Bug 42661 - compat-wireless disconnects randomly and locks system scheduler preventing process launch
Summary: compat-wireless disconnects randomly and locks system scheduler preventing pr...
Status: NEW
Alias: None
Product: Networking
Classification: Unclassified
Component: Wireless (show other bugs)
Hardware: All Linux
: P1 high
Assignee: networking_wireless@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2012-01-26 13:10 UTC by Orfeo Ciano
Modified: 2016-02-15 20:32 UTC (History)
9 users (show)

See Also:
Kernel Version: 3.3.0
Subsystem:
Regression: No
Bisected commit-id:


Attachments
dmesg output (23.51 KB, text/plain)
2012-01-26 13:10 UTC, Orfeo Ciano
Details
Dmesg after SysRQ -> W (66.59 KB, text/plain)
2012-01-26 15:44 UTC, Orfeo Ciano
Details
dmesg after SysRQ -> W and then plugging out the device (102.33 KB, text/plain)
2013-11-28 14:46 UTC, David Leppla-Weber
Details

Description Orfeo Ciano 2012-01-26 13:10:32 UTC
Created attachment 72204 [details]
dmesg output

I've got a TPLINK TL-WN821N dongle (0cf3:7015 Atheros Communications, Inc.).
I'm using Ubuntu 11.10, but tested also on the latest Mint (Ubuntu-based and debian-based) and this bug affects also their kernel.
The dongle connects correctly, authenticates with OPEN/WEP/WPA/WPA2 but after a random time (2-20minutes) stops working (especially if downloading a huge file with high bitrate): sometimes get switched off, sometimes stays on but doesn't transmit data. 
Firstly I tried to use the latest compat-wireless (3.3 rc1) but the problem got not solved. 
Then I tried updating linux firmwares htc_9271.fw, htc_7010.fw. Still same problem.
So I decided to leave wireless dongle plugged in also if the wireless connection was down and I noticed this strange behaviour through dmesg (see attachment).
I simply tried to run Chromium browser and nothing happens.
Comment 1 Orfeo Ciano 2012-01-26 13:22:12 UTC
Here I hot-unplugged and replugged USB:
...
[ 2961.066036] usb 1-1.4: USB disconnect, device number 4
[ 2961.189923] ath: Chip reset failed
[ 2961.189930] ath: Unable to reset hardware; reset status -22 (freq 2462 MHz)
[ 2961.200175] ath: Failed to wakeup in 500us
[ 2961.212336] ath: Failed to wakeup in 500us
[ 2961.312862] ath: Chip reset failed
[ 2961.312868] ath: Unable to reset hardware; reset status -22 (freq 2462 MHz)
[ 2961.379506] ath: Failed to wakeup in 500us
[ 2961.505305] ath: Failed to wakeup in 500us
[ 2962.450654] usb 1-1.4: ath9k_htc: USB layer deinitialized
[ 2965.349894] usb 1-1.4: new high speed USB device number 5 using ehci_hcd
...
Comment 2 Stanislaw Gruszka 2012-01-26 14:37:53 UTC
When this problem happens, can you do "echo w > /proc/sysrq-trigger" and attach dmesg ouptut here.
Comment 3 Orfeo Ciano 2012-01-26 15:44:14 UTC
Created attachment 72206 [details]
Dmesg after SysRQ -> W
Comment 4 Orfeo Ciano 2012-01-26 15:52:27 UTC
Don't know if it's important, but if I unplug the dongle everything works properly!
Comment 5 Orfeo Ciano 2012-01-27 13:16:32 UTC
Today Ubuntu released a kernel update 3.0.0-16
https://bugs.launchpad.net/ubuntu/+source/linux/+bug/907778
Looks like this series of patches fixes the scheduler freeze:
    mac80211: fix race condition caused by late addBA response
When dongle disconnects the system hangs for 10-15secs then resumes and works correctly! I'll test this update more and more to give you more feedbacks, but still the dongle continues to disconnect!
Comment 6 Orfeo Ciano 2012-01-27 13:28:27 UTC
(In reply to comment #5)
> Today Ubuntu released a kernel update 3.0.0-16
> https://bugs.launchpad.net/ubuntu/+source/linux/+bug/907778
> Looks like this series of patches fixes the scheduler freeze:
>     mac80211: fix race condition caused by late addBA response
> When dongle disconnects the system hangs for 10-15secs then resumes and works
> correctly! I'll test this update more and more to give you more feedbacks,
> but
> still the dongle continues to disconnect!

