XFS Bug fixes for 6.12-rc4

* Fix integer overflow in xrep_bmap
 * Fix stale dealloc punching for COW IO
 
 Signed-off-by: Carlos Maiolino <cem@kernel.org>
 -----BEGIN PGP SIGNATURE-----
 
 iJUEABMJAB0WIQQMHYkcUKcy4GgPe2RGdaER5QtfpgUCZw5LIwAKCRBGdaER5Qtf
 puRlAYDezbvs1dDSkKIGOt3inGdLptNAu4qniXBUkbYI9BzmtIVDueWP4Wo0dV3d
 gu3xrWQBfjFXdmEuBlwLuAFrp07AN18BVMj+DWCiEShsPHSoSPcF/IrDiz4BHvGv
 MKYq9CywFw==
 =Gj9b
 -----END PGP SIGNATURE-----

Merge tag 'xfs-6.12-fixes-4' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux

Pull xfs fixes from Carlos Maiolino:

 - Fix integer overflow in xrep_bmap

 - Fix stale dealloc punching for COW IO

* tag 'xfs-6.12-fixes-4' of git://git.kernel.org/pub/scm/fs/xfs/xfs-linux:
  xfs: punch delalloc extents from the COW fork for COW writes
  xfs: set IOMAP_F_SHARED for all COW fork allocations
  xfs: share more code in xfs_buffered_write_iomap_begin
  xfs: support the COW fork in xfs_bmap_punch_delalloc_range
  xfs: IOMAP_ZERO and IOMAP_UNSHARE already hold invalidate_lock
  xfs: take XFS_MMAPLOCK_EXCL xfs_file_write_zero_eof
  xfs: factor out a xfs_file_write_zero_eof helper
  iomap: move locking out of iomap_write_delalloc_release
  iomap: remove iomap_file_buffered_write_punch_delalloc
  iomap: factor out a iomap_last_written_block helper
  xfs: fix integer overflow in xrep_bmap

Documentation/filesystems/iomap/operations.rst

@@ -208,7 +208,7 @@ The filesystem must arrange to `cancel
 such `reservations
 <https://lore.kernel.org/linux-xfs/20220817093627.GZ3600936@dread.disaster.area/>`_
 because writeback will not consume the reservation.
-The ``iomap_file_buffered_write_punch_delalloc`` can be called from a
+The ``iomap_write_delalloc_release`` can be called from a
 ``->iomap_end`` function to find all the clean areas of the folios
 caching a fresh (``IOMAP_F_NEW``) delalloc mapping.
 It takes the ``invalidate_lock``.

fs/iomap/buffered-io.c

@@ -1145,10 +1145,36 @@ static void iomap_write_delalloc_scan(struct inode *inode,
 }
 
 /*
+ * When a short write occurs, the filesystem might need to use ->iomap_end
+ * to remove space reservations created in ->iomap_begin.
+ *
+ * For filesystems that use delayed allocation, there can be dirty pages over
+ * the delalloc extent outside the range of a short write but still within the
+ * delalloc extent allocated for this iomap if the write raced with page
+ * faults.
+ *
  * Punch out all the delalloc blocks in the range given except for those that
  * have dirty data still pending in the page cache - those are going to be
  * written and so must still retain the delalloc backing for writeback.
  *
+ * The punch() callback *must* only punch delalloc extents in the range passed
+ * to it. It must skip over all other types of extents in the range and leave
+ * them completely unchanged. It must do this punch atomically with respect to
+ * other extent modifications.
+ *
+ * The punch() callback may be called with a folio locked to prevent writeback
+ * extent allocation racing at the edge of the range we are currently punching.
+ * The locked folio may or may not cover the range being punched, so it is not
+ * safe for the punch() callback to lock folios itself.
+ *
+ * Lock order is:
+ *
+ * inode->i_rwsem (shared or exclusive)
+ *   inode->i_mapping->invalidate_lock (exclusive)
+ *     folio_lock()
+ *       ->punch
+ *         internal filesystem allocation lock
+ *
  * As we are scanning the page cache for data, we don't need to reimplement the
  * wheel - mapping_seek_hole_data() does exactly what we need to identify the
  * start and end of data ranges correctly even for sub-folio block sizes. This
@@ -1177,7 +1203,7 @@ static void iomap_write_delalloc_scan(struct inode *inode,
  * require sprinkling this code with magic "+ 1" and "- 1" arithmetic and expose
  * the code to subtle off-by-one bugs....
  */
-static void iomap_write_delalloc_release(struct inode *inode, loff_t start_byte,
+void iomap_write_delalloc_release(struct inode *inode, loff_t start_byte,
 		loff_t end_byte, unsigned flags, struct iomap *iomap,
 		iomap_punch_t punch)
 {
@@ -1185,12 +1211,13 @@ static void iomap_write_delalloc_release(struct inode *inode, loff_t start_byte,
 	loff_t scan_end_byte = min(i_size_read(inode), end_byte);
 
