Update/runc 1.3.2

[n4mlz]

Description

• Bump the vendored runc used by contest / CI pipelines to v1.3.2 (first release that carries the new log-based CPU shares → weight conversion).
  • cf. https://github.com/opencontainers/runc/releases/tag/v1.3.2
• Port that conversion into youki’s cgroup v2 and systemd backends so our runtime writes the same cpu.weight values (e.g. 22000 shares → 1204).
• Update the corresponding unit tests and contest expectation so validate-contest-runc passes with the newer runtime.

Type of Change

• Bug fix (non-breaking change that fixes an issue)
• New feature (non-breaking change that adds functionality)
• Breaking change (fix or feature that would cause existing functionality to not work as expected)
• Documentation update
• Refactoring (no functional changes)
• Performance improvement
• Test updates
• CI/CD related changes
• Other (please describe):

Testing

• Added new unit tests
• Added new integration tests
• Ran existing test suite
• Tested manually (please provide steps)

Related Issues

Additional Context

• PR Implement Linux memory policy #3230 depends on runc v1.4.0-rc.2 being available in main
  • merging this update first keeps that PR focused on its functionality changes.

[saku3]

I understand the background.

Due to the following Kubernetes issue, the conversion formula from cgroup v1 shares to cgroup v2 weight has been changed:
kubernetes/kubernetes#131216

The fix in runc is as follows.
opencontainers/cgroups#20

[saku3]

@n4mlz
Thanks for the PR!

Overall, it looks good to me.
If you could add comments similar to runc’s, I think it would make the changes even clearer.

[saku3]

Having comments would make it even easier to understand.

[saku3]

Having comments would make it even easier to understand.

[saku3]

Thanks!

LGTM once CI passes.

[saku3]

@utam0k @YJDoc2
PTAL and run CI, please 🙏

[YJDoc2]

Hey looks good to me, but I'll request you to add the runc release tag link and pr links mentioned by saku3 in a comment here. In future we might not be sure as to why we are using this particular formula , and the git link to pr might also be lost if we run formatting or file movement. The explicit links in comments would help us get the context for this.

[YJDoc2]

Also our current CI requires exactly one kind/ label to be assigned , so it works out with our release ci ; I have removed the bug label and kept test ; feel free to change it to something else as appropriate.

[n4mlz]

Sorry for the delay in my response. I’ve added some comments. Please take a look.