On Fri, 2023-11-10 at 15:55 -0800, Sean Christopherson wrote: > Restrict XSAVE in guest cpu_caps so that XSAVES dependencies on XSAVE are > automatically handled instead of manually checking for host and guest > XSAVE support. Aside from modifying XSAVE in cpu_caps, this should be a > glorified nop as KVM doesn't query guest XSAVE support (which is also why > it wasn't/isn't a bug to leave XSAVE set in guest CPUID). > > Signed-off-by: Sean Christopherson <seanjc@xxxxxxxxxx> > --- > arch/x86/kvm/svm/svm.c | 2 +- > arch/x86/kvm/vmx/vmx.c | 4 ++-- > 2 files changed, 3 insertions(+), 3 deletions(-) > > diff --git a/arch/x86/kvm/svm/svm.c b/arch/x86/kvm/svm/svm.c > index 9e3a9191dac1..6fe2d7bf4959 100644 > --- a/arch/x86/kvm/svm/svm.c > +++ b/arch/x86/kvm/svm/svm.c > @@ -4315,8 +4315,8 @@ static void svm_vcpu_after_set_cpuid(struct kvm_vcpu *vcpu) > * XSS on VM-Enter/VM-Exit. Failure to do so would effectively give > * the guest read/write access to the host's XSS. > */ > + guest_cpu_cap_restrict(vcpu, X86_FEATURE_XSAVE); > guest_cpu_cap_change(vcpu, X86_FEATURE_XSAVES, > - boot_cpu_has(X86_FEATURE_XSAVE) && > boot_cpu_has(X86_FEATURE_XSAVES) && > guest_cpu_cap_has(vcpu, X86_FEATURE_XSAVE)); > > diff --git a/arch/x86/kvm/vmx/vmx.c b/arch/x86/kvm/vmx/vmx.c > index 815692dc0aff..7645945af5c5 100644 > --- a/arch/x86/kvm/vmx/vmx.c > +++ b/arch/x86/kvm/vmx/vmx.c > @@ -7752,8 +7752,8 @@ static void vmx_vcpu_after_set_cpuid(struct kvm_vcpu *vcpu) > * to the guest. XSAVES depends on CR4.OSXSAVE, and CR4.OSXSAVE can be > * set if and only if XSAVE is supported. > */ > - if (boot_cpu_has(X86_FEATURE_XSAVE) && > - guest_cpu_cap_has(vcpu, X86_FEATURE_XSAVE)) > + guest_cpu_cap_restrict(vcpu, X86_FEATURE_XSAVE); > + if (guest_cpu_cap_has(vcpu, X86_FEATURE_XSAVE)) > guest_cpu_cap_restrict(vcpu, X86_FEATURE_XSAVES); > else > guest_cpu_cap_clear(vcpu, X86_FEATURE_XSAVES); Reviewed-by: Maxim Levitsky <mlevitsk@xxxxxxxxxx> Best regards, Maxim Levitsky
