I just uploaded an experimental packet that updates our VirtualBox port from version 6.1.38 to version 7.2.10. The packet is built for Sculpt 26.04 and can be downloaded from my depot. What is working, tested, and what not is tracked in issue #5891.
Feel free to try out the new version. Currently, there exists no installer, but if your are ready to possibly corrupt your VM (“Trust me, I know what I’m doing”
), you can reuse any VDI or VMDK that works with our VirtualBox 6 port.
Here are the steps to reuse a previously setup VM:
- Update the index of ssumpf
- Download and install, but do not configure, ssumpf → Virtual machines → vbox7
- Copy the machine.vbox7 file from <used-fs>/depot/ssumpf/raw/vbox7/2026-06-22/machine.vbox7 into the directory that contains the VM’s machine.vbox6
- Edit machine.vbox7:
- Replace
<HardDisk uuid="{d65876fd-b855-4a35-bbfe-115447c1b9ba}" location="machine.vdi" format="VDI" type="Normal"/>with the corresponding line from your machine.vbox6 file - Replace
<Image uuid="{d65876fd-b855-4a35-bbfe-115447c1b9ba}"/>with the corresponding line from your machine.vbox6 file - Adjust the RAM size (
<Memory RAMSize="4096"/>) - if desired - Adjust the CPU count (
<CPU count="2">) - if desired. I would not configure more than four CPUs at the current state - You can also enable EFI Boot by setting
<Firmware type="EFI"/>
- Replace
- Configure and start
vbox7→ Good luck from here on!
In case you want to click together the configuration, here the deploy that was generated during my test run:
deploy
+ option board | required: yes
+ option sculpt | required: yes
+ option managed | required: yes
+ option system_clock
+ option mixer
+ option black_hole
+ option window_manager
+ option recall_fs
+ option system_shell
+ child mesa_gpu-cpu | pkg: genodelabs/pkg/mesa_gpu-cpu/2026-05-18 | priority: -3
+ connect
+ rm | + parent
+ child vbox7 | pkg: ssumpf/pkg/vbox7/2026-06-22 | priority: -3
+ affinity | xpos: 0 | ypos: 1 | width: 4 | height: 1
+ connect
+ fs vm | + child nvme-1.4.fs | prepend_resource: /vm/ubuntu-22.04
+ fs shared | + child nvme-1.4.fs | prepend_resource: /shared
+ vm | + parent
+ gpu | + child black_hole
+ gui | + child wm
+ nic | + child network
+ rom capslock | + child report | label: global_keys/capslock
+ rom platform_info | + parent | label: platform_info
+ report shape | + child wm
+ report clipboard | + child wm
+ rom clipboard | + child wm
+ rom mesa_gpu.lib.so | + child mesa_gpu-cpu
+ rm | + parent
+ rtc | + child system_clock
+ usb | + child usb
+ play | + child black_hole
+ record | + child black_hole
-
This should give some hints on what to start and what to select. I then converted vbox7 into an option and put it into the model/option folder. This way I can just start vbox7 using the System → Options dialog. Here the option:
option
+ child vbox7 | pkg: ssumpf/pkg/vbox7/2026-06-22 | priority: default | ram: 8G
+ affinity | xpos: 0 | ypos: 1 | width: 4 | height: 1
+ connect
+ fs vm | + child nvme-1.4.fs | prepend_resource: /vm/ubuntu-22.04
+ fs shared | + child nvme-1.4.fs | prepend_resource: /shared
+ vm | + parent
+ gpu | + child black_hole
+ gui | + child wm
+ nic | + child network
+ rom capslock | + child report | label: global_keys/capslock
+ rom platform_info | + parent | label: platform_info
+ report shape | + child wm
+ report clipboard | + child wm
+ rom clipboard | + child wm
+ rom mesa_gpu.lib.so | + child mesa_gpu-cpu
+ rm | + parent
+ rtc | + child system_clock
+ usb | + child usb
+ play | + child black_hole
+ record | + child black_hole
-
You can also use the option as a basis and adjust the location of the fs vm node. In case you configured more than 4GB of RAM for the VM (default), you also have to provide a ram attribute, because vbox7 will require more RAM than configured in the recipe. In the example above I set ram: 8G.
Have fun testing and feel free to report any problems here or in #5891. Cheers!