Skip to contentSkip to conversation

Rfkill buttons : bi-stable, mono-stable, tactile

DISCUSSION Lukasz Sokol 3 participants 10 messages Dec 17, 2009 – Dec 18, 2009

lore
Share & tools
lore
#0 lore raw

Hello Developers,

I'm sorry for not cc:ing to relevant people and other lists, I just encountered a situation I wanted to share.

On Acer Aspire One, A150 Aw, sold to me with Linpus but then rooted up and installed Ubuntu Karmic (yes, 9.10) and updated the BIOS to 3309 (latest). This however isn't Ubuntu-specific (actually if I googled correctly, only one distribution got it right, the original Linpus Lite...)

The issue I am talking about is : the killswitch for the wireless adaptor.

The Internet is almost 'flooded' with reports of this not working on various distributions, that's why decided to go a level higher.

I have actually found that pushing this button repeatedly, produces 2 scancodes : showkey reports keycode 238 pressed and keycode 238 released; in -s mode reports 0xe0 0xf3 and 0xe0 0x73.

Also I think the firmware (BIOS) reports these only when the operation on the WIFI card has been done, i.e. when it successfully toggled it, because when pushing the button too often (too small gap between pushes)

As this netbook gives a completely unusable WMI interface (found out by Carlos Corbacho, the acer-wmi developer) there is no other way the rfkill infrastructure can be aware of the hard-lock status;

The killswitch itself is a mono-stable sliding button, not a bi-stable slider; so I think this might also be related to some people's problems with (de)activating wifi hardware on laptops that have just a Fn+wifi toggle; This is why I think this might be a more generic issue, and why I decided to post here.

Apologies, if it really really doesn't belong here.

Lukasz

#1 lore raw

Lukasz Sokol <el.es.cr <at> gmail.com> writes:

Sorry, didn't finish a sentence :

Quoted 5 lines

Also I think the firmware (BIOS) reports these only when the operation on the WIFI card has been done, i.e. when it successfully toggled it, because when pushing the button too often (too small gap between pushes)

... then it only reports one scancode (0xe0 0xf3)

(but then that might be because of some polling interval issues and/or the fact I was doing it from a virtual console under X).

Quoted 4 lines

Lukasz

Lukasz

#2 lore raw
In reply to Lukasz Sokol
Quoted 9 lines · Lukasz Sokol

On Acer Aspire One, A150 Aw, sold to me with Linpus but then rooted up and installed Ubuntu Karmic (yes, 9.10) and updated the BIOS to 3309 (latest). This however isn't Ubuntu-specific (actually if I googled correctly, only one distribution got it right, the original Linpus Lite...)

The issue I am talking about is : the killswitch for the wireless adaptor.

The Internet is almost 'flooded' with reports of this not working on various distributions, that's why decided to go a level higher.

What do you mean by "not working"? You don't really explain what the issue you're seeing is.

Signature · 1 line
Matthew Garrett | [email protected]
#3 lore raw
Quoted 6 lines · Matthew Garrett <mjg59 <at> srcf.ucam.org> writes:
Quoted 2 lines

The Internet is almost 'flooded' with reports of this not working on various distributions, that's why decided to go a level higher.

What do you mean by "not working"? You don't really explain what the issue you're seeing is.

Hello Matthew,

I define "not working" in this case by "it switches hardware on/off without the software knowing".

i.e. the rfkill does not know, and the 'problem' is only noticed when the wifi disassociates (but Network Manager tries on to reconnect). On the other hand, invoking /sbin/rfkill [un]block wifi blocks/unblocks wifi comfortably.

#4 lore raw

(Please keep me Cc:ed - linux-kernel is high enough traffic that it's easy to lose track otherwise)

Quoted 6 lines · Lukasz Sokol

I define "not working" in this case by "it switches hardware on/off without the software knowing".

i.e. the rfkill does not know, and the 'problem' is only noticed when the wifi disassociates (but Network Manager tries on to reconnect). On the other hand, invoking /sbin/rfkill [un]block wifi blocks/unblocks wifi comfortably.

What does ls -l /sys/class/rfkill look like? What driver are you using for your wireless card?

Signature · 1 line
Matthew Garrett | [email protected]
#5 lore raw
Quoted 1 line · Matthew Garrett

(Please keep me Cc:ed - linux-kernel is high enough traffic that it's

Ok, sorry, I was posting through gmane's web interface, I like it quite, but it is stupid in the sense that it does not keep cc:s...

Quoted 7 lines
Quoted 3 lines

i.e. the rfkill does not know, and the 'problem' is only noticed when the wifi disassociates (but Network Manager tries on to reconnect). On the other hand, invoking /sbin/rfkill [un]block wifi blocks/unblocks wifi comfortably.

What does ls -l /sys/class/rfkill look like? What driver are you using for your wireless card?

lrwxrwxrwx 1 root root 0 2009-12-17 20:30 rfkill0 -> ../../devices/pci0000:00/0000:00:1c.2/0000:03:00.0/ieee80211/phy0/rfkill0

lukasz@workhorse:~$ lsmod |grep 802
mac80211              181140  1 ath5k
cfg80211               93052  3 ath5k,mac80211,ath

rfkill is built in (i.e. not a module) in this kernel, I heard somewhere.

Lukasz

#6 lore raw
Quoted 1 line · Lukasz Sokol

lrwxrwxrwx 1 root root 0 2009-12-17 20:30 rfkill0 -> ../../devices/pci0000:00/0000:00:1c.2/0000:03:00.0/ieee80211/phy0/rfkill0

Ok. You have no platform rfkill interface. ath5k needs to notice that it's been killed and update its rfkill state accordingly.

Signature · 1 line
Matthew Garrett | [email protected]
#7 lore raw
Quoted 7 lines · Matthew Garrett
Quoted 1 line · Lukasz Sokol

lrwxrwxrwx 1 root root 0 2009-12-17 20:30 rfkill0 -> ../../devices/pci0000:00/0000:00:1c.2/0000:03:00.0/ieee80211/phy0/rfkill0

Ok. You have no platform rfkill interface. ath5k needs to notice that it's been killed and update its rfkill state accordingly.

So it is up to the ath5k driver developers then ? :)

Lukasz

#8 lore raw
Quoted 9 lines · Lukasz Sokol
Quoted 7 lines · Matthew Garrett
Quoted 1 line · Lukasz Sokol

lrwxrwxrwx 1 root root 0 2009-12-17 20:30 rfkill0 -> ../../devices/pci0000:00/0000:00:1c.2/0000:03:00.0/ieee80211/phy0/rfkill0

Ok. You have no platform rfkill interface. ath5k needs to notice that it's been killed and update its rfkill state accordingly.

So it is up to the ath5k driver developers then ? :)

Yup.

Signature · 1 line
Matthew Garrett | [email protected]
#9 lore raw
Quoted 11 lines · Matthew Garrett
Quoted 8 lines · Lukasz Sokol
Quoted 6 lines · Matthew Garrett
Quoted 1 line · Lukasz Sokol

lrwxrwxrwx 1 root root 0 2009-12-17 20:30 rfkill0 -> ../../devices/pci0000:00/0000:00:1c.2/0000:03:00.0/ieee80211/phy0/rfkill0

Ok. You have no platform rfkill interface. ath5k needs to notice that it's been killed and update its rfkill state accordingly.

So it is up to the ath5k driver developers then ? :)

Yup.

Thanks Matthew, will see if somebody shows up :)

Lukasz