【BUG】CAN发送时拔插can总线，导致后续发送显示失败

发布于 2022-10-10 19:01:07 浏览：1183 订阅该版

[tocm]

- 使用KEILV5 AC6编译器 优化等级-OZ
- RTT使用V4.1.0
- 使用CAN驱动
- 芯片为STM32F429IGT6

# 问题描述
这些测试都没有使用自动重发功能，can部分代码都是最新从RTT源码拉下来的。
### 先插上CAN总线，发送10次成功。
![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221012/12d47769a47ae649d9c53fd785c81a39.png)

### 断开CAN总线，发送10次失败

![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221012/91c75880ef4e7edecc43c800a58e4c50.png)

### 再次插上CAN总线，发送10次成功

![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221012/163595c785cf7b4011073b7a64fdf872.png)

### 再次断开CAN总线，发送10次失败

![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221012/955b28f82f532133a1ddef63847e7a5b.png)

### 再次插上CAN总线，发送10次 **失败**

- 可以看出进入到被动恢复了
![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221012/7a6044a2fc31d0224e9171ac9050d808.png)

- 接上分析仪可以看出，发送是成功的。但是代码认为发送失败了
![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221012/b83b8545eb509a0ed72eb609f0bf1dc9.png)
### 跟这个帖子一样，进入被动恢复后，没有解决这个问题。
> https://club.rt-thread.org/ask/question/d19dd32f398cc2a6.html

## 原来的测试

### 正常发送，没有拔插can总线

![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221010/48ec858a7927ea2acd535177b26e11b1.png)

### 拔出总线

![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221010/b7c331f03e84590dea6eafc8d55a741b.png)

### 重新插回总线,显示失败。实际上发送出来了

![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221010/19a6214240d860f65a1f2f76820ec808.png)

