1*c1eaa6d0SDavid WoodhouseXen HVM guest support 2*c1eaa6d0SDavid Woodhouse===================== 3*c1eaa6d0SDavid Woodhouse 4*c1eaa6d0SDavid Woodhouse 5*c1eaa6d0SDavid WoodhouseDescription 6*c1eaa6d0SDavid Woodhouse----------- 7*c1eaa6d0SDavid Woodhouse 8*c1eaa6d0SDavid WoodhouseKVM has support for hosting Xen guests, intercepting Xen hypercalls and event 9*c1eaa6d0SDavid Woodhousechannel (Xen PV interrupt) delivery. This allows guests which expect to be 10*c1eaa6d0SDavid Woodhouserun under Xen to be hosted in QEMU under Linux/KVM instead. 11*c1eaa6d0SDavid Woodhouse 12*c1eaa6d0SDavid WoodhouseSetup 13*c1eaa6d0SDavid Woodhouse----- 14*c1eaa6d0SDavid Woodhouse 15*c1eaa6d0SDavid WoodhouseXen mode is enabled by setting the ``xen-version`` property of the KVM 16*c1eaa6d0SDavid Woodhouseaccelerator, for example for Xen 4.10: 17*c1eaa6d0SDavid Woodhouse 18*c1eaa6d0SDavid Woodhouse.. parsed-literal:: 19*c1eaa6d0SDavid Woodhouse 20*c1eaa6d0SDavid Woodhouse |qemu_system| --accel kvm,xen-version=0x4000a 21*c1eaa6d0SDavid Woodhouse 22*c1eaa6d0SDavid WoodhouseAdditionally, virtual APIC support can be advertised to the guest through the 23*c1eaa6d0SDavid Woodhouse``xen-vapic`` CPU flag: 24*c1eaa6d0SDavid Woodhouse 25*c1eaa6d0SDavid Woodhouse.. parsed-literal:: 26*c1eaa6d0SDavid Woodhouse 27*c1eaa6d0SDavid Woodhouse |qemu_system| --accel kvm,xen-version=0x4000a --cpu host,+xen_vapic 28*c1eaa6d0SDavid Woodhouse 29*c1eaa6d0SDavid WoodhouseWhen Xen support is enabled, QEMU changes hypervisor identification (CPUID 30*c1eaa6d0SDavid Woodhouse0x40000000..0x4000000A) to Xen. The KVM identification and features are not 31*c1eaa6d0SDavid Woodhouseadvertised to a Xen guest. If Hyper-V is also enabled, the Xen identification 32*c1eaa6d0SDavid Woodhousemoves to leaves 0x40000100..0x4000010A. 33*c1eaa6d0SDavid Woodhouse 34*c1eaa6d0SDavid WoodhouseThe Xen platform device is enabled automatically for a Xen guest. This allows 35*c1eaa6d0SDavid Woodhousea guest to unplug all emulated devices, in order to use Xen PV block and network 36*c1eaa6d0SDavid Woodhousedrivers instead. Note that until the Xen PV device back ends are enabled to work 37*c1eaa6d0SDavid Woodhousewith Xen mode in QEMU, that is unlikely to cause significant joy. Linux guests 38*c1eaa6d0SDavid Woodhousecan be dissuaded from this by adding 'xen_emul_unplug=never' on their command 39*c1eaa6d0SDavid Woodhouseline, and it can also be noted that AHCI disk controllers are exempt from being 40*c1eaa6d0SDavid Woodhouseunplugged, as are passthrough VFIO PCI devices. 41*c1eaa6d0SDavid Woodhouse 42*c1eaa6d0SDavid WoodhouseProperties 43*c1eaa6d0SDavid Woodhouse---------- 44*c1eaa6d0SDavid Woodhouse 45*c1eaa6d0SDavid WoodhouseThe following properties exist on the KVM accelerator object: 46*c1eaa6d0SDavid Woodhouse 47*c1eaa6d0SDavid Woodhouse``xen-version`` 48*c1eaa6d0SDavid Woodhouse This property contains the Xen version in ``XENVER_version`` form, with the 49*c1eaa6d0SDavid Woodhouse major version in the top 16 bits and the minor version in the low 16 bits. 50*c1eaa6d0SDavid Woodhouse Setting this property enables the Xen guest support. 51*c1eaa6d0SDavid Woodhouse 52*c1eaa6d0SDavid Woodhouse``xen-evtchn-max-pirq`` 53*c1eaa6d0SDavid Woodhouse Xen PIRQs represent an emulated physical interrupt, either GSI or MSI, which 54*c1eaa6d0SDavid Woodhouse can be routed to an event channel instead of to the emulated I/O or local 55*c1eaa6d0SDavid Woodhouse APIC. By default, QEMU permits only 256 PIRQs because this allows maximum 56*c1eaa6d0SDavid Woodhouse compatibility with 32-bit MSI where the higher bits of the PIRQ# would need 57*c1eaa6d0SDavid Woodhouse to be in the upper 64 bits of the MSI message. For guests with large numbers 58*c1eaa6d0SDavid Woodhouse of PCI devices (and none which are limited to 32-bit addressing) it may be 59*c1eaa6d0SDavid Woodhouse desirable to increase this value. 60*c1eaa6d0SDavid Woodhouse 61*c1eaa6d0SDavid Woodhouse``xen-gnttab-max-frames`` 62*c1eaa6d0SDavid Woodhouse Xen grant tables are the means by which a Xen guest grants access to its 63*c1eaa6d0SDavid Woodhouse memory for PV back ends (disk, network, etc.). Since QEMU only supports v1 64*c1eaa6d0SDavid Woodhouse grant tables which are 8 bytes in size, each page (each frame) of the grant 65*c1eaa6d0SDavid Woodhouse table can reference 512 pages of guest memory. The default number of frames 66*c1eaa6d0SDavid Woodhouse is 64, allowing for 32768 pages of guest memory to be accessed by PV backends 67*c1eaa6d0SDavid Woodhouse through simultaneous grants. For guests with large numbers of PV devices and 68*c1eaa6d0SDavid Woodhouse high throughput, it may be desirable to increase this value. 69*c1eaa6d0SDavid Woodhouse 70*c1eaa6d0SDavid WoodhouseOS requirements 71*c1eaa6d0SDavid Woodhouse--------------- 72*c1eaa6d0SDavid Woodhouse 73*c1eaa6d0SDavid WoodhouseThe minimal Xen support in the KVM accelerator requires the host to be running 74*c1eaa6d0SDavid WoodhouseLinux v5.12 or newer. Later versions add optimisations: Linux v5.17 added 75*c1eaa6d0SDavid Woodhouseacceleration of interrupt delivery via the Xen PIRQ mechanism, and Linux v5.19 76*c1eaa6d0SDavid Woodhouseaccelerated Xen PV timers and inter-processor interrupts (IPIs). 77