 	/*
-	 * Lock the mapping to avoid races with page faults re-instantiating
+	 * The caller must hold invalidate_lock to avoid races with page faults
-	 * folios and dirtying them via ->page_mkwrite whilst we walk the
+	 * re-instantiating folios and dirtying them via ->page_mkwrite whilst
-	 * cache and perform delalloc extent removal. Failing to do this can
+	 * we walk the cache and perform delalloc extent removal.  Failing to do
-	 * leave dirty pages with no space reservation in the cache.
+	 * this can leave dirty pages with no space reservation in the cache.
 	 */
-	filemap_invalidate_lock(inode->i_mapping);
+	lockdep_assert_held_write(&inode->i_mapping->invalidate_lock);
+
 	while (start_byte < scan_end_byte) {
 		loff_t		data_end;
 
@@ -1207,7 +1234,7 @@ static void iomap_write_delalloc_release(struct inode *inode, loff_t start_byte,
 		if (start_byte == -ENXIO || start_byte == scan_end_byte)
 			break;
 		if (WARN_ON_ONCE(start_byte < 0))
-			goto out_unlock;
+			return;
 		WARN_ON_ONCE(start_byte < punch_start_byte);
 		WARN_ON_ONCE(start_byte > scan_end_byte);
 
@@ -1218,7 +1245,7 @@ static void iomap_write_delalloc_release(struct inode *inode, loff_t start_byte,
 		data_end = mapping_seek_hole_data(inode->i_mapping, start_byte,
 				scan_end_byte, SEEK_HOLE);
 		if (WARN_ON_ONCE(data_end < 0))
-			goto out_unlock;
+			return;
 
 		/*
 		 * If we race with post-direct I/O invalidation of the page cache,
@@ -1240,74 +1267,8 @@ static void iomap_write_delalloc_release(struct inode *inode, loff_t start_byte,
 	if (punch_start_byte < end_byte)
 		punch(inode, punch_start_byte, end_byte - punch_start_byte,
 				iomap);
-out_unlock:
-	filemap_invalidate_unlock(inode->i_mapping);
 }
-
+EXPORT_SYMBOL_GPL(iomap_write_delalloc_release);
-/*
- * When a short write occurs, the filesystem may need to remove reserved space
- * that was allocated in ->iomap_begin from it's ->iomap_end method. For
- * filesystems that use delayed allocation, we need to punch out delalloc
- * extents from the range that are not dirty in the page cache. As the write can
- * race with page faults, there can be dirty pages over the delalloc extent
- * outside the range of a short write but still within the delalloc extent
- * allocated for this iomap.
- *
- * This function uses [start_byte, end_byte) intervals (i.e. open ended) to
- * simplify range iterations.
- *
- * The punch() callback *must* only punch delalloc extents in the range passed
- * to it. It must skip over all other types of extents in the range and leave
- * them completely unchanged. It must do this punch atomically with respect to
- * other extent modifications.
- *
- * The punch() callback may be called with a folio locked to prevent writeback
- * extent allocation racing at the edge of the range we are currently punching.
- * The locked folio may or may not cover the range being punched, so it is not
- * safe for the punch() callback to lock folios itself.
- *
- * Lock order is:
- *
- * inode->i_rwsem (shared or exclusive)
- *   inode->i_mapping->invalidate_lock (exclusive)
- *     folio_lock()
- *       ->punch
- *         internal filesystem allocation lock
- */
-void iomap_file_buffered_write_punch_delalloc(struct inode *inode,
-		loff_t pos, loff_t length, ssize_t written, unsigned flags,
-		struct iomap *iomap, iomap_punch_t punch)
-{
-	loff_t			start_byte;
-	loff_t			end_byte;
-	unsigned int		blocksize = i_blocksize(inode);
-
-	if (iomap->type != IOMAP_DELALLOC)
-		return;
-
-	/* If we didn't reserve the blocks, we're not allowed to punch them. */
-	if (!(iomap->flags & IOMAP_F_NEW))
-		return;
-
-	/*
-	 * start_byte refers to the first unused block after a short write. If
-	 * nothing was written, round offset down to point at the first block in
-	 * the range.
-	 */
-	if (unlikely(!written))
-		start_byte = round_down(pos, blocksize);
-	else
-		start_byte = round_up(pos + written, blocksize);
-	end_byte = round_up(pos + length, blocksize);
-
-	/* Nothing to do if we've written the entire delalloc extent */
-	if (start_byte >= end_byte)
-		return;
-
-	iomap_write_delalloc_release(inode, start_byte, end_byte, flags, iomap,
-			punch);
-}
-EXPORT_SYMBOL_GPL(iomap_file_buffered_write_punch_delalloc);
 
