This document describes how to set up and run the Fuchsia emulator (FEMU), including networking and GPU support setup.
To run FEMU, you must have:
Building Fuchsia for FEMU
Before you can use FEMU, you need to build Fuchsia using
specifying a qemu board and supported product. This example uses
qemu-x64 for the board and
workstation for the product:
fx set workstation.qemu-x64 --release [--with=...]
For Fuchsia's ephemeral software to work with FEMU, you need to configure an IPv6 network.
To enable networking in FEMU using tap networking, run the following commands:
sudo ip tuntap add dev qemu mode tap user $USER
sudo ip link set qemu up
User Networking (SLIRP) is the default networking set up for FEMU on macOS. This networking set up does not support Fuchsia device discovery.
The most common way to run FEMU is with networking enabled, using the following commands.
To support device discovery without access to external networks.
fx vdl start -N
To get access to external networks:
fx vdl start -N -u FUCHSIA_ROOT/scripts/start-unsecure-internet.sh
Once you run the command, a separate window opens with the title "Fuchsia Emulator". After the Fuchsia emulator launches successfully, the terminal starts with the SSH console. You can run shell commands in this window, just like you would on a Fuchsia device.
On macOS, Fuchsia device discovery does not work. However, you can still use
fx tools such as
fx vdl start
From the output, take note of the instruction on running
fx set-device, you will need it for the steps below.
fx set-device to specify the launched Fuchsia emulator SSH port. For
SSH_PORT, use the value that the
fx vdl start --host-gpu command outputted.
fx set-device 127.0.0.1:SSH_PORT
Additional FEMU options
By default FEMU uses multi-touch input. You can add the argument
for mouse cursor input instead.
fx vdl start --pointing-device mouse
Run FEMU without GUI support
If you don't need graphics or working under the remote workflow, you can run FEMU in headless mode:
fx vdl start --headless
Specify GPU used by FEMU
By default, FEMU launcher uses software rendering using SwiftShader.
To force FEMU to use a specific graphics emulation method, use the parameters
--software-gpu to the
fx vdl start command.
These are the valid commands and options:
|GPU Emulation method||Explanation||
|Hardware (host GPU)||Uses the host machine’s GPU directly to perform GPU processing.||
|Software (host CPU)||Uses the host machine’s CPU to simulate GPU processing.||
Supported hardware for graphics acceleration
FEMU currently supports a limited set of GPUs on macOS and Linux for hardware graphics acceleration. FEMU uses a software renderer fallback for unsupported GPUs.
|Operating System||GPU Manufacturer||OS / Driver Version|
|Linux||Nvidia Quadro||Nvidia Linux Drivers 440.100+|
|macOS||Intel HD Graphics||macOS version 10.15+|
|macOS||AMD Radeon Pro||macOS version 10.15+|
To exit FEMU, run
dm poweroff in the FEMU terminal.