[META] probe-rs + LPC-Link2 LPC deploy roll-up

[zackees]

Scope

This is the focused roll-up for the fbuild-side NXP LPC deploy work that intersects probe-rs, LPC-Link2 / CMSIS-DAP, lpc21isp, and the LPC845-BRK flash -> monitor handoff.

Created from an audit of fbuild issues matching probe-rs, LPC845-BRK, LPC8xx, CMSIS-DAP, LPC-Link2, and lpc21isp.

The broader LPC8xx support tracker is #586. This issue is narrower: keep the probe-rs / LPC-Link2 deploy path coherent, avoid stale split-out tickets, and make sure the remaining fbuild integration actually lands.

Current source of truth

Touchless SWD flashing of the physical LPC845-BRK with the stock LPC-Link2 CMSIS-DAP v1.0.7 firmware was verified in #935 / #936 using the FastLED probe-rs fork:

• FastLED/probe-rs#1: DAP_Info fallback hardening
• FastLED/probe-rs#2: explicit SWD DAP_Connect plus retry
• probe-rs (fork): replace hidapi with nusb for CMSIS-DAP v1 HID transport — the actual fix for touchless LPC-Link2 deploy #936: nusb-v1 HID transport work, with hidapi fallback still the winning Windows path after a physical USB replug restored HidUsb state

That means the active fbuild work is no longer "prove probe-rs can ever flash this board." It is now "ship, integrate, and clean up the stale probe-rs/LPC tickets."

Active tracked work

• probe-rs on LPC-Link2 v1.0.7: blocker #4 — SerialNumber + remaining DAP_Info fallbacks (#930 follow-up) #931 - Audit remaining DAP_Info fallback work. If probe-rs LPC-Link2 v1.0.7: push through all blockers end-to-end, single-shot #935/probe-rs (fork): replace hidapi with nusb for CMSIS-DAP v1 HID transport — the actual fix for touchless LPC-Link2 deploy #936 fully supersede it, close it with the exact proof links. If not, rewrite it to name the one remaining query/failure signature instead of keeping the old broad blocker text.
• probe-rs on LPC-Link2 v1.0.7: SWD-attach layer quirks (#930 follow-up) #932 - Audit SWD-attach quirks after the verified touchless flash. If FastLED/probe-rs#2 plus the probe-rs LPC-Link2 v1.0.7: push through all blockers end-to-end, single-shot #935 proof satisfies it, close it as superseded. If not, update it with the current failing command/trace.
• cross-compile + host FastLED/probe-rs fork (#930 follow-up) #933 - Cross-compile and host the FastLED/probe-rs fork artifacts, or replace this with the current release mechanism. The deliverable needs provenance, checksums, and a repeatable way for fbuild to fetch the right binary.
• wire fbuild LpcDeployer to prefer probe-rs when CMSIS-DAP probe enumerates (#930 follow-up) #934 - Wire fbuild LpcDeployer to use probe-rs when a CMSIS-DAP probe is present, with explicit fallback to lpc21isp and a force/opt-out flag. This is the main remaining fbuild code path.
• [META] LPC845-BRK on-hand burn-down: cover the open LPC8xx work end-to-end #586 - Update the broad LPC845-BRK meta so it reflects the closed state of the older deploy issues and links here for the probe-rs/LPC-Link2 path.

Closed context to preserve

• probe-rs on LPC-Link2 CMSIS-DAP v1.0.7: series of upstream quirks to patch through #930 - Original probe-rs CMSIS-DAP v1.0.7 umbrella; split into probe-rs on LPC-Link2 v1.0.7: blocker #4 — SerialNumber + remaining DAP_Info fallbacks (#930 follow-up) #931-wire fbuild LpcDeployer to prefer probe-rs when CMSIS-DAP probe enumerates (#930 follow-up) #934 and closed.
• probe-rs LPC-Link2 v1.0.7: push through all blockers end-to-end, single-shot #935 - Hardware-validation umbrella; touchless probe-rs flash was verified end-to-end and the issue closed.
• probe-rs (fork): replace hidapi with nusb for CMSIS-DAP v1 HID transport — the actual fix for touchless LPC-Link2 deploy #936 - probe-rs transport/fork issue; closed after the verified flash path and Windows driver-state diagnosis.
• fbuild deploy: own nxplpc/LPC8xx flash end-to-end (FastLED is removing its direct pyocd path) #921 - fbuild-owned LPC deploy path issue; closed as redundant after the lpc21isp deployer work was found to already be merged.
• nxplpc: LpcDeployer feeds -hex to a .bin firmware and passes CMSIS-DAP speed=1000 as the lpc21isp baud #927 - Fixed the immediate lpc21isp .bin/baud issues in LpcDeployer.
• LPC845-BRK: deploy must always unwedge the shared VCOM (pyocd/SWD flash leaves CDC in Windows error 31) #565 - VCOM unwedge acceptance; closed 2026-07-07. Re-check this only if making probe-rs the default SWD deploy path reintroduces a flash -> monitor failure.
• LPC8xx Stage 7: deploy support (lpc21isp / pyocd / cmsis-dap / jlink) #551 - Original LPC8xx deploy-support tracker; closed predecessor to fbuild deploy: own nxplpc/LPC8xx flash end-to-end (FastLED is removing its direct pyocd path) #921/wire fbuild LpcDeployer to prefer probe-rs when CMSIS-DAP probe enumerates (#930 follow-up) #934.
• LPC8xx Stage 6: FastLED canary — Blink/DemoReel100/FxEngine compile for lpc845/lpc804 #550, [META] LPC8xx end-to-end support: orchestrator + Arduino framework + real FastLED sketches #487, Native support for NXP LPC8xx (LPC845-BRK, LPCXpresso804, LPCXpresso845-MAX) #513, Stage 2: implement NXP LPC8xx build orchestrator (replace stub, unstub LPC804/LPC845 CI) #477, lpc8xx: convert 5 stub workflows to real builds (Stage 8 of #487) — orchestrator + upstream proof are in #528 - prior LPC8xx build/framework/CI support issues; useful background but not blockers for this focused deploy roll-up.
• [META] FastLED #3339 LPC bring-up lessons — fbuild structural follow-up across 4 axes #699 - Closed fbuild structural follow-up from FastLED LPC bring-up lessons; background only unless a regression points back to serial/deploy architecture.

