Re: [Qemu-devel] [PATCH 3/7] KVM: i386: Add support for KVM_CAP_EXCEPTION_PAYLOAD

[Maran Wilson <maran.wilson@oracle.com>]

On 6/17/2019 10:27 AM, Paolo Bonzini wrote:
> On 17/06/19 13:34, Liran Alon wrote:
>> Putting this all together, in case kernel doesn’t support extracting
>> nested-state, there is no decent way to know if guest is running
>> nested-virtualization. Which means that in theory we always need to
>> fail migration in case kernel doesn’t support KVM_CAP_NESTED_STATE or
>> KVM_CAP_EXCEPTION_PAYLOAD and vCPU is exposed with VMX/SVM
>> capability.
> For VMX this would be okay because we had a blocker before this series,
> and this wouldn't be any worse.

I agree it shouldn't be a gating issue for this patch series, but I'd 
hate to see this discussion thread die off.

I'm still pretty interested in hearing whether anyone has any good ideas 
for how to conclusively determine whether a given L1 VM has created a 
nested L2 or not when the host is running an older Kernel that doesn't 
support KVM_CAP_NESTED_STATE. That would be a very useful capability, 
especially for CSP use cases. If anyone has any suggestions about where 
to look, I don't mind spending some time digging into it and possibly 
testing out a few ideas. Again, separate from this particular patch 
series. So far I've been drawing a blank after Liran pointed out that 
corner case problems associated with env->cr[4] & CR4_VMXE_MASK.

Thanks,
-Maran

> For SVM we can ignore the case and fix it when we have
> KVM_CAP_NESTED_STATE, as again that wouldn't be any worse.
>
> Paolo
>
>> I can condition this behaviour with a flag that can be manipulated
>> using QMP to allow user to indicate it wishes to migrate guest anyway
>> in this case. This however bring me back to the entire discussion I
>> had with Dr. David Alan Gilbert on migration backwards compatibility
>> in general and the fact that I believe we should have a generic QMP
>> command which allows to provide list of VMState subsections that can
>> be ignored in migration… See:
>> https://www.mail-archive.com/qemu-devel@nongnu.org/msg622274.html
>>
>> Paolo, What are your thoughts on how I would proceed with this?