Merge tag 'xfs-4.17-merge-1' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux

Pull xfs updates from Darrick Wong:
 "Here's the first round of fixes for XFS for 4.17.

  The biggest new features this time around are the addition of lazytime
  support, further enhancement of the on-disk inode metadata verifiers,
  and a patch to smooth over some of the AGFL padding problems that have
  intermittently plagued users since 4.5. I forsee sending a second pull
  request next week with further bug fixes and speedups in the online
  scrub code and elsewhere.

  This series has been run through a full xfstests run over the weekend
  and through a quick xfstests run against this morning's master, with
  no major failures reported.

  Summary of changes for this release:

   - Various cleanups and code fixes

   - Implement lazytime as a mount option

   - Convert various on-disk metadata checks from asserts to -EFSCORRUPTED

   - Fix accounting problems with the rmap per-ag reservations

   - Refactorings and cleanups for xfs_log_force

   - Various bugfixes for the reflink code

   - Work around v5 AGFL padding problems to prevent fs shutdowns

   - Establish inode fork verifiers to inspect on-disk metadata
     correctness

   - Various online scrub fixes

   - Fix v5 swapext blowing up on deleted inodes"

* tag 'xfs-4.17-merge-1' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux: (49 commits)
  xfs: do not log/recover swapext extent owner changes for deleted inodes
  xfs: clean up xfs_mount allocation and dynamic initializers
  xfs: remove dead inode version setting code
  xfs: catch inode allocation state mismatch corruption
  xfs: xfs_scrub_iallocbt_xref_rmap_inodes should use xref_set_corrupt
  xfs: flag inode corruption if parent ptr doesn't get us a real inode
  xfs: don't accept inode buffers with suspicious unlinked chains
  xfs: move inode extent size hint validation to libxfs
  xfs: record inode buf errors as a xref error in inobt scrubber
  xfs: remove xfs_buf parameter from inode scrub methods
  xfs: inode scrubber shouldn't bother with raw checks
  xfs: bmap scrubber should do rmap xref with bmap for sparse files
  xfs: refactor inode buffer verifier error logging
  xfs: refactor inode verifier error logging
  xfs: refactor bmap record validation
  xfs: sanity-check the unused space before trying to use it
  xfs: detect agfl count corruption and reset agfl
  xfs: unwind the try_again loop in xfs_log_force
  xfs: refactor xfs_log_force_lsn
  xfs: minor cleanup for xfs_reflink_end_cow
  ...

fs/inode.c

@@ -346,9 +346,8 @@ void inc_nlink(struct inode *inode)
 }
 EXPORT_SYMBOL(inc_nlink);
 
-void address_space_init_once(struct address_space *mapping)
+static void __address_space_init_once(struct address_space *mapping)
 {
-	memset(mapping, 0, sizeof(*mapping));
 	INIT_RADIX_TREE(&mapping->page_tree, GFP_ATOMIC | __GFP_ACCOUNT);
 	spin_lock_init(&mapping->tree_lock);
 	init_rwsem(&mapping->i_mmap_rwsem);
@@ -356,6 +355,12 @@ void address_space_init_once(struct address_space *mapping)
 	spin_lock_init(&mapping->private_lock);
 	mapping->i_mmap = RB_ROOT_CACHED;
 }
+
+void address_space_init_once(struct address_space *mapping)
+{
+	memset(mapping, 0, sizeof(*mapping));
+	__address_space_init_once(mapping);
+}
 EXPORT_SYMBOL(address_space_init_once);
 
 /*
@@ -371,7 +376,7 @@ void inode_init_once(struct inode *inode)
 	INIT_LIST_HEAD(&inode->i_io_list);
 	INIT_LIST_HEAD(&inode->i_wb_list);
 	INIT_LIST_HEAD(&inode->i_lru);
-	address_space_init_once(&inode->i_data);
+	__address_space_init_once(&inode->i_data);
 	i_size_ordered_init(inode);
 }
 EXPORT_SYMBOL(inode_init_once);
@@ -1533,7 +1538,6 @@ retry:
 	if (atomic_dec_and_lock(&inode->i_count, &inode->i_lock)) {
 		if (inode->i_nlink && (inode->i_state & I_DIRTY_TIME)) {
 			atomic_inc(&inode->i_count);
-			inode->i_state &= ~I_DIRTY_TIME;
 			spin_unlock(&inode->i_lock);
 			trace_writeback_lazytime_iput(inode);
 			mark_inode_dirty_sync(inode);

fs/sync.c

@@ -192,12 +192,8 @@ int vfs_fsync_range(struct file *file, loff_t start, loff_t end, int datasync)
 
 	if (!file->f_op->fsync)
 		return -EINVAL;
-	if (!datasync && (inode->i_state & I_DIRTY_TIME)) {
-		spin_lock(&inode->i_lock);
-		inode->i_state &= ~I_DIRTY_TIME;
-		spin_unlock(&inode->i_lock);
+	if (!datasync && (inode->i_state & I_DIRTY_TIME))
 		mark_inode_dirty_sync(inode);
-	}
 	return file->f_op->fsync(file, start, end, datasync);
 }
 EXPORT_SYMBOL(vfs_fsync_range);

fs/xfs/kmem.c

@@ -46,13 +46,13 @@ kmem_alloc(size_t size, xfs_km_flags_t flags)
 }
 
 void *
-kmem_zalloc_large(size_t size, xfs_km_flags_t flags)
+kmem_alloc_large(size_t size, xfs_km_flags_t flags)
 {
 	unsigned nofs_flag = 0;
 	void	*ptr;
 	gfp_t	lflags;
 
-	ptr = kmem_zalloc(size, flags | KM_MAYFAIL);
+	ptr = kmem_alloc(size, flags | KM_MAYFAIL);
 	if (ptr)
 		return ptr;
 
