Merge tag 'xfs-4.15-fixes-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux

Pull xfs fixes from Darrick Wong:
 "Here are some XFS fixes for 4.15-rc5. Apologies for the unusually
  large number of patches this late, but I wanted to make sure the
  corruption fixes were really ready to go.

  Changes since last update:

   - Fix a locking problem during xattr block conversion that could lead
     to the log checkpointing thread to try to write an incomplete
     buffer to disk, which leads to a corruption shutdown

   - Fix a null pointer dereference when removing delayed allocation
     extents

   - Remove post-eof speculative allocations when reflinking a block
     past current inode size so that we don't just leave them there and
     assert on inode reclaim

   - Relax an assert which didn't accurately reflect the way locking
     works and would trigger under heavy io load

   - Avoid infinite loop when cancelling copy on write extents after a
     writeback failure

   - Try to avoid copy on write transaction reservation overflows when
     remapping after a successful write

   - Fix various problems with the copy-on-write reservation automatic
     garbage collection not being cleaned up properly during a ro
     remount

   - Fix problems with rmap log items being processed in the wrong
     order, leading to corruption shutdowns

   - Fix problems with EFI recovery wherein the "remove any rmapping if
     present" mechanism wasn't actually doing anything, which would lead
     to corruption problems later when the extent is reallocated,
     leading to multiple rmaps for the same extent"

* tag 'xfs-4.15-fixes-8' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  xfs: only skip rmap owner checks for unknown-owner rmap removal
  xfs: always honor OWN_UNKNOWN rmap removal requests
  xfs: queue deferred rmap ops for cow staging extent alloc/free in the right order
  xfs: set cowblocks tag for direct cow writes too
  xfs: remove leftover CoW reservations when remounting ro
  xfs: don't be so eager to clear the cowblocks tag on truncate
  xfs: track cowblocks separately in i_flags
  xfs: allow CoW remap transactions to use reserve blocks
  xfs: avoid infinite loop when cancelling CoW blocks after writeback failure
  xfs: relax is_reflink_inode assert in xfs_reflink_find_cow_mapping
  xfs: remove dest file's post-eof preallocations before reflinking
  xfs: move xfs_iext_insert tracepoint to report useful information
  xfs: account for null transactions in bunmapi
  xfs: hold xfs_buf locked between shortform->leaf conversion and the addition of an attribute
  xfs: add the ability to join a held buffer to a defer_ops

fs/xfs/libxfs/xfs_alloc.c

@@ -702,7 +702,7 @@ xfs_alloc_ag_vextent(
 	ASSERT(args->agbno % args->alignment == 0);
 
 	/* if not file data, insert new block into the reverse map btree */
-	if (args->oinfo.oi_owner != XFS_RMAP_OWN_UNKNOWN) {
+	if (!xfs_rmap_should_skip_owner_update(&args->oinfo)) {
 		error = xfs_rmap_alloc(args->tp, args->agbp, args->agno,
 				       args->agbno, args->len, &args->oinfo);
 		if (error)
@@ -1682,7 +1682,7 @@ xfs_free_ag_extent(
 	bno_cur = cnt_cur = NULL;
 	mp = tp->t_mountp;
 
-	if (oinfo->oi_owner != XFS_RMAP_OWN_UNKNOWN) {
+	if (!xfs_rmap_should_skip_owner_update(oinfo)) {
 		error = xfs_rmap_free(tp, agbp, agno, bno, len, oinfo);
 		if (error)
 			goto error0;

fs/xfs/libxfs/xfs_attr.c

@@ -212,6 +212,7 @@ xfs_attr_set(
 	int			flags)
 {
 	struct xfs_mount	*mp = dp->i_mount;
+	struct xfs_buf		*leaf_bp = NULL;
 	struct xfs_da_args	args;
 	struct xfs_defer_ops	dfops;
 	struct xfs_trans_res	tres;
@@ -327,9 +328,16 @@ xfs_attr_set(
 		 * GROT: another possible req'mt for a double-split btree op.
 		 */
 		xfs_defer_init(args.dfops, args.firstblock);
-		error = xfs_attr_shortform_to_leaf(&args);
+		error = xfs_attr_shortform_to_leaf(&args, &leaf_bp);
 		if (error)
 			goto out_defer_cancel;
+		/*
+		 * Prevent the leaf buffer from being unlocked so that a
+		 * concurrent AIL push cannot grab the half-baked leaf
+		 * buffer and run into problems with the write verifier.
+		 */
+		xfs_trans_bhold(args.trans, leaf_bp);
+		xfs_defer_bjoin(args.dfops, leaf_bp);
 		xfs_defer_ijoin(args.dfops, dp);
 		error = xfs_defer_finish(&args.trans, args.dfops);
 		if (error)
@@ -337,13 +345,14 @@ xfs_attr_set(
 
