Enable Secure TSC for SEV-SNP
v6 cover Enable Secure TSC for SEV-SNP
The hypervisor controls TSC value calculations for the guest. A malicious hypervisor can prevent the guest from progressing. The Secure TSC feature for SEV-SNP allows guests to securely use the RDTSC and RDTSCP instructions. This ensures the guest has a consistent view of time and prevents a malicious hypervisor from manipulating time, such as making it appear to move backward or advance too quickly. For more details, refer to the "Secure Nested Paging (SEV-SNP)" section, subsection "Secure TSC" in APM Volume 2.
This patch set is also available at:
https://github.com/AMDESE/linux-kvm/tree/sectsc-host-latest
and is based on kvm/master
Testing Secure TSC
-----------------
Secure TSC guest patches are available as part of v6.14-rc1.
QEMU changes:
https://github.com/nikunjad/qemu/tree/snp-securetsc-latest
QEMU command line SEV-SNP with Secure TSC:
qemu-system-x86_64 -cpu EPYC-Milan-v2 -smp 4 \
-object memory-backend-memfd,id=ram1,size=1G,share=true,prealloc=false,reserve=false \
-object sev-snp-guest,id=sev0,cbitpos=51,reduced-phys-bits=1,secure-tsc=on,stsc-freq=2000000000 \
-machine q35,confidential-guest-support=sev0,memory-backend=ram1 \
...Changelog:
----------
v6:
* Rebased on top of kvm/master
* Collected Reviewed-by/Tested-by
* s/svm->vcpu/vcpu/ in snp_launch_update_vmsa() as vcpu pointer is already available (Tom)
* Simplify assignment of guest_protected_tsc (Tom)
v5: https://lore.kernel.org/kvm/[email protected]/
* Rebased on top of kvm/queue that includes protected TSC patches
https://lore.kernel.org/kvm/[email protected]/
* Dropped patch 4/5 as it is not required after protected TSC patches
* Set guest_tsc_protected when Secure TSC is enabled (Paolo)
* Collect Reviewed-by from Tom
* Base the desired_tsc_freq on KVM's ABI (Sean)
Ketan Chaturvedi (1):
KVM: SVM: Enable Secure TSC for SNP guests
Nikunj A Dadhania (3):
x86/cpufeatures: Add SNP Secure TSC
KVM: SVM: Add missing member in SNP_LAUNCH_START command structure
KVM: SVM: Add GUEST_TSC_FREQ MSR for Secure TSC enabled guests
arch/x86/include/asm/cpufeatures.h | 1 + arch/x86/include/asm/svm.h | 1 + arch/x86/include/uapi/asm/kvm.h | 3 ++- arch/x86/kvm/svm/sev.c | 18 +++++++++++++++++- arch/x86/kvm/svm/svm.c | 1 + arch/x86/kvm/svm/svm.h | 11 ++++++++++- include/linux/psp-sev.h | 2 ++ 7 files changed, 34 insertions(+), 3 deletions(-)
base-commit: c77eee50caa289fee6cfde146471aa7b0f311471
Signature · 1 line
2.43.0
PATCH v6 1/4 x86/cpufeatures: Add SNP Secure TSC
The Secure TSC feature for SEV-SNP allows guests to securely use the RDTSC and RDTSCP instructions, ensuring that the parameters used cannot be altered by the hypervisor once the guest is launched. For more details, refer to the AMD64 APM Vol 2, Section "Secure TSC".
--- arch/x86/include/asm/cpufeatures.h | 1 + 1 file changed, 1 insertion(+)
arch/x86/include/asm/cpufeatures.h +1 -0
@@ -449,6 +449,7 @@#define X86_FEATURE_VM_PAGE_FLUSH (19*32+ 2)/* VM Page Flush MSR is supported */ #define X86_FEATURE_SEV_ES (19*32+ 3)/* "sev_es" Secure Encrypted Virtualization - Encrypted State */ #define X86_FEATURE_SEV_SNP (19*32+ 4)/* "sev_snp" Secure Encrypted Virtualization - Secure Nested Paging */ +#define X86_FEATURE_SNP_SECURE_TSC (19*32+ 8)/* SEV-SNP Secure TSC */ #define X86_FEATURE_V_TSC_AUX (19*32+ 9)/* Virtual TSC_AUX */ #define X86_FEATURE_SME_COHERENT (19*32+10)/* hardware-enforced cache coherency */ #define X86_FEATURE_DEBUG_SWAP (19*32+14)/* "debug_swap" SEV-ES full debug state swap support */
Signature · 1 line
2.43.0
PATCH v6 2/4 KVM: SVM: Add missing member in SNP_LAUNCH_START command structure
In reply to Nikunj A DadhaniaThe sev_data_snp_launch_start structure should include a 4-byte desired_tsc_khz field before the gosvw field, which was missed in the initial implementation. As a result, the structure is 4 bytes shorter than expected by the firmware, causing the gosvw field to start 4 bytes early. Fix this by adding the missing 4-byte member for the desired TSC frequency.
