cmsis-dsp-sys

rust bindings for the prebuilt cmsis-dsp math

prereqs

Currently only requires unzip and curl cli commands. Happy to take prs to use pure rust instead. Downloads the cmsis math pack file during build.

usage

use cmsis_dsp_sys::{arm_sin_f32};

let sin = unsafe { arm_sin_f32(core::f32::consts::PI) };
assert_eq!(sin, 0f32);

the dirty details

Its complicated. Im currently not building any c files that accompany the arm dsp libraries. Bringing the necessary libc, libm and abstracting against every possible architecture and build flag seems hopeless atm. As a result you're going to have to do some work on your end.

When you start to do anything complicated you're probably going to see something like:

  = note: rust-lld: error: undefined symbol: sqrtf
          >>> referenced by arm_math.h:6841 (../../Include/arm_math.h:6841)
          >>>               arm_cmplx_mag_f32.o:(arm_cmplx_mag_f32) in archive /home/jacob/Downloads/dsp-discoveryf4-rust/lab4/libarm_cortexM4lf_math.a

Which means you need to bring some libm type function with you. Possibilitie pure rust options I've seen include libm or micromath with differing tradeoffs. Stub one in with:

use micromath::F32Ext;

//C needs access to a sqrt fn, lets use micromath
#[no_mangle]
pub extern "C" fn sqrtf(x: f32) -> f32 {
    x.sqrt()
}

Further you're going to want to use something from arm_const_structs.c like arm_cfft_sR_f32_len16 which uses tables in arm_common_tables.c. I would look those up inside the cmsis pack and translate it into some arm_common_tables.rs file in your project

#![allow(clippy::excessive_precision)]
#![allow(clippy::approx_constant)]
#![allow(non_snake_case)]
#![allow(non_upper_case_globals)]
#![allow(unused)]

pub const ARMBITREVINDEXTABLE_16_TABLE_LENGTH: u16 = 20;

pub static armBitRevIndexTable16: &[u16] = &[
    /* 8x2, size 20 */
    8, 64, 24, 72, 16, 64, 40, 80, 32, 64, 56, 88, 48, 72, 88, 104, 72, 96, 104, 112,
];

pub static twiddleCoef_16: &[f32] = &[
    1.000000000,  0.000000000,
    0.923879533,  0.382683432,
    0.707106781,  0.707106781,
    0.382683432,  0.923879533,
    0.000000000,  1.000000000,
   -0.382683432,  0.923879533,
   -0.707106781,  0.707106781,
   -0.923879533,  0.382683432,
   -1.000000000,  0.000000000,
   -0.923879533, -0.382683432,
   -0.707106781, -0.707106781,
   -0.382683432, -0.923879533,
   -0.000000000, -1.000000000,
    0.382683432, -0.923879533,
    0.707106781, -0.707106781,
    0.923879533, -0.382683432,
];

mod arm_common_tables;
use arm_common_tables::{
    armBitRevIndexTable16, twiddleCoef_16, ARMBITREVINDEXTABLE_16_TABLE_LENGTH,
};
use cmsis_dsp_sys::{arm_cfft_f32, arm_cfft_instance_f32, arm_cmplx_mag_f32};

//your data as stored as real and imaginary pairs here
let mut dtfsecoef = [0f32; 32];

let cfft = arm_cfft_instance_f32 {
    fftLen: 16,
    pTwiddle: twiddleCoef_16.as_ptr(),
    pBitRevTable: armBitRevIndexTable16.as_ptr(),
    bitRevLength: ARMBITREVINDEXTABLE_16_TABLE_LENGTH,
};

let mut mag = [0f32; 16];

//Coefficient calculation with CFFT function
unsafe {

    //CFFT calculation
    arm_cfft_f32(&cfft, dtfsecoef.as_mut_ptr(), 0, 1);

    // Magnitude calculation
    arm_cmplx_mag_f32(s.as_ptr(), mag.as_mut_ptr(), N::to_usize() as uint32_t);

}