Tech-Roy Blog

Overview

Over the years, I’ve accumulated hundreds of bookmarks. These are the ones I actually keep coming back to — stripped of the dead links, outdated tools, and fluff. Consider this my own backup, and hopefully a useful reference for fellow engineers.

Bookmarks

🌐 Network Diagnostics

🌐 Network Speed Test

🔧 Development Tools

🛠️ Online Dev Tools

📝 Tech Blogs

📝 Tech Articles

📦 GitHub Repos

📐 Algorithms & Data Structures

🔒 Security & Threat Assessment

📚 Reference

🗺️ Fun Exploration

📜 History & Culture

🛠️ Practical Tools

🛠️ Life Skills

🎬 Media

💬 Community & Forums

📊 Data & Finance

🎓 Education

🎮 Gaming

🍉 Internet Trivia

最近在V2EX论坛，有V友问我Superalink怎么用，干脆整理成一篇攻略，方便回顾。
Ref: https://www.v2ex.com/t/1210180

套餐情况

我目前用的是大陆地区 5 日套餐，参数如下：

• 有效期：5天
• 流量：每日5GB（其中热点分享上限2GB）
• 原价：200泰铢（约42元人民币）
• 折扣价：25泰铢（约5.26元人民币）——需要折扣码，结算时选泰铢货币

  折扣码：ROY000000
  网址： https://www.superalink.com/destination/aff/ROY000000

算下来每个月买6次，月资费约31.5元，可以用150GB境外流量（覆盖HK+SG），日常看视频没问题。

使用前提：你的设备得支持 eSIM

Superalink是eSIM流量卡，购买后会给你一个QR二维码或激活码(LPA 开头的字符串)，用手机导入即可。

两种设备情况：

1. 手机原生支持eSIM(iPhone XS及以后的国际版、部分安卓旗舰)
   1. 直接扫码激活，没有任何门槛
2. 手机不支持原生eSIM(比如我用的国行iPhone 13 PRO)
   1. 可以买一张实体eSIM转接卡，比如ESTK.ME的卡，原理是把eSIM profile写进实体卡里，插到手机用。市面上还有”小白卡”也是类似方案，贴牌的多。只要能扫码或者输入LPA激活码就行。
   2. ESTK官网：https://store.estk.me/
   3. 我的邀请码：UPN3ALUL53
   4. 填写邀请码可以抵扣10%OFF

一个必须注意的设置

第一次用大概率会遇到：有信号、没网络。
原因：数据漫游默认是关的。
去「设置 → 蜂窝网络 → 对应eSIM → 数据漫游」，打开就好了。

使用限制说明

• 这张卡只能在大陆境内用，买的是”大陆地区”套餐，出了大陆就没流量了。
• 流量走的是移动基站，漫游到海外流量运营商的节点，所以能绕过GFW ，访问Google、YouTube之类的没问题。
• 每次购买需要用新邮箱，老邮箱折扣码会失效——用Gmail的 + 别名（yourname+001@gmail.com）或者Cloudflare catch-all转发都能搞定。

操作节奏

每5天买一次，熟了以后整个流程3分钟以内。唯一麻烦的就是要换邮箱，不过解决了邮箱问题之后基本就是重复操作了。

我的购买步骤

输入网址

Superalink官网：https://www.superalink.com/cn/destination/aff/ROY000000

点击[领取折扣]，领取后才能用优惠价格。

选择使用流量地区

选择5天套餐，折扣后显示25泰铢

填写邮箱

PayPal支付

Superalink目前已经不支持大陆信用卡直接付款，所以我用PayPal支付做了一层中转，实际绑定的还是大陆信用卡。

支付成功

去自己邮箱检查QR Code

用ESTK App扫码

导入到ESTK eSIM并下载eSIM profile

完成下载并打开数据漫游

实现无痛上网

此处无图 XD

IP显示和流量出口

速率测速结果

网友反馈

• 来自V2EX
  
• 来自Gmail
  
• 来自小红书
  

感受

如果只是想：“低成本长期挂一个海外流量” 我觉得这个方案目前还挺能打。
有问题欢迎留言，我也是用了一段时间的普通用户，能回答的会回答。

Overview

Jellyfin is a free, open-source media server that lets you organize, manage, and stream your personal collection of movies, TV shows, music, and photos. It turns your local files into a clean, Netflix-like library accessible from phones, TVs, browsers, or any compatible client — all without subscriptions, licenses, or paywalls. You own your data completely.
Key strengths on a Pi 4B: lightweight enough to run 24/7, supports basic transcoding (though CPU-limited), multi-user accounts, plugins for metadata scraping (Douban, TMDb, etc.), and a nice web UI + apps for almost every device.

Features at a Glance

• Personal media library with automatic organization
• On-the-fly transcoding (CPU fallback on Pi)
• Clients for Android, iOS, Roku, Fire TV, web, etc.
• Extensible via plugins (metadata, skins, intro skipper, etc.)
• Poster walls, collections, watch lists, live TV/DVR support

Installation via Docker (Recommended for Pi)

Docker is the cleanest way on Raspberry Pi OS — easy updates, isolated, and portable.

Update the system (always start here)

1

sudo apt update && sudo apt upgrade -y

Create directories for persistent data

Use a dedicated folder under your home dir (or wherever you prefer).

1

mkdir -p ~/jellyfin/config ~/jellyfin/cache

Create and edit docker-compose.yml

1
2

cd ~/jellyfin
sudo nano docker-compose.yml

Here’s a solid, modern compose file (adjusted for Pi 4B realities):

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23

version: "3.8"

services:
  jellyfin:
    image: jellyfin/jellyfin:latest
    container_name: jellyfin
    restart: unless-stopped
    network_mode: bridge
    ports:
      - 8096:8096       # HTTP
      - 8920:8920       # HTTPS (optional)
    environment:
      - PUID=1000       # your user ID (id -u)
      - PGID=1000       # your group ID (id -g)
      - TZ=Asia/Taipei  # or Asia/Shanghai if preferred
    volumes:
      - ./config:/config
      - ./cache:/cache
      - /mnt/NAS8T/Movies:/media/Movies:ro       # read-only for safety
      - /mnt/NAS8T/TVShows:/media/TVShows:ro
      - /home/roy/shared/JellyfinShows:/media/NetWorkVideo:ro
    # devices:                                # commented out — see note below
    #   - /dev/dri:/dev/dri

• Important note on hardware acceleration

  1 2

  On Raspberry Pi 4B, hardware transcoding via /dev/dri (VideoCore VI) is deprecated and unreliable in recent Jellyfin versions (10.9+). Official docs dropped support because it was immature, often fell back to CPU anyway, and Pi 5 has no hardware encoder at all. Most users on Pi 4B run without it — direct play works fine for compatible clients (e.g., VLC, Infuse). If you force /dev/dri, you might get partial decode but no real gains and possible crashes. Leave it commented unless you're on an older version and have tested it.

Start the container

1
2

cd ~/jellyfin
docker compose up -d

Access the web UI

• Open your browser

  1	http://<your-pi-ip>:8096 (e.g., http://192.168.0.109:8096)

  First run → wizard sets up library paths, users, metadata languages, etc.

