CI and Testing

This document defines the continuous-integration pipelines, smoke synthesis, functional regression, benchmarking, and gating thresholds for RVSvKit, consistent with the Overview and Module Inventory.

1. CI Pipelines (in `ci/`)

Pipeline	Tool	Trigger	Purpose
verilator.yml	Verilator	PR / push to `main`	Functional lint & simulation
icarus.yml	Icarus Verilog	PR / push to `main`	RTL simulation with testbenches
vivado_docker.yml	Vivado (Docker)	Nightly / scheduled	Synthesis smoke, area & timing measurements

Each pipeline reports pass/fail status, code coverage, and synthesis metrics via badges in README.md.

2. Functional CI

Regression tests:
- Run all testbenches under modules/**/tb/ using Verilator & Icarus.
- Gate on 0 new assertions, no regressions, and ≥ 95% line coverage per module.
Coverage collection:
- Verilator generates coverage data (.gcov).
- Coverage thresholds defined in ci/thresholds.json.

3. Synthesis CI

Smoke Synthesis:
- Quick synth of each module wrapper (in ci/smoke/) at default parameters.
- Capture utilization: LUTs, FFs, BRAMs, DSPs.
Baseline thresholds:
- Defined in ci/thresholds.json (e.g., adder: LUTs ≤ 50, multiplier: LUTs ≤ 200).
- Fail CI if any metric exceeds its threshold.
Timing checks:
- Report worst negative slack (WNS) at typical PVT corner.
- Gate on WNS ≥ −1 ns margin.

4. Benchmarking & Metrics

Historical logging:
- CSV files under ci/benchmarks/ record date, module name, LUTs, FFs, WNS, coverage.
- PRs that change resource usage must include updated CSV entries.
Trend visualization:
- Scripts in scripts/sweep_params.py can plot resource trends over time.

5. Additional Quality Gates

Linting:
- Enforce SV style via svlint matching 03-coding-conventions.md.
Static Timing Analysis (STA):
- Full P&R run on examples/simple_soc scheduled weekly.
Security & Dependency Checks:
- Scan for unapproved DPI imports.
- Check Docker images for CVEs.

Packages and Interfaces Automation Optimization