x86 virtualization
x86 virtualizationis the use ofhardware-assisted virtualizationcapabilities on an x86/x86-64 CPU.
In the late 1990s x86 virtualization was achieved by complex software techniques, necessary to compensate for the processor's lack of hardware-assisted virtualization capabilities while attaining reasonable performance. In 2005 and 2006, bothIntel(VT-x) andAMD(AMD-V) introduced limited hardware virtualization support that allowed simpler virtualization software but offered very few speed benefits.[1]Greater hardware support, which allowed substantial speed improvements, came with later processor models.
Software-based virtualization
[edit]The following discussion focuses only on virtualization of the x86 architectureprotected mode.
In protected mode the operating system kernel runs at a higher privilege such asring0, and applications at a lower privilege such as ring 3.[citation needed]In software-based virtualization, a host OS has direct access to hardware while the guest OSs have limited access to hardware, just like any other application of the host OS. One approach used in x86 software-based virtualization to overcome this limitation is calledring deprivileging,which involves running the guest OS at a ring higher (lesser privileged) than 0.[2]
Three techniques made virtualization of protected mode possible:
- Binary translationis used to rewrite certain ring 0 instructions in terms of ring 3 instructions, such asPOPF,that would otherwise fail silently or behave differently when executed above ring 0,[3][4]: 3 making the classictrap-and-emulate virtualizationimpossible.[4]: 1 [5]To improve performance, the translatedbasic blocksneed to be cached in a coherent way that detectscode patching(used inVxDsfor instance), the reuse of pages by the guest OS, or evenself-modifying code.[6]
- A number of key data structures used by a processor need to beshadowed.Because most operating systems usepaged virtual memory,and granting the guest OS direct access to theMMUwould mean loss of control by thevirtualization manager,some of the work of the x86 MMU needs to be duplicated in software for the guest OS using a technique known asshadow page tables.[7]: 5 [4]: 2 This involves denying the guest OS any access to the actual page table entries by trapping access attempts and emulating them instead in software. The x86 architecture uses hidden state to storesegment descriptorsin the processor, so once the segment descriptors have been loaded into the processor, the memory from which they have been loaded may be overwritten and there is no way to get the descriptors back from the processor.Shadow descriptor tablesmust therefore be used to track changes made to the descriptor tables by the guest OS.[5]
- I/O device emulation: Unsupported devices on the guest OS must be emulated by adevice emulatorthat runs in the host OS.[8]
These techniques incur some performance overhead due to lack of MMU virtualization support, as compared to a VM running on a natively virtualizable architecture such as theIBM System/370.[4]: 10 [9]: 17 and 21
On traditional mainframes, the classictype 1hypervisor was self-standing and did not depend on any operating system or run any user applications itself. In contrast, the first x86 virtualization products were aimed at workstation computers, and ran a guest OS inside a host OS by embedding the hypervisor in a kernel module that ran under the host OS (type 2 hypervisor).[8]
There has been some controversy whether the x86 architecture with no hardware assistance is virtualizable as described byPopek and Goldberg.VMwareresearchers pointed out in a 2006ASPLOSpaper that the above techniques made the x86 platform virtualizable in the sense of meeting the three criteria of Popek and Goldberg, albeit not by the classic trap-and-emulate technique.[4]: 2–3
A different route was taken by other systems likeDenali,L4,andXen,known asparavirtualization,which involvesportingoperating systems to run on the resulting virtual machine, which does not implement the parts of the actual x86 instruction set that are hard to virtualize. The paravirtualized I/O has significant performance benefits as demonstrated in the originalSOSP'03 Xen paper.[10]
The initial version ofx86-64(AMD64) did not allow for a software-only full virtualization due to the lack of segmentation support inlong mode,which made the protection of the hypervisor's memory impossible, in particular, the protection of the trap handler that runs in the guest kernel address space.[11][12]: 11 and 20 Revision D and later 64-bit AMD processors (as a rule of thumb, those manufactured in 90 nm or less) added basic support for segmentation in long mode, making it possible to run 64-bit guests in 64-bit hosts via binary translation. Intel did not add segmentation support to its x86-64 implementation (Intel 64), making 64-bit software-only virtualization impossible on Intel CPUs, but Intel VT-x support makes 64-bit hardware assisted virtualization possible on the Intel platform.[13][14]: 4
On some platforms, it is possible to run a 64-bit guest on a 32-bit host OS if the underlying processor is 64-bit and supports the necessary virtualization extensions.
