The vulnerability originates from the Linux kernel's AMDGPU driver, where a recursive acquisition of ww_mutex locks in the function amdgpu_devcoredump_format can lead to a self‑deadlock. This represents an instance of improper locking (CWE‑667). The self‑deadlock blocks the kernel indefinitely, causing kernel hangs and flooding the system log with lockdep warnings. The bug does not directly grant code execution but can bring the system to a non‑responsive state, potentially disconnecting users and interrupting GPU‑dependent workloads.

All Linux kernel systems that use the AMDGPU driver and contain the lock‑acquisition logic before the commit that introduced the fix (identified by commit d6bf4242). Versions of the kernel where this code existed—prior to the patch—are affected, regardless of distribution or patch level. Exact version numbers are not specified in the provided data.

The issue resides at the kernel level, indicating high severity. While no publicly available exploitation path is described, the flaw can be triggered by operations involving GPU buffer dumps during hung jobs. Attackers with local kernel or GPU‑related privileges could intentionally force a deadlock, leading to denial of service. The absence of CVSS or EPSS scores makes precise quantification difficult, but the kernel‑wide impact and potential for system freeze recommend treating it as a critical risk.