Revert "KVM: nVMX: Expose load IA32_PERF_GLOBAL_CTRL VM-{Entry,Exit} control"

Browse source

This reverts commit 03a8871add.

Since commit 03a8871add ("KVM: nVMX: Expose load IA32_PERF_GLOBAL_CTRL
VM-{Entry,Exit} control"), KVM has taken ownership of the "load
IA32_PERF_GLOBAL_CTRL" VMX entry/exit control bits, trying to set these
bits in the IA32_VMX_TRUE_{ENTRY,EXIT}_CTLS MSRs if the guest's CPUID
supports the architectural PMU (CPUID[EAX=0Ah].EAX[7:0]=1), and clear
otherwise.

This was a misguided attempt at mimicking what commit 5f76f6f5ff
("KVM: nVMX: Do not expose MPX VMX controls when guest MPX disabled",
2018-10-01) did for MPX.  However, that commit was a workaround for
another KVM bug and not something that should be imitated.  Mucking with
the VMX MSRs creates a subtle, difficult to maintain ABI as KVM must
ensure that any internal changes, e.g. to how KVM handles _any_ guest
CPUID changes, yield the same functional result.  Therefore, KVM's policy
is to let userspace have full control of the guest vCPU model so long
as the host kernel is not at risk.

Now that KVM really truly ensures kvm_set_msr() will succeed by loading
PERF_GLOBAL_CTRL if and only if it exists, revert KVM's misguided and
roundabout behavior.

Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>
[sean: make it a pure revert]
Signed-off-by: Sean Christopherson <seanjc@google.com>
Message-Id: <20220722224409.1336532-6-seanjc@google.com>
Signed-off-by: Paolo Bonzini <pbonzini@redhat.com>

...

This commit is contained in:

Paolo Bonzini 2022-07-22 22:44:09 +00:00

parent 4496a6f9b4

commit 9389d5774a

3 changed files with 0 additions and 27 deletions

Show all changes Ignore whitespace when comparing lines Ignore changes in amount of whitespace Ignore changes in whitespace at EOL

Download patch file Download diff file Expand all files Collapse all files

22

arch/x86/kvm/vmx/nested.c

View file

2

arch/x86/kvm/vmx/nested.h

View file

3

arch/x86/kvm/vmx/pmu_intel.c

View file