thinknode_m6: don't drive GPS REINIT pin so the L76K is detected

[benskigomez]

Problem

On the ThinkNode M6, the L76K GPS was never detected — the companion app reported "no GPS / not supported", and on-device the GPS stayed at sat 0. The same sealed unit's GPS works fine under Meshtastic, so the hardware is good.

Root cause

Pin 29 (PIN_GPS_RESET, which is the module's REINIT line on this board) was being driven HIGH, which holds the L76K silent — it never emits any NMEA, so detection (Serial1.available()) sees zero bytes and the GPS setting is hidden.

It was driven HIGH in two places:

• variants/thinknode_m6/variant.cpp initVariant() drove it HIGH at boot.
• Because PIN_GPS_RESET was defined, MicroNMEALocationProvider::begin() also drove it HIGH (GPS_RESET = 29, GPS_RESET_FORCE = LOW, so begin() writes !LOW = HIGH).

The Meshtastic M6 variant leaves this exact pin as a floating input and never drives it — which is why GPS works there.

Verification (no logic analyzer; M6 is a sealed unit)

A standalone sketch powered the GPS and passively captured NMEA at 9600 baud while changing one variable at a time:

pin 29 state	result
driven HIGH (current firmware)	0 bytes
left floating (INPUT)	full NMEA stream ($GNGGA/$GNRMC/$GNGSA/$GNGSV)

An EN power-cycle was tried as well and made no difference — floating pin 29 alone is sufficient.

The GPS streams standard NMEA at 9600 by default and needs no $PCAS/config commands.

Fix

• variant.h: define GPS_RESET (-1) so the location provider never touches pin 29.
• variant.cpp: stop driving pin 29; leave it floating.

This matches the Meshtastic M6 variant. Confirmed on hardware: GPS is now detected and the app shows live coordinates, altitude, and map position.

Fixes #2862.

Note for maintainers

The same pattern may affect other Elecrow ThinkNode boards whose reset pin is labeled "REINIT" — notably the M3 (variant.h has PIN_GPS_RESET (25) // REINIT), which has a matching open report (#1864). I don't have M3 hardware to verify, so it's intentionally left out of this PR.

[benskigomez]

Likely the same bug on the ThinkNode M3 (#1864). For a maintainer with M3 hardware:

variants/thinknode_m3/variant.h has #define PIN_GPS_RESET (25) // REINIT (note the same "REINIT" labelling as the M6), and variant.cpp sets it as an OUTPUT. As on the M6, MicroNMEALocationProvider::begin() then drives that pin HIGH, which on the M6 is exactly what held the L76K silent. The open issue #1864 ("thinknode m3 doesnt have gps") matches the M6 symptom.

If that reproduces on an M3, the same one-line fix should apply: stop driving pin 25 / set GPS_RESET (-1) so the provider leaves it floating. I don't have an M3 to verify, so I've left it out of this PR.