CVE-2024-26672

What it actually does

Pointer to mca_funcs is dereferenced prior to any NULL validation inside amdgpu_mca_smu_get_mca_entry, enabling a null pointer dereference that can corrupt kernel memory or trigger a panic on AMD GPUs.

Why it matters to the business

AMD GPU systems may suffer unexpected crashes or stability loss, forcing unplanned reboots and kernel updates that interrupt workloads and increase maintenance overhead.

How an attack unfolds

An attacker with local access invokes the MCA error entry function with UE or CE types, hitting the unconditional dereference of a potentially null mca_funcs structure before the late guard clause executes.

How to fix it

Apply the upstream fix that relocates the NULL check ahead of any dereference; upgrade to a patched kernel. For the listed Chromium packages, install the referenced fixed builds such as 87.0.4280.88-0.1 on Debian or 85.0.4183.83-0ubuntu0.16.04.2 on Ubuntu.

How to confirm the fix worked

Rebuild or boot the updated kernel, then exercise amdgpu_mca_smu_get_mca_entry paths while confirming the absence of the dereference-before-check warning in static analysis or runtime logs.

Typical attack path

1. Obtain local execution context on an affected Linux system
2. Trigger AMDGPU MCA error path via hardware fault or crafted input
3. Force null mca_funcs dereference leading to kernel panic or corruption

How to detect it

Review kernel logs and dmesg for amdgpu_mca.c line references or null pointer dereference faults around MCA entry functions; correlate with process activity that exercises the AMDGPU MCA interface.

What your security team should do

• Prioritize kernel updates on all AMD GPU hosts within standard change windows
• Add static analysis rules to catch similar dereference-before-check patterns
• Monitor kernel oops and panic logs for amdgpu_mca related faults