Useful Docker Commands (CLI cheatsheet)

1
2
3
4
5
6
7
8
9
10
11

# Follow logs in real time
docker logs -f jellyfin

# Stop / start / restart
docker compose down
docker compose up -d
docker compose restart

# Or single container
docker stop jellyfin
docker start jellyfin

Making Jellyfin Better (Plugins & Tweaks)

Install plugins via Dashboard → Plugins → Catalog

Popular ones

• Douban metadata (for Chinese titles, actors, ratings)
• TMDb Box Sets / Collections
• Intro Skipper (auto-skip TV intros)
• MetaShark or similar for better scraping

Recommended reading (Chinese-focused)

Please following articles as below to enhance your Jellyfin

• Jellyfin影视库插件推荐，各国小姐姐家装、豆瓣刮削、动漫刮削、皮肤美化
• Jellyfin 豆瓣元数据插件

YouTube guides worth watching

• 利用NAS加Jellyfin打造个人影音库

• Jellyfin实用插件两枚 — MetaShark + intro skipper

• 群晖Jellyfin装豆瓣插件

• 家庭影院一条龙开源免费，实现本地视频自动刮削，字幕下载，海报墙展示

Live TV / IPTV Setup

Add an M3U playlist for free channels:

• Subscription URL example: https://garyshare.sharewithyou.dpdns.org/mylist.m3u
• In Jellyfin: Dashboard → Live TV → Add Tuner → M3U Tuner → paste URL → follow wizard for guide data (XMLTV if available).

References

• Official docs: Jellyfin Docs
• Docker-specific: Jellyfin Docker
• Hardware accel info: Hardware Acceleration

This setup gives you a rock-solid personal media server on Pi 4B — low power, always-on, and expandable. Direct play is king here; avoid heavy transcoding unless you upgrade to a mini PC later. Hit me up if you run into permission issues with volumes or plugin setup. Enjoy the poster wall! 🍿

Overview

yddict is a lightweight CLI dictionary tool that queries Youdao Dictionary (有道词典) directly from the terminal. It’s perfect for developers who want quick English ↔ Chinese lookups without opening a browser or dealing with heavy GUI apps full of ads.

Features

• English → Chinese translation
• Chinese → English translation
• Phonetic symbols / pronunciation hints
• Web-sourced definitions and examples
• Custom proxy support
• Custom output color
• Completely ad-free

Requirements

• Node.js ≥ 12.0.0
• npm ≥ 6.0.0

yddict is pure JavaScript, so it runs fine on ARM64 like Raspberry Pi 4B — no compilation needed.

Demo

Installation Steps

Update your system

Always a good first step on fresh Raspberry Pi OS installs.

1

sudo apt update && sudo apt upgrade -y

Install Node.js (LTS recommended)

The Node source repo gives you a clean, up-to-date version. Here we’re using v20.x (stable as of 2025–2026).

1
2

curl -fsSL https://deb.nodesource.com/setup_20.x | sudo -E bash -
sudo apt install -y nodejs

Verify versions

1
2

node -v    # should show v20.x.x
npm -v     # should show 10.x.x or similar

Install yddict globally

1

sudo npm install -g yddict

Quick check that it’s installed

1

npm list -g --depth=0

You should see yddict listed under global packages.

Optional: Verify Installation Paths

• Check the package metadata

  1	cat /usr/lib/node_modules/yddict/package.json

  Look for the bin field — it maps the yd command to index.js.

• Peek at the main script (just for curiosity)

  1	cat /usr/lib/node_modules/yddict/index.js

  You’ll see it uses request, chalk, cli-spinner, etc., to fetch and pretty-print results from Youdao.

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44

#!/usr/bin/env node

const request = require('request')
const chalk = require('chalk')
const Spinner = require('cli-spinner').Spinner
const isChinese = require('is-chinese')
const urlencode = require('urlencode')
const noCase = require('no-case')
const config = require('./lib/config')
const Parser = require('./lib/parser')

let word = process.argv.slice(2).join(' ')
if (!word) {
console.log('Usage: yd <WORD_TO_QUERY>')
process.exit()
}

const spinner = new Spinner('努力查询中... %s')

if (config.spinner) {
spinner.setSpinnerString('|/-\\')
spinner.start()
}

const isCN = isChinese(word)

word = isCN ? word : noCase(word)

const options = {
'url': config.getURL(word) + urlencode(word),
'proxy': config.proxy || null
}

const ColorOutput = chalk.keyword(config.color)
request(options, (error, response, body) => {
if (error) {
console.error(error)
}

        if (config.spinner) {
                spinner.stop(true)
        }
        console.log(ColorOutput(Parser.parse(isCN, body)))
})

Basic Usage

Run lookups directly in the terminal

1
2

yd hello
yd 世界

It handles both directions automatically (detects Chinese characters).

Convenience Tip: Shell Alias

If you prefer a shorter command:

1

nano ~/.bashrc

Add this line at the end:

1

alias y='yd'

Save, then reload:

1

source ~/.bashrc

Now you can just type:

1

y hello

Uninstall (if needed)

1

sudo npm uninstall -g yddict

Notes & Gotchas

• Proxy setup: If you’re behind a corporate/VPN/proxy, edit ~/.config/configstore/yddict.json after first run

  1 2 3 4

  { "proxy": "https://your-proxy:port", "color": "yellow" }

• No releases published on GitHub, but the repo is still active (Hacktoberfest-tagged, ongoing commits). Install via npm is the official way.
• Works great on headless Pi setups — super lightweight for quick lookups while coding or reading docs.

References

• GitHub repo: yddict
• English README: yddict English README

That’s it — you now have a fast, terminal-based dictionary on your Pi. Handy when you’re knee-deep in code and need to check a word without breaking flow. Enjoy!

Overview

This is a quick roundup of lesser-known but genuinely useful shortcuts and behaviors in Windows Task Manager that save time when troubleshooting or managing processes.

Launching Task Manager

Standard ways everyone knows

• Ctrl + Shift + Esc
• Right-click taskbar → Task Manager
• Ctrl + Alt + Del → Task Manager

Lesser-known: force the classic (Windows 10-style) top-tab UI instead of the modern sidebar layout (especially helpful on Windows 11):

1

Win + R → taskmgr -d

Modern UI (default on Win 11)

• vertical tabs on the left (Processes, Performance, App history, etc.)
  

Classic UI (-d flag)

• horizontal text tabs across the top, more compact for some workflows
  

Tip

Tip 1: Freeze Refresh with Ctrl (Pause Updates)

• Problem: On the Processes or Details tab, high-CPU or memory-hogging entries keep jumping around every second, making it hard to right-click → End task, open file location, or even read the exact name.
• Fix:
  • Switch to Processes, Details, or Performance tab.
  • Hold down the Ctrl key and keep it held.
    • → All graphs, percentages, and sort order freeze instantly.
  • Do whatever you need (select, sort manually, right-click actions).
  • Release Ctrl → live updating resumes.

This is especially useful when hunting a process that’s spiking and moving up/down the list rapidly.

Tip 2: Ctrl + Click “Run new task” → Elevated Command Prompt

Normal behavior

File → Run new task → opens a simple Run dialog (like Win + R) that runs as your current user.

Hidden behavior

• Go to File menu.
• Hold Ctrl key.
• Click “Run new task”.
  → Instead of the dialog, Windows launches an elevated (Administrator) Command Prompt (cmd.exe) directly.
• No UAC prompt if you’re already an admin; otherwise it prompts as usual.

This is faster than searching for cmd → right-click → Run as administrator, especially when Task Manager is already open because something is seriously misbehaving.

Quick Notes

• The modern sidebar UI became default in Windows 11 (and late Windows 10 builds).
• -d flag forces the legacy top-tabs layout on any recent Windows version that still supports it.
• Both Ctrl tricks work in the modern and classic UIs.