--- include/linux/psp-sev.h | 2 ++ 1 file changed, 2 insertions(+)
include/linux/psp-sev.h +2 -0
@@ -594,6 +594,7 @@ struct sev_data_snp_addr {* @imi_en: launch flow is launching an IMI (Incoming Migration Image) for the * purpose of guest-assisted migration. * @rsvd: reserved +* @desired_tsc_khz: hypervisor desired mean TSC freq in kHz of the guest * @gosvw: guest OS-visible workarounds, as defined by hypervisor */ struct sev_data_snp_launch_start {
@@ -603,6 +604,7 @@ struct sev_data_snp_launch_start { u32 ma_en:1; /* In */ u32 imi_en:1; /* In */ u32 rsvd:30; +u32 desired_tsc_khz;/* In */ u8 gosvw[16 ];/* In */ } __packed;
Signature · 1 line
2.43.0
PATCH v6 3/4 KVM: SVM: Add GUEST_TSC_FREQ MSR for Secure TSC enabled guests
In reply to Nikunj A DadhaniaIntroduce the read-only MSR GUEST_TSC_FREQ (0xc0010134) that returns guest's effective frequency in MHZ when Secure TSC is enabled for SNP guests. Disable interception of this MSR when Secure TSC is enabled. Note that GUEST_TSC_FREQ MSR is accessible only to the guest and not from the hypervisor context.
--- arch/x86/include/asm/svm.h | 1 + arch/x86/kvm/svm/sev.c | 3 +++ arch/x86/kvm/svm/svm.c | 1 + arch/x86/kvm/svm/svm.h | 11 ++++++++++- 4 files changed, 15 insertions(+), 1 deletion(-)
arch/x86/include/asm/svm.h +1 -0
@@ -290,6 +290,7 @@ static_assert((X2AVIC_MAX_PHYSICAL_ID & AVIC_PHYSICAL_MAX_INDEX_MASK) == X2AVIC_#define SVM_SEV_FEAT_RESTRICTED_INJECTION BIT(3)#define SVM_SEV_FEAT_ALTERNATE_INJECTION BIT(4)#define SVM_SEV_FEAT_DEBUG_SWAP BIT(5) +#define SVM_SEV_FEAT_SECURE_TSC BIT(9)struct vmcb_seg {u16 selector;
arch/x86/kvm/svm/sev.c +3 -0
@@ -4504,6 +4504,9 @@ static void sev_es_init_vmcb(struct vcpu_svm *svm)/* Clear intercepts on selected MSRs */ set_msr_interception (vcpu, svm->msrpm,MSR_EFER ,1 ,1 );set_msr_interception (vcpu, svm->msrpm,MSR_IA32_CR_PAT ,1 ,1 ); + +if (snp_secure_tsc_enabled (vcpu->kvm)) +set_msr_interception (vcpu, svm->msrpm,MSR_AMD64_GUEST_TSC_FREQ ,1 ,1 ); }void sev_init_vmcb (struct vcpu_svm *svm)
arch/x86/kvm/svm/svm.c +1 -0
@@ -143,6 +143,7 @@ static const struct svm_direct_access_msrs { { .index = X2APIC_MSR(APIC_TMICT), .always = false }, { .index = X2APIC_MSR(APIC_TMCCT), .always = false }, { .index = X2APIC_MSR(APIC_TDCR), .always = false }, + { .index = MSR_AMD64_GUEST_TSC_FREQ, .always = false }, { .index = MSR_INVALID, .always = false }, };
arch/x86/kvm/svm/svm.h +10 -1
@@ -44,7 +44,7 @@ static inline struct page *__sme_pa_to_page(unsigned long pa)#define IOPM_SIZE PAGE_SIZE * 3#define MSRPM_SIZE PAGE_SIZE * 2 -#define MAX_DIRECT_ACCESS_MSRS 48 +#define MAX_DIRECT_ACCESS_MSRS 49#define MSRPM_OFFSETS 32extern u32 msrpm_offsets[MSRPM_OFFSETS ]__read_mostly ;extern bool npt_enabled;
@@ -377,10 +377,19 @@ static __always_inline bool sev_snp_guest(struct kvm *kvm)return (sev->vmsa_features &SVM_SEV_FEAT_SNP_ACTIVE ) && !WARN_ON_ONCE(!sev_es_guest(kvm)); } + +static inline bool snp_secure_tsc_enabled (struct kvm *kvm) +{ +struct kvm_sev_info *sev =to_kvm_sev_info (kvm); + +return (sev->vmsa_features &SVM_SEV_FEAT_SECURE_TSC ) && + !WARN_ON_ONCE (!sev_snp_guest (kvm)); +}#else #define sev_guest (kvm) false#define sev_es_guest (kvm) false#define sev_snp_guest (kvm) false +#define snp_secure_tsc_enabled (kvm) false#endif static inline bool ghcb_gpa_is_registered (struct vcpu_svm *svm,u64 val)
