https://gcc.gnu.org/bugzilla/show_bug.cgi?id=128872
--- Comment #2 from Drea Pinski <pinskia at gcc dot gnu.org> --- unsigned short _5; _5 = .SAT_SUB (c_6(D), b_7(D)); The problem is the source is a saturating subtraction in 8bit but the return value says it is in 16bit which is wrong.
