Skip to contentSkip to conversation

KVM: nSVM: TLB_CONTROL / FLUSHBYASID "fixes"

PATCH Applied Sean Christopherson 3 participants 7 messages Oct 18, 2023 – Dec 1, 2023

Subsystems (3)
lore
Share & tools
lore

Sean Christopherson

author maintainer
#0 lore raw

cover KVM: nSVM: TLB_CONTROL / FLUSHBYASID "fixes"

Two "fixes" to play nice with running VMware Workstation on top of KVM, in quotes because patch 2 isn't really a fix.

Sean Christopherson (2):
Revert "nSVM: Check for reserved encodings of TLB_CONTROL in nested
VMCB"
KVM: nSVM: Advertise support for flush-by-ASID

 arch/x86/kvm/svm/nested.c | 15 ---------------
 arch/x86/kvm/svm/svm.c    |  1 +
 2 files changed, 1 insertion(+), 15 deletions(-)

base-commit: 437bba5ad2bba00c2056c896753a32edf80860cc

Signature · 1 line
2.42.0.655.g421f12c284-goog
author maintainer
#1 lore raw

PATCH 1/2 Revert "nSVM: Check for reserved encodings of TLB_CONTROL in nested VMCB"

Revert KVM's made-up consistency check on SVM's TLB control. The APM says that unsupported encodings are reserved, but the APM doesn't state that VMRUN checks for a supported encoding. Unless something is called out in "Canonicalization and Consistency Checks" or listed as MBZ (Must Be Zero), AMD behavior is typically to let software shoot itself in the foot.

This reverts commit 174a921b6975ef959dd82ee9e8844067a62e3ec1.

Fixes 174a921b6975 ("nSVM: Check for reserved encodings of TLB_CONTROL in nested VMCB")
---
 arch/x86/kvm/svm/nested.c | 15 ---------------
 1 file changed, 15 deletions(-)
arch/x86/kvm/svm/nested.c +0 -15
index 3fea8c47679e..60891b9ce25f 100644
@@ -247,18 +247,6 @@ static bool nested_svm_check_bitmap_pa(struct kvm_vcpu *vcpu, u64 pa, u32 size)
 	    kvm_vcpu_is_legal_gpa(vcpu, addr + size - 1);
 }
 
-static bool nested_svm_check_tlb_ctl(struct kvm_vcpu *vcpu, u8 tlb_ctl)
-{
-	/* Nested FLUSHBYASID is not supported yet.  */
-	switch(tlb_ctl) {
-		case TLB_CONTROL_DO_NOTHING:
-		case TLB_CONTROL_FLUSH_ALL_ASID:
-			return true;
-		default:
-			return false;
-	}
-}
-
 static bool __nested_vmcb_check_controls(struct kvm_vcpu *vcpu,
 					 struct vmcb_ctrl_area_cached *control)
 {
@@ -278,9 +266,6 @@ static bool __nested_vmcb_check_controls(struct kvm_vcpu *vcpu,
 					   IOPM_SIZE)))
 		return false;
 
-	if (CC(!nested_svm_check_tlb_ctl(vcpu, control->tlb_ctl)))
-		return false;
-
 	if (CC((control->int_ctl & V_NMI_ENABLE_MASK) &&
 	       !vmcb12_is_intercept(control, INTERCEPT_NMI))) {
 		return false;
Signature · 1 line
2.42.0.655.g421f12c284-goog
Reviewed-by on v1 1/2
#2 lore raw

Re: [PATCH 1/2] Revert "nSVM: Check for reserved encodings of TLB_CONTROL in nested VMCB"

Quoted 41 lines · Sean Christopherson

Revert KVM's made-up consistency check on SVM's TLB control. The APM says that unsupported encodings are reserved, but the APM doesn't state that VMRUN checks for a supported encoding. Unless something is called out in "Canonicalization and Consistency Checks" or listed as MBZ (Must Be Zero), AMD behavior is typically to let software shoot itself in the foot.

This reverts commit 174a921b6975ef959dd82ee9e8844067a62e3ec1.

Fixes 174a921b6975 ("nSVM: Check for reserved encodings of TLB_CONTROL in nested VMCB")
---
 arch/x86/kvm/svm/nested.c | 15 ---------------
 1 file changed, 15 deletions(-)
arch/x86/kvm/svm/nested.c +0 -12
index 3fea8c47679e..60891b9ce25f 100644
@@ -247,18 +247,6 @@ static bool nested_svm_check_bitmap_pa(struct kvm_vcpu *vcpu, u64 pa, u32 size)
 	    kvm_vcpu_is_legal_gpa(vcpu, addr + size - 1);
 }
 
-static bool nested_svm_check_tlb_ctl(struct kvm_vcpu *vcpu, u8 tlb_ctl)
-{
-	/* Nested FLUSHBYASID is not supported yet.  */
-	switch(tlb_ctl) {
-		case TLB_CONTROL_DO_NOTHING:
-		case TLB_CONTROL_FLUSH_ALL_ASID:
-			return true;
-		default:
-			return false;
-	}
-}
-
 static bool __nested_vmcb_check_controls(struct kvm_vcpu *vcpu,
 					 struct vmcb_ctrl_area_cached *control)
 {

arch/x86/kvm/svm/nested.c static bool __nested_vmcb_check_controls(struct kvm_vcpu *vcpu, · line 271

Quoted 9 lines
arch/x86/kvm/svm/nested.c +0 -3
@@ -278,9 +266,6 @@ static bool __nested_vmcb_check_controls(struct kvm_vcpu *vcpu,
 					   IOPM_SIZE)))
 		return false;
 