@@ -67,7 +67,7 @@ kmem_zalloc_large(size_t size, xfs_km_flags_t flags)
 		nofs_flag = memalloc_nofs_save();
 
 	lflags = kmem_flags_convert(flags);
-	ptr = __vmalloc(size, lflags | __GFP_ZERO, PAGE_KERNEL);
+	ptr = __vmalloc(size, lflags, PAGE_KERNEL);
 
 	if (flags & KM_NOFS)
 		memalloc_nofs_restore(nofs_flag);

fs/xfs/kmem.h

@@ -71,7 +71,7 @@ kmem_flags_convert(xfs_km_flags_t flags)
 }
 
 extern void *kmem_alloc(size_t, xfs_km_flags_t);
-extern void *kmem_zalloc_large(size_t size, xfs_km_flags_t);
+extern void *kmem_alloc_large(size_t size, xfs_km_flags_t);
 extern void *kmem_realloc(const void *, size_t, xfs_km_flags_t);
 static inline void  kmem_free(const void *ptr)
 {
@@ -85,6 +85,12 @@ kmem_zalloc(size_t size, xfs_km_flags_t flags)
 	return kmem_alloc(size, flags | KM_ZERO);
 }
 
+static inline void *
+kmem_zalloc_large(size_t size, xfs_km_flags_t flags)
+{
+	return kmem_alloc_large(size, flags | KM_ZERO);
+}
+
 /*
  * Zone interfaces
  */

fs/xfs/libxfs/xfs_ag_resv.c

@@ -95,13 +95,13 @@ xfs_ag_resv_critical(
 
 	switch (type) {
 	case XFS_AG_RESV_METADATA:
-		avail = pag->pagf_freeblks - pag->pag_agfl_resv.ar_reserved;
+		avail = pag->pagf_freeblks - pag->pag_rmapbt_resv.ar_reserved;
 		orig = pag->pag_meta_resv.ar_asked;
 		break;
-	case XFS_AG_RESV_AGFL:
+	case XFS_AG_RESV_RMAPBT:
 		avail = pag->pagf_freeblks + pag->pagf_flcount -
 			pag->pag_meta_resv.ar_reserved;
-		orig = pag->pag_agfl_resv.ar_asked;
+		orig = pag->pag_rmapbt_resv.ar_asked;
 		break;
 	default:
 		ASSERT(0);
@@ -126,10 +126,10 @@ xfs_ag_resv_needed(
 {
 	xfs_extlen_t			len;
 
-	len = pag->pag_meta_resv.ar_reserved + pag->pag_agfl_resv.ar_reserved;
+	len = pag->pag_meta_resv.ar_reserved + pag->pag_rmapbt_resv.ar_reserved;
 	switch (type) {
 	case XFS_AG_RESV_METADATA:
-	case XFS_AG_RESV_AGFL:
+	case XFS_AG_RESV_RMAPBT:
 		len -= xfs_perag_resv(pag, type)->ar_reserved;
 		break;
 	case XFS_AG_RESV_NONE:
@@ -160,10 +160,11 @@ __xfs_ag_resv_free(
 	if (pag->pag_agno == 0)
 		pag->pag_mount->m_ag_max_usable += resv->ar_asked;
 	/*
-	 * AGFL blocks are always considered "free", so whatever
-	 * was reserved at mount time must be given back at umount.
+	 * RMAPBT blocks come from the AGFL and AGFL blocks are always
+	 * considered "free", so whatever was reserved at mount time must be
+	 * given back at umount.
 	 */
-	if (type == XFS_AG_RESV_AGFL)
+	if (type == XFS_AG_RESV_RMAPBT)
 		oldresv = resv->ar_orig_reserved;
 	else
 		oldresv = resv->ar_reserved;
@@ -185,7 +186,7 @@ xfs_ag_resv_free(
 	int				error;
 	int				err2;
 
-	error = __xfs_ag_resv_free(pag, XFS_AG_RESV_AGFL);
+	error = __xfs_ag_resv_free(pag, XFS_AG_RESV_RMAPBT);
 	err2 = __xfs_ag_resv_free(pag, XFS_AG_RESV_METADATA);
 	if (err2 && !error)
 		error = err2;
@@ -284,15 +285,15 @@ xfs_ag_resv_init(
 		}
 	}
 
-	/* Create the AGFL metadata reservation */
-	if (pag->pag_agfl_resv.ar_asked == 0) {
+	/* Create the RMAPBT metadata reservation */
+	if (pag->pag_rmapbt_resv.ar_asked == 0) {
 		ask = used = 0;
 
 		error = xfs_rmapbt_calc_reserves(mp, agno, &ask, &used);
 		if (error)
 			goto out;
 
-		error = __xfs_ag_resv_init(pag, XFS_AG_RESV_AGFL, ask, used);
+		error = __xfs_ag_resv_init(pag, XFS_AG_RESV_RMAPBT, ask, used);
 		if (error)
 			goto out;
 	}
@@ -304,7 +305,7 @@ xfs_ag_resv_init(
 		return error;
 
 	ASSERT(xfs_perag_resv(pag, XFS_AG_RESV_METADATA)->ar_reserved +
-	       xfs_perag_resv(pag, XFS_AG_RESV_AGFL)->ar_reserved <=
+	       xfs_perag_resv(pag, XFS_AG_RESV_RMAPBT)->ar_reserved <=
 	       pag->pagf_freeblks + pag->pagf_flcount);
 #endif
 out:
@@ -325,8 +326,10 @@ xfs_ag_resv_alloc_extent(
 	trace_xfs_ag_resv_alloc_extent(pag, type, args->len);
 
 	switch (type) {
-	case XFS_AG_RESV_METADATA:
 	case XFS_AG_RESV_AGFL:
+		return;
+	case XFS_AG_RESV_METADATA:
+	case XFS_AG_RESV_RMAPBT:
 		resv = xfs_perag_resv(pag, type);
 		break;
 	default:
@@ -341,7 +344,7 @@ xfs_ag_resv_alloc_extent(
 
 	len = min_t(xfs_extlen_t, args->len, resv->ar_reserved);
 	resv->ar_reserved -= len;
-	if (type == XFS_AG_RESV_AGFL)
+	if (type == XFS_AG_RESV_RMAPBT)
 		return;
 	/* Allocations of reserved blocks only need on-disk sb updates... */
 	xfs_trans_mod_sb(args->tp, XFS_TRANS_SB_RES_FDBLOCKS, -(int64_t)len);
@@ -365,8 +368,10 @@ xfs_ag_resv_free_extent(
 	trace_xfs_ag_resv_free_extent(pag, type, len);
 
