CVE-2026-98716
NoneCVSS
—
EPSS
0.2%
Published
Oct 18, 2026
CISA due date
—
Cloudanix analysis
A bug in the Linux kernel used for running virtual machines on AMD processors could let one virtual machine improperly access hardware settings belonging to others.
- Exploitation
- Local Privilege
- Attacker needs
- Low Privilege
- Complexity
- Moderate
- Impact
- Privilege Gain
- Component
- OS kernel
- Class
- Authorization Bypass
- Remediation
- Upgrade
- Effort
- Medium
- Patch risk
- Low
What it actually does
Incorrect MSR permission bitmap selection in KVM SVM nested virtualization allows L2 guests to access MSRs that should be intercepted when L1 configures selective interception, bypassing the intended authorization checks on MSR accesses.
Why it matters to the business
Nested virtualization environments on AMD processors may experience unauthorized register access by guest workloads, increasing risk of VM escape or privilege escalation within virtualized infrastructure and requiring urgent kernel updates to maintain isolation guarantees.
How an attack unfolds
An attacker with control of an L2 guest configures MSR access patterns that exploit the mismatch between the active VMCB bitmap and the assumed vmcb02 bitmap; when L1 enables selective MSR interception, the wrong bitmap is consulted, permitting the L2 guest to read or write protected MSRs that should trigger exits to L0.
How to fix it
Apply the upstream Linux kernel patch that selects the active VMCB's MSR bitmap for interception checks; rebuild and deploy updated kernels to all affected hosts running KVM on AMD processors.
How to confirm the fix worked
Confirm the fix by inspecting the KVM SVM MSR bitmap selection logic in the running kernel source or binary for use of the active VMCB pointer instead of a hardcoded vmcb02 reference; test nested guest MSR interception behavior under L1 selective interception.
Typical attack path
- Compromise or control an L2 nested guest on an AMD SVM host
- Trigger L1 hypervisor to enable selective MSR interception via msrpm01/msrpm12 merge
- Issue MSR read/write operations that rely on the incorrect active bitmap
- Bypass interception and access privileged MSRs directly from L2
- Escalate to impact L0 host or other guests via MSR side effects
How to detect it
Search hypervisor and guest kernel logs for anomalous MSR access patterns or missing VM exits during nested virtualization; examine KVM SVM code paths handling L1/L2 MSR bitmap merges for signs of prior exploitation via unexpected guest behavior or crashes.
What your security team should do
- Prioritize kernel updates on all AMD-based KVM hosts within the next maintenance window
- Audit nested virtualization configurations and disable if not required
- Monitor for new kernel releases containing the described MSR bitmap correction
- Review hypervisor logs for unexpected MSR exit patterns after patching
The record
In the Linux kernel, the following vulnerability has been resolved: KVM: SVM: Use the active VMCB's MSR bitmap when checking if MSR is intercepted Use the MSR permission bitmap of the active VMCB instead of assuming that KVM is always using vmcb02's bitmap when L2 is active, as KVM uses msrpm02 if and only if L1 wants to intercept MSR accesses, i.e. if and only if KVM needs to merge msprm01 with msrpm12. Don't bother tracking the virtual address of the bitmap that's being used, as __va() is cheap on x86, and caching the virtual address would introduce yet another source of potentially stale information.
Affected packages
| Ecosystem | Package | Introduced | Fixed |
|---|