-	if (CC(!nested_svm_check_tlb_ctl(vcpu, control->tlb_ctl)))
-		return false;
-
 	if (CC((control->int_ctl & V_NMI_ENABLE_MASK) &&
 	       !vmcb12_is_intercept(control, INTERCEPT_NMI))) {
 		return false;

Yes, after checking Jim's comment (*) on this I still agree that revert is OK. KVM never passes through the tlb_ctl field (but does copy it to the cache), thus there is no need to sanitize it.

https://www.spinics.net/lists/kvm/msg316072.html

Best regards,
Maxim Levitsky

author maintainer
#3 lore raw

PATCH 2/2 KVM: nSVM: Advertise support for flush-by-ASID

In reply to Sean Christopherson

Advertise support for FLUSHBYASID when nested SVM is enabled, as KVM can always emulate flushing TLB entries for a vmcb12 ASID, e.g. by running L2 with a new, fresh ASID in vmcb02. Some modern hypervisors, e.g. VMWare Workstation 17, require FLUSHBYASID support and will refuse to run if it's not present.

Punt on proper support, as "Honor L1's request to flush an ASID on nested VMRUN" is one of the TODO items in the (incomplete) list of issues that need to be addressed in order for KVM to NOT do a full TLB flush on every nested SVM transition (see nested_svm_transition_tlb_flush()).

---
 arch/x86/kvm/svm/svm.c | 1 +
 1 file changed, 1 insertion(+)
arch/x86/kvm/svm/svm.c +1 -0
index 1785de7dc98b..9cf7eef161ff 100644
@@ -5083,6 +5083,7 @@ static __init void svm_set_cpu_caps(void)
 	if (nested) {
 		kvm_cpu_cap_set(X86_FEATURE_SVM);
 		kvm_cpu_cap_set(X86_FEATURE_VMCBCLEAN);
+		kvm_cpu_cap_set(X86_FEATURE_FLUSHBYASID);
 
 		if (nrips)
 			kvm_cpu_cap_set(X86_FEATURE_NRIPS);
Signature · 1 line
2.42.0.655.g421f12c284-goog
Reviewed-by on v1 2/2
#4 lore raw

Re: [PATCH 2/2] KVM: nSVM: Advertise support for flush-by-ASID

Quoted 23 lines · Sean Christopherson

Advertise support for FLUSHBYASID when nested SVM is enabled, as KVM can always emulate flushing TLB entries for a vmcb12 ASID, e.g. by running L2 with a new, fresh ASID in vmcb02. Some modern hypervisors, e.g. VMWare Workstation 17, require FLUSHBYASID support and will refuse to run if it's not present.

Punt on proper support, as "Honor L1's request to flush an ASID on nested VMRUN" is one of the TODO items in the (incomplete) list of issues that need to be addressed in order for KVM to NOT do a full TLB flush on every nested SVM transition (see nested_svm_transition_tlb_flush()).

---
 arch/x86/kvm/svm/svm.c | 1 +
 1 file changed, 1 insertion(+)
arch/x86/kvm/svm/svm.c +0 -0
index 1785de7dc98b..9cf7eef161ff 100644

arch/x86/kvm/svm/svm.c static __init void svm_set_cpu_caps(void) · line 5089

Quoted 7 lines
arch/x86/kvm/svm/svm.c +1 -0
@@ -5083,6 +5083,7 @@ static __init void svm_set_cpu_caps(void)
 	if (nested) {
 		kvm_cpu_cap_set(X86_FEATURE_SVM);
 		kvm_cpu_cap_set(X86_FEATURE_VMCBCLEAN);
+		kvm_cpu_cap_set(X86_FEATURE_FLUSHBYASID);
 
 		if (nrips)
 			kvm_cpu_cap_set(X86_FEATURE_NRIPS);

Nitpick: if you think that this is worth it,
maybe we can add a comment here on why we can 'support' the flushbyasid feature? in addition to the commit message.

Other than that:

Best regards,
Maxim Levitsky

#5 lore raw
In reply to Sean Christopherson
Quoted 3 lines · Sean Christopherson

Two "fixes" to play nice with running VMware Workstation on top of KVM, in quotes because patch 2 isn't really a fix.

Hey, thanks for providing these patches. Tested them here with ESXi 7, they work like a charm!

As a sidenote regarding my tests: had to fiddle a bit with the reverted commit as I applied them on top of the v6.2 tag, because that is still the kernel version we use over here.

Quoted 13 lines

Sean Christopherson (2):
Revert "nSVM: Check for reserved encodings of TLB_CONTROL in nested
VMCB"
KVM: nSVM: Advertise support for flush-by-ASID

 arch/x86/kvm/svm/nested.c | 15 ---------------
 arch/x86/kvm/svm/svm.c    |  1 +
 2 files changed, 1 insertion(+), 15 deletions(-)

base-commit: 437bba5ad2bba00c2056c896753a32edf80860cc --
2.42.0.655.g421f12c284-goog

author maintainer Applied
#6 lore raw
In reply to Sean Christopherson
Quoted 9 lines · Sean Christopherson

Two "fixes" to play nice with running VMware Workstation on top of KVM, in quotes because patch 2 isn't really a fix.

Sean Christopherson (2):
Revert "nSVM: Check for reserved encodings of TLB_CONTROL in nested
VMCB"
KVM: nSVM: Advertise support for flush-by-ASID

[...]

Applied to kvm-x86 svm, with a comment as suggested by Maxim.

[1/2] Revert "nSVM: Check for reserved encodings of TLB_CONTROL in nested VMCB"

      https://github.com/kvm-x86/linux/commit/a484755ab252
[2/2] KVM: nSVM: Advertise support for flush-by-ASID
      https://github.com/kvm-x86/linux/commit/176bfc5b17fe

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