Intel 3945ABG has one connection attempt, after that a reload of the ipw3945 module is needed for the interface to work at all. Connections and connecting in itself is unreliable

Bug #134515 reported by Zorael
26
Affects Status Importance Assigned to Milestone
ipw3945
Invalid
Undecided
Unassigned
linux (Ubuntu)
Won't Fix
Undecided
Unassigned
linux-source-2.6.22 (Ubuntu)
Won't Fix
Undecided
Unassigned

Bug Description

(NOTE: edited version once I realized I could modify the original description, some of the following comments of mine may no longer apply)

I have an Acer Aspire 9815 which has an integrated Intel 3945ABG networking chipset.

I only have *one* attempt at connecting to a wireless network per "load" (modprobe) of the ipw3945 module. network-manager and/or knetworkmanager seems to have some additional steps in its connection procedure rendering it impossible to connect unless I use DHCP; likely something early in the procedure "uses up" my one golden connection try, leaving the interface dumb and mute once it's time to actually connect. I replaced them with wicd which does the trick for me, managing to connect using either static or dynamic.

It works flawlessly in Windows. It worked marvelously and without hitches prior to sending the machine in for service, where they replaced the motherboard. Many others (and myself two months ago) could effortly use this networking chipset out-of-the-box, while the I of today and some others (going by the replies I've encountered on ubuntuforums.org) have to fight to keep it working at all, with varying degree of success. Perhaps there's a revision n+1 of the chipset, and/or a new firmware, incompatible to the current ipw3945?

zorael@azrael:~$ lspci | grep 3945
07:00.0 Network controller: Intel Corporation PRO/Wireless 3945ABG Network Connection (rev 02)

zorael@azrael:~$ sudo lshw -C network
  *-network
       description: Ethernet interface
       product: 88E8055 PCI-E Gigabit Ethernet Controller
       vendor: Marvell Technology Group Ltd.
       physical id: 0
       bus info: pci@02:00.0
       logical name: eth0
       version: 12
       serial: 00:a0:d1:a0:ec:6b
       capacity: 1GB/s
       width: 64 bits
       clock: 33MHz
       capabilities: bus_master cap_list ethernet physical tp 10bt 10bt-fd 100bt 100bt-fd 1000bt 1000bt-fd autonegotiation
       configuration: autonegotiation=on broadcast=yes driver=sky2 driverversion=1.13 firmware=N/A latency=0 link=no multicast=yes port=twisted pair
       resources: iomemory:cc000000-cc003fff ioport:3000-30ff irq:16
  *-network
       description: Wireless interface
       product: PRO/Wireless 3945ABG Network Connection
       vendor: Intel Corporation
       physical id: 0
       bus info: pci@07:00.0
       logical name: eth1
       version: 02
       serial: 00:18:de:6e:39:9a
       width: 32 bits
       clock: 33MHz
       capabilities: bus_master cap_list ethernet physical wireless
       configuration: broadcast=yes driver=ipw3945 driverversion=1.2.0mp firmware=14.2 1:0 () ip=192.168.0.2 latency=0 link=yes multicast=yes wireless=IEEE 802.11b
       resources: iomemory:ce000000-ce000fff irq:19

zorael@azrael:~$ iwconfig
lo no wireless extensions.

eth0 no wireless extensions.

eth1 IEEE 802.11b ESSID:"LappskoleB"
          Mode:Managed Frequency:2.437 GHz Access Point: 00:11:95:1F:68:EB
          Bit Rate:11 Mb/s Tx-Power:15 dBm
          Retry limit:15 RTS thr:off Fragment thr:off
          Power Management:off
          Link Quality=71/100 Signal level=-65 dBm Noise level=-66 dBm
          Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
          Tx excessive retries:0 Invalid misc:798 Missed beacon:0

zorael@azrael:~$ iwlist eth1 scan
eth1 Scan completed :
          Cell 01 - Address: 00:11:95:1F:68:EB
                    ESSID:"LappskoleB"
                    Protocol:IEEE 802.11b
                    Mode:Master
                    Channel:6
                    Encryption key:on
                    Bit Rates:1 Mb/s; 2 Mb/s; 5.5 Mb/s; 11 Mb/s
                    Quality=67/100 Signal level=-66 dBm Noise level=-66 dBm
                    Extra: Last beacon: 180ms ago

--------- When broken ---------

zorael@azrael:~$ iwconfig
lo no wireless extensions.

eth0 no wireless extensions.

eth1 unassociated ESSID:"LappskoleB"
          Mode:Managed Frequency=2.437 GHz Access Point: 00:11:95:1F:68:EB
          Bit Rate:11 Mb/s Tx-Power:16 dBm
          Retry limit:15 RTS thr:off Fragment thr:off
          Power Management:off
          Link Quality:0 Signal level:0 Noise level:0
          Rx invalid nwid:0 Rx invalid crypt:0 Rx invalid frag:0
          Tx excessive retries:0 Invalid misc:1162 Missed beacon:0

zorael@azrael:~$ sudo iwlist eth1 scan
eth1 No scan results

zorael@azrael:~$ sudo iwlist eth1 scan
eth1 Interface doesn't support scanning.

zorael@azrael:~$ sudo iwlist eth1 scan
eth1 No scan results

At times, it also says "Frequency=nan kHz" when it's in its non-working state.

Even when borked, it is correctly set up through ifconfig and iwconfig; it just doesn't see any available networks. And if I enter a "hidden ESSID", it still can't/won't connect. ifdown and then ifup doesn't help, disabling and then reenabling the wireless radio through the kill switch doesn't help, nor does biting one's nails. It *seems* to be set up correctly in every way imaginable, but it doesn't do anything.

Only rebooting or manually doing:

zorael@azrael:~$ sudo modprobe -r ipw3945 && sudo modprobe ipw3945

...gives me another golden attempt. Which doesn't spell well for when the machine loses connection when unattended. (For instance, if the router randomly restarts when I have class, or if the chipset itself decides to disconnect for no apparent reason. Which happens, all too often.)

NOTE that this only grants me another *try* at connecting; there's no guarantee I'll actually be able to establish one until after five more attempts. It's iffy, sketchy, unreliable.

Sometimes, if I connect with DHCP, it just freezes at Obtaining IP address.
Other times, both with DHCP and with a static ip, it connects and reports a 0% signal.
Further other times, it connects and reports a valid signal, but only one ping request goes through before the connection dies. (I usually run a 'ping 192.168.0.1' in another terminal so I know if I succeeded.)

As mentioned in a comment below: When I successfully get through but with only a 0% signal, there's no successful traffic, but it reports being connected. Workaround ritual: disconnecting (ifconfig down or through wicd), stopping/starting the module, connecting with DHCP but cancelling when obtaining IP, disconnecting, stopping/starting module, then connecting with static IP again fixes it in 90% of all cases.

Revision history for this message
Zorael (zorael) wrote :

I forgot to mention, I experience the same thing in Ubuntu/Kubuntu 7.04 as in Ubuntu 7.10 tribe 4. I only tested briefly in 7.10 tribe 4 with Gnome's network manager, since it was on a Live CD session, but if it helps the process along I can download tribe 5 and install it properly.

Revision history for this message
Zorael (zorael) wrote :

Further, seems I'm not alone in having this problem.

http://ubuntuforums.org/showthread.php?t=530352

Revision history for this message
Zorael (zorael) wrote :

Some updates.

I tried it with a properly installed 7.10 tribe 5, still the same errors there, so reinstalled 7.04 to be able to safely use it as my desktop computer.

<copy/paste from my forum thread>

I've made "some" progress.

First off, I scrapped network-manager and knetworkmanager in favor of wicd, and I've never since looked back.

Secondly, while I still only have one shot at connecting, but *now* I can reset it without having to restart the machine.

Now, off the top of my head, the ritual to restore after it's broken down (with likely some irrelevant, unneeded and redundant steps):

1) Open wicd
2) Disconnect the network to be safe (with the button)
3) Close wicd
3) Do 'sudo modprobe -r ipw3945 && sudo modprobe ipw3945' in a terminal (mayhaps just one modprobe ipw3945 call does the trick, I'm not knowledgeable enough to know)
4) Open wicd again, hit refresh
5) At this point, it discovers my essids again 90% of the time, if not I return to step 2, rinse and repeat
6) Hit connect and pray.

That'd be the discovery part. I've had shaky luck of being able to reconnect, with a success rate of ~20% or less.

Sometimes, if I do it the DHCP way, it just freezes at Obtaining IP address.
Other times, both with DHCP and with a static ip, it connects and reports a 0% signal.
Further other times, it connects and reports a valid signal, but only one ping request goes through before the connection dies. (I usually run a 'ping 192.168.0.1' in another terminal so I know if I succeeded.)

Wicd doesn't seem to want to connect successfully (or at least, not give the UI feedback that it is trying) if you haven't hit disconnect since your last attempt, and often if you haven't restarted the program itself. So a lot of opening/closing involved.

Better than nothing, I guess. This way I can keep at it and eventually get it back in shape without the constant reboots. But it's still about as broken; what if the connection breaks down when I'm not there to give an effort to beat it back into shape?

