TOPIC: VIRTUALIZATION
Windows 11 virtualisation on Linux using KVM and QEMU
5th March 2026
Windows 11 arrived in October 2021 with a requirement that posed a challenge to many virtualisation users: TPM 2.0 was mandatory, not optional. For anyone running Windows in a virtual machine, that meant their hypervisor needed to emulate a Trusted Platform Module convincingly enough to satisfy the installer.
VirtualBox, which had been my go-to choice for desktop virtualisation for years, could not do this in its 6.1.x series. Support arrived only with VirtualBox 7.0 in October 2022, meaning anyone who needed Windows 11 in a VM faced roughly a year with no straightforward path through their existing tool.
That gap prompted a look at KVM (Kernel-based Virtual Machine), which could handle the TPM requirement through software emulation. This article documents what that investigation found, what the rough edges were at the time, and how the situation has developed in the years since.
What KVM Actually Is
KVM is not a standalone application. It is a virtualisation infrastructure built directly into the Linux kernel, and has been since the module was merged between 2006 and 2007. Rather than sitting on top of the operating system as a separate layer, it turns the Linux kernel itself into a hypervisor. This makes KVM a type-1 hypervisor in practice, even when running on a desktop machine, which is part of why its performance characteristics compare favourably with hosted solutions.
In use, KVM operates alongside QEMU for hardware emulation, libvirt for virtual machine management and virt-manager as a graphical front end. The distinction matters because problems and improvements tend to originate in different parts of that stack. KVM itself is rarely the issue; QEMU and libvirt are where the day-to-day configuration lives.
To confirm that the host CPU supports hardware virtualisation before beginning, the following command checks for the relevant flags:
egrep -c '(vmx|svm)' /proc/cpuinfoAny result above zero means the hardware is capable. Intel processors expose the vmx flag and AMD processors expose svm.
Installing the Required Packages
The installation is straightforward on any major distribution.
On Debian and Ubuntu:
sudo apt install qemu-kvm libvirt-daemon-system libvirt-clients bridge-utils virt-managerOn Fedora:
sudo dnf install @virtualizationOn Arch Linux:
sudo pacman -S qemu libvirt virt-manager bridge-utilsAfter installation, the current user needs to be added to the libvirt and kvm groups before the tools will work without root privileges:
sudo usermod -aG libvirt,kvm $(whoami)Logging out and back in instates the group membership.
Configuring Network Bridging
The default network configuration in libvirt uses NAT, which is sufficient for most purposes and requires no additional setup. The VM can reach the internet and the host, but the host cannot initiate connections to the VM. For a Windows 11 guest used primarily for application compatibility, NAT works without complaint.
A bridged network, which places the VM on the same network segment as the host, requires a wired Ethernet connection. Wireless interfaces do not support bridging in the standard Linux networking stack due to how 802.11 handles MAC addresses. For those on a wired connection, a bridge can be defined with a file named bridge.xml:
<network>
<name>br0</name>
<forward mode="bridge"/>
<bridge name="br0"/>
</network>The bridge is then activated with:
sudo virsh net-define bridge.xml
sudo virsh net-start br0
sudo virsh net-autostart br0Installing Windows 11
Windows 11 requires TPM 2.0 and Secure Boot. Neither is present in a default KVM configuration, and both need to be added explicitly.
The swtpm package provides software TPM emulation:
sudo apt install swtpm swtpm-tools # Debian/Ubuntu
sudo dnf install swtpm swtpm-tools # FedoraUEFI firmware is provided by the ovmf package, which supplies the file that virt-manager needs for Secure Boot:
sudo apt install ovmf # Debian/Ubuntu
sudo dnf install edk2-ovmf # FedoraIn virt-manager, when creating the VM, the firmware should be set to UEFI x86_64: /usr/share/OVMF/OVMF_CODE.fd rather than the default BIOS option. A TPM 2.0 device should be added in the hardware configuration before the VM is started. With those two elements in place, the Windows 11 installer proceeds without complaint about the hardware requirements.
The VirtIO drivers ISO should be attached as a second virtual CD-ROM drive during installation. The installer will not find the storage device otherwise because the VirtIO disk controller is not a standard device that Windows recognises without a driver. When prompted to select an installation location and no disks appear, clicking "Load driver" and browsing to the VirtIO ISO resolves it.
During the out-of-box experience, Windows 11 requires a Microsoft account and an internet connection by default. To bypass this and create a local account instead, opening a command prompt with Shift+F10 and running the following works on the Home edition:
oobebypassNROThe machine restarts and presents an option to proceed without internet access.
