changes added for SPI CS defines as requested for BMI055 driver integration (e.g. GPIO_SPI1_CS_PORTE_PIN15 for bmi055 gyro) to avoid double declaration of same chip select pin

format check

Avoid dropouts, we have enough RAM there.

This will upgrade systems to the new .ulog format used by http://logs.px4.io and supported by Flight Plot

- fixes #6974

In case when the baudrate change failed, we should still try that
baudrate again in the next iteration.

This prevents an exception happening inside the serial stack.

   sched_note_{suspend|resume} were calling hrt_absolute_time before
   it hrt_init is called. This can lead to register access before
   clocking is enabled. The result is a hardfault.

 - Fixed V712
 - The compiler can optimize this code by creating an infinite loop, or simply deleting it.
 - There is need to add a volatile qualifier to the '_ExitFlag' and 'sim_delay' variables.

The FW_PSP_OFF parameter no longer breaches the pitch min/max limits

Fix error in scaling of population high frequency vibration metrics
Add histograms for delta angle and velocity bias data
Fix variable descriptions

Add assessment of IMU bias and mag field estimation
Reduce warning false positives by adjusting thresholds and eliminating use of peak value plots for output observer monitoring
Clear each figure after saving to reduce memory usage

Miscellaneous bug fixes and improvements including:

More conservative reporting of velocity and position accuracy when aiding is lost
Separate reporting of accuracy for local and global position
Detection and recovery from badly conditioned accel bias states and covariance values
Logic to prevent high manoeuvre levels causing bad acceleration bias learning
Reduce sensitivity of covariance prediction to timing jitter
Continue reporting of GPS quality until the later of airborne or checks pass
Add ability to perform wind estimation for multi-rotors using a specific force drag model

 Ultimate we want this changes that is being backed out herein.
 But it is breaking things because it returns the EINVAL
 when there is a rate overlap. So the rest of the pwm ioctl
 calls then fail and do not set the pulse widths on arming.

 As a secondary issue. We sould call up_pwm_servo_init() to
 establish the PWM channel allocation early. This then allows
 FMU::set_pwm_rate to properly check for improper rate request
 not isolate to one group (timer).