 static loff_t iomap_unshare_iter(struct iomap_iter *iter)
 {

fs/xfs/scrub/bmap_repair.c

@@ -801,7 +801,7 @@ xrep_bmap(
 {
 	struct xrep_bmap	*rb;
 	char			*descr;
-	unsigned int		max_bmbt_recs;
+	xfs_extnum_t		max_bmbt_recs;
 	bool			large_extcount;
 	int			error = 0;
 

fs/xfs/xfs_aops.c

@@ -116,7 +116,7 @@ xfs_end_ioend(
 	if (unlikely(error)) {
 		if (ioend->io_flags & IOMAP_F_SHARED) {
 			xfs_reflink_cancel_cow_range(ip, offset, size, true);
-			xfs_bmap_punch_delalloc_range(ip, offset,
+			xfs_bmap_punch_delalloc_range(ip, XFS_DATA_FORK, offset,
 					offset + size);
 		}
 		goto done;
@@ -456,7 +456,7 @@ xfs_discard_folio(
 	 * byte of the next folio. Hence the end offset is only dependent on the
 	 * folio itself and not the start offset that is passed in.
 	 */
-	xfs_bmap_punch_delalloc_range(ip, pos,
+	xfs_bmap_punch_delalloc_range(ip, XFS_DATA_FORK, pos,
 				folio_pos(folio) + folio_size(folio));
 }
 

fs/xfs/xfs_bmap_util.c

@@ -442,11 +442,12 @@ out_unlock_iolock:
 void
 xfs_bmap_punch_delalloc_range(
 	struct xfs_inode	*ip,
+	int			whichfork,
 	xfs_off_t		start_byte,
 	xfs_off_t		end_byte)
 {
 	struct xfs_mount	*mp = ip->i_mount;
-	struct xfs_ifork	*ifp = &ip->i_df;
+	struct xfs_ifork	*ifp = xfs_ifork_ptr(ip, whichfork);
 	xfs_fileoff_t		start_fsb = XFS_B_TO_FSBT(mp, start_byte);
 	xfs_fileoff_t		end_fsb = XFS_B_TO_FSB(mp, end_byte);
 	struct xfs_bmbt_irec	got, del;
@@ -474,11 +475,14 @@ xfs_bmap_punch_delalloc_range(
 			continue;
 		}
 
-		xfs_bmap_del_extent_delay(ip, XFS_DATA_FORK, &icur, &got, &del);
+		xfs_bmap_del_extent_delay(ip, whichfork, &icur, &got, &del);
 		if (!xfs_iext_get_extent(ifp, &icur, &got))
 			break;
 	}
 
