ipw3945 not coming up after second suspend

Bug #58247 reported by Gilbert
44
Affects Status Importance Assigned to Milestone
acpi-support (Ubuntu)
Invalid
Undecided
Unassigned
linux-source-2.6.17 (Ubuntu)
Fix Released
Undecided
Unassigned

Bug Description

i'm using dapper but with edgy's 2.6.17-6 kernel (compiled from source deb package in order to keep any Ubuntu patches and using gcc 4.0) as I have to use acpi=off with dapper's stock kernel, so no suspend in stock dapper. everything works great with the exception of suspend annoyances as always =)

when resuming from the first suspend, network manager is able connect to the wireless AP without problems.

resuming after the second suspend, there is no listed network interface in ifconfig or iwconfig. the card still exists in lspci. i've tried stopping ipw3945d-`uname -r`, unloading the module then reloading the module (noticed that ipw3945d starts on its own after the module is loaded) with the same results.

don't really think its a kernel issue but may be wrong. any help is appreciated.

Revision history for this message
Gilbert (r1z83) wrote :

to lessen confusion, by second suspend i mean suspending after resuming from a previous suspend and by no listed network interface in ifconfig or iwconfig it only applies to the wifi card. i can still see/use the wired network interface.

forgot to mention i have to reboot after the second suspend in order to get the wireless network interface back.

Revision history for this message
Gilbert (r1z83) wrote :

performed a clean install of knot2 (daily 20060914) and got hiberante working with the following added to /etc/acpi/suspend.d/70-modules-unload.sh

killall ipw3945d-`uname -r`
rm /var/run/ipw3945d*pid

and adding ipw3945 to the module list in /etc/default/acpi-support along with removing the splash screen in grub.

however I still have issues after a second resume from suspend to disk as the ipw3945 wireless card does not appear in ifconfig or iwconfig even though it's listed in lspci and module/daemon appear to start without problems (nothing in dmesg even with debug=1 when modprobing ipw3945 manunally).

Revision history for this message
Gilbert (r1z83) wrote :

Still having the same issue on a fresh 6.10 beta install with updates (2.6.17-10-generic)

Revision history for this message
Gilbert (r1z83) wrote :

nothing has changed in ubuntu 6.10 release 2.6.17-10-generic. the wifi devices disappears after a second suspend and reloading the module/daemon doesnt help. tried adding the module to /etc/default/acpi-support and appending "ipw3945d-`uname -r` --kill" in /etc/acpi/suspend.d/55-down-interfaces.sh

will likely try suspend2 and see how that goes.

works fine in fedora core 6 even after 5 resumes from suspend to disk and/or memory :(

Revision history for this message
Quanta (yohann-hamon) wrote :

Same problem for me (Fujitsu siemens AmiloProV3505 with Kubuntu Edgy - generic kernel).

To have my wifi card working I have first to start under Windows, then to reboot my computer with Ubuntu (at this point the wifi light on my laptop is on). Then my wifi card work perfectly.
If I reboot mu computer once again directly in with ubuntu the wifi card does not exist any more (and the wifi light stay off all along).

I read something about a problem with the kill switch turned on or of...

Revision history for this message
Gilbert (r1z83) wrote :

the wifi light stays on for me through out the entire ordeal.

Revision history for this message
Peter Whittaker (pwwnow) wrote :

Thanks this bug report. To make progress, some more information will be needed.

Please attach (by commenting on the bug) to the bug report the output of the following commands:

uname -a

sudo lspci -vv

sudo lspci -vvn

sudo dmidecode

It would also be helpful if you could try to hibernate/suspend and after that fails, restart your system and attach /var/log/kern.log.0 as well.

Since there will be a lot of output, please do be sure to attach rather than commenting -in-line. Thanks again!

Changed in linux-source-2.6.17:
status: Unconfirmed → Needs Info
Revision history for this message
Elliot Hughes (elliot-hughes) wrote :

Rejecting because of no response to the request for more information. If you have any more information please reopen it. Thanks once more for reporting this bug.

Changed in acpi-support:
assignee: nobody → elliot-hughes
status: Unconfirmed → Rejected
Revision history for this message
FiNeX (finex) wrote :

I've got the same problem on feisty fawn (7.04): after the second resume from hibernation, iwconfig show:
  lo no wireless extensions.
  eth1 no wireless extensions.
  eth0 no wireless extensions.

Simply restarting the ipw3945 daemon solved the problem. (from a root shell I've just typed: "ipw3945d-2.6.20-15-generic"). The output of this command is:

ipw3945d - regulatory daemon
Copyright (C) 2005-2006 Intel Corporation. All rights reserved.
version: 1.7.22
Intel PRO/Wireless 3945ABG Network Connection found at:
 /sys/bus/pci/drivers/ipw3945/0000:04:00.0
Daemon launched as pid 10332. Exiting.

