ipw3945 produces 99% cpu, making my laptop unusable

I'm running Dapper with kernel version 2.6.5-23-686 on an HP Pavillion dv8000 and an Intel pro/wireless 3945abg. I'm getting the same problem, except it's only running one of my cores at 100% (also a core duo). The problem doesn't seem to happen unless I use the hardware toggle switch to disable the wireless card and leave the computer sitting for a couple hours.

Any help would be welcome.

Hello!

I have the same problems running Ubuntu 6.06 Dapper on an IBM Thinkpad Z61m with the Kernel
2.6.15-26-686 #1 SMP PREEMPT Fri Jul 7 19:48:22 UTC 2006 i686 GNU/Linux

One processor is at 100%, the other is in normal operation.

I haven't found a solution for this when usign the smp-kernel together with these restricted modules.

As far as I observed now the ipw3945/0 does NOT run everytime I boot my laptop. It happens from time to time, but I can't say when it happens and when not.

Hello,

This happens on my laptop, too.
(I had to disable the boot-time loading of this module, but it's not a solution)

Clevo M550N, CoreDuo, ipw3945.

Same for me on acer aspire 9412 (CoreDuo).
dmesg output:

Aug 6 12:57:18 clt kernel: [17179668.060000] ipw3945: Intel(R) PRO/Wireless 3945 Network Connection driver for Linux, 1.0.5m
Aug 6 12:57:18 clt kernel: [17179668.060000] ipw3945: Copyright(c) 2003-2006 Intel Corporation
Aug 6 12:57:18 clt kernel: [17179668.064000] ipw3945: Detected Intel PRO/Wireless 3945ABG Network Connection
Aug 6 12:57:18 clt kernel: [17179668.992000] ipw3945: Detected geography ABG (13 802.11bg channels, 23 802.11a channels)
Aug 6 12:57:20 clt kernel: [17179669.456000] ipw3945: Microcode SW error detected. Restarting.
Aug 6 12:57:20 clt kernel: [17179669.456000] ipw3945: Error Reply type 0x00000005 cmd LEDS_CMD (0x48) seq 0x0408 ser 0x004A0000
Aug 6 12:57:20 clt kernel: [17179669.456000] ipw3945: Can't stop Rx DMA.

The last four lines repeat till I kill /sbin/ipw3945d-2.6.15-26-686

I tried the ipw3945 driver version 1.1.0, too even with ieee80211 driver 1.1.14 to no avail. Still the same dmesg output.

This error occurs on:
  Linux version 2.6.15-26-686 (buildd@terranova) (gcc version 4.0.3 (Ubuntu 4.0.3-1ubuntu5)) #1 SMP PREEMPT Thu Aug 3 03:13:28 UTC 2006

So I tried the 386 kernel image to verify if it is specific to the smp kernel and I got the exactly same results on the 386 kernel, too.

Best Regards,
Christian

Update:

There appears another repeating error message in the log output:
"17197492.596000] ipw3945: Error Reply type 0x00000005 cmd ADD_STA (0x18) seq 0x0407 ser 0x004A0000"

Full sequence of error output:
[17197477.064000] ipw3945: Detected Intel PRO/Wireless 3945ABG Network Connection
[17197477.880000] ipw3945: Detected geography ABG (13 802.11bg channels, 23 802.11a channels)
[17197488.768000] ipw3945: Intel(R) PRO/Wireless 3945 Network Connection driver for Linux, 1.0.5m
[17197488.768000] ipw3945: Copyright(c) 2003-2006 Intel Corporation
[17197488.768000] ipw3945: Detected Intel PRO/Wireless 3945ABG Network Connection
[17197489.568000] ipw3945: Detected geography ABG (13 802.11bg channels, 23 802.11a channels)
[17197490.028000] ipw3945: Microcode SW error detected. Restarting.
[17197490.028000] ipw3945: Error Reply type 0x00000005 cmd LEDS_CMD (0x48) seq 0x0408 ser 0x004A0000
[17197490.028000] ipw3945: Can't stop Rx DMA.
[17197492.140000] ipw3945: Detected geography ABG (13 802.11bg channels, 23 802.11a channels)
[17197492.596000] ipw3945: Microcode SW error detected. Restarting.
[17197492.596000] ipw3945: Error Reply type 0x00000005 cmd ADD_STA (0x18) seq 0x0407 ser 0x004A0000
[17197492.596000] ipw3945: Can't stop Rx DMA.
...

