On 20.08.2020 17:08, Roger Pau Monne wrote: > @@ -181,6 +182,18 @@ int guest_rdmsr(struct vcpu *v, uint32_t msr, uint64_t > *val) > /* Not offered to guests. */ > goto gp_fault; > > + case MSR_IA32_FEATURE_CONTROL: > + if ( !(cp->x86_vendor & X86_VENDOR_INTEL) ) > + goto gp_fault;
Can we really do it this way round, rather than raising #GP when we know the MSR isn't there (AMD / Hygon)? I realized code e.g. ... > + *val = IA32_FEATURE_CONTROL_LOCK; > + if ( vmce_has_lmce(v) ) > + *val |= IA32_FEATURE_CONTROL_LMCE_ON; > + if ( nestedhvm_enabled(d) ) > + *val |= IA32_FEATURE_CONTROL_ENABLE_VMXON_OUTSIDE_SMX; > + break; > + > + > case MSR_IA32_PLATFORM_ID: > if ( !(cp->x86_vendor & X86_VENDOR_INTEL) || > !(boot_cpu_data.x86_vendor & X86_VENDOR_INTEL) ) ... in context right here does it the same way, but I still wonder whether we wouldn't better switch existing instances, too. Jan
