Skip to content

Docker/Qemu Based Linux Kernel Learning, Development and Testing Environment; New Linux ELF Video Course from this project author: https://www.cctalk.com/m/group/88089283

License

Notifications You must be signed in to change notification settings

tinyclub/linux-lab

Repository files navigation

Linux Lab Logo

Subscribe Wechat

Wechat Public

Table of Content

1. Linux Lab Overview

1.1 Project Introduction

This project aims to create a Docker and QEMU based Linux development Lab to easier the learning, development and testing of Linux Kernel.

Linux Lab is open source with no warranty – use at your own risk.

Docker QEMU Linux Lab

Friendly Notice: TinyLab Community have developed Linux Lab Disk, you can buy from TinyLab.org's Taobao Shop or Bilibili Shop.

1.2 Project Homepage

Related Projects:

1.3 Demonstration

1.3.1 Free Video Courses In Chinese

1.3.2 Non-Free Video Courses In Chinese

1.4 Project Functions

Now, Linux Lab becomes an intergrated Linux learning, development and testing environment, it supports:

Items Description
Boards QEMU based, 7+ main Architectures, 20+ popular boards; Several real boards supported too
Components Uboot, Linux / Modules, Buildroot, Qemu, Linux v0.11, v2.6.10 ~ 5.x supported
Prebuilt All of above components have been prebuilt
Rootfs Support include initrd, harddisk, mmc and nfs, Debian availab for ARM
Docker Cross toolchains from gcc-4.3 available in one command, external ones configurable
Access Accessible from local or remote, include CLI and GUI, support bash, ssh, vnc, web vnc and web ssh
Network Builtin bridge networking, every board has network (except Raspi3)
Boot Support serial port, curses (bash/ssh friendly) and graphic booting
Testing Support automatic testing via make test target
Debugging debuggable via make debug target

Continue reading for more features and usage.

1.5 Project History

1.5.1 Project Origins

About 10 years ago (2010), a tinylinux proposal: Work on Tiny Linux Kernel accepted by Embedded Linux Foundation, therefore I have worked on this project for serveral months.

1.5.2 Problems Solved

During the project cycle, several scripts written to verify if the adding tiny features (e.g. gc-sections) breaks the other kernel features on the main CPU architectures.

These scripts uses qemu-system-ARCH as the cpu/board simulator, basic boot+function tests have been done for ftrace+perf, accordingly, defconfigs, rootfs, test scripts have been prepared, at that time, all of them were simply put in a directory, without a design or holistic consideration.

1.5.3 Project Born

They have slept in my harddisk for several years without any attention, until one day, docker and novnc came to my world, at first, Linux 0.11 Lab was born, after that, Linux Lab was designed to unify all of the above scripts, defconfigs, rootfs and test scripts.

2. Linux Lab Installation

Linux Lab uses Docker, if have already installed Docker and configured the best mirror site of docker images, it is very easy to install Linux Lab.

If really a Linux newbie or simply don't want to spend time on boring installation, buy the instant Linux Lab Disk:

Linux Lab Disk

It supports:

  • Capacity
    • From 32G to 512G and even 1T, 2T, 4T
  • Products
    • High Speed U Disk, Solid U Disk, Portable disk, Solid disk (NVME / SATA)
  • Systems
    • Top6 Linux Distributions and even more based on your requirement
    • Include Ubuntu 18.04-22.04, Deepin 20.8+, Fedora 37+, Mint 21.1+, Kali, Manjaro
  • Features
    • Boot from any powered-off 64bit X86 Machine, include PC, Laptop and MacBook
    • Boot from any running Windows, Linux and run in parallel with them
    • Switch from or to any running Windows, Linux without poweroff
    • Multiple Linux Lab Disks can boot or switch from/to each other
    • Support timezone setting of different systems transparently, without manual setting
    • Share files and clipboards automatically between the main system and our disk system
    • Support transparent compress, use 128G as ~256G capacity
    • Support memory compiling, speedup compiling and save disk erase life
    • Support factory restore, allow restore factory system in some cases
    • Support volatile memory booting, allow read and write from memory, faster and longer
    • Merged in many labs, such as Linux Lab, Linux 0.11 Lab, be able to learn Linux kernel, embedded Linux, Uboot, Assembly, C, Python, Database, Network and so forth
  • Where to buy
  • Product details

2.1 Hardware and Software Requirement

Linux Lab is a full embedded Linux development system, it needs enough calculation capacity and disk & memory storage space, to avoid potential extension issues, here is the recommended configuration:

Hardware Requirement Description
Processor X86_64, > 1.5GHz Must choose 64bit X86 while using virtual machine
Disk >= 50G System (25G), Docker Images(~5G), Linux Lab (20G)
Memory >= 4G Lower than 4G may have many unpredictable exceptions

If often use, please increase disk storage to 100G~200G, memory storage to 8G, cpu cores to 4 and above.

Currently, all of the X86_64 systems support Docker should be able to run Linux Lab, include Windows, Linux and MacOS, all of the popular Linux distributions may have been tried by different users.

Welcome to take a look at the systems running Linux Lab and share yours, for example:

$ cd /path/to/cloud-lab
$ tools/docker/env
System: Ubuntu 16.04.6 LTS
Linux: 4.4.0-176-generic
Docker: Docker version 18.09.4, build d14af54

2.2 Docker Installation

Docker is required by Linux Lab, please install it at first:

  • Linux, Mac OSX, Windows 10

    Docker CE

  • older Windows (include some older Windows 10)

    Docker Toolbox; Install Ubuntu via Virtualbox or Vmware Virtual Machine

Before running Linux Lab, please refer to section 6.1.4 and make sure the following command works without sudo and without any issue:

$ docker run hello-world

In China, to use docker service normally, please must configure one of chinese docker mirror sites, for example:

More docker related issues, such as download slowly, download timeout and download errors, are cleary documented in the 6.1 section of FAQs.

The other issues, please read the official docker docs.

Notes for Ubuntu Users

Notes for Arch Users

Notes for Manjaro Users

Notes for Windows Users:

  • Please make sure your Windows version support docker: Official Docker Documentation and determine Docker Desktop or Docker Toolbox should be used

  • Linux Lab only tested with 'Git Bash' in Windows, please must use with it

    • After installing Git For Windows, "Git Bash Here" will come out in right-button press menu

2.3 Choose a working directory

Please simply choose one directory in ~/Downloads or ~/Documents or create a new ~/Develop directory.

$ mkdir ~/Develop
$ cd ~/Develop

For Windows and Mac OSX, to compile Linux normally, please refer to section 5.7.1 and enable building cache:

2.4 Switch to normal user

Before downloading Linux Lab, please MUST switch to normal user.

Check who am i, 0 means root, non-zero means normal user:

$ id -u `whoami`
1000

If current user is root, switch to a normal one:

# id -u `whoami`
0
# sudo -su <USER>

If no normal user exists, create new:

$ sudo useradd --create-home --shell /bin/bash --user-group --groups adm,sudo laber
$ sudo passwd laber
$ sudo -su laber
$ whoami
laber

2.5 Download the lab

Use Ubuntu system as an example:

Download cloud lab framework with normal user, pull images and checkout linux-lab repository:

$ git clone https://gitee.com/tinylab/cloud-lab.git
$ cd cloud-lab/

If cloned source code with root account, please MUST switch to normal user and change their owner:

$ sudo -su <USER>
$ sudo chown -R <USER>:<USER> -R cloud-lab/{*,.git}

2.6 Run and login the lab

Launch the lab and login with the user and password printed in the console:

$ tools/docker/run linux-lab

Login with Bash:

$ tools/docker/bash

Re-login the lab via web browser:

$ tools/docker/webvnc

The other login methods:

$ tools/docker/vnc
$ tools/docker/ssh
$ tools/docker/webssh

Choose one of the methods:

$ tools/docker/login list  # List, choose and record
$ tools/docker/login vnc  # Choose one directly and record for late login

Summary of login methods:

Login Method Description Default User Where
bash docker bash Ubuntu localhost
ssh normal ssh Ubuntu localhost
vnc normal vnc Ubuntu localhost+VNC client
webvnc web desktop Ubuntu anywhere via internet
webssh web ssh Ubuntu anywhere via internet

Since vnc clients differs from operating systems, we use webvnc by default to make sure auto login vnc for all systems.

If really want to use local vnc clients, please install a vnc client, for example: vinagre, then specify it like this:

$ tools/docker/vnc vinagre

If the above command not work normally, based on the information printed above, please configure the vnc client yourself.

Notes:

  • vinagre has fullscreen mode, but not enabled by default, which can be enabled through menu: View -> Fullscreen, but must enable Keyboard shortcuts at first, otherwise, no way exit fullscreen except via sudo pkill x11vnc.
  • The directly connected ssh and vnc may not always work, please use one of the other three methods instead.