---

So there seems to be *something* awry, perhaps with the ipw3945 module, in the way it "releases" the interface, forcing me to reload the module to be able to once again discover essids and establish connections. Likely the iffy connecting is also a result of this?

Zorael (zorael)
description: updated
Revision history for this message
MrYutz (gregbazar) wrote : I can confirm this behavior

I have an Acer TravelMate 8210 with the same wireless chipset on Gusty - and have the same isssue.

Revision history for this message
Zorael (zorael) wrote : Re: Intel PRO/Wireless 3945ABG has one connection attempt, after that a reload of the ipw3945 module is needed to discover ESSIDs, connecting remains iffy business

Workaround for when I only end up with a 0% strength connection (at least for static ips): connect with DHCP, cancel it when it's attempting to obtain IP (letting it finish takes forever and a day), stop and reload the module, connect again with static ip. It seems to do the trick every time.

Zorael (zorael)
description: updated
Zorael (zorael)
description: updated
Revision history for this message
Andy Leeman (andy-leeman) wrote :

I have posted a similar, possibly related bug;
https://bugs.launchpad.net/ubuntu/+bug/139642

Revision history for this message
TheKind (meda-ugo) wrote :

I have the same issue with my Dell 1520, Wifi pre-N, on Gusty.

Revision history for this message
RainerT (tr-ml) wrote :

The same here, Dell Latitude 630, Gutsy.

Revision history for this message
Alexandre Campo (alexandre-campo) wrote :

Similar problem, Dell xps m1330, Gutsy

Revision history for this message
muszek (muszek) wrote :

Dell 1720 - a similar or exactly the same problem. I submitted the bug report almost a month ago - https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.22/+bug/154549 .

You guys might consider yourselves lucky - pair these problems with a router that goes down ever more often and you have a really frustrating situation - you never know which thing died. It's been 2 months since I moved to this house and reported the problem to the landlord :/

I'll add just one thing that I noticed lately and might be relevant: my router likes to die mainly in two usage scenarios: p2p and skype... what is interesting is that my girlfriend's skype (on Windows) doesn't cause problems. Lastest stable (I think it's 1.4) on my computer doesn't do any harm either. Skype beta 2.0.0.13 kills the router every 5-10 minutes. Could it be something related to Skype using UDP?
P2P case: all torrent clients I've tried (utorrent on girlfriend's windows, deluge and azureus on my gutsy) manage to kill it. But azureus seems to do it like a pro - Right after a program starts, when I'm connected to 2 peers (there aren't many more in the swarm) or... when I stop downloading a torrent (meaning "I disconnect many connections at once).

Anyways, I've seen 3 similar bug reports so far (Andy Leeman's being the third one), quite a few people confirming and not a single word from devs :/

Has anyone had any luck with other distros? Preferably apt-based... .

Revision history for this message
muszek (muszek) wrote :

I'm sorry, I forgot to mention one thing.
Often when I boot the computer and log in to Gnome (when NetworkManager tries to connect automatically) router gets killed (noticed on my girlfriend's computer. Her computer never does that.

Revision history for this message
Borlox (masterluke-729) wrote :

Same problem with Fujistu Siemens Amilo Pro V3505, Gutsy

Revision history for this message
guille (guizzzmo) wrote :

I'm experiencing the same problem with Toshiba Satellite Pro centrino ipw2200

Borlox (masterluke-729)
Changed in ipw3945:
status: New → Confirmed
Revision history for this message
Brian Murray (brian-murray) wrote :

Reviewing this bug report and its comments it seems that multiple people are experiencing this problem. Subsequently, I am confirming this bug report. For future reference you can manage the status of bug reports by clicking on the current status in the yellow line and then choosing a new status in the drop down box. You can learn more about bug statuses at http://wiki.ubuntu.com/Bugs/Status .

Revision history for this message
Brian Murray (brian-murray) wrote :

Thank you for taking the time to report this bug and helping to make Ubuntu better. This bug did not have a package associated with it, which is important for ensuring that it gets looked at by the proper developers. You can learn more about finding the right package at https://wiki.ubuntu.com/Bugs/FindRightPackage . I have classified this bug as a bug in linux-source-2.6.22.

Revision history for this message
Tokolosh (local-tosh) wrote :

I have the same problem, i was going to buy a 4965AGN wifi card anyway so i will see if that fixes the problem.

Revision history for this message
Paul Dufresne (paulduf) wrote :

Coming here because of bug #159064.
In http://ubuntuforums.org/showthread.php?t=205586&page=2 the initial bug reporter Zorael suggest to try:
sudo iwpriv set_power 0 (well, my interpretation)
like mentioned in section 3.9 of http://ipw3945.sourceforge.net/README.ipw3945
This is linked to http://www.intel.com/support/wireless/wlan/sb/cs-006205.htm.

