- Package:
- network-manager
- Source:
- network-manager
- Description:
- network management framework (daemon and userspace tools)
- Submitter:
- Sam Morris
- Date:
- 2021-05-21 22:03:14 UTC
- Severity:
- normal
When trying to connect to a wireless network, I get the message "wireless is
disabled". But if I run 'iwlist scan', I can see several wireless networks.
This is a Dell Vostro 1320.
# lspci -s 0e:00 -vnn
0e:00.0 Network controller [0280]: Intel Corporation Wireless WiFi Link 5100 [8086:4232]
Subsystem: Intel Corporation Wireless WiFi Link 5100 [8086:1321]
Flags: bus master, fast devsel, latency 0, IRQ 32
Memory at fa000000 (64-bit, non-prefetchable) [size=8K]
Capabilities: [c8] Power Management version 3
Capabilities: [d0] MSI: Enable+ Count=1/1 Maskable- 64bit+
Capabilities: [e0] Express Endpoint, MSI 00
Capabilities: [100] Advanced Error Reporting
Capabilities: [140] Device Serial Number 00-22-fb-ff-ff-ab-b8-54
Kernel driver in use: iwlagn
Kernel modules: iwlagn
Toggling the hardware wlan enable switch makes an LED turn on and off, but
does not change NetworkManager's opinion on whether the hardware is disabled.
$ ls -l /sys/class/rfkill/
total 0
lrwxrwxrwx 1 root root 0 2009-08-29 18:03 rfkill0 -> ../../devices/virtual/rfkill/rfkill0
lrwxrwxrwx 1 root root 0 2009-08-29 18:03 rfkill1 -> ../../devices/pci0000:00/0000:00:1c.2/0000:0e:00.0/rfkill/rfkill1
$ cat /sys/class/rfkill/rfkill*/type
wlan
wlan
$ cat /sys/class/rfkill/rfkill*/state
2
1
$ nm-tool
NetworkManager Tool
State: connected
- Device: eth0 [Ifupdown (eth0)] ----------------------------------------------
Type: Wired
Driver: r8169
State: connected
Default: yes
HW Address: 00:24:E8:C2:9E:5B
Capabilities:
Carrier Detect: yes
Speed: 1000 Mb/s
Wired Properties
Carrier: on
IPv4 Settings:
Address: 192.168.0.231
Prefix: 24 (255.255.255.0)
Gateway: 192.168.0.1
DNS: 192.168.0.1
- Device: wlan0 ----------------------------------------------------------------
Type: 802.11 WiFi
Driver: iwlagn
State: unavailable
Default: no
HW Address: 00:22:FB:AB:B8:54
Capabilities:
Wireless Properties
WEP Encryption: yes
WPA Encryption: yes
WPA2 Encryption: yes
Wireless Access Points
N.B., I changed the version number in the bug report to the version in
stable because I tried that, and only upgraded to the version in backports
to see if that would fix this bug.
severity 544217 normal thanks Sam Morris wrote: Is this a regression or has NM worked before? Problem might be, that you are mixing bpo and stable, like kernel etc. The bpo version of NM is not officially supported. Could you try to reproduce the problem with an up-to-date sid system please. Michael
Not on this hardware. I just installed it today. Like I said, I only installed the version from backports.org to see if it would fix the bug. I've now purged and downgraded the networkmanager packages to versions from lenny. As for the kernel; there's no point downgrading to the kernel in stable, since it lacks drivers for the wireless interface, and even the driver for the ethernet interface doesn't work in 2.6.26. :( Unfortunately I can't do this as I need to run stable on this system. I'm just setting it up for my Dad to use. -- Sam Morris https://robots.org.uk/ PGP key id 1024D/5EA01078 3412 EA18 1277 354B 991B C869 B219 7FDB 5EA0 1078
Sam Morris wrote: If the rfkill state is not correctly detected, it's more likely a problem of hal, so you might start checking there first. Michael
affects 544319 network-manager
affects 544321 network-manager
block 544217 by 544319
block 544217 by 544321
thanks
The applet makes the following dbus call to find out if the wireless
hardware is disabled:
$ dbus-send --print-reply --system --dest=org.freedesktop.NetworkManager /org/freedesktop/NetworkManager org.freedesktop.NetworkManager.getWirelessEnabled
method return sender=:1.98 -> dest=:1.237 reply_serial=2
boolean false
boolean false
These two values correspond to the wireless_enabled and hw_rf_enabled
values from NetworkManager. These values are populated based on stuff
from HAL.
$ hal-find-by-capability --capability killswitch | while read line; do echo $line; dbus-send --system --print-reply --dest=org.freedesktop.Hal $line org.freedesktop.Hal.Device.KillSwitch.GetPower; done
/org/freedesktop/Hal/devices/computer_rfkill_dell_wifi_wlan
method return sender=:1.103 -> dest=:1.285 reply_serial=2
int32 0
/org/freedesktop/Hal/devices/pci_8086_4232_rfkill_5100AGN_wlan
method return sender=:1.103 -> dest=:1.287 reply_serial=2
int32 1
/org/freedesktop/Hal/devices/dell_wlan_switch
int32 0
According to the HAL spec, GetPower returns 1 iff the power to the radio
is on.
At this point I took a closer look at these three killswitch devices.
The full output of 'lshal' is attached to the original bug report.
