STM32 Code Generation

LibXR provides the xr_cubemx_cfg command to automatically generate C++ initialization code from an STM32CubeMX project. This command wraps several tools including configuration parsing, code generation, and CMake integration.

---

Quick Start

In the root directory of your STM32CubeMX project, run:

xr_cubemx_cfg -d .

This command will perform the following steps automatically:

1. Initialize or update the libxr submodule
2. Locate the .ioc file and convert it into .config.yaml
3. Generate app_main.cpp with initialization code
4. Modify CMakeLists.txt to integrate LibXR

---

Example Output

[INFO] LibXR submodule already exists. Checking for updates...
[INFO] [OK] cd . && git submodule update --init --recursive
[INFO] LibXR submodule updated.
Found .ioc file: ./atom.ioc
Parsing .ioc file...
[INFO] [OK] xr_parse_ioc -d . -o ./.config.yaml
Generating C++ code...
[INFO] [OK] xr_gen_code_stm32 -i ./.config.yaml -o ./User/app_main.cpp
[INFO] [OK] xr_stm32_cmake .
[INFO] [Pass] All tasks completed successfully!

---

Output Structure

After execution, your project directory will contain the following generated or modified files:

.
├── .config.yaml                      # Parsed configuration from .ioc
├── User/
│   ├── app_main.cpp                  # Main initialization code
│   ├── app_main.h                    # Header for app_main
│   ├── libxr_config.yaml             # LibXR runtime configuration
│   └── flash_map.hpp                 # Flash address mapping table
├── cmake/LibXR.CMake                 # CMake build config for LibXR
├── CMakeLists.txt                    # Modified to include LibXR
└── Middlewares/Third_Party/LibXR     # LibXR as a Git submodule

---

app_main.cpp

The app_main.cpp file contains the LibXR initialization function app_main().

You can safely insert your own code between the User Code Begin xxx and User Code End xxx sections; these parts will not be overwritten when the code is regenerated.

Note:
This function should never return. If it does, all peripheral objects (such as usart1) will be destructed and resources released, which will cause a crash if you try to access them afterwards.

It is recommended to pass base class pointers of peripheral objects to your threads or tasks and operate on them there.
A better way to implement this will be introduced in the section "Integrate with XRobot" later in this chapter.

#include "app_main.h"

#include "libxr.hpp"
#include "main.h"
#include "stm32_adc.hpp"
#include "stm32_can.hpp"
#include "stm32_dac.hpp"
#include "stm32_gpio.hpp"
#include "stm32_i2c.hpp"
......

using namespace LibXR;

/* User Code Begin 1 */

/* User Code End 1 */
/* External HAL Declarations */
extern ADC_HandleTypeDef hadc1;
extern CAN_HandleTypeDef hcan1;
extern I2C_HandleTypeDef hi2c1;
extern SPI_HandleTypeDef hspi1;
extern TIM_HandleTypeDef htim1;
......

/* DMA Resources */
static uint16_t adc1_buf[64];
static uint8_t spi1_tx_buf[32];
static uint8_t spi1_rx_buf[32];
static uint8_t usart1_tx_buf[128];
static uint8_t usart1_rx_buf[128];
static uint8_t i2c1_buf[32];
......

extern "C" void app_main(void) {
  /* User Code Begin 2 */
  
  /* User Code End 2 */
  STM32TimerTimebase timebase(&htim2);
  PlatformInit(2, 2048);
  STM32PowerManager power_manager;

  /* GPIO Configuration */
  STM32GPIO USER_KEY(USER_KEY_GPIO_Port, USER_KEY_Pin, EXTI0_IRQn);
  STM32GPIO LED_B(LED_B_GPIO_Port, LED_B_Pin);
  STM32PWM pwm_tim1_ch1(&htim1, TIM_CHANNEL_1, false);
  STM32SPI spi1(&hspi1, spi1_rx_buf, spi1_tx_buf, 3);
  STM32UART usart1(&huart1,
              usart1_rx_buf, usart1_tx_buf, 5);
  STM32I2C i2c1(&hi2c1, i2c1_buf, 3);
  STM32CAN can1(&hcan1, 5);
  /* User Code Begin 3 */
  while (1) {
      LibXR::Thread::Sleep(1000);
  }
  /* User Code End 3 */
}

---

How to Use

Call app_main() in the entry function (StartDefaultTask) of the main thread.

Bare-metal Project

#include "app_main.h"

int main() {
    HAL_Init();
    SystemClock_Config();
    ...
    app_main();  // Initialize peripherals and start application
    while (1) {
        ...
    }
}

FreeRTOS Project

Call app_main() in the main thread entry function (such as StartDefaultTask).
Make sure to adjust the initial thread stack size in STM32CubeMX to avoid stack overflow.

---

Optional Arguments

Argument	Description
-d	Specify STM32 project root directory
-t	Set terminal peripheral (e.g. usart1)
--xrobot	Generate glue code for XRobot modules

---

Toolchain Switch

If you need to switch between GCC/Clang compilers or change the Clang standard library, use the following commands:

xr_stm32_toolchain_switch gcc
xr_stm32_toolchain_switch clang -g
xr_stm32_toolchain_switch clang --newlib
xr_stm32_toolchain_switch clang --picolibc

Command execution will automatically modify CMakePresets.json and cmake/starm-clang.cmake, restart VSCode to take effect.

---

Project Requirements

• Must be a CMake project exported from STM32CubeMX
• Must contain a valid .ioc file
• Must enable Mutex when using FreeRTOS(configUSE_MUTEXES)

---

Build Optimization

By default, the generated CMake configuration applies the -O2 optimization option to libraries such as LibXR, HAL and FreeRTOS in Debug mode, while using the -Og optimization option for code in the User directory and XRobot modules. This approach helps reduce FLASH usage.

---

Subcommands (Internally used by xr_cubemx_cfg, can also be run separately)

Tool	Description
xr_parse_ioc	Parses .ioc and generates .config.yaml
xr_gen_code_stm32	Generates app_main.cpp from the YAML config
xr_stm32_cmake	Integrates LibXR into the project build system
xr_stm32_toolchain_switch	Switch toolchain and standard library

---

References

LibXR STM32 Code Generation Tool Test Project (Github Action)

LibXR CLI tool and documentation (PyPI)