Bug 1375893

The following kernel options are available: Hostdev passthrough & SR-IOV - Enables support for IOMMU in kernel. The option itself is *not* enough to enable whole hostdev and SR-IOV, one must consult documentation regarding bios setting and hardware support for these features. Roughly translates to intel_iommu (Intel) or amd_iommu (AMD) on kernel command line. Nested Virtualization - Enables support for nested virtualization. The option only configures the KVM module, to use nested virtualization the nestedvt hook must be installed (to propagate the vmx (Intel) and svm (AMD) cpu flags). Unsafe Interrupts - Enables hostdev passthrough on host without proper hardware protection for device assignment. This option should be left unchecked by default, and only considered if VM with hostdev passthrough fails to start and output of dmesg on the host contains " No interrupt remapping support. Use the module param "allow_unsafe_interrupts" to enable VFIO IOMMU support on this platform ". Only trusted guests should be run in this configuration[1] PCI Reallocation - Enables PCI reallocation, sometimes enabling SR-IOV devices that could not create VFs to do so. The root cause of the issue is kernel's inability to allocate enough MMIO space due to a limitation or BIOS issue. [1] http://vfio.blogspot.cz/2014/08/vfiovga-faq.html