vtol_att_control move to WQ with uORB callback scheduling

[dagar]

Requires #12207 before merging.
Background - #12207

[dannyfpv]

tested on Pixhawk4 mini V5 v-tol
Modes Tested

mission Mode: no issue
Altitude Mode: no issue
Stabilized Mode: no issue
land mode: no issue
pr log:
https://review.px4.io/plot_app?log=59aa9861-a0c8-4f15-aec5-61e3e39be905
master log:
https://review.px4.io/plot_app?log=1c38da3d-622f-4c2b-b526-e1435c0867e7

[dagar]

@PX4/testflights could you give this one another go?
@RomanBapst @sfuhrer if you're flying VTOL in the near future could you give this a quick flight as well?

Other than potential regressions the thing I'd like to review is cpu and memory usage (vs master as of a few days ago) and end-to-end latency for both the FW and MC actuators.

[sfuhrer]

@PX4/testflights could you give this one another go?
@RomanBapst @sfuhrer if you're flying VTOL in the near future could you give this a quick flight as well?

Other than potential regressions the thing I'd like to review is cpu and memory usage (vs master as of a few days ago) and end-to-end latency for both the FW and MC actuators.

@dagar Tested on our FunCub VTOL (Pixhawk4, quadplane). Flew MC/FW in stab and pos. Felt as always. log:
https://review.px4.io/plot_app?log=2bfa950a-80b7-45d2-bf22-b581c683a067

[dagar]

Thanks @sfuhrer. I did a quick comparison with 2 other logs from the same vehicle (testing #12626)

• https://logs.px4.io/plot_app?log=6ee671da-d7ce-463b-9483-17ecbddd0585
• https://logs.px4.io/plot_app?log=c510e280-078d-480b-8b5a-34188a8bdec6

Differences I noticed.

• overall cpu usage down slightly (maybe a couple percent)
• px4io control latency (MAIN outputs) about the same
• px4fmu control latency (AUX outputs) down significantly (> 4000 us to ~700 us)
• memory usage reduced by somewhere around 17 kB (this is a huge chunk for older boards)