References

• Originally spotted and confirmed in this V2EX thread (Chinese tech forum, lots of good replies)
  • https://www.v2ex.com/t/1187198#reply44

These are small quality-of-life things, but once you know them they become muscle memory for sysadmin/ troubleshooting sessions.

Overview

This is a battle-tested workflow for exporting Docker images from a Raspberry Pi (or any ARM-based Linux box) to an exFAT-formatted USB drive, then loading them back on a fresh install or different machine. It avoids network pulls, preserves exact image tags and IDs, and sidesteps common permission headaches when dealing with auto-mounted removable media.
Tested on Raspberry Pi 4B running Raspberry Pi OS / Debian-based distro, but the commands are generic to any modern Linux with Docker.

Why This Matters

Docker images on embedded devices eat up space fast. TF cards and small SSDs fill quickly, and rebuilding everything from scratch over slow or unreliable internet is painful. A reliable offline backup lets you:

• Wipe/reinstall the OS without losing your container environment
• Migrate to new hardware quickly
• Keep a cold spare of critical images (e.g. transmission, cloudreve, custom builds)

Exporting Images to USB

Plug in the USB drive and identify it.

• execute: lsblk
  

  Look for something like /dev/sda1 or /dev/sdb1 with exFAT filesystem. Never guess — double-check size and label.

Unmount any auto-mount (devmon/udisks2 usually mounts to /media//)Bash

• execute: sudo umount /media/devmon/DockerIMGs
• adjust path to match your system
  

Create a clean mount point

• execute: sudo mkdir -p /mnt/docker-usb
  

Mount the drive manually with proper ownership and permissions

This is the most important step — auto-mounts almost always give root:root or 1000:1000 with restrictive umask, making it annoying to write from non-root.

• execute:

  1 2 3

  sudo mount -t exfat \ -o uid=1000,gid=1000,umask=022 \ /dev/sdb1 /mnt/docker-usb

• uid=1000,gid=1000 → files appear owned by your regular user
• umask=022 → creates files as 644, directories as 755 (normal user-writable)
  
  

Create backup directory on the drive

• execute

  1 2 3

  cd /mnt/docker-usb sudo mkdir docker_images_backup sudo chown $USER:$USER docker_images_backup # optional but cleaner

Export images

Use docker save to create a single .tar per image (or multi-image tar if you prefer).
Replace with your actual image:tag pairs.

1
2

sudo docker save -o /mnt/docker-usb/docker_images_backup/transmission_latest.tar lscr.io/linuxserver/transmission:latest
sudo docker save -o /mnt/docker-usb/docker_images_backup/cloudreve_4.10.1-arm.tar cloudreve:4.10.1-arm

Run as your user if the mount is correctly owned; otherwise prefix with sudo.

Verify sizes and integrity

• execute:

  1 2	ls -lh /mnt/docker-usb/docker_images_backup du -sh /mnt/docker-usb/docker_images_backup/*

Safely unmount

• execute:

  1 2	cd ~ sudo umount /mnt/docker-usb

  Wait for the command to return, then physically remove the drive.

Restoring Images on a New / Reinstalled System

Plug in the USB drive.

It will probably auto-mount again (often to /media// or /mnt/docker-usb if you left the directory).

Locate the mount point and check contents

• execute:

  1 2	lsblk -f ls -lh /mnt/docker-usb/docker_images_backup # adjust path

Load the images

• execute

  1	sudo docker load -i /media/devmon/ImgBackU/docker_images_backup/openwrt-aarch64.tar

• You’ll see output like Loaded image: buddyfly/openwrt-aarch64:latest
  

Verify everything came back intact

• execute: sudo docker images
• Check that repository:tag and IMAGE ID match exactly what you had before. If IDs match, layers were preserved perfectly.
  

(Optional) Unmount when done

• execute:
  • sudo umount /mnt/usb
• or wherever it mounted

Quick Tips & Gotchas

• Always use docker save / docker load — never copy /var/lib/docker directly. The latter is dangerous and usually breaks.
• exFAT is ideal for cross-platform (Linux ↔ Windows/Mac) backups. If you only stay on Linux, ext4 works too but isn’t as portable.
• If mount -t exfat fails → install exfatprogs or exfat-utils (sudo apt install exfatprogs).
• Large images (>10–20 GB) take time to save/load — be patient and watch iotop or htop if curious.
• For bulk export of all images:

  1	docker images --format '{{.Repository}}:{{.Tag}}' | xargs -n1 docker save -o backup.tar

  but single-file tar gets huge; per-image is usually better for manageability.

References

• Docker官方文档 - docker save
• Docker官方文档 - docker load
• Linux mount exFAT 文件系统（Explain the usage of uid/gid/umask parameters）

This workflow has survived multiple Pi reimages and migrations for me. Keep your USB drive in a safe place — it’s now your offline Docker lifeline.

Overview

This guide explains the boot chain for early 2017 Nintendo Switch units (unpatched, vulnerable to Fusee-Gelee). It covers how the console is forced into a user-controlled bootloader environment instead of the stock Horizon OS (HOS).
The short version: We’re exploiting a hardware-level vulnerability in the Tegra X1 bootROM to inject Hekate as the primary bootloader. This lets you chainload custom firmware (Atmosphère), manage NAND/emuMMC, run backups, boot Android/Linux, and more — all while keeping the original system intact if you use proper isolation.

Key Terms

Unpatched Switch (2017 models)

Only early units with serials in specific ranges are vulnerable to Fusee-Gelee (a permanent bootROM exploit). Later models (2018+, Mariko chip: V2, Lite, OLED) are hardware-patched and require a modchip. No software update or NAND restore can fix the vulnerability — it’s in silicon.

RCM (Recovery Mode)

NVIDIA’s factory recovery mode on the Tegra X1. When triggered, the console halts boot before loading any OS and waits for a signed payload over USB. This is the entry point for all softmods on unpatched units.

RCM Jig

A simple shorting tool (often a metal clip or 3D-printed piece) that bridges pin 10 on the right Joy-Con rail to ground. This tricks the console into entering RCM on power-on.

Payload Injector / RCM Loader

Hardware dongle (e.g., small USB device with blue LED) that stores and sends the Hekate payload (hekate_ctcaer_x.x.x.bin) over USB when the console is in RCM. Without one, you need a PC/phone running TegraRcmGUI, Rekado, etc., every boot.

Hekate

• A full-featured, GUI-based bootloader (comparable to GRUB on PC). It handles:
  • Chainloading CFW (Atmosphère via fusee or package3)
  • eMMC/emuMMC backup & restore
  • emuMMC creation/migration/repair
  • SD partitioning for multi-OS setups
  • USB mass storage (UMS) mode for SD/eMMC
  • Hardware info readout
  • Auto-boot configuration

Atmosphère (CFW)

The de-facto custom firmware layer. It patches Horizon OS on-the-fly rather than replacing it. Enables homebrew, sysmodules, overlays, emuMMC booting, etc.

Tesla Overlay

A modular overlay framework (accessed via key combo in-game or menu). Allows real-time system tweaks, monitoring, and plugin loading without replacing the UI.

Verifying Compatibility

Only unpatched (V1) Switches support this method. Check your serial number (System Settings → System, or on the bottom sticker)
If your unit is patched, this guide does not apply — look into modchip solutions instead.