Hardware-assisted virtualization
[edit]In 2005 and 2006,IntelandAMD(working independently) created newprocessor extensionsto the x86 architecture. The first generation of x86 hardware virtualization addressed the issue of privileged instructions. The issue of low performance of virtualized system memory was addressed withMMUvirtualization that was added to the chipset later.
Central processing unit
[edit]Virtual 8086 mode
[edit]Because theIntel 80286could not run concurrent DOS applications well by itself in protected mode, Intel introduced thevirtual 8086 modein their80386chip, which offered virtualized 8086 processors on the 386 and later chips. Hardware support for virtualizing the protected mode itself, however, became available 20 years later.[15]
AMD virtualization (AMD-V)
[edit]![](https://upload.wikimedia.org/wikipedia/commons/thumb/6/66/AMD_Phenom_die_equalized.png/220px-AMD_Phenom_die_equalized.png)
AMD developed its first generation virtualization extensions under the code name "Pacifica", and initially published them as AMD Secure Virtual Machine (SVM),[16]but later marketed them under the trademarkAMD Virtualization,abbreviatedAMD-V.
On May 23, 2006, AMD released the Athlon 64 ("Orleans"), the Athlon 64 X2 ("Windsor") and the Athlon 64 FX ("Windsor") as the first AMD processors to support this technology.
AMD-V capability also features on theAthlon 64andAthlon 64 X2family of processors with revisions "F" or "G" onsocket AM2,Turion 64 X2,andOpteron2nd generation[17]and third-generation,[18]PhenomandPhenom IIprocessors. TheAPU Fusionprocessors support AMD-V. AMD-V is not supported by any Socket 939 processors. The onlySempronprocessors which support it are APUs andHuron,Regor,Sargasdesktop CPUs.
AMD Opteron CPUs beginning with the Family 0x10 Barcelona line, and Phenom II CPUs, support a second generation hardware virtualization technology calledRapid Virtualization Inde xing(formerly known as Nested Page Tables during its development), later adopted by Intel asExtended Page Tables(EPT).
As of 2019, allZen-based AMD processors support AMD-V.
TheCPU flagfor AMD-V is "svm". This may be checked inBSD derivativesviadmesgorsysctland inLinuxvia/proc/cpuinfo
.[19]Instructions in AMD-V include VMRUN, VMLOAD, VMSAVE, CLGI, VMMCALL, INVLPGA, SKINIT, and STGI.
With somemotherboards,users must enable AMD SVM feature in theBIOSsetup before applications can make use of it.[20]
Intel virtualization (VT-x)
[edit]![](https://upload.wikimedia.org/wikipedia/commons/thumb/8/87/Intel_Core_i7-940_bottom.jpg/220px-Intel_Core_i7-940_bottom.jpg)
Previously codenamed "Vanderpool", VT-x represents Intel's technology for virtualization on the x86 platform. On November 13, 2005, Intel released two models ofPentium 4(Model 662 and 672) as the first Intel processors to support VT-x. The CPU flag for VT-x capability is "vmx"; in Linux, this can be checked via/proc/cpuinfo
,or inmacOSviasysctl machdep.cpu.features
.[19]
"VMX" stands for Virtual Machine Extensions, which adds 13 new instructions: VMPTRLD, VMPTRST, VMCLEAR, VMREAD, VMWRITE, VMCALL, VMLAUNCH, VMRESUME, VMXOFF, VMXON, INVEPT, INVVPID, and VMFUNC.[21]These instructions permit entering and exiting a virtual execution mode where the guest OS perceives itself as running with full privilege (ring 0), but the host OS remains protected.