 		/*
 		 * Commit the leaf transformation.  We'll need another (linked)
-		 * transaction to add the new attribute to the leaf.
+		 * transaction to add the new attribute to the leaf, which
+		 * means that we have to hold & join the leaf buffer here too.
 		 */
-
 		error = xfs_trans_roll_inode(&args.trans, dp);
 		if (error)
 			goto out;
-
+		xfs_trans_bjoin(args.trans, leaf_bp);
+		leaf_bp = NULL;
 	}
 
 	if (xfs_bmap_one_block(dp, XFS_ATTR_FORK))
@@ -374,8 +383,9 @@ xfs_attr_set(
 
 out_defer_cancel:
 	xfs_defer_cancel(&dfops);
-	args.trans = NULL;
 out:
+	if (leaf_bp)
+		xfs_trans_brelse(args.trans, leaf_bp);
 	if (args.trans)
 		xfs_trans_cancel(args.trans);
 	xfs_iunlock(dp, XFS_ILOCK_EXCL);

fs/xfs/libxfs/xfs_attr_leaf.c

@@ -735,10 +735,13 @@ xfs_attr_shortform_getvalue(xfs_da_args_t *args)
 }
 
 /*
- * Convert from using the shortform to the leaf.
+ * Convert from using the shortform to the leaf.  On success, return the
+ * buffer so that we can keep it locked until we're totally done with it.
  */
 int
-xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
+xfs_attr_shortform_to_leaf(
+	struct xfs_da_args	*args,
+	struct xfs_buf		**leaf_bp)
 {
 	xfs_inode_t *dp;
 	xfs_attr_shortform_t *sf;
@@ -818,7 +821,7 @@ xfs_attr_shortform_to_leaf(xfs_da_args_t *args)
 		sfe = XFS_ATTR_SF_NEXTENTRY(sfe);
 	}
 	error = 0;
-
+	*leaf_bp = bp;
 out:
 	kmem_free(tmpbuffer);
 	return error;

fs/xfs/libxfs/xfs_attr_leaf.h

@@ -48,7 +48,8 @@ void	xfs_attr_shortform_create(struct xfs_da_args *args);
 void	xfs_attr_shortform_add(struct xfs_da_args *args, int forkoff);
 int	xfs_attr_shortform_lookup(struct xfs_da_args *args);
 int	xfs_attr_shortform_getvalue(struct xfs_da_args *args);
-int	xfs_attr_shortform_to_leaf(struct xfs_da_args *args);
+int	xfs_attr_shortform_to_leaf(struct xfs_da_args *args,
+			struct xfs_buf **leaf_bp);
 int	xfs_attr_shortform_remove(struct xfs_da_args *args);
 int	xfs_attr_shortform_allfit(struct xfs_buf *bp, struct xfs_inode *dp);
 int	xfs_attr_shortform_bytesfit(struct xfs_inode *dp, int bytes);

fs/xfs/libxfs/xfs_bmap.c

@@ -5136,7 +5136,7 @@ __xfs_bunmapi(
 	 * blowing out the transaction with a mix of EFIs and reflink
 	 * adjustments.
 	 */
-	if (xfs_is_reflink_inode(ip) && whichfork == XFS_DATA_FORK)
+	if (tp && xfs_is_reflink_inode(ip) && whichfork == XFS_DATA_FORK)
 		max_len = min(len, xfs_refcount_max_unmap(tp->t_log_res));
 	else
 		max_len = len;

fs/xfs/libxfs/xfs_defer.c

@@ -249,6 +249,10 @@ xfs_defer_trans_roll(
 	for (i = 0; i < XFS_DEFER_OPS_NR_INODES && dop->dop_inodes[i]; i++)
 		xfs_trans_log_inode(*tp, dop->dop_inodes[i], XFS_ILOG_CORE);
 
+	/* Hold the (previously bjoin'd) buffer locked across the roll. */
+	for (i = 0; i < XFS_DEFER_OPS_NR_BUFS && dop->dop_bufs[i]; i++)
+		xfs_trans_dirty_buf(*tp, dop->dop_bufs[i]);
+
 	trace_xfs_defer_trans_roll((*tp)->t_mountp, dop);
 