• Use a checker like https://ismyswitchpatched.com/
• Safe ranges (unpatched): XAW1xxxxxx–XAW7xxxxxx, etc.
• “Possibly patched” ranges require testing with a jig.
• Anything XKWxxxxxx, XKJxxxxxx, or later is patched (Mariko).
  

Step-by-Step Boot Process (Unpatched V1 Only)

1. Format your microSD card to FAT32 (use a tool like guiformat or Rufus if needed; 128GB+ recommended for emuMMC).
2. Copy the full Atmosphère package (e.g., latest stable release + sigpatches) to the SD root. Typical structure includes /atmosphere/, /bootloader/, /switch/, etc. Ensure Hekate files are present in /bootloader/.
3. Insert the SD card into the Switch. Remove the right Joy-Con.
4. Power on normally (briefly) to ensure it reaches the lock screen or menu, then power off completely.
5. Insert the RCM jig into the right Joy-Con rail.
6. Connect your payload injector / RCM loader (or USB cable to PC/phone). The LED should turn blue/indicate successful injection when Hekate loads.
7. Power off fully (hold POWER button).
8. Hold VOL+ and tap POWER once. The console enters RCM → Hekate boots.

You should now see the Hekate splash, then the Nyx GUI menu.

Hekate Nyx GUI Breakdown

Home / Launch

• Main screen with boot entries (e.g., CFW (emuMMC), CFW (sysNAND), Stock, Android/Linux if configured). Select to boot.
  
  

Tools

• Backup/Restore eMMC or emuMMC (BOOT0/BOOT1/GPP partitions — always do full backups first).
• Partition SD Card (for multi-boot setups).
• USB Mass Storage: Mount SD/eMMC/emuMMC as a drive on PC (eject properly via “Close” before unplugging).
• USB Gamepad mode.
• Archive bit fixer, touch calibration, benchmarks, AutoRCM toggle (avoid on patched units).
  
  

USB Tools

• Dedicated section for UMS modes (SD, eMMC raw partitions) and gamepad emulation.
  
  

Console Info

• Detailed readout: SoC revision, fuse status, RAM config, display/touch, eMMC/SD health, battery/PSU/charger stats.
  

Options

• Set autoboot entry (e.g., emuMMC CFW) and boot delay (bootwait in seconds).
• Backlight level, themes (via nyx.ini).
• Boot protection, auto-patching flags (NoGC, etc.).
• Recommended: Enable autoboot to emuMMC with 3–5s delay for quick normal boots while still allowing VOL- interrupt to enter menu.
  

What This Actually Achieves

• Boot control — Hekate intercepts boot before Horizon loads, giving you root-level choice over what runs.
• Isolation via emuMMC — Keep a clean sysNAND for online/stock use; run CFW/plugins only in emulated NAND on SD. Minimizes ban risk if you block Nintendo servers properly (exosphere, DNS-MITM, 90DNS).
• Hardware freedom — Turns the Switch into a general-purpose ARM device capable of running Linux, Android, bare-metal tools, etc.

This isn’t about piracy — it’s about owning the boot process and exploring what the hardware can really do.

References

• Hekate GitHub
  • https://github.com/CTCaer/hekate
• Atmosphère GitHub
  • https://github.com/Atmosphere-NX/Atmosphere
• Detailed introduction to Hekate interface
  • https://docs.qq.com/doc/DVVpQcmd1RFNpUnNL
• NH Switch Guide (authoritative)
  • https://switch.hacks.guide/
• Fusee-Gelee explanation (Chinese)
  • https://github.com/Ginurx/fusee_gelee_explained_in_chinese
• Console Hacking Glossary
  • http://gledos.science/hacking-game-console-word-list.html#%E5%9B%BD%E4%BA%A7%E8%8A%AF%E7%89%87
• Switch jailbreak history
  • https://xmper.cc/2023/02/21/Switch-jailbreaking-history/
• Tesla - The Nintendo Switch Overlay Menu
  • https://gbatemp.net/threads/tesla-the-nintendo-switch-overlay-menu.557362/
• Tesla-Menu
  • https://github.com/WerWolv/Tesla-Menu
• Tesla overview
  • https://switch.hacks.guide/homebrew/tesla-menu.html

Always verify downloads from official repos. Keep backups current. Test in emuMMC first.

Prerequisites

• A modded Nintendo Switch with working Hekate bootloader (recommend using the latest Hekate via fusee or similar; version 6.0.3 or newer required).
• LineageOS for nx (Android TV build) – currently LineageOS 21 (Android 14) or 22.2 (Android 15) depending on the latest available from switchroot / LineageOS.
• SD card formatted as FAT32 or exFAT with enough free space.
• MindTheGapps package for the corresponding Android version and architecture (arm64 ATV variant).
  

About MindTheGapps

MindTheGapps is a lightweight, custom Google Apps package maintained for LineageOS. For Android TV builds, use the ATV-specific variant (full or minimal). It includes core Google services without the heavy bloat found in stock OpenGApps.

Download MindTheGapps – a custom Google Apps package for Android

• Website:
  • https://mindthegapps.com/
    

Why Choose MindTheGapps?

• Smaller footprint compared to full OpenGApps.
• Better compatibility with LineageOS Android TV builds on Switch hardware.
• Officially recommended on the LineageOS GApps page for recent versions.
  

How to Install MindTheGapps?

Installing GApps

If you’ve already booted into Android TV once without GApps, installing them directly will usually cause bootloops or crashes due to mismatched system properties and missing framework hooks.

Recommended procedure

1. Wipe data if the system has already booted once
   2. Boot into Hekate（Hold Volume Up while launching Hekate） → More Configs → Hold VOL+ to select LineageOS Recovery (or TWRP if using that).
   3. In recovery: Wipe → Factory Reset / Format data / factory reset.
      
      
   4. This clears the Android data and cache partitions only — it does not touch Horizon OS (HOS) or your emuMMC/sysMMC.
2. Download the correct GApps package
   3. Go to https://wiki.lineageos.org/gapps/ or directly to MindTheGapps GitHub releases.
      4. https://github.com/MindTheGapps/14.0.0-arm64-ATV/releases/tag/MindTheGapps-14.0.0-arm64-ATV-20240523_192151
   4. For Android 14 (LineageOS 21): MindTheGapps-14.0.0-arm64-ATV-full-[date].zip
      
      
   5. For Android 15 (LineageOS 22): MindTheGapps-15.0.0-arm64-ATV-[variant].zip
   6. Place the zip on the root of your SD card.
3. Flash GApps in recovery
   4. In LineageOS Recovery: Install → Apply Update → Choose from SWITCH SD (or SD card).
      
      
      
   5. Select the MindTheGapps zip.
   6. When you see “Signature verification failed”, choose Yes to proceed (normal for sideloaded packages).
   7. Let it install completely.
4. Reboot
   5. Return to main recovery menu → Reboot system now.
   6. The device should now boot into Android TV with Google services available for the first-time setup.
5. Conclusion
   

Magisk (Root)

To gain root access on the Android installation

• For most models (especially OLED / Mariko units), use Magisk v25.2 (newer versions may have compatibility issues with older boot images or SELinux policies on Switch hardware).
• Download: Magisk-v25.2.apk from https://github.com/topjohnwu/Magisk/releases/tag/v25.2
• Install the APK like any app once Android is running, then follow the standard Magisk patching procedure (patch boot image via the app or manually via recovery if preferred).

