Skip to contentSkip to conversation

KVM: VMX: UMIP emulation related cleanups

PATCH Applied Sean Christopherson 1 participant 4 messages Apr 13, 2023 – Jun 2, 2023

Subsystems (3)
lore
Share & tools
lore

Sean Christopherson

author maintainer
#0 lore raw

cover KVM: VMX: UMIP emulation related cleanups

Two cleanups for UMIP related code, identified by Robert. These are *almost* purely cosmetic, I don't expect any functional impact on real world setups.

Sean Christopherson (2):
KVM: VMX: Treat UMIP as emulated if and only if the host doesn't have
UMIP
KVM: VMX: Use proper accessor to read guest CR4 in handle_desc()

 arch/x86/kvm/vmx/capabilities.h |  4 ++--
 arch/x86/kvm/vmx/nested.c       |  3 +--
 arch/x86/kvm/vmx/vmx.c          | 10 ++++++++--
 3 files changed, 11 insertions(+), 6 deletions(-)

base-commit: 4d663d83a561344d237c77a77b651f55eb7f7356

Signature · 1 line
2.40.0.634.g4ca3ef3211-goog
author maintainer
#1 lore raw

PATCH 1/2 KVM: VMX: Treat UMIP as emulated if and only if the host doesn't have UMIP

Advertise UMIP as emulated if and only if the host doesn't natively support UMIP, otherwise vmx_umip_emulated() is misleading when the host _does_ support UMIP. Of the four users of vmx_umip_emulated(), two already check for native support, and the logic in vmx_set_cpu_caps() is relevant if and only if UMIP isn't natively supported as UMIP is set in KVM's caps by kvm_set_cpu_caps() when UMIP is present in hardware.

That leaves KVM's stuffing of X86_CR4_UMIP into the default cr4_fixed1 value enumerated for nested VMX. In that case, checking for (lack of) host support is actually a bug fix of sorts, as enumerating UMIP support based solely on descriptor table exiting works only because KVM doesn't sanity check MSR_IA32_VMX_CR4_FIXED1. E.g. if a (very theoretical) host supported UMIP in hardware but didn't allow UMIP+VMX, KVM would advertise UMIP but not actually emulate UMIP. Of course, KVM would explode long before it could run a nested VM on said theoretical CPU, as KVM doesn't modify host CR4 when enabling VMX, i.e. would load an "illegal" value into vmcs.HOST_CR4.

---
 arch/x86/kvm/vmx/capabilities.h | 4 ++--
 arch/x86/kvm/vmx/nested.c       | 3 +--
 arch/x86/kvm/vmx/vmx.c          | 2 +-
 3 files changed, 4 insertions(+), 5 deletions(-)
arch/x86/kvm/vmx/capabilities.h +2 -2
index 45162c1bcd8f..d0abee35d7ba 100644
@@ -152,8 +152,8 @@ static inline bool cpu_has_vmx_ept(void)
 
 static inline bool vmx_umip_emulated(void)
 {
-	return vmcs_config.cpu_based_2nd_exec_ctrl &
-		SECONDARY_EXEC_DESC;
+	return !boot_cpu_has(X86_FEATURE_UMIP) &&
+	       (vmcs_config.cpu_based_2nd_exec_ctrl & SECONDARY_EXEC_DESC);
 }
 
 static inline bool cpu_has_vmx_rdtscp(void)
arch/x86/kvm/vmx/nested.c +1 -2
index 96ede74a6067..4fadf6c661d5 100644
@@ -2328,8 +2328,7 @@ static void prepare_vmcs02_early(struct vcpu_vmx *vmx, struct loaded_vmcs *vmcs0
 		 * Preset *DT exiting when emulating UMIP, so that vmx_set_cr4()
 		 * will not have to rewrite the controls just for this bit.
 		 */
-		if (!boot_cpu_has(X86_FEATURE_UMIP) && vmx_umip_emulated() &&
-		    (vmcs12->guest_cr4 & X86_CR4_UMIP))
+		if (vmx_umip_emulated() && (vmcs12->guest_cr4 & X86_CR4_UMIP))
 			exec_control |= SECONDARY_EXEC_DESC;
 
 		if (exec_control & SECONDARY_EXEC_VIRTUAL_INTR_DELIVERY)
arch/x86/kvm/vmx/vmx.c +1 -1
index 44fb619803b8..57464c762c79 100644
@@ -3404,7 +3404,7 @@ void vmx_set_cr4(struct kvm_vcpu *vcpu, unsigned long cr4)
 	else
 		hw_cr4 |= KVM_PMODE_VM_CR4_ALWAYS_ON;
 
