LSC1-08: use fixed scalar RESULT byte mux

[Th0rgal]

Scope

This is exactly one bounded LSC1-08 slice from 79d926462cbaf51a2c87c3fa112527dd28f610aa.

• replace only the scalar RESULT serializer's variable offset/index selection with explicit fixed byte offsets for the zero-write, deferred-write, and one-write shapes
• preserve the existing scalar semantic staging, payload lengths/bytes, CRC path, handshakes, RETIRE/BUSY behavior, and BLAKE3 serializer
• retarget the existing scalar byte-selection and stable-staging mutation anchors without weakening their behavioral kills
• update SHA256SUMS for the two changed tracked sources

Verification on exact clean head

Semantic verification head: d1be00a6d6abe5de30fa6d8daeedd7ee1983dfc2

Evidence publication head: 83f864aa399ba4a60996ab9dd44740a0c1957ada (receipts/checksum metadata only; no RTL semantic change)

All required lanes passed:

• make -C test/packet_frontend sim
• make -C test/packet_frontend mutation
• make -C test/packet_frontend differential-mutation
• make -C test/packet_frontend lifecycle-mutation
• PYTHONDONTWRITEBYTECODE=1 PYTHONPATH=. python3 -m unittest sim.test_packet_frontend_rtl_differential -v (17 tests)
• make lsc1-authored-rtl-contract
• make lsc1-host-authored-rtl-boundary
• make sim
• make check (379 simulation/model tests plus the remaining repository gates)
• make lean (build and required mutation checks; existing linter warnings unchanged)
• make checksum-check

The full LSC1 netlist-assurance lane also passed using a private cache outside the checkout. The complete machine-readable receipt is now committed at evidence/lsc1-08-s3/full-lsc1-netlist-receipt.json, and its digest is pinned by the repository SHA256SUMS:

• receipt status: pass
• receipt SHA-256: b0afb9a5a826e8a468f3bae65991c624abb01fa3796b889fc854e8cc76872a9e
• generic netlist SHA-256: e2656fa0093bfb8ad5634639ce7aab3914a6655abfc592224a85e842145c84c0
• checkout head recorded by receipt: d1be00a6d6abe5de30fa6d8daeedd7ee1983dfc2
• 3-edge whole-design bounded check: pass
• observable correspondence mutation: killed
• classification: bounded-pass-unbounded-blocked
• unbounded blocker: the 15-second whole-design temporal-induction attempt expired; no unbounded correspondence claim is made

Generic Yosys structure observation

The committed machine-readable Yosys structure receipt records the Yosys 0.68 banner, pinned OSS CAD Suite CI setup/manifest digest, exact commands, source SHAs, log digests, warnings, and full before/after cell-type trees. Its summarized generic proc; flatten; opt; fsm; opt; memory; opt; check; stat counts are:

Structure	Base	Head
wires	4,795	4,732
wire bits	136,681	134,587
cells	4,440	4,381
$mux	2,721	2,659
$shiftx	28	18
$sub	22	16

Both runs report seven warnings. The existing ext_word latch warning remains because ext_word is still used by the untouched BLAKE3 serializer branch.

These are unmapped generic structure counts only. This PR makes no PPA, mapped-area, timing, netlist-unbounded, P&R, FPGA, or hardware claim, and it does not establish universal cycle equivalence or arbitrary ready/valid liveness.

---

Note

Medium Risk
Changes packet frontend RESULT byte serialization on the TX path; extensive sim, mutation, and bounded netlist checks mitigate protocol regressions, but this is still correctness-critical RTL.

Overview
LSC1-08 replaces the scalar RESULT TX path (tx_external_kind == 3) in lsc1_packet_frontend.sv: variable tx_payload_index math and indexed scalar_staged_access[] reads become nested case statements with fixed per-byte slices for zero-write, one-write, and deferred payload shapes. Three wires (scalar_staged_access_0–2) expose staged access words for those explicit muxes; BLAKE3 serialization is unchanged.

Assurance: .gitignore whitelists evidence/lsc1-08-s3/*.json; checked-in full LSC1 netlist and Yosys structure receipts document bounded-pass verification and slightly lower generic $mux/$shiftx counts. test/packet_frontend/Makefile scalar-result-mutation anchors target the new 17: tx_payload_external_data = scalar_staged_write_value[7:0]; line. SHA256SUMS updated for touched artifacts.

^{Reviewed by Cursor Bugbot for commit 83f864aa399ba4a60996ab9dd44740a0c1957adaa399ba4a60996ab9dd44740a0c1957ada. Bugbot is set up for automated code reviews on this repo. Configure here.}

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_04fbd5dd-378a-4c3d-a505-d816119267e0)

[cursor]

Bugbot couldn't run - usage limit reached

Bugbot is counted against Cursor usage for this user or team, and this run hit a usage or spend limit.

A user or team admin can review and increase usage limits in the Cursor dashboard.

(requestId: serverGenReqId_a012715b-d2ee-47ad-b2d0-d7999db464ce)

[Th0rgal]

@codex review

Fresh exact-head review requested for 83f864aa399ba4a60996ab9dd44740a0c1957ada.

Evidence-only repair receipts:

• full LSC1 netlist receipt: evidence/lsc1-08-s3/full-lsc1-netlist-receipt.json
• full receipt SHA-256: b0afb9a5a826e8a468f3bae65991c624abb01fa3796b889fc854e8cc76872a9e
• Yosys 0.68 before/after structure receipt: evidence/lsc1-08-s3/yosys-structure-receipt.json
• structure receipt SHA-256: cfa734602cf69a09a9d7d2a5219e59dc4ec89e73afbf9007b00b0b2d24c0de8c
• semantic RTL head recorded by both receipts: d1be00a6d6abe5de30fa6d8daeedd7ee1983dfc2
• current head changes only receipt/checksum metadata; no RTL semantic change

Immediately before this request, local HEAD, remote branch HEAD, and PR HEAD matched exactly; the worktree was clean and make checksum-check passed.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.

[Th0rgal]

@codex review

Fresh exact-head review requested for 83f864aa399ba4a60996ab9dd44740a0c1957ada — CI run 32019487089 is now fully terminal green (48/48 substantive checks success, 1 neutral usage-limited bot). The prior 10:18:32Z request for this same head returned usage-limited without producing a review. Head is unchanged since.

[chatgpt-codex-connector]

You have reached your Codex usage limits for code reviews. You can see your limits in the Codex usage dashboard.
To continue using code reviews, add credits to your account and enable them for code reviews in your settings.