Re: [PATCH v6 12/12] KVM: SVM: Add support for virtualizating Bus Lock Detect

[sashiko-bot@kernel.org]

Thank you for your contribution! Sashiko AI review found 1 potential issue(s) to consider:
- [Medium] kvm: svm: missing host_initiated check in MSR_IA32_DEBUGCTLMSR restoration
--

commit 6a801e9cb8e116e2aead8b9c8e78497f605d7a2c
Author: Ravi Bangoria <ravi.bangoria@amd.com>

KVM: SVM: Add support for virtualizating Bus Lock Detect

This commit advertises support for virtualizing Bus Lock Detect on SVM.
It allows the guest to enable BUS_LOCK_DETECT via MSR_IA32_DEBUGCTLMSR
if LBR Virtualization is enabled.

> diff --git a/arch/x86/kvm/svm/svm.h b/arch/x86/kvm/svm/svm.h
> index c3380d865b751..6d0242966dc87 100644
> --- a/arch/x86/kvm/svm/svm.h
> +++ b/arch/x86/kvm/svm/svm.h
[ ... ]
> @@ -789,7 +789,12 @@ BUILD_SVM_MSR_BITMAP_HELPERS(void, set, __set)
>  
>  static inline u64 svm_get_supported_debugctl(struct kvm_vcpu *vcpu)
>  {
> -	return DEBUGCTLMSR_LBR;
> +	u64 debugctl = DEBUGCTLMSR_LBR;
> +
> +	if (guest_cpu_cap_has(vcpu, X86_FEATURE_BUS_LOCK_DETECT))
> +		debugctl |= DEBUGCTLMSR_BUS_LOCK_DETECT;
> +

[Severity: Medium]
Does this restrict host-initiated MSR restoration? By making the
DEBUGCTLMSR_BUS_LOCK_DETECT bit strictly dependent on guest_cpu_cap_has(),
this might reject host-initiated writes to MSR_IA32_DEBUGCTLMSR if the VMM
restores MSRs before initializing the guest CPUID.

In svm_set_msr(), the rejection happens unconditionally:

svm_set_msr()
    if (data & ~svm_get_supported_debugctl(vcpu))
        return 1;

Since svm_set_msr() does not appear to check msr->host_initiated before
enforcing this capability limit, could this regression break live
migration when userspace sends KVM_SET_MSRS before the CPUID is fully
configured?

> +	return debugctl;
>  }

-- 
Sashiko AI review · https://sashiko.dev/#/patchset/20260930193634.3415089-1-seanjc@google.com?part=12