普冉PY32系列(二) Ubuntu GCC Toolchain和VSCode开发环境

戚英逸

2023-12-01

项目模板

GitHub 仓库地址: https://github.com/IOsetting/py32f0-template

仓库文件结构

├── Build                       # 编译结果
├── Docs                        # 数据手册和用户手册
├── Examples
│   ├── FreeRTOS                # FreeRTOS 例子
│   ├── HAL                     # 非 FreeRTOS 的例子, 基于HAL外设库
│   └── LL                      # 非 FreeRTOS 的例子, 基于LL外设库
├── Libraries
│   ├── BSP                     # delay 和 printf 的 BSP 库, 用于 HAL外设库
│   ├── BSP_LL                  # delay 和 printf 的 BSP 库, 用于 LL外设库
│   ├── CMSIS
│   ├── FreeRTOS                # FreeRTOS library
│   ├── LDScripts               # 编译连接描述文件
│   ├── PY32F0xx_HAL_Driver     # HAL 外设驱动库
│   └── PY32F0xx_LL_Driver      # LL(low layer)外设驱动库
├── Makefile                    # Make设置
├── Misc
│   ├── Flash
│   │   ├── JLinkDevices        # JLink Flash 算法文件
│   │   └── Sources             # Flash 算法文件源代码
│   ├── Puya.PY32F0xx_DFP.x.pack # DFP pack文件, 用于 PyOCD 烧录
│   └── SVD                     # SVD 文件, 用于Debug
├── README.md
├── rules.mk                    # 预置的 make 规则
└── User                        # 用户项目代码目录

环境准备

硬件方面

PY32F0 开发板, 或任何基于 PY32F002/003/030 系列的电路
烧录工具(任一)
- J-Link: J-Link OB programmer
- PyOCD: DAPLink or J-Link

注: STLink测试不能直接用, 写入会有Timeout错误

软件方面

SEGGER J-Link 软件 https://www.segger.com/downloads/jlink/
或者 PyOCD https://pyocd.io/
GNU Arm Embedded Toolchain

环境配置和编译过程

1. 安装 GNU Arm Embedded Toolchain

根据你的PC架构, 从 Arm GNU Toolchain Downloads 下载工具链, 然后解压文件到合适的目录下, 例如

sudo mkdir -p /opt/gcc-arm/
sudo tar xvf arm-gnu-toolchain-12.2.rel1-x86_64-arm-none-eabi.tar.xz -C /opt/gcc-arm/
cd /opt/gcc-arm/
sudo chown -R root:root arm-gnu-toolchain-12.2.rel1-x86_64-arm-none-eabi/

2. (选项一) 安装 SEGGER J-Link

从 J-Link / J-Trace Downloads 下载并安装 SEGGER JLink

# installation command for .deb
sudo dpkg -i JLink_Linux_V784f_x86_64.deb
# uncompression command for .tar.gz
sudo tar xvf JLink_Linux_V784f_x86_64.tgz -C [target folder]

默认的安装路径是 /opt/SEGGER

将目录 [项目目录]/Misc/Flash/JLinkDevices 复制到 [User home]/.config/SEGGER/JLinkDevices/ 目录下, 对于最新版本 V7.67之后的 JLink, 不再需要单独编辑 JLinkDevices.xml 了

cd py32f0-template
cp -r Misc/Flash/JLinkDevices/ ~/.config/SEGGER/

关于如果给JLink配置新型号的芯片: https://wiki.segger.com/J-Link_Device_Support_Kit

2. (选项二): 安装 PyOCD

不要使用Ubuntu自带的apt仓库里的PyOCD, 这个版本 0.13.1+dfsg-1 太低, 无法识别 JLink OB

从Pip安装PyOCD

pip uninstall pyocd

这样会将 PyOCD 安装到这些目录

/home/[user]/.local/bin/pyocd
/home/[user]/.local/bin/pyocd-gdbserver
/home/[user]/.local/lib/python3.10/site-packages/pyocd-0.34.2.dist-info/*
/home/[user]/.local/lib/python3.10/site-packages/pyocd/*

在 Ubuntu 中, .profile 会自动把 .local 加入 PATH, 所以只需要执行一下 source ~/.profile 就能用 pyocd 命令了

