The kernel GPU driver where the shrinker walk over non‑system memory does not take a runtime power‑management reference. When the GPU is runtime‑suspended, shrinking a buffer object in non‑system memory invalidates GPU mappings, causing the page‑table zap to trigger an assertion and return -ENODEV. This can lead to kernel warnings, possible false page faults, and in the worst case a service disruption or device failure, effectively denying correct operation of the GPU subsystem.

The issue is present in the Linux kernel; all distributions using the affected kernel versions inherit the defect. No specific kernel release range is listed, so any kernel build that includes the xe_shrinker implementation is potentially impacted.

The vulnerability is not listed in CISA's KEV catalog and has an EPSS score of <1%, indicating very low but non‑zero exploitation probability. The CVSS score of 5.5 indicates moderate severity. It is a local problem that requires the target device to be in a runtime‑suspended state and an operation that triggers memory shrinkage, such as certain user‑space GPU allocations or igt test scenarios. Based on the description, it is inferred that the likely attack vector is local privileged access. An attacker with escalated privileges on the system could deliberately trigger the bug to cause a denial of service or destabilize the GPU driver.