- Feb 01, 2017
-
-
Siddharth Bharat Purohit authored
-
- Jan 31, 2017
-
-
Beat Küng authored
Fixes the following case: a board that has several bus_options enabled, and wants to start a specific device (eg mpu6000 -T 20608 start). In that case the given device_type is never actually checked, and since the start routine iterates over all bus_options, several can match, and thus start multiple instances of the driver (note that the whoami check is performed against the value provided with -T and thus the check will pass). This happens on Pixracer (PX4_SPIDEV_MPU and PX4_SPIDEV_ICM_20608 are both set).
-
Beat Küng authored
This is what's used according to https://pixhawk.org/modules/pixracer.
-
Lucas De Marchi authored
It uses a shared I2C bus with MS65611 which causes noise on the baro reads. This will rely on the external compass instead of the internal one.
-
- Jan 30, 2017
-
-
Daniel Agar authored
-
Daniel Agar authored
-
Daniel Agar authored
- fixes #6469
-
Daniel Agar authored
-
Daniel Agar authored
-
Lorenz Meier authored
Snapdragon RC driver: Move RC struct out of main loop to preserve channel values. Initialize raw rc count.
-
Lorenz Meier authored
-
- Jan 29, 2017
-
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
This is to ensure that if sensor IDs are fixed we do not end up in a state where the system is partially calibrated and the need for re-calibration is not properly communicated to the user.
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
Before this change the MPU6K driver would register only the MPU6K family, but not the sub-type, which prevented telling individual sensors apart. This is a breaking change because users will have to perform their accel and gyro calibration again. However, it is unavoidable since right now the different sensors can end up with the same ID and the wrong offsets can be applied to the wrong sensor.
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Lorenz Meier authored
This implementation is a baseline implementation and makes no attempt to be tamper-proof. A monotonic counter like the W25R64FV or a similar HW facility would be required to achieve this.
-
Lorenz Meier authored
-
Lorenz Meier authored
-
Khoi Tran authored
-
Khoi Tran authored
-
Lorenz Meier authored
-
Daniel Agar authored
-
Daniel Agar authored
-