3. 导出这个模板仓库

Clone到本地目录下

git clone https://github.com/IOsetting/py32f0-template.git

4. 修改 Makefile

根据自己本地环境, 修改Makefile

USE_LL_LIB 普冉的开发包中提供了两套外设驱动库, HAL 和 LL, 设置 USE_LL_LIB ?= y 就会切换到LL.
ENABLE_PRINTF_FLOAT 设置为 y 会在编译中添加 -u _printf_float 参数, 这样可以支持 printf 打印浮点数, 但是这也会增加最后固件的大小.
USE_FREERTOS 设置USE_FREERTOS ?= y用于在项目中包含 FreeRTOS
USE_DSP 在项目中包含 CMSIS DSP
FLASH_PROGRM
- 如果用 J-Link, FLASH_PROGRM 可以设为 jlink 或 pyocd
- 如果用的是 DAPLink, 设为 FLASH_PROGRM ?= pyocd
- ST-LINK 暂时还不支持, 等能解决问题的大佬出手
ARM_TOOCHAIN 确认这个指向了正确的 arm-none-eabi-gcc 所在路径

##### Project #####

PROJECT           ?= app
# The path for generated files
BUILD_DIR         = Build


##### Options #####

# Use LL library instead of HAL, y:yes, n:no
USE_LL_LIB        ?= n
# Enable printf float %f support, y:yes, n:no
ENABLE_PRINTF_FLOAT ?= n
# Build with FreeRTOS, y:yes, n:no
USE_FREERTOS      ?= n
# Build with CMSIS DSP functions, y:yes, n:no
USE_DSP           ?= n
# Programmer, jlink or pyocd
FLASH_PROGRM      ?= pyocd

##### Toolchains #######
ARM_TOOCHAIN      ?= /opt/gcc-arm/arm-gnu-toolchain-12.2.rel1-x86_64-arm-none-eabi/bin

# path to JLinkExe
JLINKEXE          ?= /opt/SEGGER/JLink/JLinkExe
# JLink device type, options:
#   PY32F002AX5, PY32F002X5, 
#   PY32F003X4, PY32F003X6, PY32F003X8, 
#   PY32F030X4, PY32F030X6, PY32F030X7, PY32F030X8
JLINK_DEVICE      ?= PY32F030X6
# path to PyOCD, 
PYOCD_EXE         ?= pyocd
# PyOCD device type, options: 
#     py32f002ax5, py32f002x5, 
#   py32f003x4,  py32f003x6, py32f003x8, 
#   py32f030x3,  py32f030x4, py32f030x6, py32f030x7, py32f030x8
#   py32f072xb
PYOCD_DEVICE      ?= py32f030x8


##### Paths ############

# Link descript file: py32f002x5.ld, py32f003x6.ld, py32f003x8.ld, py32f030x6.ld, py32f030x8.ld
LDSCRIPT          = Libraries/LDScripts/py32f030x8.ld
# Library build flags: 
#   PY32F002x5, PY32F002Ax5, 
#   PY32F003x4, PY32F003x6, PY32F003x8, 
#   PY32F030x3, PY32F030x4, PY32F030x6, PY32F030x7, PY32F030x8, 
#   PY32F072xB
LIB_FLAGS         = PY32F030x6

5. 编译和烧录

编译执行

# clean source code
make clean
# build
make
# or make with verbose output
V=1 make

写入, 会根据前面的配置调用对应的烧录方法

# flash
make flash

这个例子会展示

驱动 PB5 上的LED灯亮灭, 如果你使用的型号不带 PB5 pin, 需要修改一下 main.c, 换成其它的pin.
PA2, PA3 通过UART输出echo字符串, 需要接在 USB2TTL 上通过PC端串口软件查看, Ubuntu下可以用 Screen 或 CuteCOM 等软件

尝试其它例子

在 Examples 目录下有更多的代码示例, 可以复制替换掉 User 目录下的文件, 然后编译烧录查看运行效果

配置 VSCode 开发环境

安装好 VSCode 的 c/cpp 扩展后, 直接打开项目所在目录就好了, 代码高亮, 仓库方法都是可用的, 需要的额外配置一个是c/cpp配置, 另一个是task

