[05/11] xfs: forcibly convert unwritten blocks within an rt extent before sharing

Commit Message

Darrick J. Wong Dec. 13, 2024, 1:22 a.m. UTC

From: Darrick J. Wong <[e-mail]>

As noted in the previous patch, XFS can only unmap and map full rt
extents.  This means that we cannot stop mid-extent for any reason,
including stepping around unwritten/written extents.  Second, the
reflink and CoW mechanisms were not designed to handle shared unwritten
extents, so we have to do something to get rid of them.

If the user asks us to remap two files, we must scan both ranges
beforehand to convert any unwritten extents that are not aligned to rt
extent boundaries into zeroed written extents before sharing.

Signed-off-by: "Darrick J. Wong" <[e-mail]>
---
 fs/xfs/xfs_reflink.c |   19 +++++++++++++++++++
 1 file changed, 19 insertions(+)

Patch

diff --git a/fs/xfs/xfs_reflink.c b/fs/xfs/xfs_reflink.c
index 4f87f7041995c4..82ceec8517a020 100644
--- a/fs/xfs/xfs_reflink.c
+++ b/fs/xfs/xfs_reflink.c
@@ -1666,6 +1666,25 @@  xfs_reflink_remap_prep(
 	if (ret)
 		goto out_unlock;
 
+	/*
+	 * Now that we've marked both inodes for reflink, make sure that all
+	 * allocation units (AU) mapped into either files' ranges are either
+	 * wholly written, wholly unwritten, or holes.  The bmap code requires
+	 * that we align all unmap and remap requests to an AU.  We've already
+	 * flushed the page cache and finished directio for the range that's
+	 * being remapped, so we can convert the mappings directly.
+	 */
+	if (xfs_inode_has_bigrtalloc(src)) {
+		ret = xfs_convert_rtbigalloc_file_space(src, pos_in, *len);
+		if (ret)
+			goto out_unlock;
+	}
+	if (xfs_inode_has_bigrtalloc(dest)) {
+		ret = xfs_convert_rtbigalloc_file_space(dest, pos_out, *len);
+		if (ret)
+			goto out_unlock;
+	}
+
 	/*
 	 * If pos_out > EOF, we may have dirtied blocks between EOF and
 	 * pos_out. In that case, we need to extend the flush and unmap to cover