CI on UNIBO premises

[FrancescoConti]

This PR moves CI to local UNIBO premises in place of shared IIS runners.
It also removes regressions running on ri5cy.

[rfiorani]

LGTM!

[FrancescoConti]

I could not test the push from GitHub because GitHub actions are currently broken, but the pipeline seems to be now working fine on GitLab side. Needs a bit of cleanup though (squashing a few "update nonfree" commits in particular)

[FrancescoConti]

Currently failing only in elab (looks like a legitimate problem on our servers, to be checked)

[rfiorani]

I updated the nonfree commit, fixing the elaboration issue and improving the CI script.
The vsim regression tests for the C32E40X are temporarily commented out.
The GitHub Action timed out after 6 hours. Should we increase the timeout limit?

[AlessandroNadalini]

I updated the nonfree commit, fixing the elaboration issue and improving the CI script. The vsim regression tests for the C32E40X are temporarily commented out. The GitHub Action timed out after 6 hours. Should we increase the timeout limit?

Unfortunately 6h is a hard cap for GitHub actions, so at the moment there's no real solution for the timeout failing. Re-triggering the action solves the problem since it simply checks the status of the offloaded pipeline for the commit that triggered the CI.

[AlessandroNadalini]

The GitHub Action timed out after 6 hours. Should we increase the timeout limit?

I noticed that the timeout in .github/workflows/gitlab-ci.yml is already set to 16 hours, so I have no idea why it timed out after 6.

All GitHub-hosted runners for actions have a maximum timeout limit of 6h (https://docs.github.com/en/actions/reference/limits)

[FrancescoConti]

Not much to do about the timeout (unless we change the CI to use Verilator later, which is both faster and potentially more parallel). I'll merge.