 	/* Roll the transaction. */
@@ -264,6 +268,12 @@ xfs_defer_trans_roll(
 	for (i = 0; i < XFS_DEFER_OPS_NR_INODES && dop->dop_inodes[i]; i++)
 		xfs_trans_ijoin(*tp, dop->dop_inodes[i], 0);
 
+	/* Rejoin the buffers and dirty them so the log moves forward. */
+	for (i = 0; i < XFS_DEFER_OPS_NR_BUFS && dop->dop_bufs[i]; i++) {
+		xfs_trans_bjoin(*tp, dop->dop_bufs[i]);
+		xfs_trans_bhold(*tp, dop->dop_bufs[i]);
+	}
+
 	return error;
 }
 
@@ -295,6 +305,31 @@ xfs_defer_ijoin(
 		}
 	}
 
+	ASSERT(0);
+	return -EFSCORRUPTED;
+}
+
+/*
+ * Add this buffer to the deferred op.  Each joined buffer is relogged
+ * each time we roll the transaction.
+ */
+int
+xfs_defer_bjoin(
+	struct xfs_defer_ops		*dop,
+	struct xfs_buf			*bp)
+{
+	int				i;
+
+	for (i = 0; i < XFS_DEFER_OPS_NR_BUFS; i++) {
+		if (dop->dop_bufs[i] == bp)
+			return 0;
+		else if (dop->dop_bufs[i] == NULL) {
+			dop->dop_bufs[i] = bp;
+			return 0;
+		}
+	}
+
+	ASSERT(0);
 	return -EFSCORRUPTED;
 }
 
@@ -493,9 +528,7 @@ xfs_defer_init(
 	struct xfs_defer_ops		*dop,
 	xfs_fsblock_t			*fbp)
 {
-	dop->dop_committed = false;
-	dop->dop_low = false;
-	memset(&dop->dop_inodes, 0, sizeof(dop->dop_inodes));
+	memset(dop, 0, sizeof(struct xfs_defer_ops));
 	*fbp = NULLFSBLOCK;
 	INIT_LIST_HEAD(&dop->dop_intake);
 	INIT_LIST_HEAD(&dop->dop_pending);

fs/xfs/libxfs/xfs_defer.h

@@ -59,6 +59,7 @@ enum xfs_defer_ops_type {
 };
 
 #define XFS_DEFER_OPS_NR_INODES	2	/* join up to two inodes */
+#define XFS_DEFER_OPS_NR_BUFS	2	/* join up to two buffers */
 
 struct xfs_defer_ops {
 	bool			dop_committed;	/* did any trans commit? */
@@ -66,8 +67,9 @@ struct xfs_defer_ops {
 	struct list_head	dop_intake;	/* unlogged pending work */
 	struct list_head	dop_pending;	/* logged pending work */
 
-	/* relog these inodes with each roll */
+	/* relog these with each roll */
 	struct xfs_inode	*dop_inodes[XFS_DEFER_OPS_NR_INODES];
+	struct xfs_buf		*dop_bufs[XFS_DEFER_OPS_NR_BUFS];
 };
 
 void xfs_defer_add(struct xfs_defer_ops *dop, enum xfs_defer_ops_type type,
@@ -77,6 +79,7 @@ void xfs_defer_cancel(struct xfs_defer_ops *dop);
 void xfs_defer_init(struct xfs_defer_ops *dop, xfs_fsblock_t *fbp);
 bool xfs_defer_has_unfinished_work(struct xfs_defer_ops *dop);
 int xfs_defer_ijoin(struct xfs_defer_ops *dop, struct xfs_inode *ip);
+int xfs_defer_bjoin(struct xfs_defer_ops *dop, struct xfs_buf *bp);
 
 /* Description of a deferred type. */
 struct xfs_defer_op_type {

fs/xfs/libxfs/xfs_iext_tree.c

@@ -632,8 +632,6 @@ xfs_iext_insert(
 	struct xfs_iext_leaf	*new = NULL;
 	int			nr_entries, i;
 
-	trace_xfs_iext_insert(ip, cur, state, _RET_IP_);
-
 	if (ifp->if_height == 0)
 		xfs_iext_alloc_root(ifp, cur);
 	else if (ifp->if_height == 1)
@@ -661,6 +659,8 @@ xfs_iext_insert(
 	xfs_iext_set(cur_rec(cur), irec);
 	ifp->if_bytes += sizeof(struct xfs_iext_rec);
 
