Skip to contentSkip to conversation

KVM: VMX: PT (RTIT) bug fix and cleanup

PATCH Applied Sean Christopherson 3 participants 6 messages Dec 10, 2019 – Jan 15, 2020

Subsystems (3)
lore
Share & tools
lore
#0 lore raw

cover KVM: VMX: PT (RTIT) bug fix and cleanup

Add a missing non-canonical check on writes to the RTIT address MSRs and tack on a cleanup patch.

** ALL PATCHES ARE COMPLETELY UNTESTED **

Untested due to lack of hardware.

Sean Christopherson (2):
KVM: VMX: Add non-canonical check on writes to RTIT address MSRs
KVM: VMX: Add helper to consolidate up PT/RTIT WRMSR fault logic

 arch/x86/kvm/vmx/vmx.c | 57 ++++++++++++++++++++++++------------------
 1 file changed, 33 insertions(+), 24 deletions(-)
Signature · 1 line
2.24.0
#1 lore raw

PATCH 1/2 KVM: VMX: Add non-canonical check on writes to RTIT address MSRs

Reject writes to RTIT address MSRs if the data being written is a non-canonical address as the MSRs are subject to canonical checks, e.g. KVM will trigger an unchecked #GP when loading the values to hardware during pt_guest_enter().

---
 arch/x86/kvm/vmx/vmx.c | 2 ++
 1 file changed, 2 insertions(+)
arch/x86/kvm/vmx/vmx.c +2 -0
index 51e3b27f90ed..9aa2006dbe04 100644
@@ -2152,6 +2152,8 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 			(index >= 2 * intel_pt_validate_cap(vmx->pt_desc.caps,
 					PT_CAP_num_address_ranges)))
 			return 1;
+		if (is_noncanonical_address(data, vcpu))
+			return 1;
 		if (index % 2)
 			vmx->pt_desc.guest.addr_b[index / 2] = data;
 		else
Signature · 1 line
2.24.0
#2 lore raw

RE: [PATCH 1/2] KVM: VMX: Add non-canonical check on writes to RTIT address MSRs

Quoted 25 lines

-----Original Message-----
From: Christopherson, Sean J <[email protected]>
Sent: Wednesday, December 11, 2019 7:25 AM
To: Paolo Bonzini <[email protected]>
Cc: Christopherson, Sean J <[email protected]>; Vitaly Kuznetsov <[email protected]>; Wanpeng Li <[email protected]>; Jim Mattson <[email protected]>; Joerg Roedel <[email protected]>; [email protected]; linux- [email protected]; Chao Peng <[email protected]>; Kang, Luwei <[email protected]>
Subject: [PATCH 1/2] KVM: VMX: Add non-canonical check on writes to RTIT address MSRs

Reject writes to RTIT address MSRs if the data being written is a non-canonical address as the MSRs are subject to canonical checks, e.g.
KVM will trigger an unchecked #GP when loading the values to hardware during pt_guest_enter().

---
 arch/x86/kvm/vmx/vmx.c | 2 ++
 1 file changed, 2 insertions(+)

arch/x86/kvm/vmx/vmx.c static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info) · line 2156

Quoted 5 lines
arch/x86/kvm/vmx/vmx.c +2 -0
@@ -2152,6 +2152,8 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 			(index >= 2 * intel_pt_validate_cap(vmx->pt_desc.caps,
 					PT_CAP_num_address_ranges)))
 			return 1;
+		if (is_noncanonical_address(data, vcpu))
+			return 1;

Is this for live migrate a VM with 5 level page table to the VM with 4 level page table?

Thanks,
Luwei Kang

Quoted 5 lines
arch/x86/kvm/vmx/vmx.c +0 -1
 		if (index % 2)
 			vmx->pt_desc.guest.addr_b[index / 2] = data;
 		else
--

2.24.0

#3 lore raw

Re: [PATCH 1/2] KVM: VMX: Add non-canonical check on writes to RTIT address MSRs

Quoted 10 lines · Luwei
Quoted 7 lines
arch/x86/kvm/vmx/vmx.c index renamed from arch/x86/kvm/vmx/vmx.c +0 -0

51e3b27f90ed..9aa2006dbe04 100644 --- a/arch/x86/kvm/vmx/vmx.c +++ b/arch/x86/kvm/vmx/vmx.c @@ -2152,6 +2152,8 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info) (index >= 2 * intel_pt_validate_cap(vmx->pt_desc.caps, PT_CAP_num_address_ranges)))
return 1; + if (is_noncanonical_address(data, vcpu)) +
return 1;

Is this for live migrate a VM with 5 level page table to the VM with 4 level page table?

This is orthogonal to live migration or 5-level paging. Unless I'm missing something, KVM simply fails to validate the incoming address.

#4 lore raw

PATCH 2/2 KVM: VMX: Add helper to consolidate up PT/RTIT WRMSR fault logic

In reply to Sean Christopherson

Add a helper to consolidate the common checks for writing PT MSRs, and opportunistically clean up the formatting of the affected code.

No functional change intended.

