Bug 41382 - ath5k: rfkill problems
Summary: ath5k: rfkill problems
Status: NEEDINFO
Alias: None
Product: Networking
Classification: Unclassified
Component: Wireless (show other bugs)
Hardware: All Linux
: P1 normal
Assignee: networking_wireless@kernel-bugs.osdl.org
URL:
Keywords:
Depends on:
Blocks:
 
Reported: 2011-08-19 06:27 UTC by Jan Franke
Modified: 2012-08-30 10:39 UTC (History)
5 users (show)

See Also:
Kernel Version: 3.5.1
Subsystem:
Regression: No
Bisected commit-id:


Attachments

Description Jan Franke 2011-08-19 06:27:34 UTC
Hello,

My distributor (kanotix) adviced me to report here...

Something seems to be wrong with the ath5k wireless driver in kernels higher than 2.6.32.21. My Atheros networkcard doesn`t work with these kernels. Until 2.6.32.21 it works fine.
Any advices?

"lspci -vnn" says:
02:00.0 Ethernet controller [0200]: Atheros Communications Inc. AR242x / AR542x Wireless Network Adapter (PCI-Express) [168c:001c] (rev 01)
Subsystem: Hewlett-Packard Company Device [103c:137b]
Physical Slot: 1
Flags: bus master, fast devsel, latency 0, IRQ 17
Memory at d2600000 (64-bit, non-prefetchable) [size=64K]
Capabilities: <access denied>
Kernel driver in use: ath5k 

"dmesg" under:

http://paste.debian.net/126593/

I will use this oppetunity to thank you all for your work.
Regards from Germany from a normal user, not an expert...
:)
Comment 1 John W. Linville 2011-08-22 13:29:45 UTC
That link seems to be wrong.  It says there is "some error".  Could you try adding the dmesg output as an attachment here?
Comment 2 Bob Copeland 2011-08-22 13:49:10 UTC
So v2.6.32.21 is the first bad kernel, or last good kernel and 2.6.32.22 is the first bad kernel?  Or is 2.6.33 the first bad kernel?

Not much happened in that range specifically for ath5k:

$ git log v2.6.32.20..v2.6.32.23 drivers/net/wireless/ath/ath5k | git shortlog
John W. Linville (1):
      ath5k: check return value of ieee80211_get_tx_rate

Maxim Levitsky (1):
      ath5k: disable ASPM L0s for all cards

I don't think either of those would be guilty.  v2.6.33 had a number of potential candidates.
Comment 3 Jan Franke 2011-08-23 10:07:11 UTC
Thanks for answering.
32.21 is the highest 32er I tried.
The first bad kernel I tried is called 2.6.33-02063307-generic. It` s an ubuntu-Kernel, modified (but not the ath5k) by Kano from the Kanotix-team.

http://kanotix.com/files/fix/mainline/

regards
___________________________________________________________
Schon gehört? WEB.DE hat einen genialen Phishing-Filter in die
Toolbar eingebaut! http://produkte.web.de/go/toolbar
Comment 4 Jan Franke 2011-08-23 10:16:33 UTC
Yes of course.
3dmesgLinux 3.1.0-0301rc2-generic i686[../email/download?si=kkwMPr3wyN9S5eQVNhx*025&draft=b6f942f57bd52fcf032fd6986e82f6db&file=0fc0093cf89785d04b3b70af5191b82c] is from the not working 3.1.0.Kernel
32dmesg is from the working 32.21 kernel.
thanks for answering.

-----Ursprüngliche Nachricht-----
Von: bugzilla-daemon@bugzilla.kernel.org
Gesendet: 22.08.2011 15:30:20
An: m.f.janfranke@web.de
Betreff: [Bug 41382] ath5k

>https://bugzilla.kernel.org/show_bug.cgi?id=41382
>
>
>John W. Linville <linville@tuxdriver.com> changed:
>
> What |Removed |Added
>----------------------------------------------------------------------------
> CC| |linville@tuxdriver.com,
> | |me@bobcopeland.com,
> | |mickflemm@gmail.com
>
>
>
>
>--- Comment #1 from John W. Linville <linville@tuxdriver.com> 2011-08-22
>13:29:45 ---
>That link seems to be wrong. It says there is "some error". Could you try
>adding the dmesg output as an attachment here?
>
>--
>Configure bugmail: https://bugzilla.kernel.org/userprefs.cgi?tab=email
>------- You are receiving this mail because: -------
>You reported the bug.


___________________________________________________________
Schon gehört? WEB.DE hat einen genialen Phishing-Filter in die
Toolbar eingebaut! http://produkte.web.de/go/toolbar
Comment 5 Bob Copeland 2011-08-23 10:57:15 UTC
(In reply to comment #4)
> Yes of course.
> 3dmesgLinux 3.1.0-0301rc2-generic
>
> i686[../email/download?si=kkwMPr3wyN9S5eQVNhx*025&draft=b6f942f57bd52fcf032fd6986e82f6db&file=0fc0093cf89785d04b3b70af5191b82c]
> is from the not working 3.1.0.Kernel
> 32dmesg is from the working 32.21 kernel.
> thanks for answering.


I am not sure how you tried to attach this (email?) but it didn't work -- you should add it on bugzilla.kernel.org.

I'm afraid to get further, we will need you to build and test some kernels on your hardware.  2.6.33 is quite old now, and the range of 2.6.32 to 3.1 is simply too large.  git-bisect would be the best option, or you can try reverting patches from 2.6.33 until it works acceptably, then try the same on later kernels.  Of course some issues that were introduced in the .33 time frame might be fixed in later kernels and other things regressed, so YMMV.

Here are some candidates for issues in 2.6.33:

commit e5e2647fd6ceef2cdc479954b84517535eb7febd
Author: Bob Copeland <me@bobcopeland.com>
Date:   Wed Oct 14 14:16:30 2009 -0400

    ath5k: use noise calibration from madwifi hal

commit f1cf2dbd0f798b71b1590e7aca6647f2caef1649
Author: Lukáš Turek <8an@praha12.net>
Date:   Thu Nov 19 23:02:02 2009 +0100

    ath5k: Fix I/Q calibration
Comment 6 John W. Linville 2012-01-13 19:14:59 UTC
Does this problem persist with 3.2 kernels?
Comment 7 Seb 2012-05-01 14:46:31 UTC
Yes and with 3.3.4 too.
When I boot (with WIFI switch ON), WIFI doesn't work, "rfkill list" gives :
    Soft blocked: no
    Hard blocked: yes

I'm forced to do (after each boot or suspend) :

rmmod -f ath5k
rfkill unblock all
modprobe ath5k

and then it works.
Comment 8 Seb 2012-07-08 14:35:40 UTC
Still exists with 3.4.4 kernel.
Comment 9 Seb 2012-08-15 11:24:02 UTC
Still exists with 3.5.1 kernel.

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