FS#21683 - [kernel26] 2.6.36 panic when iwlagn (IPW4965) and rt2x00pci (RT2500) are put into monitor mode

Attached to Project: Arch Linux
Opened by Borromini (Borromini) - Saturday, 13 November 2010, 08:11 GMT
Last edited by Ionut Biru (wonder) - Sunday, 19 June 2011, 18:15 GMT
Task Type Bug Report
Category Upstream Bugs
Status Closed
Assigned To Tobias Powalowski (tpowa)
Thomas Bächler (brain0)
Architecture i686
Severity Medium
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 9
Private No

Details

Description:
Trying to put my Intel IPW4965 wireless card into monitor mode results in a kernel panic with kernel26 2.6.36 from [Testing]. Putting the card into monitor mode works fine - bringing it up in that mode will cause a kernel panic though, X shuts down, I get dumped to a console (mouse cursor still there, but doesn't react to any input though). This happens both with the mac80211 stack/iwlagn driver part of kernel26 and with the compat-wireless mac80211 stack/iwlagn driver (I tested 2.6.36* patches). I have tested this various times - it breaks every single time.

I have had similar behaviour with 2.6.36 on my desktop which has a Ralink RT2500 PCI card; there I won't get a kernel panic, but the system will lock up - sometimes I can bring up the interface in monitor mode, but most of the time it will just lock up while trying to do so; if it does not lock up while bringing up the interface, it will do so in the next few seconds that follow. As with my laptop, rolling back to 2.6.35.8 works just fine.

Related topic on the BBS: https://bbs.archlinux.org/viewtopic.php?id=108139


Additional info:
* tested with stock kernel26 package from [Testing] (2.6.36-ARCH)
* tested with vanilla kernel package (based on the kernel26 PKGBUILD, without Arch patches, just upstream source code); I suspect it to be a configuration problem (arch-dev-public shows quite some additional stuff enabled in 2.6.36?)
* works fine with 2.6.35.8 (latest kernel26 package from [core])


Steps to reproduce:
- Install kernel26 from [Testing]
- Bring down the wireless interface (ifconfig wlan0 down)
- Put wireless interface into monitor mode (iwconfig wlan0 mode monitor)
- Bring up wireless interface again (ifconfig wlan0 up)
This task depends upon

Closed by  Ionut Biru (wonder)
Sunday, 19 June 2011, 18:15 GMT
Reason for closing:  Fixed
Additional comments about closing:  since 2.6.37
Comment by Marek Niklicek (Dogmeat) - Saturday, 13 November 2010, 14:29 GMT
It happens with Atheros AR928X Wireless Network Adapter (ath9k module) as well under Linux 2.6.36 (testing/kernel26 2.6.36-3).
Comment by Adrian Goll (goll) - Monday, 22 November 2010, 03:05 GMT
Can confirm the same bug with my intel 5100 wireless card and testing/kernel26 2.6.36-3
Comment by Tobias Powalowski (tpowa) - Wednesday, 24 November 2010, 10:34 GMT
Please try latest kernel from testing, thanks.
Comment by Adrian Goll (goll) - Wednesday, 24 November 2010, 12:03 GMT
Unfortunately, I get the same kernel panic trying to go into monitor mode, using testing/kernel26-2.6.36.1-3
Comment by spock (spock) - Friday, 26 November 2010, 15:41 GMT
same thing happening for me with rtl8187 on kernel26-2.6.36.1-3 on i686. also tried compat-wireless builds from aur but the panics just kept coming.
Comment by Tiago Teixeira (TigTex) - Saturday, 27 November 2010, 23:05 GMT
I can confirm that bug with rtl8187 on kernel26 2.6.36.1 (i686) like spock... once I enable monitor mode I have a kernel panic.
Comment by jwbirdsong (jwbirdsong) - Sunday, 28 November 2010, 00:51 GMT
As much as I dislike "Me too" posts:
Same with Broadcom 4311 and 4312 chips.