+	trace_xfs_iext_insert(ip, cur, state, _RET_IP_);
+
 	if (new)
 		xfs_iext_insert_node(ifp, xfs_iext_leaf_key(new, 0), new, 2);
 }

fs/xfs/libxfs/xfs_refcount.c

@@ -1488,27 +1488,12 @@ __xfs_refcount_cow_alloc(
 	xfs_extlen_t		aglen,
 	struct xfs_defer_ops	*dfops)
 {
-	int			error;
-
 	trace_xfs_refcount_cow_increase(rcur->bc_mp, rcur->bc_private.a.agno,
 			agbno, aglen);
 
 	/* Add refcount btree reservation */
-	error = xfs_refcount_adjust_cow(rcur, agbno, aglen,
+	return xfs_refcount_adjust_cow(rcur, agbno, aglen,
 			XFS_REFCOUNT_ADJUST_COW_ALLOC, dfops);
-	if (error)
-		return error;
-
-	/* Add rmap entry */
-	if (xfs_sb_version_hasrmapbt(&rcur->bc_mp->m_sb)) {
-		error = xfs_rmap_alloc_extent(rcur->bc_mp, dfops,
-				rcur->bc_private.a.agno,
-				agbno, aglen, XFS_RMAP_OWN_COW);
-		if (error)
-			return error;
-	}
-
-	return error;
 }
 
 /*
@@ -1521,27 +1506,12 @@ __xfs_refcount_cow_free(
 	xfs_extlen_t		aglen,
 	struct xfs_defer_ops	*dfops)
 {
-	int			error;
-
 	trace_xfs_refcount_cow_decrease(rcur->bc_mp, rcur->bc_private.a.agno,
 			agbno, aglen);
 
 	/* Remove refcount btree reservation */
-	error = xfs_refcount_adjust_cow(rcur, agbno, aglen,
+	return xfs_refcount_adjust_cow(rcur, agbno, aglen,
 			XFS_REFCOUNT_ADJUST_COW_FREE, dfops);
-	if (error)
-		return error;
-
-	/* Remove rmap entry */
-	if (xfs_sb_version_hasrmapbt(&rcur->bc_mp->m_sb)) {
-		error = xfs_rmap_free_extent(rcur->bc_mp, dfops,
-				rcur->bc_private.a.agno,
-				agbno, aglen, XFS_RMAP_OWN_COW);
-		if (error)
-			return error;
-	}
-
-	return error;
 }
 
 /* Record a CoW staging extent in the refcount btree. */
@@ -1552,11 +1522,19 @@ xfs_refcount_alloc_cow_extent(
 	xfs_fsblock_t			fsb,
 	xfs_extlen_t			len)
 {
+	int				error;
+
 	if (!xfs_sb_version_hasreflink(&mp->m_sb))
 		return 0;
 
-	return __xfs_refcount_add(mp, dfops, XFS_REFCOUNT_ALLOC_COW,
+	error = __xfs_refcount_add(mp, dfops, XFS_REFCOUNT_ALLOC_COW,
 			fsb, len);
+	if (error)
+		return error;
+
+	/* Add rmap entry */
+	return xfs_rmap_alloc_extent(mp, dfops, XFS_FSB_TO_AGNO(mp, fsb),
+			XFS_FSB_TO_AGBNO(mp, fsb), len, XFS_RMAP_OWN_COW);
 }
 
 /* Forget a CoW staging event in the refcount btree. */
@@ -1567,9 +1545,17 @@ xfs_refcount_free_cow_extent(
 	xfs_fsblock_t			fsb,
 	xfs_extlen_t			len)
 {
+	int				error;
+
 	if (!xfs_sb_version_hasreflink(&mp->m_sb))
 		return 0;
 
+	/* Remove rmap entry */
+	error = xfs_rmap_free_extent(mp, dfops, XFS_FSB_TO_AGNO(mp, fsb),
+			XFS_FSB_TO_AGBNO(mp, fsb), len, XFS_RMAP_OWN_COW);
+	if (error)
+		return error;
+
 	return __xfs_refcount_add(mp, dfops, XFS_REFCOUNT_FREE_COW,
 			fsb, len);
 }

