The Linux kernel contains a defect in the IOMMU/VT‑d subsystem that prevents proper handling of PASID attachment for a nested domain when dirty tracking is enabled. The kernel prematurely rejects the attachment, leaving any dirty pages in the nested domain unwiped, which can result in lost or corrupted data during device passthrough operations. This flaw does not provide remote code execution or privilege escalation but does compromise data integrity for virtualized workloads that rely on nested IOMMU domains.

The issue is present in all Linux kernels that expose the IOMMU/VT‑d subsystem for PASID attachment in nested virtualization setups. No specific version range is given in the advisory, so any affected kernel build that does not yet incorporate the patch should be considered at risk. The affected product is the Linux kernel itself.

With a CVSS score of 5.5 and an EPSS score of less than 1%, the vulnerability presents a moderate risk rating and is unlikely to see widespread exploitation at present. The vulnerability is not listed in CISA’s KEV catalog. Exploitation would require a system that is both running a vulnerable kernel and utilizing nested IOMMU domains with dirty tracking. Since the defect triggers a kernel error and blocks attachment, the attack surface is limited to privileged operations that manage device assignments; it therefore does not enable arbitrary code execution or compromise beyond data loss. The likely attack vector is inferred to be internal privileged interactions with the IOMMU subsystem, as the flaw manifests during the PASID attachment process.