- 逻辑分析仪显示由ACK回复

![screenshot_image.png](https://oss-club.rt-thread.org/uploads/20221011/462d0bca7af047412e74e0619776c690.png.webp)
##代码如下
```c
/* USER CODE BEGIN Header */
/**
  ******************************************************************************
  * @file           : 
  * @brief          : 
  * @date           :
  ******************************************************************************
  * @attention
  * @author
  ******************************************************************************
  */
/* USER CODE END Header */
/* Includes ------------------------------------------------------------------*/

/* Private includes ----------------------------------------------------------*/
#include <rtthread.h>
#include "rtdevice.h"

#define CAN_DEV_NAME        "can1"      /* CAN 设备名称 */
/*ulog include*/
#define LOG_TAG             CAN_DEV_NAME
#define LOG_LVL              DBG_INFO
#include <ulog.h>
/* Private typedef -----------------------------------------------------------*/

/* Private define ------------------------------------------------------------*/

/* Private macro -------------------------------------------------------------*/
/* Private define ------------------------------------------------------------*/
/* 监控线程配置 */
#define THREAD_PRIORITY      12  //线程优先级
#define THREAD_TIMESLICE     10  //线程时间片
#define THREAD_STACK_SIZE    1024//栈大小
/* Private variables ---------------------------------------------------------*/

/* Private function prototypes -----------------------------------------------*/
static struct rt_semaphore rx_sem;     /* 用于接收消息的信号量 */
static rt_device_t can_dev;            /* CAN 设备句柄 */
/**
  * @brief  接收数据回调函数
  * @param  None
  * @retval None
  * @note   
*/
static rt_err_t can_rx_call(rt_device_t dev, rt_size_t size)
{
    /* CAN 接收到数据后产生中断，调用此回调函数，然后发送接收信号量 */
    rt_sem_release(&rx_sem);

return RT_EOK;
}
/**
  * @brief  can接收线程
  * @param  None
  * @retval None
  * @note   
*/
static void can_rx_thread(void *parameter)
{
    int i;
    rt_err_t res;
    struct rt_can_msg rxmsg = {0};

/* 设置接收回调函数 */
    rt_device_set_rx_indicate(can_dev, can_rx_call);

#ifdef RT_CAN_USING_HDR
    struct rt_can_filter_item items[5] =
    {
        RT_CAN_FILTER_ITEM_INIT(0x100, 0, 0, 0, 0x700, RT_NULL, RT_NULL), /* std,match ID:0x100~0x1ff，hdr 为 - 1，设置默认过滤表 */
        RT_CAN_FILTER_ITEM_INIT(0x300, 0, 0, 0, 0x700, RT_NULL, RT_NULL), /* std,match ID:0x300~0x3ff，hdr 为 - 1 */
        RT_CAN_FILTER_ITEM_INIT(0x211, 0, 0, 0, 0x7ff, RT_NULL, RT_NULL), /* std,match ID:0x211，hdr 为 - 1 */
        RT_CAN_FILTER_STD_INIT(0x486, RT_NULL, RT_NULL),                  /* std,match ID:0x486，hdr 为 - 1 */
        {0x555, 0, 0, 0, 0x7ff, 7,}                                       /* std,match ID:0x555，hdr 为 7，指定设置 7 号过滤表 */
    };
    struct rt_can_filter_config cfg = {5, 1, items}; /* 一共有 5 个过滤表 */
    /* 设置硬件过滤表 */
    res = rt_device_control(can_dev, RT_CAN_CMD_SET_FILTER, &cfg);
    RT_ASSERT(res == RT_EOK);
#endif

while (1)
    {
        /* hdr 值为 - 1，表示直接从 uselist 链表读取数据 */
        rxmsg.hdr = -1;
        /* 阻塞等待接收信号量 */
        rt_sem_take(&rx_sem, RT_WAITING_FOREVER);
        /* 从 CAN 读取一帧数据 */
        rt_device_read(can_dev, 0, &rxmsg, sizeof(rxmsg));
        /* 打印数据 ID 及内容 */
        rt_kprintf("\n");
        rt_kprintf("ID:%x:", rxmsg.id);
        for (i = 0; i < 8; i++)
        {
            rt_kprintf("%2x", rxmsg.data[i]);
        }

rt_kprintf("\n");
    }
}
/**
  * @brief  CAN初始化函数
  * @param  None
  * @retval None
  * @note   
*/
int can1_init(void)
{
    struct rt_can_msg msg = {0};
    rt_err_t res;
    rt_size_t  size;
    /* 查找 CAN 设备 */
    can_dev = rt_device_find(CAN_DEV_NAME);
    if (!can_dev)
    {
        LOG_E("find %s failed", CAN_DEV_NAME);
        return RT_ERROR;
    }

/* 初始化 CAN 接收信号量 */
    rt_sem_init(&rx_sem, "rx_sem", 0, RT_IPC_FLAG_FIFO);

/* 以中断接收及发送方式打开 CAN 设备 */
    res = rt_device_open(can_dev, RT_DEVICE_FLAG_INT_TX | RT_DEVICE_FLAG_INT_RX);
    RT_ASSERT(res == RT_EOK);
    /* 创建线程 放在最后 避免定时器还为初始化*/
    rt_thread_t thread = rt_thread_create( "can_rx",           /* 线程名字 */
                                           can_rx_thread,  /* 线程入口函数 */
                                           RT_NULL,               /* 线程入口函数参数 */
                                           THREAD_STACK_SIZE,     /* 线程栈大小 */
                                           THREAD_PRIORITY,       /* 线程的优先级 */
                                           THREAD_TIMESLICE);     /* 线程时间片 */
    /* 创建成功则启动线程 */
    if (thread != RT_NULL)
    {
        rt_thread_startup(thread);
    }
    else
    {
        LOG_E("Monitor thread created failed.");
        res =  RT_ERROR;
    }

msg.id = 0x78;              /* ID 为 0x78 */
    msg.ide = RT_CAN_STDID;     /* 标准格式 */
    msg.rtr = RT_CAN_DTR;       /* 数据帧 */
    msg.len = 8;                /* 数据长度为 8 */
    /* 待发送的 8 字节数据 */
    msg.data[0] = 0x00;
    msg.data[1] = 0x11;
    msg.data[2] = 0x22;
    msg.data[3] = 0x33;
    msg.data[4] = 0x44;
    msg.data[5] = 0x55;
    msg.data[6] = 0x66;
    msg.data[7] = 0x77;
    /* 发送一帧 CAN 数据 */
    size = rt_device_write(can_dev, 0, &msg, sizeof(msg));
    if (size == 0)
    {
        LOG_E("can dev write data failed!");
    }

return res;
}
INIT_APP_EXPORT(can1_init);
/**
  * @brief  CAN发送测试
  * @param  None
  * @retval None
  * @note   
*/
void can_send_test(void)
{
    struct rt_can_msg msg = {0};
    rt_size_t  size;
    static rt_uint8_t num = 0;
    /* 查找 CAN 设备 */
    can_dev = rt_device_find(CAN_DEV_NAME);
    if (!can_dev)
    {
        LOG_E("find %s failed", CAN_DEV_NAME);
    }
    msg.id = 0x78;              /* ID 为 0x78 */
    msg.ide = RT_CAN_STDID;     /* 标准格式 */
    msg.rtr = RT_CAN_DTR;       /* 数据帧 */
    msg.len = 8;                /* 数据长度为 8 */

/* 待发送的 8 字节数据 */
    msg.data[0] = 0x00;
    msg.data[1] = num++;
    msg.data[2] = 0x22;
    msg.data[3] = 0x33;
    msg.data[4] = num++;
    msg.data[5] = 0x55;
    msg.data[6] = 0x66;
    msg.data[7] = 0x77;
    /* 发送一帧 CAN 数据 */
    size = rt_device_write(can_dev, 0, &msg, sizeof(msg));
    if (size == 0)
    {
        rt_kprintf("can dev write data failed!\n");
    }
}
/* 导出到 msh 命令列表中 */
MSH_CMD_EXPORT(can_send_test, can send test);
```
# 解决办法
- 查找到这个帖子 https://club.rt-thread.org/ask/article/5cedb728813e6fd8.html
- 发现主要问题在最新的RTT中已经解决了。
## 但是断开总线再次插入发送失败的问题没有解决
- 查看帖子下面的评论
@mhxsoft 的回复
> 还发现一个地方，drv_can.c里，当CAN总线拔掉发送时，会进入ACK错误中断SCE，这时TSR的TXOK都是0，另CAN_FLAG_TXOK0/1/2值的定义，还要移位才能对应，READ_BIT没有进行移位，所以下面这三个条件并不会执行，改为下面方式就可以执行了，并且清理了发送请求位CAN_TSR_RQCP0/1/2

```c
CAN1_SCE_IRQHandler(void)
{
    ......
    case RT_CAN_BUS_ACK_ERR:/* attention !!! test ack err's unit is transmit unit */
        drv_can1.device.status.ackerrcnt++;
        if (!READ_BIT(drv_can1.CanHandle.Instance->TSR, CAN_FLAG_TXOK0))
            rt_hw_can_isr(&drv_can1.device, RT_CAN_EVENT_TX_FAIL | 0 << 8);
        else if (!READ_BIT(drv_can1.CanHandle.Instance->TSR, CAN_FLAG_TXOK1))
            rt_hw_can_isr(&drv_can1.device, RT_CAN_EVENT_TX_FAIL | 1 << 8);
        else if (!READ_BIT(drv_can1.CanHandle.Instance->TSR, CAN_FLAG_TXOK2))
            rt_hw_can_isr(&drv_can1.device, RT_CAN_EVENT_TX_FAIL | 2 << 8);
        break;
    ......
}
```