2.7 Update and rerun the lab

Usually, only need to update Linux Lab itself, to get the new boards support or related fixups:

$ cd /path/to/cloud-lab/labs/linux-lab/
$ git checkout master
$ git pull

Sometimes, need to update Cloud Lab, to fix up potential running issues or getting newer docker image:

$ cd /path/to/cloud-lab
$ git checkout master
$ git pull

If modified the running environment of Linux Lab locally and want to reuse it in the future, save the container (very slow, not recommend if not necessary):

$ tools/docker/save linux-lab
$ git checkout -- configs/linux-lab/docker/name

Then rerurn Linux lab:

$ tools/docker/rerun linux-lab

2.8 Quickstart: Boot a board

Get into the lab environment, switch directory:

$ cd /labs/linux-lab

Issue the following command to boot the prebuilt kernel and rootfs on the default vexpress-a9 board:

$ make boot

Login as root user without password(password is empty), just input root and press Enter:

Welcome to Linux Lab

linux-lab login: root

# uname -a
Linux linux-lab 5.1.0 #3 SMP Thu May 30 08:44:37 UTC 2019 armv7l GNU/Linux
#
# poweroff
#

Shutdown the board with the poweroff command.

*Notes: If some boards not support poweroff, please press CTRL+a x. Of course, open another terminal and issue kill or pkill command also can quit qemu.

3. Linux Lab Kickstart

3.1 Using boards

3.1.1 List available boards

List builtin boards:

$ make list
[ aarch64/raspi3 ]:
      ARCH    = arm64
      CPU    ?= cortex-a53
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/mmcblk0 /dev/ram0
      ROOTDEV ?= /dev/mmcblk0
[ aarch64/virt ]:
      ARCH    = arm64
      CPU    ?= cortex-a57
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/sda /dev/vda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/vda
[ arm/mcimx6ul-evk ]:
      ARCH    = arm
      CPU    ?= cortex-a9
      LINUX  ?= v5.4
      ROOTDEV_LIST := /dev/mmcblk0 /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/mmcblk0
[ arm/versatilepb ]:
      ARCH    = arm
      CPU    ?= arm926t
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/sda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0
[ arm/vexpress-a9 ]:
      ARCH    = arm
      CPU    ?= cortex-a9
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/mmcblk0 /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0
[ i386/pc ]:
      ARCH    = x86
      CPU    ?= qemu32
      LINUX  ?= v5.1
      ROOTDEV_LIST ?= /dev/hda /dev/ram0 /dev/nfs
      ROOTDEV_LIST[LINUX_v2.6.34.9] ?= /dev/sda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/hda
[ mips64el/ls2k ]:
      ARCH    = mips
      CPU    ?= mips64r2
      LINUX  ?= loongnix-release-1903
      LINUX[LINUX_loongnix-release-1903] := 04b98684
      ROOTDEV_LIST := /dev/sda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0
[ mips64el/ls3a7a ]:
      ARCH    = mips
      CPU    ?= mips64r2
      LINUX  ?= loongnix-release-1903
      LINUX[LINUX_loongnix-release-1903] := 04b98684
      ROOTDEV_LIST ?= /dev/sda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0
[ mipsel/ls1b ]:
      ARCH    = mips
      CPU    ?= mips32r2
      LINUX  ?= v5.2
      ROOTDEV_LIST ?= /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0
[ mipsel/ls232 ]:
      ARCH    = mips
      CPU    ?= mips32r2
      LINUX  ?= v2.6.32-r190726
      ROOTDEV_LIST := /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0
[ mipsel/malta ]:
      ARCH    = mips
      CPU    ?= mips32r2
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/hda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0
[ ppc/g3beige ]:
      ARCH    = powerpc
      CPU    ?= generic
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/hda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0
[ riscv32/virt ]:
      ARCH    = riscv
      CPU    ?= any
      LINUX  ?= v5.0.13
      ROOTDEV_LIST := /dev/vda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/vda
[ riscv64/virt ]:
      ARCH    = riscv
      CPU    ?= any
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/vda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/vda
[ x86_64/pc ]:
      ARCH    = x86
      CPU    ?= qemu64
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/hda /dev/ram0 /dev/nfs
      ROOTDEV_LIST[LINUX_v3.2] := /dev/sda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0
[ csky/virt ]:
      ARCH    = csky
      CPU    ?= ck810
      LINUX  ?= v4.9.56
      ROOTDEV ?= /dev/nfs

ARCH, FILTER arguments are supported:

$ make list ARCH=arm
$ make list FILTER=virt

and more:

$ make list-board        # only ARCH
$ make list-short        # ARCH and Linux
$ make list-base          # no plugin
$ make list-plugin        # only plugin
$ make list-full          # everything
$ make list-real          # real hardware boards
$ make list-virt          # only virtual boards

3.1.2 Choosing a board

3.1.2.1 Real board

From version v0.6, to support learn external devices, Linux Lab adds real hardware board support, to use such boards, please buy them and connect them to your develop host correctly.

Only list real boards:

$ make list-real
[ arm/ebf-imx6ull ]:
  ARCH    = arm
  CPU    ?= cortex-a9
  LINUX  ?= v4.19.35
  ROOTDEV_LIST := /dev/mmcblk0 /dev/ram0 /dev/nfs
  ROOTDEV ?= /dev/mmcblk0

Because real hardware boards differs from each other, so, board specific document are recommended, for example: boards/arm/ebf-imx6ull/README.md.

Linux Lab Board - IMX6ULL

3.1.2.2 Virtual board

By default, the default virtual board: vexpress-a9 is used, we can configure, build and boot for a specific board with BOARD, for example:

$ make BOARD=malta
$ make boot

If several boards have the same name, please specify the architecture to distinguish:

$ make BOARD=mipsel/malta

Currently, such boards have the same name:

$ make list FILTER=virt
[ aarch64/virt ]:
      ARCH    = arm64
      CPU    ?= cortex-a57
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/sda /dev/vda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/vda
[ riscv32/virt ]:
      ARCH    = riscv
      CPU    ?= any
      LINUX  ?= v5.0.13
      ROOTDEV_LIST := /dev/vda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/vda
[ riscv64/virt ]:
      ARCH    = riscv
      CPU    ?= any
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/vda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/vda

$ make list FILTER=/pc
[ i386/pc ]:
      ARCH    = x86
      CPU    ?= qemu32
      LINUX  ?= v5.1
      ROOTDEV_LIST ?= /dev/hda /dev/ram0 /dev/nfs
      ROOTDEV_LIST[LINUX_v2.6.34.9] ?= /dev/sda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/hda
[ x86_64/pc ]:
      ARCH    = x86
      CPU    ?= qemu64
      LINUX  ?= v5.1
      ROOTDEV_LIST := /dev/hda /dev/ram0 /dev/nfs
      ROOTDEV_LIST[LINUX_v3.2] := /dev/sda /dev/ram0 /dev/nfs
      ROOTDEV ?= /dev/ram0

Use them like this:

$ make BOARD=x86_64/pc
$ make BOARD=riscv64/virt

If using board, it only works on-the-fly, the setting will not be saved, this is helpful to run multiple boards at the same and not to disrupt each other:

$ make board=malta boot

This allows to run multi boards in different terminals or background at the same time.

Check the board specific configuration:

$ cat boards/arm/vexpress-a9/Makefile

Notes: More money are required to maintain this project, only one virtual board is free now, the others are non-free, buy one as you want.

3.1.2.3 Buy one

All supported real hardware boards, virtual hardware boards support and the related Linux Lab Disk will be put in TinyLab.org's Taobao Shop or Bilibili Shop, after bought them, please contact with wechat: tinylab and join in the development group.

3.1.3 Using as plugins

The 'Plugin' feature is supported by Linux Lab, to allow boards being added and maintained in standalone git repositories. Standalone repository is very important to ensure Linux Lab itself not grow up big and big while more and more boards being added in.

Book examples or the boards with a whole new CPU architecture benefit from such feature a lot, for book examples may use many boards and a new CPU architecture may need require lots of new packages (such as cross toolchains and the architecture specific QEMU system tool).

Here maintains the available plugins:

The Loongson plugin has been merged into v5.0.

3.1.4 Configure boards

Every board has its own configuration, some can be changed on demand, for example, memory size, linux version, buildroot version, qemu version and the other external devices, such as serial port, network devices and so on.

The configure method is very simple, just edit it by referring to current values (boards/<BOARD>/Makefile), this command open local configuration (boards/<BOARD>/.labconfig) via vim:

$ make edit

But please don't make a big change once, we often only need to tune Linux version, this command is better for such case:

$ make list-linux
v4.12 v4.5.5 v5.0.10 [v5.1]
$ make config LINUX=v5.0.10
$ make list-linux
v4.12 v4.5.5 [v5.0.10] v5.1

If want to upstream your local changes, please use board-edit and board-config, otherwise, edit and config are preferrable, for they will avoid conflicts while pulling remote updates.

3.2 Build in one command

v0.3+ version add target dependency by default, so, if want to compile a kernel, just run:

$ make kernel-build

Or

$ make build kernel

It will do everything required, of course, we still be able to run the targets explicitly.

And futher, with the timestamping support, finished targets will not be run again during the late operations, if still want, just clean the stamp and run it again:

$ make cleanstamp kernel-build
$ make kernel-build

Or

$ make force-kernel-build

To clean all of the stamp files:

$ make cleanstamp kernel

This function also support uboot, root and qemu.