 	switch (type) {
-	case XFS_AG_RESV_METADATA:
 	case XFS_AG_RESV_AGFL:
+		return;
+	case XFS_AG_RESV_METADATA:
+	case XFS_AG_RESV_RMAPBT:
 		resv = xfs_perag_resv(pag, type);
 		break;
 	default:
@@ -379,7 +384,7 @@ xfs_ag_resv_free_extent(
 
 	leftover = min_t(xfs_extlen_t, len, resv->ar_asked - resv->ar_reserved);
 	resv->ar_reserved += leftover;
-	if (type == XFS_AG_RESV_AGFL)
+	if (type == XFS_AG_RESV_RMAPBT)
 		return;
 	/* Freeing into the reserved pool only requires on-disk update... */
 	xfs_trans_mod_sb(tp, XFS_TRANS_SB_RES_FDBLOCKS, len);

fs/xfs/libxfs/xfs_ag_resv.h

@@ -32,4 +32,35 @@ void xfs_ag_resv_alloc_extent(struct xfs_perag *pag, enum xfs_ag_resv_type type,
 void xfs_ag_resv_free_extent(struct xfs_perag *pag, enum xfs_ag_resv_type type,
 		struct xfs_trans *tp, xfs_extlen_t len);
 
+/*
+ * RMAPBT reservation accounting wrappers. Since rmapbt blocks are sourced from
+ * the AGFL, they are allocated one at a time and the reservation updates don't
+ * require a transaction.
+ */
+static inline void
+xfs_ag_resv_rmapbt_alloc(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agno)
+{
+	struct xfs_alloc_arg	args = {0};
+	struct xfs_perag	*pag;
+
+	args.len = 1;
+	pag = xfs_perag_get(mp, agno);
+	xfs_ag_resv_alloc_extent(pag, XFS_AG_RESV_RMAPBT, &args);
+	xfs_perag_put(pag);
+}
+
+static inline void
+xfs_ag_resv_rmapbt_free(
+	struct xfs_mount	*mp,
+	xfs_agnumber_t		agno)
+{
+	struct xfs_perag	*pag;
+
+	pag = xfs_perag_get(mp, agno);
+	xfs_ag_resv_free_extent(pag, XFS_AG_RESV_RMAPBT, NULL, 1);
+	xfs_perag_put(pag);
+}
+
 #endif	/* __XFS_AG_RESV_H__ */

fs/xfs/libxfs/xfs_alloc.c

@@ -53,6 +53,23 @@ STATIC int xfs_alloc_ag_vextent_size(xfs_alloc_arg_t *);
 STATIC int xfs_alloc_ag_vextent_small(xfs_alloc_arg_t *,
 		xfs_btree_cur_t *, xfs_agblock_t *, xfs_extlen_t *, int *);
 
+/*
+ * Size of the AGFL.  For CRC-enabled filesystes we steal a couple of slots in
+ * the beginning of the block for a proper header with the location information
+ * and CRC.
+ */
+unsigned int
+xfs_agfl_size(
+	struct xfs_mount	*mp)
+{
+	unsigned int		size = mp->m_sb.sb_sectsize;
+
+	if (xfs_sb_version_hascrc(&mp->m_sb))
+		size -= sizeof(struct xfs_agfl);
+
+	return size / sizeof(xfs_agblock_t);
+}
+
 unsigned int
 xfs_refc_block(
 	struct xfs_mount	*mp)
@@ -550,7 +567,7 @@ xfs_agfl_verify(
 	if (bp->b_pag && be32_to_cpu(agfl->agfl_seqno) != bp->b_pag->pag_agno)
 		return __this_address;
 
-	for (i = 0; i < XFS_AGFL_SIZE(mp); i++) {
+	for (i = 0; i < xfs_agfl_size(mp); i++) {
 		if (be32_to_cpu(agfl->agfl_bno[i]) != NULLAGBLOCK &&
 		    be32_to_cpu(agfl->agfl_bno[i]) >= mp->m_sb.sb_agblocks)
 			return __this_address;
@@ -1564,7 +1581,6 @@ xfs_alloc_ag_vextent_small(
 	int		*stat)	/* status: 0-freelist, 1-normal/none */
 {
 	struct xfs_owner_info	oinfo;
-	struct xfs_perag	*pag;
 	int		error;
 	xfs_agblock_t	fbno;
 	xfs_extlen_t	flen;
@@ -1616,18 +1632,13 @@ xfs_alloc_ag_vextent_small(
 			/*
 			 * If we're feeding an AGFL block to something that
 			 * doesn't live in the free space, we need to clear
-			 * out the OWN_AG rmap and add the block back to
-			 * the AGFL per-AG reservation.
+			 * out the OWN_AG rmap.
 			 */
 			xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_AG);
 			error = xfs_rmap_free(args->tp, args->agbp, args->agno,
 					fbno, 1, &oinfo);
 			if (error)
 				goto error0;
-			pag = xfs_perag_get(args->mp, args->agno);
-			xfs_ag_resv_free_extent(pag, XFS_AG_RESV_AGFL,
-					args->tp, 1);
-			xfs_perag_put(pag);
 
 			*stat = 0;
 			return 0;
@@ -1911,14 +1922,12 @@ xfs_free_ag_extent(
 	XFS_STATS_INC(mp, xs_freex);
 	XFS_STATS_ADD(mp, xs_freeb, len);
 
-	trace_xfs_free_extent(mp, agno, bno, len, type == XFS_AG_RESV_AGFL,
-			haveleft, haveright);
+	trace_xfs_free_extent(mp, agno, bno, len, type, haveleft, haveright);
 
 	return 0;
 
  error0:
-	trace_xfs_free_extent(mp, agno, bno, len, type == XFS_AG_RESV_AGFL,
-			-1, -1);
+	trace_xfs_free_extent(mp, agno, bno, len, type, -1, -1);
 	if (bno_cur)
 		xfs_btree_del_cursor(bno_cur, XFS_BTREE_ERROR);
 	if (cnt_cur)
@@ -2053,6 +2062,93 @@ xfs_alloc_space_available(
 	return true;
 }
 
