Now that softfloat-native is gone, there is no real point on not always
enabling floatx80 and float128 support.
Reviewed-by: NPeter Maydell <peter.maydell@linaro.org>
Signed-off-by: NAurelien Jarno <aurelien@aurel32.net>