A critical Linux kernel vulnerability designated CVE-2026-64600 and named “RefluXFS” enables an unprivileged local user to achieve full root access by exploiting a race condition in the XFS filesystem’s copy-on-write path. Discovered by Qualys Threat Research Unit, the flaw bypasses SELinux enforcement and leaves no trace in kernel logs, making it particularly dangerous for multi-tenant and internet-facing systems.
How the RefluXFS Race Condition Works
The vulnerability resides in the XFS filesystem code that handles copy-on-write operations for reflinked files. When two concurrent O_DIRECT writes target the same reflinked file, the kernel briefly drops the inode lock while waiting for transaction log space. A second writer can exploit this window to remap the file and alter the reference count, causing the first writer to overwrite the original disk block using a stale reference. Because O_DIRECT bypasses page cache and skips revalidation, the corrupted write becomes permanent on disk.
What Systems Are Vulnerable to CVE-2026-64600
Any Linux system running kernel version 4.11 (released in 2017) or later without the patch is potentially affected, provided three conditions are met: an XFS filesystem with reflink=1 enabled, a directory writable by an unprivileged user, and a high-value target file such as a SUID-root binary or system configuration file. Qualys estimates over 16.4 million systems worldwide may be exposed.
Confirmed affected distributions include Red Hat Enterprise Linux 8, 9, and 10, CentOS Stream 8 through 10, Oracle Linux 8 through 10, Rocky and AlmaLinux 8 through 10, CloudLinux 8 through 10, Amazon Linux 2023 and AL2 (December 2022 onward), and Fedora Server 31 and later. Debian, Ubuntu, and SUSE are at lower risk unless XFS was manually selected with reflink=1 enabled.
Why Existing Security Controls Cannot Block RefluXFS
RefluXFS operates at the filesystem allocation layer, below the reach of conventional kernel hardening mechanisms. KASLR, SMEP, and SMAP target different attack surfaces. Kernel lockdown does not restrict O_DIRECT or FICLONE calls. SELinux does not block the vulnerable code path. Container isolation techniques, including user-namespace restrictions and capability limits, also fail to mitigate the flaw. No reliable workaround exists short of applying the vendor-supplied kernel patch.
The AI-Assisted Discovery Process
The vulnerability was identified through a joint research initiative between Qualys and Anthropic, integrating the Claude Mythos Preview model into Qualys TRU’s audit workflow under Project Glasswing. Researchers tasked the model with hunting for Dirty COW-style race conditions. After iterative prompting, it identified the XFS flaw and generated a working proof-of-concept, which Qualys engineers independently verified before coordinating disclosure with kernel maintainers.
Broader Context: A Pattern of Kernel Write Flaws
RefluXFS is the fourth significant Linux privilege-escalation vulnerability disclosed in roughly three months, following Copy Fail (CVE-2026-31431), Dirty Frag (CVE-2026-43284 and CVE-2026-43500), and DirtyClone (CVE-2026-43503). Each targets a different kernel component but shares a common theme: tricking the kernel into writing attacker-controlled data into protected memory or disk blocks that back setuid binaries or configuration files. These disclosures represent an escalating pattern of researchers finding novel ways to exploit timing flaws and logic gaps in the kernel’s write paths.
What Organizations Should Do Now
Vendor-fixed kernels are already available and being backported across enterprise distributions including RHEL, Oracle Linux, AlmaLinux, Rocky Linux, and Fedora. Organizations should prioritize patching internet-facing and multi-tenant systems immediately. Apply the relevant distribution security update and perform a full reboot to confirm the fix is active, since no interim mitigation or configuration workaround currently neutralizes the flaw. For ongoing protection, organizations should invest in a comprehensive vulnerability management platform that provides continuous asset inventory, patch prioritization, and automated remediation workflows to reduce exposure windows for critical kernel-level vulnerabilities.
