The QCE cryptographic driver in the Linux kernel allocates a buffer for the CCM AAD (associated authenticated data) that is smaller than the length used by the DMA scatterlist. This mis‑calculation means that during authentication the QCE hardware can read past the end of the allocated buffer, causing forged or corrupted authentication tags. A faulty tag can allow malicious data to be accepted as authenticated, effectively bypassing message integrity checks.

Linux kernel versions that include the QCE crypto driver before the applied patch are affected. The issue is present in the kernel’s qce_aead_ccm_prepare_buf_assoclen() function which handles CCM mode for AES. No specific version numbers are listed in the CNA data, but any kernel using the unpatched QCE implementation is vulnerable.

The vulnerability can be exploited by any process that can invoke CCM mode with a custom AAD length, which may include legitimate applications or malicious local code. Because the flaw lies in kernel driver memory handling, it requires userland control of cryptographic operations but does not provide network‑exposed entry points. The exploit would lead to authentication forgery rather than code execution. Exploitability is therefore limited to local privilege escalation or data integrity compromise. No CVSS or EPSS scores are supplied, but the absence of a KEV listing suggests that widespread exploitation is not yet confirmed. Nonetheless, given the potential for message forgery, the risk is not negligible.