Problems with CAN Libraries

RE: Problems with CAN Libraries

Sigurd — Thu, 19 Oct 2023 11:27:45 GMT

Hi!

We are not familiar with this MCP2515 chip from Microchip. Maybe someone in the Zephyr discord can help out. https://docs.zephyrproject.org/latest/develop/getting_started/index.html#asking-for-help

RE: Problems with CAN Libraries

babos — Mon, 02 Oct 2023 07:57:03 GMT

Hi,

I discovered that the problem arises because at a certain point, I no longer receive data from the ISR queue of the MCP2515 chip (I still don't know the real reason). However, I would like to restart the 2515 chip through software, stopping all the specific threads and reconfiguring it to temporarily circumvent the problem. I have tried to stop the threads, remove the filters, stop the CAN, and restart it using API commands, but it does not reboot.I'm currently using NVIC_SystemReset(); for error handling, which restarts the whole system and gets it back up and running. However, this causes a Bluetooth connection reset, which takes too long to reconnect. The alternatives would be to either find the root issue or only restart the CAN chip or perform a warm reboot without disconnecting the Bluetooth (if that's even possible). I haven't been able to do it yet. Is there anyone who can support me?
I'm attaching the codes that I have tested:

/// @brief re-initialize can devices and start relative thread(s)
/// @param
void can_reinit()
{
    /* Do NOT WORK */
    can_stop(can_dev_0);
    k_thread_abort(get_state_tid);
    k_thread_abort(rx_tid);
    can_api_initialized = false;
    can_api_init();

    /* al posto di 'NVIC_SystemReset()' reinizializzare il solo chip CAN eventualmente bloccato (TODO) */
    // NVIC_SystemReset(); // reset nrf52840
    LOG_INF("CAN API re-initialized");
}

Attached is the '.c' file

thank you

Sdevzone.nordicsemi.com/.../can_5F00_trouble.c

RE: Problems with CAN Libraries

babos — Wed, 27 Sep 2023 14:07:12 GMT

in fact I have two called within ISR; one during the CAN interrupt, where I should pass what is highlighted below in the code:

K_SEM_DEFINE(uart_sem, 1, 1);
/// @brief
/// @param work
void can_refresh_send(struct k_work *work)
{
    if (uart_dev == NULL || k_sem_take(&uart_sem, K_NO_WAIT) != 0)
    {
        return;
    }

    /* atomic_get(&amsgCAN_array[3]); */
    sprintf(msg, "events cd %04X%02X%04X%02X%04X%04X%04X\r\n",
            atomic_get(&amsgCAN[8]) + atomic_get(&amsgCAN[9]) * 256, // consumption lt/h
            atomic_get(&amsgCAN[7]),                                 // motor load percentage
            atomic_get(&amsgCAN[5]) + atomic_get(&amsgCAN[6]) * 256, // battery voltage
            atomic_get(&amsgCAN[4]),                                 // refrigerant temperature
            atomic_get(&amsgCAN[2]) + atomic_get(&amsgCAN[3]) * 256, // speed Km/h
            atomic_get(&amsgCAN[0]) + atomic_get(&amsgCAN[1]) * 256, // rpm
            atomic_get(&amsgCAN[13]) + atomic_get(&amsgCAN[14]) * 256);

    LOG_INF("cd event");
    // uart_tx(uart_dev, msg, strlen(msg), 10);
    int len = strlen(msg);
    for (int i = 0; i < len; i++)
    {
        uart_poll_out(uart_dev, msg[i]);
    }
    k_sem_give(&uart_sem);
}

and the one already posted above. It is not necessary to send them during the callback of either one or the other, in fact I would put the messages in a queue and process them in a separate thread. However, I don't really understand how I can use the uart queues.

Thx

Babos

RE: Problems with CAN Libraries

Hakon — Wed, 27 Sep 2023 10:54:56 GMT

[quote user="babos"]I need to send a string to UART1 during a UART0 callback, as shown in the attached code, and another one within the CAN chip's interrupt.[/quote]

Do you need to send the string during the callback, or can you do it after? That way you won't trigger the assertion.

RE: Problems with CAN Libraries

babos — Mon, 25 Sep 2023 05:41:15 GMT

/// @brief
/// @param dev
/// @param evt
/// @param user_data
static void uart_callback(const struct device *dev, struct uart_event *evt,
						  void *user_data)
{
	int dev_idx = (int)user_data;
	struct uart_data_event *event;
	struct uart_rx_buf *buf;
	int err;

	switch (evt->type)
	{
	case UART_RX_RDY:
		uart_rx_buf_ref(evt->data.rx.buf);
		event = new_uart_data_event();
		event->dev_idx = dev_idx;
		event->buf = &evt->data.rx.buf[evt->data.rx.offset];
		event->len = evt->data.rx.len;
		/* insert decode CAN data */
		char *compare = NULL;
		size_t length = event->len;

		compare = malloc(length + 1); // Alloca memoria per la variabile 'compare'
		if (compare != NULL)
		{
			strncpy(compare, event->buf, length);
			compare[length] = '\0'; // Aggiungi il terminatore di stringa

			if (strstr((const char *)compare, "[M1]") && dev_idx == 1)
			{
				uint16_t numero;
				sscanf(compare, "[%*2s]%hu", &numero);
				rpm_memo = numero;// no offset / 0.125;
				flag_memo_rpm = true;
				//err = uart_tx(devices[0], compare, length, 0); // opzionale
			}
			else if (strstr((const char *)compare, "[M0]") && dev_idx == 1)
			{
				flag_memo_rpm = false;
				rpm_memo = 0;
				//err = uart_tx(devices[0], compare, length, 0); // opzionale
			}
			free(compare); // Libera la memoria allocata per 'compare'
		}
		else
		{
			// Gestisci l'errore di allocazione della memoria
		}
		APP_EVENT_SUBMIT(event);
		break;

Hello guys, the previous issue lies in the fact that I'm using serial port transmissions incorrectly within ISR, and this is causing my program to lock up. I need to send a string to UART1 during a UART0 callback, as shown in the attached code, and another one within the CAN chip's interrupt. This is why I'm experiencing abnormal behavior. How can I work around this issue? Thank you.

BAbos