+	if (whichfork == XFS_COW_FORK && !ifp->if_bytes)
+		xfs_inode_clear_cowblocks_tag(ip);
+
 out_unlock:
 	xfs_iunlock(ip, XFS_ILOCK_EXCL);
 }
@@ -580,7 +584,7 @@ xfs_free_eofblocks(
 	 */
 	if (ip->i_diflags & (XFS_DIFLAG_PREALLOC | XFS_DIFLAG_APPEND)) {
 		if (ip->i_delayed_blks) {
-			xfs_bmap_punch_delalloc_range(ip,
+			xfs_bmap_punch_delalloc_range(ip, XFS_DATA_FORK,
 				round_up(XFS_ISIZE(ip), mp->m_sb.sb_blocksize),
 				LLONG_MAX);
 		}

fs/xfs/xfs_bmap_util.h

@@ -30,7 +30,7 @@ xfs_bmap_rtalloc(struct xfs_bmalloca *ap)
 }
 #endif /* CONFIG_XFS_RT */
 
-void	xfs_bmap_punch_delalloc_range(struct xfs_inode *ip,
+void	xfs_bmap_punch_delalloc_range(struct xfs_inode *ip, int whichfork,
 		xfs_off_t start_byte, xfs_off_t end_byte);
 
 struct kgetbmap {

fs/xfs/xfs_file.c

@@ -347,10 +347,83 @@ xfs_file_splice_read(
 	return ret;
 }
 
+/*
+ * Take care of zeroing post-EOF blocks when they might exist.
+ *
+ * Returns 0 if successfully, a negative error for a failure, or 1 if this
+ * function dropped the iolock and reacquired it exclusively and the caller
+ * needs to restart the write sanity checks.
+ */
+static ssize_t
+xfs_file_write_zero_eof(
+	struct kiocb		*iocb,
+	struct iov_iter		*from,
+	unsigned int		*iolock,
+	size_t			count,
+	bool			*drained_dio)
+{
+	struct xfs_inode	*ip = XFS_I(iocb->ki_filp->f_mapping->host);
+	loff_t			isize;
+	int			error;
+
+	/*
+	 * We need to serialise against EOF updates that occur in IO completions
+	 * here. We want to make sure that nobody is changing the size while
+	 * we do this check until we have placed an IO barrier (i.e. hold
+	 * XFS_IOLOCK_EXCL) that prevents new IO from being dispatched.  The
+	 * spinlock effectively forms a memory barrier once we have
+	 * XFS_IOLOCK_EXCL so we are guaranteed to see the latest EOF value and
+	 * hence be able to correctly determine if we need to run zeroing.
+	 */
+	spin_lock(&ip->i_flags_lock);
+	isize = i_size_read(VFS_I(ip));
+	if (iocb->ki_pos <= isize) {
+		spin_unlock(&ip->i_flags_lock);
+		return 0;
+	}
+	spin_unlock(&ip->i_flags_lock);
+
+	if (iocb->ki_flags & IOCB_NOWAIT)
+		return -EAGAIN;
+
+	if (!*drained_dio) {
+		/*
+		 * If zeroing is needed and we are currently holding the iolock
+		 * shared, we need to update it to exclusive which implies
+		 * having to redo all checks before.
+		 */
+		if (*iolock == XFS_IOLOCK_SHARED) {
+			xfs_iunlock(ip, *iolock);
+			*iolock = XFS_IOLOCK_EXCL;
+			xfs_ilock(ip, *iolock);
+			iov_iter_reexpand(from, count);
+		}
+
+		/*
+		 * We now have an IO submission barrier in place, but AIO can do
+		 * EOF updates during IO completion and hence we now need to
+		 * wait for all of them to drain.  Non-AIO DIO will have drained
+		 * before we are given the XFS_IOLOCK_EXCL, and so for most
+		 * cases this wait is a no-op.
+		 */
+		inode_dio_wait(VFS_I(ip));
+		*drained_dio = true;
+		return 1;
+	}
+
+	trace_xfs_zero_eof(ip, isize, iocb->ki_pos - isize);
+
+	xfs_ilock(ip, XFS_MMAPLOCK_EXCL);
+	error = xfs_zero_range(ip, isize, iocb->ki_pos - isize, NULL);
+	xfs_iunlock(ip, XFS_MMAPLOCK_EXCL);
+
+	return error;
+}
+
 /*
  * Common pre-write limit and setup checks.
  *
- * Called with the iolocked held either shared and exclusive according to
+ * Called with the iolock held either shared and exclusive according to
  * @iolock, and returns with it held.  Might upgrade the iolock to exclusive
  * if called for a direct write beyond i_size.
  */
@@ -360,13 +433,10 @@ xfs_file_write_checks(
 	struct iov_iter		*from,
 	unsigned int		*iolock)
 {
-	struct file		*file = iocb->ki_filp;
+	struct inode		*inode = iocb->ki_filp->f_mapping->host;
-	struct inode		*inode = file->f_mapping->host;
-	struct xfs_inode	*ip = XFS_I(inode);
-	ssize_t			error = 0;
 	size_t			count = iov_iter_count(from);
 	bool			drained_dio = false;
-	loff_t			isize;
+	ssize_t			error;
 