It is true, but after 3-4 minutes the system hangs and freezes the scheduler again! (sorry for spamming, i'll stop doing it)
Comment 7 shafi 2012-01-27 13:48:35 UTC
please make sure you are using
http://linuxwireless.org/download/htc_fw/1.3/
i am using this USB dongle for my network now + few distros downloading, let me see if i can get these issues,
Comment 8 Orfeo Ciano 2012-01-27 13:55:46 UTC
> please make sure you are using
> http://linuxwireless.org/download/htc_fw/1.3/
> i am using this USB dongle for my network now + few distros downloading, let
> me
> see if i can get these issues,

dmesg says:
[   16.584882] ath9k_htc 1-1.4:1.0: ath9k_htc: FW Version: 1.3

Maybe a WPA2 issue??? Let me know if you have got the same issues and what kind of network encryption you are using. Maybe WPA2/PSK is the problem!?
Comment 9 shafi 2012-01-27 14:05:46 UTC
(In reply to comment #8)
> > please make sure you are using
> > http://linuxwireless.org/download/htc_fw/1.3/
> > i am using this USB dongle for my network now + few distros downloading,
> let me
> > see if i can get these issues,
> 
> dmesg says:
> [   16.584882] ath9k_htc 1-1.4:1.0: ath9k_htc: FW Version: 1.3
> 
> Maybe a WPA2 issue??? Let me know if you have got the same issues and what
> kind
> of network encryption you are using. Maybe WPA2/PSK is the problem!?


sure i will try with WPA2 security
Comment 10 Orfeo Ciano 2012-01-30 11:57:48 UTC
(In reply to comment #9)
> (In reply to comment #8)
> > > please make sure you are using
> > > http://linuxwireless.org/download/htc_fw/1.3/
> > > i am using this USB dongle for my network now + few distros downloading,
> let me
> > > see if i can get these issues,
> > 
> > dmesg says:
> > [   16.584882] ath9k_htc 1-1.4:1.0: ath9k_htc: FW Version: 1.3
> > 
> > Maybe a WPA2 issue??? Let me know if you have got the same issues and what
> kind
> > of network encryption you are using. Maybe WPA2/PSK is the problem!?
> 
> 
> sure i will try with WPA2 security

News? This bug affects you too? Have you discovered if it is related to my own dongle (maybe the device is defected)?
Comment 11 shafi 2012-01-30 15:41:58 UTC
till now i was using 3.3.0-rc1-wl+ and using it for browsing etc. parallely ran a wget script to download distros. got some other warning but not related to this bug.
no most likely the problem is in the driver.
Comment 12 Orfeo Ciano 2012-02-07 11:19:19 UTC
Uhm just discovered that I haven't got 
TPLINK TL-WN821N
but
TPLINK TL-WN821N V3
I think that now all the things will change...
Comment 13 Oleksij Rempel 2013-10-18 08:24:56 UTC
Is this bug reproducible with latest kernel?
Comment 14 David Leppla-Weber 2013-11-28 14:46:11 UTC
Yes, it is. Since I switched from Win7 to Linux I'm having this with my TL-WN821N V3 (Bus 001 Device 005: ID 0cf3:7015 Atheros Communications, Inc. TP-Link TL-WN821N v3 802.11n [Atheros AR7010+AR9287]).
I'm now using:
Linux arch64 3.12.1-2-ARCH #1 SMP PREEMPT Mon Nov 25 12:19:01 CET 2013 x86_64 GNU/Linux
It feels like I'm not getting the issue as often as before a few months, but it's still there.
After doing iwconfig interface txpower 0 it feels like it at least takes longer until the wifi stops working but I'm not sure. Everytime the problem occurs, I have to plug out the device, plug it in and reconnect to wifi. Normally, the LED of the device stops blinking and lights without delays when the problem occurs, just sometimes the LED completely stops blinking.
I attached the dmesg output after running echo w > /proc/sysrq-trigger and then plugging out the device.
Comment 15 David Leppla-Weber 2013-11-28 14:46:20 UTC
Created attachment 116561 [details]
dmesg after SysRQ -> W and then plugging out the device

Note You need to log in before you can comment on or make changes to this bug.