Message ID | 20210225030951.17099-2-weijiang.yang@intel.com (mailing list archive) |
---|---|
State | New, archived |
Headers | show |
Series | [v2] KVM: nVMX: Sync L2 guest CET states between L1/L2 | expand |
+Vitaly On Thu, Feb 25, 2021, Yang Weijiang wrote: > These fields are rarely updated by L1 QEMU/KVM, sync them when L1 is trying to > read/write them and after they're changed. If CET guest entry-load bit is not > set by L1 guest, migrate them to L2 manaully. > > Suggested-by: Sean Christopherson <seanjc@google.com> > Signed-off-by: Yang Weijiang <weijiang.yang@intel.com> > > change in v2: > - Per Sean's review feedback, change CET guest states as rarely-updated fields. > And also migrate L1's CET states to L2 if the entry-load bit is not set. > - Opportunistically removed one blank line. > --- > arch/x86/kvm/cpuid.c | 1 - > arch/x86/kvm/vmx/nested.c | 29 +++++++++++++++++++++++++++++ > arch/x86/kvm/vmx/vmx.h | 3 +++ > 3 files changed, 32 insertions(+), 1 deletion(-) > > diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c > index 46087bca9418..afc97122c05c 100644 > --- a/arch/x86/kvm/cpuid.c > +++ b/arch/x86/kvm/cpuid.c > @@ -143,7 +143,6 @@ void kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu) > } > vcpu->arch.guest_supported_xss = > (((u64)best->edx << 32) | best->ecx) & supported_xss; > - > } else { > vcpu->arch.guest_supported_xss = 0; > } > diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c > index 9728efd529a1..1703b8874fad 100644 > --- a/arch/x86/kvm/vmx/nested.c > +++ b/arch/x86/kvm/vmx/nested.c > @@ -2516,6 +2516,12 @@ static void prepare_vmcs02_rare(struct vcpu_vmx *vmx, struct vmcs12 *vmcs12) > vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, vmx->msr_autoload.guest.nr); > > set_cr4_guest_host_mask(vmx); > + > + if (kvm_cet_supported()) { This needs to be conditioned on CET coming from vmcs12, it's on the loading of host state on VM-Exit that is unconditional (if CET is supported). if (kvm_cet_supported() && vmx->nested.nested_run_pending && (vmcs12->vm_entry_controls & VM_ENTRY_LOAD_CET_STATE)) { I also assume these should be guarded by one of the eVMCS fields, though a quick search of the public docs didn't provide a hit on the CET fields. Vitaly, any idea if these will be GUEST_GRP2 or something else? > + vmcs_writel(GUEST_SSP, vmcs12->guest_ssp); > + vmcs_writel(GUEST_S_CET, vmcs12->guest_s_cet); > + vmcs_writel(GUEST_INTR_SSP_TABLE, vmcs12->guest_ssp_tbl); > + } > }
On Mon, Mar 01, 2021 at 09:46:19AM -0800, Sean Christopherson wrote: > +Vitaly > > On Thu, Feb 25, 2021, Yang Weijiang wrote: > > These fields are rarely updated by L1 QEMU/KVM, sync them when L1 is trying to > > read/write them and after they're changed. If CET guest entry-load bit is not > > set by L1 guest, migrate them to L2 manaully. > > > > Suggested-by: Sean Christopherson <seanjc@google.com> > > Signed-off-by: Yang Weijiang <weijiang.yang@intel.com> > > > > change in v2: > > - Per Sean's review feedback, change CET guest states as rarely-updated fields. > > And also migrate L1's CET states to L2 if the entry-load bit is not