3.3 Detailed Operations

3.3.1 Downloading

Download board specific package and the kernel, buildroot source code:

$ make source APP=bsp,kernel,root,uboot
Or
$ make source APP=all
Or
$ make source all

Download one by one:

$ make bsp-source
$ make kernel-source
$ make root-source
$ make uboot-source

Or

$ make source bsp
$ make source kernel
$ make source root
$ make source uboot

After v0.5, the source code are downloaded in src/, before, they are saved in the root directory of Linux Lab.

Notes: the source code will be downloaded to build/src when CACHE_SRC or ONESHOT is configured to 1, please save or backup the data inside manually, otherwise, they will be lost after system poweroff.

3.3.2 Checking out

Checkout the target version of kernel and builroot:

$ make checkout APP=kernel,root

Checkout them one by one:

$ make kernel-checkout
$ make root-checkout

Or

$ make checkout kernel
$ make checkout root

If checkout not work due to local changes, save changes and run to get a clean environment:

$ make kernel-cleanup
$ make root-cleanup

Or

$ make cleanup kernel
$ make cleanup root

The same to QEMU and U-Boot.

3.3.3 Patching

Apply available patches in boards/<BOARD>/bsp/patch/linux and src/patch/linux/:

$ make kernel-patch

Or

$ make patch kernel

3.3.4 Configuration

3.3.4.1 Default Configuration

Configure kernel and buildroot with defconfig:

$ make defconfig APP=kernel,root

Configure one by one, by default, use the defconfig in boards/<BOARD>/bsp/:

$ make kernel-defconfig
$ make root-defconfig

Or

$ make defconfig kernel
$ make defconfig root

Configure with specified defconfig:

$ make B=raspi3
$ make kernel-defconfig bcmrpi3_defconfig
$ make root-defconfig raspberrypi3_64_defconfig

If only defconfig name specified, search boards/ at first, and then the default configs path of buildroot, u-boot and linux-stable respectivly: src/buildroot/configs, src/u-boot/configs, src/linux-stable/arch//configs.

3.3.4.2 Manual Configuration

$ make kernel-menuconfig
$ make root-menuconfig

Or

$ make menuconfig kernel
$ make menuconfig root

3.3.4.3 Old default configuration

$ make kernel-olddefconfig
$ make root-olddefconfig
$ make uboot-olddefconfig

Or

$ make olddefconfig kernel
$ make olddefconfig root
$ make olddefconfig uboot

3.3.5 Building

Build kernel and buildroot together:

$ make build APP=kernel,root

Build them one by one:

$ make kernel-build  # make kernel
$ make root-build    # make root

Or

$ make build kernel
$ make build root

After v0.5, the building result are stored in build/, before they are put in output/.

3.3.6 Saving

Save all of the configs and rootfs/kernel/dtb images:

$ make save APP=kernel,root
$ make saveconfig APP=kernel,root

Save configs and images to boards/<BOARD>/bsp/:

$ make kernel-saveconfig
$ make root-saveconfig
$ make root-save
$ make kernel-save

Or

$ make saveconfig kernel
$ make saveconfig root
$ make save kernel
$ make save root

3.3.7 Booting

Boot with serial port (nographic) by default, exit with CTRL+a x, poweroff, reboot or pkill qemu (See poweroff hang):

$ make boot

Boot with graphic (Exit with CTRL+ALT+2 quit):

$ make b=pc boot G=1 LINUX=v5.1 BUILDROOT=2019.11
$ make b=versatilepb boot G=1 LINUX=v5.1 BUILDROOT=2016.05
$ make b=g3beige boot G=1 LINUX=v5.1 BUILDROOT=2016.05
$ make b=malta boot G=1 LINUX=v2.6.36 BUILDROOT=2016.05
$ make b=vexpress-a9 boot G=1 LINUX=v4.6.7 BUILDROOT=2016.05 // LINUX=v3.18.39 works too

Note:

  • real graphic boot require LCD and keyboard drivers, the above boards work well, with Linux v5.1, raspi3 and malta has tty0 console but without keyboard input.

  • new buildroot config files set tty console to serial with (BR2_TARGET_GENERIC_GETTY_PORT="ttyAMA0"), to enable console with G=1, please change the getty line in /etc/inittab, for example, replace ttyAMA0 with console, we can also simply switch to the serial console via the Qemu 'View' menu.

vexpress-a9 and virt has no LCD support by default, but for the latest qemu, it is able to boot with G=1 and switch to serial console via the 'View' menu, this can not be used to test LCD and keyboard drivers. QOPTS specify the additional QEMU options.

$ make b=vexpress-a9 CONSOLE=ttyAMA0 boot G=1 LINUX=v5.1
$ make b=raspi3 CONSOLE=ttyAMA0 QOPTS="-serial vc -serial vc" boot G=1 LINUX=v5.1

Boot with curses graphic (friendly to bash/ssh login, not work for all boards, exit with ESC+2 quit or ALT+2 quit):

$ make b=pc boot G=2 LINUX=v4.6.7

Boot with PreBuilt Kernel, Dtb and Rootfs:

$ make boot kernel=old dtb=old root=old

Boot with new kernel, dtb and rootfs if exists:

$ make boot kernel=new dtb=new root=new

Boot with new kernel and uboot, build them if not exists:

$ make boot BUILD=kernel,uboot

Boot without Uboot (only versatilepb and vexpress-a9 boards tested):

$ make boot U=0

Boot with different rootfs (depends on board, check /dev/ after boot):

$ make boot ROOTDEV=ram0     // support by all boards, basic boot method
$ make boot ROOTDEV=nfs      // depends on network driver, only raspi3 not work
$ make boot ROOTDEV=sda
$ make boot ROOTDEV=mmcblk0
$ make boot ROOTDEV=vda      // virtio based block device

Boot with extra kernel command line (KCLI = Additional Kernel Command LIne):

$ make boot ROOTDEV=nfs KCLI="init=/bin/bash"

List supported options:

$ make list ROOTDEV
$ make list BOOTDEV
$ make list CCORI
$ make list NETDEV
$ make list linux
$ make list uboot
$ make list qemu

And more <xxx>-list are also supported with list <xxx>, for example:

$ make list features
$ make list modules
$ make list gcc

4. Linux Lab Advance

4.1 Using Linux Kernel

4.1.1 non-interactive configuration

A tool named scripts/config in Linux kernel is helpful to get/set the kernel config options non-interactively, based on it, both of kernel-getconfig and kernel-setconfig are added to tune the kernel options, with them, we can simply "enable/disable/setstr/setval/getstate" of a kernel option or many at the same time:

Get state of a kernel module:

$ make kernel-getconfig m=minix_fs
Getting kernel config: MINIX_FS ...

build/aarch64/linux-v5.1-virt/.config:CONFIG_MINIX_FS=m

Enable a kernel module:

$ make kernel-setconfig m=minix_fs
Setting kernel config: m=minix_fs ...

build/aarch64/linux-v5.1-virt/.config:CONFIG_MINIX_FS=m

Enable new kernel config: minix_fs ...

More control commands of kernel-setconfig including y, n, c, o, s, v:

Option Description
y build the modules in kernel or enable anther kernel options.
c build the modules as pluginable modules, just like m.
o build the modules as pluginable modules, just like m.
n disable a kernel option.
s RTC_SYSTOHC_DEVICE="rtc0", set the rtc device to rtc0
v PANIC_TIMEOUT=5, set the kernel panic timeout to 5 secs.

Operates many options in one command line:

$ make kernel-setconfig m=tun,minix_fs y=ikconfig v=panic_timeout=5 s=DEFAULT_HOSTNAME=linux-lab n=debug_info
$ make kernel-getconfig o=tun,minix,ikconfig,panic_timeout,hostname

4.1.2 using kernel modules

Build all internel kernel modules:

$ make modules
$ make modules-install
$ make root-rebuild    // not need for nfs boot
$ make boot

List available modules in src/modules/, boards/<BOARD>/bsp/modules/:

$ make modules-list

If m argument specified, list available modules in src/modules/, boards/<BOARD>/bsp/modules/ and src/linux-stable/:

$ make modules-list m=hello
    1      m=hello ; M=$PWD/src/modules/hello
$ make modules-list m=tun,minix
    1      c=TUN ; m=tun ; M=drivers/net
    2      c=MINIX_FS ; m=minix ; M=fs/minix

Enable one kernel module:

$ make kernel-getconfig m=minix_fs
Getting kernel config: MINIX_FS ...

build/aarch64/linux-v5.1-virt/.config:CONFIG_MINIX_FS=m

$ make kernel-setconfig m=minix_fs
Setting kernel config: m=minix_fs ...

build/aarch64/linux-v5.1-virt/.config:CONFIG_MINIX_FS=m

Enable new kernel config: minix_fs ...