 restart:
 	error = generic_write_checks(iocb, from);
@@ -389,7 +459,7 @@ restart:
 	 * exclusively.
 	 */
 	if (*iolock == XFS_IOLOCK_SHARED && !IS_NOSEC(inode)) {
-		xfs_iunlock(ip, *iolock);
+		xfs_iunlock(XFS_I(inode), *iolock);
 		*iolock = XFS_IOLOCK_EXCL;
 		error = xfs_ilock_iocb(iocb, *iolock);
 		if (error) {
@@ -400,64 +470,24 @@ restart:
 	}
 
 	/*
-	 * If the offset is beyond the size of the file, we need to zero any
+	 * If the offset is beyond the size of the file, we need to zero all
 	 * blocks that fall between the existing EOF and the start of this
-	 * write.  If zeroing is needed and we are currently holding the iolock
+	 * write.
-	 * shared, we need to update it to exclusive which implies having to
-	 * redo all checks before.
 	 *
-	 * We need to serialise against EOF updates that occur in IO completions
+	 * We can do an unlocked check for i_size here safely as I/O completion
-	 * here. We want to make sure that nobody is changing the size while we
+	 * can only extend EOF.  Truncate is locked out at this point, so the
-	 * do this check until we have placed an IO barrier (i.e.  hold the
+	 * EOF can not move backwards, only forwards. Hence we only need to take
-	 * XFS_IOLOCK_EXCL) that prevents new IO from being dispatched.  The
+	 * the slow path when we are at or beyond the current EOF.
-	 * spinlock effectively forms a memory barrier once we have the
-	 * XFS_IOLOCK_EXCL so we are guaranteed to see the latest EOF value and
-	 * hence be able to correctly determine if we need to run zeroing.
-	 *
-	 * We can do an unlocked check here safely as IO completion can only
-	 * extend EOF. Truncate is locked out at this point, so the EOF can
-	 * not move backwards, only forwards. Hence we only need to take the
-	 * slow path and spin locks when we are at or beyond the current EOF.
 	 */
-	if (iocb->ki_pos <= i_size_read(inode))
+	if (iocb->ki_pos > i_size_read(inode)) {
-		goto out;
+		error = xfs_file_write_zero_eof(iocb, from, iolock, count,
-
+				&drained_dio);
-	spin_lock(&ip->i_flags_lock);
+		if (error == 1)
-	isize = i_size_read(inode);
-	if (iocb->ki_pos > isize) {
-		spin_unlock(&ip->i_flags_lock);
-
-		if (iocb->ki_flags & IOCB_NOWAIT)
-			return -EAGAIN;
-
-		if (!drained_dio) {
-			if (*iolock == XFS_IOLOCK_SHARED) {
-				xfs_iunlock(ip, *iolock);
-				*iolock = XFS_IOLOCK_EXCL;
-				xfs_ilock(ip, *iolock);
-				iov_iter_reexpand(from, count);
-			}
-			/*
-			 * We now have an IO submission barrier in place, but
-			 * AIO can do EOF updates during IO completion and hence
-			 * we now need to wait for all of them to drain. Non-AIO
-			 * DIO will have drained before we are given the
-			 * XFS_IOLOCK_EXCL, and so for most cases this wait is a
-			 * no-op.
-			 */
-			inode_dio_wait(inode);
-			drained_dio = true;
 			goto restart;
-		}
-
-		trace_xfs_zero_eof(ip, isize, iocb->ki_pos - isize);
-		error = xfs_zero_range(ip, isize, iocb->ki_pos - isize, NULL);
 		if (error)
 			return error;
-	} else
+	}
-		spin_unlock(&ip->i_flags_lock);
 
