Re: [PATCH RFC 3/3] x86/kvm/hyper-v: don't allow to turn on unsupported VMX controls for nested guests

[Paolo Bonzini]

On 16/01/20 17:21, Sean Christopherson wrote:
> On Thu, Jan 16, 2020 at 09:55:57AM +0100, Vitaly Kuznetsov wrote:
>> Liran Alon <liran.alon@xxxxxxxxxx> writes:
>>
>>>> On 15 Jan 2020, at 19:10, Vitaly Kuznetsov <vkuznets@xxxxxxxxxx> wrote:
>>>>
>>>> Sane L1 hypervisors are not supposed to turn any of the unsupported VMX
>>>> controls on for its guests and nested_vmx_check_controls() checks for
>>>> that. This is, however, not the case for the controls which are supported
>>>> on the host but are missing in enlightened VMCS and when eVMCS is in use.
>>>>
>>>> It would certainly be possible to add these missing checks to
>>>> nested_check_vm_execution_controls()/_vm_exit_controls()/.. but it seems
>>>> preferable to keep eVMCS-specific stuff in eVMCS and reduce the impact on
>>>> non-eVMCS guests by doing less unrelated checks. Create a separate
>>>> nested_evmcs_check_controls() for this purpose.
>>>>
>>>> Signed-off-by: Vitaly Kuznetsov <vkuznets@xxxxxxxxxx>
>>>> ---
>>>> arch/x86/kvm/vmx/evmcs.c  | 56 ++++++++++++++++++++++++++++++++++++++-
>>>> arch/x86/kvm/vmx/evmcs.h  |  1 +
>>>> arch/x86/kvm/vmx/nested.c |  3 +++
>>>> 3 files changed, 59 insertions(+), 1 deletion(-)
>>>>
>>>> diff --git a/arch/x86/kvm/vmx/evmcs.c b/arch/x86/kvm/vmx/evmcs.c
>>>> index b5d6582ba589..88f462866396 100644
>>>> --- a/arch/x86/kvm/vmx/evmcs.c
>>>> +++ b/arch/x86/kvm/vmx/evmcs.c
>>>> @@ -4,9 +4,11 @@
>>>> #include <linux/smp.h>
>>>>
>>>> #include "../hyperv.h"
>>>> -#include "evmcs.h"
>>>> #include "vmcs.h"
>>>> +#include "vmcs12.h"
>>>> +#include "evmcs.h"
>>>> #include "vmx.h"
>>>> +#include "trace.h"
>>>>
>>>> DEFINE_STATIC_KEY_FALSE(enable_evmcs);
>>>>
>>>> @@ -378,6 +380,58 @@ void nested_evmcs_filter_control_msr(u32 msr_index, u64 *pdata)
>>>> 	*pdata = ctl_low | ((u64)ctl_high << 32);
>>>> }
>>>>
>>>> +int nested_evmcs_check_controls(struct vmcs12 *vmcs12)
>>>> +{
>>>> +	int ret = 0;
>>>> +	u32 unsupp_ctl;
>>>> +
>>>> +	unsupp_ctl = vmcs12->pin_based_vm_exec_control &
>>>> +		EVMCS1_UNSUPPORTED_PINCTRL;
>>>> +	if (unsupp_ctl) {
>>>> +		trace_kvm_nested_vmenter_failed(
>>>> +			"eVMCS: unsupported pin-based VM-execution controls",
>>>> +			unsupp_ctl);
>>>
>>> Why not move "CCâ macro from nested.c to nested.h and use it here as-well instead of replicating itâs logic?
>>>
>>
>> Because error messages I add are both human readable and conform to SDM!
>> :-)
>>
>> On a more serious not yes, we should probably do that. We may even want
>> to use it in non-nesting (and non VMX) code in KVM.
> 
> No, the CC() macro is short for Consistency Check, i.e. specific to nVMX.
> Even if KVM ends up adding nested_evmcs_check_controls(), which I'm hoping
> can be avoided, I'd still be hesitant to expose CC() in nested.h.
> 

For now let's keep Vitaly's patch as is.  It's definitely a good one as
it would catch Hyper-V's issue immediately even without patch 2 (which
is the only really contentious change).

Paolo