CVE-2026-18107 in Red Hat
Summary
by MITRE • 07/28/2026
A flaw was found in CRIU's handling of restartable sequences (rseq) during checkpoint/restore. A malicious process inside a container can register an rseq critical section that hijacks CRIU's parasite code injection during checkpoint, allowing it to spoof the process credentials saved in the checkpoint image. On restore, the container process gains elevated capabilities and zeroed UIDs/GIDs.
The practical impact on Red Hat products is limited by several factors: checkpoint/restore requires root privileges (podman) or cluster-admin RBAC (OpenShift) to trigger and cannot be initiated from within the container itself; on OpenShift prior to 4.17 the feature required explicit opt-in, and on 4.17+ the kubelet checkpoint API RBAC is not configured by default; OpenShift enforces user namespaces by default for regular workloads (hostUsers is gated behind admin-only SCCs), which makes the spoofed capabilities namespace-scoped and ineffective for privilege escalation; SELinux type enforcement (container_t) blocks privilege transitions independently of capabilities; seccomp filters persist through checkpoint/restore and cannot be corrupted via the parasite; and kernel mount namespace ownership checks on RHEL 9/10 kernels prevent mount-based container escape even with spoofed capabilities.
If you want to get best quality of vulnerability data, you may have to visit VulDB.
Analysis
by VulDB Data Team • 07/28/2026
The vulnerability in CRIU's handling of restartable sequences represents a sophisticated attack vector targeting containerized environments through checkpoint and restore operations. This flaw specifically exploits the rseq critical section mechanism to manipulate CRIU's parasite code injection process during checkpoint operations. The technical implementation leverages the fact that rseq allows processes to register critical sections that can be hijacked by malicious code, enabling attackers to modify how credentials are saved in checkpoint images. When the container subsequently restores, it inherits spoofed privileges including elevated capabilities and zeroed user identifiers, effectively bypassing normal security boundaries.
This vulnerability falls under CWE-284 Access Control Bypass, as it allows unauthorized privilege escalation through manipulation of credential storage during container state transitions. The attack vector aligns with ATT&CK technique T1564.001 Stealthy Persistence via Process Injection and T1068 Privilege Escalation through exploitation of system-level code injection mechanisms. The flaw demonstrates a complex interaction between kernel-level memory management, process state persistence, and container security boundaries that requires deep understanding of both operating system internals and container orchestration platforms.
The practical impact on Red Hat products remains significantly constrained by multiple defensive measures implemented at various layers of the container stack. The fundamental requirement for root privileges or cluster-admin RBAC access to initiate checkpoint/restore operations creates a substantial barrier for attackers, as these permissions are not typically available within standard container contexts. Additionally, OpenShift's default security configurations further limit exploitability through explicit opt-in requirements for older versions and default RBAC restrictions in newer releases, preventing unauthorized access to the vulnerable APIs.
The container security model provides additional protection through user namespace enforcement, which makes any spoofed capabilities effectively scoped to individual namespaces rather than providing global privilege escalation. This defense mechanism ensures that even if credential spoofing were successful, the attacker would be limited by namespace boundaries that separate different workload contexts. The SELinux type enforcement system adds another layer of protection by blocking privilege transitions independent of capability manipulation, while seccomp filters maintain their integrity across checkpoint/restore operations, preventing corruption of security policies.
Kernel-level protections further mitigate potential exploitation through mount namespace ownership checks present in RHEL 9 and 10 systems, which prevent mount-based container escape attempts even when capabilities have been spoofed. These kernel safeguards ensure that malicious processes cannot leverage compromised credentials to gain unauthorized access to host resources through filesystem mounting operations. The combination of these defense-in-depth mechanisms creates multiple barriers that significantly reduce the exploitability of this vulnerability in production Red Hat environments, though the underlying flaw remains a serious concern for container security architecture and requires careful monitoring and patching strategies.