Intel Pro/Wireless 4965 AG stops transferring data within a few minutes

Bug #277634 reported by Jonathan Ernst on 2008-10-03
52
This bug affects 8 people
Affects Status Importance Assigned to Milestone
linux (Ubuntu)
Medium
Unassigned
Nominated for Intrepid by Jonathan Ernst
Declined for Jaunty by Brian Murray
Nominated for Karmic by abtabdn
linux-meta (Ubuntu)
Undecided
Unassigned
Nominated for Intrepid by Jonathan Ernst
Declined for Jaunty by Brian Murray
Nominated for Karmic by abtabdn

Bug Description

On a Lenovo T61 the wireless NIC stops transmitting packets after some time. Then you have to wait ~1 minute and it's working again for some time.

The driver used is iwlagn

You can see some details here : http://www.intellinuxwireless.org/bugzilla/show_bug.cgi?id=1700

Changed in linux-meta:
assignee: nobody → ubuntu-kernel-team
Jonathan Ernst (jonathan.ernst) wrote :

Zhu Yi from Intel says he thinks it has been fixed in 2.6.27 rc1 or rc2... However it is still not working in Intrepid, against what rc is linux-image-generic based ?

description: updated
description: updated
Christian Doczkal (thyrus) wrote :
Download full text (3.8 KiB)

I have the same problem on an FSC Lifebook. The adapter is:

Intel Corporation PRO/Wireless 4965 AG or AGN [Kedron] Network Connection (rev 61)
and the driver is: iwlagn as well.

during these "freeze" times route hangs

chris@laptop:~$ route
Kernel-IP-Routentabelle
Ziel Router Genmask Flags Metric Ref Use Iface
192.168.178.0 * 255.255.255.0 U 2 0 0 wlan0
link-local * 255.255.0.0 U 1000 0 0 wlan0
------- here route hangs -------
default fritz.fonwlan.b 0.0.0.0 UG 0 0 0 wlan0

I have attached my dmesg.

Furthermore the disconnects seem to be very regular (every 10 minutes) and i get the following in my syslog each time

Nov 4 22:30:01 laptop /USR/SBIN/CRON[8055]: (root) CMD ([ -x /usr/sbin/update-motd ] && /usr/sbin/update-motd 2>/dev/null)
Nov 4 22:33:02 laptop ntpd[7986]: synchronized to 130.149.17.21, stratum 1
Nov 4 22:39:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Nov 4 22:39:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7
Nov 4 22:40:01 laptop /USR/SBIN/CRON[8365]: (root) CMD ([ -x /usr/sbin/update-motd ] && /usr/sbin/update-motd 2>/dev/null)
Nov 4 22:48:04 laptop ntpd[7986]: synchronized to 130.149.17.8, stratum 1
Nov 4 22:49:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Nov 4 22:49:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7
Nov 4 22:50:01 laptop /USR/SBIN/CRON[8531]: (root) CMD ([ -x /usr/sbin/update-motd ] && /usr/sbin/update-motd 2>/dev/null)
Nov 4 22:51:46 laptop ntpd[7986]: synchronized to 130.149.17.21, stratum 1
Nov 4 22:55:29 laptop ntpd[7986]: synchronized to 130.149.17.8, stratum 1
Nov 4 22:59:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Nov 4 22:59:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7
Nov 4 23:00:01 laptop /USR/SBIN/CRON[8636]: (root) CMD ([ -x /usr/sbin/update-motd ] && /usr/sbin/update-motd 2>/dev/null)
Nov 4 23:03:36 laptop ntpd[7986]: synchronized to 130.149.17.21, stratum 1
Nov 4 23:09:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Nov 4 23:09:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7
Nov 4 23:10:01 laptop /USR/SBIN/CRON[8756]: (root) CMD ([ -x /usr/sbin/update-motd ] && /usr/sbin/update-motd 2>/dev/null)
Nov 4 23:16:58 laptop ntpd[7986]: synchronized to 130.149.17.8, stratum 1
Nov 4 23:17:01 laptop /USR/SBIN/CRON[8825]: (root) CMD ( cd / && run-parts --report /etc/cron.hourly)
Nov 4 23:19:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Nov 4 23:19:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7
Nov 4 23:20:01 laptop /USR/SBIN/CRON[8906]: (root) CMD ([ -x /usr/sbin/update-motd ] && /usr/sbin/update-motd 2>/dev/null)
Nov 4 23:29:35 laptop NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Nov 4 23:29:35 laptop NetworkManager:...

