Re: [PATCH 2/7] KVM: Intel virtual mode extensions definitions

[Avi Kivity <avi@qumranet.com>]

Steven Rostedt wrote:
> On Thu, 2006-10-19 at 15:48 +0200, Avi Kivity wrote:
>   
>> Add some constants for the various bits defined by Intel's VT extensions.
>>
>> Most of this file was lifted from the Xen hypervisor.
>>
>> Signed-off-by: Yaniv Kamay <yaniv@qumranet.com>
>> Signed-off-by: Avi Kivity <avi@qumranet.com>
>>
>> Index: linux-2.6/drivers/kvm/vmx.h
>> ===================================================================
>> --- /dev/null
>> +++ linux-2.6/drivers/kvm/vmx.h
>> @@ -0,0 +1,287 @@
>>     
>
> This entire file is also very specific to an architecture. Couldn't it
> be put somewhere in arch/x86_64 and not in drivers?
>
> I know that this is all currently focused on Intel and AMD
> virtualization platforms, but could you split out the x86_64 specific
> stuff and make the rest more generic. Perhaps in the future this will
> make it easier for other platforms to use this code as well.
>   

We're already doing some splitting since Intel and AMD have incompatible 
extensions for doing this.  The result however will still be x86 (-64 
and i386) specific.

> It's hard to do a generic approach when developing it new, but if you
> don't think about that now, it will be magnitudes larger in difficulty
> to make generic when this is all done.
>   

I don't know enough about ppc and ia64 virtualization for that.  Perhaps 
someone would like to comment.

-- 
error compiling committee.c: too many arguments to function