gcc 10.2 warns about switch cases that can fall through without explicitly allowing it to fall through.
In ckddasd.c, line 1461 I think gcc has found a bug, because all the cases there fall through to the case that ultimately sets shift to 4. That doesn't look like it's intended.