---
 arch/x86/kvm/vmx/vmx.c | 55 ++++++++++++++++++++++++------------------
 1 file changed, 31 insertions(+), 24 deletions(-)
arch/x86/kvm/vmx/vmx.c +31 -24
index 9aa2006dbe04..22e05e3fb0a3 100644
@@ -1057,6 +1057,12 @@ static unsigned long segment_base(u16 selector)
 }
 #endif
 
+static inline bool pt_can_write_msr(struct vcpu_vmx *vmx)
+{
+	return (pt_mode == PT_MODE_HOST_GUEST) &&
+	       !(vmx->pt_desc.guest.ctl & RTIT_CTL_TRACEEN);
+}
+
 static inline void pt_load_msr(struct pt_ctx *ctx, u32 addr_range)
 {
 	u32 i;
@@ -2110,47 +2116,48 @@ static int vmx_set_msr(struct kvm_vcpu *vcpu, struct msr_data *msr_info)
 		pt_update_intercept_for_msr(vmx);
 		break;
 	case MSR_IA32_RTIT_STATUS:
-		if ((pt_mode != PT_MODE_HOST_GUEST) ||
-			(vmx->pt_desc.guest.ctl & RTIT_CTL_TRACEEN) ||
-			(data & MSR_IA32_RTIT_STATUS_MASK))
+		if (!pt_can_write_msr(vmx))
+			return 1;
+		if (data & MSR_IA32_RTIT_STATUS_MASK)
 			return 1;
 		vmx->pt_desc.guest.status = data;
 		break;
 	case MSR_IA32_RTIT_CR3_MATCH:
-		if ((pt_mode != PT_MODE_HOST_GUEST) ||
-			(vmx->pt_desc.guest.ctl & RTIT_CTL_TRACEEN) ||
-			!intel_pt_validate_cap(vmx->pt_desc.caps,
-						PT_CAP_cr3_filtering))
+		if (!pt_can_write_msr(vmx))
+			return 1;
+		if (!intel_pt_validate_cap(vmx->pt_desc.caps,
+					   PT_CAP_cr3_filtering))
 			return 1;
 		vmx->pt_desc.guest.cr3_match = data;
 		break;
 	case MSR_IA32_RTIT_OUTPUT_BASE:
-		if ((pt_mode != PT_MODE_HOST_GUEST) ||
-			(vmx->pt_desc.guest.ctl & RTIT_CTL_TRACEEN) ||
-			(!intel_pt_validate_cap(vmx->pt_desc.caps,
-					PT_CAP_topa_output) &&
-			 !intel_pt_validate_cap(vmx->pt_desc.caps,
-					PT_CAP_single_range_output)) ||
-			(data & MSR_IA32_RTIT_OUTPUT_BASE_MASK))
+		if (!pt_can_write_msr(vmx))
+			return 1;
+		if (!intel_pt_validate_cap(vmx->pt_desc.caps,
+					   PT_CAP_topa_output) &&
+		    !intel_pt_validate_cap(vmx->pt_desc.caps,
+					   PT_CAP_single_range_output))
+			return 1;
+		if (data & MSR_IA32_RTIT_OUTPUT_BASE_MASK)
 			return 1;
 		vmx->pt_desc.guest.output_base = data;
 		break;
 	case MSR_IA32_RTIT_OUTPUT_MASK:
-		if ((pt_mode != PT_MODE_HOST_GUEST) ||
-			(vmx->pt_desc.guest.ctl & RTIT_CTL_TRACEEN) ||
-			(!intel_pt_validate_cap(vmx->pt_desc.caps,
-					PT_CAP_topa_output) &&
-			 !intel_pt_validate_cap(vmx->pt_desc.caps,
-					PT_CAP_single_range_output)))
+		if (!pt_can_write_msr(vmx))
+			return 1;
+		if (!intel_pt_validate_cap(vmx->pt_desc.caps,
+					   PT_CAP_topa_output) &&
+		    !intel_pt_validate_cap(vmx->pt_desc.caps,
+					   PT_CAP_single_range_output))
 			return 1;
 		vmx->pt_desc.guest.output_mask = data;
 		break;
 	case MSR_IA32_RTIT_ADDR0_A ... MSR_IA32_RTIT_ADDR3_B:
+		if (!pt_can_write_msr(vmx))
+			return 1;
 		index = msr_info->index - MSR_IA32_RTIT_ADDR0_A;
-		if ((pt_mode != PT_MODE_HOST_GUEST) ||
-			(vmx->pt_desc.guest.ctl & RTIT_CTL_TRACEEN) ||
-			(index >= 2 * intel_pt_validate_cap(vmx->pt_desc.caps,
-					PT_CAP_num_address_ranges)))
+		if (index >= 2 * intel_pt_validate_cap(vmx->pt_desc.caps,
+						       PT_CAP_num_address_ranges))
 			return 1;
 		if (is_noncanonical_address(data, vcpu))
 			return 1;
Signature · 1 line
2.24.0
maintainer Applied
#5 lore raw
In reply to Sean Christopherson
Quoted 5 lines · Sean Christopherson

Add a missing non-canonical check on writes to the RTIT address MSRs and tack on a cleanup patch.

** ALL PATCHES ARE COMPLETELY UNTESTED **

Quoted 9 lines

Untested due to lack of hardware.

Sean Christopherson (2):
KVM: VMX: Add non-canonical check on writes to RTIT address MSRs
KVM: VMX: Add helper to consolidate up PT/RTIT WRMSR fault logic

 arch/x86/kvm/vmx/vmx.c | 57 ++++++++++++++++++++++++------------------
 1 file changed, 33 insertions(+), 24 deletions(-)

Queued, thanks.

Paolo