Video References

• Dual-boot Atmosphere + Android setup and root guide (older but concepts still apply)
  • 莫老师的附件表
    • https://zxmls.lol/#/?id=ep150-双系统切换！给switch刷入大气层及android11系统，并root
• Recent 2025 Android 15 flashing tutorial
• Comprehensive Switch Android install covering OLED compatibility

Official Resources

• Switchroot downloads
  • https://download.switchroot.org/
• LineageOS device page (nx / nx_tab variants
  • https://download.lineageos.org/devices/nx/builds
• Switchroot LineageOS Downloader tool
  • https://github.com/sthetix/Switchroot-LineageOS-Downloader
• Main documentation
  • https://wiki.switchroot.org/wiki/
  • https://wiki.lineageos.org/
  • https://wiki.lineageos.org/devices/nx/variant4/
  • Android 14/15 builds info: https://wiki.switchroot.org/wiki/android/android-14-15
  • GApps: https://wiki.lineageos.org/gapps/

Notes

• Always double-check your Switch model (Erista vs Mariko / OLED) and download the matching build variant (nx for TV on some, nx_tab for tablet-style).
• Android TV builds are optimized for docked/controller use; touch works poorly or not at all.
• Backup your SD card contents before repartitioning or major changes.
• If you run into signature mismatches or boot issues after GApps, repeat the data wipe + clean flash sequence.

  This should get you a clean, functional Android TV environment on your Switch. Test thoroughly after first boot.

Overview

For Windows

• This script is designed to collect detailed system information from Windows machines.
• It is written in Bash and can run in command-line environments such as Git Bash, Cygwin, or WSL.
• The script retrieves system configuration data by calling native Windows tools and commands.

For Linux

• This script is designed for Linux and macOS systems, also written in Bash.
• It leverages typical Linux system features such as the /proc filesystem and standard command-line utilities to collect detailed system information.

Purpose

For Windows

• Quick system diagnostics
  • Helps users and IT administrators quickly understand hardware and software configuration.
• System information archiving
  • Records system status before deployment, troubleshooting, or upgrades.
• Remote technical support
  • Allows engineers to collect system details remotely for issue analysis.
• System health monitoring
  • Tracks key system metrics such as memory usage, disk space, and CPU status.
• Network configuration inspection
  • Displays public and private IP addresses, network interface information, and approximate location.

For Linux

• Comprehensive system overview
  • Provides a complete snapshot of system information from hardware to software.
• Performance baseline creation
  • Establishes performance baselines for later monitoring and optimization.
• Server status inspection
  • Useful for server administration, especially for uptime and load monitoring.
• Troubleshooting assistance
  • Provides useful diagnostic data for debugging system or performance issues.
• Compatibility validation
  • Checks whether the system meets the requirements for specific software or applications.

Use Cases

For Windows

• IT administrators performing system audits and asset management
• Support engineers collecting client system data for troubleshooting
• Developers verifying hardware configuration of development environments
• Compatibility checks before system migration or upgrades
• Security auditing and compliance verification

For Linux

• Server administration and monitoring
• Cloud instance configuration validation
• Base image inspection in container environments
• Environment checks before distributed system deployment
• Performance tuning and capacity planning
• Security hardening and compliance auditing

Shared Features of Both Scripts

Technical Characteristics

• Cross-platform design
  • Both scripts target different operating systems but use standard Bash syntax.
• No additional dependencies
  • The scripts rely only on built-in system tools.
• Colorized output
  • Green-colored output improves readability.
• Modular output structure
  • Information is grouped by category for clarity.
• Basic error handling
  • Handles command failures gracefully.

Information Collected

• Basic information
  • Hostname, username, operating system version
• Hardware information
  • CPU model, memory size, disk configuration
• Network information
  • Public and private IP addresses, network interfaces, geolocation
• System status
  • Uptime, boot time, timezone
• Storage information
  • Disk partitions, mount points, usage statistics

Benefits

• Improved efficiency
  • Automatically collects system data instead of manual inspection.
• Standardized output format
  • Consistent output makes comparison and analysis easier.
• Historical reference
  • Output can be saved for system change tracking.
• Extensibility
  • Additional checks can be added easily if needed.
• Educational value
  • Helps users understand system structure and configuration.

Example Output

For Windows

For Linux

Windows Version Script

Collect Windows System Information

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169

#!/bin/bash

# Windows 主机系统信息收集脚本
# 颜色设置
echo -e "\e[32m"

echo "=== Windows主机系统信息 ==="

# 主机名和用户名
echo "主机名: $(hostname)"
echo "用户名: ${USERNAME:-$USER}"

# 网络地址信息
echo "=== 网络地址信息 ==="

# 公网IP
if command -v curl &>/dev/null; then
public_ip=$(curl -s --max-time 5 ifconfig.me 2>/dev/null || curl -s --max-time 5 ipinfo.io/ip 2>/dev/null || echo "无法获取")
else
public_ip="无法获取 (curl未安装)"
fi
echo "公网IP地址: $public_ip"

# IP地理位置
if command -v curl &>/dev/null && [[ "$public_ip" != "无法获取" && "$public_ip" != "无法获取 (curl未安装)" ]]; then
echo -n "IP地理位置: "
location=$(curl -s --max-time 3 "ipinfo.io/$public_ip/json" 2>/dev/null | grep -E '"city"|"region"|"country"' | \
sed 's/.*: "//;s/",//' | tr '\n' ' ' | sed 's/ $//')
[[ -n "$location" ]] && echo "$location" || echo "未知"
else
echo "IP地理位置: 需要网络连接"
fi

# 内网IP地址
echo "内网IP地址:"
ipconfig 2>/dev/null | grep -E "(IPv4|IP.v4)" | grep -v "169.254." | while read line; do
interface=$(echo "$line" | cut -d' ' -f1 | tr -d ':')
ip=$(echo "$line" | cut -d: -f2 | sed 's/^[ \t]*//')
[[ -n "$ip" ]] && echo "  $interface: $ip"
done

# 系统详细信息
echo "=== 系统详细信息 ==="

if command -v systeminfo &>/dev/null; then
sysinfo=$(systeminfo 2>/dev/null)

    # 操作系统版本
    os_version=$(echo "$sysinfo" | grep "OS Name" | head -1 | cut -d: -f2 | sed 's/^[ \t]*//')
    echo "操作系统版本: $os_version"
    
    echo "内核版本: $(uname -r)"
    
    # 启动设备
    boot_device=$(echo "$sysinfo" | grep -i "Boot Device" | head -1 | cut -d: -f2- | sed 's/^[ \t]*//')
    [[ -n "$boot_device" ]] && echo "启动设备: $boot_device"
    
    # 时区
    time_zone=$(echo "$sysinfo" | grep -i "Time Zone" | head -1 | cut -d: -f2- | sed 's/^[ \t]*//')
    [[ -n "$time_zone" ]] && echo "时区: $time_zone"
    
    # 系统运行时间
    boot_time=$(echo "$sysinfo" | grep -i "System Boot Time" | head -1 | cut -d: -f2- | sed 's/^[ \t]*//')
    if [[ -n "$boot_time" ]]; then
        echo "系统启动时间: $boot_time"
        
