
+ * repair on-disk data structures.  That task was left to the xfs_check
+ * and xfs_repair tools, both of which require taking the filesystem
+ * offline for a thorough but time consuming examination.  Online
+ * scrub & repair, on the other hand, enables us to check the metadata
+ * for obvious errors while carefully stepping around the filesystem's
+ * ongoing operations, locking rules, etc.
+ *
+ * Given that most XFS metadata consist of records stored in a btree,
+ * most of the checking functions iterate the btree blocks themselves
+ * looking for irregularities.  When a record block is encountered, each
+ * record can be checked for obviously bad values.  Record values can
+ * also be cross-referenced against other btrees to look for potential
+ * misunderstandings between pieces of metadata.
+ *
+ * It is expected that the checkers responsible for per-AG metadata
+ * structures will lock the AG headers (AGI, AGF, AGFL), iterate the
+ * metadata structure, and perform any relevant cross-referencing before
+ * unlocking the AG and returning the results to userspace.  These
+ * scrubbers must not keep an AG locked for too long to avoid tying up
+ * the block and inode allocators.
+ *
+ * Block maps and b-trees rooted in an inode present a special challenge
+ * because they can involve extents from any AG.  The general scrubber
+ * structure of lock -> check -> xref -> unlock still holds, but AG
+ * locking order rules /must/ be obeyed to avoid deadlocks.  The
+ * ordering rule, of course, is that we must lock in increasing AG
+ * order.  Helper functions are provided to track which AG headers we've
+ * already locked.  If we detect an imminent locking order violation, we
+ * can signal a potential deadlock, in which case the scrubber can jump
+ * out to the top level, lock all the AGs in order, and retry the scrub.
+ *
+ * For file data (directories, extended attributes, symlinks) scrub, we
+ * can simply lock the inode and walk the data.  For btree data
+ * (directories and attributes) we follow the same btree-scrubbing
+ * strategy outlined previously to check the records.
+ *
+ * We use a bit of trickery with transactions to avoid buffer deadlocks
+ * if there is a cycle in the metadata.  The basic problem is that
+ * travelling down a btree involves locking the current buffer at each
+ * tree level.  If a pointer should somehow point back to a buffer that
+ * we've already examined, we will deadlock due to the second buffer
+ * locking attempt.  Note however that grabbing a buffer in transaction
+ * context links the locked buffer to the transaction.  If we try to
+ * re-grab the buffer in the context of the same transaction, we avoid
+ * the second lock attempt and continue.  Between the verifier and the
+ * scrubber, something will notice that something is amiss and report
+ * the corruption.  Therefore, each scrubber will allocate an empty
+ * transaction, attach buffers to it, and cancel the transaction at the
+ * end of the scrub run.  Cancelling a non-dirty transaction simply
+ * unlocks the buffers.
+ *
+ * There are four pieces of data that scrub can communicate to
+ * userspace.  The first is the error code (errno), which can be used to
+ * communicate operational errors in performing the scrub.  There are
+ * also three flags that can be set in the scrub context.  If the data
+ * structure itself is corrupt, the CORRUPT flag will be set.  If
+ * the metadata is correct but otherwise suboptimal, the PREEN flag
+ * will be set.
+ *
+ * We perform secondary validation of filesystem metadata by
+ * cross-referencing every record with all other available metadata.
+ * For example, for block mapping extents, we verify that there are no
+ * records in the free space and inode btrees corresponding to that
+ * space extent and that there is a corresponding entry in the reverse
+ * mapping btree.  Inconsistent metadata is noted by setting the
+ * XCORRUPT flag; btree query function errors are noted by setting the
+ * XFAIL flag and deleting the cursor to prevent further attempts to
+ * cross-reference with a defective btree.
+ */
+
+/*
+ * Scrub probe -- userspace uses this to probe if we're willing to scrub
+ * or repair a given mountpoint.  This will be used by xfs_scrub to
+ * probe the kernel's abilities to scrub (and repair) the metadata.  We
+ * do this by validating the ioctl inputs from userspace, preparing the
+ * filesystem for a scrub (or a repair) operation, and immediately
+ * returning to userspace.  Userspace can use the returned errno and
+ * structure state to decide (in broad terms) if scrub/repair are
+ * supported by the running kernel.