+/*
+ * Check the agfl fields of the agf for inconsistency or corruption. The purpose
+ * is to detect an agfl header padding mismatch between current and early v5
+ * kernels. This problem manifests as a 1-slot size difference between the
+ * on-disk flcount and the active [first, last] range of a wrapped agfl. This
+ * may also catch variants of agfl count corruption unrelated to padding. Either
+ * way, we'll reset the agfl and warn the user.
+ *
+ * Return true if a reset is required before the agfl can be used, false
+ * otherwise.
+ */
+static bool
+xfs_agfl_needs_reset(
+	struct xfs_mount	*mp,
+	struct xfs_agf		*agf)
+{
+	uint32_t		f = be32_to_cpu(agf->agf_flfirst);
+	uint32_t		l = be32_to_cpu(agf->agf_fllast);
+	uint32_t		c = be32_to_cpu(agf->agf_flcount);
+	int			agfl_size = xfs_agfl_size(mp);
+	int			active;
+
+	/* no agfl header on v4 supers */
+	if (!xfs_sb_version_hascrc(&mp->m_sb))
+		return false;
+
+	/*
+	 * The agf read verifier catches severe corruption of these fields.
+	 * Repeat some sanity checks to cover a packed -> unpacked mismatch if
+	 * the verifier allows it.
+	 */
+	if (f >= agfl_size || l >= agfl_size)
+		return true;
+	if (c > agfl_size)
+		return true;
+
+	/*
+	 * Check consistency between the on-disk count and the active range. An
+	 * agfl padding mismatch manifests as an inconsistent flcount.
+	 */
+	if (c && l >= f)
+		active = l - f + 1;
+	else if (c)
+		active = agfl_size - f + l + 1;
+	else
+		active = 0;
+
+	return active != c;
+}
+
+/*
+ * Reset the agfl to an empty state. Ignore/drop any existing blocks since the
+ * agfl content cannot be trusted. Warn the user that a repair is required to
+ * recover leaked blocks.
+ *
+ * The purpose of this mechanism is to handle filesystems affected by the agfl
+ * header padding mismatch problem. A reset keeps the filesystem online with a
+ * relatively minor free space accounting inconsistency rather than suffer the
+ * inevitable crash from use of an invalid agfl block.
+ */
+static void
+xfs_agfl_reset(
+	struct xfs_trans	*tp,
+	struct xfs_buf		*agbp,
+	struct xfs_perag	*pag)
+{
+	struct xfs_mount	*mp = tp->t_mountp;
+	struct xfs_agf		*agf = XFS_BUF_TO_AGF(agbp);
+
+	ASSERT(pag->pagf_agflreset);
+	trace_xfs_agfl_reset(mp, agf, 0, _RET_IP_);
+
+	xfs_warn(mp,
+	       "WARNING: Reset corrupted AGFL on AG %u. %d blocks leaked. "
+	       "Please unmount and run xfs_repair.",
+	         pag->pag_agno, pag->pagf_flcount);
+
+	agf->agf_flfirst = 0;
+	agf->agf_fllast = cpu_to_be32(xfs_agfl_size(mp) - 1);
+	agf->agf_flcount = 0;
+	xfs_alloc_log_agf(tp, agbp, XFS_AGF_FLFIRST | XFS_AGF_FLLAST |
+				    XFS_AGF_FLCOUNT);
+
+	pag->pagf_flcount = 0;
+	pag->pagf_agflreset = false;
+}
+
 /*
  * Decide whether to use this allocation group for this allocation.
  * If so, fix up the btree freelist's size.
@@ -2114,6 +2210,10 @@ xfs_alloc_fix_freelist(
 		}
 	}
 
+	/* reset a padding mismatched agfl before final free space check */
+	if (pag->pagf_agflreset)
+		xfs_agfl_reset(tp, agbp, pag);
+
 	/* If there isn't enough total space or single-extent, reject it. */
 	need = xfs_alloc_min_freelist(mp, pag);
 	if (!xfs_alloc_space_available(args, need, flags))
@@ -2266,10 +2366,11 @@ xfs_alloc_get_freelist(
 	bno = be32_to_cpu(agfl_bno[be32_to_cpu(agf->agf_flfirst)]);
 	be32_add_cpu(&agf->agf_flfirst, 1);
 	xfs_trans_brelse(tp, agflbp);
-	if (be32_to_cpu(agf->agf_flfirst) == XFS_AGFL_SIZE(mp))
+	if (be32_to_cpu(agf->agf_flfirst) == xfs_agfl_size(mp))
 		agf->agf_flfirst = 0;
 
 	pag = xfs_perag_get(mp, be32_to_cpu(agf->agf_seqno));
+	ASSERT(!pag->pagf_agflreset);
 	be32_add_cpu(&agf->agf_flcount, -1);
 	xfs_trans_agflist_delta(tp, -1);
 	pag->pagf_flcount--;
@@ -2377,10 +2478,11 @@ xfs_alloc_put_freelist(
 			be32_to_cpu(agf->agf_seqno), &agflbp)))
 		return error;
 	be32_add_cpu(&agf->agf_fllast, 1);
-	if (be32_to_cpu(agf->agf_fllast) == XFS_AGFL_SIZE(mp))
+	if (be32_to_cpu(agf->agf_fllast) == xfs_agfl_size(mp))
 		agf->agf_fllast = 0;
 
 	pag = xfs_perag_get(mp, be32_to_cpu(agf->agf_seqno));
+	ASSERT(!pag->pagf_agflreset);
 	be32_add_cpu(&agf->agf_flcount, 1);
 	xfs_trans_agflist_delta(tp, 1);
 	pag->pagf_flcount++;
@@ -2395,7 +2497,7 @@ xfs_alloc_put_freelist(
 
 	xfs_alloc_log_agf(tp, agbp, logflags);
 
