XEN
XEN Distributions
| 2003 | Xen Project | University of Cambridge | https://xenproject.org/ | https://xenbits.xen.org/gitweb/ |
| 2007 | Citrix XenServer | Citrix acquired XenSource | ||
| 2010 | XCP (Xen Cloud Platform) | Citrix opensource (hypervisor + management) | ||
| 2018 | XCP-ng | forked from XCP after citrix restrictions.
Xen API (XAPI) / Xen Orchestra |
https://xenproject.org/projects/xcp-ng/ | |
| 2022 | XenServer | Citrix | https://www.xenserver.com/story |
XEN is a kernel, booted by GRUB, which can then boot a Linux kernel as Dom0. The first operating system that can control the XEN kernel and boot a privileged virtual machine as DomU. VM's are called domains, which have a name and an id. On the DomU (host), XEN can be controlled on the DomU with the command xe, xm or xl depending on the Xen version. The unprivileged vm's are named Dom0
xl info xl create /etc/xen/win.hvm xl list xl destroy 2
XEN can emulate devices (HVM) or paravirtualize devices (PV). With PV you need the guest (Dom0) to be aware that it is a virtual machine and have the guest tools with the drivers involved. PV is faster than HVM.
https://wiki.xenproject.org/wiki/Understanding_the_Virtualization_Spectrum
XEN supports the fully emulated HVM network interface Intel E1000 and Realtek RTL8137C2+, which emulates the hardware, so that the default OS drivers can be used. The interface will appear as vif25.1-emu on the host.
XEN PV paravirtualized devices (VIF) that use XEN drivers to talk directly to the XEN backends (e.g. XENQEMU) will show up as vif25.0
vif = ['bridge=lan,model=e1000,rate=10Gb/s','bridge=lan,type=vif']
https://wiki.xenproject.org/wiki/Xen_Networking
https://xenbits.xen.org/docs/unstable/man/xl-network-configuration.5.html
https://xenproject.org/resources/downloads/
https://github.com/xcp-ng/win-pv-drivers/releases
The drivers are not signed, to install them on windows the testsigning must be switch on, otherwise the driver installation will fail without error message.
bcdedit /set testsigning on
xl network-list 25 Idx BE Mac Addr. handle state evt-ch tx-/rx-ring-ref BE-path 0 0 00:16:3e:5b:b7:e2 0 4 -1 -1/-1 /local/domain/0/backend/vif/25/0 1 0 00:16:3e:4d:95:03 1 4 -1 -1/-1 /local/domain/0/backend/vif/25/1
Alpine Linux
Alpine Linux downloads come with a XEN ISO that can be used, but using a regulare version can also run XEN
https://wiki.alpinelinux.org/wiki/Xen_Dom0
setup-xen-dom0
apk add xen-hypervisor apk add bridge apk add seabios ovmf
echo "xen-netback" >> /etc/modules echo "xen-blkback" >> /etc/modules echo "tun" >> /etc/modules
rc-update add xenconsoled rc-update add xendomains rc-update add xenqemu rc-update add xenstored
Contents of /etc/default/grub
grub-mkconfig -o /boot/grub/grub.cfg grub-set-default "$(grep ^menuentry /boot/grub/grub.cfg | grep Xen | cut -d \' -f 2 | head -1)"
efibootmgr -c -d /dev/sda -p 1 -l /EFI/boot/xen.efi -L "XEN"
efibootmgr -o 2,0,4
https://wiki.xenproject.org/wiki/Xen_EFI
https://wiki.archlinux.org/title/xen
https://xcp-ng.org/docs/architecture.html#api
PCI passthrough was indeed necessary for the software I'm using to pick up all network traffic properly (not bridging through the virtual switch). Doing so requires enabling IOMMU and using the legacy linux bridge backend networking (instead of the default Open vSwitch backend networking).
xl network-list opnsense Idx BE Mac Addr. handle state evt-ch tx-/rx-ring-ref BE-path 0 0 00:16:3e:3f:4f:61 0 5 -1 -1/-1 /local/domain/0/backend/vif/11/0
https://xenbits.xen.org/docs/4.15-testing/man/xl-network-configuration.5.html
https://wiki.xenproject.org/wiki/Device_Model_Stub_Domains
type=ioemu vifspec model=e1000 rate=10Gb/s
vif = ['bridge=lan,model=e1000,rate=10Gb/s']
device_model_stubdomain_override