GitLab

Minor fix on bullet point in new features list

1f127905

Colibri race specific code into the mpu6500 driver.

Merge branch 'feature/make-top-level-test' of https://github.com/Echelon9/cleanflight into Echelon9-feature/make-top-level-test

Remove duplicated -mfpu=fpv4-sp-d16 under cortex-m4 based targets.

Fix upper bound clamping of floating point PID settings upon MSP read

Just reverse the min/max arguments.

Unit tested.

Fixes #1132.

Failsafe - Add rxfail `auto` mode.

Failsafe - Fixes for RX loss detection.

code.

Tested with X8R in SBus and PWM, and Spek Sat, GR-24 PPM, PWM and SUMD,
Spek PPM

channels, not just aux channel.

ports, previously only softserial was working. - See #1021

Fix small typos in README.md

Merge branch 'rx_sumd_added_crc_check' of https://github.com/ProDrone/cleanflight into ProDrone-rx_sumd_added_crc_check

Conflicts:
	src/main/rx/sumd.c

RX - Corrections to RX fail detection, hold and preset, when using a PWM connection

Because a required change for PWM disturbed the PPM mode.

Add ability to get/set mag calibration data

Modified VBatt functionality to enable better precision.

RX - SUMD channels are received as 16 bit values instead of 32

Officially the CRC check is part of the SUMD protocol.

Framing errors are already covered and i expect these to occur around
the same time as CRC errors, so i am not sure if someone will ever
notice the difference...

Logic for valid flightchannel detection is inverted in order to detect
the first flightchannel failure instead of waiting to check them all.

Clear PWM channel capture on read.

This invalidates the control channels on read. They are validated by
receiving a good value BEFORE the aux channels are received. This is
done because control channels configures to go OFF on failsafe are
detected with a PWM capture time-out. Time-out took so long that all aux
channels where overwritten by their RX configured failsafe values BEFORE
the invalid (=OFF) control channel was detected.

Preserve 32 bytes of RAM.

GPS speed was 10x too big on frsky telemetry