+ */
+static int
+xfs_scrub_probe(
+	struct xfs_scrub_context	*sc)
+{
+	int				error = 0;
+
+	if (xfs_scrub_should_terminate(sc, &error))
+		return error;
+
+	return 0;
+}
+
+/* Scrub setup and teardown */
+
+/* Free all the resources and finish the transactions. */
+STATIC int
+xfs_scrub_teardown(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip_in,
+	int				error)
+{
+	xfs_scrub_ag_free(sc, &sc->sa);
+	if (sc->tp) {
+		xfs_trans_cancel(sc->tp);
+		sc->tp = NULL;
+	}
+	if (sc->ip) {
+		if (sc->ilock_flags)
+			xfs_iunlock(sc->ip, sc->ilock_flags);
+		if (sc->ip != ip_in &&
+		    !xfs_internal_inum(sc->mp, sc->ip->i_ino))
+			iput(VFS_I(sc->ip));
+		sc->ip = NULL;
+	}
+	if (sc->buf) {
+		kmem_free(sc->buf);
+		sc->buf = NULL;
+	}
+	return error;
+}
+
+/* Scrubbing dispatch. */
+
+static const struct xfs_scrub_meta_ops meta_scrub_ops[] = {
+	[XFS_SCRUB_TYPE_PROBE] = {	/* ioctl presence test */
+		.type	= ST_NONE,
+		.setup	= xfs_scrub_setup_fs,
+		.scrub	= xfs_scrub_probe,
+	},
+	[XFS_SCRUB_TYPE_SB] = {		/* superblock */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_fs,
+		.scrub	= xfs_scrub_superblock,
+	},
+	[XFS_SCRUB_TYPE_AGF] = {	/* agf */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_fs,
+		.scrub	= xfs_scrub_agf,
+	},
+	[XFS_SCRUB_TYPE_AGFL]= {	/* agfl */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_fs,
+		.scrub	= xfs_scrub_agfl,
+	},
+	[XFS_SCRUB_TYPE_AGI] = {	/* agi */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_fs,
+		.scrub	= xfs_scrub_agi,
+	},
+	[XFS_SCRUB_TYPE_BNOBT] = {	/* bnobt */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_ag_allocbt,
+		.scrub	= xfs_scrub_bnobt,
+	},
+	[XFS_SCRUB_TYPE_CNTBT] = {	/* cntbt */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_ag_allocbt,
+		.scrub	= xfs_scrub_cntbt,
+	},
+	[XFS_SCRUB_TYPE_INOBT] = {	/* inobt */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_ag_iallocbt,
+		.scrub	= xfs_scrub_inobt,
+	},
+	[XFS_SCRUB_TYPE_FINOBT] = {	/* finobt */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_ag_iallocbt,
+		.scrub	= xfs_scrub_finobt,
+		.has	= xfs_sb_version_hasfinobt,
+	},
+	[XFS_SCRUB_TYPE_RMAPBT] = {	/* rmapbt */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_ag_rmapbt,
+		.scrub	= xfs_scrub_rmapbt,
+		.has	= xfs_sb_version_hasrmapbt,
+	},
+	[XFS_SCRUB_TYPE_REFCNTBT] = {	/* refcountbt */
+		.type	= ST_PERAG,
+		.setup	= xfs_scrub_setup_ag_refcountbt,
+		.scrub	= xfs_scrub_refcountbt,
+		.has	= xfs_sb_version_hasreflink,
+	},
+	[XFS_SCRUB_TYPE_INODE] = {	/* inode record */
+		.type	= ST_INODE,
+		.setup	= xfs_scrub_setup_inode,
+		.scrub	= xfs_scrub_inode,
+	},
+	[XFS_SCRUB_TYPE_BMBTD] = {	/* inode data fork */
+		.type	= ST_INODE,
+		.setup	= xfs_scrub_setup_inode_bmap,
+		.scrub	= xfs_scrub_bmap_data,
+	},
+	[XFS_SCRUB_TYPE_BMBTA] = {	/* inode attr fork */
+		.type	= ST_INODE,
+		.setup	= xfs_scrub_setup_inode_bmap,
+		.scrub	= xfs_scrub_bmap_attr,
+	},
+	[XFS_SCRUB_TYPE_BMBTC] = {	/* inode CoW fork */
+		.type	= ST_INODE,
+		.setup	= xfs_scrub_setup_inode_bmap,
+		.scrub	= xfs_scrub_bmap_cow,
+	},
+	[XFS_SCRUB_TYPE_DIR] = {	/* directory */
+		.type	= ST_INODE,
+		.setup	= xfs_scrub_setup_directory,
+		.scrub	= xfs_scrub_directory,
+	},
+	[XFS_SCRUB_TYPE_XATTR] = {	/* extended attributes */
+		.type	= ST_INODE,
+		.setup	= xfs_scrub_setup_xattr,
+		.scrub	= xfs_scrub_xattr,
+	},
+	[XFS_SCRUB_TYPE_SYMLINK] = {	/* symbolic link */
+		.type	= ST_INODE,
+		.setup	= xfs_scrub_setup_symlink,
+		.scrub	= xfs_scrub_symlink,
+	},
+	[XFS_SCRUB_TYPE_PARENT] = {	/* parent pointers */
+		.type	= ST_INODE,
+		.setup	= xfs_scrub_setup_parent,
+		.scrub	= xfs_scrub_parent,
+	},
+	[XFS_SCRUB_TYPE_RTBITMAP] = {	/* realtime bitmap */
+		.type	= ST_FS,
+		.setup	= xfs_scrub_setup_rt,
+		.scrub	= xfs_scrub_rtbitmap,
+		.has	= xfs_sb_version_hasrealtime,
+	},
+	[XFS_SCRUB_TYPE_RTSUM] = {	/* realtime summary */
+		.type	= ST_FS,
+		.setup	= xfs_scrub_setup_rt,
+		.scrub	= xfs_scrub_rtsummary,
