MAV_CMD_DO_SET_MODE - add support

[hamishwillee]

This adds support for MAV_CMD_DO_SET_MODE, which replaces SET_MODE (though I have not removed support for that)

Literally all it does is publish the command, which is all that the SET_MODE message hander does in https://github.com/PX4/PX4-Autopilot/blob/main/src/modules/mavlink/mavlink_receiver.cpp#L920-L946 too (after you factor out conversions).

I have tested this and the mode can be changed. An ACK is received.

@julianoes would appreciate your advice here. The ACK is always accepted, even though it should be denied in some cases.
For example, I tried to trigger the default path here by passing an auto mode with an invalid submode - in this case: mode: 29, custom_mode: 4,custom_sub_mode: 20 (20 is greater than the mode.

When I run logging I can see that this is triggered, and main_ret = TRANSITION_DENIED;

However further down this is tested to set the return value:

			if (main_ret != TRANSITION_DENIED) {
				cmd_result = vehicle_command_ack_s::VEHICLE_CMD_RESULT_ACCEPTED;

			} else {
				cmd_result = vehicle_command_ack_s::VEHICLE_CMD_RESULT_TEMPORARILY_REJECTED;
			}

At this point main_ret has magically changed -1, which is why this is accepted. I can't see it being set - advice?

Upshot, I think my changes are good, but there is something wrong with the existing code.

[hamishwillee]

@julianoes This now passes the formatting tests and build tests.

Would be great if you could help me debug why it isn't returning the right ACK, as per #24704 (comment) (specifically, the code detects that a mode can't be sent, and then ACKs anyway).

[dakejahl]

bump, what is the impetus? Should we merge?

[hamishwillee]

The blocker is that it returns the wrong ACK in the pre-existing checker code #24704 (comment)

I was hoping for debug help from @julianoes to get this in since I can't see how it can be being set.

I'll have another look though.

[julianoes]

How do I reproduce the problem?

[hamishwillee]

@julianoes Actually this does work. I was confused. I can set modes quite happy. It will return accepted for modes you might not expect, but I'm pretty sure that's because the base modes are valid.

[hamishwillee]

Do you think we might get this in v1.16 @julianoes - or is that really heading towards stable?

[julianoes]

Do you think we might get this in v1.16 @julianoes - or is that really heading towards stable?

I wouldn't know. If this is tested and low risk, then just make a backport pull request and tag me.

[hamishwillee]

Thanks! Done #25334