Better way to avoid race condition, only notification lives in CAN task

- Runs only once
- Only critical notification is inside task
- CAN message during boot will not be lost, instead put inside FIFO0

Core/Src/canlog.c

@@ -47,14 +47,11 @@ void CAN_Logger_Init(CAN_HandleTypeDef *hcan1, CAN_HandleTypeDef *hcan2)
 
     filter.FilterBank = 0;
     if (HAL_CAN_ConfigFilter(hcan1, &filter) != HAL_OK) Error_Handler();
+    if (HAL_CAN_Start(hcan1) != HAL_OK) Error_Handler();
+
     filter.FilterBank = 14;
     if (HAL_CAN_ConfigFilter(hcan2, &filter) != HAL_OK) Error_Handler();
-
-    if (HAL_CAN_Start(hcan1) != HAL_OK) Error_Handler();
     if (HAL_CAN_Start(hcan2) != HAL_OK) Error_Handler();
-
-    if (HAL_CAN_ActivateNotification(hcan1, CAN_IT_RX_FIFO0_MSG_PENDING) != HAL_OK) Error_Handler();
-    if (HAL_CAN_ActivateNotification(hcan2, CAN_IT_RX_FIFO0_MSG_PENDING) != HAL_OK) Error_Handler();
 }
 /* END CAN_Logger_Init */
 
@@ -102,12 +99,10 @@ void vCANLoggerListen(void *argument)
 {
   /* CODE BEGIN */
   CAN_HandleTypeDef *hcan = (CAN_HandleTypeDef*)argument;
-  osMessageQueueId_t queue;
+  osMessageQueueId_t queue = (hcan->Instance == CAN1) ? xCAN1RxQueue : xCAN2RxQueue;
   CanMessage_t message;
 
-  CAN_Logger_Init(&hcan1, &hcan2);
+  HAL_CAN_ActivateNotification(hcan, CAN_IT_RX_FIFO0_MSG_PENDING);
-
-  queue = (hcan->Instance == CAN1) ? xCAN1RxQueue : xCAN2RxQueue;
 
   for (;;)
   {

Core/Src/main.c

@@ -110,6 +110,7 @@ int main(void)
   MX_SDIO_SD_Init();
 
   /* USER CODE BEGIN 2 */
+  CAN_Logger_Init(&hcan1, &hcan2);
   /* USER CODE END 2 */
 
   /* Init scheduler */
@@ -184,7 +185,6 @@ int main(void)
   /* USER CODE END RTOS_EVENTS */
 
   /* USER CODE BEGIN 3 */
-  CAN_Logger_Init(&hcan1, &hcan2);
   /* USER CODE END 3 */
 
   /* Start scheduler */