Performance Considerations
KVM performance for a Windows 11 guest is generally good, but one factor specific to Windows 11 is worth understanding. Memory Integrity, also referred to as Hypervisor-Protected Code Integrity (HVCI), is a Windows security feature that uses virtualisation to protect the kernel. Running it inside a virtual machine creates nested virtualisation overhead because the guest is attempting to run its own virtualisation layer inside the host's. The performance impact is more pronounced on processors predating Intel Kaby Lake or AMD Zen 2, where the hardware support for nested virtualisation is less capable.
The CPU type selection in virt-manager also matters more than it might appear. Setting the CPU model to host-passthrough exposes the actual host CPU flags to the guest, which improves performance compared to emulated CPU models, at the cost of reduced portability if the VM image is ever moved to a different machine.
Host File System Access and Clipboard Sharing
This was where the experience diverged most noticeably from VirtualBox. VirtualBox Guest Additions handle shared folders and clipboard integration as a single installation, and the result works reliably with minimal configuration. KVM requires separate solutions for each, and in 2022 neither was as seamless as it has since become.
Clipboard Sharing via SPICE
Clipboard sharing uses the SPICE display protocol rather than VNC. The VM needs a SPICE display and a virtio-serial controller, which virt-manager adds automatically when SPICE is selected. Within the Windows guest, the installer for SPICE guest tools provides the clipboard agent. Once installed, clipboard text passes between host and guest in both directions.
The critical dependency that caused problems in 2022 was the virtio-serial channel. Without a com.redhat.spice.0 character device present in the VM configuration, the clipboard agent installs successfully but does nothing. Virt-manager now adds this automatically when SPICE is selected, which removes one of the more common failure points.
Host Directory Sharing via Virtiofs
At the time of this investigation, the practical option for sharing files between the Linux host and a Windows guest was WebDAV, which worked but felt like a workaround. The proper solution, virtiofs, existed but was not yet well-supported on Windows guests. The situation has since improved to the point where virtiofs is now the standard recommended approach.
It requires three components: the virtiofsd daemon on the host (included in recent QEMU packages), the virtiofs driver from the VirtIO Windows drivers package and WinFsp, which is the Windows equivalent of FUSE. Once configured through virt-manager's file system hardware settings, the shared directory appears as a mapped drive in Windows Explorer. The virtiofsd daemon was also rewritten in Rust in the intervening period, improving both its reliability and performance.
To configure a shared directory, shared memory must first be enabled in the VM's memory settings, then a file system device added with the driver set to virtiofs, a source path on the host and an arbitrary mount tag. The corresponding libvirt XML looks like this:
<memoryBacking>
<source type='memfd'/>
<access mode='shared'/>
</memoryBacking>
<filesystem type='mount' accessmode='passthrough'>
<driver type='virtiofs' queue='1024'/>
<source dir='/home/user/shared'/>
<target dir='host_share'/>
</filesystem>This was the area where VirtualBox held a clear practical advantage in 2022. The gap has since narrowed considerably.
Migrating from VirtualBox
Moving existing VirtualBox VMs to KVM is possible using qemu-img, which converts between disk image formats. The straightforward conversion from VDI to QCOW2 is:
qemu-img convert -f vdi -O qcow2 windows11.vdi windows11.qcow2For large images or where reliability is a concern, converting via an intermediate RAW format reduces the risk of issues:
qemu-img convert -f vdi -O raw windows11.vdi windows11.raw
qemu-img convert -f raw -O qcow2 windows11.raw windows11.qcow2The resulting QCOW2 file can then be used when creating a new VM in virt-manager, selecting "Import existing disk image" rather than creating a new one.
How the Landscape Has Shifted Since
The investigation described here took place during a specific window: VirtualBox 6.1.x was the current release, Windows 11 had just launched, and KVM was the most practical route to TPM emulation on Linux. That context has changed in several ways worth noting for anyone reading this in 2026.
VirtualBox 7.0 arrived in October 2022 with TPM 1.2 and 2.0 support, Secure Boot and a number of additional improvements. The original reason for investigating KVM was resolved, and for those who had moved across during the gap period, returning to VirtualBox for Windows guests made sense given its more straightforward Guest Additions integration.
QEMU reached version 10.0 in April 2025, a significant milestone reflecting years of accumulated improvements to hardware emulation, storage performance and x86 guest support. Libvirt has kept pace, adding reliable internal snapshots for UEFI-based VMs, evdev input device hot plug and improved unprivileged user support. The virtiofs situation for Windows guests has moved from "technically possible but awkward" to "the recommended approach with good documentation and a rewritten daemon", which addresses the most significant practical shortcoming from 2022 directly.
The broader desktop virtualisation landscape shifted when VMware Workstation Pro became free for all users, including commercial ones, in November 2024. VMware Workstation Player was discontinued as a separate product at the same time, having become redundant once Workstation Pro was available at no cost. This gave desktop users a third credible option alongside VirtualBox and KVM, with VMware's historically strong Windows guest integration now accessible without a licence fee, though users of the free version are not entitled to support through the global support team.
