This patch and another one that you have [1] cause a ton of build
warnings with clang, as reported by the 0day bot, because of the
implicit conversions from floating points to integer. Is this
intentional? Would you mind taking a look at them and addressing them if
you do another revision?