| Message ID | 1565854883-27019-3-git-send-email-pbonzini@redhat.com (mailing list archive) |
|---|---|
| State | New, archived |
| Headers | show |
| Series | KVM: x86: fixes for AMD speculation bug CPUID leaf | expand |
On Thu, Aug 15, 2019 at 09:41:23AM +0200, Paolo Bonzini wrote: > Even though it is preferrable to use SPEC_CTRL (represented by > X86_FEATURE_AMD_SSBD) instead of VIRT_SPEC, VIRT_SPEC is always > supported anyway because otherwise it would be impossible to > migrate from old to new CPUs. Make this apparent in the > result of KVM_GET_SUPPORTED_CPUID as well. > > Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> > Reported-by: Eduardo Habkost <ehabkost@redhat.com> > Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> > --- > arch/x86/kvm/cpuid.c | 10 ++++++---- > 1 file changed, 6 insertions(+), 4 deletions(-) > > diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c > index 145ec050d45d..5865bc73bbb5 100644 > --- a/arch/x86/kvm/cpuid.c > +++ b/arch/x86/kvm/cpuid.c > @@ -747,11 +747,13 @@ static inline int __do_cpuid_func(struct kvm_cpuid_entry2 *entry, u32 function, > entry->ebx &= kvm_cpuid_8000_0008_ebx_x86_features; > cpuid_mask(&entry->ebx, CPUID_8000_0008_EBX); > /* > - * The preference is to use SPEC CTRL MSR instead of the > - * VIRT_SPEC MSR. > + * VIRT_SPEC is only implemented for AMD processors, > + * but the host could set AMD_SSBD if it wanted even > + * for Intel processors. > */ > - if (boot_cpu_has(X86_FEATURE_LS_CFG_SSBD) && > - !boot_cpu_has(X86_FEATURE_AMD_SSBD)) > + if ((boot_cpu_has(X86_FEATURE_LS_CFG_SSBD) || > + boot_cpu_has(X86_FEATURE_AMD_SSBD)) && > + boot_cpu_has(X86_FEATURE_SVM)) Would it be desirable to move this code to svm_set_supported_cpuid(), or is there a reason for keeping this in cpuid.c? > entry->ebx |= F(VIRT_SSBD); > break; > } > -- > 1.8.3.1 >
On 15/08/19 19:14, Eduardo Habkost wrote: > On Thu, Aug 15, 2019 at 09:41:23AM +0200, Paolo Bonzini wrote: >> Even though it is preferrable to use SPEC_CTRL (represented by >> X86_FEATURE_AMD_SSBD) instead of VIRT_SPEC, VIRT_SPEC is always >> supported anyway because otherwise it would be impossible to >> migrate from old to new CPUs. Make this apparent in the >> result of KVM_GET_SUPPORTED_CPUID as well. >> >> Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> >> Reported-by: Eduardo Habkost <ehabkost@redhat.com> >> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> >> --- >> arch/x86/kvm/cpuid.c | 10 ++++++---- >> 1 file changed, 6 insertions(+), 4 deletions(-) >> >> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c >> index 145ec050d45d..5865bc73bbb5 100644 >> --- a/arch/x86/kvm/cpuid.c >> +++ b/arch/x86/kvm/cpuid.c >> @@ -747,11 +747,13 @@ static inline int __do_cpuid_func(struct kvm_cpuid_entry2 *entry, u32 function, >> entry->ebx &= kvm_cpuid_8000_0008_ebx_x86_features; >> cpuid_mask(&entry->ebx, CPUID_8000_0008_EBX); >> /* >> - * The preference is to use SPEC CTRL MSR instead of the >> - * VIRT_SPEC MSR. >> + * VIRT_SPEC is only implemented for AMD processors, >> + * but the host could set AMD_SSBD if it wanted even >> + * for Intel processors. >> */ >> - if (boot_cpu_has(X86_FEATURE_LS_CFG_SSBD) && >> - !boot_cpu_has(X86_FEATURE_AMD_SSBD)) >> + if ((boot_cpu_has(X86_FEATURE_LS_CFG_SSBD) || >> + boot_cpu_has(X86_FEATURE_AMD_SSBD)) && >> + boot_cpu_has(X86_FEATURE_SVM)) > > Would it be desirable to move this code to > svm_set_supported_cpuid(), or is there a reason for keeping this > in cpuid.c? Yes, of course. Forgot about it. Paolo
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c index 145ec050d45d..5865bc73bbb5 100644 --- a/arch/x86/kvm/cpuid.c +++ b/arch/x86/kvm/cpuid.c @@ -747,11 +747,13 @@ static inline int __do_cpuid_func(struct kvm_cpuid_entry2 *entry, u32 function, entry->ebx &= kvm_cpuid_8000_0008_ebx_x86_features; cpuid_mask(&entry->ebx, CPUID_8000_0008_EBX); /* - * The preference is to use SPEC CTRL MSR instead of the - * VIRT_SPEC MSR. + * VIRT_SPEC is only implemented for AMD processors, + * but the host could set AMD_SSBD if it wanted even + * for Intel processors. */ - if (boot_cpu_has(X86_FEATURE_LS_CFG_SSBD) && - !boot_cpu_has(X86_FEATURE_AMD_SSBD)) + if ((boot_cpu_has(X86_FEATURE_LS_CFG_SSBD) || + boot_cpu_has(X86_FEATURE_AMD_SSBD)) && + boot_cpu_has(X86_FEATURE_SVM)) entry->ebx |= F(VIRT_SSBD); break; }
Even though it is preferrable to use SPEC_CTRL (represented by X86_FEATURE_AMD_SSBD) instead of VIRT_SPEC, VIRT_SPEC is always supported anyway because otherwise it would be impossible to migrate from old to new CPUs. Make this apparent in the result of KVM_GET_SUPPORTED_CPUID as well. Cc: Konrad Rzeszutek Wilk <konrad.wilk@oracle.com> Reported-by: Eduardo Habkost <ehabkost@redhat.com> Signed-off-by: Paolo Bonzini <pbonzini@redhat.com> --- arch/x86/kvm/cpuid.c | 10 ++++++---- 1 file changed, 6 insertions(+), 4 deletions(-)