ONLY mentioned because broadcom not mentioned in here or BBS thread.
Comment by kolias (koliasforever) - Monday, 29 November 2010, 18:32 GMT
confirm with intel3945 and the iwl3945 driver. Kernel panic when i try to put in monitor mode
Comment by Fdo Lagos (zerial) - Thursday, 02 December 2010, 19:50 GMT
I confirm with iwlagn and rtl8187.
I can put in monitor mode but when I run "airodump", crash with kernel panic.

[root@balcebu ~]# dmesg |tail -1
usbcore: registered new interface driver rtl8187
[root@balcebu ~]# iwconfig wlan1 mode monitor
[root@balcebu ~]# iwconfig wlan1
wlan1 IEEE 802.11bg Mode:Monitor Tx-Power=0 dBm
Retry long limit:7 RTS thr:off Fragment thr:off
Power Management:off

[root@balcebu ~]# airodump-ng wlan1
- kernel panic -


Comment by Brian Johnson (alienvenom) - Sunday, 05 December 2010, 04:47 GMT
Can confirm as well with iwlagn/Intel 4965.

The person sitting next to me with Gentoo can do it just fine with 2.6.36. OP may be correct in that it may be a configuration issue. I'm getting his config and will report back with my findings.
Comment by Adrian Goll (goll) - Sunday, 05 December 2010, 10:50 GMT
There seems to be a patch that fixes the issue, and alienvenom, it is a kernel bug: https://bugzilla.kernel.org/show_bug.cgi?id=24102

Patch can be found here: http://www.spinics.net/lists/netdev/msg148705.html
Comment by jwbirdsong (jwbirdsong) - Monday, 06 December 2010, 03:22 GMT
YES!! Nice find.
Rebuilt using supplied patch, All is well in broadcom land as far as Moniter Mode goes.
I can at least enter moniter mode, Injection works and exit Moniter mode. With no segfault/crash/error.
Comment by spock (spock) - Tuesday, 14 December 2010, 09:03 GMT
the patch fixed the problem for me too.

however while testing my computer completely froze when I disconnected the usb wifi adapter. no kernel panic or anything, just a hard freeze. rebooted but didn't find anything interesting in the logs at all.
Comment by Thomas Bächler (brain0) - Tuesday, 14 December 2010, 09:33 GMT
spock, I recently read about problems with rt2x00 with disconnecting a USB wifi. There are new patches to fix this, but they are not in compat-wireless yet: http://marc.info/?l=linux-wireless&m=129224620129461&w=4
Comment by spock (spock) - Tuesday, 14 December 2010, 16:59 GMT
this freeze happened with an rtl8187 device. but perhaps those fixes are general and not just for the rt2x00 drivers?
Comment by spock (spock) - Wednesday, 15 December 2010, 05:15 GMT
thomas, this morning I tried again to replicate the crash yesterday but today it seems the problem is gone. I really have no idea how to explain the crash yesterday..
I did note in dmesg however there are 12 lines with "hpet1: lost N rtc interrupts".. N being anything from 1 to 24. but I dont know if this is relevant here.
all this testing was done with aircrack-ng on kernel26-2.6.36.2 with the above patch and driver rtl8187 from compat-wireless-patched from aur (http://aur.archlinux.org/packages.php?ID=41472).

anyway personally I'd say what happened yesterday was just some weirdness. but I'll share this here in case it happens to someone else with this patch.
Comment by Lyubomir Grigorov (bgalakazam) - Sunday, 19 December 2010, 02:46 GMT
I am having the exact same issue, with ipw4965. I am having trouble applying the mentioned patch. Can someone five directions?
Comment by ksha nargaroth (ksha) - Wednesday, 29 December 2010, 13:45 GMT
I'm having the same problem with mode monitor, details of my system:

- Atheros Communications Inc. AR9285
- 2.6.36 and 2.6.35 from pkgbuild aur and normal installation
Comment by Greg (dolby) - Monday, 02 May 2011, 12:35 GMT
Is this still a problem?
Comment by Jelle van der Waa (jelly) - Saturday, 18 June 2011, 08:38 GMT
again is this still a problem?
Comment by Holoduke (Holoduke) - Saturday, 18 June 2011, 12:15 GMT
I had this problem and it was gone with 2.6.37.

Loading...