I'm running into the exact same issue. I've checked some other forums regarding the driver and it seems like there are quite a few people experiencing this issue. Has anyone here heard of any fixes yet? I'm kind of trolling everywhere hoping that there's some magic somewhere ;-) I need this wireless working mostly as a matter of convenience for work :-/

Some of my ouput:

DMESG
------------
[4298156.380000] ipw3945: Detected geography ABG (11 802.11bg channels, 13
802.11a channels)
[4298156.626000] ipw3945: Microcode SW error detected. Restarting.
[4298156.626000] ipw3945: Error Reply type 0x000000FF cmd LEDS_CMD (0x48) seq
0x040A ser 0x00030000
[4298156.626000] ipw3945: Can't stop Rx DMA.
[4298157.481000] ipw3945: Detected geography ABG (11 802.11bg channels, 13
802.11a channels)
[4298157.788000] ipw3945: Microcode SW error detected. Restarting.
[4298157.788000] ipw3945: Error Reply type 0x000000FF cmd TX (0x1C) seq 0x0000
ser 0x00030000
[4298157.788000] ipw3945: Can't stop Rx DMA.
[4298158.643000] ipw3945: Detected geography ABG (11 802.11bg channels, 13
802.11a channels)
[4298158.898000] ipw3945: Microcode SW error detected. Restarting.
[4298158.898000] ipw3945: Error Reply type 0x000000FF cmd TX (0x1C) seq 0x0000
ser 0x00030000
[4298158.898000] ipw3945: Can't stop Rx DMA.
[4298159.751000] ipw3945: Detected geography ABG (11 802.11bg channels, 13
802.11a channels)
[4298160.015000] ipw3945: Microcode SW error detected. Restarting.
[4298160.015000] ipw3945: Error Reply type 0x000000FF cmd TX (0x1C) seq 0x0000
ser 0x00030000
[4298160.015000] ipw3945: Can't stop Rx DMA.
[4298160.871000] ipw3945: Detected geography ABG (11 802.11bg channels, 13
802.11a channels)
[4298161.125000] ipw3945: Microcode SW error detected. Restarting.
[4298161.125000] ipw3945: Error Reply type 0x000000FF cmd TX (0x1C) seq 0x0000
ser 0x00030000
[4298161.126000] ipw3945: Can't stop Rx DMA.
[4298161.980000] ipw3945: Detected geography ABG (11 802.11bg channels, 13
802.11a channels)
[4298162.227000] ipw3945: Microcode SW error detected. Restarting.
[4298162.227000] ipw3945: Error Reply type 0x000000FF cmd DAEMON (0x10) seq
0x0406 ser 0x00030000

MESSAGES
-----------------
Aug 7 19:01:01 localhost kernel: [4298149.650000] ipw3945: Detected geography
ABG (11 802.11bg channels, 13 802.11a channels)
Aug 7 19:01:02 localhost kernel: [4298150.757000] ipw3945: Detected geography
ABG (11 802.11bg channels, 13 802.11a channels)
Aug 7 19:01:03 localhost kernel: [4298151.911000] ipw3945: Detected geography
ABG (11 802.11bg channels, 13 802.11a channels)
Aug 7 19:01:04 localhost kernel: [4298153.027000] ipw3945: Detected geography
ABG (11 802.11bg channels, 13 802.11a channels)
Aug 7 19:01:05 localhost kernel: [4298154.131000] ipw3945: Detected geography
ABG (11 802.11bg channels, 13 802.11a channels)

etc

The WiFi light goes crazy for a while while these plaster my logs and dmesg then it all seems to settle and work for a bit. I've tried generating tons of traffic, starting multiple sessions accross different networks, leaving it alone ... nothing seems to really kick off this activity it just *happens* randomly

Hi SpEnTBoY,

It seems that the guy running the site http://www.die-welt.net has made the Wireless card in the IBM Thinkpad Z61m working. This is the same notebook I'm using (see above). But he uses Debian.

