Yocto 项目（可生成K1可用的SDCard 镜像）

K1
1

链接：
github仓库

进迭时空K1 Yocto SDK

本项目为 SpacemiT K1 平台提供一套 Yocto 层与构建脚本（通过 meta-k1/setup.sh 初始化），用于生成可刷写到 SDCard 的镜像。

功能简介

minimal镜像

极简 Linux，适用于快速启动、串口调试与基本网络测试。

minimal
minimal1920×1231 221 KB

weston镜像

带 Weston（Wayland）合成器的图形环境，支持在 K1 上运行简单图形应用（可使用K1上的GPU进行加速）。

weston
weston1920×1080 182 KB

固件下载与刷机

  • 下载：

  • 刷入示例（写入 SD 卡，操作前请确认目标设备）：

    sudo dd if=core-image-minimal.wic of=/dev/sdX bs=4M conv=fsync status=progress
sudo sync

  • 固件 root 用户密码：空（输入密码时直接按下回车）。

从本项目编译固件

运行环境与版本说明

  • 建议：Ubuntu 22.04/24.04、WSL2 或等效 Linux 环境。
  • 必要工具：git、bash、python3 及 Yocto 宿主依赖（参考 Yocto 官方文档）。

步骤

git clone https://github.com/yingjie-liu-spacemit/spacemit-yocto.git/'
cd spacemit-yocto
git submodule update --init --recursive
source meta-k1/setup.sh
bitbake core-image-minimal # 或 bitbake core-image-weston

构建产物

  • 存放在build/tmp/deploy/images/k1/（包含 .wic、bootfs等）。
1 个赞
2

This is all great!

I just have a question: Will this program work with Bianbu Star 2.1.7 and Bianbu 2.2.1?

3

Thank you for your interest in our project. Regarding your inquiry about running the Yocto build system natively on RISC-V, we would like to highlight the following key challenges:

  1. Tooling Incompatibility: Most Yocto “native” tools are optimized for x86_64. Native builds on RISC-V often face missing patches and toolchain errors that require significant manual effort to fix.
  2. Performance Constraints: Build processes are resource-heavy. RISC-V hardware currently lacks the processing power and I/O speed of x86 servers, resulting in extremely long build times (up to 10x–50x slower).
  3. Limited Support: The industry standard is cross-compilation.

Currently, we have not attempted a native build on RISC-V ourselves. We suggest maintaining the cross-compilation model for stability. However, we are open to further discussion and would be happy to share insights if either side makes progress on this front in the future.

1 个赞
4

Thank you for your detailed and comprehensive response!
I hope that eventually this project will be able to run directly on the RISCV architecture.