Archived/Xen Project Release Features 4.0 to 4.3

From Xen
Revision as of 11:36, 27 June 2017 by Lars.kurth (talk | contribs) (Device Models and Virtual Firmware: missed 1)

Legend: : Feature : Partial : Removed


Xen 4.0 Xen 4.1 Xen 4.2 Xen 4.3
Initial Release 7 April 2010 25 March 2011 17 Sept 2012 9 July 2013
Supported until ???[ 21 ] ???[ 21 ] ???[ 21 ] Jan 2015
Security support until ???[ 21 ] ???[ 21 ] Sept 2015 July 2016
Feature List FL 4.2 FL 4.3
Release Notes RN 4.0 RN 4.1 RN 4.2 RN 4.3
Supported Mainline Architectures for the hypervisor (Host)
IA-A32 removed
Itanium (ia64) deprecated in this release deprecated removed
ARM v7 + Virtualization Extensions preview [ 6 ]
ARM v8 preview [ 6 ]
Guest Types
For X86 Architectures
Traditional Xen Project PV guest
HVM Guest [ 1 ]
Fully virtualised guest using hardware virtualisation extensions
PV-on-HVM Guest [ 1 ]
Fully virtualised guest using PV extensions/drivers for improved performance
For ARM Architectures
ARM Guest
Optimal combination of full virtualization and PV extensions
preview [ 6 ]


Xen 4.0 Xen 4.1 Xen 4.2 Xen 4.3
Host Limits
For X86 Architectures
Physical CPUs 128 [ 0 ] >255 4095 4095
Physical RAM 1TB 5TB 5TB 16TB
For ARM Architectures
Physical CPUs 8
Physical RAM 16GB
Guest Limits
X86 PV Guest Limits
Virtual CPUs 128 >255 512 512
Virtual RAM 512GB 512GB 512GB 512GB
X86 HVM Guest Limits
Virtual CPUs 128 128 128 128
Virtual RAM 1TB 1TB 1TB 1TB
ARM Guest Limits
Virtual CPUs 8
Virtual RAM 16GB
Number of event channels
Event Channels 1024/4096 1024/4096 1024/4096 1024/4096

Toolstack and Tools

Xen 4.0 Xen 4.1 Xen 4.2 Xen 4.3
xend / xm deprecated in this release deprecated
XL initial implementation preview release
Qemu based disk backend (qdisk) for XL
[ 5 ] [ 5 ] [ 5 ]
XL Open vSwitch integration preview
3rd Party
libvirt driver for XL
debugger to debug ELF guests
Virtual Performance Management Unit for HVM guests
[ 4 ] [ 4 ] [ 4 ] [ 4 ]
Serial console Add EHCI debug support
performance analysis


Xen 4.0 Xen 4.1 Xen 4.2 Xen 4.3
Advanced Memory Management
Memory Ballooning
Memory Sharing
allow sharing of identical pages between HVM guests
preview preview preview [ 3 ] preview [ 3 ]
Memory Paging
allow pages belonging to HVM guests to be paged to disk
preview preview preview [ 3 ] preview [ 3 ]
TMEM - Transcendent Memory experimental [ 2 ] experimental [ 2 ] experimental [ 2 ] experimental [ 2 ]
Resource Management
advanced partitioning
Credit 2 Scheduler
designed for latency-sensitive workloads and very large systems.
prototype prototype experimental
NUMA scheduler affinity
1GB/2MB super page support
Deliver events to PVHVM guests using Xen event channels
Fair locks (ticket-locks)
High Availability and Fault Tolerance
Live Migration, Save & Restore
Remus Fault Tolerance
X86 vMCE
Forward Machine Check Exceptions to Appropriate guests
Network and Storage
Online resize of virtual disks
Security (also see this presentation or this document)
Driver Domains
Device Model Stub Domains
Virtual Machine Introspection (Memaccess and VM Event APIs)
enabling integration of 3rd party security solutions into Xen virtualized environments
mandatory access control policy providing fine-grained controls over Xen domains, similar to SELinux
XSM & FLASK support for IS_PRIV
vTPM Support

updates and new functionality

Interoperability / Hardware Support

Xen 4.0 Xen 4.1 Xen 4.2 Xen 4.3
X86 Nested Virtualisation
Running a hypervisor inside an HVM guest
experimental experimental
HVM PXE Stack (x86) gPXE iPXE iPXE iPXE
Physical CPU Hotplug
Physical Memory Hotplug
Support for PV kernels in bzImage format
PCI Passthrough [ 12 ]
X86 Advanced Vector eXtension (AVX) [ 4 ]

