Re: [PATCH] KVM: SVM: Assume a 64-bit hypercall for guests with protected state

[Vitaly Kuznetsov <vkuznets@redhat.com>]

Tom Lendacky <thomas.lendacky@amd.com> writes:

> On 5/24/21 6:53 AM, Vitaly Kuznetsov wrote:
>> Tom Lendacky <thomas.lendacky@amd.com> writes:
>> 
>>> When processing a hypercall for a guest with protected state, currently
>>> SEV-ES guests, the guest CS segment register can't be checked to
>>> determine if the guest is in 64-bit mode. For an SEV-ES guest, it is
>>> expected that communication between the guest and the hypervisor is
>>> performed to shared memory using the GHCB. In order to use the GHCB, the
>>> guest must have been in long mode, otherwise writes by the guest to the
>>> GHCB would be encrypted and not be able to be comprehended by the
>>> hypervisor. Given that, assume that the guest is in 64-bit mode when
>>> processing a hypercall from a guest with protected state.
>>>
>>> Fixes: f1c6366e3043 ("KVM: SVM: Add required changes to support intercepts under SEV-ES")
>>> Reported-by: Sean Christopherson <seanjc@google.com>
>>> Signed-off-by: Tom Lendacky <thomas.lendacky@amd.com>
>>> ---
>>>  arch/x86/kvm/x86.c | 7 ++++++-
>>>  1 file changed, 6 insertions(+), 1 deletion(-)
>>>
>>> diff --git a/arch/x86/kvm/x86.c b/arch/x86/kvm/x86.c
>>> index 9b6bca616929..e715c69bb882 100644
>>> --- a/arch/x86/kvm/x86.c
>>> +++ b/arch/x86/kvm/x86.c
>>> @@ -8403,7 +8403,12 @@ int kvm_emulate_hypercall(struct kvm_vcpu *vcpu)
>>>  
>>>  	trace_kvm_hypercall(nr, a0, a1, a2, a3);
>>>  
>>> -	op_64_bit = is_64_bit_mode(vcpu);
>>> +	/*
>>> +	 * If running with protected guest state, the CS register is not
>>> +	 * accessible. The hypercall register values will have had to been
>>> +	 * provided in 64-bit mode, so assume the guest is in 64-bit.
>>> +	 */
>>> +	op_64_bit = is_64_bit_mode(vcpu) || vcpu->arch.guest_state_protected;
>>>  	if (!op_64_bit) {
>>>  		nr &= 0xFFFFFFFF;
>>>  		a0 &= 0xFFFFFFFF;
>> 
>> While this is might be a very theoretical question, what about other
>> is_64_bit_mode() users? Namely, a very similar to the above check exists
>> in kvm_hv_hypercall() and kvm_xen_hypercall().
>
> Xen doesn't support SEV, so I think this one is ok until they do. Although
> I guess we could be preemptive and hit all those call sites. The other
> ones are in arch/x86/kvm/hyperv.c.
>
> Thoughts?

Would it hurt if we just move 'vcpu->arch.guest_state_protected' check
to is_64_bit_mode() itself? It seems to be too easy to miss this
peculiar detail about SEV in review if new is_64_bit_mode() users are to
be added.

-- 
Vitaly