Docker Android - Run QEMU Android x86 and Android ARM in a Docker! X11 Forwarding! CI/CD for Android!
:5555
localhost:50922
)localhost:5555
)/dev/video0
)/dev/snd
)/dev/dri
)docker commit
This project is maintained by @sickcodes Sick.Codes. (Twitter)
Additional credits can be found here: https://github.com/sickcodes/dock-droid/blob/master/CREDITS.md
Epic thanks to @BlissRoms who maintain absolutely incredible Android x86 images. If you love their images, consider donating to the project: https://blissos.org/!
Special thanks to @zhouziyang who maintains an even more native fork Redroid!
This project is heavily based on Docker-OSX: https://github.com/sickcodes/Docker-OSX
https://github.com/sickcodes/droid-native
https://github.com/sickcodes/dock-droid
https://github.com/sickcodes/anbox-modules-dkms
Before you do anything else, you will need to turn on hardware virtualization in your BIOS. Precisely how will depend on your particular machine (and BIOS), but it should be straightforward.
Then, you'll need QEMU and some other dependencies on your host:
# ARCH
sudo pacman -S qemu libvirt dnsmasq virt-manager bridge-utils flex bison iptables-nft edk2-ovmf
# UBUNTU DEBIAN
sudo apt install qemu qemu-kvm libvirt-clients libvirt-daemon-system bridge-utils virt-manager
# CENTOS RHEL FEDORA
sudo yum install libvirt qemu-kvm
Then, enable libvirt and load the KVM kernel module:
sudo systemctl enable --now libvirtd
sudo systemctl enable --now virtlogd
echo 1 | sudo tee /sys/module/kvm/parameters/ignore_msrs
sudo modprobe kvm
You can run the Live OS image, or install to disk.
Connect to the WiFi network called VirtWifi
.
docker run -it \
--device /dev/kvm \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-p 5555:5555 \
sickcodes/dock-droid:latest
Headlessly (on a server, or locally)
docker run -it \
--device /dev/kvm \
-e EXTRA="-display none -vnc 0.0.0.0:99,password=on" \
-p 5555:5555 \
-p 5999:5999 \
sickcodes/dock-droid:latest
For headless, in the QEMU console, type change vnc password user
And then connect on localhost:5999
, or the server IP, or Docker IP.
docker run -it \
--device /dev/kvm \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-v "${PWD}/android.qcow2:/home/arch/dock-droid/android.qcow2" \
-p 5555:5555 \
sickcodes/dock-droid:naked
This will boot, but currently does not "work".
Change CPU
to Penryn
, which is normally host
Change ENABLE_KVM
, which is normally -enable-kvm
Change KVM
, which is normally accel=kvm:tcg
Change CPUID_FLAGS
, which is normally very long.
# use a spacebar in quotes
-e CPU=qemu64 \
-e ENABLE_KVM=' ' \
-e KVM=' ' \
-e CPUID_FLAGS=' ' \
For example (Work in Progress):
docker run -it \
-e CPU=Penryn \
-e ENABLE_KVM=' ' \
-e KVM=' ' \
-e CPUID_FLAGS=' ' \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-p 5555:5555 \
sickcodes/dock-droid:latest
Increase RAM by adding this line: -e RAM=10 \
for 10GB.
Want to use your Laptop/USB WebCam and Audio too?
There are two options: usb passthrough, or usb redirect (network).
v4l2-ctl --list-devices
lsusb
Find the hostbus
and hostaddr
:
Bus 003 Device 003: ID 13d3:56a2 IMC Networks USB2.0 HD UVC WebCam
Would be -device usb-host,hostbus=3,hostaddr=3
docker run -it \
--device /dev/kvm \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-p 5555:5555 \
-e EXTRA='-device usb-host,hostbus=3,hostaddr=3' \
sickcodes/dock-droid:latest
lsusb
# Bus 003 Device 003: ID 13d3:56a2 IMC Networks USB2.0 HD UVC WebCam
Vendor ID is 13d3
Product ID is 56a2
In one Terminal on host:
sudo usbredirserver -p 7700 13d3:56a2
In another Terminal on host:
# 172.17.0.1 is the IP of the Docker Bridge, usually the host, but you can change this to anything.
PORT=7700
IP_ADDRESS=172.17.0.1
docker run -it \
--device /dev/kvm \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-p 5555:5555 \
-e EXTRA="-chardev socket,id=usbredirchardev1,port=${PORT},host=${IP_ADDRESS} -device usb-redir,chardev=usbredirchardev1,id=usbredirdev1,bus=ehci.0,debug=4" \
sickcodes/dock-droid:latest
Currently in development by BlissOS team: mesa graphics card + OpenGL3.2.
