Skip to contentSkip to conversation

KVM: nSVM: Gate CET state copy in svm_copy_vmrun_state() on guest SHSTK

PATCH Superseded Jim Mattson 3 participants 6 messages Nov 12 – Nov 13

Subsystems (3)

v1 of 2 Superseded Latest is v2 →

lore
Share & tools
lore
#0 lore raw

PATCH 1/1 KVM: nSVM: Gate CET state copy in svm_copy_vmrun_state() on guest SHSTK

Like nested_vmcb02_prepare_save() and nested_svm_vmexit_update_vmcb12(), svm_copy_vmrun_state() should gate copying s_cet, isst_addr, and ssp on guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK) (i.e. the feature is advertised to the guest). Instead, copying is gated on kvm_cpu_cap_has(X86_FEATURE_SHSTK) (i.e. the host supports shadow stacks).

Check guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK) instead of kvm_cpu_cap_has(X86_FEATURE_SHSTK) in svm_copy_vmrun_state().

Fixes c5ba49458513 ("KVM: nSVM: Save/load CET Shadow Stack state to/from vmcb12/vmcb02")
---
 arch/x86/kvm/svm/nested.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kvm/svm/nested.c +1 -1
index 1299fdadeb73..b73bc9c3ebdb 100644
@@ -1248,7 +1248,7 @@ void svm_copy_vmrun_state(struct kvm_vcpu *vcpu, struct vmcb_save_area *to_save,
 	to_save->rip = from_save->rip;
 	to_save->cpl = 0;
 
-	if (kvm_cpu_cap_has(X86_FEATURE_SHSTK)) {
+	if (guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK)) {
 		to_save->s_cet  = from_save->s_cet;
 		to_save->isst_addr = from_save->isst_addr;
 		to_save->ssp = from_save->ssp;
Signature · 1 line
2.56.1.703.gf848ed807c-goog
#1 lore raw
Thank you for your contribution! Sashiko AI review found 1 potential issue to consider: - [High] KVM: nSVM: undeclared vcpu in svm_copy_vmrun_state() breaks build --

Thank you for your contribution! Sashiko AI review found 1 potential issue to consider:
- [High] KVM: nSVM: undeclared vcpu in svm_copy_vmrun_state() breaks build
--

commit 11b89023661c173dccf1cc07088879a6b4ff0bef
Author: Jim Mattson <[email protected]>

KVM: nSVM: Gate CET state copy in svm_copy_vmrun_state() on guest SHSTK

This commit changes svm_copy_vmrun_state() to check guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK) instead of kvm_cpu_cap_has(X86_FEATURE_SHSTK) before copying CET shadow stack fields.

arch/x86/kvm/svm/nested.c void svm_copy_vmrun_state(struct vmcb_save_area *to_save, · line 1224

Quoted 13 lines
arch/x86/kvm/svm/nested.c +1 -1
index f9090b601efaa..2a1c279bd54a1 100644
@@ -1218,7 +1218,7 @@ void svm_copy_vmrun_state(struct vmcb_save_area *to_save,
 	to_save->rip = from_save->rip;
 	to_save->cpl = 0;
 
-	if (kvm_cpu_cap_has(X86_FEATURE_SHSTK)) {
+	if (guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK)) {
 		to_save->s_cet  = from_save->s_cet;
 		to_save->isst_addr = from_save->isst_addr;
 		to_save->ssp = from_save->ssp;

[Severity: High]
Does this fail to build because vcpu is undeclared in svm_copy_vmrun_state()? The function only takes to_save and from_save, so its signature and callers in svm_set_nested_state(), svm_enter_smm(), and svm_leave_smm() need to pass vcpu as well.

Signature · 1 line
Sashiko AI review · https://sashiko.dev/#/patchset/[email protected]?part=1
#2 lore raw
Quoted 24 lines · On Thu, Nov 12, 2026 at 1:32 PM <[email protected]> wrote:

Thank you for your contribution! Sashiko AI review found 1 potential issue to consider:
- [High] KVM: nSVM: undeclared vcpu in svm_copy_vmrun_state() breaks build
--

commit 11b89023661c173dccf1cc07088879a6b4ff0bef
Author: Jim Mattson <[email protected]>

KVM: nSVM: Gate CET state copy in svm_copy_vmrun_state() on guest SHSTK

This commit changes svm_copy_vmrun_state() to check guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK) instead of kvm_cpu_cap_has(X86_FEATURE_SHSTK) before copying CET shadow stack fields.

Quoted 8 lines
arch/x86/kvm/svm/nested.c +0 -0
index f9090b601efaa..2a1c279bd54a1 100644
@@ -1218,7 +1218,7 @@ void svm_copy_vmrun_state(struct vmcb_save_area *to_save,
      to_save->rip = from_save->rip;
      to_save->cpl = 0;
 
Quoted 11 lines

arch/x86/kvm/svm/nested.c void svm_copy_vmrun_state(struct kvm_vcpu *vcpu, struct vmcb_save_area *to_save, · line 1254

Quoted 5 lines
arch/x86/kvm/svm/nested.c +1 -1
-     if (kvm_cpu_cap_has(X86_FEATURE_SHSTK)) {
+     if (guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK)) {
              to_save->s_cet  = from_save->s_cet;
              to_save->isst_addr = from_save->isst_addr;
              to_save->ssp = from_save->ssp;

[Severity: High]
Does this fail to build because vcpu is undeclared in svm_copy_vmrun_state()? The function only takes to_save and from_save, so its signature and callers in svm_set_nested_state(), svm_enter_smm(), and svm_leave_smm() need to pass vcpu as well.

