On my laptop i have a built in Centrino wireless 1000 wifi adapter that uses the iwlagn driver. up through kernel 2.6.38, i am able to suspend my laptop and resume without problems. When i update to a newer kernel, 2.6.39 2.6.40 or 3.0 and higher, upon resume the device fails to associate with the previous access point. It complains that it cannot flush all fifo queues(or pipes cant remember which) and that direct probing fails. It scans properly, and i can attempt to connect however it always fails. I have tried stopping and restarting the network, unloading and reloading the module, after suspend, and unloading before and reloading after, and still no luck. I'm running gentoo on my laptop but i have confirmed that it is not a bug with gentoo since i have confirmed the same behavior on fedora with a 3.0+ kernel. Also, i have tried it with different access points so its not a network problem, and i ruled out a bug in NetworkManager because I have confirmed its failure to associate with both wpa_gui (wpa_supplicant) and wicd. I'm wondering if anyone else has had the same problem, or if anyone knows a way to fix it, or has any other suggestions of what to try. I've done searches on this subject and havent seen any results with the new kernel, so possibly its a simple fix that i've overlooked?

Any help is appreciated.

Yes, I have exactly the same problem. Laptop is hp pavilion dv7-6195us. I have reported this to both Ubuntu and Fedora bugzilla and, as suggested by Ubuntu, to LKML. There seemed to be a lot of interest at first on LKML but, so far, no additional responses or actions. Here is the link to my LKML post that also links into the bugzilla reports:
(remove spaces)
https:// lkml. org / lkml /2011 /10/26/6
I suspect that this dv7 is not used by many linux users. (It also uses a ton of power, 50w plus, in linux but hardly any in windows.)

-gene

well I think i've narrowed the problem down to the wireless card/driver iwlagn. know for a fact that it was revised between kernel 2.6.38 and 2.6.39. It was merged with the iwlcore module, or at least the functions were. At some point i may try to compile the old version of iwlagn with the 3.0 kernel and see if that makes a difference.

What wireless do you have? mine is an intel centrino wireless-N 1000. I'm curious to see if its the module problem or just hardware specific.

Mine is same type as yours. My hardware details can be seen here:
https://bugs.launchpad.net/ubuntu/+s...ux/+bug/880648
e.g.,
0d:00.0 Network controller: Intel Corporation Centrino Wireless-N 1000
Subsystem: Intel Corporation Centrino Wireless-N 1000 BGN
Flags: bus master, fast devsel, latency 0, IRQ 53
Memory at c4500000 (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 8c-a9-82-ff-ff-a0-f4-7a
Kernel driver in use: iwlagn
Kernel modules: iwlagn

Glad you are looking at this! What type of laptop is yours?

https://lkml.org/lkml/2011/12/12/256

wwguy in above thread suggested a fix that works for me.

-gene

excellent, i can confirm that it works for me as well. also, if you save a .conf file in /etc/modprobe.d with bt_coex_active=0 then it will auto load on resume so you dont have to add the option when you modprobe, if you haven't done this already. Thanks for the help. and thanks to wwguy as well.

Glad it worked for you too. Yes, I also created an /etc/modprobe.d file called iwlagn.conf containing the single line

options iwlagn bt_coex_active=N

so I don't have to manually run the modprobe commands.

Unfortunately the latest f16 kernel, 3.1.6-1.fc.x86_64, breaks it again and can't reconnect wifi after sleep without a complete reboot. Still works with older kernels, e.g., 3.1.5-6 and 3.1.5-2.

for the options change the module name to iwlwifi. it doesnt seem to load the iwlagn module anymore for me, though my kernel is 3.2
I think it might be the same problem for you.

Yes, found out about this a while back but forgot to write back about what I found. Thanks for posting the solution.
Re: https://lkml.org/lkml/2012/1/7/119