C/C++配置

Ctrl+Shift+P调出快捷菜单, 在里面选择 C/C++ Edit Configurations(JSON), 会创建对应的初始化配置, 然后根据自己的环境修改, 以下是我用的环境. 如果换了不同型号的芯片, defines 里面的参数也要跟着换

{
    "configurations": [
        {
            "name": "Linux",
            "includePath": [
                "${workspaceFolder}/**",
                "${workspaceFolder}/User/**",
                "${workspaceFolder}/Libraries/CMSIS/Include",
                "/opt/gcc-arm/arm-gnu-toolchain-12.2.mpacbti-bet1-x86_64-arm-none-eabi/arm-none-eabi/include",
                "/opt/gcc-arm/arm-gnu-toolchain-12.2.mpacbti-bet1-x86_64-arm-none-eabi/lib/gcc/arm-none-eabi/12.2.0/include"
            ],
            "defines": [
                "PY32F002Ax5"
            ],
            "compilerPath": "/opt/gcc-arm/arm-gnu-toolchain-12.2.mpacbti-bet1-x86_64-arm-none-eabi/bin/arm-none-eabi-gcc",
            "cStandard": "gnu99",
            "cppStandard": "gnu++14",
            "intelliSenseMode": "gcc-arm",
            "configurationProvider": "ms-vscode.makefile-tools"
        }
    ],
    "version": 4
}

Task 配置

同样唤出快捷菜单后, 在Task中创建默认模板进行配置, 我使用的tasks.json内容为

{
    // See https://go.microsoft.com/fwlink/?LinkId=733558
    // for the documentation about the tasks.json format
    "version": "2.0.0",
    "tasks": [
        {
            "label": "clean, build",
            "type": "shell",
            "command": "make clean;make",
            "problemMatcher": []
        },
        {
            "label": "build, download",
            "type": "shell",
            "command": "make;make flash",
            "problemMatcher": []
        },
        {
            "label": "download",
            "type": "shell",
            "command": "make flash",
            "problemMatcher": []
        },
        {
            "label": "build",
            "type": "shell",
            "command": "make",
            "problemMatcher": []
        },
        {
            "label": "clean",
            "type": "shell",
            "command": "make clean",
            "problemMatcher": []
        }
    ]
}

用 Shift + Alt + F10 调出 Task 菜单, 选择对应的条目进行编译或烧录

Debug 配置

在VSCode中安装 Cortex Debug 扩展后, 在 launch.json 中新增一个运行配置, 例如

{
    "armToolchainPath": "/opt/gcc-arm/arm-gnu-toolchain-12.2.rel1-x86_64-arm-none-eabi/bin/",
    "toolchainPrefix": "arm-none-eabi",
    "name": "Cortex Debug",
    "cwd": "${workspaceFolder}",
    "executable": "${workspaceFolder}/Build/app.elf",
    "request": "launch",        // can be launch or attach
    "type": "cortex-debug",
    "runToEntryPoint": "Reset_Handler", // "main" or other function name. runToMain is deprecated
    "servertype": "jlink",  // jlink, openocd, pyocd, pe and stutil
    "device": "PY32F030X8",
    "interface": "swd",
    "preLaunchTask": "build",  // Set this to run a task from tasks.json before starting a debug session
    // "preLaunchCommands": ["Build all"], // Uncomment this if not using preLaunchTask
    "svdFile": "${workspaceFolder}/Misc/SVD/py32f030xx.svd",  // svd for this part number
    "showDevDebugOutput": "vscode", // parsed, raw, vscode:vscode log and raw
    "swoConfig":
    {
        "enabled": true,
        "cpuFrequency": 8000000, // Target CPU frequency in Hz
        "swoFrequency":  4000000,
        "source": "probe", // either be “probe” to get directly from the debug probe, 
                           // or a serial port device to use a serial port external to the debug probe.
        "decoders":
        [
            {
                "label": "ITM port 0 output",
                "type": "console",
                "port": 0,
                "showOnStartup": true,
                "encoding": "ascii"
            }
        ]
    }
}

如果 Cortex Debug 无法找到 JLinkGDBServerCLExe, 可以在 settings.json 中增加以下的配置

"cortex-debug.JLinkGDBServerPath": "/opt/SEGGER/JLink/JLinkGDBServerCLExe",