[bug] cgroup.freeze: no such device

[bitoku]

Description

When cri-o tries running runc state, it rarely gets

read /sys/fs/cgroup/<cgroup>/cgroup.freeze: no such device

or

seek /sys/fs/cgroup/<cgroup>/cgroup.freeze: no such device

I don't know the cause, but it looks similar to containers/crun#539

This breaks some CI in k/k.

• runc needs to be upgraded ([Flaking Test] [sig-node] Hostname of Pod [NodeConformance] a pod without FQDN is not affected by SetHostnameAsFQDN field) kubernetes/kubernetes#132030
• [Flaking Test][sig-node] Container Runtime blackbox test on terminated container should report termination message from file when pod succeeds and TerminationMessagePolicy FallbackToLogsOnError is set [NodeConformance] [Conformance] kubernetes/kubernetes#129760

https://storage.googleapis.com/kubernetes-ci-logs/logs/ci-kubernetes-node-swap-fedora/1940313391126024192/artifacts/n1-standard-2-fedora-coreos-42-20250609-3-0-gcp-x86-64-dc23837a/crio.log

Steps to reproduce the issue

We don't have reproducer yet, but it happens repeatedly in some test in k/k.

Describe the results you received and expected

It shouldn't get error.

What version of runc are you using?

runc version 1.3.0, commit: v1.3.0-0-g4ca628d1

Host OS information

No response

Host kernel information

No response

[kannon92]

cc @kolyshkin

[lifubang]

Maybe the regression was introduced here: #2955