linux-rootfs-arm64

In order to enable qemu's net service, you need to create follow files:

➜  ~ mkdir /etc/qemu
➜  ~ echo 'allow all' > /etc/qemu/bridge.conf
➜  ~ mkdir -p /dev/net
➜  ~ mknod /dev/net/tun c 10 200
➜  ~ chmod 600 /dev/net/tun

This branch is A simple root file system built by busybox for linux-arm64.

You may see some folders with only .gitkeep files, this is to prevent git from ignoring these empty directories. These empty directories are all necessary in the root filesystem.

There is a MAKEDEV file in initramfs/dev/, you need to go into this directory and run MAKEDEV before all:

➜  linux-rootfs cd initramfs/dev 
➜  dev sh MAKEDEV

Then please ensure that the execute permission for the files has been enabled:

➜  linux-rootfs sudo chmod a+x initramfs/init
➜  linux-rootfs sudo chmod a+x initramfs/etc/init.d/rcS

This root file system includes a mount directory mnt/. This directory is mounted to the /root/mnt_path/ directory on the physical machine. Files between them will be synchronized.

Below is my build process.

Compiling BusyBox from Source

Download the BusyBox source code: https://busybox.net/downloads/busybox-1.36.0.tar.bz2

Before compiling BusyBox, change a setting:

➜  ~ cd busybox-1.36.0
➜  busybox-1.36.0 make menuconfig

Turn on the setting -> [*] Build static binary (no shared libs).

➜  busybox-1.36.0 make -j 40
➜  busybox-1.36.0 make install

Building the Root File System

➜  ~ cd busybox-1.36.0/_install
➜  _install mkdir etc dev lib proc sys tmp mnt
➜  _install touch init
➜  _install chmod a+x init
➜  _install rm linuxrc

Write the following in the init file:

➜  _install cat init
#!/bin/busybox sh
exec /sbin/init

Create the following files in the etc/ directory:

➜  _install cd etc/
➜  etc cat profile 
#!/bin/sh
export HOSTNAME=kuankuanQAQ
export USER=root
export HOME=/home
export PS1="[$USER@$HOSTNAME \W]\# "
PATH=/bin:/sbin:/usr/bin:/usr/sbin
LD_LIBRARY_PATH=/lib:/usr/lib:$LD_LIBRARY_PATH
export PATH LD_LIBRARY_PATH

➜  etc cat inittab 
::sysinit:/etc/init.d/rcS
::respawn:-/bin/sh
::askfirst:-/bin/sh
::ctrlaltdel:/bin/umount -a -r

➜  etc cat fstab 
#device  mount-point    type     options   dump   fsck order
proc /proc proc defaults 0 0
tmpfs /tmp tmpfs defaults 0 0
sysfs /sys sysfs defaults 0 0
tmpfs /dev tmpfs defaults 0 0
debugfs /sys/kernel/debug debugfs defaults 0 0
mount-1 /mnt 9p trans=virtio,version=9p2000.L,posixacl,msize=104857600 0 0

➜  etc cat init.d/rcS 
touch /etc/passwd
touch /etc/group
echo root:x:0:0:root:/root:/bin/sh > /etc/passwd
echo root:x:0: > /etc/group
chmod 4755 /bin/busybox

/bin/mount -a
mdev -s

Create device files in the dev/ directory:

➜  _install cd dev/
➜  dev mknod console c 5 1

Then, create the initramfs:

➜  linux-rootfs mkdir initramfs
➜  linux-rootfs cd initramfs
➜  initramfs cp ~/busybox-1.36.0/_install/* -rf ./

Finally, package the initramfs:

➜  linux-rootfs cd initramfs
➜  initramfs find . -print0 | cpio --null -ov --format=newc | gzip -9 > ../initramfs.cpio.gz

权限问题

如果并不是以root身份构造busybox，则有可能会出现：mount: you must be root 报错。因此需要将 bin 目录下所有文件修改拥有者为root：

➜  bin sudo chown root * -R

bullsyes根文件系统

create-image.sh来源：https://github.com/google/syzkaller/blob/master/tools/create-image.sh

这是一个用于创建最小化Debian Linux镜像的脚本，主要用于syzkaller（一个内核模糊测试工具）项目。我没去掉里面syzkaller相关的部分。

但是做了如下小修改：

# 可能会需要某种修改！把enp0s1改成ip link show查询出的设备名
printf '\nauto enp0s1\niface enp0s1 inet dhcp\n' | sudo tee -a $DIR/etc/network/interfaces

如果你搜索启动日志，会看到这样一条：virtio_net virtio1 enp0s1: renamed from eth0。这是因为使用systemd的发行版Linux采用了"可预测网络接口命名"机制。它不再使用传统的 eth0、eth1 这样的名称，而是根据设备的物理/虚拟特性来命名。因此/etc/network/interfaces中的eth0要改成enp0s1。

但是我最近遇到了ip link show什么也查不出来的情况，可能是内核配置没有开相应的选项。那么如果启动不了就不是这里的问题了。

/etc/fstab修改：

# 把qemu虚拟机中的/mnt挂在到物理机上
mount-1 /mnt 9p trans=virtio,version=9p2000.L,posixacl,msize=512000 0 0

然后把qemu启动命令加上：

    -fsdev local,path=/root/mnt_path/,security_model=mapped,id=dev-1 \
    -device virtio-9p-device,fsdev=dev-1,mount_tag=mount-1 \