Guest handling of IA32_DEBUGCTL MSR

Guest handling of IA32_DEBUGCTL MSR

[Nadav Amit]

It seems strange that the guest is allowed to set IA32_DEBUGCTL MSR for the
nested VM and get this value to the physical IA32_DEBUGCTL (see
prepare_vmcs02), while it cannot set IA32_DEBUGCTL for itself (see
kvm_set_msr_common).

Am I missing something?

Regards,
Nadav

Re: Guest handling of IA32_DEBUGCTL MSR

[Paolo Bonzini]

On 28/04/2015 13:42, Nadav Amit wrote:
> It seems strange that the guest is allowed to set IA32_DEBUGCTL MSR for the
> nested VM and get this value to the physical IA32_DEBUGCTL (see
> prepare_vmcs02), while it cannot set IA32_DEBUGCTL for itself (see
> kvm_set_msr_common).
> 
> Am I missing something?

No, it makes no sense.

Paolo

Re: Guest handling of IA32_DEBUGCTL MSR

[Jan Kiszka]

Am 2015-04-28 um 13:43 schrieb Paolo Bonzini:
> 
> 
> On 28/04/2015 13:42, Nadav Amit wrote:
>> It seems strange that the guest is allowed to set IA32_DEBUGCTL MSR for the
>> nested VM and get this value to the physical IA32_DEBUGCTL (see
>> prepare_vmcs02), while it cannot set IA32_DEBUGCTL for itself (see
>> kvm_set_msr_common).
>>
>> Am I missing something?
> 
> No, it makes no sense.

Are you sure that vmx is not allowing direct access to that MSR while in
guest mode? We do save/restore it on all Intel CPUs, see
setup_vmcs_config. Not sure about the AMD situation, though.

Jan

-- 
Siemens AG, Corporate Technology, CT RTC ITP SES-DE
Corporate Competence Center Embedded Linux

Re: Guest handling of IA32_DEBUGCTL MSR

[Nadav Amit]

Jan Kiszka <jan.kiszka@siemens.com> wrote:

> Am 2015-04-28 um 13:43 schrieb Paolo Bonzini:
>> On 28/04/2015 13:42, Nadav Amit wrote:
>>> It seems strange that the guest is allowed to set IA32_DEBUGCTL MSR for the
>>> nested VM and get this value to the physical IA32_DEBUGCTL (see
>>> prepare_vmcs02), while it cannot set IA32_DEBUGCTL for itself (see
>>> kvm_set_msr_common).
>>> 
>>> Am I missing something?
>> 
>> No, it makes no sense.
> 
> Are you sure that vmx is not allowing direct access to that MSR while in
> guest mode? We do save/restore it on all Intel CPUs, see
> setup_vmcs_config. Not sure about the AMD situation, though.

Unless you explicitly disable the interception of the MSR in the bitmap, you
trap WRMSR to this MSR. I guess the original intention of saving/restoring
it was to address situations in which the CPU itself implicitly changes it
(e.g., clearing LBR when “freezing LBR on PMI” is set). I think that these
situations do not apply when the MSR is zeroed, as it happens to be in KVM.

Nadav

Re: Guest handling of IA32_DEBUGCTL MSR

[Paolo Bonzini]

On 28/04/2015 14:00, Jan Kiszka wrote:
> > > It seems strange that the guest is allowed to set IA32_DEBUGCTL MSR for the
> > > nested VM and get this value to the physical IA32_DEBUGCTL (see
> > > prepare_vmcs02), while it cannot set IA32_DEBUGCTL for itself (see
> > > kvm_set_msr_common).
> > >
> > > Am I missing something?
> > 
> > No, it makes no sense.
> 
> Are you sure that vmx is not allowing direct access to that MSR while in
> guest mode? We do save/restore it on all Intel CPUs, see
> setup_vmcs_config. Not sure about the AMD situation, though.

Right now, L1 could for example enable LBR for the L2 guest, but it
wouldn't make sense since the L1 guest cannot then read the last branch
MSRs.

Paolo