Why is the Dell killswitch interface turning *off* all hardware BT switches?

[Michael_E_Brown at Dell.com]

Bastien,
  I'm curious as to what "stupid things" that dellWirelessCtl might be doing. The statement by itself isnt exceedingly enlightening or helpful.

  The BIOS interface either supports the full runtime interface or none of it, as far as I know. Ie. if dellWirelessCtl says it supports any of the runtime interface, it supports the full capabilities.

I wrote dellWirelessCtl a long time ago and I remember running into one limitation with setting switch settings, but I dont remember the exact details (a BIOS limitation). Something like changing the switch settings always turned BT on or off. It was a BIOS bug or limitation or whatever you want to call it. But the solution would most likely be to adjust what HAL is doing as suggested by Mario. I've been messing with it on my system for a little bit to try to remember what the problem was, but I cant get the exact sequence that hits what I remember hitting.
--
Michale


-----Original Message-----
From: Mario Limonciello [mailto:mario_limonciello at dell.com]
Sent: Thu 10/23/2008 2:14 PM
To: Bastien Nocera
Cc: hal at lists.freedesktop.org; Brown, Michael E
Subject: Re: Why is the Dell killswitch interface turning *off* all hardware BT switches?
 
Hi Bastien:

This information (about the ability to switch via software interface)
isn't necessarily reported by the BIOS.  I'm adding on Michael Brown. 
Michael, libsmbios can't reliably report this information can it?

Why was the hal killswitch interface mucking with a hardware switch's
settings in the first place?

Regards

Bastien Nocera wrote:
> On Thu, 2008-10-23 at 13:03 -0500, Mario Limonciello wrote:
>   
>> Hi:
>>
>> I've been trying to track down why on some machines the hardware BT
>> switch suddenly stops working periodically and then figured out what's
>> actually happening.  HAL is turning it off every time.  Well
>> unfortunately, on a lot of Dell platforms, the software switch simply
>> doesn't work.  This is a BIOS limitation on those platforms.  What
>> happens then is somehow this kill switch interface gets triggered,
>> turns off the hardware switch, and suddenly the user has no way to
>> turn bluetooth back on.  I feel the better solution is to only turn
>> off via the software switch.  If the platform ends up supporting it,
>> it will work, if not, nothing happens.
>>
>> I'm attaching a patch to adjust this behavior, but I'm open to
>> constructive criticism.
>>     
>
> How about fixing libsmbios' dellWirelessCtl to not do stupid things?
>
> Cheers
>
>   

-- 
Mario Limonciello
*Dell | Linux Engineering*
mario_limonciello at dell.com