Device Models and Virtual Firmware

Xen 4.0 Xen 4.1 Xen 4.2 Xen 4.3
Device Models and Virtual Firmware for HVM guests
For X86 Architectures
Traditional Device Model
Device emulator based on Xen fork of Qemu
Qemu Upstream Device Model
Device emulator based on upstream Qemu
preview default, unless stubdomains are used
BIOS used with traditional device model only
BIOS used with upstream qemu device model and XL only
UEFI Firmware used with upstream qemu device model and XL only
experimental [ 4 ] experimental [ 4 ]
PV Bootloader support
For X86 Architectures
PyGrub support for GRUB 2
PyGrub support for /boot on ext4
pvnetboot support
Bootloader supporting network boot of PV guests


[ 0 ] Compile time limit, can be increased
[ 1 ] Requires hardware virtualisation support
[ 2 ] Disabled by default (enable with hypervisor command line option). Not supported in production and doesn't have full security support
[ 3 ] Preview, due to limited tools support. Hypervisor side in good shape
[ 4 ] Disabled by default (enable with hypervisor command line option)
[ 5 ] Used as a fallback if blkback and/or blktap2 are not available
[ 6 ] See Xen ARM with Virtualization Extensions; basic functionality implemented; lack of real hardware to validate against
[ 7 ] Limited by supported host memory
[ 8 ] Tested on Broadcom 7445D0 A15, Midway (Calxeda), Odroid-XU, OMAP5, OMAP6, DRA7 (Texas Instrument), Exynos5 (Samsung), SunXI (AllWinner) aka A20/A21, CubieTruck, CubieBoard
[ 9 ] Tested on Vexpress (ARM Ltd.), Seattle (AMD), Mustang (Applied Micro-X-Gene), McDivitt aka HP Moonshot cartridge (Applied Micro X-Gene)
[ 10 ] Remus has been ported to XL and has been tested with COLO which is currently outside the Xen tree, but planned to be included in Xen 4.6. Also see Remus Toolkit Differences explaining differences between Remus with XM (not supported in Xen 4.5) and XL
[ 11 ] There are a number of reported crashes that have not yet been fixed and some known limitations
[ 12 ] Not yet supported on ARM; for a more detailed update on what is supported on ARM see here and here
[ 13 ] The PCI passthrough features did not make it on time for Xen 4.5, but doing passthrough of MMIO regions did. In the ARM world it is quite common to have no PCIe devices and to only access devices using MMIO regions. As such this feature allows us to have driver domains be in charge of network or storage devices.
[ 14 ] Added the following platforms to [ 8 ]: Renesas R-Car Gen2, Huawei hip04-d04
[ 15 ] Added the following platforms to [ 9 ]: Thunder X (partial), Xilinx ZynqMP SoC
[ 16 ] Now tested as part of Xen Project Test Lab and OpenStack CI Loop
[ 17 ] Now tested as part of Xen Project Test Lab
[ 18 ] Migration v2: see [1] and [2]. VM migration using libvirt between two hosts is now tested. Live Migration between hosts of different Xen versions is now tested.
[ 19 ] The memory event subsystem has been reworked and extended to a new VM event subsystem. The new VM event subsystems supports both the ARM and x86 architectures.
[ 20 ] Passthrough for non-PCI devices allows users to passthrough devices via partial device trees.
[ 21 ] Dates for historical releases have not be researched.
[ 22 ] The code for hypervisor live patching is NOT enabled by default. When compiling the hypervisor one must change the config to build it. Note that Live Patching is currently only supported on x86. For more information on live patching, see LivePatch.
[ 23 ] CPU ID Levelling enables migration of VM’s between a larger range of non-identical hosts than previously supported.
[ 24 ] The COLO Manager has been fully integrated as an experimental feature into Xen 4.7. Note that the COLO Block Replication and COLO Proxy components, both of which are QEMU components, are currently still reviewed by the QEMU community. Both components are available as out-of-tree add-ons to the Xen Project Hypervisor, until fully integrated into QEMU.
[ 25 ] Added experimental support for VT-d Posted Interrupts, VMX TSC Scaling, Memory Protection Keys
[ 26 ] Removed support for Huawei hip04-d04