Acceptance

This meta is done when:

• probe-rs on LPC-Link2 v1.0.7: blocker #4 — SerialNumber + remaining DAP_Info fallbacks (#930 follow-up) #931 and probe-rs on LPC-Link2 v1.0.7: SWD-attach layer quirks (#930 follow-up) #932 are either closed as superseded by probe-rs LPC-Link2 v1.0.7: push through all blockers end-to-end, single-shot #935/probe-rs (fork): replace hidapi with nusb for CMSIS-DAP v1 HID transport — the actual fix for touchless LPC-Link2 deploy #936 or rewritten with current, reproducible remaining work.
• cross-compile + host FastLED/probe-rs fork (#930 follow-up) #933 has produced a usable probe-rs artifact channel, or has been replaced by an issue that names the actual artifact source fbuild should consume.
• wire fbuild LpcDeployer to prefer probe-rs when CMSIS-DAP probe enumerates (#930 follow-up) #934 is implemented with tests for probe-rs dispatch, lpc21isp fallback, and the force/opt-out path.
• fbuild deploy tests/platform/lpc845brk -e lpc845brk --port COM10 can use the probe-rs path when the LPC-Link2 CMSIS-DAP probe is present and still falls back clearly when it is not.
• [META] LPC845-BRK on-hand burn-down: cover the open LPC8xx work end-to-end #586 is updated so the broader LPC tracker no longer carries stale deploy status.

Do not close this just because #935/#936 are closed; those proved the hardware path. This meta closes only after the remaining fbuild integration and issue hygiene are actually complete.

[zackees]

Closeout

All active children in this focused probe-rs + LPC-Link2 roll-up are resolved.

• probe-rs on LPC-Link2 v1.0.7: blocker #4 — SerialNumber + remaining DAP_Info fallbacks (#930 follow-up) #931: closed after the current FastLED probe-rs path reached SWD attach and COM10 reopened.
• probe-rs on LPC-Link2 v1.0.7: SWD-attach layer quirks (#930 follow-up) #932: closed as superseded by the verified SWD attach + touchless flash path.
• cross-compile + host FastLED/probe-rs fork (#930 follow-up) #933: closed by Install FastLED probe-rs for LPC SWD deploy #1001. fbuild now pins and verifies the latest FastLED probe-rs release, fastled-v0.31.2-nusb-v1-transport, with per-host artifact names and SHA256 digests.
• wire fbuild LpcDeployer to prefer probe-rs when CMSIS-DAP probe enumerates (#930 follow-up) #934: closed by Install FastLED probe-rs for LPC SWD deploy #1001. LpcDeployer now prefers probe-rs SWD for the LPC-Link2 CMSIS-DAP probe, falls back to lpc21isp, and supports fbuild deploy --no-probe-rs.
• [META] LPC845-BRK on-hand burn-down: cover the open LPC8xx work end-to-end #586: updated with the final status in [META] LPC845-BRK on-hand burn-down: cover the open LPC8xx work end-to-end #586 (comment) and left open for broader LPC work outside this focused probe-rs/LPC-Link2 slice.

Merged PR: #1001 (cc319808).

Validation on merged origin/main:

• soldr --no-cache cargo build -p fbuild-cli -p fbuild-daemon passed.
• FBUILD_DEV_MODE=1 .\target\x86_64-pc-windows-msvc\debug\fbuild.exe deploy tests\platform\lpc845brk -e lpc845brk --port COM10 --verbose exited 0 on the attached LPC845-BRK.
• The dev daemon log shows the probe-rs SWD path: chip=LPC845M301JBD48, probe=1fc9:0132, firmware=tests\platform\lpc845brk\.fbuild\build\release\firmware.elf.
• No fallback-to-lpc21isp line appeared in the merged-main validation log.
• COM10 reopened successfully after deploy; the dev daemon was stopped afterward.

Closing this focused meta as complete.

[zackees]

Follow-up audit/meta opened as #1002 after concern that this focused roll-up may have closed before all probe-rs/LPC follow-ups were independently reconciled. Use #1002 as the active focused tracker for the probe-rs + LPC deploy audit.