Re: [PATCH RESEND] pwm: th1520: Remove requirement for mul_u64_u64_div_u64_roundup

["Gary Guo" <gary@garyguo.net>]

On Mon May 25, 2026 at 4:51 PM BST, Maurice Hieronymus wrote:
> On Mon, 2026-05-25 at 16:36 +0200, Michal Wilczynski wrote:
>> 
>> To move forward, we should fix this by implementing the proper
>> generic
>> 64-bit math helper in the Rust abstractions, rather than narrowing
>> the
>> types here. Are you open to looking into adding the Rust equivalent
>> for
>> mul_u64_u64_div_u64_roundup instead?
>
> I would be definitely down, to help with that.
>
> Before I start working on it, a couple of clarification questions.
>
> Should I implement this as a thin FFI wrapper calling the C-Functions?
> Or do you prefer to have a re-implementation in Rust.
>
> Since mul_u64_u64_div_u64_roundup is a macro: Is it okay if this will
> be a small Rust function or should it be a macro as well?

This should definitely not be a macro. You should only use Rust macros when it
needs to be a macro. This function should just be a method of `u64` added via
an extension trait, so you could use `u64::mul_div_ceil` to invoke it.

The implementation of `mul_u64_add_u64_div_u64`is non-trivial, and thus you
should defer to the C function to avoid re-implementing the same code. However,
I think it'll be fine to first wrap the `mul_u64_add_u64_div_u64` as
`u64::mul_add_div()` and then have `u64::mul_div_ceil()` be

    #[inline]
    fn mul_div_ceil(self, mul: Self, div: Self) -> Self {
        self.mul_add_div(mul, div - 1, div)
    }

Best,
Gary