-	ASSERT(be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp));
+	ASSERT(be32_to_cpu(agf->agf_flcount) <= xfs_agfl_size(mp));
 
 	agfl_bno = XFS_BUF_TO_AGFL_BNO(mp, agflbp);
 	blockp = &agfl_bno[be32_to_cpu(agf->agf_fllast)];
@@ -2428,9 +2530,9 @@ xfs_agf_verify(
 	if (!(agf->agf_magicnum == cpu_to_be32(XFS_AGF_MAGIC) &&
 	      XFS_AGF_GOOD_VERSION(be32_to_cpu(agf->agf_versionnum)) &&
 	      be32_to_cpu(agf->agf_freeblks) <= be32_to_cpu(agf->agf_length) &&
-	      be32_to_cpu(agf->agf_flfirst) < XFS_AGFL_SIZE(mp) &&
-	      be32_to_cpu(agf->agf_fllast) < XFS_AGFL_SIZE(mp) &&
-	      be32_to_cpu(agf->agf_flcount) <= XFS_AGFL_SIZE(mp)))
+	      be32_to_cpu(agf->agf_flfirst) < xfs_agfl_size(mp) &&
+	      be32_to_cpu(agf->agf_fllast) < xfs_agfl_size(mp) &&
+	      be32_to_cpu(agf->agf_flcount) <= xfs_agfl_size(mp)))
 		return __this_address;
 
 	if (be32_to_cpu(agf->agf_levels[XFS_BTNUM_BNO]) < 1 ||
@@ -2588,6 +2690,7 @@ xfs_alloc_read_agf(
 		pag->pagb_count = 0;
 		pag->pagb_tree = RB_ROOT;
 		pag->pagf_init = 1;
+		pag->pagf_agflreset = xfs_agfl_needs_reset(mp, agf);
 	}
 #ifdef DEBUG
 	else if (!XFS_FORCED_SHUTDOWN(mp)) {

fs/xfs/libxfs/xfs_alloc.h

@@ -26,6 +26,8 @@ struct xfs_trans;
 
 extern struct workqueue_struct *xfs_alloc_wq;
 
+unsigned int xfs_agfl_size(struct xfs_mount *mp);
+
 /*
  * Freespace allocation types.  Argument to xfs_alloc_[v]extent.
  */

fs/xfs/libxfs/xfs_alloc_btree.c

@@ -74,18 +74,13 @@ xfs_allocbt_alloc_block(
 	int			error;
 	xfs_agblock_t		bno;
 
-	XFS_BTREE_TRACE_CURSOR(cur, XBT_ENTRY);
-
 	/* Allocate the new block from the freelist. If we can't, give up.  */
 	error = xfs_alloc_get_freelist(cur->bc_tp, cur->bc_private.a.agbp,
 				       &bno, 1);
-	if (error) {
-		XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
+	if (error)
 		return error;
-	}
 
 	if (bno == NULLAGBLOCK) {
-		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
 		*stat = 0;
 		return 0;
 	}
@@ -95,7 +90,6 @@ xfs_allocbt_alloc_block(
 	xfs_trans_agbtree_delta(cur->bc_tp, 1);
 	new->s = cpu_to_be32(bno);
 
-	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
 	*stat = 1;
 	return 0;
 }

fs/xfs/libxfs/xfs_bmap.c

@@ -1244,8 +1244,9 @@ xfs_iread_extents(
 			xfs_warn(ip->i_mount,
 				"corrupt dinode %Lu, (btree extents).",
 				(unsigned long long) ip->i_ino);
-			XFS_CORRUPTION_ERROR(__func__,
-				XFS_ERRLEVEL_LOW, ip->i_mount, block);
+			xfs_inode_verifier_error(ip, -EFSCORRUPTED,
+					__func__, block, sizeof(*block),
+					__this_address);
 			error = -EFSCORRUPTED;
 			goto out_brelse;
 		}
@@ -1261,11 +1262,15 @@ xfs_iread_extents(
 		 */
 		frp = XFS_BMBT_REC_ADDR(mp, block, 1);
 		for (j = 0; j < num_recs; j++, frp++, i++) {
+			xfs_failaddr_t	fa;
+
 			xfs_bmbt_disk_get_all(frp, &new);
-			if (!xfs_bmbt_validate_extent(mp, whichfork, &new)) {
-				XFS_ERROR_REPORT("xfs_bmap_read_extents(2)",
-						 XFS_ERRLEVEL_LOW, mp);
+			fa = xfs_bmap_validate_extent(ip, whichfork, &new);
+			if (fa) {
 				error = -EFSCORRUPTED;
+				xfs_inode_verifier_error(ip, error,
+						"xfs_iread_extents(2)",
+						frp, sizeof(*frp), fa);
 				goto out_brelse;
 			}
 			xfs_iext_insert(ip, &icur, &new, state);
@@ -6154,3 +6159,39 @@ xfs_bmap_finish_one(
 
 	return error;
 }
+
+/* Check that an inode's extent does not have invalid flags or bad ranges. */
+xfs_failaddr_t
+xfs_bmap_validate_extent(
+	struct xfs_inode	*ip,
+	int			whichfork,
+	struct xfs_bmbt_irec	*irec)
+{
+	struct xfs_mount	*mp = ip->i_mount;
+	xfs_fsblock_t		endfsb;
+	bool			isrt;
+
+	isrt = XFS_IS_REALTIME_INODE(ip);
+	endfsb = irec->br_startblock + irec->br_blockcount - 1;
+	if (isrt) {
+		if (!xfs_verify_rtbno(mp, irec->br_startblock))
+			return __this_address;
+		if (!xfs_verify_rtbno(mp, endfsb))
+			return __this_address;
+	} else {
+		if (!xfs_verify_fsbno(mp, irec->br_startblock))
+			return __this_address;
+		if (!xfs_verify_fsbno(mp, endfsb))
+			return __this_address;
+		if (XFS_FSB_TO_AGNO(mp, irec->br_startblock) !=
+		    XFS_FSB_TO_AGNO(mp, endfsb))
+			return __this_address;
+	}
+	if (irec->br_state != XFS_EXT_NORM) {
+		if (whichfork != XFS_DATA_FORK)
+			return __this_address;
+		if (!xfs_sb_version_hasextflgbit(&mp->m_sb))
+			return __this_address;
+	}
+	return NULL;
+}

fs/xfs/libxfs/xfs_bmap.h

@@ -274,4 +274,7 @@ static inline int xfs_bmap_fork_to_state(int whichfork)
 	}
 }
 