fs/xfs/libxfs/xfs_rmap.c

@@ -367,6 +367,51 @@ xfs_rmap_lookup_le_range(
 	return error;
 }
 
+/*
+ * Perform all the relevant owner checks for a removal op.  If we're doing an
+ * unknown-owner removal then we have no owner information to check.
+ */
+static int
+xfs_rmap_free_check_owner(
+	struct xfs_mount	*mp,
+	uint64_t		ltoff,
+	struct xfs_rmap_irec	*rec,
+	xfs_fsblock_t		bno,
+	xfs_filblks_t		len,
+	uint64_t		owner,
+	uint64_t		offset,
+	unsigned int		flags)
+{
+	int			error = 0;
+
+	if (owner == XFS_RMAP_OWN_UNKNOWN)
+		return 0;
+
+	/* Make sure the unwritten flag matches. */
+	XFS_WANT_CORRUPTED_GOTO(mp, (flags & XFS_RMAP_UNWRITTEN) ==
+			(rec->rm_flags & XFS_RMAP_UNWRITTEN), out);
+
+	/* Make sure the owner matches what we expect to find in the tree. */
+	XFS_WANT_CORRUPTED_GOTO(mp, owner == rec->rm_owner, out);
+
+	/* Check the offset, if necessary. */
+	if (XFS_RMAP_NON_INODE_OWNER(owner))
+		goto out;
+
+	if (flags & XFS_RMAP_BMBT_BLOCK) {
+		XFS_WANT_CORRUPTED_GOTO(mp, rec->rm_flags & XFS_RMAP_BMBT_BLOCK,
+				out);
+	} else {
+		XFS_WANT_CORRUPTED_GOTO(mp, rec->rm_offset <= offset, out);
+		XFS_WANT_CORRUPTED_GOTO(mp,
+				ltoff + rec->rm_blockcount >= offset + len,
+				out);
+	}
+
+out:
+	return error;
+}
+
 /*
  * Find the extent in the rmap btree and remove it.
  *
@@ -444,33 +489,40 @@ xfs_rmap_unmap(
 		goto out_done;
 	}
 
-	/* Make sure the unwritten flag matches. */
-	XFS_WANT_CORRUPTED_GOTO(mp, (flags & XFS_RMAP_UNWRITTEN) ==
-			(ltrec.rm_flags & XFS_RMAP_UNWRITTEN), out_error);
+	/*
+	 * If we're doing an unknown-owner removal for EFI recovery, we expect
+	 * to find the full range in the rmapbt or nothing at all.  If we
+	 * don't find any rmaps overlapping either end of the range, we're
+	 * done.  Hopefully this means that the EFI creator already queued
+	 * (and finished) a RUI to remove the rmap.
+	 */
+	if (owner == XFS_RMAP_OWN_UNKNOWN &&
+	    ltrec.rm_startblock + ltrec.rm_blockcount <= bno) {
+		struct xfs_rmap_irec    rtrec;
+
+		error = xfs_btree_increment(cur, 0, &i);
+		if (error)
+			goto out_error;
+		if (i == 0)
+			goto out_done;
+		error = xfs_rmap_get_rec(cur, &rtrec, &i);
+		if (error)
+			goto out_error;
+		XFS_WANT_CORRUPTED_GOTO(mp, i == 1, out_error);
+		if (rtrec.rm_startblock >= bno + len)
+			goto out_done;
+	}
 
 	/* Make sure the extent we found covers the entire freeing range. */
 	XFS_WANT_CORRUPTED_GOTO(mp, ltrec.rm_startblock <= bno &&
-		ltrec.rm_startblock + ltrec.rm_blockcount >=
-		bno + len, out_error);
+			ltrec.rm_startblock + ltrec.rm_blockcount >=
+			bno + len, out_error);
 
-	/* Make sure the owner matches what we expect to find in the tree. */
-	XFS_WANT_CORRUPTED_GOTO(mp, owner == ltrec.rm_owner ||
-				    XFS_RMAP_NON_INODE_OWNER(owner), out_error);
-
-	/* Check the offset, if necessary. */
-	if (!XFS_RMAP_NON_INODE_OWNER(owner)) {
-		if (flags & XFS_RMAP_BMBT_BLOCK) {
-			XFS_WANT_CORRUPTED_GOTO(mp,
-					ltrec.rm_flags & XFS_RMAP_BMBT_BLOCK,
-					out_error);
-		} else {
-			XFS_WANT_CORRUPTED_GOTO(mp,
-					ltrec.rm_offset <= offset, out_error);
-			XFS_WANT_CORRUPTED_GOTO(mp,
-					ltoff + ltrec.rm_blockcount >= offset + len,
-					out_error);
-		}
-	}
+	/* Check owner information. */
+	error = xfs_rmap_free_check_owner(mp, ltoff, &ltrec, bno, len, owner,
+			offset, flags);
+	if (error)
+		goto out_error;
 
 	if (ltrec.rm_startblock == bno && ltrec.rm_blockcount == len) {
 		/* exact match, simply remove the record from rmap tree */
@@ -664,6 +716,7 @@ xfs_rmap_map(
 		flags |= XFS_RMAP_UNWRITTEN;
 	trace_xfs_rmap_map(mp, cur->bc_private.a.agno, bno, len,
 			unwritten, oinfo);
+	ASSERT(!xfs_rmap_should_skip_owner_update(oinfo));
 
