[Debian-eeepc-devel] Bug#485236: eeepc-acpi-scripts: Fn+F2 sometimes fails to switch wireless on

[Trevor]

Alan Jenkins wrote:
> On 8/7/09, Trevor <trevor.chart at gmail.com> wrote:
>>
>> Andreas Schreiner wrote:
>>> Quoting Trevor <trevor.chart at gmail.com>:
>>>
>>>> This is a constant bug for me on 900a. Fn+F2 will switch wireless off,
>>>> but not back on. Running the wireless scripts manually results in:
>>>> off:
>>>> trevor at Myri:/etc/acpi/actions$ sudo sh wireless.sh off
>>>> Detected eeepc-wlan as rfkill /sys/class/rfkill/rfkill0
>>>> Detected WLAN module ath5k on wlan0
>>>>
>>>> On:
>>>> trevor at Myri:/etc/acpi/actions$ sudo sh wireless.sh on
>>>> Detected eeepc-wlan as rfkill /sys/class/rfkill/rfkill0
>>>> Detected WLAN module ath5k on wlan0
>>>> wlan0: ERROR while getting interface flags: No such device
>>>> exec: 51: wireless.sh: not found
>>>>
>>>> the led comes back on, but I get no connectivity and scanning with wicd
>>>>
>>>> anything else I can provide?
>>>
>>> I also use the EeePC 900a but not with pure Debian but Sidux instead. I
>>> have not yet tried this with "real" Debian Kernels and this is why I
>>> didt not yet open a bug report.
>>>
>>> But the story is as follows! The EeePC 900A still needs pciehp_force=1
>>> when loading the pciehp kernel module - even with the most up-to-date
>>> kernels. WLAN toggle will not work without pciehp_force=1 but it will
>>> work with it!
>>>
>>> Here is what you can do to test this:
>>> Reboot, then open a console and become root:
>>>
>>> Unload the pciehp module:
>>> # modprobe -r pciehp
>>>
>>> Load it again with pciehp_force
>>> # modprobe pciehp pciehp_force=1
>>>
>>> You will now see that WLAN toggle works again without problems by
>>> pressing fn-F2
>>>
>>>
>>> If you want to fix this permanently until the next eeepc-acpi-scripts
>>> update you can edit the /etc/init.d/eeepc-acpi-scripts file accordingly.
>>> This is what I have done but I would only recommend it if you at least
>>> understand how the file works and what to change - otherwise might break
>>> something.
>>>
>>>
>>> I hope I will find some time soon to test this with original Debian
>>> Kernels (stable, testing and unstable). Then it'll be easy to patch
>>> eeepc-acpi-scripts accordingly to specially treat the 900A.
>>>
>>> Regards
>>> Andreas Schreiner
>>>
>>>
>> Hey! It works great! Thanks Andreas!
> 
> [CC asus acpi mailing list]
> 
> Sorry to butt in late.

I wouldn't call it butting in - getting this to work right would be 
great, and you've probably got some insights that haven't yet been voiced.

> 			  On 2.6.30, eeepc-laptop is intended to replace
> the force-loading of pciehp.  It would be good to know why it doesn't
> work on the 900A when it works on other machines.  Maybe we can fix
> it.

I don't have any idea why, but it's definitely not working here. I'm 
happy to help however I can

> Does dmesg show any errors during the initialisation of eeepc-laptop?
> 
> I would be interested in the full output of dmesg and acpidump.
> 
> Regards
> Alan

I don't see anything out-of-place in dmesg, but I've never seen the 
successful version, so my eyes are likely not the best to look at it. 
full output of dmesg is attached, how would I get you the acpidump?

Trevor
-------------- next part --------------
An embedded and charset-unspecified text was scrubbed...
Name: dmesg.txt
URL: <http://lists.alioth.debian.org/pipermail/debian-eeepc-devel/attachments/20090811/0645e1d8/attachment-0001.txt>