Build one kernel module (e.g. minix.ko):

$ make modules M=fs/minix/
Or
$ make modules m=minix

Install and clean the module:

$ make modules-install M=fs/minix/
$ make modules-clean M=fs/minix/

More flexible usage:

$ make kernel-setconfig m=tun
$ make kernel tun.ko M=drivers/net
$ make kernel drivers/net/tun.ko

Build external kernel modules (the same as internel modules):

$ make modules m=hello
Or
$ make kernel x=$PWD/modules/hello/hello.ko

4.1.3 using kernel features

4.1.3.1 list supported kernel features

Kernel features are abstracted in src/feature/linux/, including their configurations patchset, it can be used to manage both of the out-of-mainline and in-mainline features.

$ make feature-list
[ /labs/linux-lab/src/feature/linux ]:
  + 9pnet
  + core
    - debug
    - module
  + ftrace
    - v2.6.36
      * env.g3beige
      * env.malta
      * env.pc
      * env.versatilepb
    - v2.6.37
      * env.g3beige
  + gcs
    - v2.6.36
      * env.g3beige
      * env.malta
      * env.pc
      * env.versatilepb
  + kft
    - v2.6.36
      * env.malta
      * env.pc
  + uksm
    - v2.6.38

Verified boards and Linux versions are recorded there, so, it should work without any issue if the environment not changed.

4.1.3.2 using kernel modules

For example, to enable kernel modules support, simply do:

// only upper case 'FEATURE' will be saved
$ make feature FEATURE=module
$ make kernel-olddefconfig
$ make kernel

4.1.3.3 using rust feature

Use x86_64/pc as an example:

$ make BOARD=x86_64/pc

switch to v6.1.1 Linux:

$ make config LINUX=v6.1.1

Compile the kernel, and test it with one of the simplest module - rust_minimal:

// clean up everything for a whole new test
$ make kernel-cleanall

// this 'f' variable will not be saved for standalone make targets
$ make test f=rust m=rust_minimal

4.1.3.4 using kft feature

For kft feature in v2.6.36 for malta board:

$ make cleanall b=malta
$ make test b=malta f=kft LINUX=v2.6.36

4.1.3.5 using rt feature

Linux officially provide RT Preemption support, but many patches are outside of mainline kernel, to use it:

$ make feature-list f=rt
$ make test b=i386/pc f=rt LINUX=v5.2

4.1.3.6 persist or clear feature setting

Clear feature setting (reset feature saved in .labconfig):

$ make feature FEATURE=rust
$ make feature FEATURE=

The above function is the same as 'make config'.

4.1.4 Create new development branch

If want to use a new development branch, please follow such steps:

At first, Get into src/linux-stable or another directory specified with KERNEL_SRC, checkout a development branch from a specific version:

$ cd src/linux-stable
$ git checkout -b linux-v5.1-dev v5.1

And then, clone the necessary configurations and directories for our new branch.

$ make kernel-clone LINUX=v5.1 LINUX_NEW=linux-v5.1-dev

The v5.1 must be the already supported version, if not, please use the near one in supported list, for example, i386/pc board support such versions:

$ make b=i386/pc list linux
v2.6.10 v2.6.11.12 v2.6.12.6 v2.6.21.5 v2.6.24.7 v2.6.34.9 v2.6.35.14 v2.6.36 v4.6.7 [v5.1] v5.2

If want to develop v2.6.38, please try to clone one from v2.6.36:

$ cd src/linux-stable
$ git checkout -b linux-v2.6.38-dev v2.6.38
$ make kernel-clone LINUX=v2.6.36 LINUX_NEW=linux-v2.6.38-dev

In development, please commit asap, and also, please use such commands carefully to avoid destroy your important changes:

  • kernel-checkout, checkout a specified kernel version, may override your changes
  • kernel-cleanup, clean up git repository, may remove your changes
  • kernel-clean, clean building history
  • kernel-cleanall, clean both of the building history and the source code changes

4.1.5 Use standalone git repository

v0.8 starts to add KERNEL_FORK, allows to configure the third party Linux source code repository, has added openEuler and wsl2, both of them support x86_64/pc and the former support aarch64/virt too.

For example, to compile wsl2 kernel, switch KERNEL_FORK to wsl2 directly:

$ make BOARD=x86_64/pc
$ make config KERNEL_FORK=wsl2
$ make kernel

To configure the wsl2 kernel version, configure it as following:

$ make edit
LINUX[KERNEL_FORK_wsl2]  := linux-msft-wsl-5.10.74.3

The value should be one of the available tag in git tag list.

4.2 Using U-Boot Bootloader

Choose one of the tested boards: versatilepb and vexpress-a9.

$ make BOARD=vexpress-a9

Download Uboot:

$ make uboot-source

Checkout the specified version:

$ make uboot-checkout

Patching with necessary changes, BOOTDEV and ROOTDEV available, use flash by default.

$ make uboot-patch

Use tftp, sdcard or flash explicitly, should run make U-Boot-checkout before a new uboot-patch:

$ make uboot-patch BOOTDEV=tftp
$ make uboot-patch BOOTDEV=sdcard
$ make uboot-patch BOOTDEV=flash

BOOTDEV is used to specify where to store and load the images for uboot, ROOTDEV is used to tell kernel where to load the rootfs.

Configure:

$ make uboot-defconfig
$ make uboot-menuconfig

Building:

$ make uboot

Boot with BOOTDEV and ROOTDEV, use flash by default:

$ make boot U=1

Use tftp, sdcard or flash explicitly:

$ make boot U=1 BOOTDEV=tftp
$ make boot U=1 BOOTDEV=sdcard
$ make boot U=1 BOOTDEV=flash

We can also change ROOTDEV during boot, for example:

$ make boot U=1 BOOTDEV=flash ROOTDEV=nfs

Clean images if want to update ramdisk, dtb and uImage:

$ make uboot-images-clean
$ make uboot-clean

Save U-Boot images and configs:

$ make uboot-save
$ make uboot-saveconfig

4.3 Using QEMU Emulator

Builtin QEMU may not work with the newest Linux kernel, so, we need compile and add external prebuilt qemu, this has been tested on vexpress-a9 and virt board.

At first, build qemu-system-ARCH:

$ make B=vexpress-a9
$ make qemu
$ make qemu-save

QEMU-ARCH-static and qemu-system-ARCH can not be compiled together. to build QEMU-ARCH-static, please enable QEMU_US=1 in board specific Makefile and rebuild it.

If QEMU and QTOOL specified, the one in bsp submodule will be used in advance of one installed in system, but the first used is the one just compiled if exists.

While porting to newer kernel, Linux 5.0 hangs during boot on QEMU 2.5, after compiling a newer QEMU 2.12.0, no hang exists. please take notice of such issue in the future kernel upgrade.

If already download QEMU and its submodules and don't want to upadte the submodules, just skip it:

$ make qemu git_module_status=0

4.4 Using Toolchains

The pace of Linux mainline is very fast, builtin toolchains can not keep up, to reduce the maintaining pressure, external toolchain feature is added. for example, ARM64/virt, CCVER and CCPATH has been added for it.

List available prebuilt toolchains:

$ make gcc-list

Download, decompress and enable the external toolchain:

$ make gcc

Switch compiler version if exists, for example:

$ make gcc-switch CCORI=internal GCC=4.8

$ make gcc-switch CCORI=linaro

If not external toolchain there, the builtin will be used back.

If no builtin toolchain exists, please must use this external toolchain feature, currently, aarch64, arm, riscv, mipsel, ppc, i386, x86_64 support such feature.

GCC version can be configured in board specific Makefile for Linux, Uboot, Qemu and Root, for example:

GCC[LINUX_v2.6.11.12] = 4.4

With this configuration, GCC will be switched automatically during defconfig and compiling of the specified Linux v2.6.11.12.

To build host tools, host gcc should be configured too(please specify b=i386/pc explicitly):

$ make gcc-list b=i386/pc
$ make gcc-switch CCORI=internal GCC=4.8 b=i386/pc

4.5 Using Rootfs

Builtin rootfs is minimal, is not enough for complex application development, which requires modern Linux distributions.

Such a type of rootfs has been introduced and has been released as docker image, ubuntu 18.04 is added for arm32v7 at first, more later.

Run it via docker directly:

$ docker run -it tinylab/arm32v7-ubuntu

Extract it out and run in Linux Lab:

(host)$ sudo apt-get install -y qemu-user-static

ARM32/vexpress-a9 (user: root, password: root):

(host)$ tools/root/docker/extract.sh tinylab/arm32v7-ubuntu arm
(lab )$ make boot b=arm/vexpress-a9 U=0 V=1 MEM=1024M ROOTDEV=nfs ROOTFS=$PWD/prebuilt/fullroot/tmp/tinylab-arm32v7-ubuntu

ARM64/raspi3 (user: root, password: root):

(host)$ tools/root/docker/extract.sh tinylab/arm64v8-ubuntu arm
(lab )$ make boot b=aarch64/virt V=1 ROOTDEV=nfs ROOTFS=$PWD/prebuilt/fullroot/tmp/tinylab-arm64v8-ubuntu

More rootfs from docker can be found:

$ docker search arm64 | egrep "ubuntu|debian"
arm64v8/ubuntu  Ubuntu is a Debian-based Linux operating system  25
arm64v8/debian  Debian is a Linux distribution that's composed  20

4.6 Debugging Linux and U-Boot

4.6.1 Debugging Linux

Compile the kernel with debugging options:

$ make feature FEATURE=debug
$ make kernel-olddefconfig
$ make kernel

Compile with one thread:

$ make kernel JOBS=1

And then debug it directly:

$ make debug

The above command will use tmux to split into two terminals, each running QEMU and gdb respectively, and load the script from .gdb/kernel.default.

