Force TX after CCA failure on EZSP v7+

[puddly]

EMBER_VALUE_FORCE_TX_AFTER_FAILED_CCA_ATTEMPTS seems to work in the latest Gecko SDK release! This allows the radio to force a transmit, even after CCA failures.

Normally this sounds like a bad idea but the alternative is to allow the radio to silently fail with a generic "delivery error" (TODO: figure out a way to infer the actual error code).

[codecov]

Codecov Report

Patch coverage: 100.00% and no project coverage change.

Comparison is base (f52d73e) 99.77% compared to head (e0ce7d9) 99.77%.

Additional details and impacted files

@@           Coverage Diff           @@
##              dev     #576   +/-   ##
=======================================
  Coverage   99.77%   99.77%           
=======================================
  Files          67       67           
  Lines        4837     4855   +18     
=======================================
+ Hits         4826     4844   +18     
  Misses         11       11           

Files Changed	Coverage Δ
bellows/ezsp/config.py	100.00% <100.00%> (ø)
bellows/ezsp/protocol.py	97.56% <100.00%> (+0.26%)	⬆️

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

[puddly]

It appears that this change is compatible with many more versions of EmberZNet than I thought: reading the value with getValue fails with EzspStatus.ERROR_INVALID_ID, but the write will still succeed. I've confirmed that the CCA bypassing works as far back as EmberZNet 7.1.1.0, and I'm sure previous releases too, meaning this PR will affect most modern EZSP adapters.

[puddly]

And with the adapter in a noisy environment, here are some traffic captures:

With CCA bypass (coordinator responds to beacon request)

---

Without CCA bypass (coordinator doesn't respond)

---

A device was only able to join (successfully might I add) in the first capture.

[MattWestb]

Great finding and work done our master HACKER !!!!