The miniature PC market also expanded considerably from 2023 onwards, with Intel N100-based and AMD Ryzen Embedded systems offering enough performance to run Windows natively at modest cost. For many people, that proves a cleaner solution than any hypervisor, eliminating the integration limitations entirely by giving Windows its own dedicated hardware.
Final Assessment
KVM handled Windows 11 competently during a period when the alternatives could not, and the platform has continued to improve in the years since. The two areas that fell short in 2022, host file sharing and clipboard integration, have been addressed by developments in virtiofs and the SPICE tooling, and a new user starting today may find the experience noticeably smoother.
Whether KVM is the right choice in 2026 depends on the use case. For Linux-native workloads and server-style VM management, it remains the strongest option on Linux. For a Windows desktop guest where ease of integration matters most, VirtualBox 7.x and VMware Workstation Pro are both strong alternatives, with the latter now free to use for both commercial and personal purposes. The question that drove this investigation was answered by VirtualBox itself in October 2022. KVM provided a workable solution in the meantime, and the platform has only become more capable since then.
Additional Reading
How To Convert VirtualBox Disk Image (VDI) to Qcow2 format
How to enable TPM and secure boot on KVM?
Windows 11 on KVM – How to Install Step by Step?
Enable Virtualization-based Protection of Code Integrity in Microsoft Windows
VirtualBox memory allocation error: Solving Linux Mint host issues after LLM usage
22nd March 2025
It happened to me today when I tried starting up Windows virtual machines in VirtualBox on my main Linux Mint workstation as a host after a long layover for these. They failed to start, only for these messages to appear:
Out of memory condition when allocating memory with low physical backing. (VERR_NO_LOW_MEMORY).
Result Code:
NS_ERROR_FAILURE (0x80004005)
Component:
ConsoleWrap
Interface:
IConsole {6ac83d89-6ee7-4e33-8ae6-b257b2e81be8}
Since the messages are cryptic in the circumstances, I had to seek out their meaning. The system has plenty of memory, so it could be that. Various suggestions came my way like installing the VirtualBox Extension Pack or reinstalling VirtualBox Extensions in the affected VM. The first had no effect, while the second was impossible.
However, there was one more suggestion: fragmentation of memory, much like file fragmentation on a disk drive. Thus, I opted for a reboot, which sorted out things, making it look as if that were the problem. If it comes up again, I might try compacting the memory with the following command, leaving for a while to complete due to any temporary system slowdown:
echo 1 > /proc/sys/vm/compact_memory
Because there had been some on-machine usage of an LLM, I now reckon that caused the malaise. These can be as heavy on memory as they are on processors, so fragmentation can result. That is yet another likely lesson learned from experimenting with this much-hyped technology.
Upheaval and miniaturisation
4th March 2025
The ongoing AI boom got me refreshing my computer assets. One was a hefty upgrade to my main workstation, still powered by Linux. Along the way, I learned a few lessons:
- Processing with LLM's only works on a graphics card when everything can remain within its onboard memory. It is all too easy to revert to system memory and CPU usage, given the amount of memory you get on consumer graphics cards. That applies even with the latest and greatest from Nvidia, when the main use case is for gaming. Things become prohibitively expensive when you go on from there.
- Even with water cooling, keeping a top of the range CPU cool and its fans running quietly remains a challenge, more so than when I last went for a major upgrade. It takes time for things to settle down.
- My Iiyama monitor now feels flaky with input from the latest technology. This is enough to make me look for a replacement, and it is waking up from dormancy that is the real issue. While it was always slow, plugging out from mains electricity and then back in again is a hack that is needed all too often.
- KVM switches may need upgrading to work with the latest graphical input. The monitor may have been a culprit with the problems that I was getting, yet things were smoother once I replaced the unit that I had been using with another that is more modern.
- AMD Ryzen 9 chips now have onboard graphics, a boon when things are not proceeding too well with a dedicated graphics card. Even though this was not the case when the last major upgrade happened, there were no issues like what I faced this time around.
- Having LED's on a motherboard to tell what might be stopping system startup is invaluable. This helped in July 2021 and averted confusion this time around as well. While only four of them were on offer, knowing which of CPU, DRAM, GPU or system boot needs attention is a big help.
- Optical drives are not needed any longer. Booting off a USB drive was enough to get Linux Mint installed, once I got the image loaded on there properly. Rufus got used, and I needed to select the low-level writing option before things proceeded as I had hoped.
Just like 2021, the 2025 upgrade cycle needed a few weeks for everything to settle down. The previous cycle was more challenging, and this was not just because of an accompanying heatwave. The latest one was not so bedevilled.
