Rootlinux17: Hypervisors on ARM - Overview and Design Choices by Julien Grall, ARM
•
6 likes•3,525 views
Hypervisors are used in a broad range of domains ranging from Embedded systems, Automotive to big iron servers. The choice of hypervisor has a strong impact on the overall design of your project and its performance. This talk introduces the state of virtualization on ARM, and provides a description of three popular open source hypervisors: KVM, Jailhouse and Xen. Julien Grall explains respective key features, technical differences and suitability of the hypervisor for different application domains. Julien Grall is a Software Virtualisation Engineer at ARM. The talk was delivered at Root Linux Conference 2017. Learn more: http://linux.globallogic.com/materials. The video recording is available at https://www.youtube.com/watch?v=jZNXtqFJpuc
![About me
Working on ARM virtualization for the past 4 years
With ARM since 2016
Co-maintaining Xen on ARM - with Stefano Stabellini [Aporeto]
2 © ARM 2017](data:image/gif;base64,R0lGODlhAQABAIAAAAAAAP///yH5BAEAAAAALAAAAAABAAEAAAIBRAA7)
Recommended
Recommended
More Related Content
What's hot
What's hot (20)
Similar to Rootlinux17: Hypervisors on ARM - Overview and Design Choices by Julien Grall, ARM
Similar to Rootlinux17: Hypervisors on ARM - Overview and Design Choices by Julien Grall, ARM (20)
More from The Linux Foundation
More from The Linux Foundation (20)
Recently uploaded
Recently uploaded (20)
Rootlinux17: Hypervisors on ARM - Overview and Design Choices by Julien Grall, ARM
- 1. Hypervisors on ARM Overview and Design choices Julien Grall <julien.grall@arm.com> Root Linux Conference 2017 © ARM 2017
- 2. About me Working on ARM virtualization for the past 4 years With ARM since 2016 Co-maintaining Xen on ARM - with Stefano Stabellini [Aporeto] 2 © ARM 2017
- 3. Virtualization, what is it? 3 © ARM 2017
- 4. Virtualization refers to the act of creating a virtual version of something Wikipedia 4 © ARM 2017
- 5. 5 © ARM 2017
- 6. Use cases 6 © ARM 2017
- 7. Type of hypervisors Figure: From wikipedia 7 © ARM 2017
- 8. Kind of virtualization Full hardware virtualization OS is running unmodified Guest I/O are either emulated handled by virtualization-aware hardware Para-virtualization OS is aware of the hypervisor Privilege instruction are replaced by hooks Devices (network, block...) are para-virtualized The trend is a mix of both Use as much as possible hardware-assisted virtualization Devices (network, block...) para-virtualized or passthrough-ed Emulation very limited 8 © ARM 2017
- 9. ARM virtualization 9 © ARM 2017
- 10. ARMv8-A Privilege Model Support both AArch32 and AArch64 execution modes 32-64bit inter-working limited to exception boundaries AArch64 always has a higher privilege than AArch32 AArch64 state is a superset of AArch32 state 10 © ARM 2017
- 11. ARM virtualization Introduced with the latest version of ARMv7 architecture New hypervisor execution state Non-Secure world, higher privilege than EL1 11 © ARM 2017
- 12. Virtualization in a nutshell Second stage of memory translation Adds an extra level of indirection between guests and physical memory TLBs are tagged by Virtual Machine ID (VMID) Ability to trap access of most system registers The hypervisor decides what it wants to trap Can handle IRQs, FIQs and asynchronous aborts The guest doesn’t see physical interrupts firing, for example Guests can call into EL2 mode (HVC instruction) Allows para-virtualizated services Standard architecture peripherals are virtualization-aware GIC and timer have specific features to help virtualization 12 © ARM 2017