As of 2015[update],almost all newer server, desktop and mobile Intel processors support VT-x, with some of theIntel Atomprocessors as the primary exception.[22]With somemotherboards,users must enable Intel's VT-x feature in theBIOSsetup before applications can make use of it.[23]
Intel started to includeExtended Page Tables(EPT),[24]a technology for page-table virtualization,[25]since theNehalemarchitecture,[26][27]released in 2008. In 2010,Westmereadded support for launching the logical processor directly inreal mode– a feature called "unrestricted guest", which requires EPT to work.[28][29]
Since theHaswellmicroarchitecture (announced in 2013), Intel started to includeVMCS shadowingas a technology that acceleratesnested virtualizationof VMMs.[30] Thevirtual machine control structure(VMCS) is adata structurein memory that exists exactly once per VM, while it is managed by the VMM. With every change of the execution context between different VMs, the VMCS is restored for the current VM, defining the state of the VM's virtual processor.[31]As soon as more than one VMM or nested VMMs are used, a problem appears in a way similar to what required shadow page table management to be invented, as describedabove.In such cases, VMCS needs to be shadowed multiple times (in case of nesting) and partially implemented in software in case there is no hardware support by the processor. To make shadow VMCS handling more efficient, Intel implemented hardware support for VMCS shadowing.[32]
VIA virtualization (VIA VT)
[edit]VIA Nano3000 Series Processors and higher support VIA VT virtualization technology compatible with Intel VT-x.[33]EPT is present inZhaoxinZX-C, a descendant ofVIA QuadCore-E&Eden X4similar to NanoC4350AL.[34]
Interrupt virtualization (AMD AVIC and Intel APICv)
[edit]In 2012, AMD announced theirAdvanced Virtual Interrupt Controller(AVIC) targeting interrupt overhead reduction in virtualization environments.[35]This technology, as announced, does not supportx2APIC.[36] In 2016, AVIC is available on the AMD family 15h models 6Xh (Carrizo) processors and newer.[37]
Also in 2012, Intel announced a similar technology for interrupt andAPICvirtualization, which did not have a brand name at its announcement time.[38] Later, it was branded asAPIC virtualization(APICv)[39] and it became commercially available in theIvy Bridge EPseries of Intel CPUs, which is sold as Xeon E5-26xx v2 (launched in late 2013) and as Xeon E5-46xx v2 (launched in early 2014).[40]
Graphics processing unit
[edit]Graphics virtualization is not part of the x86 architecture. IntelGraphics Virtualization Technology(GVT) provides graphics virtualization as part of more recent Gen graphics architectures. AlthoughAMD APUsimplement thex86-64instruction set, they implement AMD's own graphics architectures (TeraScale,GCNandRDNA) which do not support graphics virtualization.[citation needed]Larrabeewas the only graphicsmicroarchitecturebased on x86, but it likely did not include support for graphics virtualization.
Chipset
[edit]Memory and I/O virtualization is performed by thechipset.[41]Typically these features must be enabled by the BIOS, which must be able to support them and also be set to use them.
I/O MMU virtualization (AMD-Vi and Intel VT-d)
[edit]![](https://upload.wikimedia.org/wikipedia/commons/thumb/c/c4/AMD-Vi_boot_log_screenshot.png/330px-AMD-Vi_boot_log_screenshot.png)
An input/output memory management unit (IOMMU) allows guestvirtual machinesto directly useperipheraldevices, such as Ethernet, accelerated graphics cards, and hard-drive controllers, throughDMAandinterruptremapping. This is sometimes calledPCI passthrough.[42]
An IOMMU also allows operating systems to eliminate bounce buffers needed to allow themselves to communicate with peripheral devices whose memory address spaces are smaller than the operating system's memory address space, by using memory address translation. At the same time, an IOMMU also allows operating systems and hypervisors to prevent buggy or malicious hardware fromcompromising memory security.