set. > > - Opportunistically removed one blank line. > > --- > > arch/x86/kvm/cpuid.c | 1 - > > arch/x86/kvm/vmx/nested.c | 29 +++++++++++++++++++++++++++++ > > arch/x86/kvm/vmx/vmx.h | 3 +++ > > 3 files changed, 32 insertions(+), 1 deletion(-) > > > > diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c > > index 46087bca9418..afc97122c05c 100644 > > --- a/arch/x86/kvm/cpuid.c > > +++ b/arch/x86/kvm/cpuid.c > > @@ -143,7 +143,6 @@ void kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu) > > } > > vcpu->arch.guest_supported_xss = > > (((u64)best->edx << 32) | best->ecx) & supported_xss; > > - > > } else { > > vcpu->arch.guest_supported_xss = 0; > > } > > diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c > > index 9728efd529a1..1703b8874fad 100644 > > --- a/arch/x86/kvm/vmx/nested.c > > +++ b/arch/x86/kvm/vmx/nested.c > > @@ -2516,6 +2516,12 @@ static void prepare_vmcs02_rare(struct vcpu_vmx *vmx, struct vmcs12 *vmcs12) > > vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, vmx->msr_autoload.guest.nr); > > > > set_cr4_guest_host_mask(vmx); > > + > > + if (kvm_cet_supported()) { > > This needs to be conditioned on CET coming from vmcs12, it's on the loading of > host state on VM-Exit that is unconditional (if CET is supported). > > if (kvm_cet_supported() && vmx->nested.nested_run_pending && > (vmcs12->vm_entry_controls & VM_ENTRY_LOAD_CET_STATE)) { > Thanks Sean! Will change it. > I also assume these should be guarded by one of the eVMCS fields, though a quick > search of the public docs didn't provide a hit on the CET fields. > I got some description from MSFT as below, do you mean that: GuestSsp uses clean field GUEST_BASIC (bit 10) GuestSCet/GuestInterruptSspTableAddr uses GUEST_GRP1 (bit 11) HostSCet/HostSsp/HostInterruptSspTableAddr uses HOST_GRP1 (bit 14) If it is, should these go into separate patch series for Hyper-v nested support? I have some pending patches for the enabling. > Vitaly, any idea if these will be GUEST_GRP2 or something else? > > > + vmcs_writel(GUEST_SSP, vmcs12->guest_ssp); > > + vmcs_writel(GUEST_S_CET, vmcs12->guest_s_cet); > > + vmcs_writel(GUEST_INTR_SSP_TABLE, vmcs12->guest_ssp_tbl); > > + } > > }
On 02/03/21 10:05, Yang Weijiang wrote: > I got some description from MSFT as below, do you mean that: > > GuestSsp uses clean field GUEST_BASIC (bit 10) > GuestSCet/GuestInterruptSspTableAddr uses GUEST_GRP1 (bit 11) > HostSCet/HostSsp/HostInterruptSspTableAddr uses HOST_GRP1 (bit 14) > > If it is, should these go into separate patch series for Hyper-v nested > support? I have some pending patches for the enabling. Yes, it should be a separate patch. The main patch however should add the CET fields to EVMCS1_UNSUPPORTED_VMENTRY_CTRL and EVMCS1_UNSUPPORTED_VMEXIT_CTRL. Thanks, Paolo
