Fix MeshInterface::get_default_instance()

[SeppoTakalo]

Description

Fix MeshInterface::get_default_instance()
Both 6LoWPAN and Thread interfaces were returning and referencing wrong types. Pointers instead of reference, and objects where pointer should have been.

Also enable targets like MCR20A and NCS36510 to provide the default Nanostack RF driver.

Tested quickly on K64F+Atmel shield, K64F+MCR20 shield, MCR20A.

Requires following PRs to be usable:

• Allow driver to provide default NanostackRfPhy PelionIoT/atmel-rf-driver#65
• Allow driver to provide default NanostackRfPhy PelionIoT/mcr20a-rf-driver#18

Pull request type

[X] Fix
[ ] Refactor
[ ] Target update
[ ] Feature
[ ] Breaking change

[SeppoTakalo]

@kjbracey-arm @mikaleppanen Please review.

[kjbracey]

On this, I think (but am not totally certain) that "MCR20A" should be in "extra_labels". That seems to be where specific parts are named, whereas "device_has" is for general technologies.

[SeppoTakalo]

Hmm.. extra_labels means that it would pick up folder like TARGET_MCR20A but it would not be available as a build time flag.

device_has is a build time macro, but not visible in the mbed_config.h

[kjbracey]

@sg- suggests that this - driver providing an implementation - is the intended use of the new component, and that should work both ways. Worth a try?

[SeppoTakalo]

Removed the DEVICE_MCR20A flag and rebased on top of master

[SeppoTakalo]

For KW24D to be fully usable, it requires #7814

[AnttiKauppila]

This is targeted to 5.10.0-rc1
@kjbracey-arm can you review this again.

[0xc0170]

/morph build

[mbed-ci]

Build : FAILURE

Build number : 2889
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7778/

[0xc0170]

Please review KW24D build failures

[SeppoTakalo]

Rebased on top of master and dropped the KW24D changes.
Those are in #7814

[cmonr]

/morph build

[cmonr]

@SeppoTakalo Please take a look at the single linker failure.

[stevew817]

@SeppoTakalo If this impacts NCS36510, then targets providing SL_RAIL most probably need a similar update. Please don't leave us behind.

[SeppoTakalo]

My bad, the previous commit did not disable NCS36510 builds.
I had to remove the test config as well.

Now "should" build with IAR as well

+------+--------+-----------+------------+-------------+
| name | target | toolchain | static_ram | total_flash |
+------+--------+-----------+------------+-------------+
+------+--------+-----------+------------+-------------+


Build successes:
  * NCS36510::IAR::MBED-BUILD


Build skips:
  * NCS36510::IAR::MBED-OS-TESTS-NETSOCKET-DNS
  * NCS36510::IAR::MBED-OS-TESTS-NETSOCKET-TCP
  * NCS36510::IAR::MBED-OS-TESTS-NETSOCKET-UDP

It skips the build as it is supposed to.

[SeppoTakalo]

@stevew817 Good point,
Can you do this for the SL_RAIL as well
9836b9b

So targets that use 802.15.4 as a default, should provide NanostackRfPhy &NanostackRfPhy::get_default_instance() method.
Then once the device does:

	        "overrides": {
            "network-default-interface-type": "MESH"
        }

in the targets.json and has 802_15_4_PHY in the device_has list, it will provide mesh network as a NetworkInterface::get_default_instance() call.

[0xc0170]

/morph build

[mbed-ci]

Build : SUCCESS

Build number : 2963
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/7778/

Triggering tests

/morph test
/morph export-build
/morph mbed2-build

[mbed-ci]

Exporter Build : FAILURE

Build number : 2579
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7778/

[cmonr]

Exporter node got stuck, had to abort.

/morph export-build

[mbed-ci]

Exporter Build : SUCCESS

Build number : 2580
Build artifacts/logs : http://mbed-os.s3-website-eu-west-1.amazonaws.com/?prefix=builds/exporter/7778/

[mbed-ci]

Test : SUCCESS

Build number : 2715
Test logs :http://mbed-os-logs.s3-website-us-west-1.amazonaws.com/?prefix=logs/7778/2715