+		.has	= xfs_sb_version_hasrealtime,
+	},
+	[XFS_SCRUB_TYPE_UQUOTA] = {	/* user quota */
+		.type	= ST_FS,
+		.setup	= xfs_scrub_setup_quota,
+		.scrub	= xfs_scrub_quota,
+	},
+	[XFS_SCRUB_TYPE_GQUOTA] = {	/* group quota */
+		.type	= ST_FS,
+		.setup	= xfs_scrub_setup_quota,
+		.scrub	= xfs_scrub_quota,
+	},
+	[XFS_SCRUB_TYPE_PQUOTA] = {	/* project quota */
+		.type	= ST_FS,
+		.setup	= xfs_scrub_setup_quota,
+		.scrub	= xfs_scrub_quota,
+	},
+};
+
+/* This isn't a stable feature, warn once per day. */
+static inline void
+xfs_scrub_experimental_warning(
+	struct xfs_mount	*mp)
+{
+	static struct ratelimit_state scrub_warning = RATELIMIT_STATE_INIT(
+			"xfs_scrub_warning", 86400 * HZ, 1);
+	ratelimit_set_flags(&scrub_warning, RATELIMIT_MSG_ON_RELEASE);
+
+	if (__ratelimit(&scrub_warning))
+		xfs_alert(mp,
+"EXPERIMENTAL online scrub feature in use. Use at your own risk!");
+}
+
+static int
+xfs_scrub_validate_inputs(
+	struct xfs_mount		*mp,
+	struct xfs_scrub_metadata	*sm)
+{
+	int				error;
+	const struct xfs_scrub_meta_ops	*ops;
+
+	error = -EINVAL;
+	/* Check our inputs. */
+	sm->sm_flags &= ~XFS_SCRUB_FLAGS_OUT;
+	if (sm->sm_flags & ~XFS_SCRUB_FLAGS_IN)
+		goto out;
+	/* sm_reserved[] must be zero */
+	if (memchr_inv(sm->sm_reserved, 0, sizeof(sm->sm_reserved)))
+		goto out;
+
+	error = -ENOENT;
+	/* Do we know about this type of metadata? */
+	if (sm->sm_type >= XFS_SCRUB_TYPE_NR)
+		goto out;
+	ops = &meta_scrub_ops[sm->sm_type];
+	if (ops->setup == NULL || ops->scrub == NULL)
+		goto out;
+	/* Does this fs even support this type of metadata? */
+	if (ops->has && !ops->has(&mp->m_sb))
+		goto out;
+
+	error = -EINVAL;
+	/* restricting fields must be appropriate for type */
+	switch (ops->type) {
+	case ST_NONE:
+	case ST_FS:
+		if (sm->sm_ino || sm->sm_gen || sm->sm_agno)
+			goto out;
+		break;
+	case ST_PERAG:
+		if (sm->sm_ino || sm->sm_gen ||
+		    sm->sm_agno >= mp->m_sb.sb_agcount)
+			goto out;
+		break;
+	case ST_INODE:
+		if (sm->sm_agno || (sm->sm_gen && !sm->sm_ino))
+			goto out;
+		break;
+	default:
+		goto out;
+	}
+
+	error = -EOPNOTSUPP;
+	/*
+	 * We won't scrub any filesystem that doesn't have the ability
+	 * to record unwritten extents.  The option was made default in
+	 * 2003, removed from mkfs in 2007, and cannot be disabled in
+	 * v5, so if we find a filesystem without this flag it's either
+	 * really old or totally unsupported.  Avoid it either way.
+	 * We also don't support v1-v3 filesystems, which aren't
+	 * mountable.
+	 */
+	if (!xfs_sb_version_hasextflgbit(&mp->m_sb))
+		goto out;
+
+	/* We don't know how to repair anything yet. */
+	if (sm->sm_flags & XFS_SCRUB_IFLAG_REPAIR)
+		goto out;
+
+	error = 0;
+out:
+	return error;
+}
+
+/* Dispatch metadata scrubbing. */
+int
+xfs_scrub_metadata(
+	struct xfs_inode		*ip,
+	struct xfs_scrub_metadata	*sm)
+{
+	struct xfs_scrub_context	sc;
+	struct xfs_mount		*mp = ip->i_mount;
+	bool				try_harder = false;
+	int				error = 0;
+
+	BUILD_BUG_ON(sizeof(meta_scrub_ops) !=
+		(sizeof(struct xfs_scrub_meta_ops) * XFS_SCRUB_TYPE_NR));
+
+	trace_xfs_scrub_start(ip, sm, error);
+
+	/* Forbidden if we are shut down or mounted norecovery. */
+	error = -ESHUTDOWN;
+	if (XFS_FORCED_SHUTDOWN(mp))
+		goto out;
+	error = -ENOTRECOVERABLE;
+	if (mp->m_flags & XFS_MOUNT_NORECOVERY)
+		goto out;
+
+	error = xfs_scrub_validate_inputs(mp, sm);
+	if (error)
+		goto out;
+
+	xfs_scrub_experimental_warning(mp);
+
+retry_op:
+	/* Set up for the operation. */
+	memset(&sc, 0, sizeof(sc));
+	sc.mp = ip->i_mount;
+	sc.sm = sm;
+	sc.ops = &meta_scrub_ops[sm->sm_type];
+	sc.try_harder = try_harder;
+	sc.sa.agno = NULLAGNUMBER;
+	error = sc.ops->setup(&sc, ip);
+	if (error)
+		goto out_teardown;
+
+	/* Scrub for errors. */
+	error = sc.ops->scrub(&sc);
+	if (!try_harder && error == -EDEADLOCK) {
+		/*
+		 * Scrubbers return -EDEADLOCK to mean 'try harder'.
+		 * Tear down everything we hold, then set up again with
+		 * preparation for worst-case scenarios.
