CVE-2026-89482

A flaw was found in the nvme-tcp module of the Linux kernel. This vulnerability arises from improper handling of C2HData for REQ_OP_WRITE_ZEROES commands, where the system fails to adequately validate the data length. A remote attacker could exploit this by sending specially crafted C2HData, leading to a wild-memory-access. This memory corruption can cause a kernel crash, resulting in a Denial of Service (DoS) for the affected system.

A wild memory write in the NVMe/TCP host receive path can occur because C2HData validation used blk_rq_payload_bytes() without also checking the driver private request state. For WRITE_ZEROES, a request can have non zero bytes but no physical segments, leaving the current receive iterator uninitialized while stale iterator state from an earlier command on the same tag may still exist. A malicious or compromised NVMe/TCP target can send C2HData for this command state and cause nvme_tcp_recv_data to copy attacker supplied data into the stale iterator destination, resulting in memory corruption or a kernel crash. For the CVSS the PR:N is used because the attacker does not need local privileges on the victim host when acting as the remote NVMe/TCP target already used by the host. AV:A is used because NVMe/TCP is normally deployed on storage or data center networks rather than the public Internet. Impact is at least denial of service and data corruption. For the paranoid score, C:H/I:H/A:H is defensible because the bug is a kernel memory write through stale iterator state and should not be treated as DoS only.

The CVSS score(s) listed for this vulnerability do not reflect the associated product's status, and are included for informational purposes to better understand the severity of this vulnerability.

Mitigation for this issue is either not available or the currently available options don't meet the Red Hat Product Security criteria comprising ease of use and deployment, applicability to widespread installation base or stability.