FS#28784 - [linux] 3.2.9-1-ARCH kernel: wlan0 link *never* becomes available

Attached to Project: Arch Linux
Opened by Stenten (stenten) - Monday, 05 March 2012, 07:38 GMT
Last edited by Thomas Bächler (brain0) - Thursday, 15 March 2012, 16:58 GMT
Task Type Bug Report
Category Packages: Testing
Status Closed
Assigned To Tobias Powalowski (tpowa)
Thomas Bächler (brain0)
Architecture i686
Severity Low
Priority Normal
Reported Version
Due in Version Undecided
Due Date Undecided
Percent Complete 100%
Votes 0
Private No

Details

Description:

I can connect to my WPA-EAP wireless network through netcfg using 3.2.8-1-ARCH kernel but not the 3.2.9-1-ARCH kernel.

Under 3.2.9's dmesg I get this every time I try to connect with netcfg:

Mar 5 01:12:37 localhost kernel: [ 351.016898] b43-phy0: Loading firmware version 666.2 (2011-02-23 01:15:07)
Mar 5 01:12:37 localhost kernel: [ 351.077596] ADDRCONF(NETDEV_UP): wlan0: link is not ready

But with the working 3.2.8:

Mar 5 02:04:41 localhost kernel: [ 11.033405] b43-phy0: Loading firmware version 666.2 (2011-02-23 01:15:07)
Mar 5 02:04:41 localhost kernel: [ 11.113574] ADDRCONF(NETDEV_UP): wlan0: link is not ready
Mar 5 02:04:43 localhost kernel: [ 13.266933] b43-phy0: Loading firmware version 666.2 (2011-02-23 01:15:07)
Mar 5 02:04:44 localhost kernel: [ 13.334264] ADDRCONF(NETDEV_UP): wlan0: link is not ready
Mar 5 02:04:47 localhost kernel: [ 16.836945] wlan0: direct probe to 00:24:6c:bf:81:41 (try 1/3)
Mar 5 02:04:47 localhost kernel: [ 17.036918] wlan0: direct probe to 00:24:6c:bf:81:41 (try 2/3)
Mar 5 02:04:47 localhost kernel: [ 17.043545] wlan0: direct probe responded
Mar 5 02:04:47 localhost kernel: [ 17.050236] wlan0: authenticate with 00:24:6c:bf:81:41 (try 1)
Mar 5 02:04:47 localhost kernel: [ 17.068189] wlan0: authenticated
Mar 5 02:04:47 localhost kernel: [ 17.068364] wlan0: associate with 00:24:6c:bf:81:41 (try 1)
Mar 5 02:04:47 localhost kernel: [ 17.078032] wlan0: RX AssocResp from 00:24:6c:bf:81:41 (capab=0x431 status=0 aid=1)
Mar 5 02:04:47 localhost kernel: [ 17.078036] wlan0: associated
Mar 5 02:04:47 localhost kernel: [ 17.090417] ADDRCONF(NETDEV_CHANGE): wlan0: link becomes ready
Mar 5 02:04:49 localhost kernel: [ 18.765712] Intel AES-NI instructions are not detected.
Mar 5 02:04:58 localhost kernel: [ 27.870227] wlan0: no IPv6 routers present



Additional info:
* package version(s)
netcfg 2.6.8-1
wpa_supplicant 0.7.3-5

* config and/or log files etc.
ATTACHED:
kernel.log.txt: most recent 2 kernel.logs. The first boot is the 3.2.9 kernel with the bug, and the second, most recent boot, is the 3.2.8 kernel which I'm connected to currently.
lspci.txt: output of 'lspci -vvnn'.
netcfg_debug.txt: netcfg connect command under 3.2.9 with NETCFG_DEBUG="yes".
wpa_supplicant.conf.txt: my /etc/wpa_supplicant.conf.
secure1x.txt: my netcfg profile from /etc/network.d/secure1x (with censored info).



Steps to reproduce:
'sudo pacman -U linux-3.2.9-1-i686.pkg.tar.xz'
Reboot.
'sudo netcfg secure1x'.

I can switch kernels back and forth from 3.2.8-1-ARCH to 3.2.9-1-ARCH and reproduce it every time: always works with 3.2.8, never with 3.2.9.
   kernel.log.txt (169.9 KiB)
   lspci.txt (10.3 KiB)
   netcfg_debug.txt (0.9 KiB)
   wpa_supplicant.conf.txt (0.1 KiB)
   secure1x.txt (0.4 KiB)
This task depends upon

Closed by  Thomas Bächler (brain0)
Thursday, 15 March 2012, 16:58 GMT
Reason for closing:  Fixed
Additional comments about closing:  ... and it will forever remain a mystery what happened here.
Comment by mangust (mangust) - Tuesday, 06 March 2012, 08:53 GMT
Hi!

You have this problem because of Broadcom module. I have similar problem in my logs (the detailed information a bit different). I have Asus 1215b, which has Broadcom BCM4313 (rev 01). Moreover, if you try to suspend/hibernate, you may have problems (system hangs, kernel panic etc.), at least I do.

The same problem has got even Linus Torvalds (on Mac Air). See here: http://comments.gmane.org/gmane.linux.network/217370

According to the thread, he fixed it on his laptop. And the patch is merged and we use it in 3.2.8.


So far, I have no idea how to solve it.

mangust
Comment by Phillip Wood (phil) - Friday, 09 March 2012, 16:46 GMT
Same problem here with a BCM4306 as well, has anyone reported this upstream? the macbook uses a different chip and linus had problems with the brcmsmac driver, which looks like this upstream bug https://bugzilla.kernel.org/show_bug.cgi?id=42591
Comment by Phillip Wood (phil) - Saturday, 10 March 2012, 14:14 GMT
Reported upstream - https://bugzilla.kernel.org/show_bug.cgi?id=42896
Comment by Phillip Wood (phil) - Sunday, 11 March 2012, 18:50 GMT
@stenten, found your ubuntu bug report this afternoon on google. It's great that they're working on it, looks like you should have it pinned down soon. I had started a bisection myself, perhaps in future you could post a link if you report a bug elsewhere (especially if it's being actively worked on) to avoid such duplication of effort. For others it is available at https://bugs.launchpad.net/ubuntu/+source/linux/+bug/947751
Comment by Thomas Bächler (brain0) - Monday, 12 March 2012, 14:05 GMT
This must be a problem with brcmsmac, I cannot reproduce it on my ath9k.

Unfortunately, I cannot find any Linux commits for brcmsmac between 3.2.8 and 3.2.9 and there were no configuration changes on our end. I am clueless.
Comment by Phillip Wood (phil) - Thursday, 15 March 2012, 16:51 GMT
Works fine with 3.2.11 from testing, what ever it was has been fixed.

Loading...