This sounds more like a network problem than a phone problem. What makes you think it is a phone problem? My parents have a mesh network - it works fine on my 10 III.
It worked perfectly fine with sfos 4.5 and now it doesn’t. I also read in oneof the 5.x release threads about this problem. It works perfectly fine with my laptop and my old Nexus5 running a very old Lineage version.
What do you mean by “Switch to another ap”? Are you switching this on the phone or are you walking near a different access point?
I have a mesh network and it is 100% transparent to any device* - if anything, I would attempt to resolve from the mesh’s perspective.
(*) With 1000s of quirks, wireless technology doesn’t scale very well and the standards aren’t very well respected.
I’d advise to check the logs of your mesh and report to the manufacturer.
I also recognize this problem since update from 4.6 to 5… Do you also have a FritzBox Router? Similar problem has occured before (years ago), problem back then was that the router was set to automaticly change between different bandwidth
For me it works when I set my fritzbox wifi to WPA2 mode. If I use WPA2&WPA3 mode the C2 has problems. But it is not alone. My heat pump (Daikin) has similar problems.
Could be connected to connman GitHub - sailfishos/connman: Sailfish OS fork of https://01.org/connman . Seems that was updated 10 months ago to 1.38. Not sure why they picked a 5+ year old release instead of the recent one, maybe 1.42 at the time. But maybe that is not what ends up on our devices.
Looking at connman upstream there are many jolla contributions even in recent day code.
FWIW, I use WPA3 Personal with an Asus mesh and works perfectly fine. In order for us to establish whether this is SFOS or not, I wonder if the device running Android behaves differently when connected to the problematic mesh.
There are many historical reasons to this because ours is a fork that has diverged a lot. And upstream develops constantly, so we need to be quite careful in upgrading to newer versions. We decided to take it in steps, to first go to 1.38 that seemed to be a stable one at that time when we updated. Aim is to keep updating but history has dragged lots of future conflicts to be part of this process.
Lots of upstream content has been cherry-picked into our fork, so we are somewhere between 1.38 and 1.42 at the moment. But we will keep updating it, getting closer to recent upstream.
ConnMan, wpa_supplicant and the kernel used on the device have had a lot of updates after that. If you’d enable some logging on wpa_supplicant that might help. You could set the journalctl to use permanent logging first, like having an additional conf file at: /etc/systemd/journald.conf.d/debug.conf with content:
and systemctl restart systemd-journald after creating (or removing) the config. then use wpa_cli to increase logging level:
wpa_cli log_level DEBUG
EXCESSIVE, MSGDUMP, DEBUG, INFO, WARNING, ERROR being other options. You can check the initial log level with wpa_cli log_level which you might want to set after testing.
journalctl -b -u wpa_supplicant > wpa_supplicant.log gets you all the logs gotten with the log_level but it will contain all SSID and MACs of the networks you use, so, you could strip them off first. Or just use something, like journalctl -b -u wpa_supplicant|grep -i -e error -e warn -e err -e fail > wpa_supplicant.log.
I canged my FritzBox Setting to no longer change WLAN Frequencies between 2,4 and 5 Ghz - now WLAN uses 2,4 Ghz permanently. My first impression is, that it solved the problem for me - my Xperia 10iii does not loose connection anymore when moving around the house (Mesh with 3 Powerline Senders)
The issue doesn’t occur for me with a fritzbox mesh with 1 repeater, and dynamic band selection. I have crossover band enabled, and have a xperia 10 II, which might make a difference.