The first device is an rfkill device. If I read
/sys/devices/virtual/rfkill/rfkill0/state then I get back '2', meaning
that the radio has been disabled in hardware--incorrect! If I remove the
dell_wireless kernel module then this device disappears. Bug #544319
filed.
The second device is another rfkill device, and it seems to work
fine--its state is '1' which means the radio is on.
The third device has its 'killswitch.access_method' set to 'dell', which
causes HAL to run the dellWirelessCtl program:
$ dellWirelessCtl --st_wlan; echo $?
Radio Status for WLAN:
WLAN supported
WLAN installed
WLAN disabled
Status Code: 1
1
Bug #544321 filed.
There are a few options for working around these bugs. I list them ehre
for the benefit of other users with the same problem:
1. Tell HAL to ignore all rfkill devices. I tried to do this by
setting the 'info.ignore' property on each of the killswitch
devices. Unfortunately, this didn't stop NetworkManager from
seeing them. I guess info.ignore isn't supposed to be used for
this purpose. Is there another way to deal with this at the HAL
level?
2. Patch NetworkManager to override the information coming from
HAL. The 'nm-ignore-killswitches.patch' is attached, and it does
work. Of course, it's not suitable as anything other than a
crude hack to workaround the issue.
3. Fix the broken killswitch devices. The first broken device comes
from the kernel's dell_laptop module, so that's one bug. The
second comes from HAL's device database, but is actually queried
with the dellWirelessCtl utility, which returns the wrong
result. I have filed separate bugs for these issues.
4. Work around the broken killswitch devices. The dell_laptop
kernel module can simply be blacklisted, and removing the
libsmbios-bin package that contains dellWirelessCtl causes
querying HAL for its status to return an error; which in turn
causes NetworkManager to assume that the radio is enabled.
5. Fix the hardware. Dell have posted a BIOS update for my laptop,
version A03, the release notes of which include "Fix kill switch
sttatus report fail". Unfortunately, it is impossible to
actually install this update without installing Windows, so I
can't test that out. :(
BTW, I booted the Ubuntu 9.04 LiveCD and wireless worked fine. I ran
lshal (output attached) and noticed that they don't have the broken
computer_rfkill_dell_wifi_wlan device. Maybe this is because they have
some super secret fixes in their packages, or maybe it's because this is
a bug in a later kernel than their 2.6.28 kernel (or they don't build
the dell_laptop kernel module).
I have this same exact problem on sid, after my notebook "woke up" from suspend state, my network manager says "Networking disabled" and no wireless connections are being displayed. I suggest you change the bug status back to grave, since it's incredibly serious bug, and I see they have same problem on ubuntu as well. This is happening on just updated Sid, hardware Lenovo X300 absinthe@havoc:~$ lspci | grep -i "wireless" 03:00.0 Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN [Kedron] Network Connection (rev 61) absinthe@havoc:~$ ls -l /sys/class/rfkill/ total 0 lrwxrwxrwx 1 root root 0 Dec 13 19:40 rfkill0 -> ../../devices/platform/thinkpad_acpi/rfkill/rfkill0 lrwxrwxrwx 1 root root 0 Dec 13 19:40 rfkill1 -> ../../devices/pci0000:00/0000:00:1a.0/usb3/3-1/3-1:1.0/bluetooth/hci0/rfkill1 lrwxrwxrwx 1 root root 0 Dec 13 19:40 rfkill2 -> ../../devices/pci0000:00/0000:00:1c.1/0000:03:00.0/ieee80211/phy0/rfkill2 absinthe@havoc:~$ cat /sys/class/rfkill/rfkill*/type bluetooth bluetooth wlan absinthe@havoc:~$ cat /sys/class/rfkill/rfkill*/state 1 1 1 absinthe@havoc:~$ nm-tool NetworkManager Tool State: asleep Adnan
I got the same problem after updating my sid (I didn't update it for ages). I have a Dell Latitude E6400. There are basically two ways to make it work again : - the simplest : blacklist dell_laptop echo "blacklist dell_laptop" > /etc/modprobe.d/blacklist_dell_laptop.conf inconvenient : dell_laptop isn't loaded. I don't know what it does. - somewhat more difficult : update your bios Since I had some problems to be able to update my bios on Debian, I wrote some detailed walkthrough on my blog : http://www.everlong.org/blog/index.php/post/2009/12/How-to-update-the-BIOS-firmware-on-a-Dell-Laptop-using-only-Debian-Linux I guess this resolves also the problems in bug #544319 and bug #544321. What should we do in Debian ? Should we try to workaround this broken bios ? The problem is that... it used to work before, even if it wasn't perfect (I had to restart network-manager each time I used the rfkill switch, now it works perfectly).
I can confirm this bug on a Dell E6410. What fixed the issue for me was to downgrade network-manager and network-manager-gnome to the version from stable (I'm running testing). I'm just posting this since other people might want to resort to a simpler solution to have this work for them :-).
Salaam, Almana Financial Broker, is the home of discerning investors. We offer independent financial advice and assist our clients to make sound investment decisions from the array of investment opportunities available. We are structured to render personalized services to our clients thereby ensuring safety of capital, adequate returns on investments. Our investors are ready to provide funding for your business expansion as Debt/Equity finance. If you require funding, we would be able to partner with you. We look forward to your response, thanks and stay safe, Abdulkader Abdi, CFA Almana Financial Broker Dubai, United Arab Emirates