Linux KVM Flaw CVE-2026-99312 Lets Guest VMs Access Host Memory on AMD Hosts
CVE-2026-99312 in Linux KVM allows guest VMs to read host memory on AMD systems with nested virtualization.
A flaw in Linux KVM (CVE-2026-99312) enables guest virtual machines to access up to 127 GB of host memory on AMD processors with nested virtualization enabled, potentially exposing other customers' VMs on shared cloud servers. The vulnerability stems from an incomplete fix for CVE-2022-3344 that left a pointer referencing freed memory, which attackers with root access inside a guest VM could exploit to forge memory table entries. Fixed kernels shipped upstream on November 18, 2026, and DigitalOcean patched its infrastructure; operators running KVM on AMD hardware should update kernels or disable nested virtualization.
AI summary of the linked source. Not reviewed by an editor. How this works
Run DigitalOcean?
Get an email when a new story names DigitalOcean, usually within the hour.
Free. Your list stays private and never appears in a subject line. One click stops it. How Vendor Watch worksPrivacy
The lesson from this incident
Full lessonAn incomplete fix for a 2022 Linux KVM vulnerability left a dangling pointer that attackers with root access inside a guest VM could exploit to forge memory table entries and read up to 127 GB of host memory on AMD processors with nested virtualization enabled. This is a classic case of a partial patch introducing a residual risk that persists for years before being discovered and fully remediated. In multi-tenant cloud environments, the blast radius is especially severe because one compromised or malicious tenant could access memory belonging to other customers. The incident underscores that patch quality and regression testing are as important as patch speed, and that hypervisor-level vulnerabilities demand priority treatment given the privileged boundary they protect.
What helps against this
How we list vendors- Vulnerability scanning and prioritisation
You know which systems are vulnerable, and which ones to fix first.
- Automated patch management
Known holes get closed in days, not months.
- Network segmentation and microsegmentation
One compromised machine can no longer reach everything else.
Indicators of Compromise
Addresses and file hashes are copied from the publisher's page by pattern matching. Names are picked by AI. On stories from before 17 November 2026, some addresses and hashes were picked by AI too. Not reviewed by an editor. How this works
- cve: CVE-2022-3344
- cve: CVE-2026-99312