Add kernel_kselftest.sh

[bmastbergen]

This is a helper script for running kselftests against a kernel under test. It attempts to install the appropriate dependencies for different versions of Rocky Linux and attempts to run appropriate tests, with appropriate permissions, for a few select kernel versions.

This is just a starting point. Hopefully this continues to evolve, increasing the number of successful that can reliably be run.

[Copilot]

Pull Request Overview

This PR adds a comprehensive helper script for running kernel self-tests (kselftests) against different kernel versions on Rocky Linux systems. The script automates dependency installation and test execution with appropriate permissions.

• Implements version-specific dependency installation for Rocky Linux 8, 9, and 10
• Provides kernel version-specific test configurations with different target sets and skip lists
• Includes automated test execution with both regular and sudo permissions for different test categories

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[ciq-it-service-account]

🎉 Snyk checks have passed. No issues have been found so far.

✅ code/snyk check is complete. No issues have been found. (View Details)

[Copilot]

Pull Request Overview

Copilot reviewed 1 out of 1 changed files in this pull request and generated 6 comments.

---

_{Tip: Customize your code reviews with copilot-instructions.md. Create the file or learn how to get started.}

[roxanan1996]

Minor question: do we need to add a Signed-off -by for all commits? Or is it just for the kernel sources?
Personally, I am so used to checking Signed-off-by everywhere, so I would encourage people to use it everywhere.
This can be achieved by using git commit -s.

[bmastbergen]

Minor question: do we need to add a Signed-off -by for all commits? Or is it just for the kernel sources? Personally, I am so used to checking Signed-off-by everywhere, so I would encourage people to use it everywhere. This can be achieved by using git commit -s.

No, we haven't stipulated sign-offs for any repos other than the kernel. Certainly doesn't hurt though.

[roxanan1996]

Hmm, I would exit here. Because otherwise it will run the kselftests for an unknown version.

I would imagine we run into this if we test a new kernel. So an early exit would catch this and then the person running this would add a case here for the new kernel version. Otherwise the kselftests would run and this warning may be missed.

What do you think?

[bmastbergen]

Yea, I almost made it exit here. I could go either way.

[roxanan1996]

~/selftest-$(uname -r)-$run.log does not work for me. And I believe it would be a bit hard to differentiate between logs if we save all of them in the same place, even though we do include the version in the name.

The folder structure for each vm (kernel type) I have is:

kernel-dist-git
kernel-src-tree
kernel-src-tree-tools -> /home/rnicolescu/ciq/kernel-src-tree-tools
kernel-tools -> /home/rnicolescu/ciq/kernel-tools
build-logs
kselftests-logs

So for me ../kselftests-logs/selftest-$(uname -r)-$run.log would work. Or something similar, happy to adjust it.

We can also add a param for this, but I would keep it simple. I think it would be super easier for our tooling if we all have the same tree structure for kernels.
What do you think?

[bmastbergen]

yea, ../kselftests-logs makes sense. or....err... ../kselftest-logs I never know whether to use kselftest or kselftests

[roxanan1996]

I am also bad at naming. Pick one and stick with it :)

[roxanan1996]

Yesterday I triggered another run by mistake and I wish I saved the previous log.
If you also think it's valuable, I would check if a file with the same name exists and adjust the run var accordingly.

[bmastbergen]

or maybe timestamping logs makes more sense than numbering?

[roxanan1996]

Either way, I will not know easily which one is which if they have long names. Not sure here which one is the best at the moment. Maybe this is over engineering at the moment.