        # 计算运行时间
        boot_date=$(date -d "$boot_time" "+%s" 2>/dev/null || echo "")
        if [[ -n "$boot_date" ]]; then
            current_date=$(date "+%s")
            uptime_seconds=$((current_date - boot_date))
            days=$((uptime_seconds / 86400))
            hours=$(( (uptime_seconds % 86400) / 3600 ))
            minutes=$(( (uptime_seconds % 3600) / 60 ))
            echo "系统运行时间: ${days}天 ${hours}小时 ${minutes}分钟"
        fi
    fi
else
echo "操作系统版本: 无法获取 (systeminfo未找到)"
echo "内核版本: $(uname -r)"
fi

# CPU信息
echo "=== CPU信息 ==="

if command -v wmic &>/dev/null; then
cpu_info=$(wmic cpu get name 2>/dev/null | tail -2 | head -1 | sed 's/^[ \t]*//;s/[ \t]*$//')
[[ -n "$cpu_info" && ! "$cpu_info" =~ "Name" ]] && echo "CPU型号: $cpu_info" || echo "CPU型号: 无法获取"
else
echo "CPU型号: 无法获取 (wmic未找到)"
fi

# 内存信息
echo "=== 内存信息 ==="

if command -v wmic &>/dev/null; then
memory_bytes=$(wmic computersystem get TotalPhysicalMemory 2>/dev/null | grep '[0-9]' | awk '{print $1}')
if [[ -n "$memory_bytes" ]]; then
memory_gb=$(echo "$memory_bytes" | awk '{printf "%.2f", $1/1024/1024/1024}')
echo "内存大小: ${memory_gb}GB"

        # 获取可用内存
        free_bytes=$(wmic OS get FreePhysicalMemory 2>/dev/null | grep '[0-9]' | awk '{print $1 * 1024}')
        if [[ -n "$free_bytes" ]]; then
            free_gb=$(echo "$free_bytes" | awk '{printf "%.2f", $1/1024/1024/1024}')
            used_gb=$(echo "$memory_bytes $free_bytes" | awk '{printf "%.2f", ($1-$2)/1024/1024/1024}')
            echo "内存使用: 已用 ${used_gb}GB / 可用 ${free_gb}GB"
        fi
    else
        echo "内存大小: 无法获取"
    fi
else
echo "内存大小: 无法获取 (wmic未找到)"
fi

# 磁盘信息
echo "=== 磁盘分区信息 ==="

echo "逻辑磁盘信息:"
if command -v wmic &>/dev/null; then
wmic logicaldisk where "drivetype=3" get deviceid,size,freespace,volumename /format:list 2>/dev/null | \
tr -d '\r' | awk -F'=' '
/DeviceID/ {drive=$2}
/Size/ {size=$2/1024/1024/1024}
/FreeSpace/ {free=$2/1024/1024/1024; used=size-free}
/VolumeName/ {volumename=$2}
/^$/ && drive {
if (volumename == "") volumename="本地磁盘";
printf "  %s (%s): 总大小=%.2fGB, 已用=%.2fGB, 可用=%.2fGB\n",
drive, volumename, size, used, free;
drive=""; volumename=""
}
'
else
echo "  WMIC命令不可用"
fi

# 使用DF命令查看挂载点（备选方法）
if command -v df &>/dev/null; then
echo -e "\n使用DF命令查看挂载点:"
df -h 2>/dev/null | grep -E "^[A-Z]:" | while read line; do
echo "  $line" | awk '{
printf "  %s: 总大小=%s, 已用=%s, 可用=%s, 使用率=%s",
$1, $2, $3, $4, $5;
if (NF > 5) {
printf ", 挂载点=";
for(i=6;i<=NF;i++) printf "%s ", $i;
}
print ""
}'
done
fi

echo -e "\n=== 物理磁盘信息 ==="
if command -v wmic &>/dev/null; then
wmic diskdrive get model,size /format:list 2>/dev/null | \
tr -d '\r' | awk -F'=' '
/Model/ {model=$2}
/Size/ {size=$2/1024/1024/1024; printf "  %s: %.2fGB\n", model, size}
' | head -10
else
echo "  WMIC命令不可用"
fi

# 用户信息
echo -e "\n=== 当前登录用户 ==="
echo "  当前用户: ${USERNAME:-$USER}"

echo -e "\e[0m"

Linux Version Script

Collect Linux System Information

1
2
3
4
5
6
7
8
9
10
11
12
13
14
15
16
17
18
19
20
21
22
23
24
25
26
27
28
29
30
31
32
33
34
35
36
37
38
39
40
41
42
43
44
45
46
47
48
49
50
51
52
53
54
55
56
57
58
59
60
61
62
63
64
65
66
67
68
69
70
71
72
73
74
75
76
77
78
79
80
81
82
83
84
85
86
87
88
89
90
91
92
93
94
95
96
97
98
99
100
101
102
103
104
105
106
107
108
109
110
111
112
113
114
115
116
117
118
119
120
121
122
123
124
125
126
127
128
129
130
131
132
133
134
135
136
137
138
139
140
141
142
143
144
145
146
147
148
149
150
151
152
153
154
155
156
157
158
159
160
161
162
163
164
165
166
167
168
169
170
171
172
173
174
175
176
177
178
179
180
181
182
183
184
185
186
187
188
189
190
191
192
193
194
195
196
197
198
199
200
201
202
203
204
205
206
207
208
209
210
211
212
213
214
215
216
217
218
219
220
221
222
223
224
225
226
227
228
229
230
231
232
233
234
235
236
237
238
239
240
241
242
243
244
245
246
247
248
249
250
251
252
253
254
255
256
257

#!/bin/bash

# Linux/macOS 主机系统信息收集脚本
# 颜色设置
echo -e "\e[32m"

# 检测是Linux还是macOS
if [[ "$(uname)" == "Darwin" ]]; then
    OS_TYPE="macOS"
else
    OS_TYPE="Linux"
fi

echo "=== ${OS_TYPE}主机系统信息 ==="

# 主机名和用户名
echo "主机名: $(hostname)"
echo "用户名: $USER"

# 网络地址信息
echo "=== 网络地址信息 ==="

# 公网IP
if command -v curl &>/dev/null; then
    public_ip=$(curl -s --max-time 5 ifconfig.me 2>/dev/null || curl -s --max-time 5 ipinfo.io/ip 2>/dev/null || echo "无法获取")
else
    public_ip="无法获取 (curl未安装)"
fi
echo "公网IP地址: $public_ip"

# IP地理位置
if command -v curl &>/dev/null && [[ "$public_ip" != "无法获取" && "$public_ip" != "无法获取 (curl未安装)" ]]; then
    echo -n "IP地理位置: "
    location=$(curl -s --max-time 3 "ipinfo.io/$public_ip/json" 2>/dev/null | grep -E '"city"|"region"|"country"' | \
        sed 's/.*: "//;s/",//' | tr '\n' ' ' | sed 's/ $//')
    [[ -n "$location" ]] && echo "$location" || echo "未知"
else
    echo "IP地理位置: 需要网络连接"
fi

# 内网IP地址
echo "内网IP地址:"
if command -v ip &>/dev/null; then
    ip -4 addr show 2>/dev/null | grep -v "127.0.0.1" | grep inet | awk '{print "  "$NF": "$2}'
elif command -v ifconfig &>/dev/null; then
    ifconfig 2>/dev/null | grep -E "inet " | grep -v "127.0.0.1" | awk '{print "  "$1": "$2}'
else
    echo "  无法获取网络接口信息"
fi

