Arch Linux

Please read this before reporting a bug:
https://wiki.archlinux.org/title/Bug_reporting_guidelines

Do NOT report bugs when a package is just outdated, or it is in the AUR. Use the 'flag out of date' link on the package page, or the Mailing List.

REPEAT: Do NOT report bugs for outdated packages!
Tasklist

FS#29339 - [linux] kernel oops followed by system freeze when shutting down wireless interface

Attached to Project: Arch Linux
Opened by Marisa Kirisame (Sayachan) - Sunday, 08 April 2012, 15:54 GMT
Last edited by Tobias Powalowski (tpowa) - Monday, 23 April 2012, 09:57 GMT
Task Type Bug Report
Category Kernel
Status Closed
Assigned To Tobias Powalowski (tpowa)
Thomas Bächler (brain0)
Architecture All
Severity Critical
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 2
Private No

Details

Description:

After updating from 3.2.6 to 3.2.14 due to the addition of patches for module r8712u (I'm using an USB wireless adapter with RTL8188SU chipset), I have run into a big problem which causes the system to freeze completely upon shutting down the wireless interface. This can be very annoying since not only it can happen during normal use, but it also interrupts the system shutdown process, making it hang indefinitely.

Additional info:

All installed packages are up to date as of 8th April 2012, 17:50 (CEST).

Using netcfg + wpa_supplicant (network has WEP encryption).

I am unable to capture the output of the oops message, since moments after that, the screen gets flooded with similar messages related to the CPU hanging, plus keyboard and mouse become unresponsive.

I can only provide output from a virtual machine (which may not be accurate, sorry for the inconvenience).
http://i.imgur.com/x1FWN.png <- the initial message when shutting down the connection
http://i.imgur.com/G2vma.png <- this is all I see after a while

This is my profile configuration (key removed for obvious reasons)
---
CONNECTION='wireless'
DESCRIPTION='R-wlanBJ82'
INTERFACE='wlan0'
SECURITY='wep'
ESSID='R-wlanBJ82'
KEY='s:#############'
IP='static'
IFOPTS="192.168.0.20 netmask 255.255.255.0 broadcast 192.168.0.255"
GATEWAY=192.168.0.1
DNS1=92.243.8.139
DNS2=82.237.169.10
# Uncomment this if your ssid is hidden
#HIDDEN=yes
---

Steps to reproduce:

Upon having connected to a wireless network loading its profile (netcfg up <profile>), bring the interface down (netcfg down <profile>).
This task depends upon

Closed by  Tobias Powalowski (tpowa)
Monday, 23 April 2012, 09:57 GMT
Reason for closing:  Fixed
Additional comments about closing:  3.3.3-1
Comment by Jelle van der Waa (jelly) - Monday, 09 April 2012, 09:12 GMT
And with 3.3?
Comment by Marisa Kirisame (Sayachan) - Monday, 09 April 2012, 09:47 GMT
Updated to 3.3. Looks like the mirror I had been using all this time was outdated.

The error seems identical, however, this time only the tty where I have initiated the interface shutdown appears to have hanged, mouse and keyboard still work and I can switch to a different one.

Powering off is still impossible from software. It's still hanging.

I was able to catch the dmesg output this time.
   dmesg (77.9 KiB)
Comment by Atilla ÖNTAŞ (tarakbumba) - Monday, 09 April 2012, 21:53 GMT
I have same symtomps with 3.2.14 and 3.3.1 kernels. But i can not catch dmesg but kernel.log shows the problem i believe. See Apr 10 00:33:48 log in kern.log.
   kern.log (101.8 KiB)
Comment by Ben Ruijl (revelation60) - Saturday, 14 April 2012, 09:12 GMT
I have the same problem with 3.3.1. Attached is the log with the call trace.

edit: I've notified Larry Finger, one of the developers of the r8712u module.
   messages (12.4 KiB)
Comment by Larry Finger (lwfinger) - Saturday, 14 April 2012, 18:26 GMT
There are two patches (attached here) that likely affect this issue. Both of these are in kernel 3.4-rc2, and are scheduled to be backported to stable, but it seems that they are not in your source.
   patch_2080913e (2.2 KiB)
   patch_9f4bc8c (2.2 KiB)
Comment by Marisa Kirisame (Sayachan) - Sunday, 15 April 2012, 11:49 GMT
Thanks, it appears that the problem is completely gone now. (by the way, I only needed the first patch)

Edit: I tested it on most of my 64-bit machines, but cross-compiling the kernel for one of my 32-bit machines made it unbootable. That is unrelated to this issue, though.
Comment by Atilla ÖNTAŞ (tarakbumba) - Thursday, 19 April 2012, 09:19 GMT
patch_2080913e worked here. Arch devs, please consider using this patch in official linux package until upstream backported kernel arrives. Thanks Larry Finger for your dedicated and good work.
Comment by Ben Ruijl (revelation60) - Saturday, 21 April 2012, 09:23 GMT
Why wasn't that patch included in the Arch version of 3.3.2-1?
Comment by Larry Finger (lwfinger) - Saturday, 21 April 2012, 18:16 GMT
I forgot to mark it as being needed for the stable kernels. it should be in kernel 3.3.4.

Sorry for the oversight,

Larry

Loading...