Skip to content
[ SIGNAL VERIFIED ] INDEPENDENT CYBERCRIME JOURNALISMCLASSIFIED // THE UNDERGROUND HAS A NEWSROOM / RSS
.COM
Latest News
ALL NEWS
Newsroom

NoisyTCP Linux Kernel Flaw CVE-2026-99288 Allows Remote Server Crashes Before Login

The NoisyTCP bug, disclosed by Hyunwoo Kim, can cause a Linux kernel crash through reachable TCP services. A kernel patch exists; exposure and update status vary by distribution.

Newsroom editorial artwork
PUBLISHED

IPCRIME NEWSROOM | NOVEMBER 16, 2026

New disclosure, old underlying bug

Security researcher Hyunwoo Kim publicly disclosed NoisyTCP, tracked as CVE-2026-99288, in an oss-security advisory dated November 16 in South Korea. The race-condition flaw has existed in Linux kernel code since October 2015, according to the researcher. The defect sits in TCP socket handling, which means public-facing services can expose an affected kernel even when the application itself is correctly configured.

What the impact is

Kim describes a pre-authentication remote denial-of-service condition under specific concurrent packet-processing circumstances. An attack may cause a kernel panic and take a host offline. In a containerized environment the relevant kernel belongs to the host, so the consequences can extend beyond a single application container. The available reporting does not establish that this vulnerability gives an attacker arbitrary code execution or access to private files.

Where exposure depends on configuration

The published analysis identifies multiqueue network interfaces or Receive Packet Steering configurations as important prerequisites to the race. Common TCP listeners such as web servers or SSH can be exposed where the underlying kernel is affected. Reproducibility varies with network and processor scheduling; the researcher’s tests included substantial differences between virtual machines and quiet physical servers.

Patch first, then validate deployed kernels

Kim identifies kernel commit 4e2b3ef19e0e, merged November 3, as the correction. Administrators should obtain fixed kernels through the support channel for their distribution or provider, not assume that installing a newer user-space web server addresses a kernel defect. Check the actual running kernel after the required reboot, including cloud nodes and container hosts. A downstream vendor may backport the fix without changing a headline kernel version.

Treat temporary mitigations as trade-offs

The advisories discuss temporarily altering SYN-cookie or receive-queue behavior, but those changes affect resilience against other network attacks or performance. They should be evaluated by the systems/network team and rolled back after supported patches are installed. As of the sources reviewed, the public advisory documents the vulnerability and patch; it is not evidence that a widespread exploitation campaign is already underway.

Sources and documentation

IPCRIME distinguishes primary statements, independently reported details and unresolved findings. Publication here reflects the November 16 research review, not necessarily the original date of each underlying event.

Editorial note

IPCRIME distinguishes public records and documented findings from allegations, analysis and commentary. Report corrections through our contact page.