Signature · 1 line
2.43.0
Sean Christopherson
maintainerRe: [PATCH v6 3/4] KVM: SVM: Add GUEST_TSC_FREQ MSR for Secure TSC enabled guests
Quoted 68 lines · Nikunj A Dadhania
Introduce the read-only MSR
GUEST_TSC_FREQ(0xc0010134) that returns guest's effective frequency in MHZ when Secure TSC is enabled for SNP guests. Disable interception of this MSR when Secure TSC is enabled. Note thatGUEST_TSC_FREQMSR is accessible only to the guest and not from the hypervisor context.Reviewed-by Tom Lendacky <[email protected]>Tested-by Vaishali Thakkar <[email protected]>Signed-off-by Nikunj A Dadhania <[email protected]>--- arch/x86/include/asm/svm.h | 1 + arch/x86/kvm/svm/sev.c | 3 +++ arch/x86/kvm/svm/svm.c | 1 + arch/x86/kvm/svm/svm.h | 11 ++++++++++- 4 files changed, 15 insertions(+), 1 deletion(-)arch/x86/include/asm/svm.h +1
-0@@ -290,6 +290,7 @@ static_assert((X2AVIC_MAX_PHYSICAL_ID & AVIC_PHYSICAL_MAX_INDEX_MASK) == X2AVIC_#define SVM_SEV_FEAT_RESTRICTED_INJECTION BIT(3)#define SVM_SEV_FEAT_ALTERNATE_INJECTION BIT(4)#define SVM_SEV_FEAT_DEBUG_SWAP BIT(5) +#define SVM_SEV_FEAT_SECURE_TSC BIT(9)struct vmcb_seg {u16 selector;arch/x86/kvm/svm/sev.c +3
-0@@ -4504,6 +4504,9 @@ static void sev_es_init_vmcb(struct vcpu_svm *svm)/* Clear intercepts on selected MSRs */ set_msr_interception (vcpu, svm->msrpm,MSR_EFER ,1 ,1 );set_msr_interception (vcpu, svm->msrpm,MSR_IA32_CR_PAT ,1 ,1 ); + +if (snp_secure_tsc_enabled (vcpu->kvm)) +set_msr_interception (vcpu, svm->msrpm,MSR_AMD64_GUEST_TSC_FREQ ,1 ,1 ); }void sev_init_vmcb (struct vcpu_svm *svm)arch/x86/kvm/svm/svm.c +1
-0@@ -143,6 +143,7 @@ static const struct svm_direct_access_msrs { { .index = X2APIC_MSR(APIC_TMICT), .always = false }, { .index = X2APIC_MSR(APIC_TMCCT), .always = false }, { .index = X2APIC_MSR(APIC_TDCR), .always = false }, + { .index = MSR_AMD64_GUEST_TSC_FREQ, .always = false }, { .index = MSR_INVALID, .always = false }, };arch/x86/kvm/svm/svm.h +1
-1@@ -44,7 +44,7 @@ static inline struct page *__sme_pa_to_page(unsigned long pa)#define IOPM_SIZE PAGE_SIZE * 3#define MSRPM_SIZE PAGE_SIZE * 2 -#define MAX_DIRECT_ACCESS_MSRS 48 +#define MAX_DIRECT_ACCESS_MSRS 49#define MSRPM_OFFSETS 32extern u32 msrpm_offsets[MSRPM_OFFSETS ]__read_mostly ;extern bool npt_enabled;
arch/x86/kvm/svm/svm.h static __always_inline bool sev_snp_guest(struct kvm *kvm) · line 381
Quoted 5 lines
arch/x86/kvm/svm/svm.h +2
-0@@ -377,10 +377,19 @@ static __always_inline bool sev_snp_guest(struct kvm *kvm)return (sev->vmsa_features &SVM_SEV_FEAT_SNP_ACTIVE ) && !WARN_ON_ONCE (!sev_es_guest (kvm)); } + +static inline bool snp_secure_tsc_enabled (struct kvm *kvm)
This is only ever used in sev.c, it has no business living in svm.c. And there's especially no reason to have a stub for CONFIG_KVM_AMD_SEV=n.
