Skip to content

coord: synth v0.50 (13 skips) + REQ-PIX-007 → implemented (both halves proven) + gale k_msgq kernel-seam need - #165

Merged
avrabe merged 1 commit into
mainfrom
coord/synth050-reqpix007-galemsgq
Jul 23, 2026
Merged

coord: synth v0.50 (13 skips) + REQ-PIX-007 → implemented (both halves proven) + gale k_msgq kernel-seam need#165
avrabe merged 1 commit into
mainfrom
coord/synth050-reqpix007-galemsgq

Conversation

@avrabe

@avrabe avrabe commented Jul 23, 2026

Copy link
Copy Markdown
Contributor

Architecture-sync tick — the one genuinely-unblocked feature (grounding the crossing on gale's verified k_msgq) turned out gale-kernel-seam-gated, so this reconciles the state instead.

STEP 1 — release-watch

synth v0.50.0 on falcon-v1.127 = 14 → 13 skips. v0.50 is mostly deps + verified-codegen infra (VCR-WASM-001 i64 exec wiring; RV32 RA validator), not falcon-lane float fixes; the incidental −1 is an i64 op. My #345 repro isn't actioned yet (3× #345 remain). run-stabilization still lowers. Gate still open (13>0). AFD-035 refreshed.

STEP 2 — REQ-PIX-007 draft → implemented

Both halves now have green CI-gated evidence:

  1. WIT-typed contract is now structured records (VehicleState.nav = 16 f32, ActuatorCmd.pwm = 4 f32) — TEST-PIX-023.
  2. RPC cross-core over shared SRAM — the 16×f32 NavState crosses M4→SHMEM→M7 byte-exact — TEST-PIX-030.

Remaining to verified: ground the transport on gale's verified k_msgq.

STEP 3 — gale msgq integration: found blocked, recorded (not forced)

gale-wasm-msgq-v0.4.0-cortex-m4f.o has 9 undefined scheduler-seam symbols (gale_w_msgq_pend/ready_thread/reschedule/thread_swap_data/unpend_first_thread/…) — the gale kernel, which the v0.4.0 release does not ship as standalone objects. So consuming k_msgq needs gale to also expose the kernel-seam .o (or jess links the fused gale-nano core). Recorded on the DD-023 GALE row as the concrete gale ask to file when jess can bind (jess is float-gated, not ready-to-bind — per the DD-022/023 sequencing rule).

rivet validate → PASS.

🤖 Generated with Claude Code

…es proven) + gale k_msgq kernel-seam need recorded

Architecture-sync tick (no unblocked feature — the gale msgq integration is gale-kernel-seam-gated).

STEP 1 (release-watch): synth v0.50.0 on falcon-v1.127 = 14 -> 13 skips. v0.50 is mostly
deps + verified-codegen infra (VCR-WASM-001 i64 exec; RV32 RA validator), NOT falcon-lane
float fixes; the incidental -1 is an i64 op. #345 (my minimal repro) not yet actioned.
run-stabilization still lowers. Gate STILL OPEN (13 > 0). AFD-035 refreshed.

STEP 2 (architecture-sync): REQ-PIX-007 draft -> IMPLEMENTED. Both halves now have green
CI-gated evidence: (1) the WIT-typed contract is now STRUCTURED records (VehicleState.nav =
16 f32, ActuatorCmd.pwm = 4 f32; TEST-PIX-023) and (2) the NavState crosses M4->SHMEM->M7
byte-exact (TEST-PIX-030). Remaining to `verified`: ground the transport on gale's verified
k_msgq.

STEP 3 (gale msgq — found blocked, recorded): the gale-wasm-msgq-v0.4.0-cortex-m4f.o has 9
undefined scheduler-seam symbols (gale_w_msgq_pend/ready_thread/reschedule/... = the gale
kernel), which the v0.4.0 release does NOT ship as standalone objects. So consuming k_msgq
needs gale to also expose the kernel-seam .o (or jess links the fused gale-nano core). RECORD
not file (jess is float-gated, not ready-to-bind) — added to the DD-023 GALE row as the
concrete gale ask to file when jess can bind.

rivet validate PASS.

Co-Authored-By: Claude Opus 4.8 <noreply@anthropic.com>
@avrabe
avrabe merged commit 58190c7 into main Jul 23, 2026
5 checks passed
@avrabe
avrabe deleted the coord/synth050-reqpix007-galemsgq branch July 23, 2026 14:58
Sign up for free to join this conversation on GitHub. Already have an account? Sign in to comment

Labels

None yet

Projects

None yet

Development

Successfully merging this pull request may close these issues.

1 participant