Fix GPS capability poll blocking for 500ms every 5 seconds

[sensei-hacker]

Summary

Fixes a bug where gpsProcessNewSolutionData was not called for 500ms every 5 seconds due to unnecessary blocking in the u-blox capability polling loop.

Problem

In gpsProtocolStateThread, the periodic MON-VER and MON-GNSS polls (every GPS_CAPA_INTERVAL = 5s) waited on _ack_state for ACK/NAK:

pollGnssCapabilities();
ptWaitTimeout((_ack_state == UBX_ACK_GOT_ACK || _ack_state == UBX_ACK_GOT_NAK), GPS_CFG_CMD_TIMEOUT_MS);

In the u-blox protocol, MON-class messages respond with data, not ACK/NAK (only CFG-class messages produce ACK/NAK). So the condition was never satisfied and ptWaitTimeout always hit the full 500ms (GPS_CFG_CMD_TIMEOUT_MS) timeout. During this window the protothread was blocked and could not loop back to call gpsProcessNewSolutionData.

At 10 Hz GPS, this means ~5 position fixes go unprocessed every 5 seconds.

Fix

Remove the ptWaitTimeout calls entirely. The receiver thread parses MON-VER/MON-GNSS responses asynchronously regardless of whether the state thread waits, and nothing between the polls and the next loop iteration depends on the updated data. The GPS_CAPA_INTERVAL (5s) guard already prevents re-polling too frequently.

This matches ArduPilot's approach to capability polling — fire-and-forget, with responses handled asynchronously as they arrive.

Testing

• SITL build compiles cleanly
• Code reviewed before PR
• Note: this was discovered by breadoven during testing of PR Position estimator sensor correction refactor #11322 (position estimator refactor)

Changes

• src/main/io/gps_ublox.c — Remove blocking ptWaitTimeout calls from periodic MON-VER and MON-GNSS polls in gpsProtocolStateThread main loop

[qodo-code-review]

ⓘ You are approaching your monthly quota for Qodo. Upgrade your plan

PR Compliance Guide 🔍

All compliance sections have been disabled in the configurations.

[breadoven]

A quick test shows this fixes the problem. Updates occur at the 10 Hz rate for the GPS used with no noticeable delay, Hopefully the glitch should be gone especially for multirotors which suffered from a regular 5s nodding twitch at speed.