Quoted 17 lines
arch/x86/kvm/svm/svm.h +7
-0+{ +struct kvm_sev_info *sev =to_kvm_sev_info (kvm); + +return (sev->vmsa_features &SVM_SEV_FEAT_SECURE_TSC ) && + !WARN_ON_ONCE (!sev_snp_guest (kvm)); +}#else #define sev_guest (kvm) false#define sev_es_guest (kvm) false#define sev_snp_guest (kvm) false +#define snp_secure_tsc_enabled (kvm) false#endif static inline bool ghcb_gpa_is_registered (struct vcpu_svm *svm,u64 val)Signature · 1 line
2.43.0
Re: [PATCH v6 3/4] KVM: SVM: Add GUEST_TSC_FREQ MSR for Secure TSC enabled guests
Quoted 10 lines · Sean Christopherson
arch/x86/kvm/svm/svm.h static __always_inline bool sev_snp_guest(struct kvm *kvm) · line 381
Quoted 6 lines · Nikunj A Dadhania
arch/x86/kvm/svm/svm.h +2
-0@@ -377,10 +377,19 @@ static __always_inline bool sev_snp_guest(struct kvm *kvm)return (sev->vmsa_features &SVM_SEV_FEAT_SNP_ACTIVE ) && !WARN_ON_ONCE (!sev_es_guest (kvm)); } + +static inline bool snp_secure_tsc_enabled (struct kvm *kvm)This is only ever used in
sev.c, it has no business living insvm.c. And there's especially no reason to have a stub forCONFIG_KVM_AMD_SEV=n.
Ack, will move this to sev.c and remove the stub as well.
Regards
Nikunj
PATCH v6 4/4 KVM: SVM: Enable Secure TSC for SNP guests
In reply to Nikunj A DadhaniaFrom: Ketan Chaturvedi <[email protected]>
Add support for Secure TSC, allowing userspace to configure the Secure TSC feature for SNP guests. Use the SNP specification's desired TSC frequency parameter during the SNP_LAUNCH_START command to set the mean TSC frequency in KHz for Secure TSC enabled guests.
As the frequency needs to be set in the SNP_LAUNCH_START command, userspace should set the frequency using the KVM_CAP_SET_TSC_KHZ VM ioctl instead of the VCPU ioctl. The desired_tsc_khz defaults to kvm->arch.default_tsc_khz.
--- arch/x86/include/uapi/asm/kvm.h | 3 ++- arch/x86/kvm/svm/sev.c | 15 ++++++++++++++- 2 files changed, 16 insertions(+), 2 deletions(-)
arch/x86/include/uapi/asm/kvm.h +2 -1
@@ -839,7 +839,8 @@ struct kvm_sev_snp_launch_start {__u64 policy;__u8 gosvw[16 ];__u16 flags; -__u8 pad0[6 ]; +__u8 pad0[2 ]; +__u32 desired_tsc_khz;__u64 pad1[4 ]; };
arch/x86/kvm/svm/sev.c +14 -1
@@ -2205,6 +2205,14 @@ static int snp_launch_start(struct kvm *kvm, struct kvm_sev_cmd *argp) start.gctx_paddr =__psp_pa (sev->snp_context); start.policy = params.policy; + +if (snp_secure_tsc_enabled (kvm)) { +if (!kvm->arch.default_tsc_khz) +return -EINVAL ; + + start.desired_tsc_khz = kvm->arch.default_tsc_khz; + } +memcpy (start.gosvw, params.gosvw,sizeof (params.gosvw)); rc =__sev_issue_cmd (argp->sev_fd,SEV_CMD_SNP_LAUNCH_START , &start, &argp->error);if (rc) {
@@ -2445,7 +2453,9 @@ static int snp_launch_update_vmsa(struct kvm *kvm, struct kvm_sev_cmd *argp)return ret; } - svm->vcpu.arch.guest_state_protected = true; + vcpu->arch.guest_state_protected = true; + vcpu->arch.guest_tsc_protected =snp_secure_tsc_enabled (kvm); +/* * SEV-ES (and thus SNP) guest mandates LBR Virtualization to * be _always_ ON. Enable it only after setting
@@ -3059,6 +3069,9 @@ void __init sev_hardware_setup(void) sev_supported_vmsa_features =0 ;if (sev_es_debug_swap_enabled) sev_supported_vmsa_features |=SVM_SEV_FEAT_DEBUG_SWAP ; + +if (sev_snp_enabled &&cpu_feature_enabled (X86_FEATURE_SNP_SECURE_TSC )) + sev_supported_vmsa_features |=SVM_SEV_FEAT_SECURE_TSC ; }void sev_hardware_unsetup (void )
Signature · 1 line
2.43.0
Sean Christopherson
maintainerRe: [PATCH v6 4/4] KVM: SVM: Enable Secure TSC for SNP guests
The previous patch to add GUEST_TSC_FREQ needs to squashed with this patch. It's impossible to review the snp_secure_tsc_enabled() logic in particular without the details added in this patch.
