drivers/sensors: Legacy sensor warning

[linguini1]

Summary

This commit implements a compile-time warning and in-code comment warning for legacy sensor drivers. This is intended to:

• Warn users that legacy drivers may eventually be removed
• Warn developers that they should not use a legacy driver as a reference for their new driver contributions

Closes #18434

Impact

All legacy drivers have a compile-time warning.

Ideally, this makes it easier for contributors to be aware of what to use as a reference.

Testing

N/A: only a comment and #warning directive was added, so if the CI passes all is well.

[acassis]

@raiden00pl maybe qe, zerocross and other "sensors" that are not really a sensor and doesn't make sense to be in drivers/sensors/ should be moved to other place. See, qe in some way it similar to pulse capture that is in drivers/timers/ so maybe we could move it to there. The zerocross could be there too or in drivers/misc/

Or other alternative is creating a drivers/sensors/misc/ or other better name and move it to there. BTW we can do it during the 13.x release

[acassis]

@linguini1 @raiden00pl I think the support to fixed math to sensors should come with 13.x, so in that in this new release "legacy" drivers will be removed (when its count-part uORB exists) or converted.

@jerpelea seems like we have a lot of work before 13.x be released :-/

[linguini1]

@linguini1 @raiden00pl I think the support to fixed math to sensors should come with 13.x, so in that in this new release "legacy" drivers will be removed (when its count-part uORB exists) or converted.

I think getting the fixed math support for 13.0.0 is a good idea, but I don't know when we'll be able to retire the legacy drivers. I don't have any of those devices to test with/convert to uORB.

@jerpelea seems like we have a lot of work before 13.x be released :-/

Yeah....but we can do our best! Maybe some things will need to wait for a 14.0.0 :)

[raiden00pl]

The initial fixed-math implementation is ready, and porting subsequent sensors can be done one by one. The only problem is that I haven't had a chance to finally verify it on HW yet so it's still marked as draft :)

[linguini1]

The initial fixed-math implementation is ready, and porting subsequent sensors can be done one by one. The only problem is that I haven't had a chance to finally verify it on HW yet so it's still marked as draft :)

I have a limited set of sensors available to me (IMU, mag, baro and GNSS) but I would be willing to help perform some testing for it! I have an RP2040-based flight computer from my L1 rocket flight, and I would be very curious to see the performance improvement on it without having to use floating point for uORB!