Want to use SwiftShader acceleration?
docker run -it \
--privileged \
--device /dev/kvm \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-p 5555:5555 \
-p 50922:10022 \
--device=/dev/dri \
--group-add video \
-e EXTRA='-display sdl,gl=on' \
sickcodes/dock-droid:latest
# get container name from
docker ps -a
# copy out the image
docker cp container_name:/home/arch/dock-droid/android.qcow2 .
Use any generic ISO or use your own Android AOSP raw image or qcow2
Where, "${PWD}/disk.qcow2"
is your image in the host system.
docker run -it \
-v "${PWD}/android.qcow2:/home/arch/dock-droid/android.qcow2" \
--privileged \
--device /dev/kvm \
--device /dev/video0 \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-p 5555:5555 \
-p 50922:10022 \
-e EXTRA='-device usb-host,hostbus=3,hostaddr=3' \
sickcodes/dock-droid:latest
Add the following: -bios /usr/share/OVMF/x64/OVMF.fd \
to Launch.sh
Or as a docker run
argument:
UEFI Boot
docker run -it \
--device /dev/kvm \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-p 5555:5555 \
-e EXTRA='-bios /usr/share/OVMF/x64/OVMF.fd' \
sickcodes/dock-droid:latest
To use an alternative CDROM
, you have two choices: runtime or buildtime.
You can add your image to the Dockerfile during the build:
CDROM_IMAGE_URL='https://sourceforge.net/projects/blissos-x86/files/Official/bleeding_edge/Generic%20builds%20-%20Pie/11.13/Bliss-v11.13--OFFICIAL-20201113-1525_x86_64_k-k4.19.122-ax86-ga-rmi_m-20.1.0-llvm90_dgc-t3_gms_intelhd.iso'
docker build \
-t dock-droid-custom \
-e CDROM_IMAGE_URL="${CDROM_IMAGE_URL}" .
OR you can add it during runtime to the docker hub images as follows.
-v "${CDROM}:/cdrom" \
-e CDROM=/cdrom \
For example:
# full path to your image on the host
CDROM="${HOME}/Downloads/image.iso"
docker run -it \
--device /dev/kvm \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-p 5555:5555 \
-v "${CDROM}:/cdrom" \
-e CDROM=/cdrom \
sickcodes/dock-droid:latest
-boot d
will force QEMU to boot from the CDROM.
-e EXTRA='-boot d ' \
Reduces the image size by 600Mb if you are using a local directory disk image:
docker cp image_name /home/arch/dock-droid/android.qcow2 .
The following groups of commands is for editing android.qcow2
.
First, we will mount the main qcow2
file using libguestfstools
Then, inside that qcow2 image, there are:
system.img
ramdisk.img
GRUB is also in there.
Mount the qcow2 using:
# on the host
# enable qemu-nbd for network device mounting
# wget -O android.qcow2 https://image.sick.codes/android.BlissOS_Stable.qcow2
sudo modprobe nbd
sudo qemu-nbd --connect=/dev/nbd0 android.qcow2 -f qcow2
sudo fdisk /dev/nbd0 -l
mkdir -p /tmp/image
sudo mount /dev/nbd0p1 /tmp/image
Now you can mount the internal disks in other places...
# make a folder to mount the whole resizable image
# make another to mount the raw Android image within that resizable image.
mkdir -p /tmp/system
sudo mount /tmp/image/bliss-x86-11.13/system.img /tmp/system
ls /tmp/system
ls /tmp/image/bliss-x86-11.13
# don't forget to unmount
ndk_translation
Android x86Extract the native-bridge using the following: https://github.com/sickcodes/Droid-NDK-Extractor
Thanks to Frank from Redroid for the idea!