 	/*
 	 * For the initial lookup, look for an exact match or the left-adjacent

fs/xfs/libxfs/xfs_rmap.h

@@ -61,7 +61,21 @@ static inline void
 xfs_rmap_skip_owner_update(
 	struct xfs_owner_info	*oi)
 {
-	oi->oi_owner = XFS_RMAP_OWN_UNKNOWN;
+	xfs_rmap_ag_owner(oi, XFS_RMAP_OWN_NULL);
+}
+
+static inline bool
+xfs_rmap_should_skip_owner_update(
+	struct xfs_owner_info	*oi)
+{
+	return oi->oi_owner == XFS_RMAP_OWN_NULL;
+}
+
+static inline void
+xfs_rmap_any_owner_update(
+	struct xfs_owner_info	*oi)
+{
+	xfs_rmap_ag_owner(oi, XFS_RMAP_OWN_UNKNOWN);
 }
 
 /* Reverse mapping functions. */

fs/xfs/xfs_extfree_item.c

@@ -538,7 +538,7 @@ xfs_efi_recover(
 		return error;
 	efdp = xfs_trans_get_efd(tp, efip, efip->efi_format.efi_nextents);
 
-	xfs_rmap_skip_owner_update(&oinfo);
+	xfs_rmap_any_owner_update(&oinfo);
 	for (i = 0; i < efip->efi_format.efi_nextents; i++) {
 		extp = &efip->efi_format.efi_extents[i];
 		error = xfs_trans_free_extent(tp, efdp, extp->ext_start,

fs/xfs/xfs_fsops.c

@@ -571,6 +571,11 @@ xfs_growfs_data_private(
 		 * this doesn't actually exist in the rmap btree.
 		 */
 		xfs_rmap_ag_owner(&oinfo, XFS_RMAP_OWN_NULL);
+		error = xfs_rmap_free(tp, bp, agno,
+				be32_to_cpu(agf->agf_length) - new,
+				new, &oinfo);
+		if (error)
+			goto error0;
 		error = xfs_free_extent(tp,
 				XFS_AGB_TO_FSB(mp, agno,
 					be32_to_cpu(agf->agf_length) - new),

fs/xfs/xfs_icache.c

@@ -870,7 +870,7 @@ xfs_eofblocks_worker(
  * based on the 'speculative_cow_prealloc_lifetime' tunable (5m by default).
  * (We'll just piggyback on the post-EOF prealloc space workqueue.)
  */
-STATIC void
+void
 xfs_queue_cowblocks(
 	struct xfs_mount *mp)
 {
@@ -1536,8 +1536,23 @@ xfs_inode_free_quota_eofblocks(
 	return __xfs_inode_free_quota_eofblocks(ip, xfs_icache_free_eofblocks);
 }
 
+static inline unsigned long
+xfs_iflag_for_tag(
+	int		tag)
+{
+	switch (tag) {
+	case XFS_ICI_EOFBLOCKS_TAG:
+		return XFS_IEOFBLOCKS;
+	case XFS_ICI_COWBLOCKS_TAG:
+		return XFS_ICOWBLOCKS;
+	default:
+		ASSERT(0);
+		return 0;
+	}
+}
+
 static void
-__xfs_inode_set_eofblocks_tag(
+__xfs_inode_set_blocks_tag(
 	xfs_inode_t	*ip,
 	void		(*execute)(struct xfs_mount *mp),
 	void		(*set_tp)(struct xfs_mount *mp, xfs_agnumber_t agno,
@@ -1552,10 +1567,10 @@ __xfs_inode_set_eofblocks_tag(
 	 * Don't bother locking the AG and looking up in the radix trees
 	 * if we already know that we have the tag set.
 	 */
-	if (ip->i_flags & XFS_IEOFBLOCKS)
+	if (ip->i_flags & xfs_iflag_for_tag(tag))
 		return;
 	spin_lock(&ip->i_flags_lock);
-	ip->i_flags |= XFS_IEOFBLOCKS;
+	ip->i_flags |= xfs_iflag_for_tag(tag);
 	spin_unlock(&ip->i_flags_lock);
 
