Skip to contentSkip to conversation

KVM: arm64: Expose CSV3 to guests on running on Meltdown-safe HW

PATCH In review Marc Zyngier 2 participants 4 messages Dec 4, 2020 – Dec 10, 2020

Subsystems (3)
lore
Share & tools
lore

Marc Zyngier

author maintainer
#0 lore raw

v2 cover KVM: arm64: Expose CSV3 to guests on running on Meltdown-safe HW

Will recently pointed out that when running on big-little systems that are known not to be vulnerable to Metldown, guests are not presented with the CSV3 property if the physical HW include a core that doesn't have CSV3, despite being known to be safe (it is on the kpti_safe_list).

Since this is valuable information that can be cheaply given to the guest, let's just do that. The scheme is the same as what we do for CSV2, allowing userspace to change the default setting if this doesn't advertise a safer setting than what the kernel thinks it is.

* From v1:
- Fix the clearing of ID_AA64PFR0_EL1.CSV3 on update from userspace
- Actually store the userspace value

Marc Zyngier (2):
arm64: Make the Meltdown mitigation state available
KVM: arm64: Advertise ID_AA64PFR0_EL1.CSV3=1 if the CPUs are
Meltdown-safe

 arch/arm64/include/asm/kvm_host.h |  1 +
 arch/arm64/include/asm/spectre.h  |  2 ++
 arch/arm64/kernel/cpufeature.c    | 20 +++++++++++++++++---
 arch/arm64/kvm/arm.c              |  6 ++++--
 arch/arm64/kvm/sys_regs.c         | 16 +++++++++++++---
 5 files changed, 37 insertions(+), 8 deletions(-)
Signature · 6 lines
2.28.0

_______________________________________________
kvmarm mailing list
[email protected]
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
author maintainer
#2 lore raw

PATCH v2 2/2 KVM: arm64: Advertise ID_AA64PFR0_EL1.CSV3=1 if the CPUs are Meltdown-safe

Cores that predate the introduction of ID_AA64PFR0_EL1.CSV3 to the ARMv8 architecture have this field set to 0, even of some of them are not affected by the vulnerability.

The kernel maintains a list of unaffected cores (A53, A55 and a few others) so that it doesn't impose an expensive mitigation uncessarily.

As we do for CSV2, let's expose the CSV3 property to guests that run on HW that is effectively not vulnerable. This can be reset to zero by writing to the ID register from userspace, ensuring that VMs can be migrated despite the new property being set.

---
 arch/arm64/include/asm/kvm_host.h |  1 +
 arch/arm64/kvm/arm.c              |  6 ++++--
 arch/arm64/kvm/sys_regs.c         | 16 +++++++++++++---
 3 files changed, 18 insertions(+), 5 deletions(-)
arch/arm64/include/asm/kvm_host.h +1 -0
index 0cd9f0f75c13..147347028a20 100644
@@ -120,6 +120,7 @@ struct kvm_arch {
 	unsigned int pmuver;
 
 	u8 pfr0_csv2;
+	u8 pfr0_csv3;
 };
 
 struct kvm_vcpu_fault_info {
arch/arm64/kvm/arm.c +4 -2
index c0ffb019ca8b..dc3fa6a0f9e5 100644
@@ -102,7 +102,7 @@ static int kvm_arm_default_max_vcpus(void)
 	return vgic_present ? kvm_vgic_get_max_vcpus() : KVM_MAX_VCPUS;
 }
 
-static void set_default_csv2(struct kvm *kvm)
+static void set_default_spectre(struct kvm *kvm)
 {
 	/*
 	 * The default is to expose CSV2 == 1 if the HW isn't affected.
@@ -114,6 +114,8 @@ static void set_default_csv2(struct kvm *kvm)
 	 */
 	if (arm64_get_spectre_v2_state() == SPECTRE_UNAFFECTED)
 		kvm->arch.pfr0_csv2 = 1;
+	if (arm64_get_meltdown_state() == SPECTRE_UNAFFECTED)
+		kvm->arch.pfr0_csv3 = 1;
 }
 
 /**
@@ -141,7 +143,7 @@ int kvm_arch_init_vm(struct kvm *kvm, unsigned long type)
 	/* The maximum number of VCPUs is limited by the host's GIC model */
 	kvm->arch.max_vcpus = kvm_arm_default_max_vcpus();
 
-	set_default_csv2(kvm);
+	set_default_spectre(kvm);
 
 	return ret;
 out_free_stage2_pgd:
arch/arm64/kvm/sys_regs.c +13 -3
index c1fac9836af1..5fee43dc66cf 100644
@@ -1122,6 +1122,8 @@ static u64 read_id_reg(const struct kvm_vcpu *vcpu,
 		val &= ~(0xfUL << ID_AA64PFR0_AMU_SHIFT);
 		val &= ~(0xfUL << ID_AA64PFR0_CSV2_SHIFT);
 		val |= ((u64)vcpu->kvm->arch.pfr0_csv2 << ID_AA64PFR0_CSV2_SHIFT);
+		val &= ~(0xfUL << ID_AA64PFR0_CSV3_SHIFT);
+		val |= ((u64)vcpu->kvm->arch.pfr0_csv3 << ID_AA64PFR0_CSV3_SHIFT);
 	} else if (id == SYS_ID_AA64PFR1_EL1) {
 		val &= ~(0xfUL << ID_AA64PFR1_MTE_SHIFT);
 	} else if (id == SYS_ID_AA64ISAR1_EL1 && !vcpu_has_ptrauth(vcpu)) {
@@ -1209,9 +1211,9 @@ static int set_id_aa64pfr0_el1(struct kvm_vcpu *vcpu,
 			       const struct kvm_one_reg *reg, void __user *uaddr)
 {
 	const u64 id = sys_reg_to_index(rd);
+	u8 csv2, csv3;
 	int err;
 	u64 val;
-	u8 csv2;
 
 	err = reg_from_user(&val, uaddr, id);
 	if (err)
@@ -1227,13 +1229,21 @@ static int set_id_aa64pfr0_el1(struct kvm_vcpu *vcpu,
 	    (csv2 && arm64_get_spectre_v2_state() != SPECTRE_UNAFFECTED))
 		return -EINVAL;
 
