Description of problem: I'm often seeing this message in dmesg [ 1201.969500] ipw2200: Firmware error detected. Restarting. Followed by a network disconnect and sometimes the system is not able to recover at all, requiring a system restart. Version-Release number of selected component (if applicable): 2.6.38.6-26.rc1.fc15.i686.PAE ipw2200-firmware-3.1-5.fc15.noarch How reproducible: randomly happens every few hours Actual results: network disconnects Expected results: uninterrupted network connection Additional info: 03:07.0 Network controller [0280]: Intel Corporation PRO/Wireless 2200BG [Calexico2] Network Connection [8086:4220] (rev 05)
I have a ipw2200-equipped box I use for some hobby work. I've seen the firmware restarts from time to time over the years, but it always recovers quickly. FWIW, I can't remember the last time ipw2200 had a patch. Perhaps you could try to replicate with an upstream git tree close to the Fedora kernel version, then try to bisect it?
Hi John, so you don't think it's a firmware issue? That's what I would have expected from the error message. I have also seen this with a vanilla 2.6.39. This is the first time I'm using fedora on that machine so I can't tell if older version are any better. But I might try some older fedora kernel.
It might be a firmware issue, or it might be the driver not talking nicely enough to the firmware. In any case, we won't be getting a firmware update. The error has been around for a long time, but I don't recall it being permanent for anyone in the past.
This has only become a killer with Kernel 2.6.40-4.fc15.i686. Wireless works for a while and then locks up (the "connected" icon in the NetworkManager panel has an "x" on it and there is no connection). I get the following messages in dmesg: ipw2200: Firmware error detected. Restarting. ipw2200: Failed to up device ipw2200 0000:02:03.0: PCI INT A disabled ipw2200: Intel(R) PRO/Wireless 2200/2915 Network Driver, 1.2.2kmprq ipw2200: Copyright(c) 2003-2006 Intel Corporation ipw2200 0000:02:03.0: PCI INT A -> GSI 17 (level, low) -> IRQ 17 ipw2200: Detected Intel PRO/Wireless 2200BG Network Connection ipw2200: Detected geography ZZM (11 802.11bg channels, 0 802.11a channels) /dev/vmnet: open called by PID 1203 (vmnet-bridge) /dev/vmnet: hub 0 does not exist, allocating memory. /dev/vmnet: port on hub 0 successfully opened bridge-eth1: device is wireless, enabling SMAC bridge-eth1: up bridge-eth1: attached bridge-eth1: disabling the bridge on dev down bridge-eth1: down bridge-eth1: detached /dev/vmnet: open called by PID 1203 (vmnet-bridge) ... Doing "sudo rmmod ipw2200; sudo modprobe ipw2200" fixes it for a random number of hours. It seems a little unlikely that this is really a firmware error, since it is only a problem with the new kernel 2.6.40-4.fc15.i686. If I restart NetworkManager without removing and reloading ipw2200 I get one or two kworker processes taking close to 100% cpu and a reboot is necessary. This too is only with kernel 2.6.40-4.fc15.i686 and not earlier kernels.
This is fixed for me with kernel 2.6.40.4-5.fc15.i686
Sorry, take it back. Still a problem with kernel 2.6.40.4-5.fc15.i686
I'm seeing this here as well... ;( 2.6.40.4-5.fc15.i686 ipw2200-firmware-3.1-5.fc15.noarch
(In reply to comment #7) > I'm seeing this here as well... ;( > > 2.6.40.4-5.fc15.i686 > ipw2200-firmware-3.1-5.fc15.noarch
This bug makes a bit annoying to use fedora on laptops with the widely used intel centrino. Any news?
The work-around is describe in comment 4. It sucks, but it is all we've got. There will be no firmware update -- Intel has abandoned this device. The firmware hasn't changed in a couple of years. Whatever is causing these restarts (or preventing recovery from them) is in the kernel itself. If someone would like to do a git bisect to pinpoint the source of the problem, we might have a chance to fix it. Any takers?
I'd like to help out. Unfortunately I haven't found a reliable way to reproduce it. Suggestions welcome.
I've been using the script below to try and collect data (and keep my wireless up). It puts markers in /var/log/messages and dmesg when the bug happens. I can't see anything in the results which looks useful, but maybe someone else can. #!/bin/sh reloadipw() { date >> /var/log/messages echo reloading ipw2200 >> /var/log/messages date >> /dev/kmsg echo reloading ipw2200 >> /dev/kmsg rmmod ipw2200 modprobe ipw2200 } while true do sleep 60 ping -c1 192.168.1.1 | grep "0 received" >/dev/null && reloadipw done # Change the ip address to suit your system
*** Bug 736887 has been marked as a duplicate of this bug. ***
Just ran into this on a Dell Inspiron 9300 laptop with a nightly build of Fedora 17 between the Alpha-and-Beta release with the following kernel: kernel-3.3.0-0.rc6.git0.2.fc17.i686 I've been running Fedora on this laptop for a couple of years and the ipw2200 has worked fine the whole time. I tried the fix listed in #4 and it seems to work... but having to manually do that after every boot just is no fun. I have gone to plugging in a USB wireless adapter that works flawlessly (Bus 001 Device 002: ID 083a:4505 Accton Technology Corp. SMCWUSB-G 802.11bg). This seems like some odd timing issue that I hope can get fixed.
Switched back to Fedora 16 fully updated with kernel-3.2.9-1.fc16.i686 on the same hardware, and the ipw2200 is working fine. This does NOT seem to be a hardware nor a firmware problem.
Just wanted to mention that I tried Fedora 17 Beta RC3 and it has the same issue on the previously mentioned Dell Inspiron 9300 laptop... the wireless is not detected and doesn't show up at all with NetworkManager. Sorry but I didn't notice what exact kernel build it was. Again switching back to Fedora 16 with the current kernel at time of writing (kernel-3.3.1-3.fc16.i686) and it works fine. Would it help if I submitted a new bug against Fedora 17? I'm guessing that since Fedora 15 is so close to EOL that this bug will be ignored.
Tested on a Dell Latitude D600... which is also an older (Pentium M) system... and it has the same problem. Removing the module and modprobing it works.
This problem seems to have gone away for me on Fedora 17. Anyone else still having this issue with Fedora 17?
Closing since no one has said they were still having issues on F17
This problem is back in FC17 with recent kernels (starting around 3.6.7 or 3.6.8). Can we reopen?