[Feature] 自动配置can波特率的可行性

Describe problem solved by the proposed feature

目前的bsp配置CAN通信波特率时使用了大量预定义的表格 不同的芯片如果当前不支持 就要新增宏定义
是否可以考虑实现一个函数自动计算适配的波特率参数呢 例如

#include <math.h>
#include <float.h>
#include <stdint.h>
#include <stdio.h>

typedef struct
{
    uint32_t prescaler;
    uint8_t  bs1;
    uint8_t  bs2;
    float    baud_error;
    float    sample_point;
} CAN_TimingConfig;

/**
 * @brief 
 * 
 * @param F_pclk CAN的APB1时钟
 * @param target_baudrate 目标波特率
 * @return CAN_TimingConfig 
 */
CAN_TimingConfig calculate_can_timing(uint32_t F_pclk,
                                      uint32_t target_baudrate)
{
    CAN_TimingConfig best                = {0};
    float            min_error           = FLT_MAX;
    float            target_sample_point = 0.8f;
    const uint8_t    tq_min = 8, tq_max = 25;
    const uint8_t    bs1_max = 16, bs2_max = 8;

    // 扩展分频范围避免遗漏
    uint32_t prescale_max = (F_pclk + target_baudrate * tq_min - 1) / (target_baudrate * tq_min);
    uint32_t prescale_min = F_pclk / (target_baudrate * tq_max);
    prescale_min          = prescale_min > 0 ? prescale_min : 1;
    prescale_max          = prescale_max > 1024 ? 1024 : prescale_max;

    // 逆序循环优先大分频值
    for (uint32_t prescale = prescale_max; prescale >= prescale_min; --prescale)
    {
        float   total_tq = (float)F_pclk / (target_baudrate * prescale);
        uint8_t tq       = (uint8_t)(total_tq + 0.5f);

        if (tq < tq_min || tq > tq_max) continue;

        uint8_t bs1 = (uint8_t)(target_sample_point * tq - 1 + 0.5f);
        uint8_t bs2 = tq - 1 - bs1;

        if (bs1 < 1 || bs1 > bs1_max || bs2 < 1 || bs2 > bs2_max) continue;

        float actual_baud  = (float)F_pclk / (prescale * (1 + bs1 + bs2));
        float baud_err     = fabsf(actual_baud - target_baudrate) / target_baudrate;
        float sample_point = (1.0f + bs1) / (1 + bs1 + bs2);
        float sample_err   = fabsf(sample_point - target_sample_point);
        float total_err    = 0.6f * baud_err + 0.4f * sample_err;

        // 误差更小 或 误差相同但分频更大 或 分频相同但TQ更小
        if ((total_err < min_error) || (total_err == min_error && ((prescale > best.prescaler) || (prescale == best.prescaler && tq < (best.bs1 + best.bs2 + 1)))))
        {
            min_error         = total_err;
            best.prescaler    = prescale;
            best.bs1          = bs1;
            best.bs2          = bs2;
            best.baud_error   = baud_err;
            best.sample_point = sample_point;
        }
    }
    return best;
}

int main()
{
    CAN_TimingConfig cfg = calculate_can_timing(36000000, 500 * 1000);
    printf("Prescaler: %u\nBS1: %u\nBS2: %u\nSample: %.1f%%\nBaud error: %.4f%%\n",
           cfg.prescaler, cfg.bs1, cfg.bs2,
           cfg.sample_point * 100, cfg.baud_error * 100);
}

这个函数可以考虑放在dev_can.c中 给drv_can.c传参时只需要相应的分频值和两个bs的TQ数(默认SWJ 1TQ)

Describe your preferred solution

No response

Describe possible alternatives

No response

Start by reading the existing CAN baud-rate tables and the CAN configuration flow in dev_can.c and drv_can.c. Check how prescaler, BS1, and BS2 values are passed to the driver, then define how the proposed calculation should work across supported chips. Done means a validated approach replaces or supplements the duplicated predefined parameters without breaking existing CAN configurations.