Both AMD and Intel have released their IOMMU specifications:
- AMD's I/O Virtualization Technology, "AMD-Vi", originally called "IOMMU"[43]
- Intel's "Virtualization Technology for Directed I/O" (VT-d),[44]included in most high-end (but not all) newer Intel processors since the Core 2 architecture.[45]
In addition to the CPU support, bothmotherboardchipsetand system firmware (BIOSorUEFI) need to fully support the IOMMU I/O virtualization functionality for it to be usable. Only thePCIorPCI Expressdevices supportingfunction level reset(FLR) can be virtualized this way, as it is required for reassigning variousdevice functionsbetween virtual machines.[46][47]If a device to be assigned does not supportMessage Signaled Interrupts(MSI), it must not shareinterruptlines with other devices for the assignment to be possible.[48] Allconventional PCIdevices routed behind a PCI/PCI-X-to-PCI Express bridge can be assigned to a guest virtual machine only all at once; PCI Express devices have no such restriction.
Network virtualization (VT-c)
[edit]- Intel's "Virtualization Technology for Connectivity" (VT-c).[49]
PCI-SIG Single Root I/O Virtualization (SR-IOV)
[edit]PCI-SIG Single Root I/O Virtualization (SR-IOV)provides a set of general (non-x86 specific) I/O virtualization methods based onPCI Express(PCIe) native hardware, as standardized by PCI-SIG:[50]
- Address translation services (ATS)supports native IOV across PCI Express via address translation. It requires support for new transactions to configure such translations.
- Single-root IOV(SR-IOV or SRIOV)supports native IOV in existing single-root complex PCI Express topologies. It requires support for new device capabilities to configure multiple virtualized configuration spaces.[51]
- Multi-root IOV (MR-IOV)supports native IOV in new topologies (for example, blade servers) by building on SR-IOV to provide multiple root complexes which share a common PCI Express hierarchy.
In SR-IOV, the most common of these, a host VMM configures supported devices to create and allocate virtual "shadows" of their configuration spaces so that virtual machine guests can directly configure and access such "shadow" device resources.[52]With SR-IOV enabled, virtualized network interfaces are directly accessible to the guests,[53] avoiding involvement of the VMM and resulting in high overall performance;[51]for example, SR-IOV achieves over 95% of thebare metalnetwork bandwidth inNASA's virtualized datacenter[54]and in theAmazon Public Cloud.[55][56]
See also
[edit]- Comparison of application virtualization software
- Comparison of platform virtualization software
- Hardware-assisted virtualization
- Hypervisor
- I/O virtualization
- Network virtualization
- OS-level virtualization
- Timeline of virtualization development
- Virtual machine
- List of IOMMU-supporting hardware
- Second Level Address Translation(SLAT)
- Message Signaled Interrupts(MSI)
References
[edit]- ^A Comparison of Software and Hardware Techniques for x86 Virtualization, Keith Adams and Ole Agesen, VMware, ASPLOS’06 October 21–25, 2006, San Jose, California, USAArchived2010-08-20 at theWayback Machine"Surprisingly, we find that the first-generation hardware support rarely offers performance advantages over existing software techniques. We ascribe this situation to high VMM/guest transition costs and a rigid programming model that leaves little room for software flexibility in managing either the frequency or cost of these transitions.
- ^"Intel Virtualization Technology Processor Virtualization Extensions and Intel Trusted execution Technology"(PDF).Intel. 2007.Archived(PDF)from the original on 2015-05-21.Retrieved2016-12-12.
- ^"USENIX Technical Program - Abstract - Security Symposium - 2000".Usenix.org. 2002-01-29.Archivedfrom the original on 2010-06-10.Retrieved2010-05-02.
- ^abcde"A Comparison of Software and Hardware Techniques for x86 Virtualization"(PDF).VMware.Archived(PDF)from the original on 20 August 2010.Retrieved8 September2010.
- ^abU.S. patent 6,397,242
- ^U.S. patent 6,704,925
- ^"Virtualization: architectural considerations and other evaluation criteria"(PDF).VMware.Archived(PDF)from the original on 6 February 2011.Retrieved8 September2010.
- ^abU.S. patent 6,496,847
- ^"VMware and Hardware Assist Technology"(PDF).Archived(PDF)from the original on 2011-07-17.Retrieved2010-09-08.
- ^"Xen and the Art of Virtualization"(PDF).Archived(PDF)from the original on 2014-09-29.
- ^"How retiring segmentation in AMD64 long mode broke VMware".Pagetable. 2006-11-09.Archivedfrom the original on 2011-07-18.Retrieved2010-05-02.