See this site:
http://www.die-welt.net/index.php/blog/66/Debian_Linux_on_an_IBM_Lenovo_ThinkPad_Z61m

This should give you some hope that it's possible in general to make this card working.

hi infu,

I've pretty much done everything he's mentioned plus I've tried about 4 different distros' default AND compiled the new driver manually. I'm running the latest Ubuntu currently but have tried many. The issue is, it does work periodically ... but when it doesn't it makes the WiFi light on my Dell D620 go crazy while it fills the logs with what I posted ... eventually it levels off and the WiFi light is solid and everything is ok for a bit until the light goes crazy again and more errors in the log file and dmesg.

I'm hoping the machine is just too new and requires a newer firmware or something. I've scoured the net and there are quite a few people who've run into this issue :-)

--
SpEnT

I have the same issue with dapper on lg s1 mduo laptop (686 kernel, restricted modules installed). Periodically, ipw3945 is choking one of my cores (constant 50% cpu). I don't know how to stop the process, as it seems to be unkillable... I sure hope this is sorted out.

I noticed that this driver works fine with a Linksys WRT54G WLAN router but produces those microcode software errors with a Dlink DWL-650G.

Regards,
Christian

Hei Christian, may I have your contact? Either e-mail or IM? I'm buying the same laptop and using ubuntu dapper 6.06.

This problem surely occurs to me aswell. Not just that, the SATA driver don't work. It work but this is what I get from /var/log/messages :

Aug 18 03:19:27 stubbe kernel: [4298880.086000] ata1 is slow to respond, please be patient
Aug 18 03:19:56 stubbe kernel: [4298906.385000] ata1: status=0x50 { DriveReady SeekComplete }
Aug 18 03:19:56 stubbe kernel: [4298906.419000] sda: Current: sense key: No Sense
Aug 18 03:19:56 stubbe kernel: [4298906.419000] Additional sense: No additional sense information
Aug 18 03:20:22 stubbe kernel: [4298934.928000] ata1 is slow to respond, please be patient
Aug 18 03:20:51 stubbe kernel: [4298961.227000] ata1: status=0x50 { DriveReady SeekComplete }
Aug 18 03:20:51 stubbe kernel: [4298961.261000] sda: Current: sense key: No Sense
Aug 18 03:20:51 stubbe kernel: [4298961.261000] Additional sense: No additional sense information
Aug 18 03:20:51 stubbe kernel: [4298961.263000] sr: Current [descriptor]: sense key: Aborted Command
Aug 18 03:20:51 stubbe kernel: [4298961.263000] Additional sense: Scsi parity error
Aug 18 03:26:37 stubbe kernel: [4299309.674000] ata1 is slow to respond, please be patient
Aug 18 03:27:06 stubbe kernel: [4299335.973000] ata1: status=0x50 { DriveReady SeekComplete }
Aug 18 03:27:06 stubbe kernel: [4299336.007000] sda: Current: sense key: No Sense
Aug 18 03:27:06 stubbe kernel: [4299336.007000] Additional sense: No additional sense information
Aug 18 03:27:06 stubbe kernel: [4299336.009000] sr: Current [descriptor]: sense key: Aborted Command
Aug 18 03:27:06 stubbe kernel: [4299336.009000] Additional sense: Scsi parity error
Aug 18 03:38:06 stubbe kernel: [4299999.400000] ata1 is slow to respond, please be patient
Aug 18 03:38:35 stubbe kernel: [4300025.699000] ata1: status=0x50 { DriveReady SeekComplete }
Aug 18 03:38:35 stubbe kernel: [4300025.732000] sda: Current: sense key: No Sense
Aug 18 03:38:35 stubbe kernel: [4300025.732000] Additional sense: No additional sense information
Aug 18 03:59:13 stubbe kernel: [4301266.402000] ata1 is slow to respond, please be patient
Aug 18 03:59:42 stubbe kernel: [4301292.701000] ata1: status=0x50 { DriveReady SeekComplete }
Aug 18 03:59:42 stubbe kernel: [4301292.735000] sda: Current: sense key: No Sense
Aug 18 03:59:42 stubbe kernel: [4301292.735000] Additional sense: No additional sense information
Aug 18 04:14:38 stubbe kernel: [4302192.008000] ata1 is slow to respond, please be patient
Aug 18 04:14:45 stubbe kernel: [4302198.307000] sr 0:0:1:0: ioctl_internal_command return code = 8000002
Aug 18 04:14:45 stubbe kernel: [4302198.307000] : Current [descriptor]: sense key: Aborted Command
Aug 18 04:14:45 stubbe kernel: [4302198.308000] Additional sense: Scsi parity error
Aug 18 04:14:45 stubbe kernel: [4302198.358000] ata1: status=0x50 { DriveReady SeekComplete }
Aug 18 04:14:48 stubbe kernel: [4302198.358000] sda: Current: sense key: No Sense
Aug 18 04:14:48 stubbe kernel: [4302198.358000] Additional sense: No additional sense information
Aug 18 04:20:41 stubbe kernel: [4302554.722000] ata1...

