SB2026101442 - Multiple vulnerabilities in containerd

Description

This security bulletin contains information about 3 vulnerabilities.

1) Allocation of Resources Without Limits or Throttling (CVE-ID: N/A)

CWE-ID: CWE-770 - Allocation of Resources Without Limits or Throttling

CVSSv4: 8.7 [CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:H/SC:N/SI:N/SA:N]

The vulnerability allows a remote attacker to cause a denial of service.

The vulnerability exists due to allocation of resources without limits in containerd's image pull process when downloading image layers from an untrusted registry. A remote attacker can return a crafted Content-Length header to cause a denial of service.

Exploitation requires parallel layer fetch to be enabled through the concurrent_layer_fetch_buffer setting. The crafted header triggers an unhandled memory allocation panic that crashes the daemon. Existing running containers remain unaffected.

2) Uncaught Exception (CVE-ID: CVE-2026-95839)

CWE-ID: CWE-248 - Uncaught Exception

CVSSv4: 0 [CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:N/SC:N/SI:N/SA:N]

The vulnerability allows a remote attacker to cause a denial of service.

The vulnerability exists due to an unrecovered panic in containerd's image manifest processing when pulling or importing an image through the transfer service whose child descriptor contains the io.containerd.manifest.subject annotation and a malformed digest. A remote attacker can supply an image manifest containing such a descriptor to cause a denial of service.

The transfer service includes default CRI PullImage requests. In Kubernetes environments, kubelet automatically retries failed image pulls with backoff for pods referencing an untrusted image.

3) Missing Release of Resource after Effective Lifetime (CVE-ID: N/A)

CWE-ID: CWE-772 - Missing Release of Resource after Effective Lifetime

CVSSv4: 0 [CVSS:4.0/AV:N/AC:L/AT:N/PR:N/UI:N/VC:N/VI:N/VA:N/SC:N/SI:N/SA:N]

The vulnerability allows a local user to cause a denial of service.

The vulnerability exists due to missing release of read-only mounts and pinned snapshots in containerd's CRI plugin when restoring a container from a checkpoint image fails. A local user can repeatedly create containers from checkpoint images with malformed metadata or a base image that cannot be pulled to cause a denial of service.

The issue occurs on Linux with CRI checkpoint and restore support enabled. In Kubernetes, exhaustion of the host mount table and disk space can prevent containerd and the kubelet from pulling images or starting new pods on the node.

Remediation

Install update from vendor's website.

References

• https://github.com/containerd/containerd/security/advisories/GHSA-8647-2jxj-p2rc
• https://github.com/containerd/containerd/security/advisories/GHSA-vwhp-8p66-mq4f
• https://github.com/containerd/containerd/security/advisories/GHSA-5p7c-hggf-c748