https://github.com/remote-android/redroid-doc/tree/master/native_bridge
# warning, this will extract overwriting /etc/system/... so make sure you're in /tmp
cd /tmp \
&& sudo wget https://github.com/sickcodes/dock-droid/raw/master/native-bridge.tar.gz \
&& sudo tar -xzvf native-bridge.tar.gz \
&& sudo rm native-bridge.tar.gz
sudo touch /tmp/system/vendor/etc/init/nativebridge.rc
sudo tee /tmp/system/vendor/etc/init/nativebridge.rc <<EOF
on early-init
setprop ro.product.cpu.abilist x86_64,arm64-v8a,x86,armeabi-v7a,armeabi
setprop ro.product.cpu.abilist64 x86_64,arm64-v8a
setprop ro.product.cpu.abilist32 x86,armeabi-v7a,armeabi
setprop ro.dalvik.vm.isa.arm x86
setprop ro.dalvik.vm.isa.arm64 x86_64
setprop ro.enable.native.bridge.exec 1
setprop ro.dalvik.vm.native.bridge libndk_translation.so
setprop ro.ndk_translation.version 0.2.2
EOF
# # Enable native bridge for target executables
# on early-init
# mount binfmt_misc binfmt_misc /proc/sys/fs/binfmt_misc
# on property:ro.enable.native.bridge.exec=1
# copy /system/etc/binfmt_misc/arm_exe /proc/sys/fs/binfmt_misc/register
# copy /system/etc/binfmt_misc/arm_dyn /proc/sys/fs/binfmt_misc/register
# on property:ro.enable.native.bridge.exec64=1
# copy /system/etc/binfmt_misc/arm64_exe /proc/sys/fs/binfmt_misc/register
# copy /system/etc/binfmt_misc/arm64_dyn /proc/sys/fs/binfmt_misc/register
# add armeabi
sudo sed -i -e 's/abilist32\=x86\,armeabi\-v7a/abilist32\=x86\,armeabi\-v7a\,armeabi/g' /tmp/system/vendor/build.prop
sudo sed -i -e 's/abilist\=x86_64\,x86\,arm64\-v8a\,armeabi\-v7a/abilist\=x86_64\,x86\,arm64\-v8a\,armeabi\-v7a\,armeabi/g' /tmp/system/vendor/build.prop
sudo rm /tmp/system/bin/enable_nativebridge \
/tmp/system/etc/binfmt_misc/arm_exe \
/tmp/system/etc/binfmt_misc/arm64_dyn \
/tmp/system/etc/binfmt_misc/arm_dyn \
/tmp/system/etc/binfmt_misc/arm64_exe
sudo rm /tmp/system/vendor/etc/binfmt_misc/*
sudo rm /tmp/system/etc/init/houdini.rc
sudo tee -a /tmp/system/product/build.prop \
-a /tmp/system/vendor/build.prop \
-a /tmp/system/build.prop <<EOF
ro.product.cpu.abilist64=x86_64,arm64-v8a
ro.product.cpu.abilist32=x86,armeabi-v7a,armeabi
ro.dalvik.vm.isa.arm=x86
ro.dalvik.vm.isa.arm64=x86_64
ro.enable.native.bridge.exec=1
ro.dalvik.vm.native.bridge=libndk_translation.so
ro.secure=0
ro.debuggable=1
ro.adb.secure=0
service.adb.root=1
EOF
# also swap libhoudini to libndk_translation in the ramdisk
mkdir -p /tmp/ramdisk
sudo /bin/bash -c "cd /tmp/ramdisk \
&& zcat /tmp/image/bliss-x86-11.13/ramdisk.img \
| cpio -iud"
sed -i -e 's/libhoudini/libndk_translation/g' /tmp/ramdisk/default.prop
touch /tmp/image/bliss-x86-11.13/ramdisk.img
find . | cpio -o -H newc | gzip > /tmp/ramdisk.img.new
mv /tmp/ramdisk.img.new /tmp/image/bliss-x86-11.13/ramdisk.img
# sudo tee -a /tmp/system/build.prop <<'EOF'
# ro.dalvik.vm.native.bridge=libndk_translation.so
# ro.product.cpu.abilist=x86_64,arm64-v8a,x86,armeabi-v7a,armeabi
# ro.product.cpu.abilist32=x86,armeabi-v7a,armeabi
# ro.ndk_translation.version=0.2.2
# EOF
# don't forget to unmount
Inside the ramdisk.img
, we would like to overwrite init
with rusty-magisk
mkdir -p /tmp/ramdisk
sudo /bin/bash -c "
cd /tmp/ramdisk
zcat /tmp/image/bliss-x86-11.13/ramdisk.img | cpio -iud && mv /tmp/ramdisk/init /tmp/ramdisk/init.real
wget -O /tmp/ramdisk/init https://github.com/axonasif/rusty-magisk/releases/download/v0.1.7/rusty-magisk_x86_64
chmod a+x /tmp/ramdisk/init
touch /tmp/image/bliss-x86-11.13/ramdisk.img
/bin/bash -c 'find . | cpio -o -H newc | sudo gzip > /tmp/image/bliss-x86-11.13/ramdisk.img'
"
sudo rm -rf /tmp/ramdisk
# don't forget to unmount
During the next boot you will have Magisk installed.