I should have mentioned that the base-commit is
dea4f6380f4ec96e8f3b0aaf2343764bd0d25ef4.

Commit cd0e6f4d7cec ("KVM: SVM: Add a vCPU-aware helper to get supported DEBUGCTL bits") changed the signature of svm_copy_vmrun_state().

Odd that you quote me as writing:

@@ -1218,7 +1218,7 @@ void svm_copy_vmrun_state(struct vmcb_save_area *to_save,

when, in fact, I wrote:

@@ -1248,7 +1248,7 @@ void svm_copy_vmrun_state(struct kvm_vcpu *vcpu, struct vmcb_save_area *to_save,

#3 lore raw
In reply to Jim Mattson
Quoted 6 lines · Jim Mattson

Like nested_vmcb02_prepare_save() and nested_svm_vmexit_update_vmcb12(), svm_copy_vmrun_state() should gate copying s_cet, isst_addr, and ssp on guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK) (i.e. the feature is advertised to the guest). Instead, copying is gated on kvm_cpu_cap_has(X86_FEATURE_SHSTK) (i.e. the host supports shadow stacks).

This is a bit confusing, nested_vmcb02_prepare_save() and nested_svm_vmexit_update_vmcb12() are about vmcb12 <-> vmcb02, so L2's state, while svm_copy_vmrun_state() is for L1's state. I think the patch is correct, but the examples are confusing.

Perhaps a better example would be __kvm_set_msr() and __kvm_get_msr()?

Quoted 26 lines

Check guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK) instead of kvm_cpu_cap_has(X86_FEATURE_SHSTK) in svm_copy_vmrun_state().

Fixes c5ba49458513 ("KVM: nSVM: Save/load CET Shadow Stack state to/from vmcb12/vmcb02")
---
 arch/x86/kvm/svm/nested.c | 2 +-
 1 file changed, 1 insertion(+), 1 deletion(-)
arch/x86/kvm/svm/nested.c +1 -2
index 1299fdadeb73..b73bc9c3ebdb 100644
@@ -1248,7 +1248,7 @@ void svm_copy_vmrun_state(struct kvm_vcpu *vcpu, struct vmcb_save_area *to_save,
        to_save->rip = from_save->rip;
        to_save->cpl = 0;
 
-       if (kvm_cpu_cap_has(X86_FEATURE_SHSTK)) {
+       if (guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK)) {
                to_save->s_cet  = from_save->s_cet;
                to_save->isst_addr = from_save->isst_addr;
                to_save->ssp = from_save->ssp;
--

2.56.1.703.gf848ed807c-goog

#4 lore raw
Quoted 9 lines · Yosry Ahmed
Quoted 6 lines · Jim Mattson

Like nested_vmcb02_prepare_save() and nested_svm_vmexit_update_vmcb12(), svm_copy_vmrun_state() should gate copying s_cet, isst_addr, and ssp on guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK) (i.e. the feature is advertised to the guest). Instead, copying is gated on kvm_cpu_cap_has(X86_FEATURE_SHSTK) (i.e. the host supports shadow stacks).

Quoted 6 lines

This is a bit confusing, nested_vmcb02_prepare_save() and nested_svm_vmexit_update_vmcb12() are about vmcb12 <-> vmcb02, so L2's state, while svm_copy_vmrun_state() is for L1's state. I think the patch is correct, but the examples are confusing.

Perhaps a better example would be __kvm_set_msr() and __kvm_get_msr()?

Or, perhaps I should just drop the examples?

Reviewed-by
#5 lore raw
Quoted 11 lines · Jim Mattson
Quoted 9 lines · Yosry Ahmed
Quoted 6 lines · Jim Mattson

Like nested_vmcb02_prepare_save() and nested_svm_vmexit_update_vmcb12(), svm_copy_vmrun_state() should gate copying s_cet, isst_addr, and ssp on guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK) (i.e. the feature is advertised to the guest). Instead, copying is gated on kvm_cpu_cap_has(X86_FEATURE_SHSTK) (i.e. the host supports shadow stacks).

Quoted 8 lines
Quoted 6 lines

This is a bit confusing, nested_vmcb02_prepare_save() and nested_svm_vmexit_update_vmcb12() are about vmcb12 <-> vmcb02, so L2's state, while svm_copy_vmrun_state() is for L1's state. I think the patch is correct, but the examples are confusing.

Perhaps a better example would be __kvm_set_msr() and __kvm_get_msr()?

Or, perhaps I should just drop the examples?

It's probably useful to mention that without guest_cpu_cap_has(vcpu, X86_FEATURE_SHSTK), the MSRs cannot be set, but I am fine either way. With either changes to the changelog: