Critical Linux kernel flaw allows remote root attacks on MPTCP servers

Security researchers at Anthropic have developed an exploit for a critical Linux kernel vulnerability that allows unauthenticated attackers to remotely execute commands as root on servers running Multipath TCP (MPTCP) listeners.
Security researchers at Anthropic have developed an exploit for a critical Linux kernel vulnerability that allows unauthenticated attackers to remotely execute commands as root on servers running Multipath TCP (MPTCP) listeners.
The flaw, tracked as CVE-2026-98432, also affects ordinary TCP connections, although remote code execution has not been publicly demonstrated on servers without an MPTCP listener.
Multipath TCP (MPTCP) is a TCP feature that lets a single connection use multiple network paths at the same time to improve performance and reliability. It is commonly used in environments where devices have multiple network connections, such as Wi-Fi and cellular networks.
CVE-2026-98432 is a use-after-free vulnerability in the kernel's IPv4 TCP networking code that received a severity score of 9.8 out of 10 from the Linux kernel CVE team.
The vulnerability affects Linux kernels dating back to version 4.4, including widely used distributions such as Ubuntu, Red Hat Enterprise Linux (RHEL), and Rocky Linux.
According to the Linux kernel advisory, the vulnerability stems from a race condition in how the kernel handles TCP connection requests, potentially causing a use-after-free condition.
"tcp_v4_syn_recv_sock() transfers ownership of ireq->ireq_opt to the child socket (newinet->inet_opt) without copying it," reads the advisory.
"Another cpu can concurrently process a retransmitted SYN for the same request socket, and send a SYNACK from tcp_check_req(). tcp_v4_send_synack() and inet_csk_route_req() read ireq->ireq_opt under rcu_read_lock() only, and ip_build_and_send_pkt() and ip_options_build() then read opt->optlen twice."
A similar issue can occur during TCP request socket migration when net.ipv4.tcp_migrate_req is enabled, or a BPF_SK_REUSEPORT_SELECT_OR_MIGRATE program is used.
Although triggering the vulnerability requires no authentication, Red Hat says exploiting the race condition is more complex than just sending a malicious network packet.
Red Hat classifies the vulnerability as a potential denial-of-service issue that could cause a kernel crash, assigning it a lower severity score of 5.9 because exploitation is complex.
However, researchers at Anthropic demonstrated that the vulnerability could be exploited for remote code execution under specific conditions.
An Anthropic report explained that an unreleased Claude AI model developed an exploit that bypassed kernel address space layout randomization (KASLR) and executed arbitrary commands as root on a Linux server running an MPTCP listener.
According to Anthropic, the exploit took about 53 minutes to achieve remote root code execution, showing the vulnerability could be exploited for more than denial-of-service attacks.
Source: BleepingComputer. This article summarizes the linked reporting; figures and quotes are attributed to the original source.