GPU VulnDB

Database/NVIDIA / GPU stack

Linux kernel drm/i915: HDCP stream array bounds check ran after the overflowing write

CVE-2026-68253NVIDIA / GPU stackcurated

Impact

In the i915 HDCP path the overflow check on data->streams[] was performed after the write that overflows it, so the out-of-bounds write had already happened by the time the warning fired. The fix moves the check ahead of the write. Exposure is narrow: this is the display content-protection path, which requires a local user on a machine driving HDCP-capable outputs through i915. On a headless GPU compute node there is no display pipe running HDCP and this code is not reached, so for most datacenter fleets this is a routine kernel update rather than a maintenance-window event. It matters more on i915-backed workstations or remote-desktop hosts that actually light up displays. The issue was found by AI-assisted static analysis and confirmed by Intel Product Security.

Who can reach it

A local user on a host with an i915 display device active on an HDCP-capable output. Not reachable over the network and not reachable on a headless server configuration.

What to do

Pick up the fix from the stable trees listed in the record and reboot into the new kernel; there is no module-reload path that avoids a reboot for a live GPU driver. Because the exposure is display-only, most datacenter operators can fold this into the next scheduled kernel roll rather than draining nodes for it. No CVSS-relevant mitigation is offered beyond the patch.

References

Related entries

All NVIDIA / GPU stack entries

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.