+		 */
+		error = xfs_scrub_teardown(&sc, ip, 0);
+		if (error)
+			goto out;
+		try_harder = true;
+		goto retry_op;
+	} else if (error)
+		goto out_teardown;
+
+	if (sc.sm->sm_flags & (XFS_SCRUB_OFLAG_CORRUPT |
+			       XFS_SCRUB_OFLAG_XCORRUPT))
+		xfs_alert_ratelimited(mp, "Corruption detected during scrub.");
+
+out_teardown:
+	error = xfs_scrub_teardown(&sc, ip, error);
+out:
+	trace_xfs_scrub_done(ip, sm, error);
+	if (error == -EFSCORRUPTED || error == -EFSBADCRC) {
+		sm->sm_flags |= XFS_SCRUB_OFLAG_CORRUPT;
+		error = 0;
+	}
+	return error;
+}
diff --git a/fs/xfs/scrub/scrub.h b/fs/xfs/scrub/scrub.h
new file mode 100644
index 000000000000..0d92af86f67a
--- /dev/null
+++ b/fs/xfs/scrub/scrub.h
@@ -0,0 +1,152 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef __XFS_SCRUB_SCRUB_H__
+#define __XFS_SCRUB_SCRUB_H__
+
+struct xfs_scrub_context;
+
+/* Type info and names for the scrub types. */
+enum xfs_scrub_type {
+	ST_NONE = 1,	/* disabled */
+	ST_PERAG,	/* per-AG metadata */
+	ST_FS,		/* per-FS metadata */
+	ST_INODE,	/* per-inode metadata */
+};
+
+struct xfs_scrub_meta_ops {
+	/* Acquire whatever resources are needed for the operation. */
+	int		(*setup)(struct xfs_scrub_context *,
+				 struct xfs_inode *);
+
+	/* Examine metadata for errors. */
+	int		(*scrub)(struct xfs_scrub_context *);
+
+	/* Decide if we even have this piece of metadata. */
+	bool		(*has)(struct xfs_sb *);
+
+	/* type describing required/allowed inputs */
+	enum xfs_scrub_type	type;
+};
+
+/* Buffer pointers and btree cursors for an entire AG. */
+struct xfs_scrub_ag {
+	xfs_agnumber_t			agno;
+
+	/* AG btree roots */
+	struct xfs_buf			*agf_bp;
+	struct xfs_buf			*agfl_bp;
+	struct xfs_buf			*agi_bp;
+
+	/* AG btrees */
+	struct xfs_btree_cur		*bno_cur;
+	struct xfs_btree_cur		*cnt_cur;
+	struct xfs_btree_cur		*ino_cur;
+	struct xfs_btree_cur		*fino_cur;
+	struct xfs_btree_cur		*rmap_cur;
+	struct xfs_btree_cur		*refc_cur;
+};
+
+struct xfs_scrub_context {
+	/* General scrub state. */
+	struct xfs_mount		*mp;
+	struct xfs_scrub_metadata	*sm;
+	const struct xfs_scrub_meta_ops	*ops;
+	struct xfs_trans		*tp;
+	struct xfs_inode		*ip;
+	void				*buf;
+	uint				ilock_flags;
+	bool				try_harder;
+
+	/* State tracking for single-AG operations. */
+	struct xfs_scrub_ag		sa;
+};
+
+/* Metadata scrubbers */
+int xfs_scrub_tester(struct xfs_scrub_context *sc);
+int xfs_scrub_superblock(struct xfs_scrub_context *sc);
+int xfs_scrub_agf(struct xfs_scrub_context *sc);
+int xfs_scrub_agfl(struct xfs_scrub_context *sc);
+int xfs_scrub_agi(struct xfs_scrub_context *sc);
+int xfs_scrub_bnobt(struct xfs_scrub_context *sc);
+int xfs_scrub_cntbt(struct xfs_scrub_context *sc);
+int xfs_scrub_inobt(struct xfs_scrub_context *sc);
+int xfs_scrub_finobt(struct xfs_scrub_context *sc);
+int xfs_scrub_rmapbt(struct xfs_scrub_context *sc);
+int xfs_scrub_refcountbt(struct xfs_scrub_context *sc);
+int xfs_scrub_inode(struct xfs_scrub_context *sc);
+int xfs_scrub_bmap_data(struct xfs_scrub_context *sc);
+int xfs_scrub_bmap_attr(struct xfs_scrub_context *sc);
+int xfs_scrub_bmap_cow(struct xfs_scrub_context *sc);
+int xfs_scrub_directory(struct xfs_scrub_context *sc);
+int xfs_scrub_xattr(struct xfs_scrub_context *sc);
+int xfs_scrub_symlink(struct xfs_scrub_context *sc);
+int xfs_scrub_parent(struct xfs_scrub_context *sc);
+#ifdef CONFIG_XFS_RT
+int xfs_scrub_rtbitmap(struct xfs_scrub_context *sc);
+int xfs_scrub_rtsummary(struct xfs_scrub_context *sc);
+#else
+static inline int
+xfs_scrub_rtbitmap(struct xfs_scrub_context *sc)
+{
+	return -ENOENT;
+}
+static inline int
+xfs_scrub_rtsummary(struct xfs_scrub_context *sc)
+{
+	return -ENOENT;
+}
+#endif
+#ifdef CONFIG_XFS_QUOTA
+int xfs_scrub_quota(struct xfs_scrub_context *sc);
+#else
+static inline int
+xfs_scrub_quota(struct xfs_scrub_context *sc)
+{
+	return -ENOENT;
+}
+#endif
+
+/* cross-referencing helpers */
+void xfs_scrub_xref_is_used_space(struct xfs_scrub_context *sc,
+		xfs_agblock_t agbno, xfs_extlen_t len);
+void xfs_scrub_xref_is_not_inode_chunk(struct xfs_scrub_context *sc,
