Bug 271188 - iwlwifi with lagg added to bridge will not associate to AP; and a network restart may result in a system crash
Summary: iwlwifi with lagg added to bridge will not associate to AP; and a network res...
Status: Closed Overcome By Events
Alias: None
Product: Base System
Classification: Unclassified
Component: kern (show other bugs)
Version: 13.2-STABLE
Hardware: Any Any
: --- Affects Only Me
Assignee: Bjoern A. Zeeb
URL:
Keywords: crash, regression
Depends on:
Blocks: iwlwifi
  Show dependency treegraph
 
Reported: 2023-05-01 22:53 UTC by js
Modified: 2023-10-28 18:03 UTC (History)
3 users (show)

See Also:


Attachments

Note You need to log in before you can comment on or make changes to this bug.
Description js 2023-05-01 22:53:22 UTC
As described in bug #271182 one can add iwlwifi to lagg but must restart network after boot in order to associate to an AP.  However, if you also add your lagg device to a bridge, you can no longer associate to an AP even after restarting network.  This was not a problem in 13.1-STABLE from late December 2022 but is now a problem with 13.2-STABLE. Furthermore, multiple network restarts result in a kernel panic.

ifconfig wlan0 scan results in valid AP list zero or more times, but at no point does wlan0 associate with AP.  As soon as scan shows no AP's, it will not show AP's again. 'service netif restart' may result in scanning working again, or it may result in a kernel panic.

Relevant rc.conf entries:

kld_list="if_iwlwifi i915kms"
ifconfig_re0="ether c4:75:ab:f3:f9:6c"
wlans_iwlwifi0="wlan0"
ifconfig_wlan0="WPA"
cloned_interfaces="lagg0 bridge0 tap0"
ifconfig_lagg0="laggproto failover laggport re0 laggport wlan0 DHCP"
ifconfig_bridge0="addm lagg0 addm tap0"
Comment 1 js 2023-08-20 18:52:19 UTC
Has anyone been able to take a look at this?  I'm happy to do more testing if there's anything else I should try.
Comment 2 Graham Perrin 2023-09-23 10:10:21 UTC
(In reply to js from comment #0)

> … multiple network restarts result in a kernel panic. …

Do any of these match your case? 

<https://bugs.freebsd.org/bugzilla/buglist.cgi?bug_status=__open__&keywords=crash&keywords_type=allwords&list_id=642784&longdesc=iwlwifi&longdesc_type=allwordssubstr&order=bug_id&query_format=advanced>

If no match, and if still reproducible with the latest stable/13: 

* for kernel panics, a separate bug report may be appropriate

– then we can review the summary and metadata here. 

----

^Triage (interim): 

* status
* component kern, for a kernel panic
* keyword crash, for a system crash (kernel panic)
* keyword regression, for the second sentence in opening comment 0.
Comment 3 js 2023-09-24 21:55:26 UTC
I upgraded to latest stable commit today (stable/13-n256387-7af7a0d1b14 AMD64 amd64) and re-tested.  Everything is now working as expected.  Thank you!
Comment 4 Bjoern A. Zeeb freebsd_committer freebsd_triage 2023-09-25 10:56:30 UTC
Can you please report back in a week or so if still everything is fine.  From my side nothing has changed since the last reports (possibly not from you) came in, so I am suspicious as to what has changed to make it work now...


[ please keep PR open ]
Comment 5 js 2023-09-26 15:56:05 UTC
When I ran mergemaster yesterday, I happened to notice the following change:
https://github.com/freebsd/freebsd-src/commit/052211e08c0e227277d0c4dc603bba2253eb3d73

Could this have been the problem, and removing the wpa_poststart function fixed it?
Comment 6 Bjoern A. Zeeb freebsd_committer freebsd_triage 2023-10-25 21:22:30 UTC
(In reply to js from comment #5)

It should not have been the cause but a possible extra interface toggle could have confused lagg?  Someone with more knowledge of today's lagg code would probably know.

Has the problem come back or stayed "solved"?
Comment 7 js 2023-10-27 02:00:35 UTC
(In reply to Bjoern A. Zeeb from comment #6)
This problem has not happened again.  iwlwifi on 13-STABLE has been rock solid for me over the past month.
Comment 8 Bjoern A. Zeeb freebsd_committer freebsd_triage 2023-10-28 18:03:17 UTC
Thanks and enjoy and please let us know if the problem comes back.