On 01/31/2018 01:27 AM, Jim Mattson wrote:
On Tue, Jan 30, 2018 at 4:19 PM, Paolo Bonzini <pbonzini@xxxxxxxxxx> wrote:
The new code in nested_vmx_merge_msr_bitmap should be conditional on
vmx->save_spec_ctrl_on_exit.
But then if L1 doesn't use MSR_IA32_SPEC_CTRL itself and it uses the
VM-entry MSR load list to set up L2's MSR_IA32_SPEC_CTRL, you will
never set vmx->save_spec_ctrl_on_exit, and L2's accesses to the MSR
will always be intercepted by L0.
I can add another variable (actually two) to indicate if msr
interception should be disabled or not for SPEC_CTRL and PRED_CMD in
nested case.
That would allow us to have a fast alternative to guest_cpuid_has in
nested_vmx_merge_msr_bitmap and at the same time maintain the current
semantics of save_spec_ctrl_on_exit (i.e we would still differentiate
between set_msr that is called from the loading MSRs for the emulated
vm-entry vs L2 actually writing to it).
What do you think?
Amazon Development Center Germany GmbH
Berlin - Dresden - Aachen
main office: Krausenstr. 38, 10117 Berlin
Geschaeftsfuehrer: Dr. Ralf Herbrich, Christian Schlaeger
Ust-ID: DE289237879
Eingetragen am Amtsgericht Charlottenburg HRB 149173 B