+		xfs_agblock_t agbno, xfs_extlen_t len);
+void xfs_scrub_xref_is_inode_chunk(struct xfs_scrub_context *sc,
+		xfs_agblock_t agbno, xfs_extlen_t len);
+void xfs_scrub_xref_is_owned_by(struct xfs_scrub_context *sc,
+		xfs_agblock_t agbno, xfs_extlen_t len,
+		struct xfs_owner_info *oinfo);
+void xfs_scrub_xref_is_not_owned_by(struct xfs_scrub_context *sc,
+		xfs_agblock_t agbno, xfs_extlen_t len,
+		struct xfs_owner_info *oinfo);
+void xfs_scrub_xref_has_no_owner(struct xfs_scrub_context *sc,
+		xfs_agblock_t agbno, xfs_extlen_t len);
+void xfs_scrub_xref_is_cow_staging(struct xfs_scrub_context *sc,
+		xfs_agblock_t bno, xfs_extlen_t len);
+void xfs_scrub_xref_is_not_shared(struct xfs_scrub_context *sc,
+		xfs_agblock_t bno, xfs_extlen_t len);
+#ifdef CONFIG_XFS_RT
+void xfs_scrub_xref_is_used_rt_space(struct xfs_scrub_context *sc,
+		xfs_rtblock_t rtbno, xfs_extlen_t len);
+#else
+# define xfs_scrub_xref_is_used_rt_space(sc, rtbno, len) do { } while (0)
+#endif
+
+#endif	/* __XFS_SCRUB_SCRUB_H__ */
diff --git a/fs/xfs/scrub/symlink.c b/fs/xfs/scrub/symlink.c
new file mode 100644
index 000000000000..3aa3d60f7c16
--- /dev/null
+++ b/fs/xfs/scrub/symlink.c
@@ -0,0 +1,92 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#include "xfs.h"
+#include "xfs_fs.h"
+#include "xfs_shared.h"
+#include "xfs_format.h"
+#include "xfs_trans_resv.h"
+#include "xfs_mount.h"
+#include "xfs_defer.h"
+#include "xfs_btree.h"
+#include "xfs_bit.h"
+#include "xfs_log_format.h"
+#include "xfs_trans.h"
+#include "xfs_sb.h"
+#include "xfs_inode.h"
+#include "xfs_inode_fork.h"
+#include "xfs_symlink.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+#include "scrub/trace.h"
+
+/* Set us up to scrub a symbolic link. */
+int
+xfs_scrub_setup_symlink(
+	struct xfs_scrub_context	*sc,
+	struct xfs_inode		*ip)
+{
+	/* Allocate the buffer without the inode lock held. */
+	sc->buf = kmem_zalloc_large(XFS_SYMLINK_MAXLEN + 1, KM_SLEEP);
+	if (!sc->buf)
+		return -ENOMEM;
+
+	return xfs_scrub_setup_inode_contents(sc, ip, 0);
+}
+
+/* Symbolic links. */
+
+int
+xfs_scrub_symlink(
+	struct xfs_scrub_context	*sc)
+{
+	struct xfs_inode		*ip = sc->ip;
+	struct xfs_ifork		*ifp;
+	loff_t				len;
+	int				error = 0;
+
+	if (!S_ISLNK(VFS_I(ip)->i_mode))
+		return -ENOENT;
+	ifp = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	len = ip->i_d.di_size;
+
+	/* Plausible size? */
+	if (len > XFS_SYMLINK_MAXLEN || len <= 0) {
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		goto out;
+	}
+
+	/* Inline symlink? */
+	if (ifp->if_flags & XFS_IFINLINE) {
+		if (len > XFS_IFORK_DSIZE(ip) ||
+		    len > strnlen(ifp->if_u1.if_data, XFS_IFORK_DSIZE(ip)))
+			xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+		goto out;
+	}
+
+	/* Remote symlink; must read the contents. */
+	error = xfs_readlink_bmap_ilocked(sc->ip, sc->buf);
+	if (!xfs_scrub_fblock_process_error(sc, XFS_DATA_FORK, 0, &error))
+		goto out;
+	if (strnlen(sc->buf, XFS_SYMLINK_MAXLEN) < len)
+		xfs_scrub_fblock_set_corrupt(sc, XFS_DATA_FORK, 0);
+out:
+	return error;
+}
diff --git a/fs/xfs/scrub/trace.c b/fs/xfs/scrub/trace.c
new file mode 100644
index 000000000000..86daed0e3a45
--- /dev/null
+++ b/fs/xfs/scrub/trace.c
@@ -0,0 +1,58 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#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_defer.h"
+#include "xfs_da_format.h"
+#include "xfs_inode.h"
+#include "xfs_btree.h"
+#include "xfs_trans.h"
+#include "xfs_bit.h"
+#include "scrub/xfs_scrub.h"
+#include "scrub/scrub.h"
+#include "scrub/common.h"
+
+/* Figure out which block the btree cursor was pointing to. */
+static inline xfs_fsblock_t
+xfs_scrub_btree_cur_fsbno(
+	struct xfs_btree_cur		*cur,
+	int				level)
+{
+	if (level < cur->bc_nlevels && cur->bc_bufs[level])
+		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);
+	else if (!(cur->bc_flags & XFS_BTREE_LONG_PTRS))
+		return XFS_AGB_TO_FSB(cur->bc_mp, cur->bc_private.a.agno, 0);
+	return NULLFSBLOCK;
+}
+
+/*
+ * We include this last to have the helpers above available for the trace
+ * event implementations.
