If the populate-volatile.sh script encounters a blank line, it duplicates the previous one due to none of the column variables getting new values. Under normal circumstances, this will not manifest in any visible manner. However, if a read-only-rootfs is being used, this causes the host directory path to be leaked into the root file system. This is caused by TNAME getting ROOT_DIR prepended to it twice, once for the actual parsed line, and then again when the blank line is "parsed".

For example, if your rootfs working directory is at /home/me/rootfs, and you have a volatiles file like:

 d root root 0755 /foo/bar
 

Then your rootfs in the working directory will get:

 /home/me/rootfs/foo/bar

And then another entry for the blank line that looks like

 /home/me/rootfs/home/me/rootfs/foo/bar

Then end effect of which is that the final root file system has the paths
 /foo/bar
 /home/me/rootfs/foo/bar