NVIDIA GPU Display Driver Kernel Mode Layer Contains Denial of Service and Information Disclosure Vulnerability
Overview
NVIDIA released a security bulletin addressing multiple vulnerabilities in their GPU Display Driver for both Windows and Linux. One of the most severe flaws, CVE-2026-0091, exists in the kernel mode driver component responsible for managing GPU memory and scheduling. The vulnerability is an out-of-bounds write condition that can be triggered by a user-mode application sending specially crafted data to the driver through the IOCTL interface. An unprivileged local attacker could exploit this flaw to write data outside of the intended buffer. This can corrupt kernel memory structures, leading to a system crash and a denial of service (DoS). Furthermore, in specific scenarios, the vulnerability could be chained with other techniques to achieve arbitrary code execution in the kernel, resulting in a full privilege escalation from a low-integrity user to SYSTEM/root. This poses a significant risk to multi-tenant AI/ML infrastructure, such as cloud-based GPU instances or on-premise Kubernetes clusters, where multiple users share the same physical GPU. A malicious user or a compromised container could exploit this vulnerability to disrupt service for all tenants on the node or potentially break out of their container to gain control of the host system. The issue affects a wide range of GeForce, Quadro, and Tesla series GPUs.
Affected Systems
Testing Guide
1. Check your current NVIDIA driver version. On Linux, run `nvidia-smi`. On Windows, open the NVIDIA Control Panel and check the System Information page. 2. Compare your installed driver version against the patched versions listed in the NVIDIA security bulletin (e.g., 565.30 for Windows, 565.25.02 for Linux). 3. If your version is lower, your system is affected and should be updated.
Mitigation Steps
1. Update NVIDIA drivers to the latest version available from the NVIDIA driver downloads page or through your cloud provider's marketplace image. 2. For enterprise environments, apply the security updates provided in the latest NVIDIA driver branches. 3. In multi-tenant environments, consider using stronger isolation technologies like gVisor or Kata Containers for GPU workloads, although driver-level vulnerabilities may still pose a risk. 4. Restrict user access to GPU nodes to only trusted individuals and processes.
Patch Details
Patches are available in R565 drivers (versions 565.30 and 565.25.02 for Windows and Linux respectively) and subsequent releases.