-	if (!boot_cpu_has(X86_FEATURE_UMIP) && vmx_umip_emulated()) {
+	if (vmx_umip_emulated()) {
 		if (cr4 & X86_CR4_UMIP) {
 			secondary_exec_controls_setbit(vmx, SECONDARY_EXEC_DESC);
 			hw_cr4 &= ~X86_CR4_UMIP;
Signature · 1 line
2.40.0.634.g4ca3ef3211-goog
author maintainer
#2 lore raw

PATCH 2/2 KVM: VMX: Use proper accessor to read guest CR4 in handle_desc()

In reply to Sean Christopherson

Use kvm_is_cr4_bit_set() to read guest CR4.UMIP when sanity checking that a descriptor table VM-Exit occurs if and only if guest.CR4.UMIP=1. UMIP can't be guest-owned, i.e. using kvm_read_cr4_bits() to decache guest- owned bits isn't strictly necessary, but eliminating raw reads of vcpu->arch.cr4 is desirable as it makes it easy to visually audit KVM for correctness.

Opportunistically add a compile-time assertion that UMIP isn't guest-owned as letting the guest own UMIP isn't compatible with emulation (or any CR4 bit that is emulated by KVM).

Opportunistically change the WARN_ON() to a ONCE variant. When the WARN fires, it fires _a lot_, and spamming the kernel logs ends up doing more harm than whatever led to KVM's unnecessary emulation.

---
 arch/x86/kvm/vmx/vmx.c | 8 +++++++-
 1 file changed, 7 insertions(+), 1 deletion(-)
arch/x86/kvm/vmx/vmx.c +7 -1
index 57464c762c79..f0975e790f7f 100644
@@ -5402,7 +5402,13 @@ static int handle_set_cr4(struct kvm_vcpu *vcpu, unsigned long val)
 
 static int handle_desc(struct kvm_vcpu *vcpu)
 {
-	WARN_ON(!(vcpu->arch.cr4 & X86_CR4_UMIP));
+	/*
+	 * UMIP emulation relies on intercepting writes to CR4.UMIP, i.e. this
+	 * and other code needs to be updated if UMIP can be guest owned.
+	 */
+	BUILD_BUG_ON(KVM_POSSIBLE_CR4_GUEST_BITS & X86_CR4_UMIP);
+
+	WARN_ON_ONCE(!kvm_is_cr4_bit_set(vcpu, X86_CR4_UMIP));
 	return kvm_emulate_instruction(vcpu, 0);
 }
 
Signature · 1 line
2.40.0.634.g4ca3ef3211-goog
author maintainer Applied
#3 lore raw
In reply to Sean Christopherson
Quoted 10 lines · Sean Christopherson

Two cleanups for UMIP related code, identified by Robert. These are *almost* purely cosmetic, I don't expect any functional impact on real world setups.

Sean Christopherson (2):
KVM: VMX: Treat UMIP as emulated if and only if the host doesn't have
UMIP
KVM: VMX: Use proper accessor to read guest CR4 in handle_desc()

[...]

Applied to kvm-x86 vmx, thanks!

[1/2] KVM: VMX: Treat UMIP as emulated if and only if the host doesn't have UMIP

      https://github.com/kvm-x86/linux/commit/3243b93c16d9
[2/2] KVM: VMX: Use proper accessor to read guest CR4 in handle_desc()
      https://github.com/kvm-x86/linux/commit/023cfa6fc200

--
https://github.com/kvm-x86/linux/tree/next
https://github.com/kvm-x86/linux/tree/fixes