To switch tmux panes, use CTRL+b followed by the arrow key (e.g., ←).

To customize kernel gdbinit script, simply copy one and edit it manually:

$ cp .gdb/kernel.default .gdb/kernel.user

It equals to:

$ make debug linux

to automate debug testing:

$ make test-debug linux

find out the code line of a kernel panic address:

$ make kernel-calltrace func+offset/length

if the debug port has been used, please try to find out who used the port and kill it:

$ sudo netstat -tlp | grep 1234
tcp        0      0 0.0.0.0:1234            0.0.0.0:*              LISTEN      3943/qemu-xxx
$ sudo kill -9 3943

4.6.2 Debugging U-Boot

To debug U-Boot with .gdb/uboot.default:

$ make debug uboot

The above command will use tmux to split into two terminals, each running QEMU and gdb respectively.

To switch tmux panes, use CTRL+b followed by the arrow key (e.g., ←).

To automate U-Boot debug testing:

$ make test-debug uboot

The same to kernel gdbinit script, customize one for uboot:

$ cp .gdb/uboot.default .gdb/uboot.user

4.7 Test Automation

Use aarch64/virt as the demo board here.

$ make BOARD=virt

Prepare for testing, install necessary files/scripts in src/system/:

$ make rootdir
$ make root-rebuild

Simply boot and poweroff (See poweroff hang):

$ make test

Don't poweroff after testing:

$ make test TEST_FINISH=echo

Run guest test case:

$ make test TEST_CASE=/tools/ftrace/trace.sh

Run guest test cases (COMMAND_LINE_SIZE must be big enough, e.g. 4096, see cmdline_size feature below):

$ make test TEST_BEGIN=date TEST_END=date TEST_CASE='ls /;echo hello world'

Reboot the guest system for several times:

$ make test TEST_REBOOT=2

NOTE: reboot may 1) hang, 2) continue; 3) timeout killed, TEST_TIMEOUT=30; 4) timeout continue, TIMEOUT_CONTINUE=1

Test a feature of a specified Linux version on a specified board (cmdline_size feature is for increase COMMAND_LINE_SIZE to 4096):

$ make test f=kft LINUX=v2.6.36 b=malta TEST_PREPARE=board-init,kernel-cleanup

NOTE: board-init and kernel-cleanup make sure test run automatically, but kernel-cleanup is not safe, please save your code before use it!!

Test a kernel module:

$ make test m=hello

Test multiple kernel modules:

$ make test m=exception,hello

Test modules with specified ROOTDEV, nfs boot is used by default, but some boards may not support network:

$ make test m=hello,exception TEST_RD=ram0

Run test cases while testing kernel modules (test cases run between insmod and rmmod):

$ make test m=exception TEST_BEGIN=date TEST_END=date TEST_CASE='ls /root;echo hello world' TEST_PREPARE=board-init,kernel-cleanup f=cmdline_size

Run test cases while testing internal kernel modules:

$ make kernel-setconfig y=debug_fs
$ make test m=lkdtm TEST_BEGIN='mount -t debugfs debugfs /mnt' TEST_CASE='echo EXCEPTION > /mnt/provoke-crash/DIRECT'

Run test cases while testing internal kernel modules, pass kernel arguments:

$ make test m=lkdtm lkdtm_args='cpoint_name=DIRECT cpoint_type=EXCEPTION'

Run test without feature-init (save time if not necessary):

$ make test m=lkdtm lkdtm_args='cpoint_name=DIRECT cpoint_type=EXCEPTION' TEST_INIT=0
Or
$ make raw-test m=lkdtm lkdtm_args='cpoint_name=DIRECT cpoint_type=EXCEPTION'

Run test with module and the module's necessary dependencies (check with make kernel-menuconfig):

$ make test m=lkdtm y=runtime_testing_menu,debug_fs lkdtm_args='cpoint_name=DIRECT cpoint_type=EXCEPTION' LINUX=v5.1 TEST_PREPARE=kernel-cleanup

Run test without feature-init, boot-init, boot-finish and no TEST_PREPARE:

$ make boot-test m=lkdtm lkdtm_args='cpoint_name=DIRECT cpoint_type=EXCEPTION'

Test a kernel module and make some targets before testing:

$ make test m=exception TEST=kernel-checkout,kernel-patch,kernel-defconfig

Test everything in one command (from download to poweroff, see poweroff hang):

$ make test TEST=kernel,root TEST_PREPARE=board-init,kernel-cleanup,root-cleanup

Test everything in one command (with U-Boot while support, e.g. vexpress-a9):

$ make test TEST=kernel,root,uboot TEST_PREPARE=board-init,kernel-cleanup,root-cleanup,uboot-cleanup

Test kernel hang during boot, allow to specify a timeout, timeout must happen while system hang:

$ make test TEST_TIMEOUT=30s

Test kernel debug:

$ make test DEBUG=1

Notes: The above tests may fail on some boards with some Linux versions, please upgrade the kernel versions if necessary.

4.8 File Sharing

To transfer files between QEMU Board and Host, three methods are supported by default:

4.8.1 Install files to rootfs

Simply put the files with a relative path in src/system/, install and rebuild the rootfs:

$ mkdir src/system/root/
$ touch src/system/root/new_file
$ make root-rebuild
$ make boot

4.8.2 Share with NFS

Boot the board with ROOTDEV=nfs:

$ make boot ROOTDEV=nfs

Host:

$ make env-dump VAR=ROOTDIR
ROOTDIR="/labs/linux-lab/boards/<BOARD>/bsp/root/<BUILDROOT_VERSION>/rootfs"

4.8.3 Transfer via tftp

Using tftp server of host from the QEMU board with the tftp command.

Host:

$ ifconfig br0
inet addr:172.17.0.3  Bcast:172.17.255.255  Mask:255.255.0.0
$ cd tftpboot/
$ ls tftpboot
kft.patch kft.log

QEMU Board:

$ ls
kft_data.log
$ tftp -g -r kft.patch 172.17.0.3
$ tftp -p -r kft.log -l kft_data.log 172.17.0.3

Note: while put file from QEMU board to host, must create an empty file in host firstly. Buggy?

4.8.4 Share with 9p virtio

To enable 9p virtio for a new board, please refer to qemu 9p setup. qemu must be compiled with --enable-virtfs, and kernel must enable the necessary options.

Reconfigure the kernel with:

CONFIG_NET_9P=y
CONFIG_NET_9P_VIRTIO=y
CONFIG_NET_9P_DEBUG=y (Optional)
CONFIG_9P_FS=y
CONFIG_9P_FS_POSIX_ACL=y
CONFIG_PCI=y
CONFIG_VIRTIO_PCI=y
CONFIG_PCI_HOST_GENERIC=y (only needed for the QEMU Arm 'virt' board)

If using -virtfs or -device virtio-9p-pci option for qemu, must enable the above PCI related options, otherwise will not work:

9pnet_virtio: no channels available for device hostshare
mount: mounting hostshare on /hostshare failed: No such file or directory

-device virtio-9p-device requires less kernel options.

To enable the above options, please simply type:

$ make feature FEATURE=9pnet
$ make kernel-olddefconfig

Docker host:

$ modprobe 9pnet_virtio
$ lsmod | grep 9p
9pnet_virtio          17519  0
9pnet                  72068  1 9pnet_virtio

Host:

$ make BOARD=virt

$ make root-rebuild

$ touch hostshare/test    # Create a file in host

$ make boot U=0 ROOTDEV=ram0 PBR=1 SHARE=1

$ make boot SHARE=1 SHARE_DIR=src/modules  # for external modules development

$ make boot SHARE=1 SHARE_DIR=build/aarch64/linux-v5.1-virt/  # for internal modules learning

$ make boot SHARE=1 SHARE_DIR=src/examples  # for c/assembly learning

QEMU Board:

$ ls /hostshare/      # Access the file in guest
test
$ touch /hostshare/guest-test  # Create a file in guest

Verified boards with Linux v5.1:

boards Status
aarch64/virt virtio-9p-device (virtio-9p-pci breaks nfsroot)
arm/vexpress-a9 only work with virtio-9p-device and without U-Boot booting
arm/versatilepb only work with virtio-9p-pci
x86_64/pc only work with virtio-9p-pci
i386/pc only work with virtio-9p-pci
riscv64/virt work with virtio-9p-pci and virtio-9p-dev
riscv32/virt work with virtio-9p-pci and virtio-9p-dev

4.9 Learning Assembly

Linux Lab has added many assembly examples in src/examples/assembly:

$ cd src/examples/assembly
$ ls
aarch64 arm mips64el mipsel powerpc powerpc64 riscv32 riscv64 x86 x86_64
$ make -s -C aarch64/
Hello, ARM64!