 	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
@@ -1587,13 +1602,13 @@ xfs_inode_set_eofblocks_tag(
 	xfs_inode_t	*ip)
 {
 	trace_xfs_inode_set_eofblocks_tag(ip);
-	return __xfs_inode_set_eofblocks_tag(ip, xfs_queue_eofblocks,
+	return __xfs_inode_set_blocks_tag(ip, xfs_queue_eofblocks,
 			trace_xfs_perag_set_eofblocks,
 			XFS_ICI_EOFBLOCKS_TAG);
 }
 
 static void
-__xfs_inode_clear_eofblocks_tag(
+__xfs_inode_clear_blocks_tag(
 	xfs_inode_t	*ip,
 	void		(*clear_tp)(struct xfs_mount *mp, xfs_agnumber_t agno,
 				    int error, unsigned long caller_ip),
@@ -1603,7 +1618,7 @@ __xfs_inode_clear_eofblocks_tag(
 	struct xfs_perag *pag;
 
 	spin_lock(&ip->i_flags_lock);
-	ip->i_flags &= ~XFS_IEOFBLOCKS;
+	ip->i_flags &= ~xfs_iflag_for_tag(tag);
 	spin_unlock(&ip->i_flags_lock);
 
 	pag = xfs_perag_get(mp, XFS_INO_TO_AGNO(mp, ip->i_ino));
@@ -1630,7 +1645,7 @@ xfs_inode_clear_eofblocks_tag(
 	xfs_inode_t	*ip)
 {
 	trace_xfs_inode_clear_eofblocks_tag(ip);
-	return __xfs_inode_clear_eofblocks_tag(ip,
+	return __xfs_inode_clear_blocks_tag(ip,
 			trace_xfs_perag_clear_eofblocks, XFS_ICI_EOFBLOCKS_TAG);
 }
 
@@ -1724,7 +1739,7 @@ xfs_inode_set_cowblocks_tag(
 	xfs_inode_t	*ip)
 {
 	trace_xfs_inode_set_cowblocks_tag(ip);
-	return __xfs_inode_set_eofblocks_tag(ip, xfs_queue_cowblocks,
+	return __xfs_inode_set_blocks_tag(ip, xfs_queue_cowblocks,
 			trace_xfs_perag_set_cowblocks,
 			XFS_ICI_COWBLOCKS_TAG);
 }
@@ -1734,6 +1749,6 @@ xfs_inode_clear_cowblocks_tag(
 	xfs_inode_t	*ip)
 {
 	trace_xfs_inode_clear_cowblocks_tag(ip);
-	return __xfs_inode_clear_eofblocks_tag(ip,
+	return __xfs_inode_clear_blocks_tag(ip,
 			trace_xfs_perag_clear_cowblocks, XFS_ICI_COWBLOCKS_TAG);
 }

fs/xfs/xfs_icache.h

@@ -81,6 +81,7 @@ void xfs_inode_clear_cowblocks_tag(struct xfs_inode *ip);
 int xfs_icache_free_cowblocks(struct xfs_mount *, struct xfs_eofblocks *);
 int xfs_inode_free_quota_cowblocks(struct xfs_inode *ip);
 void xfs_cowblocks_worker(struct work_struct *);
+void xfs_queue_cowblocks(struct xfs_mount *);
 
 int xfs_inode_ag_iterator(struct xfs_mount *mp,
 	int (*execute)(struct xfs_inode *ip, int flags, void *args),

fs/xfs/xfs_inode.c

@@ -1487,6 +1487,24 @@ xfs_link(
 	return error;
 }
 
+/* Clear the reflink flag and the cowblocks tag if possible. */
+static void
+xfs_itruncate_clear_reflink_flags(
+	struct xfs_inode	*ip)
+{
+	struct xfs_ifork	*dfork;
+	struct xfs_ifork	*cfork;
+
+	if (!xfs_is_reflink_inode(ip))
+		return;
+	dfork = XFS_IFORK_PTR(ip, XFS_DATA_FORK);
+	cfork = XFS_IFORK_PTR(ip, XFS_COW_FORK);
+	if (dfork->if_bytes == 0 && cfork->if_bytes == 0)
+		ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK;
+	if (cfork->if_bytes == 0)
+		xfs_inode_clear_cowblocks_tag(ip);
+}
+
 /*
  * Free up the underlying blocks past new_size.  The new size must be smaller
  * than the current size.  This routine can be used both for the attribute and
@@ -1583,15 +1601,7 @@ xfs_itruncate_extents(
 	if (error)
 		goto out;
 
