Maybe @A-Paul wants to look over the unittests, too. But be aware that I wrote them from the stand-point that one at least knows the make-up of a IEEE 802.15.4 frame header (little-endian, so first byte/bit in table is least significant), which should be expected if you write a driver for a network device for that protocol ;-):

where FCF is a bitfield of the following format

For the specific values of the FCF fields please see the ieee802154.h

Rebased and squashed

@aeneby #4645 (comment): if you want those you need to provide them... My knowledge about IEEE802.15.4 is not sufficient enough for that :/

@authmillenon something like:

#define IEEE802154_MIN_FREQ         2405 /**< Min. frequency (2405 MHz) */
#define IEEE802154_MAX_FREQ         2480 /**< Max. frequency (2480 MHz) */

#define IEEE802154_MIN_CHANNEL      11   /**< Min. channel (2405 MHz) */
#define IEEE802154_MAX_CHANNEL      26   /**< Max. channel (2480 MHz) */
#define IEEE802154_CHANNEL_SPACING  5    /**< Channel spacing in MHz  */

And if you want conversion macros:

#define IEEE802154_CHAN2FREQ(chan)  ( IEEE802154_MIN_FREQ + ((chan) - IEEE802154_MIN_CHANNEL) * IEEE802154_CHANNEL_SPACING )
#define IEEE802154_FREQ2CHAN(freq)  ( IEEE802154_MIN_CHANNEL + ((freq) - IEEE802154_MIN_FREQ) / IEEE802154_CHANNEL_SPACING )

Isn't this excluding Sub-GHz 802.15.4 radios?

@authmillenon @aeneby
These are some other bands which we need to support as well:

780-786 MHz (China)
868.3 MHz (EU)
906-924 MHz (North America)
Ultra wide band:
249.6 MHz to 749.6 MHz
3.1 GHz to 4.8 GHz
6.0 GHz to 10.6 GHz

And something around 950 MHz as well

Sorry, but that's the point where I say, that this is way to complicated for a "simple" helper macro ;-)

Alright, I guess I'll have to read the spec regarding the other bands too. I still think it's a good idea to have these defined eventually, though. Otherwise they have to be defined per driver.

@authmillenon said offline, that I should write: 'this should be simlified' -> so, 'this should be simplified' :-)

Well I just quoted you ad.lib. ;-)

Reworked based on discussion at H&A with @haukepetersen (but did not adapt all unittests yet)

Addressed @thomaseichinger's comments in #4645 (comment) and #4645 (comment)

Someone care to review?

@OlegHahm you maybe since you already did some heavy reviewing in the dependent PRs?

I didn't do any review, I just tested.

Rebased to master and squashed.

@authmillenon I'd like to review. Will go through it in the next days.

Rebased to current master.

@authmillenon I'd like to review. Will go through it in the next days.

@kaspar030, found time?

@kaspar030 can I have a drink to celebrate tonight too? :P

Looking through this I can't find anything to complain about, so ACK from my side (once Travis agrees of course).

Rebased and squashed.

ACK from my side, too. This doesn't touch any existing code, and RIOT CI looks good, so feel free to merge!

I thought we say we prefer still Travis for this week.

I thought we take RIOT CI's decisions with a grain of salt and actually take a look at it's result.
This one looks fine. Your call.

Travis is done now to, so merge ;-P