pkg/nordic_softdevice_ble: Use MAC48 as hardware address#10514
pkg/nordic_softdevice_ble: Use MAC48 as hardware address#10514cgundogan merged 1 commit intoRIOT-OS:masterfrom
Conversation
…to gnrc_netif/enh/iid-conversion
b68e4ae to
afbb774
Compare
|
Rebased to current master. |
|
I just noticed, that this PR isn't possible without #10521, so I'll shuffle around some dependencies tomorrow =D. |
863f1d7 to
3485467
Compare
|
Rebased and adapted for current #10521. |
3485467 to
0349528
Compare
|
Rebased to current master and dependencies |
0349528 to
4090be1
Compare
|
Rebased to current master. No longer depending on any PRs |
23e4652 to
f826de1
Compare
|
@kaspar030 do you have the hardware to test? I tested it now and it should work (@haukepetersen reported earlier offline that the softdevice did not work, but for me it worked in both this PR and in current master in communication with an Ubuntu |
|
@aabadie maybe you can test? |
|
@PeterKietzmann and I tried to use the softdevice but unfortunately, we couldn't get the bt0 interface up and running. In fact, it didn't even show up in the interface list. Nevertheless, there was one difference with this PR compared to master, though: Without this PR, we can connect to the RIOT device: and with this PR, the connection cannot be established, meaning, |
|
OK. I'll have a look. |
|
When on master I don't see the device using (neither do I see it in this branch). However, in this branch I can successfully establish a connection as described in the README and ping the node. |
|
Would have also surprised me because I did not change any behavior of the actual driver. Just how RIOT sees [edit]and handles addresses in interaction with[/edit] the soft device. |
|
Accidentally hit the wrong button ^^" |
|
Ah now on master when I am connected it shows me something. |
|
But so it does on this branch.... seems to me |
|
Ok, I did some research, but I'm not sure. The |
|
Any chance to get your hands on a lesser current version of Linux to test this ^^? |
cgundogan
left a comment
There was a problem hiding this comment.
Martine and I tested this PR on an ubuntu 16.04 with a 4.4 kernel. I can confirm that the address is shortened to 6 bytes and that pinging still works between linux <-> RIOT+softdevice.
ACK. Please squash!
f826de1 to
85ee1d2
Compare
|
Squashed |
|
@miri64 Murdock complains about doxygen |
This is just a compatibility issue waiting to happen as soon as there is support for a more standard-compliant implementation of BLE (like e.g. NimBLE ;-)).
85ee1d2 to
7f7bc8f
Compare
|
Fixed and directly amended. |
Contribution description
This is just a compatibility issue waiting to happen as soon as there is support for a more standard-compliant implementation of BLE (like e.g. NimBLE ;-)).
Testing procedure
gnrc_networkingshould still work withnordic_softdevice_ble(see README). I don't currently have the hardware at hand to test, so I pushed untested.Issues/PRs references
#10521 and should also help with @haukepetersen's NimBLE to GNRC integration work.