4.10 Learning C

4.10.1 Host build and Run

Use hello as example:

$ cd src/examples/c/hello
$ make
gcc -fno-stack-protector -fomit-frame-pointer -fno-asynchronous-unwind-tables -fno-pie -no-pie -m32 -Wall -Werror -g -o hello hello.c
Hello, World!

4.10.2 Cross build and Run

Use X32 (Code for x86-64, int/long/pointer to 32bits), ARM, MIPS, PPC and RISC-V as example:

$ sudo apt-get update -y

$ sudo apt-get install -y libc6-x32 libc6-dev-x32 libx32gcc-8-dev
$ gcc -mx32 -o hello hello.c
$ ./hello
Hello, World!

$ sudo apt-get install -y libc6-dev-armel-cross libc6-armel-cross
$ arm-linux-gnueabi-gcc -o hello hello.c
$ qemu-arm -L /usr/arm-linux-gnueabi/ ./hello
Hello, World!

$ sudo apt-get install -y libc6-dev-mipsel-cross libc6-mipsel-cross
$ mipsel-linux-gnu-gcc -o hello hello.c
$ qemu-mipsel -L /usr/mipsel-linux-gnu/ ./hello
Hello, World!

$ sudo apt-get install -y libc6-dev-powerpc-cross libc6-powerpc-cross
// Must use -static for Linux Lab v0.6, otherwise, there will be segmentation fault
$ powerpc-linux-gnu-gcc -static -o hello hello.c
$ qemu-ppc -L /usr/powerpc-linux-gnu/ ./hello
Hello, World!

$ sudo apt-get install -y libc6-riscv64-cross libc6-dev-riscv64-cross
$ riscv64-linux-gnu-gcc -o hello hello.c
$ qemu-riscv64 -L /usr/riscv64-linux-gnu/ ./hello
Hello, World!

Above run through qemu-user, to run on target boards, please copy the binaries to target boards' rootfs with help from section 4.8.1.

The main packages are libc6-dev, libc6 or libgcc, but x32 is an expection, it is libx32gcc. please list them via apt-cache search.

4.11 Running any make goals

Linux Lab allows to access Makefile goals of the APPS easily, for example:

$ make kernel help
$ make kernel menuconfig

$ make root help
$ make root busybox-menuconfig

$ make uboot help
$ make uboot menuconfig

Or

$ make kernel-help
$ make kernel-menuconfig

$ make root-help
$ make root-busybox-menuconfig

$ make uboot-help
$ make uboot-menuconfig

Allows to run sub-make goals of kernel, root and U-Boot directly without entering into their own building directory.

4.12 Speed up kernel development

4.12.1 Speed up compiling and save disk life

Notes:This operation may lose data, please take care!

This feature aims to create a ram based temporary filesystem as the 'build' directory, to store the building data, If not backup them, they will be lost after shutting down the machine.

Create temporary building cache:

$ make build cache

Check the status of building cache:

$ make build status

Use the cache for building speedup:

$ time make kernel

Backup the cache to a persistent file (If the building file are important to you):

$ make build backup

Stop the building cache, revert back to use the build directory on the disk:

$ make build uncache

Use the backup as the build directory:

$ sudo mount /path/to/backup-file /labs/linux-lab/build/

4.12.2 ONESHOT Mode

v0.9 adds a ONESHOT switch, it can be used to enable such functions:

  • Auto cache build/ in memory
  • Auto cache src/ in memory
  • Auto enable fast fetch, a.k.a git shallow fetch

It is good for:

  • Disposable, destroy after using

    • If want, please save kernel and its config with kernel-save and kernel-saveconfig
  • Better for big-memory, small-disk and slow-CPU host machines

    • Both src/ and build/ are put in memory, not in disk
  • Good for instant kernel downloading and building

    • If target host has no Linux kernel source code, and the network is slow

To use it, please simply run this before others:

$ export ONESHOT=1

If want to make it persistent, just configure it in .labinit:

ONESHOT := 1

4.12.3 Nolibc Mode

v1.2-rc2 adds Nolibc mode, allows to build ultra small kernel and application, and package them together via initrd, to achieve "Kernel-only" deployments.

Nolibc adds two types of files:

  • Small kernel config file: boards/<ARCH>/<BOARD>/bsp/configs/linux_v6.x_nolibc_defconfig
  • Small nolibc application: src/examples/nolibc/hello.c

Just similar to ONESHOT, before developing, just run this to enable NOLIBC mode:

$ export NOLIBC=1

Or, write to .labinit to let it always work:

NOLIBC := 1

To change the target nolibc aplication, we can configure NOLIBC_SRC, otherwise, the above hello.c will be used by default:

$ make nolibc-clean
$ make kernel NOLIBC_SRC=$PWD/src/examples/nolibc/hello.c

It is very good for pure kernel development.

4.12.4 Tiny Mode

Based on Nolibc mode, v1.4-rc2 adds Tiny mode, allows to build ultra small kernel but with initrd boot support.

Usage:

$ export KCFG=linux.tiny.config
$ make kernel
$ make boot ROOTDEV=ram0

Compare to defconfig, it only enables minimal config options and makes sure initrd boot with an interactive shell, so, the compiling speed is x10 faster.

It is very good for kernel features testing, development and research.

4.13 More Usage

Read more:

5. Linux Lab Development

This introduces how to add a new board for Linux Lab.

5.1 Choose a board supported by QEMU

list the boards, use arm as an example:

$ qemu-system-arm -M ?

5.2 Create the board directory

Use vexpress-a9 as an example:

$ mkdir boards/arm/vexpress-a9/

5.3 Clone a Makefile from an existing board

Use versatilepb as an example:

$ cp boards/arm/versatilebp/Makefile boards/arm/vexpress-a9/Makefile

5.4 Configure the variables from scratch

Comment everything, add minimal ones and then others.

Please refer to doc/qemu/qemu-doc.html or the online one https://www.qemu.org/docs/master/.

5.5 At the same time, prepare the configs

We need to prepare the configs for linux, buildroot and even uboot.

Buildroot has provided many examples about buildroot and kernel configuration:

buildroot: src/buildroot/configs/qemu_ARCH_BOARD_defconfig
kernel: src/buildroot/board/qemu/ARCH-BOARD/linux-VERSION.config

U-Boot has also provided many default configs:

uboot: src/u-boot/configs/vexpress_ca9x4_defconfig

Kernel itself also:

kernel: src/linux-stable/arch/arm/configs/vexpress_defconfig

Linux Lab itself also provide many working configs too, the xxx-clone target is a good helper to utilize existing configs:

$ make list kernel
v4.12 v5.0.10 v5.1
$ make kernel-clone LINUX=v5.1 LINUX_NEW=v5.4
$ make kernel-menuconfig
$ make kernel-saveconfig

$ make list root
2016.05 2019.02.2
$ make root-clone BUILDROOT=2019.02.2 BUILDROOT_NEW=2019.11
$ make root-menuconfig
$ make root-saveconfig

Edit the configs and Makefile until they match our requirements.

$ make kernel-menuconfig
$ make root-menuconfig
$ make board-edit

The configuration must be put in boards/<BOARD>/ and named with necessary version info, use raspi3 as an example:

$ make kernel-saveconfig
$ make root-saveconfig
$ ls boards/aarch64/raspi3/bsp/configs/
buildroot_2019.02.2_defconfig  linux_v5.1_defconfig

2019.02.2 is the buildroot version, v5.1 is the kernel version, both of these variables should be configured in boards/<BOARD>/Makefile.

More usage about the xxx-clone commands:

$ make qemu-clone QEMU=<old_version> QEMU_NEW=<new_version>
$ make uboot-clone UBOOT=<old_version> UBOOT_NEW=<new_version>
$ make kernel-clone LINUX=<old_version> LINUX_NEW=<new_version>
$ make root-clone BUILDROOT=<old_version> BUILDROOT_NEW=<new_version>

5.6 Choose the versions of kernel, rootfs and U-Boot

Please use tag instead of branch, use kernel as an example:

$ cd src/linux-stable
$ git tag
...
v5.0
...
v5.1
..
v5.1.1
v5.1.5
...

If want v5.1 kernel, just put a line "LINUX = v5.1" in boards/<BOARD>/Makefile.

Or clone a kernel config from the old one or the official defconfig:

$ make kernel-clone LINUX_NEW=v5.3 LINUX=v5.1

Or

$ make B=i386/pc
$ pushd Linux-stable && git checkout v5.4 && popd
$ make kernel-clone LINUX_NEW=v5.4 KCFG=i386_defconfig

If no tag existed, a virtual tag name with the real commmit number can be configured as following:

LINUX = v2.6.11.12
LINUX[LINUX_v2.6.11.12] = 8e63197f

Linux version specific ROOTFS are also supported:

ROOTFS[LINUX_v2.6.12.6]  ?= $(BSP_ROOT)/$(BUILDROOT)/rootfs32.cpio.gz

5.7 Configure, build and boot them

Use kernel as an example:

$ make kernel-defconfig
$ make kernel-menuconfig
$ make kernel
$ make boot

The same to Rootfs, U-Boot and even QEMU.