Read more...

Christian Doczkal (thyrus) wrote :

Here is the full syslog

Christian Doczkal (thyrus) wrote :

OK further testing reveals that downloading a lot of data reduces the time between connection hangs to just a handful of seconds. The observation with the route command however persists.

Augusten (grilledeasterbunny) wrote :

Same Problem here. I am on a ThinkPad T61p running Ubuntu 8.10 upgraded from hardy; I too have the 4965 AGN wireless card (home network, Belkin N router, wep2), iwlagn driver, am using the most current kernel and can confirm this issue. After following the "Known Issues" caution (about kernel panic and the 4965 agn on the release notes), I installed the linux backports which enabled wifi with this card. however, the "hang" has always been there. I am am idiot brand new to Ubuntu so only after blamming firefox and opera and reducing them almost to text browsers did I read the error logs and see the problem. Many people across many boards are experiencing the identical issue, often under the impression it is something else.

This appears to be bug 1700 on Bugzilla: http://www.intellinuxwireless.org/bugzilla/show_bug.cgi?id=1700

As well as bug 179698 on launchpad: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/179698

And it's very similar to this -which was fixed allegedly: https://bugs.launchpad.net/ubuntu/+source/linux/+bug/264104

I don't use bluetooth so I removed it; one of the posts somewhere spoke of success after killing bluetooth. My error logs continue to show the supplicant state change every few minutes:

Nov 24 02:26:21 augusten-T61p NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7 Nov 24 06:19:13 augusten-T61p NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Nov 24 06:19:13 augusten-T61p NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7
Nov 24 06:52:29 augusten-T61p NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Nov 24 06:52:29 augusten-T61p NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7
Nov 24 07:25:45 augusten-T61p NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Nov 24 07:25:45 augusten-T61p NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7

But since removing Bluetooth, I have not observed any interruption in service. But it remains to be seen whether the disabling of bluetooth has, indeed, made a difference. A full day of use should tell me.

-ab

Augusten (grilledeasterbunny) wrote :

As a follow up, problem continues -as before- after removal of bluetooth.
ab

Augusten (grilledeasterbunny) wrote :

Looking at the syslog, I saw a pattern. Just prior to each supplicant connection state change the system initiated this CRON:

Nov 24 09:20:01 augusten-T61p /USR/SBIN/CRON[11941]: (root) CMD ([ -x /usr/sbin/update-motd ] && /usr/sbin/update-motd 2>/dev/null)

I removed the me-specific information and did a Google search -what i found was very interesting: messages like the one on this page; people experiencing some sort of wireless hangs while using the 4965 agn (and I think others). I don't know anything about debugging or programming or Ubuntu -I can operate a can opener and that's it. But there had to be a pattern. So I snooped around and it didn't seem like the system would die if I disabled that CRON so I did. But upon restart is was back.

So using Synaptic, I removed the 'update-motd 2' completely. Since then, I have experienced no breaks or inrurruptions in wireless service. This is hardly a fix, it's a crude workaround but it does appear to work.

ab

Erik Reuter (misc71) wrote :

Is anyone looking into this bug? What information does the bug team need that has not been provided?

I have a Dell Inspiron 1720 with an Intel 4965 AG wireless card

0c:00.0 Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN [Kedron] Network Connection (rev 61)

  *-network
       description: Wireless interface
       product: PRO/Wireless 4965 AG or AGN [Kedron] Network Connection
       vendor: Intel Corporation
       physical id: 0
       bus info: pci@0000:0c:00.0
       logical name: wmaster0
       version: 61
       serial: 00:1d:e0:53:42:b5
       width: 64 bits
       clock: 33MHz
       capabilities: bus_master cap_list logical ethernet physical wireless configuration: broadcast=yes driver=iwlagn ip=192.168.0.158 latency=0 module=iwlagn multicast=yes wireless=IEEE 802.11abgn

running Ubuntu 8.10 and kernel 2.6.27-10 (but I also saw the issue with 2.6.27-9).

