> set_unique_reg note already makes a number of checks so that its > multitude of callers doesn't have to. E.g. it checks that there is > indeed only one set, not only one set of a live register.
Indeed, but not on the DATUM. > As you can see in the PR, there are different pieces in the compiler that > end up in the same bogosity as they add notes with the wrong mode of data. This is the bug: you shouldn't call set_unique_reg_note on a DATUM with a mode different from that of the destination. This usually means that the logic in the caller is confused. This is the same interface as emit_move_insn. -- Eric Botcazou
