btrfs: make validate_extent_map() catch ram_bytes mismatch

Previously validate_extent_map() is only to catch bugs related to
extent_map member cleanups.

But with recent btrfs-check enhancement to catch ram_bytes mismatch with
disk_num_bytes, it would be much better to catch such extent maps
earlier.

So this patch adds extra ram_bytes validation for extent maps.

Please note that, older filesystems with such mismatch won't trigger this error:

- extent_map::ram_bytes is already fixed
  Previous patch has already fixed the ram_bytes for affected file
  extents.

So this enhanced sanity check should not affect end users.

Reviewed-by: Filipe Manana <fdmanana@suse.com>
Signed-off-by: Qu Wenruo <wqu@suse.com>
Signed-off-by: David Sterba <dsterba@suse.com>

diff --git a/fs/btrfs/extent_map.c b/fs/btrfs/extent_map.c
index b869a0ee24d2826c6ec6da55dd49b0e04eda821a..6961cc73fe3fc9cfeb98fd5b176869f0455959ae 100644 (file)
--- a/fs/btrfs/extent_map.c
+++ b/fs/btrfs/extent_map.c
@@ -317,6 +317,11 @@ static void validate_extent_map(struct btrfs_fs_info *fs_info, struct extent_map
                if (em->offset + em->len > em->disk_num_bytes &&
                    !extent_map_is_compressed(em))
                        dump_extent_map(fs_info, "disk_num_bytes too small", em);
+               if (!extent_map_is_compressed(em) &&
+                   em->ram_bytes != em->disk_num_bytes)
+                       dump_extent_map(fs_info,
+               "ram_bytes mismatch with disk_num_bytes for non-compressed em",
+                                       em);
        } else if (em->offset) {
                dump_extent_map(fs_info, "non-zero offset for hole/inline", em);
        }