“To EL2 and Beyond!” is a 2017 presentation about how ARM’s Virtualization Host Extensions (VHE) could change the way Linux and KVM share the processor. Its central contrast is between a split design, with Linux at EL1 and a small hypervisor component at EL2, and a VHE design that lets an EL1-oriented Linux host run at EL2. The slides also describe changes to KVM’s vCPU run loop and timer handling. They are a historical account—not evidence of current kernel behavior or a present-day performance guarantee.
What are EL1 and EL2?
ARM processors organize execution into exception levels, which provide different privilege and responsibility boundaries. In the presentation’s framing, EL1 is where an ordinary operating-system kernel runs, while EL2 is a separate level intended for hypervisor functionality. A hypervisor uses that more privileged position to manage virtual machines and their access to processor resources.
As an Amazon Associate I earn from qualifying purchases.
The distinction matters for virtualization: when a guest operating system runs, the host must manage transitions between host and guest execution and maintain the relevant processor state. The talk describes limitations in the traditional arrangement, where the host Linux kernel runs at EL1 and a smaller KVM hypervisor component runs at EL2.
What does VHE change for KVM/ARM?
ARMv8.1 Virtualization Host Extensions were presented as a way to run an otherwise EL1-oriented operating system at EL2 with expanded EL2 functionality. Rather than requiring the host kernel to remain at EL1 while a separate, smaller component handles EL2 work, the VHE design discussed in the slides places Linux and KVM at EL2.
#1 Best Overall
- Massive 8-Bay Storage for Demanding Workloads: Engineered for high-capacity needs, this chassis supports eight 3.5-inch HDDs, providing terabytes of space for NAS, media servers, and data archives
- Seamless Compatibility with Standard ATX Motherboards: Built to accommodate standard ATX motherboards, offering flexibility and cost savings for your server build without the need for proprietary components
- High-Speed Data Transfers with Front Panel USB-C: Features a front-panel USB 3.2 Gen Type-C port for ultra-fast data transfers, simplifying backups and connectivity with modern peripherals
- Efficient Cooling System with PWM Fans: Equipped with three 80mm PWM fans that provide optimal airflow and temperature control to keep your server components running reliably
- Professional 2U Rackmount Design: Compact 2U form factor fits standard server racks and supports 2U/CRPS power supply units for efficient space utilization in data centers and server rooms
The presentation describes several aspects of this arrangement: backward compatibility when VHE is disabled, additional EL2 capabilities, support for userspace at EL0, and system-register redirection. Together, these are part of the mechanism for adapting an operating system designed around EL1 to the VHE execution model. They should not be read as a complete description of how current Linux kernels implement arm64 virtualization.
| Design in the presentation | Where host Linux runs | EL2 role |
|---|---|---|
| Split-mode KVM/ARM | EL1 | A small hypervisor component runs at EL2. |
| KVM/ARM with VHE | EL2 | Linux and KVM run at EL2, using the expanded functionality described in the slides. |
This is a historical design comparison, not a recommendation for a particular processor or kernel. Whether VHE is available depends on processor support; the talk does not establish a current hardware shortlist or the status of its implementation work today.
Which KVM optimizations did the talk discuss?
Moving work out of the vCPU run loop
The slides describe moving some work from the vCPU run loop into load/put handling. In practical terms, that design aims to perform relevant host or vCPU setup and cleanup around the periods when a vCPU is loaded or put, rather than repeating all such work in the central run loop. The presentation treats this as an optimization idea; it does not establish that the same implementation remains in current Linux code.
Quick wins for a faster PC:
Clear out junk files and repair common Windows errorsFree Scan →Scan for outdated or missing drivers - takes under a minuteDriver Scan →Repair Windows errors before they cause bigger problemsFix Now →Rank #2
- Advanced KVM Solution: Sipeed NanoKVM Pro features second screen capability and LED strip integration for enhanced system monitoring
- 4K HDMI Output: Supports high-resolution display up to 4K for enhanced visual experience and crystal-clear remote viewing
- Remote Server Control: Enables IP-KVM access for homelab and NAS management from anywhere with internet connectivity
- PoE Powered: Simplifies setup with power-over-Ethernet support for NanoKVM Pro, eliminating the need for separate power adapters
- WiFi6 and GbE Connectivity: Ensures fast and stable network performance with dual connectivity options for flexible deployment
Changing timer handling
The talk also discusses timer handling that allows timer work to be managed while KVM is running. Timer behavior is important to virtual machines because guests rely on timely interrupts and timekeeping. The slides present their approach as part of the KVM/ARM design work, not as a universal description of present-day arm64 timer handling.
What hardware and performance evidence did the presentation provide?
The slides state that their experimental system was an AMD Seattle B0 ARM server with a 2.0 GHz AMD A1100 CPU, eight-way SMP, 16 GB of RAM, and 10 GB Ethernet passthrough. These are test-system details reported by the presentation, not a current server recommendation. The deck identifies Linux v4.16 as a target for the work.
It includes a hypercall comparison labeled “3.181” for non-VHE and “3.045” for VHE. The available indexed text does not make the units or enough methodological context clear to treat those numbers as a reliable standalone statistic, so they should not be used to claim a specific speedup. The slides also characterize microbenchmark and application benchmark results positively and compare performance characteristics with x86, but those claims are historical and have not been independently validated here. They do not establish how current processors or kernels perform.
Rank #3
- Advanced KVM Solution: Sipeed NanoKVM Pro features second screen capability and LED strip integration for enhanced system monitoring
- 4K HDMI Output: Supports high-resolution display up to 4K for enhanced visual experience and crystal-clear remote viewing
- Remote Server Control: Enables IP-KVM access for homelab and NAS management from anywhere with internet connectivity
- PoE Powered: Simplifies setup with power-over-Ethernet support for NanoKVM Pro, eliminating the need for separate power adapters
- WiFi6 and GbE Connectivity: Ensures fast and stable network performance with dual connectivity options for flexible deployment
The Linux KVM project’s archived presentation record reports a December 22, 2017 revision date for the 61-page PDF. The presentation and its performance discussion can be read in the slide deck; the KVM project file record gives the archive details. An indexed copy is also available on Scribd, but it does not resolve the ambiguity around the hypercall values.
What’s actually slowing this PC down?
Pick the symptom - the matching free tool is one click away.
How should you interpret “and Beyond” today?
The phrase points to a set of architectural and implementation ideas: using VHE to place an EL1-oriented host at EL2, adapting system-register and userspace behavior, and reducing overhead through run-loop and timer changes. The presentation’s target was Linux v4.16, and its implementation details and measurements belong to that period. The sources cited here do not establish whether particular patches remain upstream, supported, or representative of current kernels.
For a current technical decision, separate four questions rather than relying on the talk’s benchmark claims:
Quick Recap
- Does the processor support VHE?
- At which exception level does the host kernel run in the configuration you care about?
- How does that kernel handle host/guest state transitions and timers?
- What workload and measurement method support any performance comparison?
Product prices and availability are accurate as of the date/time indicated and are subject to change. Any price and availability information displayed on Amazon at the time of purchase will apply.