5.8 Save the images and configs

$ make root-save
$ make kernel-save
$ make uboot-save

$ make root-saveconfig
$ make kernel-saveconfig
$ make uboot-saveconfig

5.9 Upload everything

At last, upload the images, defconfigs, patchset to board specific bsp submodule repository.

Firstly, get the remote bsp repository address as following:

$ git remote show origin
* remote origin
  Fetch URL: https://gitee.com/tinylab/qemu-aarch64-raspi3/
  Push  URL: https://gitee.com/tinylab/qemu-aarch64-raspi3/
  HEAD branch: master
  Remote branch:
    master tracked
  Local branch configured for 'git pull':
    master merges with remote master
  Local ref configured for 'git push':
    master pushes to master (local out of date)

Then, fork this repository from gitee.com, upload your changes, and send your pull request.

6. FAQs

6.1 Docker Issues

6.1.1 Speed up docker images downloading

To optimize docker images download speed, please refer to section 6.1.6.

6.1.2 Docker network conflicts with LAN

Cloud Lab use a default 172.20.0.0/16 subnet, if this conflicts with another one, please change it like this:

$ tools/docker/rm-all
$ vim configs/linux-lab/docker/subnet
$ cat configs/linux-lab/docker/subnet
172.23.0.0/16
$ tools/docker/run linux-lab

If lab network still not work, please try another private network address and eventually to avoid conflicts with LAN address.

6.1.3 Why not allow running Linux Lab in local host

The full function of Linux Lab depends on the full docker environment managed by Cloud Lab, so, please really never try and therefore please don't complain about why there are lots of packages missing failures and even the other weird issues.

Linux Lab is designed to use pre-installed environment with the docker technology and save our life by avoiding the packages installation issues in different systems, so, Linux Lab would never support local host using even in the future.

6.1.4 Run tools without sudo

To use the tools under tools without sudo, please make sure add your account to the docker group and reboot your system to take effect:

$ sudo usermod -aG docker <USER>
$ newgrp docker

If get error: "newgrp: group 'docker' does not exist", please add 'docker' group manually:

$ sudo groupadd docker

Notes: Currently, root user is not allowed to run Linux Lab。

6.1.5 Network not work

If ping not work, please check one by one:

  • DNS issue

    if ping 8.8.8.8 work, please check /etc/resolv.conf and make sure it is the same as your host configuration.

  • IP issue

    if ping not work, please refer to network conflict issue and change the ip range of docker containers.

6.1.6 Client.Timeout exceeded while waiting headers

This means must configure one of the following docker mirror sites:

Potential methods of configuration in Ubuntu, depends on docker and Ubuntu versions:

/etc/default/docker:

echo "DOCKER_OPTS=\"\$DOCKER_OPTS --registry-mirror=<your accelerate address>\""

/lib/systemd/system/docker.service:

ExecStart=/usr/bin/dockerd -H fd:// --registry-mirror=<your accelerate address>

/etc/docker/daemon.json:

{
    "registry-mirrors": ["<your accelerate address>"]
}

Please restart docker service after change the accelerate address:

$ sudo service docker restart

For the other Linux systems, Windows and macOS System, please refer to Aliyun Mirror Speedup Document.

IF still slow, please check if the mirror site is configured normally and without typos:

$ docker info | grep -A1 -i Mirrors
Registry Mirrors:
  https://XXXXX.mirror.aliyuncs.com/

6.1.7 Restart Linux Lab after host system shutdown or reboot

If want to restore the installed softwares and related configurations, please save the container manually:

$ tools/docker/save linux-lab

After host system (include virtual machine) shutdown or reboot, you can restart the lab via the "Linux Lab" icon on the desktop, or just like before, issue this command:

$ tools/docker/run linux-lab

Current implementation doesn't support the direct 'docker start' command, please learn it.

If the above methods still not restart the lab, please refer to the methods mentioned in the 6.3.9 section.

If resume from a suspended host system, the lab will restore automatically, no need to do anything to restart it, just use one of the 4 login methods mentioned in the 2.4 section, for example, start a web browser to connect it:

$ tools/docker/webvnc

6.1.8 the following directives are specified both as a flag and in the configuration file

If getting such error:

unable to configure the Docker daemon with file /etc/docker/daemon.json: the
following directives are specified both as a flag and in the configuration
file: registry-mirrors: (from flag: [https://docker.mirrors.ustc.edu.cn/], from
file: [https://xxx.mirror.aliyuncs.com])

Means both /etc/docker/daemon.json and /etc/default/docker configured registry-mirrors, please comment the late one and restart docker:

$ sudo service docker restart

6.1.9 pathspec FETCH_HEAD did not match any file known to git

If get such error while running make boot, it means network issue, please refer to section 6.1.5。

Could not resolve host: gitee.com
error: pathspec 'FETCH_HEAD' dit not match any file(s) known to git

6.1.10 Docker not work in Ubuntu 20.04

If docker not work in Ubuntu 20.04, please use doc/install/daemon.json and clean up the arguments of dockerd, learn more from docker daemon:

$ sudo cat /etc/systemd/system/docker.service.d/docker.conf
[Service]
ExecStart=
ExecStart=/usr/bin/dockerd

$ sudo cp /etc/docker/daemon.json /etc/docker/daemon.json.bak
$ sudo cp doc/install/daemon.json /etc/docker/
$ sudo service docker restart

Please make sure using the best registry-mirrors for better download speed.

6.1.11 Error creating aufs mount

If not work with failure like "error creating aufs mount to ... invalid arguments", that means the storage driver used by docker is not supported by current system, please choose another one from this page, and configure it in /etc/docker/daemon.json, for example:

$ sudo vim /etc/docker/daemon.json
{
  "registry-mirrors": ["https://docker.mirrors.ustc.edu.cn"],
  "storage-driver": "devicemapper"
}

This issue is related to kernel version, the same system may upgrade kernel version and therefore support different storage driver.

6.2 QEMU Issues

6.2.1 Why kvm speedding up is disabled

kvm only supports both of qemu-system-i386 and qemu-system-x86_64 currently, and it also requires the CPU and bios support, otherwise, you may get this error log:

modprobe: ERROR: could not insert 'kvm_intel': Operation not supported

Check CPU virtualization support, if nothing output, then, cpu not support virtualization:

$ cat /proc/cpuinfo | egrep --color=always "vmx|svm"

If CPU supports, we also need to make sure it is enabled in bios features, simply reboot your computer, press 'Delete' to enter bios, please make sure the 'Intel virtualization technology' feature is 'enabled'.

6.2.2 Poweroff hang

Both of the poweroff and reboot commands not work on these boards currently (LINUX=v5.1):

  • mipsel/malta (exclude LINUX=v2.6.36)
  • mipsel/ls232
  • mipsel/ls1b
  • mips64el/ls2k
  • mips64el/ls3a7a
  • aarch64/raspi3
  • arm/versatilepb

System will directly hang there while running poweroff or reboot, to exit qemu, please pressing CTRL+a x or using pkill qemu.

To test such boards automatically, please make sure setting TEST_TIMEOUT, e.g. make test TEST_TIMEOUT=50.

Welcome to fix up them.

6.2.3 How to exit QEMU

Where How
Serial Port Console CTRL+a x
Curses based Graphic ESC+2 quit Or ALT+2 quit
X based Graphic CTRL+ALT+2 quit
Generic Methods poweroff, reboot, kill, pkill

6.2.4 Boot with missing sdl2 libraries failure

That's because the docker image is not updated, just enter into cloud-lab and rerun the lab (please must not use tools/docker/restart here for it not using the new docker image):

$ tools/docker/rerun linux-lab

6.3 Environment Issues

6.3.1 NFS/tftpboot not work

If nfs or tftpboot not work, please run modprobe nfsd in host side and restart the net services via /configs/tools/restart-net-servers.sh in guest side and please make sure not use tools/docker/trun.

6.3.2 How to switch Windows in VIM

CTRL+w is used in both of browser and vim, to switch from one window to another, please use CTRL+Left or CTRL+Right key instead, Linux Lab has remapped CTRL+Right to CTRL+w and CTRL+Left to CTRL+p.

6.3.3 How to delete typo in shell command line

Long keypress not work in novnc client currently, so, long Delete not work, please use alt+delete or alt+backspace instead, more tips:

Function VIM Bash
begin/end ^/$ Ctrl + a/e
forward/backward w/b Ctrl + Home/end
cut one word backword db Alt + Delete/backspace
cut one word forward dw Alt + d
cut all to begin d^ Ctrl + u
cut all to end d$ Ctrl + k
paste all cutted p Ctrl + y

6.3.4 Language input switch shortcuts

In order to switch English/Chinese input method, please use CTRL+s shortcuts, it is used instead of CTRL+space to avoid conflicts with local system.

6.3.5 How to tune the screen size

There are tow methods to tune the screen size, one is auto scaling by noVNC, another is pre-setting during launching.

