[v6,00/33] KVM: VMX: Support updated eVMCSv1 revision + use vmcs_config for L1 VMX MSRs

Message

Vitaly Kuznetsov Aug. 30, 2022, 1:37 p.m. UTC

Changes since "[RFC PATCH v6 00/36] KVM: x86: eVMCS rework":
- Drop the most controversial TSC_SCALING enablement for Hyper-V on KVM:
  - "KVM: nVMX: Enforce unsupported eVMCS in VMX MSRs for host accesses" patch dropped.
  - "KVM: nVMX: Support TSC scaling with enlightened VMCS" patch dropped.
  - "KVM: selftests: Enable TSC scaling in evmcs selftest" patch dropped.

Original description:

Enlightened VMCS v1 definition was updates to include fields for the
following features:
    - PerfGlobalCtrl
    - EnclsExitingBitmap
    - TSC scaling
    - GuestLbrCtl
    - CET
    - SSP
While the information is missing in the publicly available TLFS, the
updated definition comes with a new feature bit in CPUID.0x4000000A.EBX
(BIT 0) for PerfGlobalCtrl.

Add support for the new revision to KVM. SSP, CET and GuestLbrCtl
features are not currently supported by KVM. Note: for Hyper-V on KVM,
only PerfGlobalCtrl is added as it has a dedicated CPUID bit. The way
how to enable different layouts of eVMSC in a VMM friendly way is still
under discussion.

While on it, implement Sean's idea to use vmcs_config for setting up
L1 VMX control MSRs instead of re-reading host MSRs.

Jim Mattson (1):
  KVM: x86: VMX: Replace some Intel model numbers with mnemonics

Sean Christopherson (9):
  KVM: x86: Check for existing Hyper-V vCPU in kvm_hv_vcpu_init()
  KVM: x86: Report error when setting CPUID if Hyper-V allocation fails
  KVM: nVMX: Treat eVMCS as enabled for guest iff Hyper-V is also
    enabled
  KVM: nVMX: Use CC() macro to handle eVMCS unsupported controls checks
  KVM: nVMX: WARN once and fail VM-Enter if eVMCS sees VMFUNC[63:32] !=
    0
  KVM: nVMX: Don't propagate vmcs12's PERF_GLOBAL_CTRL settings to
    vmcs02
  KVM: nVMX: Always emulate PERF_GLOBAL_CTRL VM-Entry/VM-Exit controls
  KVM: VMX: Don't toggle VM_ENTRY_IA32E_MODE for 32-bit kernels/KVM
  KVM: VMX: Adjust CR3/INVPLG interception for EPT=y at runtime, not
    setup

Vitaly Kuznetsov (23):
  x86/hyperv: Fix 'struct hv_enlightened_vmcs' definition
  x86/hyperv: Update 'struct hv_enlightened_vmcs' definition
  KVM: x86: Zero out entire Hyper-V CPUID cache before processing
    entries
  KVM: nVMX: Refactor unsupported eVMCS controls logic to use 2-d array
  KVM: VMX: Define VMCS-to-EVMCS conversion for the new fields
  KVM: nVMX: Support several new fields in eVMCSv1
  KVM: x86: hyper-v: Cache HYPERV_CPUID_NESTED_FEATURES CPUID leaf
  KVM: selftests: Add ENCLS_EXITING_BITMAP{,HIGH} VMCS fields
  KVM: selftests: Switch to updated eVMCSv1 definition
  KVM: nVMX: Support PERF_GLOBAL_CTRL with enlightened VMCS
  KVM: VMX: Get rid of eVMCS specific VMX controls sanitization
  KVM: VMX: Check VM_ENTRY_IA32E_MODE in setup_vmcs_config()
  KVM: VMX: Check CPU_BASED_{INTR,NMI}_WINDOW_EXITING in
    setup_vmcs_config()
  KVM: VMX: Tweak the special handling of SECONDARY_EXEC_ENCLS_EXITING
    in setup_vmcs_config()
  KVM: VMX: Extend VMX controls macro shenanigans
  KVM: VMX: Move CPU_BASED_CR8_{LOAD,STORE}_EXITING filtering out of
    setup_vmcs_config()
  KVM: VMX: Add missing VMEXIT controls to vmcs_config
  KVM: VMX: Add missing CPU based VM execution controls to vmcs_config
  KVM: VMX: Move LOAD_IA32_PERF_GLOBAL_CTRL errata handling out of
    setup_vmcs_config()
  KVM: nVMX: Always set required-1 bits of pinbased_ctls to
    PIN_BASED_ALWAYSON_WITHOUT_TRUE_MSR
  KVM: nVMX: Use sanitized allowed-1 bits for VMX control MSRs
  KVM: VMX: Cache MSR_IA32_VMX_MISC in vmcs_config
  KVM: nVMX: Use cached host MSR_IA32_VMX_MISC value for setting up
    nested MSR

 arch/x86/include/asm/hyperv-tlfs.h            |  22 +-
 arch/x86/include/asm/kvm_host.h               |   2 +
 arch/x86/kvm/cpuid.c                          |  18 +-
 arch/x86/kvm/hyperv.c                         |  70 +++--
 arch/x86/kvm/hyperv.h                         |   6 +-
 arch/x86/kvm/vmx/capabilities.h               |  14 +-
 arch/x86/kvm/vmx/evmcs.c                      | 192 ++++++++-----
 arch/x86/kvm/vmx/evmcs.h                      |  10 +-
 arch/x86/kvm/vmx/nested.c                     |  94 ++++---
 arch/x86/kvm/vmx/nested.h                     |   2 +-
 arch/x86/kvm/vmx/vmx.c                        | 259 ++++++++----------
 arch/x86/kvm/vmx/vmx.h                        | 172 ++++++++++--
 .../selftests/kvm/include/x86_64/evmcs.h      |  45 ++-
 .../selftests/kvm/include/x86_64/vmx.h        |   2 +
 14 files changed, 584 insertions(+), 324 deletions(-)

Comments

Sean Christopherson Sept. 17, 2022, 1:05 a.m. UTC | #1

On Tue, Aug 30, 2022, Vitaly Kuznetsov wrote:
> Changes since "[RFC PATCH v6 00/36] KVM: x86: eVMCS rework":
> - Drop the most controversial TSC_SCALING enablement for Hyper-V on KVM:
>   - "KVM: nVMX: Enforce unsupported eVMCS in VMX MSRs for host accesses" patch dropped.
>   - "KVM: nVMX: Support TSC scaling with enlightened VMCS" patch dropped.
>   - "KVM: selftests: Enable TSC scaling in evmcs selftest" patch dropped.

Pushed to branch `for_paolo/6.1` at:

    https://github.com/sean-jc/linux.git

Unless you hear otherwise, it will make its way to kvm/queue "soon".

Note, the commit IDs are not guaranteed to be stable, and in fact they are guaranteed
to not be stable right now as I need to do some surgery.