In the Linux kernel, the vt‑d subsystem removes an IOMMU context entry by writing two 64‑bit halves of a 128‑bit entry without first clearing the Present bit. This race can allow the hardware to fetch a partially cleared entry where some fields are zeroed but the Present bit remains set. The result is unpredictable IOMMU behavior or spurious faults, a clear instance of a race condition (CWE‑362, CWE‑367).

All x86 Linux kernel builds that enable IOMMU and use the vt‑d context‑teardown logic are affected. The advisory shows no explicit version range; therefore, any kernel predating the commit that adds the dma_wmb() ordering and clears the Present bit first may be vulnerable. Verify that your kernel version contains the change referenced in the patch series linked in the CVE. No explicit affected version information is provided in the payload, so the user must determine vulnerability status by inspecting your kernel version against the patch commit.

The CVSS score of 7.5 indicates high severity. An EPSS score of under 1 % shows a very low probability of exploitation. The vulnerability is not listed in CISA KEV, so no active exploitation is known. Based on the description, it is inferred that the likely attack vector would require an attacker with physical access to a DMA device or control over a device that targets the IOMMU, and would be a local or privilege‑escalation attack that could compromise system reliability. The risk assessment is derived solely from inferred information.