The issue is that, at apparently random times, the wireless connection drops. It usually comes back by itself after a few minutes, but sometimes not. Whenever the connection drops, I see log entries similar to those reported by others:

Dec 14 19:04:36 i1720 NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Dec 14 19:04:36 i1720 NetworkManager: <info> (wlan0): supplicant connection state change: 7 -> 6
Dec 14 19:04:36 i1720 NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7
Dec 14 19:04:36 i1720 NetworkManager: <info> (wlan0): supplicant connection state change: 6 -> 7

Erik Reuter (misc71) wrote :

Here is someone else who appears to experience the same bug:

http://ubuntuforums.org/showpost.php?p=6362798&postcount=1

Erik Reuter (misc71) wrote :

I switched my D-Link DIR-655 wireless router from WPA2 authentication to WEP authentication. I believe this also disabled wireless N protocol since I don't think N supports WEP (how can I check in Ubuntu what wireless protocol I am using, n, g, or b?). Since I made the switch, I have had no wireless dropouts for 2 days. Before I made the switch, I would have 2 to 5 wireless dropouts per day.

I think this intermittent network drop out issue may be related either to WPA authentication (WPA supplicant bug?), or N-type wireless networking (driver issue?). Also, when I search the web for similar issues, most of the reports I see are for Ubuntu. I am wondering whether the wireless dropouts are specific to the way Ubuntu is using the driver, wpa supplicant, or network manager?

Per a decision made by the Ubuntu Kernel Team, bugs will longer be assigned to the ubuntu-kernel-team in Launchpad as part of the bug triage process. The ubuntu-kernel-team is being unassigned from this bug report. Refer to https://wiki.ubuntu.com/KernelTeamBugPolicies for more information. Thanks.

basejumper9 (basejumper9) wrote :

I have similar issues I thought it was more associated with roaming between access points but this seems to describe my problem exactly. I'm running 8.10 on a Lenovo X61 with a 4965 card. Let me know if I can upload anything to help.

Rocko (rockorequin) wrote :

Does the problem only occur with 11N networks? I haven't had any issues with my 4965AGN card on 11g networks, but with the only 11n I've found I can't get a dhcp address (see https://bugs.launchpad.net/ubuntu/+source/linux/+bug/305338).

If the dropouts are only happening on 11n networks, maybe it's worth disabling 11n in iwlagn and seeing if this stops the dropouts:

sudo modprobe -r iwlagn && sudo modprobe iwlagn 11n_disable=1

or add 'options iwlagn 11n_disable=1' to /etc/modprobe.d/options to make it permanent.

Changed in linux-meta (Ubuntu):
status: New → Invalid
Melk79 (smelkus) wrote :

I experience this problem under Jaunty beta, but only under 11n networking.

I've found when the connection stops transmitting packets if I click on another link (that hasn't been cached into memory) many times in succession (~20x), it will reconnect in about 20 seconds instead of +1 minute.

Henry (hchang858) wrote :

I had a related problem which has been similarly reported elsewhere

Running Jaunty on a T61 with Intel 4965 AGN. Wireless worked fine on Intrepid, but has had problems since upgrading to Jaunty. A fresh Jaunty install did not fix (not 100% fresh, migrated my home folder)

Wireless would always successfully connect upon startup, but after some time (varying from minutes - usually - to sometimes hours), the internet and network devices would become suddenly unavailable, although network manager showed the network as still being connected. Seems to occur with both networkmanager and wicd.

After being unable to find a reliable way to cause the problem, I read Augusten's post about removing motd, and followed that using Synaptic. So far no disconnects (a day), which seems to me to be a minor miracle.

Bryan Wu (cooloney) wrote :

Can you please try installing the linux-backports-modules-jaunty package. The linux-backports-modules-jaunty package contains and updated compat-wireless stack which may help. Be sure to reboot after installing this package in order to test. Please let us know your results.

