Re: Dealing with wifi driver crash that doesn't remove wlan interface

Bryce Johnson <[email protected]> Tue, 22 Jul 2025 07:36:35 -0600
Newsgroups dev.linux.lists.iwd
Message-ID <CADXxVSLXWPG7iV7p9OccmHztgA23y2KHuP3kcEisdDHdtiMiqg@mail.gmail.com>
Hi James,
Thanks for the response.
>
> I'm not sure there is much within IWD that can be done. I've run into
> similar problems and about all you can do is implement some kind of
> watchdog which monitors for these conditions (which sounds like you've
> done). The difficult part is detecting these conditions. I've even
> monitored journalctl and scraped IWD logs to gain more state info about
> IWD than what is advertised on DBus. This is actually why I added the
> l_notice "events" to IWD, the idea being that they are a consistently
> formatted, tool parse-able lines containing additional information, e.g.
>
> event: roam-info, bss: aa:bb:cc:dd:ee:ff, signal: -41, load: 9/255
>
Yeah I'm on watchdog v2 and thought it covered everything, but looks
like I need to add some more checks for this more rare case. (Only
seen once or twice so far)  For the most part I don't have direct
access to the devices, so I just get the logs, need to try a fix and
see if that reduces the issues.

Bryce