[V3 PATCH 0/2] Execute hypercalls from guests according to cpu

[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

 



Confidential VMs(CVMs) need to execute hypercall instruction as per the CPU
type. Normally KVM emulates the vmcall/vmmcall instruction by patching
the guest code at runtime. Such a guest memory manipulation by KVM is
not allowed with CVMs.

This series adds support of executing hypercall as per the native cpu
type queried using cpuid instruction. CPU vendor type is stored after
one time execution of cpuid instruction to be reused later.

Changes in v3:
1) Guest logic is modified to not rely on host cpu type and instead query
cpu vendor using cpuid instruction.
2) Existing callers of vmmcall/vmcall are not updated to avoid enforcing
native hypercall instruction across all users which are mostly
non-confidential usecases.

v2:
https://lore.kernel.org/all/20220915000448.1674802-1-vannapurve@xxxxxxxxxx/

More discussion around this change:
https://lore.kernel.org/lkml/Y1Hhw40H58EmZ6lK@xxxxxxxxxx/

Vishal Annapurve (2):
  KVM: selftests: x86: Cache the cpu vendor type
  KVM: selftests: x86: Add native hypercall support

 .../selftests/kvm/include/x86_64/processor.h  |  3 ++
 .../selftests/kvm/lib/x86_64/processor.c      | 51 +++++++++++++++++--
 2 files changed, 49 insertions(+), 5 deletions(-)

-- 
2.39.0.314.g84b9a713c41-goog




[Index of Archives]     [KVM ARM]     [KVM ia64]     [KVM ppc]     [Virtualization Tools]     [Spice Development]     [Libvirt]     [Libvirt Users]     [Linux USB Devel]     [Linux Audio Users]     [Yosemite Questions]     [Linux Kernel]     [Linux SCSI]     [XFree86]

  Powered by Linux