If the issue remains even after installing and testing linux-backports-modules-jaunty, could you then please try the latest upstream compat-wireless stack. See http://wireless.kernel.org/en/users/Download for more info. If the issue exists upstream, we'll want to work with the upstream developers to try to get this resolved. Also, after confirming the issue remains with the latest upstream compat-wireless stack, be sure to also run the following command from a Terminal (Applications->Accessories->Terminal). It will automatically gather and attach updated debug information to this report.

apport-collect -p linux-image-2.6.28-11-generic <bug #>

Changed in linux (Ubuntu):
importance: Undecided → Medium
status: New → Incomplete
u20 (jacquescortelyou) wrote :

Bryan nailed it.... Went into Synaptic and installed the backport modules, the overall speed increase and the drop outs are gone. Thanks!

Philip Langer (philchip) wrote :

Unfortunately the backport modules didn't resolve this issue on my machine (Dell XPS M1530 with Intel Corporation PRO/Wireless 4965 AG or AGN). I have jaunty with Kernel 2.6.28-15-generic #49-Ubuntu SMP Tue Aug 18 19:25:34 UTC 2009 x86_64 GNU/Linux.

I also installed latest upstream compat-wireless stack but the issue remains. I can connect to wifi-networks but after some time the connection gets lost.

Here is the part in the kern.log when the connection gets lost:

[ 3225.583607] wlan0: RX ReassocResp from ffff880098ca404e (capab=0x421 status=0 aid=16)
[ 3225.583609] wlan0: associated
[ 3248.293306] wlan0: beacon loss from AP ffff88011d0f6ac0 - sending probe request
[ 3361.330590] iwlagn 0000:0b:00.0: Microcode SW error detected. Restarting 0x82000000.
[ 3361.635675] Registered led device: iwl-phy0::radio
[ 3361.635689] Registered led device: iwl-phy0::assoc
[ 3361.635701] Registered led device: iwl-phy0::RX
[ 3361.635713] Registered led device: iwl-phy0::TX
[ 4002.097082] CE: hpet increasing min_delta_ns to 33750 nsec
[ 9226.152683] CE: hpet increasing min_delta_ns to 50624 nsec

I'll see if I have time to submit the debug information requested by Bryan in the next days.

Thank you!

Philip Langer (philchip) wrote :

After a little browsing I recognized that my issue might (also) be related to this bug:
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/200509 (Microcode SW error)

So, if installing the backport module fixes this issue for the others experiencing the problem (without having the Microcode SW error), my current problem might not be related to this bug, but to the #200509 bug. In this case, sorry for spamming :)

Ian Sullivan (sullivan) wrote :

I'm running the "Intel Corporation PRO/Wireless 4965 AG or AGN [Kedron] Network Connection (rev 61)" network card on an Thinkpad X61s and have been encountering these exact prolems on a fully updated Jaunty install.

I installed the backport module package as suggested earlier in the thread but am still encountering the same problems.

If it is of use, I notice that I run into more frequent connection drops when I am in an area that has many (20+) wifi networks than when I am at home and can only see 2 or 3. Could be coincidental but thought I would mention it. Further suggestions are very welcome; this bug is pretty frustrating.

This bug is still present in Ubuntu 9.10

I too am having this issue. I have recently upgraded to 9.10 and am experiencing about 5 drop-outs per day.

I'm not certain whether or not this is related to wifi interference; although I can confirm that since I upgraded to Karmic, my issue has become chronic (before, while running Hardy, it almost never disconnected).

Ian Sullivan (sullivan) wrote :

I'm in the same boat as Keith, wireless has become almost unusable since upgrading to Jaunty. The backports module has no impact. What other information is needed to reconfirm this bug?

I can confirm this bug on Ubuntu 9.10 (Karmic) AMD64, kernel 2.6.31-14-generic #48-Ubuntu SMP Fri Oct 16 14:05:01 UTC 2009 x86_64 GNU/Linux.
Laptop is a Sony Vaio VGN-AR730E with Intel Wireless WiFi Link 4965AGN REV=0x4.