+xfs_failaddr_t xfs_bmap_validate_extent(struct xfs_inode *ip, int whichfork,
+		struct xfs_bmbt_irec *irec);
+
 #endif	/* __XFS_BMAP_H__ */

fs/xfs/libxfs/xfs_bmap_btree.c

@@ -272,10 +272,10 @@ xfs_bmbt_alloc_block(
 		cur->bc_private.b.dfops->dop_low = true;
 	}
 	if (WARN_ON_ONCE(args.fsbno == NULLFSBLOCK)) {
-		XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
 		*stat = 0;
 		return 0;
 	}
+
 	ASSERT(args.len == 1);
 	cur->bc_private.b.firstblock = args.fsbno;
 	cur->bc_private.b.allocated++;
@@ -286,12 +286,10 @@ xfs_bmbt_alloc_block(
 
 	new->l = cpu_to_be64(args.fsbno);
 
-	XFS_BTREE_TRACE_CURSOR(cur, XBT_EXIT);
 	*stat = 1;
 	return 0;
 
  error0:
-	XFS_BTREE_TRACE_CURSOR(cur, XBT_ERROR);
 	return error;
 }
 

fs/xfs/libxfs/xfs_bmap_btree.h

@@ -118,18 +118,4 @@ extern int xfs_bmbt_change_owner(struct xfs_trans *tp, struct xfs_inode *ip,
 extern struct xfs_btree_cur *xfs_bmbt_init_cursor(struct xfs_mount *,
 		struct xfs_trans *, struct xfs_inode *, int);
 
-/*
- * Check that the extent does not contain an invalid unwritten extent flag.
- */
-static inline bool xfs_bmbt_validate_extent(struct xfs_mount *mp, int whichfork,
-		struct xfs_bmbt_irec *irec)
-{
-	if (irec->br_state == XFS_EXT_NORM)
-		return true;
-	if (whichfork == XFS_DATA_FORK &&
-	    xfs_sb_version_hasextflgbit(&mp->m_sb))
-		return true;
-	return false;
-}
-
 #endif	/* __XFS_BMAP_BTREE_H__ */

fs/xfs/libxfs/xfs_btree.h

@@ -473,25 +473,6 @@ static inline int xfs_btree_get_level(struct xfs_btree_block *block)
 #define	XFS_FILBLKS_MIN(a,b)	min_t(xfs_filblks_t, (a), (b))
 #define	XFS_FILBLKS_MAX(a,b)	max_t(xfs_filblks_t, (a), (b))
 
-/*
- * Trace hooks.  Currently not implemented as they need to be ported
- * over to the generic tracing functionality, which is some effort.
- *
- * i,j = integer (32 bit)
- * b = btree block buffer (xfs_buf_t)
- * p = btree ptr
- * r = btree record
- * k = btree key
- */
-#define	XFS_BTREE_TRACE_ARGBI(c, b, i)
-#define	XFS_BTREE_TRACE_ARGBII(c, b, i, j)
-#define	XFS_BTREE_TRACE_ARGI(c, i)
-#define	XFS_BTREE_TRACE_ARGIPK(c, i, p, s)
-#define	XFS_BTREE_TRACE_ARGIPR(c, i, p, r)
-#define	XFS_BTREE_TRACE_ARGIK(c, i, k)
-#define XFS_BTREE_TRACE_ARGR(c, r)
-#define	XFS_BTREE_TRACE_CURSOR(c, t)
-
 xfs_failaddr_t xfs_btree_sblock_v5hdr_verify(struct xfs_buf *bp);
 xfs_failaddr_t xfs_btree_sblock_verify(struct xfs_buf *bp,
 		unsigned int max_recs);

fs/xfs/libxfs/xfs_dir2.h

@@ -173,7 +173,7 @@ extern void xfs_dir2_data_log_unused(struct xfs_da_args *args,
 extern void xfs_dir2_data_make_free(struct xfs_da_args *args,
 		struct xfs_buf *bp, xfs_dir2_data_aoff_t offset,
 		xfs_dir2_data_aoff_t len, int *needlogp, int *needscanp);
-extern void xfs_dir2_data_use_free(struct xfs_da_args *args,
+extern int xfs_dir2_data_use_free(struct xfs_da_args *args,
 		struct xfs_buf *bp, struct xfs_dir2_data_unused *dup,
 		xfs_dir2_data_aoff_t offset, xfs_dir2_data_aoff_t len,
 		int *needlogp, int *needscanp);

