Re: [RFC] rfkill - Add support for input key to control wireless radio

[Ivo van Doorn <ivdoorn@gmail.com>]

Hi,

> > > >  2 - Hardware key that does not control the hardware radio and does not report anything to userspace
> > > 
> > > Kind of uninteresting button ;)
> > 
> > And this is the button that rfkill was originally designed for.
> > Laptops with integrated WiFi cards from Ralink have a hardware button that don't send anything to
> > userspace (unless the ACPI event is read) and does not directly control the radio itself.
> 
> My take: if there is a button on your keyboard or laptop labeled "Kill
> my radio now", it _NEEDS_ to be somehow communicated to userspace what
> happened when the user just pressed it a second ago.  Personally, I
> don't particularly care how that happens, and I don't particularly care
> what the driver does.  But if the driver, or the hardware, decides that
> the button press means turning off the transmitter on whatever device
> that button is for, a tool like NetworkManager needs to know this
> somehow.  Ideally, this would be a HAL event, and HAL would get it from
> somewhere.
>
> The current situation with NM is unacceptable, and I can't do anything
> about it because there is no standard interface for determining whether
> the wireless card was disabled/enabled via rfkill.  I simply refuse to
> code solutions to every vendor's rfkill mechanism (for ipw, reading
> iwpriv or sysfs, for example).  I don't care how HAL gets the event, but
> when HAL gets the event, it needs to broadcast it and NM needs to tear
> down the connection and release the device.
> 
> That means (a) an event gets sent to userspace in some way that HAL can
> read it, and (b) the event is clearly associated with specific piece[s]
> of hardware on your system.  If HAL can't easily figure out what device
> the event is for, then the event is also useless to both HAL and
> NetworkManager and whatever else might use it.

This would be possible with rfkill and the ideas from Dmitry.
The vendors that have a button that directly toggle the radio, should
create an input device themselves and just send the KEY_RFKILL event when toggled.

All other types should use rfkill for the toggling handling, that way HAL only needs to
listen to KEY_RFKILL coming from the input devices that are associated to the keys.

> Again, I don't care how that happens, but I like the fact that there's
> renewed interest in getting this fixed.

Ivo