+ */
+#define CREATE_TRACE_POINTS
+#include "scrub/trace.h"
diff --git a/fs/xfs/scrub/trace.h b/fs/xfs/scrub/trace.h
new file mode 100644
index 000000000000..5d2b1c241be5
--- /dev/null
+++ b/fs/xfs/scrub/trace.h
@@ -0,0 +1,500 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#undef TRACE_SYSTEM
+#define TRACE_SYSTEM xfs_scrub
+
+#if !defined(_TRACE_XFS_SCRUB_TRACE_H) || defined(TRACE_HEADER_MULTI_READ)
+#define _TRACE_XFS_SCRUB_TRACE_H
+
+#include <linux/tracepoint.h>
+#include "xfs_bit.h"
+
+DECLARE_EVENT_CLASS(xfs_scrub_class,
+	TP_PROTO(struct xfs_inode *ip, struct xfs_scrub_metadata *sm,
+		 int error),
+	TP_ARGS(ip, sm, error),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned int, type)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_ino_t, inum)
+		__field(unsigned int, gen)
+		__field(unsigned int, flags)
+		__field(int, error)
+	),
+	TP_fast_assign(
+		__entry->dev = ip->i_mount->m_super->s_dev;
+		__entry->ino = ip->i_ino;
+		__entry->type = sm->sm_type;
+		__entry->agno = sm->sm_agno;
+		__entry->inum = sm->sm_ino;
+		__entry->gen = sm->sm_gen;
+		__entry->flags = sm->sm_flags;
+		__entry->error = error;
+	),
+	TP_printk("dev %d:%d ino 0x%llx type %u agno %u inum %llu gen %u flags 0x%x error %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->type,
+		  __entry->agno,
+		  __entry->inum,
+		  __entry->gen,
+		  __entry->flags,
+		  __entry->error)
+)
+#define DEFINE_SCRUB_EVENT(name) \
+DEFINE_EVENT(xfs_scrub_class, name, \
+	TP_PROTO(struct xfs_inode *ip, struct xfs_scrub_metadata *sm, \
+		 int error), \
+	TP_ARGS(ip, sm, error))
+
+DEFINE_SCRUB_EVENT(xfs_scrub_start);
+DEFINE_SCRUB_EVENT(xfs_scrub_done);
+DEFINE_SCRUB_EVENT(xfs_scrub_deadlock_retry);
+
+TRACE_EVENT(xfs_scrub_op_error,
+	TP_PROTO(struct xfs_scrub_context *sc, xfs_agnumber_t agno,
+		 xfs_agblock_t bno, int error, void *ret_ip),
+	TP_ARGS(sc, agno, bno, error, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, type)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, bno)
+		__field(int, error)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->type = sc->sm->sm_type;
+		__entry->agno = agno;
+		__entry->bno = bno;
+		__entry->error = error;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d type %u agno %u agbno %u error %d ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->type,
+		  __entry->agno,
+		  __entry->bno,
+		  __entry->error,
+		  __entry->ret_ip)
+);
+
+TRACE_EVENT(xfs_scrub_file_op_error,
+	TP_PROTO(struct xfs_scrub_context *sc, int whichfork,
+		 xfs_fileoff_t offset, int error, void *ret_ip),
+	TP_ARGS(sc, whichfork, offset, error, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, whichfork)
+		__field(unsigned int, type)
+		__field(xfs_fileoff_t, offset)
+		__field(int, error)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->ip->i_mount->m_super->s_dev;
+		__entry->ino = sc->ip->i_ino;
+		__entry->whichfork = whichfork;
+		__entry->type = sc->sm->sm_type;
+		__entry->offset = offset;
+		__entry->error = error;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d ino 0x%llx fork %d type %u offset %llu error %d ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->whichfork,
+		  __entry->type,
+		  __entry->offset,
+		  __entry->error,
+		  __entry->ret_ip)
+);
+
+DECLARE_EVENT_CLASS(xfs_scrub_block_error_class,
+	TP_PROTO(struct xfs_scrub_context *sc, xfs_daddr_t daddr, void *ret_ip),
+	TP_ARGS(sc, daddr, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, type)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, bno)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		xfs_fsblock_t	fsbno;
+		xfs_agnumber_t	agno;
+		xfs_agblock_t	bno;
+
+		fsbno = XFS_DADDR_TO_FSB(sc->mp, daddr);
+		agno = XFS_FSB_TO_AGNO(sc->mp, fsbno);
+		bno = XFS_FSB_TO_AGBNO(sc->mp, fsbno);
+
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->type = sc->sm->sm_type;
+		__entry->agno = agno;
+		__entry->bno = bno;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d type %u agno %u agbno %u ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->type,
+		  __entry->agno,
+		  __entry->bno,
+		  __entry->ret_ip)
+)
+
+#define DEFINE_SCRUB_BLOCK_ERROR_EVENT(name) \
+DEFINE_EVENT(xfs_scrub_block_error_class, name, \
+	TP_PROTO(struct xfs_scrub_context *sc, xfs_daddr_t daddr, \
+		 void *ret_ip), \
+	TP_ARGS(sc, daddr, ret_ip))
+
+DEFINE_SCRUB_BLOCK_ERROR_EVENT(xfs_scrub_block_error);
+DEFINE_SCRUB_BLOCK_ERROR_EVENT(xfs_scrub_block_preen);
+
+DECLARE_EVENT_CLASS(xfs_scrub_ino_error_class,
+	TP_PROTO(struct xfs_scrub_context *sc, xfs_ino_t ino, void *ret_ip),
+	TP_ARGS(sc, ino, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(unsigned int, type)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->ino = ino;
+		__entry->type = sc->sm->sm_type;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d ino 0x%llx type %u ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->type,
+		  __entry->ret_ip)
+)
+
+#define DEFINE_SCRUB_INO_ERROR_EVENT(name) \
