broadcom wifi broke on antiX21 / debian?

[Member kernelpanic]

hello antiXers, I wish you all a happy new year!

I know there is already an extensive thread about wifi with connman on antix19.
my problem however is not limited to connman or a conflict between connman and ceni (I think),
so I open this new thread.

my machine: old (~13 years) acer singlecore celeron laptop with 1 MB RAM.

after booting antiX21 32bit full (sysvinit) on USB life: no wlan connection possible.
installed the system anyway, because I thought I could solve the problem once installed.
still no connection to my home-wlan (and others) possible.

wifi shows up with correct SSID, WPA2-key is challenged.
seems to connect, systray and conky show a wlan-connection.
after a short while connman cmst in system tray shows “network services: the system is offline”,
connects again after a few seconds, cmst: “network services: the system is online”.
seems to hold connection for about a minute, then the next disconnect appears.
no internet connection for a browser (firefox, dillo) in the short time-interval of “pseudo”-connection.

a thorough reading of thread “connman on antix19” didn`t end up in a working wifi,
however many thanks to dolphin_oracle and Xecure for their video tutorials!

network devices/drivers:

$ lspci vnn
04:00.0 Network controller [0280]: Broadcom Inc. and subsidiaries BCM4311 802.11b/g WLAN [14e4:4311] (rev 01)
	Subsystem: AMBIT Microsystem Corp. BCM4311 802.11b/g WLAN [1468:0422]
	Flags: bus master, fast devsel, latency 0, IRQ 17
	Memory at f8000000 (32-bit, non-prefetchable) 
	Capabilities: <access denied>
	Kernel driver in use: b43-pci-bridge
	Kernel modules: ssb

$ inxi -nz
Network:   Device-1: Broadcom NetLink BCM5787M Gigabit Ethernet PCI Express driver: tg3 
           IF: eth0 state: down mac: <filter> 
           Device-2: Broadcom BCM4311 802.11b/g WLAN driver: b43-pci-bridge 

so correct b43 is loaded.

tried to get connected with connman: no success
made sure that /etc/network/interfaces had only the 2 lines in it, as Xecure suggested:
<blockquote>
auto lo
iface lo inet loopback
</blockquote>

made sure with rfkill list that no hard/soft blocks exist:

$ rfkill list
0: acer-wireless: Wireless LAN
	Soft blocked: no
	Hard blocked: no

as suggested by Xecure:
/etc/connman/main.conf edited to read “Enable OnlineCheck = false”
or alternatevly: (with OnlineCheck enabled):
OnlineCheckIPv6URL= http://connman.net/online/status.html

after every change:

sudo service connman restart

no success!

tinkering with network-assistant to unload/blacklist the b43 driver and change to the proprietary “wl”
was not successful, with wl loaded no wifi was detected at all, so I returned to b43, first blacklisted
wl, then purged it completely.

still no success.

tried to connect with ceni:

Enter preshared key: ********************
ifup: configuring interface wlan0=wlan0 (inet)
run-parts --exit-on-error --verbose /etc/network/if-pre-up.d
run-parts: executing /etc/network/if-pre-up.d/ethtool
run-parts: executing /etc/network/if-pre-up.d/wireless-tools
run-parts: executing /etc/network/if-pre-up.d/wpasupplicant
wpa_supplicant: wpa-driver nl80211,wext (default)
wpa_supplicant: /sbin/wpa_supplicant -s -B -P /run/wpa_supplicant.wlan0.pid -i wlan0 -D nl80211,wext -C /run/wpa_supplicant
Starting /sbin/wpa_supplicant...
wpa_supplicant: waiting for "/run/wpa_supplicant.wlan0.pid":  0 (max. 5)
wpa_supplicant: creating sendsigs omission pidfile: /run/sendsigs.omit.d/wpasupplicant.wpa_supplicant.wlan0.pid
wpa_supplicant: ctrl_interface socket located at /run/wpa_supplicant/wlan0
wpa_supplicant: configuring network block -- 0
wpa_supplicant: wpa-ssid "here is my ssid" -- OK
wpa_supplicant: wpa-psk ***** -- OK
wpa_supplicant: enabling network block 0 -- OK

CLIENT="-i";  /sbin/dhclient -4 -v $CLIENT -pf /run/dhclient.wlan0.pid -lf /var/lib/dhcp/dhclient.wlan0.leases -I -df /var/lib/dhcp/dhclient6.wlan0.leases wlan0 	
Internet Systems Consortium DHCP Client 4.4.1
Copyright 2004-2018 Internet Systems Consortium.
All rights reserved.
For info, please visit https://www.isc.org/software/dhcp/

