drivers/kw2xrd/Kconfig: fix kconfig model

[MrKevinWeiss]

Contribution description

Nightlies are currently failing as there is a difference between make
and kconfig.
This tries to match the kconfig with the makefile dep.
The only issue is the

ifneq (,$(filter netdev,$(USEMODULE)))
  USEMODULE += netdev_ieee802154_submac
endif

which may have the same effect as select
HAVE_IEEE802154_RADIO_HAL_INTERFACE.

Testing procedure

NIGHTLY=1 /bin/bash -c "source .murdock; JOBS=4 compile tests/ieee802154_hal pba-d-01-kw2x:gnu" should pass.

Issues/PRs references

Follow-up from #18469

[MrKevinWeiss]

Like I said, maybe the following should also be in

diff --git a/boards/pba-d-01-kw2x/Kconfig b/boards/pba-d-01-kw2x/Kconfig
index 8068c000ca..5307eb8a25 100644
--- a/boards/pba-d-01-kw2x/Kconfig
+++ b/boards/pba-d-01-kw2x/Kconfig
@@ -28,3 +28,4 @@ config BOARD_PBA_D_01_KW2X
     select HAVE_MPL3115A2
     select HAVE_TCS37727
     select HAVE_TMP006
+    select HAVE_IEEE802154_RADIO_HAL_INTERFACE

If @jia200x could have a quick look and confirm?

[benpicco]

Since those should already be provided by the submac netdev, the removal here is correct either way.

[MrKevinWeiss]

I also added this board to the kconfig test in the .murdock so it won't be just nightlies that check it.

[MrKevinWeiss]

Just waiting for an opinion on HAVE_IEEE802154_RADIO_HAL_INTERFACE since there is a visible difference in the makefile.dep

[benpicco]

Why would a board select this? HAVE_IEEE802154_RADIO_HAL_INTERFACE is a property of the radio driver.

[MrKevinWeiss]

I thought the radio is tied to the board... @jia200x is in now and I can discuss.

[benpicco]

The radio is tied to the board, but HAVE_IEEE802154_RADIO_HAL_INTERFACE should be tied to the radio, as it's done with cc2538, nrf802154 and socket_zep.

[MrKevinWeiss]

Maybe the MODULE_KW2XRF selecting it makes more sense. Ok, pushed the fixup.

[MrKevinWeiss]

I guess it looks good. I squashed and removed the build limiting.

[MrKevinWeiss]

Thanks for the review!