- ^"VMware and CPU Virtualization Technology"(PDF).VMware.Archived(PDF)from the original on 2011-07-17.Retrieved2010-09-08.
- ^"VMware KB: Hardware and firmware requirements for 64bit guest operating systems".Kb.vmware.Archivedfrom the original on 2010-04-19.Retrieved2010-05-02.
- ^"Software and Hardware Techniques for x86 Virtualization"(PDF).Archived fromthe original(PDF)on 2010-01-05.Retrieved2010-05-02.
- ^Yager, Tom (2004-11-05)."Sending software to do hardware's job | Hardware - InfoWorld".Images.infoworld.Archivedfrom the original on 2014-10-18.Retrieved2014-01-08.
- ^"33047_SecureVirtualMachineManual_3-0.book"(PDF).Archived(PDF)from the original on 2012-03-05.Retrieved2010-05-02.
- ^"What are the main differences between Second-Generation AMD Opteron processors and first-generation AMD Opteron processors?".amd.Archived fromthe originalon April 15, 2009.Retrieved2012-02-04.
- ^"What virtualization enhancements do Quad-Core AMD Opteron processors feature?".amd.Archived fromthe originalon April 16, 2009.Retrieved2012-02-04.
- ^abTo see if your processor supports hardware virtualizationArchived2012-11-25 at theWayback MachineIntel 2012.
- ^"How to enable Intel VTx and AMD SVM?".Support.QNAP Systems, Inc. Archived fromthe originalon 2018-03-07.Retrieved2020-12-23.
- ^ INTEL (October 2019)."Intel® 64 and IA-32 Architectures Software Developer's Manual".intel.Intel Corporation.Retrieved2020-01-04.
- ^"Intel Virtualization Technology List".Ark.intel.Archivedfrom the original on 2010-10-27.Retrieved2010-05-02.
- ^"Windows Virtual PC: Configure BIOS".Microsoft. Archived fromthe originalon 2010-09-06.Retrieved2010-09-08.
- ^Neiger, Gil; A. Santoni; F. Leung; D. Rodgers; R. Uhlig (2006)."Intel Virtualization Technology: Hardware Support for Efficient Processor Virtualization"(PDF).Intel Technology Journal.10(3). Intel: 167–178.doi:10.1535/itj.1003.01.Archived fromthe original(PDF)on 2012-09-25.Retrieved2008-07-06.
- ^Gillespie, Matt (2007-11-12)."Best Practices for Paravirtualization Enhancements from Intel Virtualization Technology: EPT and VT-d".Intel Software Network.Intel.Archivedfrom the original on 2008-12-26.Retrieved2008-07-06.
- ^"First the Tick, Now the Tock: Next Generation Intel Microarchitecture (Nehalem)"(PDF)(Press release). Intel.Archived(PDF)from the original on 2009-01-26.Retrieved2008-07-06.
- ^"Technology Brief: Intel Microarchitecture Nehalem Virtualization Technology"(PDF).Intel. 2009-03-25.Archived(PDF)from the original on 2011-06-07.Retrieved2009-11-03.
- ^[1]"Intel added unrestricted guest mode on Westmere micro-architecture and later Intel CPUs, it uses EPT to translate guest physical address access to host physical address. With this mode, VMEnter without enable paging is allowed."
- ^[2]"If the “unrestricted guest” VM-execution control is 1, the “enable EPT” VM-execution control must also be 1 "
- ^"4th-Gen Intel Core vPro Processors with Intel VMCS Shadowing"(PDF).Intel.2013.Retrieved2014-12-16.
- ^Understanding Intel Virtualization Technology (VT).ArchivedSeptember 8, 2014, at theWayback MachineRetrieved 2014-09-01
- ^The 'what, where and why' of VMCS shadowing.Archived2014-09-03 at theWayback MachineRetrieved 2014-09-01
- ^VIA Introduces New VIA Nano 3000 Series ProcessorsArchivedJanuary 22, 2013, at theWayback Machine
- ^"Notebook Solution: Kaixian ZX-C Processor + VX11PH Chipset"(PDF).