———————-改为以下代码———————
```c

CAN1_SCE_IRQHandler(void)
{
    ......
    case RT_CAN_BUS_ACK_ERR:/* attention !!! test ack err's unit is transmit unit */
        drv_can1.device.status.ackerrcnt++;
        if (__HAL_CAN_GET_FLAG(hcan, CAN_FLAG_RQCP0))
        {
            if (!__HAL_CAN_GET_FLAG(hcan, CAN_FLAG_TXOK0))
            {
                rt_hw_can_isr(&drv_can1.device, RT_CAN_EVENT_TX_FAIL | 0 << 8);
            }
            SET_BIT(hcan->Instance->TSR, CAN_TSR_RQCP0);
        }
        else if (__HAL_CAN_GET_FLAG(hcan, CAN_FLAG_RQCP1))
        {
            if (!__HAL_CAN_GET_FLAG(hcan, CAN_FLAG_TXOK1))
            {
                rt_hw_can_isr(&drv_can1.device, RT_CAN_EVENT_TX_FAIL | 1 << 8);
            }
            SET_BIT(hcan->Instance->TSR, CAN_TSR_RQCP1);
        }
        else if (__HAL_CAN_GET_FLAG(hcan, CAN_FLAG_RQCP2))
        {
            if (!__HAL_CAN_GET_FLAG(hcan, CAN_FLAG_TXOK2))
            {
                rt_hw_can_isr(&drv_can1.device, RT_CAN_EVENT_TX_FAIL | 2 << 8);
            }
            SET_BIT(hcan->Instance->TSR, CAN_TSR_RQCP2);
        }
        break;
    ......
}
```

# 修改结果
- **修改后，确实不会出现该情况，在没有开启自动重传的情况下**
- **修改后，不能开启自动重传，开启自动重传后，不能有发送失败。
**如果造成发送失败，会阻塞当前发送线程。只要发送成功**
**我用MSH命令发送，就把SHELL线程卡死了。**

- 希望这个问题能够修复。由于不是我写的代码，不好意思pr。
- 另外一个就是关于can驱动部分的代码确实没看懂，所以自己并没有找到解决办法。还得搜索。