# generate keys if you don't have already
KEYNAME=adbkey
adb keygen ~/.android/"${KEYNAME}"
touch ~/.android/"${KEYNAME}.pub"
adb pubkey ~/.android/"${KEYNAME}" > ~/.android/"${KEYNAME}.pub"
# put some keys in the box and copy to your host ~/.android folder
mkdir -p /tmp/image/bliss-x86-11.13/data/.android
mkdir -p /tmp/image/bliss-x86-11.13/data/misc/adb
tee /tmp/image/bliss-x86-11.13/data/misc/adb/adb_keys < ~/.android/"${KEYNAME}.pub"
# don't forget to unmount
After completing any of the above automation, you need to unmount the disk.
# sudo mount /tmp/image/bliss-x86-11.13/ramdisk.img /tmp/ramdisk
# unmount both disks when you're done
sudo umount /tmp/ramdisk
sudo umount /tmp/system
sudo umount /tmp/image
sudo qemu-nbd -d /dev/nbd0
# choose a version from https://github.com/frida/frida/releases/
# use arm if you're debugging arm apps, via houdini or native bridge (ndk)
HOST_ARCH=x86_64
# HOST_ARCH=arm
GUEST_SYS=android
FRIDA_RELEASE=frida-server
FRIDA_RELEASES=($(curl https://github.com/frida/frida/releases | grep -Po "(?<=\<a\ href\=\")(\/frida\/frida\/releases\/download\/\d+\.\d.\d+\/${FRIDA_RELEASE}-\d+\.\d+.\d+-${GUEST_SYS}-${HOST_ARCH}.xz)(?=\"\ )"))
RELEASE_LINK="https://github.com${FRIDA_RELEASES[0]}"
# RELEASE_LINK='https://github.com/frida/frida/releases/download/15.0.8/frida-server-15.0.8-android-x86_64.xz'
mkdir -p ./frida
cd ./frida
wget "${RELEASE_LINK}"
unxz -d "$(basename "${RELEASE_LINK}")"
find -name "${FRIDA_RELEASE}-*.*.*-${GUEST_SYS}-${HOST_ARCH}" | xargs -i mv "{}" frida-server
adb -s localhost:5555 push frida-server /data/local/tmp/frida-server
adb -s localhost:5555 shell "su -c chmod 755 /data/local/tmp/frida-server"
adb -s localhost:5555 shell "/data/local/tmp/frida-server"
Great list by @eladkarako
https://gist.github.com/eladkarako/5694eada31277fdc75cee4043461372e
Boot the container.
Open Terminal Emulator
in the Android:
# on android
su
start adbd
# setprop persist.adb.tcp.port 5555
Connect to the virtual WiFi inside Android!
Now, from the host, use the new key to adb
into the guest:
# on the host
export ADB_VENDOR_KEYS=~/.android/adbkey
adb kill-server
adb connect localhost
adb -s localhost:5555 root
adb -s localhost:5555 shell
In the Android terminal emulator, run adbd
Then from the host, you can can connect using either:
adb connect localhost:5555
adb connect 172.17.0.2:5555
If you have more than "one emulator" you may have to use:
adb -s localhost:5555 shell
adb -s 172.17.0.2:5555 shell
E.g.
su
sed -i -e 's/ro\.adb\.secure\=1/ro\.adb\.secure\=0/' /default.prop
In the Android terminal emulator, run adbd
Then from the host, you can can connect using either:
adb connect localhost:5555
adb connect 172.17.0.2:5555
For more sophisticated endeavours, we offer the following support services:
In case you're interested, contact @sickcodes on Twitter or submit a contact form here.
dock-droid is licensed under the GPL v3+, also known as the GPL v3 or later License. Contributions are welcomed and immensely appreciated.
Don't be shy, the GPLv3+ allows you to use Dock-Droid as a tool to create proprietary software, as long as you follow any other license within the software.
This is a Dockerized Android setup/tutorial for conducting Android Security Research.
Product names, logos, brands and other trademarks referred to within this project are the property of their respective trademark holders. These trademark holders are not affiliated with our repository in any capacity. They do not sponsor or endorse this project in any way.