-out:
 	return kiocb_modified(iocb);
 }
 

fs/xfs/xfs_iomap.c

@@ -975,6 +975,7 @@ xfs_buffered_write_iomap_begin(
 	int			allocfork = XFS_DATA_FORK;
 	int			error = 0;
 	unsigned int		lockmode = XFS_ILOCK_EXCL;
+	unsigned int		iomap_flags = 0;
 	u64			seq;
 
 	if (xfs_is_shutdown(mp))
@@ -1145,6 +1146,11 @@ xfs_buffered_write_iomap_begin(
 		}
 	}
 
+	/*
+	 * Flag newly allocated delalloc blocks with IOMAP_F_NEW so we punch
+	 * them out if the write happens to fail.
+	 */
+	iomap_flags |= IOMAP_F_NEW;
 	if (allocfork == XFS_COW_FORK) {
 		error = xfs_bmapi_reserve_delalloc(ip, allocfork, offset_fsb,
 				end_fsb - offset_fsb, prealloc_blocks, &cmap,
@@ -1162,19 +1168,11 @@ xfs_buffered_write_iomap_begin(
 	if (error)
 		goto out_unlock;
 
-	/*
-	 * Flag newly allocated delalloc blocks with IOMAP_F_NEW so we punch
-	 * them out if the write happens to fail.
-	 */
-	seq = xfs_iomap_inode_sequence(ip, IOMAP_F_NEW);
-	xfs_iunlock(ip, lockmode);
 	trace_xfs_iomap_alloc(ip, offset, count, allocfork, &imap);
-	return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, IOMAP_F_NEW, seq);
-
 found_imap:
-	seq = xfs_iomap_inode_sequence(ip, 0);
+	seq = xfs_iomap_inode_sequence(ip, iomap_flags);
 	xfs_iunlock(ip, lockmode);
-	return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, 0, seq);
+	return xfs_bmbt_to_iomap(ip, iomap, &imap, flags, iomap_flags, seq);
 
 convert_delay:
 	xfs_iunlock(ip, lockmode);
@@ -1188,20 +1186,20 @@ convert_delay:
 	return 0;
 
 found_cow:
-	seq = xfs_iomap_inode_sequence(ip, 0);
 	if (imap.br_startoff <= offset_fsb) {
-		error = xfs_bmbt_to_iomap(ip, srcmap, &imap, flags, 0, seq);
+		error = xfs_bmbt_to_iomap(ip, srcmap, &imap, flags, 0,
+				xfs_iomap_inode_sequence(ip, 0));
 		if (error)
 			goto out_unlock;
-		seq = xfs_iomap_inode_sequence(ip, IOMAP_F_SHARED);
+	} else {
-		xfs_iunlock(ip, lockmode);
+		xfs_trim_extent(&cmap, offset_fsb,
-		return xfs_bmbt_to_iomap(ip, iomap, &cmap, flags,
+				imap.br_startoff - offset_fsb);
-					 IOMAP_F_SHARED, seq);
 	}
 
-	xfs_trim_extent(&cmap, offset_fsb, imap.br_startoff - offset_fsb);
+	iomap_flags |= IOMAP_F_SHARED;
+	seq = xfs_iomap_inode_sequence(ip, iomap_flags);
 	xfs_iunlock(ip, lockmode);
-	return xfs_bmbt_to_iomap(ip, iomap, &cmap, flags, 0, seq);
+	return xfs_bmbt_to_iomap(ip, iomap, &cmap, flags, iomap_flags, seq);
 
