NoisyTCP (CVE-2026-99288): 11-year-old Linux kernel TCP race lets remote unauthenticated attackers panic servers; fixed in 6.12.117, 6.18.63, 7.3.4
Matéria consolidada. Reúne 2 matérias que o Tifon tinha publicado separadamente sobre a mesma história. A IA reescreveu o relatório com o que as 4 fontes trouxeram em 16/11/2026 17:00; os links antigos trazem para cá.
Fontes: theregister.com ↗ bleepingcomputer.com ↗ gbhackers.com ↗ securityonline.info ↗
NoisyTCP (CVE-2026-99288) is a race condition in the Linux kernel TCP established-connection lookup, found and disclosed by researcher Hyunwoo Kim (@v4bel). It dates to commit 079096f103fa of October 3, 2015 (gbhackers.com). A remote, unauthenticated attacker can send crafted TCP traffic to trigger a NULL pointer dereference and kernel panic, causing denial of service on hosts running default Nginx, Apache or SSH listeners. A second path could lead to a use-after-free, but it is described as hard to trigger. Kim has built a limited lab exploit against VMs but has not demonstrated general-purpose RCE (theregister.com). No exploitation in the wild is reported, the flaw is not in CISA KEV, and no public exploit code exists (securityonline.info, bleepingcomputer.com). The upstream fix, commit 4e2b3ef19e0e, landed on November 3, 2026, but as of November 15 many distributions had not shipped patched kernels (gbhackers.com). Red Hat rates the flaw Important with a CVSS score of 7.5 (theregister.com). The sources are press articles without primary-source research.
Entenda a falha
A vulnerabilidade
CVE-2026-99288 (NoisyTCP) is a race condition in the Linux kernel TCP stack, in the lockless lookup of established connections. A remote, unauthenticated attacker who can reach an ordinary TCP listener (default Nginx, Apache or SSH) can trigger a NULL pointer dereference and a kernel panic, taking down the whole host, including a container host. The target needs two or more RX queues, or RPS spreading packets across CPUs. A second path could lead to a use-after-free but is hard to trigger. Affected: kernel 4.4 and later before 6.12.117, and the 6.18 series before 6.18.63. Fixed in 6.12.117, 6.18.63, 7.3.4 and 7.4-rc2. No CVSS score is given in the sources.
Como funciona
The kernel looks up established sockets in the ehash without locks, under RCU. One CPU can hold a stale lookup result while another RX context frees a request socket and reallocates the same slab address as a SYN cookie request with the same tuple. The stale lookup then reaches the new object, which passes the tuple recheck. That request has no listener attached, so the TCP receive path dereferences a NULL listener and panics. The attacker spreads packet processing across CPUs by using the different receive hash of IPv4 fragments and regular TCP packets. In the second scenario a full child socket is recycled, which needs nested races.
Por que é vulnerável
Socket memory can be freed and reused before an RCU reader that kept the old address has finished. The kernel rechecks the connection tuple after the lookup but never checks that the socket it found is actually hashed. The bug was introduced by a 2015 commit and stayed for about 11 years. The articles do not say why it went unnoticed. The fix makes established lookups reject unhashed sockets.
Onde está
The flaw was introduced by commit 079096f103fa (October 3, 2015) in the kernel's TCP established-connection lookup; the upstream fix is commit 4e2b3ef19e0e, "inet: reject unhashed sockets during established lookups".
Como foi descoberta
Researcher Hyunwoo Kim (@v4bel) found and reported the bug. After an agreed embargo with Linux distributions ended, he published the full advisory on the oss-security mailing list. The fix landed upstream on November 3, 2026, and Red Hat published its own description of the race. As of November 15 many distributions had not shipped patched kernels. No exploitation in the wild is reported and the flaw is not in CISA KEV. Kim previously disclosed Dirty Frag and Januscape.
PoC existente, privada
O exploit existe, mas quem o tem não o publicou.
Kim has not released exploit code; his repository holds only the write-up. He says he wrote only a very limited, lab-specific exploit aimed at VM servers. In his tests, default Nginx, Apache and SSH servers in QEMU/KVM VMs crashed in about 5 seconds to 2 minutes. A finely timed PoC against a quiet bare-metal Nginx server with 8 RX queues took about 14 hours. It demonstrates only the denial-of-service panic, not code execution. Urgency is moderate: no public exploit exists, but virtualized servers are quick to crash, so patch or apply the stopgaps.
Explicação gerada por IA a partir da matéria e dos artigos, em 16/11/2026 12:19.
Detalhes técnicos
- The lockless RCU lookup of established sockets can see socket memory that has been freed and reused. The kernel rechecks the connection tuple but does not check that the new socket is hashed (securityonline.info). Red Hat describes it as recycled socket memory being accessed before it is initialized and hashed (bleepingcomputer.com).
- A stale ehash lookup must run concurrently with the freeing and reallocation of a request socket in a different RX context (gbhackers.com, bleepingcomputer.com).