And once you rebase on kvm-x86 next (i.e. the MSR interception rework), adding support for GUEST_TSC_FREQ will be like three lines of code, i.e. not worth landing in a separate patch.
Quoted 40 lines · Nikunj A Dadhania
From: Ketan Chaturvedi <[email protected]>
Add support for Secure TSC, allowing userspace to configure the Secure TSC feature for SNP guests. Use the SNP specification's desired TSC frequency parameter during the
SNP_LAUNCH_STARTcommand to set the mean TSC frequency in KHz for Secure TSC enabled guests.As the frequency needs to be set in the
SNP_LAUNCH_STARTcommand, userspace should set the frequency using theKVM_CAP_SET_TSC_KHZVM ioctl instead of the VCPU ioctl. Thedesired_tsc_khzdefaults tokvm->arch.default_tsc_khz.Signed-off-by Ketan Chaturvedi <[email protected]>Co-developed-by Nikunj A Dadhania <[email protected]>Tested-by Vaishali Thakkar <[email protected]>Reviewed-by Tom Lendacky <[email protected]>Signed-off-by Nikunj A Dadhania <[email protected]>--- arch/x86/include/uapi/asm/kvm.h | 3 ++- arch/x86/kvm/svm/sev.c | 15 ++++++++++++++- 2 files changed, 16 insertions(+), 2 deletions(-)arch/x86/include/uapi/asm/kvm.h +2
-1@@ -839,7 +839,8 @@ struct kvm_sev_snp_launch_start {__u64 policy;__u8 gosvw[16 ];__u16 flags; -__u8 pad0[6 ]; +__u8 pad0[2 ]; +__u32 desired_tsc_khz;__u64 pad1[4 ]; };arch/x86/kvm/svm/sev.c +0
-0
arch/x86/kvm/svm/sev.c static int snp_launch_start(struct kvm *kvm, struct kvm_sev_cmd *argp) · line 2210
Quoted 6 lines
arch/x86/kvm/svm/sev.c +3
-0@@ -2205,6 +2205,14 @@ static int snp_launch_start(struct kvm *kvm, struct kvm_sev_cmd *argp) start.gctx_paddr =__psp_pa (sev->snp_context); start.policy = params.policy; + +if (snp_secure_tsc_enabled (kvm)) { +if (!kvm->arch.default_tsc_khz)
Hmm, so there's an existing flaw related to the TSC frequency. Ideally, KVM shouldn't allow KVM_SET_TSC_KHZ on a vCPU with a "secure" TSC, i.e. on a TDX vCPU or on a newfangled SNP vCPU. I'm not sure that's worth addressing though, because it doesn't put KVM in any danger, it can only cause problems for guest timing. Yeah, I guess we leave it, because it's not really any different than enumerating a TSC frequency in CPUID 0x15 and then telling KVM something different.
Quoted 20 lines
arch/x86/kvm/svm/sev.c +8
-1+return -EINVAL ; + + start.desired_tsc_khz = kvm->arch.default_tsc_khz; + } +memcpy (start.gosvw, params.gosvw,sizeof (params.gosvw)); rc =__sev_issue_cmd (argp->sev_fd,SEV_CMD_SNP_LAUNCH_START , &start, &argp->error);if (rc) {@@ -2445,7 +2453,9 @@ static int snp_launch_update_vmsa(struct kvm *kvm, struct kvm_sev_cmd *argp)return ret; } - svm->vcpu.arch.guest_state_protected = true; + vcpu->arch.guest_state_protected = true; + vcpu->arch.guest_tsc_protected =snp_secure_tsc_enabled (kvm); +/* * SEV-ES (and thus SNP) guest mandates LBR Virtualization to * be _always_ ON. Enable it only after setting
arch/x86/kvm/svm/sev.c void __init sev_hardware_setup(void) · line 3074
Quoted 6 lines
arch/x86/kvm/svm/sev.c +3
-0@@ -3059,6 +3069,9 @@ void __init sev_hardware_setup(void) sev_supported_vmsa_features =0 ;if (sev_es_debug_swap_enabled) sev_supported_vmsa_features |=SVM_SEV_FEAT_DEBUG_SWAP ; + +if (sev_snp_enabled &&cpu_feature_enabled (X86_FEATURE_SNP_SECURE_TSC )) + sev_supported_vmsa_features |=SVM_SEV_FEAT_SECURE_TSC ;
I don't see anything in here that prevents userspace from stuffing SECURE_TSC into vmsa_features, which means the WARN_ON_ONCE() in snp_secure_tsc_enabled is user-triggerable.
