@sera, a note on your last patch-set. Did a build and flashed both the ubi and squash images on my v1 to the same outcome I previously reported; random reboot, cause unknown, as of yet no single discernible cause. Have not been able to collect any relevant log data as it has never flushed at an opportune time to my remote linux box. I suppose instrumenting the build is an option but then observer effect ...
Might take a shot at the 4.8 kernel this weekend.
Edit: On another note, regarding my previous observation about the jump in cpu utilization on my mamba. About the same time as the last mwlwifi release, I was modifying my build and had changed from dnsmasq to dnsmasq-full, as I wanted to possibly put dnssec into play. Right now that is looking like the likely suspect, but why an order of magnitude uptick, especially since it is not yet actually doing ipsec? Might try backing that out to prove one way or ta'other.
Edit2: Doesn't look like that was the issue. An interesting aside, where both cpu's have been ~20% from when I first noticed the change, one is now ~18%, one ~16%. So I'm guessing the hard float pushes from a day or two back had a discernible impact.
(Last edited by Villeneuve on 31 Aug 2016, 21:12)