erofs: ensure m_llen is reset to 0 if metadata is invalid

Sometimes, the on-disk metadata might be invalid due to user
interrupts, storage failures, or other unknown causes.

In that case, z_erofs_map_blocks_iter() may still return a valid
m_llen while other fields remain invalid (e.g., m_plen can be 0).

Due to the return value of z_erofs_scan_folio() in some path will
be ignored on purpose, the following z_erofs_scan_folio() could
then use the invalid value by accident.

Let's reset m_llen to 0 to prevent this.

Link: https://lore.kernel.org/r/20240629185743.2819229-1-hsiangkao@linux.alibaba.com
Signed-off-by: Gao Xiang <hsiangkao@linux.alibaba.com>

diff --git a/fs/erofs/zmap.c b/fs/erofs/zmap.c
index 9b248ee5fef2d26e0fae60d67e0c4318d8a1f0bd..74d3d7bffcf3f1976ca41c6036a62ad0b3e2d1b4 100644 (file)
--- a/fs/erofs/zmap.c
+++ b/fs/erofs/zmap.c
@@ -711,6 +711,8 @@ int z_erofs_map_blocks_iter(struct inode *inode, struct erofs_map_blocks *map,
 
        err = z_erofs_do_map_blocks(inode, map, flags);
 out:
+       if (err)
+               map->m_llen = 0;
        trace_z_erofs_map_blocks_iter_exit(inode, map, flags, err);
        return err;
 }