P3-H P1: separate SV sampling and Ethernet frame cadence

[masarray]

Tracks #139 after merged #140 and #141.

Base: main@f3a107ec4cc4eca553ee467911c979afd057ce4d
Head: 213d31235b665556e339af82c147daa1ab4100b3

Purpose

Remove the remaining semantic ambiguity between sampling cadence and Ethernet publication cadence.

IEC 61850 SCL gives smpRate, smpMod and nofASDU; these are not one interchangeable "publisher rate". A stream may carry multiple consecutive ASDUs per Ethernet publication, so sample cadence and frame cadence must be represented independently.

Model

New SvPublicationTiming authority contains:

• configured sampling basis: SmpPerSec / SmpPerPeriod / SecPerSample;
• configured smpRate;
• ASDUs per Ethernet frame;
• optional nominal system frequency context;
• rational samples/s;
• rational frames/s;
• explicit resolution state.

Rates remain exact rational values; no floating point or truncating scheduler inference is used in the engineering model.

Standards / interoperability behavior

• omitted SCL smpMod compiles as the SCL default SmpPerPeriod;
• SmpPerPeriod remains Class B until nominal frequency is explicitly supplied;
• 80 samples/period + 50 Hz resolves to 4000 sample/s;
• 80 samples/period + 60 Hz resolves to 4800 sample/s;
• 256 samples/period + 8 ASDUs/frame @ 50 Hz resolves to 12800 sample/s and 1600 frames/s;
• 4800 sample/s + 2 ASDUs/frame resolves to 2400 frames/s;
• 14400 sample/s + 6 ASDUs/frame resolves to 2400 frames/s;
• SecPerSample is represented as a rational cadence rather than guessed or rejected by the host semantic model.

Device boundary remains narrow

This does not expand the ESP32-P4 runtime claim.

classify_esp32p4_sv_profile() still requires:

• SmpPerSec;
• one ASDU/frame;
• integral bounded frame rate;
• current proven 4I+4V / 64-byte layout;
• explicit counter modulus.

Thus host semantics become broader while deployment remains fail-closed.

Compatibility

Studio and the CLI preserve the existing publisherRate / publisherRateHz compatibility key, but it is now derived from frame cadence, not copied from smpRate.

The profile/inspection schema moves to v2 and additionally exposes rational sample/frame-rate metadata.

Explicit resolution path

Class B is no longer treated as synonymous with an unresolved counter.

• Studio exposes nominal system frequency only when an SmpPerPeriod profile has unresolved timing.
• Counter-modulus confirmation remains a separate control and appears only when counter policy is unresolved.
• The CLI accepts --nominal-frequency-hz N; no 50/60 Hz value is inferred from a vendor, filename, or profile shape.
• Studio displays configured Sampling and derived Frames separately so engineers can see ASDUs/frame instead of one ambiguous rate.

Tests

Regression coverage includes:

• 4800 / 1 ASDU;
• 4800 / 2 ASDU;
• 14400 / 6 ASDU;
• explicit and omitted SmpPerPeriod;
• 50/60 Hz nominal-frequency resolution;
• 256 SPC / 8 ASDU legacy measurement shape;
• SecPerSample rational cadence.

Non-goals

No multi-ASDU ESP packet builder, no multi-stream scheduler, no new IEC 61869-9 conformance claim, and no runtime wire protocol change in this PR.

Next node after this is accepted: profile-family authority + explicit device compilation from the timing model.

[chatgpt-codex-connector]

Codex Review Summary

This comment shows the latest Codex review activity on this pull request.

Review	Status	Commit	Review trigger
📝 Code Review	✅ Completed 2026-10-06T14:12:50.730213Z	175de44	Draft marked ready

ℹ️ About Codex in GitHub

Your team has set up Codex to review pull requests in this repo. Reviews are triggered when you

• Open a pull request for review
• Mark a draft as ready
• Comment "@codex review" or "@codex security review".

Codex reacts with 👀 while any review is running, comments if it has suggestions, and reacts with 👍 once all reviews finish with no findings.