Listening on LPF/wlan0/00:1c:26:a1:7a:6b
Sending on   LPF/wlan0/00:1c:26:a1:7a:6b
Sending on   Socket/fallback
DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 5
DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 10
DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 11
DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 15
DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 15
DHCPDISCOVER on wlan0 to 255.255.255.255 port 67 interval 5
No DHCPOFFERS received.
No working leases in persistent database - sleeping.
run-parts --exit-on-error --verbose /etc/network/if-up.d
run-parts: executing /etc/network/if-up.d/000resolvconf
run-parts: executing /etc/network/if-up.d/ethtool
run-parts: executing /etc/network/if-up.d/mountnfs
run-parts: executing /etc/network/if-up.d/wpasupplicant

Press Enter key to continue ...

so still no success!

just out of curiosity:
– downloaded and created antiX21 full as 64bit (CPU is 64bit capable) on USB, booted, same outcome.
– downloaded and created on USB MX21 fluxbox edition just to see how this debian flavour would work
(they do it with network-manager instead of connman):
wlan shows up, challenges for password, tries to connect, the activity symbol in systray rotates for a while,
then nothing is happening, no connection.
– downloaded and created on USB Lubuntu: no wlan connect, did not investigate further here.

ok, back to antiX21 32bit:
set repositories to “testing”, upgraded all just to make sure that broadcom driver-/firmware was updated to newest versions.

still no success!

installed 2 additional kernels
– (original is 4.9.0-279-antix.1-486-smp)
– older: 4.4.0-264-antix.1-486-smp
– newer: 5.10.57-antix.1-486-smp

both kernels: same outcome as standard kernel

I already suspected a hardware problem with the broadcom-chip, however:
wlan in antiX17 (installed in another partition of the same machine) works perfect!!!

after all that troubleshooting I think the problem must be somewhere in the debian base
for the broadcom BCM4311 chip (driver/firmware/configuration issue).

after this odyssee my nerves are hanging out of the window 😉 …
completely out of ideas now … any help would be heavily appreciated!

• This topic was modified 1 year, 4 months ago by kernelpanic.
• This topic was modified 1 year, 4 months ago by BobC.
• This topic was modified 1 year, 4 months ago by BobC. Reason: Wifi very slow with solution

[Member stevesr0]

Helpful

Up

1

::

Perhaps comparing the dmesg output for the working and nonworking installs will provide hints?

stevesr0

• This reply was modified 1 year, 4 months ago by stevesr0.

[Moderator BobC]

Helpful

Up

1

::

Which version of antiX did you download and are you booting it with a flashdrive, DVD, frugal or installed?

PS: I realize this is a limited scope workaround, but I have a Dell D620 with BRCM4311 and the same issue, and what I’ve found is that if I install to my drive without messing with it, it works after the reboot when installed. I’m not sure if that helps your situation or not.

• This reply was modified 1 year, 4 months ago by BobC.

[Member stevesr0]

Helpful

Up

1

::

Addendum to previous post: use grep and tail with dmesg – EX. sudo dmesg | tail an sudo dmesg | grep “xyz” (xyz = bcm or device or disconnect or …)

Second thought:

Do you have more than one bcm43xx driver active -which can cause conflict. On the internet various threads note additional modules such as brcmsmac and bcma which may also be loaded.

Check lsmod |grep bcm and lsmod | grep brcmsmac. Again comparing the output between the working install and the nonworking install might show what is different.

If you find an “extra” module, see if it is reported to conflict with the bcm43-pci-bridge and remove it and see if things work (WITHOUT rebooting).

If the working system uses another module than the b43-pci-bridge, consider switching to that one. (One post on the internet states that the b43 module is “same as” b43-pci-bridge – perhaps in lsmod output at least.)

If none of this is helpful, consider looking at the linux wireless list.

stevesr0

[Member kernelpanic]

Helpful

Up

1

::

tried to post an answer here, but …
*** Forbidden. Message seems to be spam. ***
huh? 😉

[Member kernelpanic]

Helpful

Up

0

::

trying now with shorter text (without code)

@ BobC
antiX21 full 32bit sysVinit, neither connection on USB life, nor after full (no frugal) install to harddisk.

@ stevesrO
thank you for the hint with comparing dmesg!

I see the following differences (besides newer firmware version):
– cgroup: cgroup2: unknown option “nsdelegate”
– elogind-daemon[1701]: New seat seat0.
and after link becomes ready:
– cgroup: cgroup: disabling cgroup2 socket matching due to net_prio or net_cls activation
– elogind-daemon[1701]: New session 1 of user (myusername).
– wlan0: deauthenticated from e8:df:70:26:83:3a (Reason: 2=PREV_AUTH_NOT_VALID)

so the problem is deauthentication because of Reason: 2=PREV_AUTH_NOT_VALID

could this be because of elogin-daemon? (why is it creating a new session?)
I will investigate this further, at the moment I have no idea what “cgroup2 socket” means
or what elogind-daemon has to do with it.
if this does not lead to a solution, I will follow your other suggestions,
thank you very much so far!