Unless I'm missing something, this need to do something like:
arch/x86/kvm/svm/sev.c +6 -2
@@ -405,9 +405,13 @@ static int __sev_guest_init(struct kvm *kvm, struct kvm_sev_cmd *argp,struct kvm_sev_info *sev =to_kvm_sev_info (kvm);struct sev_platform_init_args init_args = {0 };bool es_active = vm_type !=KVM_X86_SEV_VM ; +bool snp_active = vm_type -=KVM_X86_SNP_VM ;u64 valid_vmsa_features = es_active ? sev_supported_vmsa_features :0 ;int ret; +if (!snp_active) + valid_vmsa_features &= ~SVM_SEV_FEAT_SECURE_TSC ; +if (kvm->created_vcpus)return -EINVAL ;
@@ -436,7 +440,7 @@ static int __sev_guest_init(struct kvm *kvm, struct kvm_sev_cmd *argp,if (sev->es_active && !sev->ghcb_version) sev->ghcb_version =GHCB_VERSION_DEFAULT ; -if (vm_type ==KVM_X86_SNP_VM ) +if (snp_active) sev->vmsa_features |=SVM_SEV_FEAT_SNP_ACTIVE ; ret =sev_asid_new (sev);
@@ -449,7 +453,7 @@ static int __sev_guest_init(struct kvm *kvm, struct kvm_sev_cmd *argp, goto e_free;/* This needs to happen after SEV/SNP firmware initialization. */ -if (vm_type ==KVM_X86_SNP_VM ) { +if (snp_active) { ret =snp_guest_req_init (kvm);if (ret) goto e_free;
Re: [PATCH v6 4/4] KVM: SVM: Enable Secure TSC for SNP guests
Quoted 7 lines · Sean Christopherson
The previous patch to add
GUEST_TSC_FREQneeds to squashed with this patch. It's impossible to review thesnp_secure_tsc_enabled()logic in particular without the details added in this patch.And once you rebase on kvm-x86 next (i.e. the MSR interception rework), adding support for
GUEST_TSC_FREQwill be like three lines of code, i.e. not worth landing in a separate patch.
Ack.
Quoted 6 lines
Quoted 6 lines
arch/x86/kvm/svm/sev.c +0
-0@@ -2205,6 +2205,14 @@ static int snp_launch_start(struct kvm *kvm, struct kvm_sev_cmd *argp)
Quoted 13 lines
arch/x86/kvm/svm/sev.c static int snp_launch_start(struct kvm *kvm, struct kvm_sev_cmd *argp) · line 2210
Quoted 5 lines
arch/x86/kvm/svm/sev.c +3
-0start.gctx_paddr =__psp_pa (sev->snp_context); start.policy = params.policy; + +if (snp_secure_tsc_enabled (kvm)) { +if (!kvm->arch.default_tsc_khz)Hmm, so there's an existing flaw related to the TSC frequency. Ideally, KVM shouldn't allow
KVM_SET_TSC_KHZon a vCPU with a "secure" TSC, i.e. on a TDX vCPU or on a newfangled SNP vCPU. I'm not sure that's worth addressing though, because it doesn't put KVM in any danger, it can only cause problems for guest timing. Yeah, I guess we leave it, because it's not really any different than enumerating a TSC frequency in CPUID0x15and then telling KVM something different.
We can prevent the host from setting KVM_SET_TSC_KHZ when arch.guest_tsc_protected is set in the vCPU IOCTL. Although, doing that in VM IOCTL will be tricky and may require to add something like kvm->arch.has_tsc_protected.
Quoted 23 lines
Quoted 23 lines
arch/x86/kvm/svm/sev.c +8
-1+return -EINVAL ; + + start.desired_tsc_khz = kvm->arch.default_tsc_khz; + } +memcpy (start.gosvw, params.gosvw,sizeof (params.gosvw)); rc =__sev_issue_cmd (argp->sev_fd,SEV_CMD_SNP_LAUNCH_START , &start, &argp->error);if (rc) {@@ -2445,7 +2453,9 @@ static int snp_launch_update_vmsa(struct kvm *kvm, struct kvm_sev_cmd *argp)return ret; } - svm->vcpu.arch.guest_state_protected = true; + vcpu->arch.guest_state_protected = true; + vcpu->arch.guest_tsc_protected =snp_secure_tsc_enabled (kvm); +/* * SEV-ES (and thus SNP) guest mandates LBR Virtualization to * be _always_ ON. Enable it only after setting @@ -3059,6 +3069,9 @@ void __init sev_hardware_setup(void) sev_supported_vmsa_features =0 ;if (sev_es_debug_swap_enabled)
Quoted 8 lines
arch/x86/kvm/svm/sev.c void __init sev_hardware_setup(void) · line 3074
Quoted 4 lines
arch/x86/kvm/svm/sev.c +3
-0sev_supported_vmsa_features |=SVM_SEV_FEAT_DEBUG_SWAP ; + +if (sev_snp_enabled &&cpu_feature_enabled (X86_FEATURE_SNP_SECURE_TSC )) + sev_supported_vmsa_features |=SVM_SEV_FEAT_SECURE_TSC ;I don't see anything in here that prevents userspace from stuffing
SECURE_TSCintovmsa_features, which means theWARN_ON_ONCE()insnp_secure_tsc_enabledis user-triggerable.
