Back Forklog Solana Foundation Alerts Validators to Critical Linux Vulnerability
A critical vulnerability, CVE-2026-98432, has been discovered in the Linux kernel, which under certain conditions allows attackers to remotely execute commands on servers without authorization. Tim Garcia, head of validator relations at the Solana Foundation, alerted Solana node operators the issue.
Validators! CVE-2026-98432 is patched in the linux kernel but not yet patched on ubuntu. Here’s what you need to do if you’re running ubuntu: — Tim Garcia (@TimGarcia0) November 2, 2026
Validators! CVE-2026-98432 is patched in the linux kernel but not yet patched on ubuntu. Here’s what you need to do if you’re running ubuntu:
— Tim Garcia (@TimGarcia0) November 2, 2026
The issue has been rated 9.8 out of 10 on the CVSS 3.1 scale.
The flaw was found in the IPv4 network subsystem, which handles TCP connections. It is caused by a race condition, where two simultaneous operations interfere with each other, resulting in the kernel freeing a memory area that is still in use.
According to the Red Hat bulletin , an attacker can send network requests with precise timing to cause kernel memory corruption. This often leads to a server crash but can potentially allow arbitrary code execution.
The vulnerability has been present in the Linux kernel since version 4.4, released in 2016. Developers have prepared patches for supported branches, but they are not yet available in all distributions.
As of November 2, Ubuntu listed versions 22.04, 24.04, 26.04, and 26.10 as vulnerable. Updates are still being prepared.
Garcia urged Solana validators using Ubuntu to temporarily restrict certain network packet processing and disable Multipath TCP and TCP connection migration features until patches are released.
CVE-2026-98432 is not directly related to Solana’s code. The threat affects the operating system of servers running validators and other network applications.
Compromise of such a node may disrupt its operation but does not automatically mean a breach of the blockchain itself.
Researchers Demonstrate Remote Exploit
On October 27, Anthropic specialists reported that their internal research model, not yet released by the company, was able to develop a working exploit for CVE-2026-98432.
In laboratory conditions, the AI created a two-step attack on a Linux server with an active Multipath TCP listener. It allowed root access without credentials or user actions.
In the first stage, the program revealed kernel memory location details and bypassed the KASLR protection mechanism. In the second, it used the same flaw to execute commands as root.
According to Anthropic, the first stage took 48 minutes, and the second an additional five minutes. The researchers did not publicly disclose the attack code.
The demonstration was conducted on a server with a specific configuration. Researchers did not claim the attack’s applicability to any Linux server.
Amazon Web Services announced the release of patches for Amazon Linux 2 and Amazon Linux 2023. The company has no evidence of CVE-2026-98432 being exploited in real-world attacks.
Previously, at the end of October, the Solflare crypto wallet reported an incident involving validator keys at a third-party staking provider. As a result, the node stopped generating income for users who had delegated approximately 1.56 million SOL, valued at $160 million.
The assets themselves were not affected. The company promised to compensate missed rewards for those who transfer their stake to a new validator.
Follow ForkLog on social media
The full story
This article is one source in a clustered incident — the cluster page carries the summary, timeline and every other outlet covering it.