+DEFINE_EVENT(xfs_scrub_ino_error_class, name, \
+	TP_PROTO(struct xfs_scrub_context *sc, xfs_ino_t ino, \
+		 void *ret_ip), \
+	TP_ARGS(sc, ino, ret_ip))
+
+DEFINE_SCRUB_INO_ERROR_EVENT(xfs_scrub_ino_error);
+DEFINE_SCRUB_INO_ERROR_EVENT(xfs_scrub_ino_preen);
+DEFINE_SCRUB_INO_ERROR_EVENT(xfs_scrub_ino_warning);
+
+DECLARE_EVENT_CLASS(xfs_scrub_fblock_error_class,
+	TP_PROTO(struct xfs_scrub_context *sc, int whichfork,
+		 xfs_fileoff_t offset, void *ret_ip),
+	TP_ARGS(sc, whichfork, offset, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, whichfork)
+		__field(unsigned int, type)
+		__field(xfs_fileoff_t, offset)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->ip->i_mount->m_super->s_dev;
+		__entry->ino = sc->ip->i_ino;
+		__entry->whichfork = whichfork;
+		__entry->type = sc->sm->sm_type;
+		__entry->offset = offset;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d ino 0x%llx fork %d type %u offset %llu ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->whichfork,
+		  __entry->type,
+		  __entry->offset,
+		  __entry->ret_ip)
+);
+
+#define DEFINE_SCRUB_FBLOCK_ERROR_EVENT(name) \
+DEFINE_EVENT(xfs_scrub_fblock_error_class, name, \
+	TP_PROTO(struct xfs_scrub_context *sc, int whichfork, \
+		 xfs_fileoff_t offset, void *ret_ip), \
+	TP_ARGS(sc, whichfork, offset, ret_ip))
+
+DEFINE_SCRUB_FBLOCK_ERROR_EVENT(xfs_scrub_fblock_error);
+DEFINE_SCRUB_FBLOCK_ERROR_EVENT(xfs_scrub_fblock_warning);
+
+TRACE_EVENT(xfs_scrub_incomplete,
+	TP_PROTO(struct xfs_scrub_context *sc, void *ret_ip),
+	TP_ARGS(sc, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, type)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->type = sc->sm->sm_type;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d type %u ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->type,
+		  __entry->ret_ip)
+);
+
+TRACE_EVENT(xfs_scrub_btree_op_error,
+	TP_PROTO(struct xfs_scrub_context *sc, struct xfs_btree_cur *cur,
+		 int level, int error, void *ret_ip),
+	TP_ARGS(sc, cur, level, error, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, type)
+		__field(xfs_btnum_t, btnum)
+		__field(int, level)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, bno)
+		__field(int, ptr);
+		__field(int, error)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		xfs_fsblock_t fsbno = xfs_scrub_btree_cur_fsbno(cur, level);
+
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->type = sc->sm->sm_type;
+		__entry->btnum = cur->bc_btnum;
+		__entry->level = level;
+		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
+		__entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
+		__entry->ptr = cur->bc_ptrs[level];
+		__entry->error = error;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d type %u btnum %d level %d ptr %d agno %u agbno %u error %d ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->type,
+		  __entry->btnum,
+		  __entry->level,
+		  __entry->ptr,
+		  __entry->agno,
+		  __entry->bno,
+		  __entry->error,
+		  __entry->ret_ip)
+);
+
+TRACE_EVENT(xfs_scrub_ifork_btree_op_error,
+	TP_PROTO(struct xfs_scrub_context *sc, struct xfs_btree_cur *cur,
+		 int level, int error, void *ret_ip),
+	TP_ARGS(sc, cur, level, error, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, whichfork)
+		__field(unsigned int, type)
+		__field(xfs_btnum_t, btnum)
+		__field(int, level)
+		__field(int, ptr)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, bno)
+		__field(int, error)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		xfs_fsblock_t fsbno = xfs_scrub_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->type = sc->sm->sm_type;
+		__entry->btnum = cur->bc_btnum;
+		__entry->level = level;
+		__entry->ptr = cur->bc_ptrs[level];
+		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
+		__entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
+		__entry->error = error;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d ino 0x%llx fork %d type %u btnum %d level %d ptr %d agno %u agbno %u error %d ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->whichfork,
+		  __entry->type,
+		  __entry->btnum,
+		  __entry->level,
+		  __entry->ptr,
+		  __entry->agno,
+		  __entry->bno,
+		  __entry->error,
+		  __entry->ret_ip)
+);
+
+TRACE_EVENT(xfs_scrub_btree_error,
+	TP_PROTO(struct xfs_scrub_context *sc, struct xfs_btree_cur *cur,
+		 int level, void *ret_ip),
+	TP_ARGS(sc, cur, level, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(unsigned int, type)
+		__field(xfs_btnum_t, btnum)
+		__field(int, level)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, bno)
+		__field(int, ptr);
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		xfs_fsblock_t fsbno = xfs_scrub_btree_cur_fsbno(cur, level);
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->type = sc->sm->sm_type;
+		__entry->btnum = cur->bc_btnum;
+		__entry->level = level;
+		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
+		__entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