- ^Wei Huang,Introduction of AMD Advanced Virtual Interrupt ControllerArchived2014-07-14 at theWayback Machine,XenSummit 2012
- ^Jörg Rödel (August 2012)."Next-generation Interrupt Virtualization for KVM"(PDF).AMD.Archived(PDF)from the original on 2016-03-04.Retrieved2014-07-12.
- ^"[Xen-devel] [RFC PATCH 0/9] Introduce AMD SVM AVIC".mail-archive.Archivedfrom the original on 2 February 2017.Retrieved4 May2018.
- ^Jun Nakajimaa (2012-12-13)."Reviewing Unused and New Features for Interrupt/APIC Virtualization"(PDF).Intel.Archived(PDF)from the original on 2015-04-21.Retrieved2014-07-12.
- ^Khang Nguyen (2013-12-17)."APIC Virtualization Performance Testing and Iozone".software.intel.Archivedfrom the original on 2014-07-14.Retrieved2014-07-12.
- ^"Product Brief Intel Xeon Processor E5-4600 v2 Product Family"(PDF).Intel. 2014-03-14.Archived(PDF)from the original on 2014-07-14.Retrieved2014-07-12.
- ^"Intel platform hardware support for I/O virtualization".Intel. 2006-08-10.Archivedfrom the original on 2007-01-20.Retrieved2012-02-04.
- ^"Linux virtualization and PCI passthrough".IBM. Archived fromthe originalon 1 November 2009.Retrieved10 November2010.
- ^"AMD I/O Virtualization Technology (IOMMU) Specification Revision 1.26"(PDF).Archived(PDF)from the original on 2011-01-24.Retrieved2011-05-24.
- ^"Intel Virtualization Technology for Directed I/O (VT-d) Architecture Specification".Archived fromthe originalon 2013-04-03.Retrieved2012-02-04.
- ^"Intel Virtualization Technology for Directed I/O (VT-d) Supported CPU List".Ark.intel. Archived fromthe originalon 2010-10-27.Retrieved2012-02-04.
- ^"PCI-SIG Engineering Change Notice: Function Level Reset (FLR)"(PDF).pcisig.2006-06-27.Archived(PDF)from the original on 2016-03-04.Retrieved2014-01-10.
- ^"Xen VT-d".xen.org.2013-06-06.Archivedfrom the original on 2014-02-09.Retrieved2014-01-10.
- ^"How to assign devices with VT-d in KVM".linux-kvm.org.2014-04-23.Archivedfrom the original on 2015-03-10.Retrieved2015-03-05.
- ^"Intel Virtualization Technology for Connectivity (VT-c)"(PDF).Intel.Archived(PDF)from the original on 2016-02-22.Retrieved2018-02-14.
- ^"PCI-SIG I/O Virtualization (IOV) Specifications".Pcisig. 2011-03-31. Archived fromthe originalon 2012-01-15.Retrieved2012-02-04.
- ^ab"Intel Look Inside: Intel Ethernet"(PDF).Intel.November 27, 2014. p. 104. Archived fromthe original(PDF)on March 4, 2016.RetrievedMarch 26,2015.
- ^Yaozu Dong; Zhao Yu; Greg Rose (2008)."SR-IOV Networking in Xen: Architecture, Design and Implementation".usenix.org.USENIX.Archivedfrom the original on 2014-01-09.Retrieved2014-01-10.
- ^Patrick Kutch; Brian Johnson; Greg Rose (September 2011)."An Introduction to Intel Flexible Port Partitioning Using SR-IOV Technology"(PDF).Intel.Archived fromthe original(PDF)on August 7, 2015.RetrievedSeptember 24,2015.
- ^"NASA's Flexible Cloud Fabric: Moving Cluster Applications to the Cloud"(PDF).Intel.Archived fromthe original(PDF)on 2012-12-22.Retrieved2014-01-08.
- ^"Enhanced Networking in the AWS Cloud".Scalable Logic. 2013-12-31.Archivedfrom the original on 2014-01-09.Retrieved2014-01-08.
- ^"Enhanced Networking in the AWS Cloud - Part 2".Scalable Logic. 2013-12-31.Archivedfrom the original on 2014-01-10.Retrieved2014-01-08.