Identify your webcam:
lsusb | grep -i cam
Bus 003 Device 003: ID 13d3:56a2 IMC Networks USB2.0 HD UVC WebCam
Using Bus
and Device
as hostbus
and hostaddr
, include the following docker command:
Waiting for public mesa
builds: https://blissos.org/
When these hardware accelerated images are released, you can follow the Issue opened by: @M1cha
See https://github.com/sickcodes/dock-droid/issues/2
the online documentation for that is very bad and mostly outdated(due to kernel and qemu updates). But here's some references that helped me set it up several times:
https://wiki.archlinux.org/title/PCI_passthrough_via_OVMF#Plain_QEMU_without_libvirt
https://www.kernel.org/doc/Documentation/vfio.txt
the general summary:
* make sure your hardware supports VT-d/AMD-VI and UEFI and linux have it enabled * figure out which devices are in the same iommu group * detach all drivers from those devices * attach vfio-pci to those devices
Add the following lines when you are ready:
--privileged \
-e EXTRA="-device vfio-pci,host=04:00.0' \
Work in progress
sudo tee -a /etc/libvirt/qemu.conf <<'EOF'
cgroup_device_acl = [
"/dev/null", "/dev/full", "/dev/zero",
"/dev/random", "/dev/urandom",
"/dev/ptmx", "/dev/kvm",
"/dev/vfio/vfio",
"/dev/dri/card0",
"/dev/dri/card1",
"/dev/dri/renderD128"
]
EOF
# --device /dev/video0 \
# --device /dev/video1 \
grep "video\|render" /etc/group
# render:x:989:
# video:x:986:sddm
sudo usermod -aG video "${USER}"
sudo systemctl restart libvirtd
docker run -it \
-v "${PWD}/android.qcow2:/home/arch/dock-droid/android.qcow2" \
--privileged \
--device /dev/kvm \
--device /dev/video1 \
-v /tmp/.X11-unix:/tmp/.X11-unix \
-e "DISPLAY=${DISPLAY:-:0.0}" \
-p 5555:5555 \
-p 50922:10022 \
--user 1000:1000 \
--group-add=966 \
--group-add=989 \
--device /dev/dri/renderD128:/dev/dri/renderD128 \
--device /dev/dri/card0:/dev/dri/card0 \
--device /dev/dri/card1:/dev/dri/card1 \
sickcodes/dock-droid:naked
# pick which graphics card
# --device /dev/dri/card0:/dev/dri/card0 \
qemu-img convert -f vdi -O qcow2 BlissOS.vdi android.qcow2
Add the following line:
-e EXTRA="-display none -vnc 0.0.0.0:99,password=on"
In the Docker terminal, press enter
until you see (qemu)
.
Type change vnc password someusername
Enter a password for your new vnc username^.
You also need the container IP: docker inspect <containerid> | jq -r '.[0].NetworkSettings.IPAddress'
Or ip n
will usually show the container IP first.
Now VNC connect using the Docker container IP, for example 172.17.0.2:5999
Remote VNC over SSH: ssh -N root@1.1.1.1 -L 5999:172.17.0.2:5999
, where 1.1.1.1
is your remote server IP and 172.17.0.2
is your LAN container IP.
Now you can direct connect VNC to any container built with this command!
This requires 250GB of REAL space.
This was previously at ./build
, but due to Docker Hub using the wrong README.md file, I have added these instructions below:
Make and add a non-root user
USERADD=user
useradd "${USERADD}" -p "${USERADD}"
tee -a /etc/sudoers <<< "${USERADD} ALL=(ALL) NOPASSWD: ALL"
mkdir -p "/home/${USERADD}"
chown "${USERADD}:${USERADD}" "/home/${USERADD}"
# passwd user <<EOF
# 1000
# 1000
# EOF
chsh -s /bin/bash "${USERADD}"
usermod -aG docker "${USERADD}"
su user
BUILD_DIRECTORY=/mnt/volume_nyc3_01
# create a persistent folder on the host for building stuff
mkdir "${BUILD_DIRECTORY}/blissos-r36"
cd "${BUILD_DIRECTORY}/blissos-r36"
wget https://raw.githubusercontent.com/sickcodes/dock-droid/master/Dockerfile.build
docker build -t blissos-builder .
docker run -it \
-e REVISION=r11-r36 \
-e MANIFEST_REPO=https://github.com/BlissRoms-x86/manifest.git \
-v "${BUILD_DIRECTORY}/blissos-r36:/blissos-r36" \
blissos-builder