See patch 3/3 for a description of exactly why we need this. I know POWER folks are also interested in making use of VFIO's external user interface from KVM and Alexey's proposed patches have a similar device tailored for SPAPR use there. I'm hoping that we can make the base device common and extend it for each architecture. The problem we're solving on x86 may or may not be architecture specific. It has to do with how the IOMMU handles No-Snoop transactions on PCIe in relation to how KVM handles vCPU cache operations. Appreciate any feedback and suggestions on how to make this more palatable for sharing with POWER. I'll also send a QEMU RFC which makes use of this. Thanks, Alex --- Alex Williamson (3): kvm: Destroy & free KVM devices on release vfio: Add check extension interface to external user support kvm: Add VFIO device for handling IOMMU cache coherency Documentation/virtual/kvm/devices/vfio.txt | 22 +++ arch/powerpc/kvm/book3s_xics.c | 1 arch/x86/include/asm/kvm_host.h | 1 arch/x86/kvm/Makefile | 2 arch/x86/kvm/vmx.c | 5 - arch/x86/kvm/x86.c | 5 - drivers/vfio/vfio.c | 8 + drivers/vfio/vfio_iommu_type1.c | 4 include/linux/kvm_host.h | 1 include/linux/vfio.h | 2 include/uapi/linux/kvm.h | 4 include/uapi/linux/vfio.h | 1 virt/kvm/kvm_main.c | 8 + virt/kvm/vfio.c | 237 ++++++++++++++++++++++++++++ 14 files changed, 295 insertions(+), 6 deletions(-) create mode 100644 Documentation/virtual/kvm/devices/vfio.txt create mode 100644 virt/kvm/vfio.c -- To unsubscribe from this list: send the line "unsubscribe kvm" in the body of a message to majordomo@xxxxxxxxxxxxxxx More majordomo info at http://vger.kernel.org/majordomo-info.html