Linux drm/amdgpu: UVD H.264 minimum dpb size derived from level rather than the decode message
Impact
The amdgpu UVD path computed the minimum decode picture buffer size for H.264 from the maximum implied by the stream level instead of the reference count actually carried in the decode message. The upstream commit is a correction to that size calculation; the record shows no demonstrated memory corruption, no CVSS score and no CWE, and the fix is a cherry-pick rather than a security-flagged change. Exposure is limited to the video decode engine, which datacenter compute workloads do not touch - it matters on nodes doing GPU transcoding with AMD parts that expose UVD, not on Instinct compute nodes running training or inference.
Who can reach it
A local user holding an open render node (/dev/dri/renderD*) who can submit a UVD decode job. No remote path and no privileged access; on a headless compute node with no video pipeline this code is not reached.
What to do
Pick up a stable kernel containing the fix (the commits are listed on the kernel.org stable links) and reboot the affected nodes on the next routine maintenance window. Given the absence of a score or a described memory-safety consequence, there is no case for an emergency drain.
References
Related entries
- Linux kernel amdgpu: unvalidated UVD decode message fields overflow the DPB size calculationCVE-2026-80908 · Linux kernel amdgpu UVD (decode message field validation)Unscored
- GPU / accelerator firmware (VBIOS, GSP, NVSwitch): GPU-resident firmware sits below the host OS and is not coveredNCVD-0000-012-gpu-accelerator-firmware-vbios-g · GPU / accelerator firmware (VBIOS, GSP, NVSwitch)Unscored
- NVIDIA Multi-Instance GPU (MIG) partitioning: MIG gives each instance its own SM slice, L2 slice, memory slice andNCVD-2020-001-nvidia-multi-instance-gpu-mig-pa · NVIDIA Multi-Instance GPU (MIG) partitioningUnscored
- NVIDIA Multi-Instance GPU (MIG) partitioning: MIG gives each instance its own SM slice, L2 slice, memory slice andNCVD-2020-003-nvidia-multi-instance-gpu-mig-pa · NVIDIA Multi-Instance GPU (MIG) partitioningUnscored
- Integrated GPU graphics data compression (Intel, AMD, Apple, Arm, Qualcomm, NVIDIA): GPUs apply data-dependent losslessNCVD-2023-003-integrated-gpu-graphics-data-com · Integrated GPU graphics data compression (Intel, AMD, Apple, Arm, Qualcomm, NVIDIA)Unscored
- NVIDIA Confidential Computing (H100/H200/B100/B200/GB200) - CC-DevTools operating mode: NVIDIA GPU confidentialNCVD-2023-004-nvidia-confidential-computing-h1 · NVIDIA Confidential Computing (H100/H200/B100/B200/GB200) - CC-DevTools operating modeUnscored
This entry is curated: imported from vendor advisories with machine assistance, not yet individually verified. Confirm against your vendor's advisory before acting, and report anything wrong.