Right, my QEMU patches enable vmsa_features only for the sev-snp-guest object. But you are right, vmsa_features is part of KVM_SEV_INIT2 and can enable SECURE_TSC causing the WARN_ON_ONCE()
arch/x86/kvm/svm/sev.c static int __sev_guest_init(struct kvm *kvm, struct kvm_sev_cmd *argp, · line 408
Quoted 12 lines
Unless I'm missing something, this need to do something like:
arch/x86/kvm/svm/sev.c +1
-0@@ -405,9 +405,13 @@ static int __sev_guest_init(struct kvm *kvm, struct kvm_sev_cmd *argp,struct kvm_sev_info *sev =to_kvm_sev_info (kvm);struct sev_platform_init_args init_args = {0 };bool es_active = vm_type !=KVM_X86_SEV_VM ; +bool snp_active = vm_type -=KVM_X86_SNP_VM ;
vm_type == KVM_X86_SNP_VM
Quoted 19 lines
arch/x86/kvm/svm/sev.c +4
-1u64 valid_vmsa_features = es_active ? sev_supported_vmsa_features :0 ;int ret; +if (!snp_active) + valid_vmsa_features &= ~SVM_SEV_FEAT_SECURE_TSC ; +if (kvm->created_vcpus)return -EINVAL ;@@ -436,7 +440,7 @@ static int __sev_guest_init(struct kvm *kvm, struct kvm_sev_cmd *argp,if (sev->es_active && !sev->ghcb_version) sev->ghcb_version =GHCB_VERSION_DEFAULT ; -if (vm_type ==KVM_X86_SNP_VM ) +if (snp_active) sev->vmsa_features |=SVM_SEV_FEAT_SNP_ACTIVE ; ret =sev_asid_new (sev);
arch/x86/kvm/svm/sev.c static int __sev_guest_init(struct kvm *kvm, struct kvm_sev_cmd *argp, · line 459
Quoted 8 lines
arch/x86/kvm/svm/sev.c +1
-1@@ -449,7 +453,7 @@ static int __sev_guest_init(struct kvm *kvm, struct kvm_sev_cmd *argp, goto e_free;/* This needs to happen after SEV/SNP firmware initialization. */ -if (vm_type ==KVM_X86_SNP_VM ) { +if (snp_active) { ret =snp_guest_req_init (kvm);if (ret) goto e_free;
I will fold this in.
Regards
Nikunj
Quoted 30 lines · Nikunj A Dadhania
The hypervisor controls TSC value calculations for the guest. A malicious hypervisor can prevent the guest from progressing. The Secure TSC feature for SEV-SNP allows guests to securely use the RDTSC and RDTSCP instructions. This ensures the guest has a consistent view of time and prevents a malicious hypervisor from manipulating time, such as making it appear to move backward or advance too quickly. For more details, refer to the "Secure Nested Paging (SEV-SNP)" section, subsection "Secure TSC" in APM Volume 2.
This patch set is also available at:
https://github.com/AMDESE/linux-kvm/tree/sectsc-host-latest
and is based on kvm/master
Testing Secure TSC
-----------------Secure TSC guest patches are available as part of v6.14-rc1.
QEMU changes:
https://github.com/nikunjad/qemu/tree/snp-securetsc-latestQEMU command line SEV-SNP with Secure TSC:
qemu-system-x86_64 -cpu EPYC-Milan-v2 -smp 4 \ -object memory-backend-memfd,id=ram1,size=1G,share=true,prealloc=false,reserve=false \ -object sev-snp-guest,id=sev0,cbitpos=51,reduced-phys-bits=1,secure-tsc=on,stsc-freq=2000000000 \ -machine q35,confidential-guest-support=sev0,memory-backend=ram1 \ ...