-	/*
-	 * Clear the reflink flag if there are no data fork blocks and
-	 * there are no extents staged in the cow fork.
-	 */
-	if (xfs_is_reflink_inode(ip) && ip->i_cnextents == 0) {
-		if (ip->i_d.di_nblocks == 0)
-			ip->i_d.di_flags2 &= ~XFS_DIFLAG2_REFLINK;
-		xfs_inode_clear_cowblocks_tag(ip);
-	}
+	xfs_itruncate_clear_reflink_flags(ip);
 
 	/*
 	 * Always re-log the inode so that our permanent transaction can keep

fs/xfs/xfs_inode.h

@@ -232,6 +232,7 @@ static inline bool xfs_is_reflink_inode(struct xfs_inode *ip)
  * log recovery to replay a bmap operation on the inode.
  */
 #define XFS_IRECOVERY		(1 << 11)
+#define XFS_ICOWBLOCKS		(1 << 12)/* has the cowblocks tag set */
 
 /*
  * Per-lifetime flags need to be reset when re-using a reclaimable inode during

fs/xfs/xfs_reflink.c

@@ -454,6 +454,8 @@ xfs_reflink_allocate_cow(
 	if (error)
 		goto out_bmap_cancel;
 
+	xfs_inode_set_cowblocks_tag(ip);
+
 	/* Finish up. */
 	error = xfs_defer_finish(&tp, &dfops);
 	if (error)
@@ -490,8 +492,9 @@ xfs_reflink_find_cow_mapping(
 	struct xfs_iext_cursor		icur;
 
 	ASSERT(xfs_isilocked(ip, XFS_ILOCK_EXCL | XFS_ILOCK_SHARED));
-	ASSERT(xfs_is_reflink_inode(ip));
 
+	if (!xfs_is_reflink_inode(ip))
+		return false;
 	offset_fsb = XFS_B_TO_FSBT(ip->i_mount, offset);
 	if (!xfs_iext_lookup_extent(ip, ifp, offset_fsb, &icur, &got))
 		return false;
@@ -610,6 +613,9 @@ xfs_reflink_cancel_cow_blocks(
 
 			/* Remove the mapping from the CoW fork. */
 			xfs_bmap_del_extent_cow(ip, &icur, &got, &del);
+		} else {
+			/* Didn't do anything, push cursor back. */
+			xfs_iext_prev(ifp, &icur);
 		}
 next_extent:
 		if (!xfs_iext_get_extent(ifp, &icur, &got))
@@ -725,7 +731,7 @@ xfs_reflink_end_cow(
 			(unsigned int)(end_fsb - offset_fsb),
 			XFS_DATA_FORK);
 	error = xfs_trans_alloc(ip->i_mount, &M_RES(ip->i_mount)->tr_write,
-			resblks, 0, 0, &tp);
+			resblks, 0, XFS_TRANS_RESERVE, &tp);
 	if (error)
 		goto out;
 
@@ -1291,6 +1297,17 @@ xfs_reflink_remap_range(
 
 	trace_xfs_reflink_remap_range(src, pos_in, len, dest, pos_out);
 
+	/*
+	 * Clear out post-eof preallocations because we don't have page cache
+	 * backing the delayed allocations and they'll never get freed on
+	 * their own.
+	 */
+	if (xfs_can_free_eofblocks(dest, true)) {
+		ret = xfs_free_eofblocks(dest);
+		if (ret)
+			goto out_unlock;
+	}
+
 	/* Set flags and remap blocks. */
 	ret = xfs_reflink_set_inode_flag(src, dest);
 	if (ret)

fs/xfs/xfs_super.c

@@ -1360,6 +1360,7 @@ xfs_fs_remount(
 			xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
 			return error;
 		}
+		xfs_queue_cowblocks(mp);
 
 		/* Create the per-AG metadata reservation pool .*/
 		error = xfs_fs_reserve_ag_blocks(mp);
@@ -1369,6 +1370,14 @@ xfs_fs_remount(
 
 	/* rw -> ro */
 	if (!(mp->m_flags & XFS_MOUNT_RDONLY) && (*flags & SB_RDONLY)) {
+		/* Get rid of any leftover CoW reservations... */
+		cancel_delayed_work_sync(&mp->m_cowblocks_work);
+		error = xfs_icache_free_cowblocks(mp, NULL);
+		if (error) {
+			xfs_force_shutdown(mp, SHUTDOWN_CORRUPT_INCORE);
+			return error;
+		}
+
 		/* Free the per-AG metadata reservation pool. */
 		error = xfs_fs_unreserve_ag_blocks(mp);
 		if (error) {