+		__entry->ptr = cur->bc_ptrs[level];
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d type %u btnum %d level %d ptr %d agno %u agbno %u ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->type,
+		  __entry->btnum,
+		  __entry->level,
+		  __entry->ptr,
+		  __entry->agno,
+		  __entry->bno,
+		  __entry->ret_ip)
+);
+
+TRACE_EVENT(xfs_scrub_ifork_btree_error,
+	TP_PROTO(struct xfs_scrub_context *sc, struct xfs_btree_cur *cur,
+		 int level, void *ret_ip),
+	TP_ARGS(sc, cur, level, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(xfs_ino_t, ino)
+		__field(int, whichfork)
+		__field(unsigned int, type)
+		__field(xfs_btnum_t, btnum)
+		__field(int, level)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, bno)
+		__field(int, ptr);
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		xfs_fsblock_t fsbno = xfs_scrub_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->type = sc->sm->sm_type;
+		__entry->btnum = cur->bc_btnum;
+		__entry->level = level;
+		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
+		__entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
+		__entry->ptr = cur->bc_ptrs[level];
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d ino 0x%llx fork %d type %u btnum %d level %d ptr %d agno %u agbno %u ret_ip %pS",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->ino,
+		  __entry->whichfork,
+		  __entry->type,
+		  __entry->btnum,
+		  __entry->level,
+		  __entry->ptr,
+		  __entry->agno,
+		  __entry->bno,
+		  __entry->ret_ip)
+);
+
+DECLARE_EVENT_CLASS(xfs_scrub_sbtree_class,
+	TP_PROTO(struct xfs_scrub_context *sc, struct xfs_btree_cur *cur,
+		 int level),
+	TP_ARGS(sc, cur, level),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(int, type)
+		__field(xfs_btnum_t, btnum)
+		__field(xfs_agnumber_t, agno)
+		__field(xfs_agblock_t, bno)
+		__field(int, level)
+		__field(int, nlevels)
+		__field(int, ptr)
+	),
+	TP_fast_assign(
+		xfs_fsblock_t fsbno = xfs_scrub_btree_cur_fsbno(cur, level);
+
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->type = sc->sm->sm_type;
+		__entry->btnum = cur->bc_btnum;
+		__entry->agno = XFS_FSB_TO_AGNO(cur->bc_mp, fsbno);
+		__entry->bno = XFS_FSB_TO_AGBNO(cur->bc_mp, fsbno);
+		__entry->level = level;
+		__entry->nlevels = cur->bc_nlevels;
+		__entry->ptr = cur->bc_ptrs[level];
+	),
+	TP_printk("dev %d:%d type %u btnum %d agno %u agbno %u level %d nlevels %d ptr %d",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->type,
+		  __entry->btnum,
+		  __entry->agno,
+		  __entry->bno,
+		  __entry->level,
+		  __entry->nlevels,
+		  __entry->ptr)
+)
+#define DEFINE_SCRUB_SBTREE_EVENT(name) \
+DEFINE_EVENT(xfs_scrub_sbtree_class, name, \
+	TP_PROTO(struct xfs_scrub_context *sc, struct xfs_btree_cur *cur, \
+		 int level), \
+	TP_ARGS(sc, cur, level))
+
+DEFINE_SCRUB_SBTREE_EVENT(xfs_scrub_btree_rec);
+DEFINE_SCRUB_SBTREE_EVENT(xfs_scrub_btree_key);
+
+TRACE_EVENT(xfs_scrub_xref_error,
+	TP_PROTO(struct xfs_scrub_context *sc, int error, void *ret_ip),
+	TP_ARGS(sc, error, ret_ip),
+	TP_STRUCT__entry(
+		__field(dev_t, dev)
+		__field(int, type)
+		__field(int, error)
+		__field(void *, ret_ip)
+	),
+	TP_fast_assign(
+		__entry->dev = sc->mp->m_super->s_dev;
+		__entry->type = sc->sm->sm_type;
+		__entry->error = error;
+		__entry->ret_ip = ret_ip;
+	),
+	TP_printk("dev %d:%d type %u xref error %d ret_ip %pF",
+		  MAJOR(__entry->dev), MINOR(__entry->dev),
+		  __entry->type,
+		  __entry->error,
+		  __entry->ret_ip)
+);
+
+#endif /* _TRACE_XFS_SCRUB_TRACE_H */
+
+#undef TRACE_INCLUDE_PATH
+#define TRACE_INCLUDE_PATH .
+#define TRACE_INCLUDE_FILE scrub/trace
+#include <trace/define_trace.h>
diff --git a/fs/xfs/scrub/xfs_scrub.h b/fs/xfs/scrub/xfs_scrub.h
new file mode 100644
index 000000000000..e00e0eadac6a
--- /dev/null
+++ b/fs/xfs/scrub/xfs_scrub.h
@@ -0,0 +1,29 @@
+/*
+ * Copyright (C) 2017 Oracle.  All Rights Reserved.
+ *
+ * Author: Darrick J. Wong <darrick.wong@oracle.com>
+ *
+ * This program is free software; you can redistribute it and/or
+ * modify it under the terms of the GNU General Public License
+ * as published by the Free Software Foundation; either version 2
+ * of the License, or (at your option) any later version.
+ *
+ * This program is distributed in the hope that it would be useful,
+ * but WITHOUT ANY WARRANTY; without even the implied warranty of
+ * MERCHANTABILITY or FITNESS FOR A PARTICULAR PURPOSE.  See the
+ * GNU General Public License for more details.
+ *
+ * You should have received a copy of the GNU General Public License
+ * along with this program; if not, write the Free Software Foundation,
+ * Inc.,  51 Franklin St, Fifth Floor, Boston, MA  02110-1301, USA.
+ */
+#ifndef __XFS_SCRUB_H__
+#define __XFS_SCRUB_H__
+
+#ifndef CONFIG_XFS_ONLINE_SCRUB
+# define xfs_scrub_metadata(ip, sm)	(-ENOTTY)
+#else
+int xfs_scrub_metadata(struct xfs_inode *ip, struct xfs_scrub_metadata *sm);
+#endif /* CONFIG_XFS_ONLINE_SCRUB */
+
+#endif	/* __XFS_SCRUB_H__ */