Fix: lis3mdl with two sensors not working

[hdiethelm]

Fix for #10740

lis3mdl was not initializing multiple sensors properly when started using:
lis3mdl start
The second sensor was just initialized but not started and there where some other bugs which where not visible at first glance.

Now, lis3mdl does the actions for all sensors when not called with -I, -X or -S.

I tested all commands with no argument, -X and -S. I was not able to test -I because I don't have an internal I2C sensor.

test_log.txt

[LorenzMeier]

Awesome, thanks!

[hdiethelm]

Note: This also fixes PREFLIGHT FAIL: MAG 1 SELFTEST FAILED in QGroundControl.

Might be there are other modules with the same problem. While searching for the problem, I found some reports about this but without a solution.

By the way: I had also PREFLIGHT FAIL: MAG SENSORS INCONSISTENT at the beginning, but the https://store.mrobotics.io/mRo-GPS-u-Blox-Neo-M8N-HMC5983-Compass-p/mro-gps003-mr.htm needs Yaw 90°. Just found it by comparing my mobile phone's compass with the compass in QGroundControl.

[dagar]

This needs to be merged into master. I'll update the PR base.

[dagar]

This doesn't apply cleanly to master. @hdiethelm could you rebase your commit to master and either change the pull request or open a new one?

[dagar]

PR needs to go into master, not stable.

[hdiethelm]

Just did a pull request for the master branch and squashed the tree commits in one. Someone knows how to change both "into" and "from" for a pull request?