# 系统详细信息
echo "=== 系统详细信息 ==="

if [[ "$OS_TYPE" == "macOS" ]]; then
    echo "操作系统: macOS $(sw_vers -productVersion 2>/dev/null || echo '未知')"
elif [[ -f /etc/os-release ]]; then
    source /etc/os-release
    echo "操作系统: $NAME $VERSION"
else
    echo "操作系统: 无法确定"
fi

echo "内核版本: $(uname -r)"
echo "系统架构: $(uname -m)"

# 系统运行时间
echo "=== 系统运行时间 ==="

if [[ "$OS_TYPE" == "macOS" ]]; then
    # macOS获取运行时间
    if command -v sysctl &>/dev/null; then
        boot_time=$(sysctl -n kern.boottime 2>/dev/null | awk '{print $4}' | sed 's/,//')
        if [[ -n "$boot_time" ]]; then
            current_time=$(date +%s)
            uptime_seconds=$((current_time - boot_time))
            days=$((uptime_seconds / 86400))
            hours=$(( (uptime_seconds % 86400) / 3600 ))
            minutes=$(( (uptime_seconds % 3600) / 60 ))
            seconds=$((uptime_seconds % 60))
            
            echo "系统运行时间: ${days}天 ${hours}小时 ${minutes}分钟 ${seconds}秒"
            
            # 显示启动时间
            boot_date=$(date -r $boot_time "+%Y-%m-%d %H:%M:%S" 2>/dev/null)
            [[ -n "$boot_date" ]] && echo "系统启动时间: $boot_date"
        else
            echo "系统运行时间: 无法获取"
        fi
    else
        echo "系统运行时间: 无法获取 (sysctl命令不可用)"
    fi
else
    # Linux获取运行时间
    if [[ -f /proc/uptime ]]; then
        uptime_seconds=$(awk '{print $1}' /proc/uptime 2>/dev/null)
        if [[ -n "$uptime_seconds" ]]; then
            days=$((uptime_seconds / 86400))
            hours=$(( (uptime_seconds % 86400) / 3600 ))
            minutes=$(( (uptime_seconds % 3600) / 60 ))
            seconds=$((uptime_seconds % 60))
            echo "系统运行时间: ${days}天 ${hours}小时 ${minutes}分钟 ${seconds}秒"
            
            # 显示启动时间
            current_time=$(date +%s)
            boot_time=$((current_time - ${uptime_seconds%.*}))
            boot_date=$(date -d "@$boot_time" "+%Y-%m-%d %H:%M:%S" 2>/dev/null)
            [[ -n "$boot_date" ]] && echo "系统启动时间: $boot_date"
        else
            echo "系统运行时间: 无法获取"
        fi
    else
        echo "系统运行时间: 无法获取 (/proc/uptime不存在)"
    fi
fi

# 显示uptime命令的输出作为补充
if command -v uptime &>/dev/null; then
    echo -n "uptime命令输出: "
    uptime_output=$(uptime 2>/dev/null)
    [[ -n "$uptime_output" ]] && echo "$uptime_output" || echo "无法获取"
fi

# 时区信息
echo "=== 时区信息 ==="

if [[ -f /etc/timezone ]]; then
    echo "时区: $(cat /etc/timezone)"
elif command -v timedatectl &>/dev/null; then
    timezone=$(timedatectl show --property=Timezone --value 2>/dev/null)
    [[ -n "$timezone" ]] && echo "时区: $timezone"
elif [[ -f /etc/localtime ]]; then
    # 尝试从/etc/localtime推断时区
    if [[ "$OS_TYPE" == "Linux" ]]; then
        timezone_link=$(readlink /etc/localtime 2>/dev/null)
        if [[ -n "$timezone_link" ]]; then
            timezone=$(echo "$timezone_link" | sed 's|.*/zoneinfo/||')
            echo "时区: $timezone"
        fi
    fi
fi

# 显示当前时间
echo "当前时间: $(date '+%Y-%m-%d %H:%M:%S')"

# CPU信息
echo "=== CPU信息 ==="

if [[ "$OS_TYPE" == "macOS" ]]; then
    cpu_info=$(sysctl -n machdep.cpu.brand_string 2>/dev/null || echo "未知")
    echo "CPU型号: $cpu_info"
    echo "CPU核心数: $(sysctl -n hw.ncpu 2>/dev/null || echo "未知")"
elif [[ -f /proc/cpuinfo ]]; then
    cpu_model=$(grep -m1 "model name" /proc/cpuinfo 2>/dev/null | cut -d: -f2 | sed 's/^[ \t]*//')
    cpu_cores=$(grep -c "^processor" /proc/cpuinfo 2>/dev/null || echo "未知")
    echo "CPU型号: ${cpu_model:-未知}"
    echo "CPU核心数: $cpu_cores"
    
    # 显示CPU频率（如果可用）
    cpu_mhz=$(grep -m1 "cpu MHz" /proc/cpuinfo 2>/dev/null | cut -d: -f2 | sed 's/^[ \t]*//')
    [[ -n "$cpu_mhz" ]] && echo "CPU频率: ${cpu_mhz} MHz"
else
    echo "CPU型号: 无法获取"
fi

# 内存信息
echo "=== 内存信息 ==="

if [[ "$OS_TYPE" == "macOS" ]]; then
    memory_bytes=$(sysctl -n hw.memsize 2>/dev/null)
    if [[ -n "$memory_bytes" ]]; then
        memory_gb=$(echo "$memory_bytes" | awk '{printf "%.2f", $1/1024/1024/1024}')
        echo "内存大小: ${memory_gb}GB"
    fi
elif [[ -f /proc/meminfo ]]; then
    total_memory=$(grep -m1 "MemTotal" /proc/meminfo 2>/dev/null | awk '{print $2}')
    if [[ -n "$total_memory" ]]; then
        memory_gb=$(echo "$total_memory" | awk '{printf "%.2f", $1/1024/1024}')
        echo "内存大小: ${memory_gb}GB"
        
        free_memory=$(grep "MemAvailable" /proc/meminfo 2>/dev/null | awk '{print $2}')
        if [[ -n "$free_memory" ]]; then
            free_gb=$(echo "$free_memory" | awk '{printf "%.2f", $1/1024/1024}')
            used_gb=$(echo "$total_memory $free_memory" | awk '{printf "%.2f", ($1-$2)/1024/1024}')
            use_percent=$(echo "$total_memory $free_memory" | awk '{printf "%.1f", 100-($2*100/$1)}')
            echo "内存使用: 已用 ${used_gb}GB / 可用 ${free_gb}GB (使用率: ${use_percent}%)"
        fi
    fi
else
    echo "内存大小: 无法获取"
fi

# 磁盘信息
echo "=== 磁盘分区信息 ==="

echo "磁盘使用情况:"
if command -v df &>/dev/null; then
    df -h 2>/dev/null | grep -E "^/dev/" | while read line; do
        echo "  $line" | awk '{
            printf "  %s: 总大小=%s, 已用=%s, 可用=%s, 使用率=%s, 挂载点=", 
                   $1, $2, $3, $4, $5;
            for(i=6;i<=NF;i++) printf "%s ", $i;
            print ""
        }'
    done
else
    echo "  df命令不可用"
fi

echo -e "\n=== 物理磁盘信息 ==="

if [[ "$OS_TYPE" == "macOS" ]]; then
    if command -v diskutil &>/dev/null; then
        diskutil list 2>/dev/null | grep -A5 "/dev/disk" | while read line; do
            if [[ "$line" =~ /dev/disk ]]; then
                echo "  $line"
            fi
        done
    else
        echo "  diskutil命令不可用"
    fi
elif command -v lsblk &>/dev/null; then
    lsblk -d -o NAME,SIZE,TYPE,MODEL 2>/dev/null | grep disk | head -5
elif [[ -f /proc/partitions ]]; then
    cat /proc/partitions 2>/dev/null | grep -v "major" | grep -E "^[[:space:]]*[0-9]" | awk '{print $4, $3}' | while read name size; do
        if [[ "$name" =~ ^[hs]d[a-z]$ ]] || [[ "$name" =~ ^nvme ]]; then
            size_gb=$(echo "$size" | awk '{printf "%.2f", $1/1024/1024}')
            echo "  $name: ${size_gb}GB"
        fi
    done
fi

