Re: Switch APIC (+nmi, +oprofile) to driver model

[Maciej W. Rozycki]

On Fri, 14 Feb 2003, Mikael Pettersson wrote:

>  >  OK, but then the question is: are the following calls:
>  > 
>  > +	driver_register(&local_apic_driver);
>  > +	return sys_device_register(&device_local_apic);
>  > 
>  > for suspend/resume exclusively?
> 
> Yes.

 OK, then.

> We could register the device also in other cases (!PM, SMP)
> but the methods would then be nullified and we'd have a device
> node with a name but no operations. I could do that, I just
> think it's pointless.

 Agreed if the interface is not going to be extended further, i.e. the
intent is to cover PM-capable devices only.

 I'd prefer the discrete APIC support not to get broken accidentally as
such systems are rare thus testing is limited. 

-- 
+  Maciej W. Rozycki, Technical University of Gdansk, Poland   +
+--------------------------------------------------------------+
+        e-mail: [e-mail], PGP key available        +