fs/xfs/libxfs/xfs_dir2_block.c

@@ -451,15 +451,19 @@ xfs_dir2_block_addname(
 	 * No stale entries, will use enddup space to hold new leaf.
 	 */
 	if (!btp->stale) {
+		xfs_dir2_data_aoff_t	aoff;
+
 		/*
 		 * Mark the space needed for the new leaf entry, now in use.
 		 */
-		xfs_dir2_data_use_free(args, bp, enddup,
-			(xfs_dir2_data_aoff_t)
-			((char *)enddup - (char *)hdr + be16_to_cpu(enddup->length) -
-			 sizeof(*blp)),
-			(xfs_dir2_data_aoff_t)sizeof(*blp),
-			&needlog, &needscan);
+		aoff = (xfs_dir2_data_aoff_t)((char *)enddup - (char *)hdr +
+				be16_to_cpu(enddup->length) - sizeof(*blp));
+		error = xfs_dir2_data_use_free(args, bp, enddup, aoff,
+				(xfs_dir2_data_aoff_t)sizeof(*blp), &needlog,
+				&needscan);
+		if (error)
+			return error;
+
 		/*
 		 * Update the tail (entry count).
 		 */
@@ -541,9 +545,11 @@ xfs_dir2_block_addname(
 	/*
 	 * Mark space for the data entry used.
 	 */
-	xfs_dir2_data_use_free(args, bp, dup,
-		(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr),
-		(xfs_dir2_data_aoff_t)len, &needlog, &needscan);
+	error = xfs_dir2_data_use_free(args, bp, dup,
+			(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr),
+			(xfs_dir2_data_aoff_t)len, &needlog, &needscan);
+	if (error)
+		return error;
 	/*
 	 * Create the new data entry.
 	 */
@@ -997,8 +1003,10 @@ xfs_dir2_leaf_to_block(
 	/*
 	 * Use up the space at the end of the block (blp/btp).
 	 */
-	xfs_dir2_data_use_free(args, dbp, dup, args->geo->blksize - size, size,
-		&needlog, &needscan);
+	error = xfs_dir2_data_use_free(args, dbp, dup,
+			args->geo->blksize - size, size, &needlog, &needscan);
+	if (error)
+		return error;
 	/*
 	 * Initialize the block tail.
 	 */
@@ -1110,18 +1118,14 @@ xfs_dir2_sf_to_block(
 	 * Add block 0 to the inode.
 	 */
 	error = xfs_dir2_grow_inode(args, XFS_DIR2_DATA_SPACE, &blkno);
-	if (error) {
-		kmem_free(sfp);
-		return error;
-	}
+	if (error)
+		goto out_free;
 	/*
 	 * Initialize the data block, then convert it to block format.
 	 */
 	error = xfs_dir3_data_init(args, blkno, &bp);
-	if (error) {
-		kmem_free(sfp);
-		return error;
-	}
+	if (error)
+		goto out_free;
 	xfs_dir3_block_init(mp, tp, bp, dp);
 	hdr = bp->b_addr;
 
@@ -1136,8 +1140,10 @@ xfs_dir2_sf_to_block(
 	 */
 	dup = dp->d_ops->data_unused_p(hdr);
 	needlog = needscan = 0;
-	xfs_dir2_data_use_free(args, bp, dup, args->geo->blksize - i,
-			       i, &needlog, &needscan);
+	error = xfs_dir2_data_use_free(args, bp, dup, args->geo->blksize - i,
+			i, &needlog, &needscan);
+	if (error)
+		goto out_free;
 	ASSERT(needscan == 0);
 	/*
 	 * Fill in the tail.
@@ -1150,9 +1156,11 @@ xfs_dir2_sf_to_block(
 	/*
 	 * Remove the freespace, we'll manage it.
 	 */
-	xfs_dir2_data_use_free(args, bp, dup,
-		(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr),
-		be16_to_cpu(dup->length), &needlog, &needscan);
+	error = xfs_dir2_data_use_free(args, bp, dup,
+			(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr),
+			be16_to_cpu(dup->length), &needlog, &needscan);
+	if (error)
+		goto out_free;
 	/*
 	 * Create entry for .
 	 */
@@ -1256,4 +1264,7 @@ xfs_dir2_sf_to_block(
 	xfs_dir2_block_log_tail(tp, bp);
 	xfs_dir3_data_check(dp, bp);
 	return 0;
+out_free:
+	kmem_free(sfp);
+	return error;
 }

fs/xfs/libxfs/xfs_dir2_data.c

@@ -932,10 +932,51 @@ xfs_dir2_data_make_free(
 	*needscanp = needscan;
 }
 
+/* Check our free data for obvious signs of corruption. */
+static inline xfs_failaddr_t
+xfs_dir2_data_check_free(
+	struct xfs_dir2_data_hdr	*hdr,
+	struct xfs_dir2_data_unused	*dup,
+	xfs_dir2_data_aoff_t		offset,
+	xfs_dir2_data_aoff_t		len)
+{
+	if (hdr->magic != cpu_to_be32(XFS_DIR2_DATA_MAGIC) &&
+	    hdr->magic != cpu_to_be32(XFS_DIR3_DATA_MAGIC) &&
+	    hdr->magic != cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) &&
+	    hdr->magic != cpu_to_be32(XFS_DIR3_BLOCK_MAGIC))
+		return __this_address;
+	if (be16_to_cpu(dup->freetag) != XFS_DIR2_DATA_FREE_TAG)
+		return __this_address;
+	if (offset < (char *)dup - (char *)hdr)
+		return __this_address;
+	if (offset + len > (char *)dup + be16_to_cpu(dup->length) - (char *)hdr)
+		return __this_address;
+	if ((char *)dup - (char *)hdr !=
+			be16_to_cpu(*xfs_dir2_data_unused_tag_p(dup)))
+		return __this_address;
+	return NULL;
+}
+
+/* Sanity-check a new bestfree entry. */
+static inline xfs_failaddr_t
+xfs_dir2_data_check_new_free(
+	struct xfs_dir2_data_hdr	*hdr,
+	struct xfs_dir2_data_free	*dfp,
+	struct xfs_dir2_data_unused	*newdup)
+{
+	if (dfp == NULL)
+		return __this_address;
+	if (dfp->length != newdup->length)
+		return __this_address;
+	if (be16_to_cpu(dfp->offset) != (char *)newdup - (char *)hdr)
+		return __this_address;
+	return NULL;
+}
+
 /*
  * Take a byte range out of an existing unused space and make it un-free.
  */
-void
+int
 xfs_dir2_data_use_free(
 	struct xfs_da_args	*args,
 	struct xfs_buf		*bp,
@@ -947,23 +988,19 @@ xfs_dir2_data_use_free(
 {
 	xfs_dir2_data_hdr_t	*hdr;		/* data block header */
 	xfs_dir2_data_free_t	*dfp;		/* bestfree pointer */
+	xfs_dir2_data_unused_t	*newdup;	/* new unused entry */
+	xfs_dir2_data_unused_t	*newdup2;	/* another new unused entry */
+	struct xfs_dir2_data_free *bf;
+	xfs_failaddr_t		fa;
 	int			matchback;	/* matches end of freespace */
 	int			matchfront;	/* matches start of freespace */
 	int			needscan;	/* need to regen bestfree */
-	xfs_dir2_data_unused_t	*newdup;	/* new unused entry */
-	xfs_dir2_data_unused_t	*newdup2;	/* another new unused entry */
 	int			oldlen;		/* old unused entry's length */
-	struct xfs_dir2_data_free *bf;
 
