This decouples the cyclic refreshing of reachability from the manual reading of the LQI tables:
Changes the reachable refresh to only poll once a minute, cycling devices that need polling. Also checking for reachable will now only read the LQI table. Also removes the coordinator node from the cyclic checks.
In case a manual refresh is triggered, both LQI and RTG tables will be read of all non-sleepy nodes, including the coordinator and without any delays in between polls.
tested this for while now. It seems still good enough and takes away quite a bit of load from the network. Additionally it gets away with the annoying warning message in the NXP backend.
This decouples the cyclic refreshing of reachability from the manual reading of the LQI tables:
Changes the reachable refresh to only poll once a minute, cycling devices that need polling. Also checking for reachable will now only read the LQI table. Also removes the coordinator node from the cyclic checks.
In case a manual refresh is triggered, both LQI and RTG tables will be read of all non-sleepy nodes, including the coordinator and without any delays in between polls.