As indicated the other day there were quite a few shortcomings. 1: re-work operand swapping for FMA4 and 4-operand XOP insns 2: re-work operand handling for 5-operand XOP insns 3: re-work operand swapping for XOP shift/rotate insns 4: FMA4 scalar insns ignore VEX.L 5: various XOP insns lack L and/or W bit decoding Jan