ci: harden Tier CI result handling and dashboard polish

[AlexChenIC]

Summary

This PR is a follow-up improvement to the CVA6 Tier CI flow introduced in PR #3285. It focuses on hardening Tier CI result handling and polishing a few dashboard presentation details.

The main changes are:

• Add a shared Tier regression script for result handling:

  • creates ci-results/run.log, ci-results/failure_summary.log, and ci-results/exit_code
  • centralizes the Tier 1 / Tier 2 regression execution logic
  • works with the existing if: always() upload steps so result artifacts are still uploaded after failures
  • checks command exit status, missing result directories, regression failure patterns, and final ISS trace status

• Simplify the Tier 1 / Tier 2 workflow YAML files:

  • replace duplicated inline shell blocks with calls to .github/scripts/run-tier-regression.sh
  • keep the existing Tier 1 / Tier 2 matrices unchanged

• Polish the Tier CI dashboard:

  • use the OpenHW logo from local dashboard assets
  • copy static dashboard assets during dashboard generation
  • rename the matrix legend entry from Not covered to Not in matrix

Motivation

The initial Tier CI flow worked, but the result handling logic was still embedded directly in the workflow YAML files and was duplicated between Tier 1 and Tier 2.

This PR makes the result handling more explicit, more robust, and easier to maintain. It also prepares the Tier CI flow for future extensions, since future Tier 1 / Tier 2 changes can reuse the same runner logic instead of duplicating shell code in multiple workflow jobs.

Validation

I validated the updated branch on my fork:

• Fork dashboard preview:
  https://alexchenic.github.io/cva6/

Validation notes:

• Tier 1 completed successfully.
• Tier 2 completed successfully.
• All uploaded Tier 1 / Tier 2 artifacts had exit_code=0.
• No failure_summary.log contained reported failures.
• I also checked ISS trace final status handling. Some raw logs contain transient *** FAILED *** lines followed by final *** SUCCESS ***; the final ISS status check correctly treats these as passing, and no final ISS FAILED status was found.

[JeanRochCoulon]

Did you pass Black and Pylint on your python files ?