CVE-2026-47531

A flaw was found in the NVIDIA GPU Display Driver. A local attacker with high privileges can trigger an invalid memory access (null pointer dereference) within the kernel mode layer. Successful exploitation leads to a Denial of Service (DoS) by causing the driver or host system to crash.

This vulnerability is rated as Moderate severity for Red Hat Enterprise Linux because triggering it requires local access with elevated administrative privileges, restricting the impact to a host denial of service without compromising confidentiality or integrity. The proprietary NVIDIA display driver is not included in default minimal installations and is only present on systems provisioned with specialized graphics hardware. Furthermore, because high privileges are a prerequisite to induce the kernel crash, the practical risk is constrained since users possessing elevated permissions already wield extensive control over the operating system.

The CVSS score(s) listed for this vulnerability do not reflect the associated product's status, and are included for informational purposes to better understand the severity of this vulnerability.

Prevent the NVIDIA kernel module from loading on systems where proprietary GPU acceleration is not required.

1. Create a modprobe configuration file to disable module loading:

echo "install nvidia /bin/true" > /etc/modprobe.d/disable-nvidia.conf

2. If the kernel module is currently loaded, a system reboot is required:

reboot

Warning: Rebooting the host will interrupt running workloads and terminate active user sessions. Applying this configuration prevents the NVIDIA graphics driver from initializing, which disables hardware-accelerated 3D rendering, graphical output via the discrete GPU, and CUDA compute workloads. Where the driver is required, restrict read and write permissions on /dev/nvidia* device nodes to trusted administrative users.