CVE-2022-42254

What it actually does

Out-of-bounds array access in the nvidia.ko kernel module allows local attackers to trigger denial of service, memory corruption for data tampering, or unauthorized information disclosure.

Why it matters to the business

Systems using NVIDIA GPUs in Linux, virtualization platforms, and cloud gaming environments face risk of crashes, corrupted workloads, or data leaks; widespread downstream use in enterprise hypervisors increases operational disruption potential despite low exploit probability.

How an attack unfolds

A local attacker with low privileges crafts inputs that reach the vulnerable array access path inside the kernel-mode NVIDIA driver, resulting in out-of-bounds reads or writes that can crash the system, alter data, or leak memory contents.

How to fix it

Upgrade the NVIDIA GPU Display Driver to the latest available version for each affected product line; apply vendor-supplied kernel module updates on Linux distributions, Citrix Hypervisor, Red Hat Enterprise Linux KVM, and VMware vSphere hosts; verify driver version after installation and reboot.

How to confirm the fix worked

Compare installed NVIDIA driver version against the fixed releases published by NVIDIA; run kernel module signature and version checks; attempt to load the module and exercise GPU workloads while monitoring for oops or panic messages referencing nvidia.ko.

Typical attack path

1. Obtain local shell or process execution on a Linux host with the vulnerable NVIDIA GPU driver loaded
2. Invoke GPU-accelerated operations or ioctls that exercise the nvidia.ko kernel module
3. Supply malformed parameters that trigger the out-of-bounds array access
4. Observe or leverage resulting kernel panic, memory corruption, or disclosed data

How to detect it

Review kernel logs (dmesg, /var/log/kern.log, journalctl) for oops or panic entries mentioning nvidia.ko or array bounds violations; correlate with process crashes involving GPU-accelerated applications; inspect for unexpected driver reloads or system reboots following GPU activity.

• nvidia.ko.*out-of-bounds
• kernel: BUG: unable to handle kernel paging request at.*nvidia
• nvidia.*general protection fault

What your security team should do

• Prioritize patching NVIDIA GPU drivers across all Linux-based virtualization and compute hosts within the next maintenance window
• Inventory all affected NVIDIA product deployments including GeForce, Quadro, Tesla, and virtual GPU offerings
• Monitor NVIDIA security bulletins for driver updates and subscribe to vendor advisories
• Restrict untrusted local code execution on systems hosting the driver until patched