I could not find the page about the reasons of ubuntu-kernel-team not being assigned to this bug, so if anyone find the current page please add the URL here.

[]s!

It is really frustrating.

I have entered in this bug now with my COMPAL clone with Lucid with 4965AGN. It started for me somewhere in the 2.6.32.X, and now I am now really sort of cabling my house....
Randomly here and there, one or two times per hour, without any network activity or with activity, connection get lost with a message:

No probe response from AP XX:XX:XX:XX:XX:XX after 500ms, disconnecting.

Then it sometimes manages to automatically reconnect, sometimes I have to disconnect/reconnect.
It did NOT happen with jaunty on same HW, and it happened very sporadically with karmic. Now on lucid after the 2.6.32 switch is REALLY frustrating. It can be considered for sure a major bug, because it drives people crazy, expecially if they (like myself) work with SSH connection. I am back to the old screen program days, but it is frustrating frustrating frustrating. Is there any way to control timeout for probe or have a less "strict" timeout check?

This bug also affect my system, a Dell Inspiron 1525 with Intel wireless card running 8.04 LTS with Kernel Linux 2.6.24-27-generic. I always install all system updates. I had trouble getting WPA to work so it is configured to use WEP. lspci shows:

0b:00.0 Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN Network Connection (rev 61)

The wireless works fine for a while (sometimes a few minutes, sometimes an hour or two), then randomly drops out. I've been disabling/reenabling wireless or even rebooting to get it back, but after reading the other reports I wonder if it's really just the elapsed time that fixes it. This symptom has been present ever since I first installed Ubuntu on the brand-new machine about a year ago. It is especially annoying because I paid extra to get an Intel card to try to avoid such problems. (My previous system with a Broadcom card used "ndist" with windows drivers and worked great, but I haven't figured out a way to do that in 8.04)

I checked /var/log/sys* and there are no "supplicant" or "motd" messages in there.

I'd be willing to try the backport module, but when I bring up the Synaptic GUI it cannot find a package named "linux-backports-modules-jaunty".

I have a similar problem on my Thinkpad X61-tablet with Intel 4965 hardware.
However, I get messages in /var/log/kern.0.log reporting "Microcode SW error"
within iwlagn. My only corrective action is reboot. Attempts to re-init the hardware short of reboot are not successful. (The hardware must need the power cycle to recover from the mu-code error.)

Tell me how, and I'll gladly activate software instrumentation.

I'm running Ubuntu Jaunty and Mint-7 (also a Jaunty derivative) with similar troubles.
My kernel is v2.6.26-18-generic and v2.6.26-11-generic respectively.

~~~ 0;-Dan

Saint DanBert (saintdanbert) wrote :

I loaded Ubuntu Lucid (v10.04) and now my intel 4965 agn wireless does not work at all under Linux. It works fine under winXP on the dual boot side.

I've used 'wavemon' and see reasonable signal strength and other details for my access points.

I use 'wicd' instead of network-manager. It shows my access points and then fails to connect saying "bad password".
I've confirmed the string being tried using the debug-mode for wicd. I use the same string from winXP and get
a good connection every time. Along with the fact that other hosts connect fine seems to rule out hardware.

I'll gladly diagnose using whatever you suggest. I'm completely stumped on this one.

John (jkrez) wrote :

Hi all,

I have Lucid 10.04.1 LTS and Network controller: Intel Corporation PRO/Wireless 4965 AG or AGN [Kedron] Network Connection (rev 61).

Problem: Network would randomly disconnect from networks after some period of time. I could not connect to that network again. It happened to be one that required a certificate, username and password. Though I could log into another network that did not require any login information after this network randomly stopped working.

Here's what fixed my problem this time
$ sudo apt-get install --reinstall linux-backports-modules-wireless-lucid-generic
$ sudo apt-get install --reinstall linux-backports-modules-headers-lucid-generic

I've had a various networking issues that I've fixed the following ways:
http://jkrez.wordpress.com/2010/08/20/

To post a comment you must log in.
This report contains Public information  Edit
Everyone can see this information.

Other bug subscribers

Remote bug watches

Bug watches keep track of this bug in other bug trackers.