Re: [PATCH/RFC] b43: remove input device usage for rfkill

[Larry Finger <Larry.Finger@lwfinger.net>]

Johannes Berg wrote:
> On Tue, 2008-07-01 at 13:50 -0300, Henrique de Moraes Holschuh wrote:
>> On Tue, 01 Jul 2008, Johannes Berg wrote:
>>>> 1. It has an input pin that sometimes people connect buttons/switches
>>>> to.
>>>>
>>>> Does that input pin act as an input for a FLIP-FLOP (and that flip-flop
>>>> output is the hardware rfkill line), or is it the hardware rfkill line
>>>> itself?
>>> Neither. It's not hw kill, but it's not just edge detect either, it does
>>> have 'kill' and 'not kill' states.
>> What exactly the input pin does?  what exactly the hardware does when
>> the input pin state changes?
> 
> The hardware does nothing, it just sets a bit high or low depending on
> the input pin.

On my HP laptop using the RFKILL code currently in wireless-testing, 
sliding the "Wireless switch" results in b43_rfkill_poll being called. 
It then issues KEY_WLAN press and release events. This changes the 
state of the wireless light and the contents of a read-only bit in the 
status register of the device. I believe that if this bit is off, the 
radio is blocked by the hardware.

Larry