CVE-2026-47565

A flaw was found in the NVIDIA GPU display driver. A privileged local user could trigger a race condition in the kernel mode layer, resulting in an out-of-bounds memory write. A successful exploit may allow the attacker to execute arbitrary code, escalate privileges, cause a denial of service (DoS), or tamper with sensitive data.

This vulnerability is classified as Moderate severity for Red Hat Enterprise Linux because exploitation requires a local user to already possess elevated administrative privileges and successfully win a complex race condition. The proprietary NVIDIA GPU display driver is an optional package not enabled or installed by default, affecting only systems specifically deployed with supported NVIDIA hardware accelerators. Although a successful exploit could allow a privileged account to trigger out-of-bounds memory access in kernel space to compromise system integrity or cause a denial of service, the high privilege requirement significantly constrains the threat landscape.

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.

To mitigate this issue on systems where NVIDIA GPU acceleration is not strictly required, prevent the NVIDIA kernel module from loading by blacklisting it in the modprobe configuration.

1. Create a configuration file to blacklist the module and prevent execution:

echo "blacklist nvidia" > /etc/modprobe.d/blacklist-nvidia.conf

echo "install nvidia /bin/false" >> /etc/modprobe.d/blacklist-nvidia.conf

2. If the module is currently loaded, unload it or restart the system:

modprobe -r nvidia

Warning: A system reboot or service restart is required to ensure the module is fully unloaded if active processes are using the device. Applying this workaround disables all NVIDIA GPU capabilities, including hardware-accelerated display output and CUDA compute workloads.