This bug is definitely still relevant.
The wifi led on my laptop (acer 5562wxmi core duo) is going wild, i can notice ipw daemon running:

#ps aux | grep ipw
root 3273 0.0 0.0 0 0 ? S< Aug22 0:00 [ipw3945/0]
root 3274 0.0 0.0 0 0 ? S< Aug22 0:00 [ipw3945/1]
root 3275 0.0 0.0 0 0 ? S< Aug22 0:00 [ipw3945/0]
root 3276 0.0 0.0 0 0 ? S< Aug22 0:00 [ipw3945/1]
root 3362 0.0 0.0 1616 352 ? S< Aug22 0:00 /sbin/ipw3945d-2.6.15-22-686 --quiet

once i kill the ipw daemon it stops.

On Mon, 21 Aug 2006 23:34:25 +0200, stubbe <email address hidden> wrote:

Hi, it the SATA bug seems to be this one:
    https://launchpad.net/bugs/38760

I could make the ata errors occur less frequently by adding a line with
the entry
  ahci
to the file /etc/modules. Those freezes didn't go away completely.

> Hei Christian, may I have your contact? Either e-mail or IM? I'm buying
> the same laptop and using ubuntu dapper 6.06.
>
> This problem surely occurs to me aswell. Not just that, the SATA driver
> don't work. It work but this is what I get from /var/log/messages :
>
> Aug 18 03:19:27 stubbe kernel: [4298880.086000] ata1 is slow to respond,
> please be patient
> Aug 18 03:19:56 stubbe kernel: [4298906.385000] ata1: status=0x50
> { DriveReady SeekComplete }
> Aug 18 03:19:56 stubbe kernel: [4298906.419000] sda: Current: sense key:
> No Sense
> Aug 18 03:19:56 stubbe kernel: [4298906.419000] Additional sense: No
> additional sense information

...

>
> Is this bug reported yet? Or do you not have the same problem?
>
> Thanks.
>

I had to do another thing to prevent the nvidia driver to hang my system
when switching to the text console - disable the framebuffer.
Look for a line in /boot/grub/menu.lst starting with:
  # defoptions= ...
and change it to:
  # defoptions=quiet video=vga16fb:off video=vesafb:off

Afterwards you have to type "update-grub" to update your boot menu.
Note: All actions as user root. It will disable the graphical boot.

What might interest you is that the wireless driver of the 2.6.15-26-686
kernel started to work fully with a WRT-54G and a Dlink DWL-650g after
I downloaded and installed all of the Acer updates for the preinstalled
Windows XP.

What a strange experience.

