On 22/03/2023 16:46, Borislav Petkov wrote: > On Mon, Mar 20, 2023 at 07:19:48PM +0000, Jeremi Piotrowski wrote: >> This series is a prerequisite for nested SNP-host support on Hyper-V > > I'm curious: what in the *world* is a sensible use case for doing this > thing at all? > This is actually not as crazy as it sounds. What this does is it allows a normal (non-SNP) VM to host confidential (SNP) VMs. I say "normal" but not every VM is going to be able to do this, it needs to be running on AMD hardware and configured to have access to VirtualizationExtensions, a "HardwareIsolation" capability, and given a number of "hardware isolated guests" that it is allowed to spawn. In practice this will result in the VM seeing a PSP device, SEV-SNP related CPUID leafs, and have access to additional memory management instructions (rmpadjust/psmash). This allows the rest of the of KVM-SNP support to work. So instead of taking a bare-metal AMD server with 128 CPUs to run confidential workloads you'll be able to provision an Azure VM with say 8 CPUs and run up to 8 SNP guests nested inside it. It's also useful for development, I participate in the kata-containers project where we're doing confidential-containers related work, and having access to test VMs to run SNP guests is going to make things much easier. If you're interested, I posted the other half of the patches required some time back: https://lore.kernel.org/lkml/20230213103402.1189285-1-jpiotrowski@xxxxxxxxxxxxxxxxxxx/#t Jeremi