Revision history for this message
Paul Dufresne (paulduf) wrote :

oh well, reading more carefully, it seems Zorael say that it power level is 6, which is equal to 0 which means never sleep.
So it is probably not the answer.

That said, bug #139642 have a surprising to me workaround:

Andy Leeman wrote:
In a terminal;
sudo gedit /etc/modprobe.d/blacklist

Then add;
#still trying to fix the gd wireless (or some other comment so you rememberwhy it was done)
blacklist ipv6

May be worth to try.

Revision history for this message
Andres Mujica (andres.mujica) wrote :

This is a duplicate of Bug #109887 . Please provide any further comments to this bug. Thanks

Revision history for this message
Andres Mujica (andres.mujica) wrote :

sorry for my last comment, this is not a duplicate form bug previously mentioned. my fault.

Revision history for this message
Zorael (zorael) wrote :

This now works using the iwl3945 driver bundled with the Hardy alpha (specifically, the Kubuntu KDE4 Alpha released at Feb 27).

I *did* get it to work using the script at http://linuxwireless.org/en/users/Download, blacklisting ipw3945 in /etc/modprobe.d/blacklist and autoloading iwl3945 by adding it to /etc/modules. Uncertain as to whether it works with the (new?) Gutsy iwl3945 driver, and hesitant to restore when I *FINALLY* got it to work.

All I know is that I tried the iwl3945 driver extensively a couple of months ago, as my earlier comments should suggest, with no luck at all.

The iwl3945 driver is the way to go, at any rate. If you can't get it to work with the Gutsy-bundled one, try the script at the above url. Else, deliverance might just come with Hardy.

Revision history for this message
faif (faifcn) wrote :

Fresh installed Kubuntu 8.04 KDE4 version on Thinkpad T60p. wireless module is Intel PRO/Wireless 3945ABG Mini-PCI Express Adapter. booting into KDE, knetworkmanager doesn't show any wireless devices available. However do ifconfig and iwconfig does show wlan interface is up. and modprobe iwl3945 does not encounter any problem.

Revision history for this message
Launchpad Janitor (janitor) wrote : This bug is now reported against the 'linux' package

Beginning with the Hardy Heron 8.04 development cycle, all open Ubuntu kernel bugs need to be reported against the "linux" kernel package. We are automatically migrating this bug to the new "linux" package. However, development has already began for the upcoming Intrepid Ibex 8.10 release. It would be helpful if you could test the upcoming release and verify if this is still an issue - http://www.ubuntu.com/testing . If the issue still exists, please update this report by changing the Status of the "linux" task from "Incomplete" to "New". We appreciate your patience and understanding as we make this transition. Thanks!

Revision history for this message
Leann Ogasawara (leannogasawara) wrote :

The Ubuntu Kernel Team is planning to move to the 2.6.27 kernel for the upcoming Intrepid Ibex 8.10 release. As a result, the kernel team would appreciate it if you could please test this newer 2.6.27 Ubuntu kernel. There are one of two ways you should be able to test:

1) If you are comfortable installing packages on your own, the linux-image-2.6.27-* package is currently available for you to install and test.

--or--

2) The upcoming Alpha5 for Intrepid Ibex 8.10 will contain this newer 2.6.27 Ubuntu kernel. Alpha5 is set to be released Thursday Sept 4. Please watch http://www.ubuntu.com/testing for Alpha5 to be announced. You should then be able to test via a LiveCD.

Please let us know immediately if this newer 2.6.27 kernel resolves the bug reported here or if the issue remains. More importantly, please open a new bug report for each new bug/regression introduced by the 2.6.27 kernel and tag the bug report with 'linux-2.6.27'. Also, please specifically note if the issue does or does not appear in the 2.6.26 kernel. Thanks again, we really appreicate your help and feedback.

Revision history for this message
Andres Mujica (andres.mujica) wrote :

Hi, i'm closing this bug report as Won't Fix, mainly because the IPW3945 module was deprecated by upstream some time ago (http://ipw3945.sourceforge.net/)

Also for the lack of response from the previous comments

Thanks in advance !!

Changed in linux (Ubuntu):
status: Incomplete → Won't Fix
Revision history for this message
Andres Mujica (andres.mujica) wrote :

Marking as invalid as upstream is deprecated

Changed in ipw3945:
status: Confirmed → Invalid
To post a comment you must log in.
This report contains Public information  
Everyone can see this information.

Other bug subscribers

Remote bug watches

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