- In the first scenario the request socket is reallocated as a SYN cookie request at the same slab address with the same tuple. It passes the tuple recheck but has no attached listener, so the kernel dereferences a NULL listener and panics (gbhackers.com, securityonline.info). On a container host this crashes the entire host.
- The second scenario recycles a full child socket and could lead to a remote use-after-free. It needs nested races (gbhackers.com).
- Targets need two or more RX queues, or RPS spreading packets across CPUs. Packet distribution relies on the receive-hash difference between IPv4 fragments and regular TCP packets (bleepingcomputer.com). The race triggers even when a server only calls listen().
- In Kim's tests, QEMU/KVM VMs crashed in about 5 seconds to 2 minutes. A quiet bare-metal Nginx server with 8 RX queues took about 14 hours (securityonline.info, gbhackers.com).
- Affected: Linux kernel 4.4 and later, before 6.12.117, and the 6.18 series before 6.18.63 (securityonline.info).
Remediação
- Install a kernel carrying commit 4e2b3ef19e0e ("inet: reject unhashed sockets during established lookups") or your distribution's backport. Fixed releases are 6.12.117, 6.18.63, 7.3.4 and 7.4-rc2 (securityonline.info). Reboot and confirm the running version with uname -r (gbhackers.com).
- Stopgap 1: set net.ipv4.tcp_syncookies to 0 in every network namespace, including containers, since it must be applied separately to containers with their own namespaces (theregister.com). This weakens SYN flood protection.
- Stopgap 2: limit external interfaces to a single RX queue and disable RPS and RFS. This can hurt network performance (theregister.com).
Relatos divergentes
- Code execution: bleepingcomputer.com says the flaw does not allow code execution or data alteration. securityonline.info and gbhackers.com say a use-after-free path exists but is hard to trigger. They add that Kim says it is not a TCP RCE so far, and that neither he nor any vendor has indicated RCE is possible. Only a DoS crash is publicly documented.
- Patch status: bleepingcomputer.com says it affects almost all Linux distributions. gbhackers.com, the most specific on this point, says many distributions had not shipped patched kernels as of November 15.
- Fix timing: bleepingcomputer.com says the fix was merged about two weeks before publication, which is consistent with the November 3 date given by gbhackers.com.
Linha do tempo
-
NoisyTCP (CVE-2026-99288): 11-year-old Linux kernel TCP flaw lets unauthenticated attackers crash servers remotely, including container hosts
The Register adds Red Hat's Important rating with a CVSS score of 7.5. It also adds that Kim has a limited lab RCE exploit against VMs, that the flaw also affects containerized workloads, and that SYN cookie mitigation must be applied per network namespace.
ler a análise desta fonte ocultar
The Register adds Red Hat's Important rating with a CVSS score of 7.5. It also adds that Kim has a limited lab RCE exploit against VMs, that the flaw also affects containerized workloads, and that SYN cookie mitigation must be applied per network namespace.
NoisyTCP (CVE-2026-99288) is a use-after-free in the Linux kernel TCP code, found by researcher Hyunwoo Kim. It dates to October 2015. An unauthenticated remote attacker with access to an open TCP port can crash the whole host, including a host running the vulnerable service in a container. The article says default Nginx, Apache and SSH services are exposed, provided the system spreads incoming traffic across multiple receive queues or CPUs. Kim's tests showed VM crashes in 5 seconds to 2 minutes, but a bare-metal Nginx server took about 14 hours. The confirmed impact is denial of service. Kim has a limited lab RCE exploit against VMs, but no general-purpose RCE has been shown. Red Hat rates the flaw Important with a CVSS of 7.5.
Technical Details
- Two kernel code paths race while processing incoming TCP traffic, and one uses memory that has already been freed. This causes a kernel panic.
- A second, much harder exploitation path involves a use-after-free.
- No special configuration is needed, but the target must process traffic across multiple RX queues or CPUs.
Remediation
- The upstream fix was committed on November 3. Install your distribution's updated kernel.
- Workaround 1: disable TCP SYN cookies. This must be applied separately in each container network namespace and weakens SYN flood protection.
- Workaround 2: restrict incoming packet processing to a single receive queue. This can reduce network performance.
-
NoisyTCP: 11-year-old Linux kernel TCP flaw (CVE-2026-99288) lets remote attackers crash servers via SYN cookie race
ler a análise desta fonte ocultar
A Linux kernel NULL pointer dereference in the TCP subsystem, dubbed NoisyTCP (CVE-2026-99288), was disclosed by researcher Hyunwoo Kim. It has existed for 11 years and affects almost all Linux distributions. An unauthenticated remote attacker can send crafted TCP packets that exploit a race between SYN cookie handling and socket hashing, causing a kernel panic and denial of service on hosts running Nginx, Apache or SSH with default configurations. It does not allow data alteration or code execution. The race window is narrow, so triggering it against a quiet bare-metal server can take a long time. The article reports no exploitation in the wild. This article was based on the article alone; no web research was performed.
