pkg/nimble/autoadv: make AD flag optional

[fjmolinas]

Contribution description

These fields can be omitted if all other FLAGS are 0 and the advertising
packets are not connectable, allowing for 3 extra bytes for advertisement
payload.

In an application I need every byte I can get out of the advertisement packet so It's useful/required to gain these 3 extra bytes.

Testing procedure

Flash with and without the flag set the FLAGS are then omitted from the payload:

CFLAGS=-DNIMBLE_AUTOADV_FLAG_FIELD=0 BOARD=nrf52840-mdk make -C examples/nimble_heart_rate_sensor/ flash term -j7

[HendrikVE]

Tested for my nrf52dk and the changes do as intended (after adding the missing C to the FLAGS at least ;P).

You are talking about saving three bytes, but I can only see two bytes that are missing. Where is the third byte hidden?

[benpicco]

Can we determine if the application is connectable based on the modules that are used?

[fjmolinas]

I'm not sure, because although it could not be advertising, it could still connect to another peripheral device, so I can't use svc_gap/svc_gat for this, and when doing info-modules I'm not sure I see anything relevant there.

[fjmolinas]

Tested for my nrf52dk and the changes do as intended (after adding the missing C to the FLAGS at least ;P).

You are talking about saving three bytes, but I can only see two bytes that are missing. Where is the third byte hidden?

Hmm I think its because nimble is then using AltBeacons where the ADFlags are encoded in only two bytes, while iBeacons use a LTV (length type value) encoding which is 3 bytes, I'll double check and update accordingly.

[fjmolinas]

Tested for my nrf52dk and the changes do as intended (after adding the missing C to the FLAGS at least ;P).
You are talking about saving three bytes, but I can only see two bytes that are missing. Where is the third byte hidden?

Hmm I think its because nimble is then using AltBeacons where the ADFlags are encoded in only two bytes, while iBeacons use a LTV (length type value) encoding which is 3 bytes, I'll double check and update accordingly.

Never mind this, I got confused myself when looking at the debug output, but the if you look at the raw payloads you can see that actually 3 bytes are saved, since LTV 0x020106 is actually gone compared with the second message.

[fjmolinas]

@HendrikVE I think your testing validates this PR since it was just confusion on both of us on how we looked at the nrfconnect log formatting, @benpicco can we move on without #16703 (comment) (I don't know how to address that)?

[HendrikVE]

@fjmolinas My bad, sorry for the confusion 😅 But yes, I had exactly the same output in my nrfconnect as the screenshots show.

[fjmolinas]

Rebase to get the tools-test fix in.