Given the above, one might be tempted to go for a less arduous path, like my acquisition of an iMac last year for another place that I own. After all, a Mac Mini packs in quite a lot of power, and it is not the only miniature option. Now that I have one, I have moved image processing off the workstation and onto it. The images are stored on the Linux machine and edited on the Mac, which has plenty of memory and storage of its own. There is also an M4 chip, so processing power is not lacking either.
It could have been used for work affairs, yet I acquired a Geekom A8 for just that. Though seeking work as I write this, my being an incorporated freelancer means that having a dedicated machine that uses my main monitor has its advantages. Virtualisation can allow drift from business affairs to business matters, that is not so easy when a separate machine is involved. There is no shortage of power either with an AMD Ryzen 9 8945HS and Radeon 780M Graphics on board. Add in 32 GB of memory and 2 TB of storage and all is commodious. It can be surprising what a small package can do.
The Iiyama's travails also pop up with these smaller machines, less so on the Geekom than with the Mac. The latter needs the HDMI cable to be removed and reinserted after a delay to sort out things. Maybe that new monitor may not be such an off the wall idea after all.
Migrating a Windows 7 Virtual Machine from VirtualBox to VMware Player
14th October 2012
Seeing how well Windows 8 was running in a VMware Player virtual machine and that was without installing VMware Tools in the guest operating system, I was reminded about how sluggish my Windows 7 VirtualBox VM had become. Therefore, I decided to try a migration of the VM from VirtualBox to VMware. My hope was this: it would be as easy as exporting to an OVA file (File > Export Appliance... in VirtualBox) and importing that into VMware (File > Open a VM in Player). However, even selecting OVF compatibility was insufficient for achieving this, and the size of the virtual disks meant that the export took a while to run as well. The solution was to create a new VM in VirtualBox from the OVA file and use the newly created VMDK files with VMware. That worked successfully to give me a speedier, more responsive Windows 7 VM for my pains.
Access to host directories needed reinstatement using a combination of the VMware Shared Folders feature and updating drive mappings in Windows 7 itself to use what appeared to it as network drives in the Shared Folders directory on the \\vmware-host domain. For that to work, VMware Tools needed to be installed in the guest OS (go to Virtual Machine > Install VMware Tools to make available a virtual CD from which the installation can be done) as I discovered when trying the same thing with my Windows 8 VM, where I dare not instate VMware Tools due to their causing trouble when I last attempted it.
Moving virtual machine software brought about its side effects, though. Software like Windows 7 detects that it's on different hardware, so reactivation can be needed. While Windows 7 reactivation was a painless online affair, it wasn't the same for Photoshop CS5. That meant that I needed help from Adobe's technical support people top get past the number of PC's for which the software already had been activated. In hindsight, deactivation should have been done before the move, but that's a lesson that I know well now. Technical support sorted my predicament politely and efficiently while reinforcing the aforementioned learning point. Moving virtual machine platform is very like moving from one PC to the next, and it hadn't clicked with me quite how real those virtual machines can be when it comes to software licensing.
Apart from that and figuring out how to do it, the move went smoothly. An upgrade to the graphics driver on the host system and getting Windows 7 to recheck the capabilities of the virtual machine even gained me a fuller Aero experience than I had before then. Full-screen operation is quite reasonable too (the CTRL + ALT + ENTER activates and deactivates it) and photo editing now feels less boxed in too.
/sbin/mount.vboxsf: mounting failed with the error: Protocol error
19th April 2009
These days, my virtualisation needs are being well served by VirtualBox 2.2. Though it may be the closed source variant, I have no complaints about it. Along with a number of Windows VM's, I also have one running Ubuntu 9.04 and, for the first time, I seem to have VirtualBox's Guest Additions playing with a Linux guest as they should. Even the Shared Folders functionality is working.
However, I did get one problem when I tried out the last feature for the first time. The procedure is to issue a command like the following in a terminal session after creating the requisite directory in the file system and adding a host directory as a shared folder:
sudo mount -t vboxsf Music /mnt/host_music/
Above, Music is the name of the folder in the VirtualBox manager and /mnt/host_music in the directory in the guest file system. However, this returned the message at the head of this post at that first attempt:
/sbin/mount.vboxsf: mounting failed with the error: Protocol error
The solution thankfully turns out to be an easy one: reinstalling the Guest Additions, which certainly did the trick for me. The cause would appear to have been an update to Ubuntu, and 9.04 is understandably in a state of flux at the moment (I suspect kernel upgrades because of my previous experiences). Regardless of this, it is good to know that it's a problem with a simple fix, and I am seeing the niceties of a larger virtual screen system together with automatic grabbing and releasing of the mouse cursor too. While there may be a chance to explore the availability of these sorts of features to other Linux guests, I have other things that I should be doing and there's sunshine outside to be enjoyed.