-	/* We can only differ with CSV2, and anything else is an error */
+	/* Same thing for CSV3 */
+	csv3 = cpuid_feature_extract_unsigned_field(val, ID_AA64PFR0_CSV3_SHIFT);
+	if (csv3 > 1 ||
+	    (csv3 && arm64_get_meltdown_state() != SPECTRE_UNAFFECTED))
+		return -EINVAL;
+
+	/* We can only differ with CSV[23], and anything else is an error */
 	val ^= read_id_reg(vcpu, rd, false);
-	val &= ~(0xFUL << ID_AA64PFR0_CSV2_SHIFT);
+	val &= ~((0xFUL << ID_AA64PFR0_CSV2_SHIFT) |
+		 (0xFUL << ID_AA64PFR0_CSV3_SHIFT));
 	if (val)
 		return -EINVAL;
 
 	vcpu->kvm->arch.pfr0_csv2 = csv2;
+	vcpu->kvm->arch.pfr0_csv3 = csv3;
 
 	return 0;
 }
Signature · 6 lines
2.28.0

_______________________________________________
kvmarm mailing list
[email protected]
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
author maintainer
#1 lore raw

PATCH v2 1/2 arm64: Make the Meltdown mitigation state available

In reply to Marc Zyngier

Our Meltdown mitigation state isn't exposed outside of cpufeature.c, contrary to the rest of the Spectre mitigation state. As we are going to use it in KVM, expose a arm64_get_meltdown_state() helper which returns the same possible values as arm64_get_spectre_v?_state().

---
 arch/arm64/include/asm/spectre.h |  2 ++
 arch/arm64/kernel/cpufeature.c   | 20 +++++++++++++++++---
 2 files changed, 19 insertions(+), 3 deletions(-)
arch/arm64/include/asm/spectre.h +2 -0
index fcdfbce302bd..52e788981f4a 100644
@@ -29,4 +29,6 @@ bool has_spectre_v4(const struct arm64_cpu_capabilities *cap, int scope);
 void spectre_v4_enable_mitigation(const struct arm64_cpu_capabilities *__unused);
 void spectre_v4_enable_task_mitigation(struct task_struct *tsk);
 
+enum mitigation_state arm64_get_meltdown_state(void);
+
 #endif	/* __ASM_SPECTRE_H */
arch/arm64/kernel/cpufeature.c +17 -3
index 6f36c4f62f69..280b10762f6b 100644
@@ -2846,14 +2846,28 @@ static int __init enable_mrs_emulation(void)
 
 core_initcall(enable_mrs_emulation);
 
+enum mitigation_state arm64_get_meltdown_state(void)
+{
+	if (__meltdown_safe)
+		return SPECTRE_UNAFFECTED;
+
+	if (arm64_kernel_unmapped_at_el0())
+		return SPECTRE_MITIGATED;
+
+	return SPECTRE_VULNERABLE;
+}
+
 ssize_t cpu_show_meltdown(struct device *dev, struct device_attribute *attr,
 			  char *buf)
 {
-	if (__meltdown_safe)
+	switch (arm64_get_meltdown_state()) {
+	case SPECTRE_UNAFFECTED:
 		return sprintf(buf, "Not affected\n");
 
-	if (arm64_kernel_unmapped_at_el0())
+	case SPECTRE_MITIGATED:
 		return sprintf(buf, "Mitigation: PTI\n");
 
-	return sprintf(buf, "Vulnerable\n");
+	default:
+		return sprintf(buf, "Vulnerable\n");
+	}
 }
Signature · 6 lines
2.28.0

_______________________________________________
kvmarm mailing list
[email protected]
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm
maintainer Acked-by on the v2 series
#3 lore raw
In reply to Marc Zyngier
Quoted 10 lines · Marc Zyngier

Will recently pointed out that when running on big-little systems that are known not to be vulnerable to Metldown, guests are not presented with the CSV3 property if the physical HW include a core that doesn't have CSV3, despite being known to be safe (it is on the kpti_safe_list).

Since this is valuable information that can be cheaply given to the guest, let's just do that. The scheme is the same as what we do for CSV2, allowing userspace to change the default setting if this doesn't advertise a safer setting than what the kernel thinks it is.

Quoted 8 lines

* From v1:
- Fix the clearing of ID_AA64PFR0_EL1.CSV3 on update from userspace
- Actually store the userspace value

Marc Zyngier (2):
arm64: Make the Meltdown mitigation state available
KVM: arm64: Advertise ID_AA64PFR0_EL1.CSV3=1 if the CPUs are
Meltdown-safe

Will
_______________________________________________
kvmarm mailing list
[email protected]
https://lists.cs.columbia.edu/mailman/listinfo/kvmarm