Google Cloud Warns Three AMD Turin CPU Flaws Could Leak Data Across Processes in VMs
Summary
Google Cloud has identified three CPU side-channel flaws affecting AMD Turin chips within its virtual machines. While mitigations are in place to prevent cross-VM and VM-to-host data leakage, an attacker with code on the same processor core could potentially infer data from other processes or the kernel within a VM. A guest OS patch and a stop-and-start of the VM are recommended for protection.
IFF Assessment
The identified CPU flaws allow for potential data inference between processes within a virtual machine, posing a risk to sensitive information for defenders.
Severity
This estimated CVSS score reflects a medium-high severity due to the potential for unauthorized information disclosure (Confidentiality Impact: High) from other processes or the kernel within a VM, although it requires local code execution and sharing a processor core (Attack Vector: Network, but realistically Local, Attack Complexity: Low, Privileges Required: Low, User Interaction: None).
Defender Context
This vulnerability highlights the ongoing risks associated with hardware-level side-channel attacks, even in cloud environments. Defenders need to stay vigilant about patching and applying recommended mitigations for CPU-related flaws to protect data within virtual machines.