Technical Details
According to Red Hat, a race condition during lockless socket lookups lets recycled socket memory be accessed before it is initialized and hashed. Triggering it requires a stale ehash lookup to run concurrently with the freeing and reallocation of a request socket in different RX contexts. Packet processing is spread using the receive-hash difference between IPv4 fragments and regular TCP packets. Targets therefore need two or more RX queues, or RPS distributing packets across CPUs.
Remediation
A fix was merged into the mainline Linux kernel about two weeks before publication. Apply the patch or upgrade the kernel. If that is not possible, disable SYN cookies, which reduces SYN flood protection, and serialize to a single RX queue to block the race.
-
NoisyTCP (CVE-2026-99288): 11-year-old Linux kernel TCP race lets remote attackers panic servers without authentication
The article adds the introducing commit 079096f103fa (October 3, 2015), the fix date of November 3, 2026, the 8-RX-queue bare-metal test, and the note that many distributions had not shipped patched kernels as of November 15. It also adds the reboot-and-verify advice.
ler a análise desta fonte ocultar
The article adds the introducing commit 079096f103fa (October 3, 2015), the fix date of November 3, 2026, the 8-RX-queue bare-metal test, and the note that many distributions had not shipped patched kernels as of November 15. It also adds the reboot-and-verify advice.
A Linux kernel TCP race condition, NoisyTCP (CVE-2026-99288), was found and patched by researcher Hyunwoo Kim (@v4bel). It lets a remote, unauthenticated attacker crash a server. The flaw dates to commit 079096f103fa of October 3, 2015. Kim published the full advisory on oss-security after the distribution embargo ended. The article says only a denial-of-service crash is publicly documented, and neither Kim nor any vendor has indicated that RCE is possible. Many distributions had not shipped patched kernels as of November 15. This summary rests on the article alone. No primary-source research was done.
Technical Details
A stale established-hash (ehash) lookup races with the freeing and reallocation of a request socket in a different RX context. In the first scenario the request is reallocated as a SYN cookie request at the same slab address with the same tuple. It passes the tuple recheck but has no attached listener, so the kernel dereferences a NULL listener and panics. This crashes the whole host, including container hosts. The second scenario recycles a full child socket and could lead to a remote use-after-free. It needs nested races and is hard to trigger. The race needs two or more RX queues, or RPS. Default Nginx, Apache and SSH listeners are exposed. In QEMU/KVM VMs, crashes took about 5 seconds to 2 minutes. A bare-metal Nginx server with 8 RX queues took about 14 hours.
Fix Details
The fix is commit 4e2b3ef19e0e, which landed on November 3, 2026. Upstream fixes are in Linux 6.12.117, 6.18.63 and 7.3.4.
Remediation
Install the patched kernel, reboot, and confirm it with uname -r. If patching is not possible, there are two stopgaps. The first is to set net.ipv4.tcp_syncookies=0 in each network namespace, which weakens SYN flood protection. The second is to limit the inbound interface to a single RX queue and disable RPS and RFS.
-
NoisyTCP (CVE-2026-99288): Linux kernel TCP race enables remote pre-auth kernel panic, fixed in 6.12.117, 6.18.63 and 7.3.4
ler a análise desta fonte ocultar
CVE-2026-99288, named NoisyTCP by researcher Hyunwoo Kim, is a race condition in the Linux kernel TCP established-connection lookup. It has existed since a 2015 commit. A remote, unauthenticated attacker can send crafted TCP traffic to trigger a kernel panic. A second path can lead to a use-after-free, but Kim calls it hard to trigger. No exploitation in the wild has been reported, and the flaw is not in CISA KEV. No public exploit code exists, and Kim says it is not a TCP RCE so far. This summary is based on the article alone. No vendor or primary-source research was done, so no CVSS score is available.
Technical Details
The lockless RCU lookup of established sockets can see socket memory that has been freed and reused. The kernel rechecks the connection tuple but does not check that the new socket is hashed. A request socket is freed and its memory is reused for a SYN cookie request. That request has no listener, so the kernel dereferences a NULL pointer and panics. The race needs two or more RX queues, or RPS spreading packets across CPUs. It triggers even when a server only calls listen(), so default Nginx, Apache and SSH are exposed. Kim triggered it in about 5 seconds to 2 minutes on QEMU/KVM VMs, and in about 14 hours on a quiet bare-metal Nginx server. A crash on a container host takes down the entire host. The affected range is Linux kernel 4.4 and later, before 6.12.117, and the 6.18 series before 6.18.63.
Remediation
Install a kernel carrying commit 4e2b3ef19e0e ("inet: reject unhashed sockets during established lookups"), or your distribution's backport. Fixed releases are 6.12.117, 6.18.63, 7.3.4 and 7.4-rc2. Two stopgaps exist. The first is to set net.ipv4.tcp_syncookies to 0 in every network namespace, including containers, which weakens SYN flood protection. The second is to limit external interfaces to a single RX queue and disable RPS and RFS.