3.73 runs the same kernel. If there were bugs I would have expected 3.75 with a minor patch to address them.
It’s a mystery. Difficult to find a common thread.
Actually, that’s exactly what happened. Version 3.75 was updated just an hour later because the initial 3.75 release completely broke the Wi-Fi, including on my device. They quickly patched whatever was wrong and pushed out an updated 3.75 build. The Wi-Fi connection came back, but it’s still suffering from the current stability issues.
This patch was discussed here.
If I’m reading the repo right 3.5 was the last release based on Dunfell and then the project moved to Scarthgap - which is a big jump… I don’t follow the Linux WiFi space typically, but my AI friend highlighted significant churn to cfg80211, mac80211 and many of the specific device drivers along with highlight a few examples of regressions patched in 6.6. It looks like I need to setup a root account so I can login and dump some logs and things which I’ll try and do this weekend or next week.
At least I know I’m not the only one seeing some weirdness.
Follow the ssh guide for root access.
/var/log has what is available.
The path was 5.10.109 to 6.12.23 LTS to 6.12.90.
This remains today in venus 3.80~39.
I was away part of last week and over the weekend but since we got home I’ve been capturing logs a few times a day - oddly so far I haven’t seen the hang… I’ll keep capturing them in hopes of finding something.
What the logs do show is that while reconnecting ConnMan is crashing while trying to restore the default route - it looks like this happens when the signal is weak and there are frequent connect/disconnections. It also looks like Venus is using a pretty old version of ConnMan, v1.33 which is like 10 releases old now.