On Wed, 2021-10-20 at 17:11 +0200, Paolo Bonzini wrote: > On 17/08/21 11:31, Robert Hoo wrote: > > + vmcs12_field_update_by_vmexit_ctrl(vmx- > > >nested.msrs.entry_ctls_high, > > + *highp, data >> 32, > > + vmx- > > >nested.vmcs12_field_existence_bitmap); > > + break; > > + case MSR_IA32_VMX_TRUE_ENTRY_CTLS: > > + vmcs12_field_update_by_vmentry_ctrl(vmx- > > >nested.msrs.exit_ctls_high, > > + *highp, data >> 32, > > + vmx- > > >nested.vmcs12_field_existence_bitmap); > > These two functions maybe could be merged into just one, since there > are > going to be duplicate checks. Can I keep them? I think this is trivial, and separating them looks more clear, from logical perspective.:-) A summary question: am I going to send v2? since I'm not sure about Sean and Jim's decision on whether to implement the interaction with shadow VMCS (which will have to consume 8KiB more memory for each vmx). And, Jim mentioned they have some virtualizing shadow vmcs patches which is going to be sent to community. Should I wait for their patches? > > Paolo >