# 用户信息
echo -e "\n=== 当前登录用户 ==="
echo "  当前用户: $USER"
if command -v who &>/dev/null; then
    users=$(who 2>/dev/null | awk '{print $1}' | sort | uniq)
    user_count=$(echo "$users" | wc -w 2>/dev/null)
    [[ "$user_count" -gt 1 ]] && echo "  其他登录用户: $(echo "$users" | tr '\n' ' ')"
    
    # 显示登录终端和登录时间
    echo -e "\n  登录会话详情:"
    who 2>/dev/null | head -5 | while read line; do
        echo "    $line"
    done
fi

# 系统负载信息
echo -e "\n=== 系统负载信息 ==="
if [[ -f /proc/loadavg ]]; then
    load=$(cat /proc/loadavg 2>/dev/null)
    [[ -n "$load" ]] && echo "  系统负载: $load" || echo "  系统负载: 无法获取"
elif command -v sysctl &>/dev/null && [[ "$OS_TYPE" == "macOS" ]]; then
    load=$(sysctl -n vm.loadavg 2>/dev/null)
    [[ -n "$load" ]] && echo "  系统负载: $load" || echo "  系统负载: 无法获取"
fi

echo -e "\e[0m"

Introduction

A bypass router architecture allows the main router to continue handling PPPoE dialing and DHCP services, while a Raspberry Pi acts as a secondary gateway within the same LAN.
By assigning a static IP and configuring routing rules, devices on the network can optionally route their traffic through the Raspberry Pi.

This approach has several advantages:

• It does not modify the existing network topology.
• The bypass router can be removed at any time without affecting the main network.
• Devices can selectively use the proxy gateway.

Objective

Run OpenWrt as a bypass gateway (transparent gateway) inside Docker on a Raspberry Pi.

• Network architecture diagram:
  

Enable Network Interface Promiscuous Mode

• Promiscuous mode allows both the host system and the virtualized OpenWrt instance to receive network packets.
• Choose one of the following depending on your connection type.
  • Wired Connection

    1	sudo ip link set dev eth0 promisc on

  • Wi-Fi Connection

    1	sudo ip link set dev wlan0 promisc on

Download the OpenWrt Docker Image

• Run the following command to download the image

  1	docker pull buddyfly/openwrt-aarch64

  
  
  
• After downloading, verify the image exists

  1	docker images

Check Whether a Docker Virtual Network Exists

• Before creating the network, check the existing Docker networks.

  1	docker network ls

• If a network with the same name already exists, remove it, e.g.:macnet

  1	docker network rm <network_name>

Create a Docker Virtual Network

• We will create a macvlan network so that the OpenWrt container can obtain an IP address within the same LAN as the main router.
• Replace the subnet and gateway values with your own network settings.
  
• Example

  1 2 3 4 5

  docker network create -d macvlan \ --subnet=10.10.10.0/24 \ --gateway=10.10.10.1 \ -o parent=eth0 \ macnet

• Example based on a typical home network:
  • Wired Connection

    1 2 3 4 5

    docker network create -d macvlan \ --subnet=192.168.0.0/24 \ --gateway=192.168.0.1 \ -o parent=eth0 \ macnet

  • WiFi Connection

    1 2 3 4 5

    docker network create -d macvlan \ --subnet=192.168.0.0/24 \ --gateway=192.168.0.1 \ -o parent=wlan0 \ macnet

How to Check Your Router Gateway (Windows)

If you are unsure about your router’s gateway address and subnet, use one of the following methods.

• Method 1: Router Admin Panel
  • Access your router’s management interface and check the LAN settings.
    
    
• Method 2: Command Line
  • Press Win + R
  • Enter: cmd
  • Run: ipconfig
  • Look for the Default Gateway field.
    
    

Verify the Docker Network

• Check whether the macvlan network was successfully created
  • docker network ls
    

Start the OpenWrt Container

• Run the following command

  1 2 3 4

  docker run --restart always -d \ --network macnet \ --privileged \ buddyfly/openwrt-aarch64:latest

  
  
• This command
  • runs the container in background
  • attaches it to the macvlan network
  • enables privileged mode

Verify the Running Container

• Check the running OpenWrt container

  1	docker ps -a | grep openwrt

  
  

Modify the OpenWrt Container IP Address

Enter the container shell

1

docker exec -it <container_id> ash

Edit the network configuration

1

sudo nano /etc/config/network

Modify the following fields

1
2
3

option ipaddr '192.168.0.252'
option gateway '192.168.0.1'
option dns '192.168.0.1'

• Explanation
  • ipaddr → the IP address of the bypass router
  • gateway → the main router IP
  • dns → usually the same as the gateway

Restart the network service

1

/etc/init.d/network restart

Test Network Connectivity

• Test connectivity with the main router

  1	ping 192.168.0.1

  If packets are received, the configuration is correct.

Permanently Enable Promiscuous Mode

• Exit the container and configure the host system.
• Edit the file
  • sudo nano /etc/network/interfaces
• Add one of the following lines.
  • Wired

    1	up ip link set eth0 promisc on

  • WiFi

    1	up ip link set wlan0 promisc on

• Verify the configuration
  • cat /etc/network/interfaces
    
• If permission issues occur

  1 2	ls -l /etc/network/interfaces sudo chmod 644 /etc/network/interfaces

Access the OpenWrt Gateway

• Open the browser and enter

  1	http://192.168.0.252

• Default credentials

  1 2	username: root password: password

• The first thing you should do is change the default password.
  

Change the System Theme

You can customize the OpenWrt interface theme through the system settings panel.

Configure the Upstream Router

Navigate to the network configuration page and set the upstream gateway to your main router.

Network Acceleration Settings

Enable hardware or software acceleration options depending on your OpenWrt build.

Configure the Proxy / Global Network Access

Install and configure the proxy plugin of your choice (for example OpenClash, Passwall, etc.).

Configure Devices to Use the Bypass Gateway

• For devices that should use the proxy
  • Set the network configuration to manual.
  • Assign an IP address within the LAN range.
  • Example

    1 2 3 4

    IP Address: 192.168.0.xxx Subnet Mask: 255.255.255.0 Gateway: 192.168.0.252 DNS: 192.168.0.252

    Devices using this gateway will route their traffic through the Raspberry Pi bypass router.

Conclusion

By running OpenWrt in Docker with a macvlan network on a Raspberry Pi, we can easily create a flexible bypass router that integrates seamlessly with an existing home network.
This architecture allows selective routing of traffic through a proxy gateway while keeping the main router configuration untouched.

Connectivity Test Commands

Run the following commands on the Raspberry Pi host to verify that the gateway and external network access are working properly.

• ip route show
• ping 8.8.8.8
• curl -Iv https://www.youtube.com --connect-timeout 10
• curl -Iv https://www.google.com --connect-timeout 10