The first one is setting noVNC before connecting.

  • Press the left sidebar of noVNC web page
  • Disconnect
  • Enable 'Auto Scaling Mode' via 'Settings -> Scaling Mode: -> Local Scaling -> Apply'
  • Connect

The second one is setting SCREEN_SIZE while running Linux Lab.

The screen size of lab is captured by xrandr, if not work, please check and set your own, for example:

Get available screen size values:

$ xrandr --current
Screen 0: minimum 1 x 1, current 1916 x 891, maximum 16384 x 16384
Virtual1 connected primary 1916x891+0+0 (normal left inverted right x axis y axis) 0mm x 0mm
  1916x891      60.00*+
  2560x1600    59.99
  1920x1440    60.00
  1856x1392    60.00
  1792x1344    60.00
  1920x1200    59.88
  1600x1200    60.00
  1680x1050    59.95
  1400x1050    59.98
  1280x1024    60.02
  1440x900      59.89
  1280x960      60.00
  1360x768      60.02
  1280x800      59.81
  1152x864      75.00
  1280x768      59.87
  1024x768      60.00
  800x600      60.32
  640x480      59.94

Update remote screen size:

$ cd /path/to/cloud-lab
$ tools/docker/resize 1280x1024  # Specifiy anyone above
$ tools/docker/resize            # If no argument, Sync with host system

If want fullscreen, follow these steps:

  1. If using virtual machine, fullscreen virtual machine at fist
  2. Run tools/docker/resize to resize remote lab screen size
  3. Enter into WebVNC Interface, Click the FullScreen button at the left sidebar

6.3.6 How to work in fullscreen mode

Open the left sidebar, press the 'Fullscreen' button.

6.3.7 How to record video

  • Enable recording

    Open the left sidebar, press the 'Settings' button, config 'File/Title/Author/Category/Tags/Description' and enable the 'Record Screen' option.

  • Start recording

    Press the 'Connect' button.

  • Stop recording

    Press the 'Disconnect' button.

  • Replay recorded video

    Press the 'Play' button.

  • Share it

    Videos are stored in 'cloud-lab/recordings', share it with help from showdesk.io.

6.3.8 Linux Lab not response

The VNC connection may hang for some unknown reasons and therefore Linux Lab may not response sometimes, to restore it, please press the flush button of web browser or re-connect after explicitly disconnect.

6.3.9 VNC login with failures

If VNC login return "Disconnect timeout", wait a while and press the left 'Connect' button again, otherwise, check as following:

At first, check the containers' status (Up: Ok, Exit: Bad):

$ docker ps -a
CONTAINER ID IMAGE COMMAND CREATED STATUS PORTS NAMES
19a61ba075b5 tinylab/linux-lab "/tools/lab/run" 4 days ago Up 4 days 22/tcp, 5900/tcp linux-lab-21575
75dae89984c9 tinylab/cloud-ubuntu-web "/startup.sh" 8 days ago Up 8 days ....443/tcp cloud-ubuntu-web

If the status is 'Exit', that means container may be shutdown or may never up, run it again to resume for the shutdown case:

$ tools/docker/run linux-lab

Otherwise, check the running logs:

$ tools/docker/logs linux-lab

If normal, that means the login account and password may have been invalid for some exceptions, please regenerte new account and password with the coming steps:

Note: The clean command will remove some containers and data, please do necessary backup before run it, for example, save the container:

$ tools/docker/save linux-lab

VNC login fails while using mismatched password, to fix up such issue, please clean up all and rerun it:

$ tools/docker/clean linux-lab
$ tools/docker/rerun linux-lab

If the above command not work, please try this one (It will clean more data, please do necessary backup)

$ tools/docker/clean-all
$ tools/docker/rerun linux-lab

6.3.10 Ubuntu Snap Issues

Users report many snap issues, please use apt-get instead:

  • users can not be added to docker group and break non-root operation.
  • snap service exhausts the /dev/loop devices and break mount operation.

6.3.11 How to exit fullscreen mode of vnc clients

The easiest method is kill the VNC server in Linux Lab:

$ sudo pkill x11vnc

6.4 Lab Issues

6.4.1 No working init found

This means the rootfs.ext2 image may be broken, please remove it and try make boot again, for example:

$ rm boards/aarch64/raspi3/bsp/root/2019.02.2/rootfs.ext2
$ make boot

make boot command can create this image automatically.

6.4.2 linux/compiler-gcc7.h: No such file or directory

This means using a newer gcc than the one Linux kernel version supported, the solution is switching to an older gcc version via make gcc-switch, use i386/pc board as an example:

$ make gcc-list
$ make gcc-switch CCORI=internal GCC=4.4

6.4.3 linux-lab/configs: Permission denied

This may happen at make boot while the repository is cloned with root user, please simply update the owner of cloud-lab/ directory:

$ cd /path/to/cloud-lab
$ sudo chown <USER>:<USER> -R ./
$ tools/docker/rerun linux-lab

Notes: To make a consistent working environment, Linux Lab only support using as general user: 'ubuntu'.

6.4.4 scripts/Makefile.headersinst: Missing UAPI file

This means MAC OSX not use Case sensitive filesystem, create one using hdiutil or Disk Utility yourself:

$ hdiutil create -type SPARSE -size 60g -fs "Case-sensitive Journaled HFS+" -volname labspace labspace.dmg
$ hdiutil attach -mountpoint ~/Develop/labspace -nobrowse labspace.dmg.sparseimage
$ cd ~/Develop/labspace

6.4.5 unable to create file: net/netfilter/xt_dscp.c

This means Windows not enable filesystem's case sensitive feature, just enable it:

$ cd /path/to/cloud-lab
$ fsutil file SetCaseSensitiveInfo ./ enable

6.4.6 how to run as root

By default, no password required to run as root with:

$ sudo -s

Notes: Please don't use the 'su' command.

6.4.7 not in supported list

Such information means the specified value is not supported currently:

$ make boot ROOTDEV=vda
ERR: /dev/vda not in supported ROOTDEV list: /dev/sda /dev/ram0 /dev/nfs, update may help: 'make bsp B=mips64el/ls3a7a'.  Stop.

$ make boot LINUX=v5.8
Makefile:594: *** ERR: v5.8 not in supported Linux list: loongnix-release-1903 v5.7, clone one please: 'make kernel-clone KERNEL_NEW=v5.8'.  Stop.

$ make boot QEMU=loongson-v1.1
Makefile:606: *** ERR: loongson-v1.1 not in supported QEMU list: loongson-v1.0, clone one please: 'make qemu-clone QEMU_NEW=loongson-v1.1'.

There are two main types:

  • One is the specified version is not there or has not been verified

    • Please clone one and verify it with the usage of xxx-clone from section 5.
  • Another is the specified value is invalid or simply not verified

    • For example, the above vda is not added in the ROOTDEV_LIST
    • This board may not support such type of device or just nobody verify and add it
    • This differs from board and kernel version

6.4.8 is not a valid rootfs directory

If using prebuilt filesystem, this error means the rootfs dir, ramdisk or harddisk creating procedure has been interrupted by CTRL+C or similar operations and it means the filesystem is not complete. If no important changes in BSP repository, reset it may help:

$ make bsp-cleanup

If using external filesystem, please make sure the filesystem architecture follows the Linux standards.

7. Contact and Sponsor

7.1 Contact Us

Welcome to join our discuss group:

  • Wechat: tinylab
  • Email: contact /AT\ tinylab /dot/ org

7.2 Give me a star

Welcome to mark our web site, star our git repositories:

7.3 Buy our products

  • Store: https://shop155917374.taobao.com

    • The taobao store of TinyLab.org community, mainly sell products around our open source projects
    • The products include Linux Lab Disk, Pocket Linux Disk and the Linux Lab boards
    • Welcome to buy some based on your requirement, you can search 'Linux Lab' in taobao application to find us
  • Circle: https://t.zsxq.com/uB2vJyF

    • The VIP knowledge channel of TinyLab.org community
    • Create 3+ years, about 1000+ shares and 20+ Linux professionals
  • Courses: https://m.cctalk.com/inst/sh8qtdag

    • TinyLab School -- The video channel of TinyLab.org community
    • Video Live, Video Courses are shared by active Linux professionals from TinyLab.org community

7.4 Sponsor

7.4.1 Sponsor via wechat

contact-sponsor

7.4.2 Sponsor list

  • 2022

  • 2021

    • Lazyparser

      • HelloGCC and HelloLLVM founder
      • 5000RMB
    • Summer 2021

      • Sponsored projects: Rust for Linux, openEuler Kernel for aarch64/virt and x86_64/pc
    • T-head

      • 1 D1 board
    • Allwinner

      • 3 D1 boards
  • 2020

    • Loongson

      • The famous Chinese Loongson CPU designer and manufacturer
      • Sponsored boards: mips64el/ls2k, mips64el/ls3a7a, mipsel/ls1b, mipsel/ls232
    • Summer 2020

      • Sponsored projects: Linux Lab docker image upgrade from Ubuntu 14.04 to Ubuntu 20.04
    • Embedfire

      • 6 imx6ull boards