NetBSD Builder

This project builds a QEMU VM Image for the freya

This project is based on cross-platform-actions/netbsd-builder GitHub action. The image contains a standard NetBSD installation without any X components. It will install the following distribution sets:

• Kernel (GENERIC)
• Kernel modules
• Base
• Configuration files
• Compiler tools
• X11 base and clients
• X11 programming
• X11 configuration
• X11 fonts
• X11 servers

In addition to the above file sets, the following packages are installed as well:

• bash
• curl
• pkgin
• rsync
• sudo
• openssl
• git

The follwoing packages are built:

• freyashell

BIOS: EFI OVMF.fd

Disk layout:

/dev/dk1 on / type ffs (read-only, local)
tmpfs on /var type tmpfs (local)
tmpfs on /tmp type tmpfs (local)
kernfs on /kern type kernfs (local)
ptyfs on /dev/pts type ptyfs (local)
procfs on /proc type procfs (local)
tmpfs on /var/shm type tmpfs (local)
tmpfs on /home/freya/.ssh type tmpfs (local)
/dev/ld1a on /home/freya/storage type ffs (local)

Attached images:

DISK1:
An image of the disk formatted as msdosfs with the following directory layout:

/KEYS - authorized_keys which will be copied to /home/freya/.ssh/


DISK2:
An image of the disk non-formatted, large enough (to fit the code and building) where 
all the files received over freyashell will be installed. The VM will format and mount
the disk manually.

DISK3:
An image of the disk non-formatted, large enough as you expect to have the swap in the system.
Optional. If this disk is not added, the system will operate without swap.

!!! Make sure that both disks are attached to VM because each is strictly binded by its order. Even if you don't need DISK1 i.e you will use default passwords, attach a dummy disk which is not necessary to format.

The / is mounted as read-only. The freya's homedir is also read-only.

Except for the root user, there's one additional user, freya, which is the user that will be running the freyashell. This user can use sudo with a password.

The default password for the root is runner.

Architectures and Versions

The following architectures and versions are supported:

Version	x86-64	ARM64
11.0	✓	✓
10.1	✓	✓
10.0	✓	✓

Building Locally

Prerequisite

UEFI firmware

This needs to be located at resources/ovmf.fd. Copy the OVMF.fd for it's install location to resources/ovmf.fd.

• Ubuntu - Install the ovmf package.
• Fedora - Install the edk2-ovmf package.
• macOS - Copy the OVMF.fd file from a Linux machine

Other

• Packer 1.7.2 or later
• QEMU

Building

1. Clone the repository:

   git clone https://github.com/4neko-org/netbsd-builder
   cd netbsd-builder
   

2. If you running it first time, probably you need to run

   packer init openbsd.pkr.hcl
   

3. Run build.sh to build the image:

   ./build.sh <version> <architecture>
   

   Where <version> and <architecture> are the any of the versions or architectures available in the above table.

   ./build.sh <version> <architecture> -var checksum=<checksum>
   

   On non-macOS platforms the display variable needs to be overridden by specifying -var display=gtk or -var display=sdl at the end when invoking the build.sh script:

   ./build.sh <version> <architecture> -var display=gtk
   

   To enable the hardware acceleration during building run

   ./build.sh <version> <architecture> -var display=gtk -var cpu_type=host
   

   Example:

   ./build.sh 10.1 x86-64 -var display=gtk -var cpu_type=host
   

The above command will build the VM image and the resulting disk image will be at the path: output/netbsd-10.1-x86-64.qcow2.

Additional Information

This VM can be shut down without any gracefull shutdown as the disk is running in read-only mode.

At startup, the image will look for a second hard drive (as described above). If it presents and it contains a file named keys at the root, it will install this file as the authorized_keys file for the runner user. The disk is expected to be formatted as FAT32. This is used as an alternative to a shared folder between the host and the guest, since this is not supported by the xhyve hypervisor. FAT32 is chosen because it's the only filesystem that is supported by both the host (macOS) and the guest (NetBSD) out of the box.

Also, at startup, the OS will look for the third hard drive (as described above). If it presents, an OS will fdisk the image and invoke newfs on the disk erasing everything which was installed previously. This disk image is a workdisk where writing is allowed.

The VM needs to be configured with the virtio-net network device. The disk needs to be configured with the GPT partitioning scheme. And the VM needs to be configured to use UEFI. All this is required for the VM image to be able to run using the xhyve hypervisor.

The qcow2 format is chosen because unused space doesn't take up any space on disk, it's compressible and easily converts the raw format.

Mounting / altering image without rebuilding

If it is required to alter something in the image (instead of rebuilding it), the following should be performed:

1. Log into the VM

2. Run the follwoing

# mount root as RW
mount -uw /

# edit the fstab
vi /etc/fstab

# set the root mount from 'ro' to 'rw' like below
NAME=2537c69f-632c-4a9d-b2e0-blabla		/	ffs	ro		 1 1
# to
NAME=2537c69f-632c-4a9d-b2e0-blabla		/	ffs	rw		 1 1

# comment the /var in order to disable tmpfs mounting like below
# tmpfs /var tmpfs   rw,-m1777,-sram%25

reboot

### DO changes

3. After making all necessary changes do the following:

# Create new image of /var
cd /

tar -cvzf var-image.tar.gz var

# in etc/fstab

# uncomment tmpfs /var line
tmpfs /var tmpfs   rw,-m1777,-sram%25

# change the RW to RO
NAME=2537c69f-632c-4a9d-b2e0-blabla		/	ffs	rw		 1 1
# to
NAME=2537c69f-632c-4a9d-b2e0-blabla		/	ffs	ro		 1 1

# reboot machine or shutdown
reboot

Startup example

/usr/bin/qemu-system-x86_64 \
    -machine type=q35,accel=hvf:kvm:tcg \
    -cpu host \
    -smp 2 \
    -m 4G \
    -device e1000,netdev=user.0,addr=0x03 \
    -netdev user,id=user.0,hostfwd=tcp::65500-:22 \
    -display sdl \
    -monitor none \
    -serial file:/tmp/NetBSD_10.1_65500.txt \
    -boot strict=off \
    --bios /usr/share/edk2/ovmf/OVMF_CODE.fd \
    -device virtio-blk-pci,drive=drive0,bootindex=0 \
    -drive if=none,file=/tmp/netbsd-10.1-x86-64.qcow2,id=drive0,cache=unsafe,discard=ignore \
    -device virtio-scsi-pci,drive=drive1,bootindex=1 \
    -drive if=none,file=/tmp/test0.qcow2,id=drive1,cache=unsafe,discard=ignore,format=qcow2 \
    -device virtio-scsi-pci,drive=drive2,bootindex=2 \
    -drive if=none,file=/tmp/test1.qcow2,id=drive2,cache=unsafe,discard=ignore,format=qcow2