Re: acpi4asus

[Bill Davidsen <davidsen@tmr.com>]

Pavel Machek wrote:

>Hi!
>
>  
>
>>>I am far from qualified to comment on this, but from a 
>>>users point of view, is it possible to not have laptop 
>>>specific code in the kernel?
>>>I have had two Linux laptops and with both I had ACPI 
>>>issues.
>>>The vendors of both laptops (Toshiba Tecra S1 and now 
>>>an Asus W3V) don't seem to be following standards. With 
>>>both I seem to need to patch ACPI to get various 
>>>functions of the laptop to work.
>>>I would love to see laptop specific functionality 
>>>definitions exist outside the kernel.
>>>
>>>      
>>>
>>I don't think that forcing laptop users to have their 
>>own code outside the kernel is really the best approach 
>>for either the developers or the users. Most users will 
>>    
>>
>
>No, we don't want that. But we do not want ibm-acpi, toshiba-acpi,
>asus-acpi, etc, when they really only differ in string constants used.
>
>We want userland to tell kernel 'mail led is controlled by AML routine
>foo', instead of having gazillion *-acpi modules.
>
>
>  
>
I see no reason why an interface to that couldn't be included in the 
kernel, with just a small table for each hardware instead of a whole 
module. Kind of a white list with detail.

-- 
bill davidsen <davidsen@tmr.com>
  CTO TMR Associates, Inc
  Doing interesting things with small computers since 1979