- 13. EL2: Not EL1++ (ARMv8.0-A) EL2 is not a superset of NS-EL1 Orthogonal mode to EL1 Allows multiplexing of NS-EL1 guests on the hardware Own translation regime Separate Stage-1 translation, no Stage-2 translation It would be difficult to run Linux in EL2 Requires too many changes to be practical EL2 could be used as a ”world switch” Between guests (barametal hypervisor/Type I) Between host and guest (hosted hypervisor/Type II) This makes the host a form of specialized guest. 13 © ARM 2017
- 14. Hypervisor architecture - Type I EL1 EL0 EL2 Guest Kernel Hypervisor Guest Kernel Guest Kernel Guest Userspace Userspace Userspace Guest Guest 14 © ARM 2017
- 15. Hypervisor architecture - Type II EL1 EL0 EL2 HYP Host Kernel Switching Code Guest Kernel Guest Kernel Host Userspace Userspace Userspace Guest Guest 15 © ARM 2017
- 16. EL2 enhancement (ARMv8.1-A) The Virtualization Host Extension (VHE) expands the capability of EL2: Designed to improve the support of the Type-2 hypervisors Allows the host OS to be run at EL2 The host OS requires minimal changes to run at EL2 User-space still runs at EL0 Host has no software running at EL1 AArch64 specific EL2 becomes a strict superset of EL1 16 © ARM 2017
- 17. Hosted hypervisor architecture on platform without VHE EL1 EL0 EL2 HYP Host Kernel Switching Code Guest Kernel Guest Kernel Host Userspace Userspace Userspace Guest Guest 17 © ARM 2017
- 18. Hosted hypervisor architecture on platform with VHE EL1 EL0 EL2 Guest Kernel Guest Kernel Host Userspace Userspace Userspace Guest Guest Host Kernel + HYP 18 © ARM 2017
- 19. Nested Virtualization (ARMv8.3-A) The Nested Virtualization extension allows an hypervisor in a VM. Unmodified guest hypervisor running in NS EL1 Implementation of a host hypervisor required Running at EL2 AArch64 specific 19 © ARM 2017
- 20. Why using ARM virtualization Robust set of virtualization features Not just about CPU virtualization Covers the whole systems architecture Scalable architecture Power to IoT-like devices ... ... all the way to server-grade systems An architecture in motion: ARMv8.1-A: https://goo.gl/Ox4thV ARMv8.2-A: https://goo.gl/0Ns37U ARMv8.3-A: https://goo.gl/CJv1n0 20 © ARM 2017
- 21. OpenSource Hypervisors 21 © ARM 2017
- 22. 22 © ARM 2017
- 23. KVM 23 © ARM 2017
- 24. Kernel-based Virtual Machine First version of KVM was merged in Linux 2.6.20 AArch32 support merged in Linux 3.9 AArch64 support merged in Linux 3.11 Source code available as GLP v2 Hosted hypervisor 24 © ARM 2017
- 25. KVM virtual machine Use of assisted hardware virtualization Devices are emulated (QEMU) para-virtualized (VIRTIO) 25 © ARM 2017
- 26. KVM architecture 26 © ARM 2017
- 27. KVM architecture with ARMv8.0-A EL1 EL0 EL2 KVM Host Kernel Switching Code Guest Kernel Guest Kernel Host Userspace Userspace Userspace Guest Guest 27 © ARM 2017
- 28. KVM architecture with ARMv8.1-A EL1 EL0 EL2 Guest Kernel Guest Kernel Host Userspace Userspace Userspace Guest Guest Host Kernel + KVM 28 © ARM 2017
- 29. Resource management All CPUs are using the same scheduler guest vCPU is a task for the host OS Resource management can be done using cgroup Standard way in Linux to control resources 29 © ARM 2017
- 30. Jailhouse 30 © ARM 2017
- 31. Jailhouse Created at Siemens in 2013 Partioning hypervisor Type-I hypervisor Linux will load Jailhouse Source code available as GPL v2 Small code base: <10K lines 31 © ARM 2017
- 32. Jailhouse architecture 32 © ARM 2017
- 33. Xen 33 © ARM 2017
- 34. Xen First released in 2003 ARM officially supported since Xen 4.4 Source code available as GPL v2 Small code base: 30K Bare-metal hypervisor 34 © ARM 2017
- 35. Xen architecture 35 © ARM 2017
- 36. Xen architecture - 2 36 © ARM 2017
- 37. Xen schedulers 37 © ARM 2017
- 38. Xen schedulers 38 © ARM 2017
- 40. Questions? 40 © ARM 2017
- 41. The trademarks featured in this presentation are registered and/or unregistered trademarks of ARM limited (or its subsidiaries) in the EU and/or elsewhere. All rights reserved. All other marks featured may be trademarks of their respective owners. Copyright © 2017 ARM Limited © ARM 2017