I also had problems with audio during video playback with xine. I had
switch xine to use the OSS emulation from the ALSA driver and to create
a file ~/.asoundrc with the content
pcm.dsp0 {
     type plug
     slave.pcm "dmix"
}
# mixer0 can stay unchanged, because
# it isn't used anyway, I guess ;)
ctl.mixer0 {
     type hw
     card 0
}

I hope there will be a kernel release fixing those nasty SATA errors.
kernel-2.6.17-2-686 from edgy does not even boot.

What else does not work:
   - Logitech webcam, but I read somewhere that a driver for the Logitech
webcam
     integrated in this notebook is in the works:
     ~$ lsusb
        Bus 005 Device 002: ID 046d:0892 Logitech, Inc.).
   - TI card reader
     ~$ lspci
        0000:0a:06.0 CardBus bridge: Texas Instruments: Unknown device 8039
        0000:0a:06.2 Mass storage controller: Texas Instruments: Unknown
device 803b

Cheers,
Christian

PS: Doom3 is perfectly playable with OSS audio output on a resolution of
900x600.
     Here is a modline (/etc/X11/xorg.conf file, section "Modes")for it:
       Modeline "960x600" 45.28 960 992 1160 1192 600 612 618 631

Wow that's one complete reply :). Thank you for spare your time for it.

  In case you don't know yet the dial-up modem also won't work yet. I read it somewhere that the developer of the chip (texas instrument or something like that) didn't release any specifications for the kernel developers.

  The wireless have already work since I first installed ubuntu the hardware was detected from the first place, but I havent really test the connection, since there're not much hotspots around here.

  Not sure what about the card readers tho, havent got a chance to try it out since I don't have any memory cards. So well that leaves us the camera and modem unusable.

  It is said that the SATA controller could use the piix driver. But i've tried both AHCI module and PIIX, still freezes, but yes, less frequent.

  Well, untill these problems solved in edgy or the latest, guess I need to stick to windows for awhile.

  BTW thank you very much and very nice to meet you man.

Christian Mayrhuber <email address hidden> wrote:
  On Mon, 21 Aug 2006 23:34:25 +0200, stubbe
wrote:

Hi, it the SATA bug seems to be this one:
https://launchpad.net/bugs/38760

I could make the ata errors occur less frequently by adding a line with
the entry
ahci
to the file /etc/modules. Those freezes didn't go away completely.

> Hei Christian, may I have your contact? Either e-mail or IM? I'm buying
> the same laptop and using ubuntu dapper 6.06.
>
> This problem surely occurs to me aswell. Not just that, the SATA driver
> don't work. It work but this is what I get from /var/log/messages :
>
> Aug 18 03:19:27 stubbe kernel: [4298880.086000] ata1 is slow to respond,
> please be patient
> Aug 18 03:19:56 stubbe kernel: [4298906.385000] ata1: status=0x50
> { DriveReady SeekComplete }
> Aug 18 03:19:56 stubbe kernel: [4298906.419000] sda: Current: sense key:
> No Sense
> Aug 18 03:19:56 stubbe kernel: [4298906.419000] Additional sense: No
> additional sense information

...

>
> Is this bug reported yet? Or do you not have the same problem?
>
> Thanks.
>

I had to do another thing to prevent the nvidia driver to hang my system
when switching to the text console - disable the framebuffer.
Look for a line in /boot/grub/menu.lst starting with:
# defoptions= ...
and change it to:
# defoptions=quiet video=vga16fb:off video=vesafb:off

Afterwards you have to type "update-grub" to update your boot menu.
Note: All actions as user root. It will disable the graphical boot.

What might interest you is that the wireless driver of the 2.6.15-26-686
kernel started to work fully with a WRT-54G and a Dlink DWL-650g after
I downloaded and installed all of the Acer updates for the preinstalled
Windows XP.

What a strange experience.

I also had problems with audio during video playback with xine. I had
switch xine to use the OSS emulation from the ALSA driver and to create
a file ~/.asoundrc with the content
pcm.dsp0 {
type plug
slave.pcm "dmix"
}
# mixer0 can stay unchanged, because
# it isn't used anyway, I guess ;)
ctl.mixer0 {
type hw
card 0
}

I hope there will be a kernel release fixing those nasty SATA errors.
kernel-2.6.17-2-686 from edgy does not even boot.

