In the Linux kernel, the following vulnerability has been resolved: KVM: arm64: Do not clear VM-wide SVE feature on vCPU init failure… (CVE-2026-98637)

The vulnerability in the Linux kernel KVM arm64 code occurs because the function pkvm_vcpu_init_sve() clears the KVM_ARM_VCPU_SVE flag in kvm->arch.vcpu_features when it fails, but the vcpu_has_sve() function checks KVM_ARCH_FLAG_GUEST_HAS_SVE, which remains set. This mismatch causes later virtual CPUs on the same VM to skip SVE setup and register with a NULL sve_state. When the guest performs its first floating-point operation, the NULL sve_state is passed to sve_load_state(), potentially causing a fault. The patch corrects this by returning the error without altering vcpu_features, preventing inconsistent state.

The vulnerability can lead to improper handling of the SVE feature state in virtual CPUs, potentially causing a NULL pointer dereference or crash during guest floating-point operations. This could affect the stability and reliability of virtual machines using KVM on arm64 architectures with SVE enabled. There is no indication of code execution or privilege escalation from the provided data.

A fix has been implemented in the Linux kernel to address this issue by ensuring that on vCPU initialization failure, the VM-wide SVE feature flags are not cleared improperly. Users should update to the fixed Linux kernel version once available. Patch status is not explicitly confirmed in the provided data; therefore, check the official Linux kernel advisories for the specific patch and remediation guidance.