NoisyTCP: Unpatched Linux Kernel Race Condition Enables Remote Server Crashes
Written by AI from the linked story. Not reviewed by an editor. How this works
A race condition in the Linux kernel's TCP stack (CVE-2026-99288, 'NoisyTCP') allows any remote attacker to crash a server by sending specially crafted packets to any open port—no authentication required. The flaw has existed since Linux kernel 4.4, meaning systems that have not been kept current are exposed to a vulnerability with years of undetected presence. Because it targets the TCP layer itself, virtually every internet-facing Linux server or cloud VM running an unpatched kernel is at risk of denial-of-service. Patches are available in upstream kernel releases 6.12.117, 6.18.63, and 7.3.4, making prompt patching the most critical remediation step. The incident underscores how foundational OS-layer vulnerabilities can have sweeping impact across entire infrastructure estates if patch cycles are slow or inconsistent.
Tactical Insight
Immediate actions
- Apply the upstream Linux kernel patches (versions 6.12.117, 6.18.63, or 7.3.4) or your distribution's backported fix as an emergency change.
- Audit all internet-facing Linux servers and cloud VMs to identify unpatched kernel versions using an authenticated vulnerability scanner.
- Apply network-level mitigations (e.g., TCP rate limiting, firewall rules restricting unexpected source IPs) as a temporary compensating control until patching is complete.
Long-term improvements
- Establish an emergency/out-of-band patching procedure with defined SLAs for critical kernel-level CVEs, separate from routine patch cycles.
- Maintain a continuously updated inventory of all Linux systems, including kernel versions, so scope of exposure can be determined within hours of a disclosure.
- Subscribe to kernel security mailing lists and vendor advisories to receive patch notifications the moment they are released.
Detection measures
- Deploy host-based monitoring to alert on unexpected kernel panics, server reboots, or TCP stack anomalies that may indicate active exploitation.
- Implement network traffic analysis to detect abnormal TCP packet patterns consistent with denial-of-service attempts targeting listening ports.