Sean Christopherson <seanjc@google.com> writes: > +Vitaly > > On Thu, Feb 25, 2021, Yang Weijiang wrote: >> These fields are rarely updated by L1 QEMU/KVM, sync them when L1 is trying to >> read/write them and after they're changed. If CET guest entry-load bit is not >> set by L1 guest, migrate them to L2 manaully. >> >> Suggested-by: Sean Christopherson <seanjc@google.com> >> Signed-off-by: Yang Weijiang <weijiang.yang@intel.com> >> >> change in v2: >> - Per Sean's review feedback, change CET guest states as rarely-updated fields. >> And also migrate L1's CET states to L2 if the entry-load bit is not set. >> - Opportunistically removed one blank line. >> --- >> arch/x86/kvm/cpuid.c | 1 - >> arch/x86/kvm/vmx/nested.c | 29 +++++++++++++++++++++++++++++ >> arch/x86/kvm/vmx/vmx.h | 3 +++ >> 3 files changed, 32 insertions(+), 1 deletion(-) >> >> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c >> index 46087bca9418..afc97122c05c 100644 >> --- a/arch/x86/kvm/cpuid.c >> +++ b/arch/x86/kvm/cpuid.c >> @@ -143,7 +143,6 @@ void kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu) >> } >> vcpu->arch.guest_supported_xss = >> (((u64)best->edx << 32) | best->ecx) & supported_xss; >> - >> } else { >> vcpu->arch.guest_supported_xss = 0; >> } >> diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c >> index 9728efd529a1..1703b8874fad 100644 >> --- a/arch/x86/kvm/vmx/nested.c >> +++ b/arch/x86/kvm/vmx/nested.c >> @@ -2516,6 +2516,12 @@ static void prepare_vmcs02_rare(struct vcpu_vmx *vmx, struct vmcs12 *vmcs12) >> vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, vmx->msr_autoload.guest.nr); >> >> set_cr4_guest_host_mask(vmx); >> + >> + if (kvm_cet_supported()) { > > This needs to be conditioned on CET coming from vmcs12, it's on the loading of > host state on VM-Exit that is unconditional (if CET is supported). > > if (kvm_cet_supported() && vmx->nested.nested_run_pending && > (vmcs12->vm_entry_controls & VM_ENTRY_LOAD_CET_STATE)) { > > I also assume these should be guarded by one of the eVMCS fields, though a quick > search of the public docs didn't provide a hit on the CET fields. > > Vitaly, any idea if these will be GUEST_GRP2 or something else? > The latest published TLFS I see is 6.0b and it doesn't list anything CET related in eVMCS v1.0 :-( So I agree with Paolo: we just need to adjust EVMCS1_UNSUPPORTED_VMENTRY_CTRL/ EVMCS1_UNSUPPORTED_VMEXIT_CTRL for now and enable it separately later. >> + vmcs_writel(GUEST_SSP, vmcs12->guest_ssp); >> + vmcs_writel(GUEST_S_CET, vmcs12->guest_s_cet); >> + vmcs_writel(GUEST_INTR_SSP_TABLE, vmcs12->guest_ssp_tbl); >> + } >> } >
On Tue, Mar 02, 2021 at 11:35:41AM +0100, Vitaly Kuznetsov wrote: > Sean Christopherson <seanjc@google.com> writes: > > > +Vitaly > > > > On Thu, Feb 25, 2021, Yang Weijiang wrote: > >> These fields are rarely updated by L1 QEMU/KVM, sync them when L1 is trying to > >> read/write them and after they're changed. If CET guest entry-load bit is not > >> set by L1 guest, migrate them to L2 manaully. > >> > >> Suggested-by: Sean Christopherson <seanjc@google.com> > >> Signed-off-by: Yang Weijiang <weijiang.yang@intel.com> > >> > >> change in v2: > >> - Per Sean's review feedback, change CET guest states as rarely-updated fields. > >> And also migrate L1's CET states to L2 if the entry-load bit is not set. > >> - Opportunistically removed one blank line. > >> --- > >> arch/x86/kvm/cpuid.c | 1 - > >> arch/x86/kvm/vmx/nested.c | 29 +++++++++++++++++++++++++++++ > >> arch/x86/kvm/vmx/vmx.h | 3 +++ > >> 3 files changed, 32 insertions(+), 1 deletion(-) > >> > >> diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c > >> index 46087bca9418..afc97122c05c 100644 > >> --- a/arch/x86/kvm/cpuid.c > >> +++ b/arch/x86/kvm/cpuid.c > >> @@ -143,7 +143,6 @@ void kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu) > >> } > >> vcpu->arch.guest_supported_xss = > >> (((u64)best->edx << 32) | best->ecx) & supported_xss; > >> - > >> } else { > >> vcpu->arch.guest_supported_xss = 0; > >> } > >> diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c > >> index 9728efd529a1..1703b8874fad 100644 > >> --- a/arch/x86/kvm/vmx/nested.c > >> +++ b/arch/x86/kvm/vmx/nested.c > >> @@ -2516,6 +2516,12 @@ static void prepare_vmcs02_rare(struct vcpu_vmx *vmx, struct vmcs12 *vmcs12) > >> vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, vmx->msr_autoload.guest.nr); > >> > >> set_cr4_guest_host_mask(vmx); > >> + > >> + if (kvm_cet_supported()) { > > > > This needs to be conditioned on CET coming from vmcs12, it's on the loading of > > host state on VM-Exit that is unconditional (if CET is supported). > > > > if (kvm_cet_supported() && vmx->nested.nested_run_pending && > > (vmcs12->vm_entry_controls & VM_ENTRY_LOAD_CET_STATE)) { > > > > I also assume these should be guarded by one of the eVMCS fields, though a quick > > search of the public docs didn't provide a hit on the CET fields. > > > > Vitaly, any idea if these will be GUEST_GRP2 or something else? > > > > The latest published TLFS I see is 6.0b and it doesn't list anything CET > related in eVMCS v1.0 :-( So I agree with Paolo: we just need to adjust > EVMCS1_UNSUPPORTED_VMENTRY_CTRL/ EVMCS1_UNSUPPORTED_VMEXIT_CTRL for now > and enable it separately later. > Thanks Vitaly and Paolo. New patches have been sent to community. > >> + vmcs_writel(GUEST_SSP, vmcs12->guest_ssp); > >> + vmcs_writel(GUEST_S_CET, vmcs12->guest_s_cet); > >> + vmcs_writel(GUEST_INTR_SSP_TABLE, vmcs12->guest_ssp_tbl); > >> + } > >> } > > > > -- > Vitaly
diff --git a/arch/x86/kvm/cpuid.c b/arch/x86/kvm/cpuid.c index 46087bca9418..afc97122c05c 100644 --- a/arch/x86/kvm/cpuid.c +++ b/arch/x86/kvm/cpuid.c @@ -143,7 +143,6 @@ void kvm_update_cpuid_runtime(struct kvm_vcpu *vcpu) } vcpu->arch.guest_supported_xss = (((u64)best->edx << 32) | best->ecx) & supported_xss; - } else { vcpu->arch.guest_supported_xss = 0; } diff --git a/arch/x86/kvm/vmx/nested.c b/arch/x86/kvm/vmx/nested.c index 9728efd529a1..1703b8874fad 100644 --- a/arch/x86/kvm/vmx/nested.c +++ b/arch/x86/kvm/vmx/nested.c @@ -2516,6 +2516,12 @@ static void prepare_vmcs02_rare(struct vcpu_vmx *vmx, struct vmcs12 *vmcs12) vmcs_write32(VM_ENTRY_MSR_LOAD_COUNT, vmx->msr_autoload.guest.nr); set_cr4_guest_host_mask(vmx); + + if (kvm_cet_supported()) { + vmcs_writel(GUEST_SSP, vmcs12->guest_ssp); + vmcs_writel(GUEST_S_CET, vmcs12->guest_s_cet); + vmcs_writel(GUEST_INTR_SSP_TABLE, vmcs12->guest_ssp_tbl); + } } /* @@ -2556,6 +2562,15 @@ static int prepare_vmcs02(struct kvm_vcpu *vcpu, struct vmcs12 *vmcs12, if (kvm_mpx_supported() && (!vmx->nested.nested_run_pending || !