Quoted 7 lines
Changelog:
----------
v6:
* Rebased on top of kvm/master
* Collected Reviewed-by/Tested-by
* s/svm->vcpu/vcpu/ insnp_launch_update_vmsa()as vcpu pointer is already available (Tom)
* Simplify assignment ofguest_protected_tsc(Tom)
A gentle reminder, any other suggestions/improvement ?
Regards
Nikunj
Quoted 39 lines · Nikunj A Dadhania
Quoted 30 lines · Nikunj A Dadhania
The hypervisor controls TSC value calculations for the guest. A malicious hypervisor can prevent the guest from progressing. The Secure TSC feature for SEV-SNP allows guests to securely use the RDTSC and RDTSCP instructions. This ensures the guest has a consistent view of time and prevents a malicious hypervisor from manipulating time, such as making it appear to move backward or advance too quickly. For more details, refer to the "Secure Nested Paging (SEV-SNP)" section, subsection "Secure TSC" in APM Volume 2.
This patch set is also available at:
https://github.com/AMDESE/linux-kvm/tree/sectsc-host-latest
and is based on kvm/master
Testing Secure TSC
-----------------Secure TSC guest patches are available as part of v6.14-rc1.
QEMU changes:
https://github.com/nikunjad/qemu/tree/snp-securetsc-latestQEMU command line SEV-SNP with Secure TSC:
qemu-system-x86_64 -cpu EPYC-Milan-v2 -smp 4 \ -object memory-backend-memfd,id=ram1,size=1G,share=true,prealloc=false,reserve=false \ -object sev-snp-guest,id=sev0,cbitpos=51,reduced-phys-bits=1,secure-tsc=on,stsc-freq=2000000000 \ -machine q35,confidential-guest-support=sev0,memory-backend=ram1 \ ...Quoted 7 lines
Changelog:
----------
v6:
* Rebased on top of kvm/master
* Collected Reviewed-by/Tested-by
* s/svm->vcpu/vcpu/ insnp_launch_update_vmsa()as vcpu pointer is already available (Tom)
* Simplify assignment ofguest_protected_tsc(Tom)
Quoted 2 lines
A gentle reminder, any other suggestions/improvement ?
A gentle reminder.
Regards
Nikunj
Quoted 10 lines · Nikunj A. Dadhania
Quoted 9 lines · Nikunj A Dadhania
Quoted 7 lines · Nikunj A Dadhania
Changelog:
----------
v6:
* Rebased on top of kvm/master
* Collected Reviewed-by/Tested-by
* s/svm->vcpu/vcpu/ insnp_launch_update_vmsa()as vcpu pointer is already available (Tom)
* Simplify assignment ofguest_protected_tsc(Tom)
Quoted 5 lines
Quoted 2 lines
A gentle reminder, any other suggestions/improvement ?
A gentle reminder.
A gentle reminder.
Regards,
Nikunj
Quoted 9 lines · Nikunj A Dadhania
The hypervisor controls TSC value calculations for the guest. A malicious hypervisor can prevent the guest from progressing. The Secure TSC feature for SEV-SNP allows guests to securely use the RDTSC and RDTSCP instructions. This ensures the guest has a consistent view of time and prevents a malicious hypervisor from manipulating time, such as making it appear to move backward or advance too quickly. For more details, refer to the "Secure Nested Paging (SEV-SNP)" section, subsection "Secure TSC" in APM Volume 2.
[...]
Applied patch 2 to kvm-x86 fixes. I'll wait for v7 to grab the others (and stating the obvious, not for 6.16).
[2/4] KVM: SVM: Add missing member in SNP_LAUNCH_START command structure
https://github.com/kvm-x86/linux/commit/51a4273dcab3
Quoted 9 lines · Sean Christopherson
Quoted 8 lines · Nikunj A Dadhania
The hypervisor controls TSC value calculations for the guest. A malicious hypervisor can prevent the guest from progressing. The Secure TSC feature for SEV-SNP allows guests to securely use the RDTSC and RDTSCP instructions. This ensures the guest has a consistent view of time and prevents a malicious hypervisor from manipulating time, such as making it appear to move backward or advance too quickly. For more details, refer to the "Secure Nested Paging (SEV-SNP)" section, subsection "Secure TSC" in APM Volume 2.
Quoted 3 lines
Quoted 1 line
[...]
Applied patch 2 to kvm-x86 fixes.
Thanks Sean.
Quoted 1 line
I'll wait for v7 to grab the others (and stating the obvious, not for 6.16).
Sure will send a v7 on top of kvm-x86 changes along with the suggested changes.
Quoted 2 lines
[2/4] KVM: SVM: Add missing member in
SNP_LAUNCH_STARTcommand structurehttps://github.com/kvm-x86/linux/commit/51a4273dcab3
Regards
Nikunj