Thinknode M6 Repeater Enable LEDs and Function Button

[AtlavoxDev]

What this fixes

On the original M6 repeater build:

• No visual LED feedback for power-on, power-off, or reset: the device gave the user no indication that a button press, shutdown command, or boot was recognized. Gave the impression of a dead device.
• No way to turn off the device: poweroff didn't actually sleep the chip. The original code called a SoftDevice API (sd_power_system_off()) that requires SoftDevice to be running — but SoftDevice isn't enabled in the repeater build, so the call silently failed and the chip stayed awake. No button mapped to poweroff.
• Function Button was disabled.: It was defined in variant.h but never read or armed as a wake source or poweroff.
• USB cable triggered power on: Charging the device would trigger the radio to boot up

What this adds

• Full LED choreography for every power event.
  • Boot: dual-LED full bright → 1 s gap → red solid + blue "loading-activity" flicker (To prevent user from attempting to press buttons during boot phase) → blue flash on boot complete.
  • Hold ~2 s: power off, with red blink ack at 0 s and 1 s + full-bright commit cue + dual-LED goodbye flash.
  • Shutdown via poweroff CLI: same shutdown cue.
• Robust SYSTEMOFF. New enterDeepSleep() helper with SoftDevice/direct-register fallback. Arms Function Button as SENSE-LOW wake source.
• Reset-cause-aware boot. RESETREAS captured in a priority-101 constructor (before Nordic's errata-136 workaround scrubs RESETPIN).
• Resilient deployment. GPREGRET2 user-intent flag distinguishes deliberate shutdown from transient resets. Survives SYSTEMOFF, cleared by full power loss. Result: device stays off when you asked it to, but self-heals from brownouts and auto-boots when a dead battery recovers.

Implementation notes

• All M6-only and #if defined(THINKNODE_M6)-guarded.
• Other M6 envs (room server, companion radio, kiss modem) inherit the SYSTEMOFF fix but not the repeater UX.
• analogWrite PWM-binds pins on the Adafruit nRF52 core — boot path use digitalWrite to keep blue compatible with the existing LoRa TX LED indicator.

Test plan

Verified on hardware:

• Hold ~2 s → power off, function button wake works
• poweroff over serial → real SYSTEMOFF
• Reset button → clean reboot
• USB plug-in to off device → stays off

Not field-verified:

• Dead-battery + solar recharge auto-boot (logic-tested only)

Out of scope

• Charging-status LED (needs EXT_PWR_DETECT / EXT_CHRG_DETECT and a wake-on-charge mode) — planned follow-up.

[AtlavoxDev]

Hey MeshCore team, this is my first ever attempt to contribute to a repo. Let me know if there are any questions or if there's something I should do differently next time. I only targeted the Repeater firmware this time, but will explore the other M6 firmware too for feature parity. Note I used Claude to assist and manually reviewed all code. Novice developer, wanted to take a stab at adding these critical features myself and contributing to the project. Tested over the past couple days.

[AtlavoxDev]

I moved as much as a could out of main.cpp. Only a couple small blocks inside main.cpp are minimal dispatch calls (board.bootComplete() and board.pollButton()) to enable boot monitoring and button functionality.

[AtlavoxDev]

I need to research the affects this has on other firmware builds for the boot LED indicator. I think this causes a regression with companion firmware builds, which would cause an indefinite LED flicker.