 	hdr = bp->b_addr;
-	ASSERT(hdr->magic == cpu_to_be32(XFS_DIR2_DATA_MAGIC) ||
-	       hdr->magic == cpu_to_be32(XFS_DIR3_DATA_MAGIC) ||
-	       hdr->magic == cpu_to_be32(XFS_DIR2_BLOCK_MAGIC) ||
-	       hdr->magic == cpu_to_be32(XFS_DIR3_BLOCK_MAGIC));
-	ASSERT(be16_to_cpu(dup->freetag) == XFS_DIR2_DATA_FREE_TAG);
-	ASSERT(offset >= (char *)dup - (char *)hdr);
-	ASSERT(offset + len <= (char *)dup + be16_to_cpu(dup->length) - (char *)hdr);
-	ASSERT((char *)dup - (char *)hdr == be16_to_cpu(*xfs_dir2_data_unused_tag_p(dup)));
+	fa = xfs_dir2_data_check_free(hdr, dup, offset, len);
+	if (fa)
+		goto corrupt;
 	/*
 	 * Look up the entry in the bestfree table.
 	 */
@@ -1008,9 +1045,9 @@ xfs_dir2_data_use_free(
 			xfs_dir2_data_freeremove(hdr, bf, dfp, needlogp);
 			dfp = xfs_dir2_data_freeinsert(hdr, bf, newdup,
 						       needlogp);
-			ASSERT(dfp != NULL);
-			ASSERT(dfp->length == newdup->length);
-			ASSERT(be16_to_cpu(dfp->offset) == (char *)newdup - (char *)hdr);
+			fa = xfs_dir2_data_check_new_free(hdr, dfp, newdup);
+			if (fa)
+				goto corrupt;
 			/*
 			 * If we got inserted at the last slot,
 			 * that means we don't know if there was a better
@@ -1036,9 +1073,9 @@ xfs_dir2_data_use_free(
 			xfs_dir2_data_freeremove(hdr, bf, dfp, needlogp);
 			dfp = xfs_dir2_data_freeinsert(hdr, bf, newdup,
 						       needlogp);
-			ASSERT(dfp != NULL);
-			ASSERT(dfp->length == newdup->length);
-			ASSERT(be16_to_cpu(dfp->offset) == (char *)newdup - (char *)hdr);
+			fa = xfs_dir2_data_check_new_free(hdr, dfp, newdup);
+			if (fa)
+				goto corrupt;
 			/*
 			 * If we got inserted at the last slot,
 			 * that means we don't know if there was a better
@@ -1084,6 +1121,11 @@ xfs_dir2_data_use_free(
 		}
 	}
 	*needscanp = needscan;
+	return 0;
+corrupt:
+	xfs_corruption_error(__func__, XFS_ERRLEVEL_LOW, args->dp->i_mount,
+			hdr, __FILE__, __LINE__, fa);
+	return -EFSCORRUPTED;
 }
 
 /* Find the end of the entry data in a data/block format dir block. */

fs/xfs/libxfs/xfs_dir2_leaf.c

@@ -877,9 +877,13 @@ xfs_dir2_leaf_addname(
 	/*
 	 * Mark the initial part of our freespace in use for the new entry.
 	 */
-	xfs_dir2_data_use_free(args, dbp, dup,
-		(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr), length,
-		&needlog, &needscan);
+	error = xfs_dir2_data_use_free(args, dbp, dup,
+			(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr),
+			length, &needlog, &needscan);
+	if (error) {
+		xfs_trans_brelse(tp, lbp);
+		return error;
+	}
 	/*
 	 * Initialize our new entry (at last).
 	 */
@@ -1415,7 +1419,8 @@ xfs_dir2_leaf_removename(
 	oldbest = be16_to_cpu(bf[0].length);
 	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
 	bestsp = xfs_dir2_leaf_bests_p(ltp);
-	ASSERT(be16_to_cpu(bestsp[db]) == oldbest);
+	if (be16_to_cpu(bestsp[db]) != oldbest)
+		return -EFSCORRUPTED;
 	/*
 	 * Mark the former data entry unused.
 	 */

fs/xfs/libxfs/xfs_dir2_node.c

@@ -387,8 +387,9 @@ xfs_dir2_leaf_to_node(
 	dp->d_ops->free_hdr_from_disk(&freehdr, free);
 	leaf = lbp->b_addr;
 	ltp = xfs_dir2_leaf_tail_p(args->geo, leaf);
-	ASSERT(be32_to_cpu(ltp->bestcount) <=
-				(uint)dp->i_d.di_size / args->geo->blksize);
+	if (be32_to_cpu(ltp->bestcount) >
+				(uint)dp->i_d.di_size / args->geo->blksize)
+		return -EFSCORRUPTED;
 
 	/*
 	 * Copy freespace entries from the leaf block to the new block.
@@ -1728,6 +1729,7 @@ xfs_dir2_node_addname_int(
 	__be16			*bests;
 	struct xfs_dir3_icfree_hdr freehdr;
 	struct xfs_dir2_data_free *bf;
+	xfs_dir2_data_aoff_t	aoff;
 
 	dp = args->dp;
 	mp = dp->i_mount;
@@ -2022,9 +2024,13 @@ xfs_dir2_node_addname_int(
 	/*
 	 * Mark the first part of the unused space, inuse for us.
 	 */
-	xfs_dir2_data_use_free(args, dbp, dup,
-		(xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr), length,
-		&needlog, &needscan);
+	aoff = (xfs_dir2_data_aoff_t)((char *)dup - (char *)hdr);
+	error = xfs_dir2_data_use_free(args, dbp, dup, aoff, length,
+			&needlog, &needscan);
+	if (error) {
+		xfs_trans_brelse(tp, dbp);
+		return error;
+	}
 	/*
 	 * Fill in the new entry and log it.
 	 */