A new iwconfig found the right interfaces:

lo no wireless extensions.
eth1 unassociated ESSID:off/any
          Mode:Managed Frequency=nan kHz Access Point: Not-Associated
          Bit Rate:0 kb/s Tx-Power:16 dBm
          Retry limit:15 RTS thr:off Fragment thr:off
          Encryption key: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:0 Missed beacon:0
eth0 no wireless extensions.

I attach the LSPCI output for more information on my hardware.

The kernel I'm using is: 2.6.20-15-generic #2 SMP Sun Apr 15 07:36:31 UTC 2007 i686 GNU/Linux

Revision history for this message
FiNeX (finex) wrote :

The bug still exist. I've posted some additional informations as asked here: https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.17/+bug/58247/comments/8

Changed in acpi-support:
status: Rejected → Confirmed
Revision history for this message
Peter Whittaker (pwwnow) wrote :

FiNeX, please refer to https://bugs.launchpad.net/ubuntu/+source/linux-source-2.6.17/+bug/58247/comments/7 - if you can provide all of the requested information under 2.6.20, this report can be confirmed and moved along; otherwise, it will be closed. Thanks!

Revision history for this message
FiNeX (finex) wrote :

The attachment contain the outputs of the following commands:

uname -a
sudo lspci -vv
sudo lspci -vvn
sudo dmidecode

Revision history for this message
FiNeX (finex) wrote :

With the latest dist-upgrade the system now work correctly: I've hibernated more than three times and, after the resume, network didn't fail.

Now Feisty (with the kernel 2.6.20-16-generic) is ok.

Thanks! :-) :-) :-)

Revision history for this message
shrd (shrd) wrote :

i see some bugs in feisty with this driver after some suspend/resume.
wifi dead, led black, one cpu at 100%, freeze of keyboard, mouse ok, but not possibility to reboot with mouse command, and no ctrl alt backspace (keyboard dead)

Revision history for this message
shrd (shrd) wrote :

of course i am not sure ipw3945 is the cause, i suspect, adobe flash player to be responsible too.

Revision history for this message
Jyrki Pulliainen (jyrki-pulliainen) wrote :

Gutsy ships with ipw3945 1.2.1 which fixes some issues with driver not waking up in some cases. It fixed suspend problems on my laptop.

I'm not sure does this driver ship with later Feisty kernels, but at least the 2.6.20-13 and previous shipped with 1.2.0. You can override this manually installing the module though.

Changed in acpi-support:
assignee: elliot-hughes → nobody
Revision history for this message
Matthew Garrett (mjg59) wrote :

Kernel issue rather than acpi-support one

Changed in acpi-support:
status: Confirmed → Invalid
Revision history for this message
FiNeX (finex) wrote :

Ok, Gutsy it's OK, we should close this bug and mark is as fixed :)

Revision history for this message
FiNeX (finex) wrote :

The bug seems being fixed on Gutsy Gibbon.

Changed in linux-source-2.6.17:
status: Incomplete → Fix Released
Revision history for this message
shrd (shrd) wrote :

good news if it is fixed on gutsy
I will test this week end after the dist upgrade.

Revision history for this message
Sarah Kowalik (hobbsee-deactivatedaccount) wrote :

hmmm, i still seem to be getting this. will have to test more. will reopen, if i can reproduce this multiple times.

Revision history for this message
xandr55 (alp05) wrote :

not sure if this helps any, but on a fresh install of Gutsy, on Dell Inspiron 5100, no wired network connects correctly after a hibernate. I tried adding STOP_SERVICES="networking" to the /etc/default/acpi-support as in bug #69426, with no luck... only a restart seems to get it to work correctly.

Revision history for this message
Paul (treepata) wrote :

Hey there,

I am no sure if the bug I am experiencing is this one, as it could also be <A href="https://launchpad.net/ubuntu/+source/network-manager/+bug/69892">bug 69892</A> or <A href="https://bugs.launchpad.net/ubuntu/+bug/69426">bug 68426</A>.

But in essence, I cannot reconnect to the internet through Network Manager after the system resumes from suspend or hibernation. It shows all the networks and I can try to connect to them, but even with both the lights green it won't eventually establish a connection.

I am running Xubuntu 8.10 on my IBM x41 and the problem first showed up when I upgraded from 8.04. I had issues with the Network Manager not being able to make a connection in the past, but I solved this by adding the following line to acpi-support:

# Add services to this list to stop them before suspend and restart them in
# the resume process.
STOP_SERVICES="networking"

However, this solution doesn't work in the current situation. And what is different, is that it still tries to establish a connection and almost succeeds in this, but it hangs after both lights turn green.

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

Duplicates of this bug

Other bug subscribers

Remote bug watches

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