CVE-2026-64508: Branch Target Reuse in Linux kernel BPF JIT memory
- Vendor
- Linux
CVE-2026-64508 is a Linux kernel BPF JIT hardening weakness associated with Branch Target Reuse (BTR), a Spectre-v2 attack. The BPF JIT allocator packs small programs into larger executable allocations and reuses freed regions without necessarily invalidating indirect branch predictions associated with their previous contents. An attacker can train an indirect branch, replace the original program with attacker-influenced JIT code at the same address, and trigger speculative execution at an obsolete target offset. This speculative execute-after-free primitive can reach misaligned disclosure gadgets, including gadgets encoded in forward-jump offsets that bypass cBPF constant blinding. The fix introduces architecture-selectable predictor-flushing infrastructure before packed JIT memory is reused; its default implementation is a no-op and requires architecture-specific activation.
Impact, mitigation & remediation
What it means and what to do now: patch path, interim mitigations, and the assume-compromise checklist.
Impact
What an attacker gets, and what they’ve been doing with it.
An unprivileged local attacker can exploit the associated BTR attack chain to disclose arbitrary kernel memory and sensitive process memory through speculative-execution side channels. End-to-end Linux cBPF demonstrations on Intel processors recovered a root password hash from a privileged process, including with constant blinding enabled. Such disclosure can expose credentials and support subsequent compromise, but it does not itself establish architectural arbitrary code execution or a complete privilege-escalation exploit. Public proof-of-concept exploits exist; active exploitation is not established.
Mitigation
If you can’t patch tonight, do this now.
Until patched, restrict execution of untrusted local code and limit unprivileged access to BPF/cBPF JIT functionality where operationally feasible. Restricting unprivileged eBPF alone does not remove the cBPF attack surface. Do not rely on constant blinding or Indirect Branch Tracking alone, because demonstrated BTR techniques bypass these defenses in tested configurations.
Remediation
Patch, then assume compromise.
Install the applicable vendor kernel update and reboot. Red Hat addresses the issue in RHSA-2026:81406 for RHEL 10, including kernel build 6.12.0-211.65.1.el10_2, and RHSA-2026:81552 for RHEL 9. Debian stable trixie provides fixes in linux 6.12.100-1 or later. For upstream or custom kernels, apply the BPF predictor-flushing infrastructure and the applicable architecture-specific implementation. On x86, deploy the companion CVE-2026-64507 change enabling IBPB flushing; the CVE-2026-64508 hook alone defaults to a no-op and is not a complete defense.
Exploits
No public exploits tracked yet. Mallory keeps watching.
Affected products & vendors
Products and vendors Mallory has correlated with this vulnerability. Open in Mallory to drill down to specific CPE configurations and version ranges.
References
Frequently asked questions
- Is CVE-2026-64508 exploited in the wild?
- As of Nov 13, 2026, Mallory has no reports of CVE-2026-64508 being exploited in the wild. It is not listed in CISA KEV.
- Is there a public exploit for CVE-2026-64508?
- As of Nov 13, 2026, Mallory has found no public exploit for CVE-2026-64508.
- Is there a patch for CVE-2026-64508?
- Install the applicable vendor kernel update and reboot. Red Hat addresses the issue in RHSA-2026:81406 for RHEL 10, including kernel build 6.12.0-211.65.1.el10_2, and RHSA-2026:81552 for RHEL 9.
- What products are affected by CVE-2026-64508?
- Linux Kernel.
- How severe is CVE-2026-64508?
- CVE-2026-64508 has a CVSS v3.1 score of 5.5 (Medium) and an EPSS score of 0.19% (8th percentile) as of Nov 13, 2026.
Recent activity
33 sources tracked across advisories, community write-ups, and news. New activity surfaces here as Mallory finds it.