 out_unlock:
 	xfs_iunlock(ip, lockmode);
@@ -1215,7 +1213,10 @@ xfs_buffered_write_delalloc_punch(
 	loff_t			length,
 	struct iomap		*iomap)
 {
-	xfs_bmap_punch_delalloc_range(XFS_I(inode), offset, offset + length);
+	xfs_bmap_punch_delalloc_range(XFS_I(inode),
+			(iomap->flags & IOMAP_F_SHARED) ?
+				XFS_COW_FORK : XFS_DATA_FORK,
+			offset, offset + length);
 }
 
 static int
@@ -1227,8 +1228,30 @@ xfs_buffered_write_iomap_end(
 	unsigned		flags,
 	struct iomap		*iomap)
 {
-	iomap_file_buffered_write_punch_delalloc(inode, offset, length, written,
+	loff_t			start_byte, end_byte;
-			flags, iomap, &xfs_buffered_write_delalloc_punch);
+
+	/* If we didn't reserve the blocks, we're not allowed to punch them. */
+	if (iomap->type != IOMAP_DELALLOC || !(iomap->flags & IOMAP_F_NEW))
+		return 0;
+
+	/* Nothing to do if we've written the entire delalloc extent */
+	start_byte = iomap_last_written_block(inode, offset, written);
+	end_byte = round_up(offset + length, i_blocksize(inode));
+	if (start_byte >= end_byte)
+		return 0;
+
+	/* For zeroing operations the callers already hold invalidate_lock. */
+	if (flags & (IOMAP_UNSHARE | IOMAP_ZERO)) {
+		rwsem_assert_held_write(&inode->i_mapping->invalidate_lock);
+		iomap_write_delalloc_release(inode, start_byte, end_byte, flags,
+				iomap, xfs_buffered_write_delalloc_punch);
+	} else {
+		filemap_invalidate_lock(inode->i_mapping);
+		iomap_write_delalloc_release(inode, start_byte, end_byte, flags,
+				iomap, xfs_buffered_write_delalloc_punch);
+		filemap_invalidate_unlock(inode->i_mapping);
+	}
+
 	return 0;
 }
 
@@ -1435,6 +1458,8 @@ xfs_zero_range(
 {
 	struct inode		*inode = VFS_I(ip);
 
+	xfs_assert_ilocked(ip, XFS_IOLOCK_EXCL | XFS_MMAPLOCK_EXCL);
+
 	if (IS_DAX(inode))
 		return dax_zero_range(inode, pos, len, did_zero,
 				      &xfs_dax_write_iomap_ops);

include/linux/iomap.h

@@ -256,6 +256,20 @@ static inline const struct iomap *iomap_iter_srcmap(const struct iomap_iter *i)
 	return &i->iomap;
 }
 
+/*
+ * Return the file offset for the first unchanged block after a short write.
+ *
+ * If nothing was written, round @pos down to point at the first block in
+ * the range, else round up to include the partially written block.
+ */
+static inline loff_t iomap_last_written_block(struct inode *inode, loff_t pos,
+		ssize_t written)
+{
+	if (unlikely(!written))
+		return round_down(pos, i_blocksize(inode));
+	return round_up(pos + written, i_blocksize(inode));
+}
+
 ssize_t iomap_file_buffered_write(struct kiocb *iocb, struct iov_iter *from,
 		const struct iomap_ops *ops, void *private);
 int iomap_read_folio(struct folio *folio, const struct iomap_ops *ops);
@@ -276,9 +290,9 @@ vm_fault_t iomap_page_mkwrite(struct vm_fault *vmf,
 
 typedef void (*iomap_punch_t)(struct inode *inode, loff_t offset, loff_t length,
 		struct iomap *iomap);
-void iomap_file_buffered_write_punch_delalloc(struct inode *inode, loff_t pos,
+void iomap_write_delalloc_release(struct inode *inode, loff_t start_byte,
-		loff_t length, ssize_t written, unsigned flag,
+		loff_t end_byte, unsigned flags, struct iomap *iomap,
-		struct iomap *iomap, iomap_punch_t punch);
+		iomap_punch_t punch);
 
 int iomap_fiemap(struct inode *inode, struct fiemap_extent_info *fieinfo,
 		u64 start, u64 len, const struct iomap_ops *ops);