dbSta: timing-accuracy report commands (AOCV/SI/PBA/CPPR/MCMM/crosstalk/noise/incremental-STA)

[saurav-fermions]

Summary

Adds the dbSta timing-accuracy command surface: a stacked series of report-only, flag-gated advanced-timing commands. Every command is additive — the default / flag-off paths are byte-identical to baseline report_checks / GBA timing.

Command surface:

• AOCV — depth-based OCV derate report (report_checks_aocv, aocv_derate_*) and depth-based derate applied during STA propagation (set_aocv_derate -propagate, default OFF).
• SI — Miller-factor coupling derate (report_coupling_si) and a timing-window-aware coupling derate (SiWindow).
• PBA — report_pba_slack (setup) + hold-path recovery + endpoint-closure decision.
• CPPR — clock reconvergence pessimism removal (report-only) + slice-2 endpoint closure with CRPR credit.
• Unified closure — report_closure.
• MCMM — multi-corner + cross-corner worst-slack, plus a per-mode dimension (report_mcmm_slack -by_mode).
• Crosstalk-aware delay — coupling-cap effective-C stage delay, per-edge direction, and crosstalk on hold paths (flag-gated).
• Noise — crosstalk-noise estimation (report_noise) and an optional noise-induced delay push.
• Incremental STA — additive, full-STA-equivalent, for the ECO/repair path.

⚠️ Dependencies (please read)

Engine hooks — OpenSTA #383. Two commits consume OpenSTA-fork hooks:

• AOCV propagation needs AocvDepthDerate / Search::deratedDelayData.
• SI timing-window needs the per-coupling-cap value setter (Parasitics::setCapacitorValue).

Both hooks ship in The-OpenROAD-Project/OpenSTA#383. CI here will fail until #383 is merged and this repo's src/sta submodule is bumped to include those hooks. This PR intentionally does not carry a src/sta gitlink bump. Related engine PRs: The-OpenROAD-Project/OpenSTA#384, The-OpenROAD-Project/OpenSTA#385.

Shared noise foundation. The crosstalk and noise commands build on a small, additive, report-only noise-estimation module (report_noise — NoiseState/NoiseVictimResult/noise_estimate_victim). It is included here as a foundation commit ("feat(dbSta): add report_noise crosstalk glitch triage").

Testing

Built the full openroad binary against src/sta carrying the #383 hooks. All 103 dbSta + rcx ctests pass (100%), including the 18 new feature regressions and the byte-identical-baseline checks (default/flag-off == baseline report_checks/GBA).

Notes

• Marked draft until the src/sta hooks land (see Dependencies).
• All commits additive / flag-gated / report-only; no default behavior change.

[gemini-code-assist]

Warning

Gemini encountered an error creating the review. You can try again by commenting /gemini review.

[saurav-fermions]

/gemini review

[gemini-code-assist]

Code Review

This pull request introduces several additive, report-only timing diagnostics and pessimism-recovery closure reports to dbSta. These features include Path-Based Analysis (PBA) pessimism recovery, Clock Reconvergence Pessimism Removal (CPPR) reporting, Multi-Corner Multi-Mode (MCMM) worst-slack reporting, a unified closure report composing PBA and CPPR, depth-based (AOCV-style) OCV derating, and incremental STA analysis to track affected fanout cones after ECO edits. Feedback on the pull request highlights a potential null pointer dereference in the incremental STA report module when retrieving the instance pin iterator, which should be defensively checked.

[gemini-code-assist]

The InstancePinIterator returned by network->pinIterator(inst) can potentially be nullptr if the instance has no pins or if the iterator creation fails. We should add a defensive null check before calling pin_iter->hasNext() and delete pin_iter to prevent a potential null pointer dereference.

    InstancePinIterator* pin_iter = network->pinIterator(inst);
    if (pin_iter != nullptr) {
      while (pin_iter->hasNext()) {
        const Pin* pin = pin_iter->next();
        if (network->isDriver(pin)) {
          Vertex* drvr = graph->pinDrvrVertex(pin);
          if (drvr != nullptr) {
            seeds.insert(drvr);
          }
        }
      }
      delete pin_iter;
    }

[saurav-fermions]

Thanks for the review (this is a draft pending the OpenSTA #383 hooks in src/sta).

On the InstancePinIterator null check at IncrementalStaReport.cc: dbNetwork::pinIterator(inst) returns new DbInstancePinIterator(...) — it never returns null (an instance with no pins just yields an iterator whose hasNext() is immediately false, which the loop already handles). No other pinIterator call site in dbSta null-checks it, so I've left it consistent with those rather than adding a guard that can't fire. Happy to reconsider if you'd prefer defensive checks added uniformly.