(vmcs12->vm_entry_controls & VM_ENTRY_LOAD_BNDCFGS))) vmcs_write64(GUEST_BNDCFGS, vmx->nested.vmcs01_guest_bndcfgs); + + if (kvm_cet_supported() && (!vmx->nested.nested_run_pending || + !(vmcs12->vm_entry_controls & VM_ENTRY_LOAD_CET_STATE))) { + vmcs_writel(GUEST_SSP, vmx->nested.vmcs01_guest_ssp); + vmcs_writel(GUEST_S_CET, vmx->nested.vmcs01_guest_s_cet); + vmcs_writel(GUEST_INTR_SSP_TABLE, + vmx->nested.vmcs01_guest_ssp_tbl); + } + vmx_set_rflags(vcpu, vmcs12->guest_rflags); /* EXCEPTION_BITMAP and CR0_GUEST_HOST_MASK should basically be the @@ -3375,6 +3390,12 @@ enum nvmx_vmentry_status nested_vmx_enter_non_root_mode(struct kvm_vcpu *vcpu, if (kvm_mpx_supported() && !(vmcs12->vm_entry_controls & VM_ENTRY_LOAD_BNDCFGS)) vmx->nested.vmcs01_guest_bndcfgs = vmcs_read64(GUEST_BNDCFGS); + if (kvm_cet_supported() && + !(vmcs12->vm_entry_controls & VM_ENTRY_LOAD_CET_STATE)) { + vmx->nested.vmcs01_guest_ssp = vmcs_readl(GUEST_SSP); + vmx->nested.vmcs01_guest_s_cet = vmcs_readl(GUEST_S_CET); + vmx->nested.vmcs01_guest_ssp_tbl = vmcs_readl(GUEST_INTR_SSP_TABLE); + } /* * Overwrite vmcs01.GUEST_CR3 with L1's CR3 if EPT is disabled *and* @@ -4001,6 +4022,9 @@ static bool is_vmcs12_ext_field(unsigned long field) case GUEST_IDTR_BASE: case GUEST_PENDING_DBG_EXCEPTIONS: case GUEST_BNDCFGS: + case GUEST_SSP: + case GUEST_INTR_SSP_TABLE: + case GUEST_S_CET: return true; default: break; @@ -4052,6 +4076,11 @@ static void sync_vmcs02_to_vmcs12_rare(struct kvm_vcpu *vcpu, vmcs_readl(GUEST_PENDING_DBG_EXCEPTIONS); if (kvm_mpx_supported()) vmcs12->guest_bndcfgs = vmcs_read64(GUEST_BNDCFGS); + if (kvm_cet_supported()) { + vmcs12->guest_ssp = vmcs_readl(GUEST_SSP); + vmcs12->guest_s_cet = vmcs_readl(GUEST_S_CET); + vmcs12->guest_ssp_tbl = vmcs_readl(GUEST_INTR_SSP_TABLE); + } vmx->nested.need_sync_vmcs02_to_vmcs12_rare = false; } diff --git a/arch/x86/kvm/vmx/vmx.h b/arch/x86/kvm/vmx/vmx.h index 9d3a557949ac..36dc4fdb0909 100644 --- a/arch/x86/kvm/vmx/vmx.h +++ b/arch/x86/kvm/vmx/vmx.h @@ -155,6 +155,9 @@ struct nested_vmx { /* to migrate it to L2 if VM_ENTRY_LOAD_DEBUG_CONTROLS is off */ u64 vmcs01_debugctl; u64 vmcs01_guest_bndcfgs; + u64 vmcs01_guest_ssp; + u64 vmcs01_guest_s_cet; + u64 vmcs01_guest_ssp_tbl; /* to migrate it to L1 if L2 writes to L1's CR8 directly */ int l1_tpr_threshold;
These fields are rarely updated by L1 QEMU/KVM, sync them when L1 is trying to read/write them and after they're changed. If CET guest entry-load bit is not set by L1 guest, migrate them to L2 manaully. Suggested-by: Sean Christopherson <seanjc@google.com> Signed-off-by: Yang Weijiang <weijiang.yang@intel.com> change in v2: - Per Sean's review feedback, change CET guest states as rarely-updated fields. And also migrate L1's CET states to L2 if the entry-load bit is not set. - Opportunistically removed one blank line. --- arch/x86/kvm/cpuid.c | 1 - arch/x86/kvm/vmx/nested.c | 29 +++++++++++++++++++++++++++++ arch/x86/kvm/vmx/vmx.h | 3 +++ 3 files changed, 32 insertions(+), 1 deletion(-)