[Member madibi]

Helpful

Up

1

::

what I did, and worked on my 15 years old Acer Extensa 5620Z, wifi BCM94311MCG, was the following procedure:

1. sudo rfkill unblock -all
2. sudo apt purge firmware-b43-installer
3. sudo apt update
4. sudo apt install --reinstall firmware-b43-installer
5. sudo modprobe -r b43
6. sudo modprobe b43

I cannot explain why, but it worked.
Results are not excellent, because the max speed is under 25 Gbit/sec (according to google speed test) in any case it is ok for me 🙂

• This reply was modified 1 year, 4 months ago by BobC. Reason: fixed typos, put code in code blocks

[Member madibi]

Helpful

Up

0

::

point 4 is
sudo apt install(space)(minus)(minus)reinstal etc 🙂

[Moderator BobC]

Helpful

Up

1

::

I tried madibi’s idea on my D620 while booted from flashdrive connected via hard wired eth0 cable and it worked.

[Member kernelpanic]

Helpful

Up

0

::

@ madibi

this did it!!!
it’s always surprising when I try to solve a problem in linux …
investigating here and there … spending a lot of time …
and then the solution is as easy as uninstalling/reinstalling,
unloading/loading firmware/modules in a certain SEQUENCE.

thanks a lot for your valuable hint! 🙂

[Member madibi]

Helpful

Up

0

::

OK, I’m happy for that. So please change the title of the tread and mark it with SOLVED 🙂

[Member stevesr0]

Helpful

Up

0

::

Hi kernelpanic,

Glad that worked.

I wonder if the dmesg logs are now (pretty much) identical for your two installs.

stevesr0

[Moderator BobC]

Helpful

Up

1

::

I was using the D620 with bcm4311 and it seems to max out at 16k/s.
The same laptop with 19.2 on the same wifi connection gets readings over 200k/s, so there is something not working well.

What can we do to diagnose and improve?

Btw, I found that all I needed to do was

sudo modprobe -r b43
sudo modprobe b43

[Moderator Brian Masinick]

Helpful

Up

0

::

BobC wrote:

I was using the D620 with bcm4311 and it seems to max out at 16k/s.
The same laptop with 19.2 on the same wifi connection gets readings over 200k/s, so there is something not working well.

What can we do to diagnose and improve?

Btw, I found that all I needed to do was

sudo modprobe -r b43
sudo modprobe b43

Coolness! I’ve either owned or used multiple variants of the Dell D600 series. I had a D600 or D610 that I got used. I used a couple of them at work. My nephew had one, and he gave it to my mother; that’s the only reason the specific models are “confused” in my mind, because over the years I think I’ve used the D600, D610, D620, and D630; they are among my favorites.

The other big favorite I’ve had is the IBM/Lenovo T series Thinkpads. To me, they might be #1 in reliability with the Dell models close.

I have a friend who’s picking up a mid-high end Dell Inspiron. He is excited about the equipment except for one thing: he claims that the quality is “garbage”. I asked why. He said that the old Latitude series and the Thinkpad series, at least the classic ones were the most solid systems ever. He said you could drop them and they were rugged enough to keep working, but he claims that the more recent Inspiron, heavy as it is, still isn’t nearly as rugged as that old classic stuff. I can’t argue with him, because this friend changes equipment like socks! He has “many pairs”!

Regarding the D series, I remember using the Broadcom B43; years ago that wasn’t always automatically configured; it was the only “quirk” regarding that system because the driver wasn’t (isn’t?) true free, open software. As long as there is a readily available source for that module, it’s very reliable and it works great.

Anyway

--
Brian Masinick

[Member kernelpanic]

Helpful

Up

0

::

stevesr0 wrote:

Hi kernelpanic,

Glad that worked.
I wonder if the dmesg logs are now (pretty much) identical for your two installs.
stevesr0

basically yes. some differences because of elogind-daemon, but this obvously had nothing to do with wifi.
in dmesg of the the broken system there was the line:

cgroup: cgroup: disabling cgroup2 socket matching due to net_prio or net_cls activation

which is now gone with the working system.

BobC wrote:

I was using the D620 with bcm4311 and it seems to max out at 16k/s.
The same laptop with 19.2 on the same wifi connection gets readings over 200k/s, so there is something not working well.

What can we do to diagnose and improve?

since this is the max speed of my connection anyway, I’m happy with it 😉

besides the “b43” (open source but needs closed source firmware) did you try the 2 other available drivers?
– proprietary “broadcom-wl” (did NOT work on my system)
– mainline “brcm80211” (didn’t test that one yet)

I found some nice infos on arch wiki:
https://wiki.archlinux.org/title/broadcom_wireless
they talk about the “brcm80211” as “kernel driver mainline version” and recommend it.
maybe that helps?

[Author]

Posts