don't use the 1kHz timer as it conflicts with other I2C devices

this allows for support of multiple sensors on a board

setup the timer start after we've started the conversion to prevent
reading faster than the MS5611 is able to produce values

this was seen on the PXF board. It isn't yet known why it happens

disabled on APM2 to save flash space

we don't need the I2C MS5611 driver

prevent waits for samples

thanks to Mike McCauley for pointing this out

we switched to GPLv3 a long time ago, but neglected to update the
per-file license headers

- Allows use of hardware floating point on the Cortex-M4.
- Added "f" suffix to floating point literals.
- Call floating point versions of stdlib math functions.

- The MPU6000 holds on to the I2C semaphore for quite some time during
  init, which caused a panic when the MS5611 is also on I2C.

- Guard I2C transactions with this semaphore in the MS5611 and
  HMC5843 drivers.

- Preparation for removing begin/end atomic.

this isn't needed as the common variables are already protected by
cli()/sei()

this changes the barometer calculations to floating point. On a MS5611
this is actually about twice as fast as the previous 64 bit
calculations, but gains us more accuracy as we are able to take
advantage of sub-bit precision when we average over 8 samples.

this returns how many samples were used to calculate the last pressure

this allows us to use the MS5611 barometer at its full 100Hz sample
rate (80Hz for pressure, 20Hz for temperature). The pressure and
temperature values are averaged between reads without adding any
latency. Previously the driver would throw away values between
readings

This also fixes a race condition in reading from the SPI bus that
could lead to bad values from the barometer

this allows the barometer driver to calibrate and return altitude and
climb rate values. This will be used by the AHRS drift correction code
for vertical velocity

The climb rate uses a 5 point average filter