...

I have been having the same ipw3945 problem as everybody else. Then today at work, the problem went away. It turns out that at home it has problems, but at work it doesn't. Here at work they are using a Cisco router, but at home, I have a WRTP54G. Earlier somebody mentioned that they got their wireless card to work with a WRTP54G. Could it be that I need to upgrade my firmware?

It turns out it didn't work with my work's sisco router. It started working, but promptly started freaking out.

It looks like your Cisco wireless access point is tripping a firmware bug.

Have you got any closer to solving this bug? My laptop overheats too, (same reason i guess) so if you need any information just ask.

Btw acpi -t gives 50-51 C that is too much right? Cpu is intel core 2 duo T2300

The same 99% CPU, fans full-speed problem occurs on my Dell Inspiron 6400 running Dapper with all updates.

I have noticed that the problem is much worse during cold weather. On winter morning, the room temperature was below 5C, and the problem would happen every time it started. If I turned turned the machine off as soon as possible, then back on, it would still run at 99% CPU. If I left the thing running for a few minutes before rebooting the problem would go away and everything would start normally: I assume that things had warmed up internally.

Now that it is spring in the southern hemisphere, the system usually starts up reliably except on very cold mornings, and the same trick works.

Perhaps there is something wrong with the temperature detection (it interprets the very cold temperature as very hot), though I have no idea how that might make ipw3945 run at full-tick.

One other point is that the fans turn on very early on in the boot process---after grub, after the splash screen appears, but before the first line of text.

Stephen

Ühel kenal päeval, T, 2006-10-31 kell 22:16, kirjutas Stephen Irons:
> The same 99% CPU, fans full-speed problem occurs on my Dell Inspiron
> 6400 running Dapper with all updates.
>
> I have noticed that the problem is much worse during cold weather. On
> winter morning, the room temperature was below 5C, and the problem would
> happen every time it started. If I turned turned the machine off as soon
> as possible, then back on, it would still run at 99% CPU. If I left the
> thing running for a few minutes before rebooting the problem would go
> away and everything would start normally: I assume that things had
> warmed up internally.
>
> Now that it is spring in the southern hemisphere, the system usually
> starts up reliably except on very cold mornings, and the same trick
> works.
>
> Perhaps there is something wrong with the temperature detection (it
> interprets the very cold temperature as very hot), though I have no idea
> how that might make ipw3945 run at full-tick.
>
> One other point is that the fans turn on very early on in the boot
> process---after grub, after the splash screen appears, but before the
> first line of text.
>
> Stephen
>

Hi,

I may have a solution to this nasty problem
http://bordeaux.freshrpms.net/rpm.html?id=1044 and from there you should
go to http://bughost.org/ipw3945/
I've not been able to monitor the activity of my laptop's CPUs after
taking the steps which are described in the bugshot page, but it seems
to me that the fan does not work as hard anymore as it used to.
What you say about cold wheather is interesting. I am not sure, but it
seems to me that the fan stops spinning so much when acpi -t gives about
49-51 C

PS applying the firmware and daemon updates may cause that you have to
load the module and the daemon manually after resuming from suspend or
hibernate but as that can be done with two simple commands it should not
be much of a trouble.

What i told about the fan stopping spinning, when temp reaches 49-51 C is probably false at least in some cases.
When I turned on my laptop this morning I noticed that the fan worked very often though temp was about 49 as usually. I suspended the machine and then resumed hoping that it would fix the problem. After resuming I started to watch acpi -t. At first temp was 39 C and the fan was calm. After some time acpi -t gave about 49-50 C and the fan kept starting and stopping after about every 30 seconds or so. The problem is actually still present. Maybe a reboot would help, but I can't try it at the moment.

Still, I'm quite sure that I've seen acpi -t giving 49-50 C and the fan has been completely still.

Maybe the fan is not using the temperature readings to decide if it should start or not?
I don't know what else could it be using, though...

Is this still an issue?

For me this nos not an issue anymore. I did not had any problems for some a longer time (at least a year, I think).

Please reopen if someone encounters this issue after all.