Avi Kivity wrote:
> Zhang, Xiantao wrote:

>> 
>> Hi Avi,
>>      Thanks, But seems it is very hard to implement it through a.o,
>> because all functions in this part should be relocated to another
>> separate address space in whole.  If it is implemented as a
>> module(ko), we can relocate it as a whole. Currenlty, we haven't
>> found an easy method to handle it, due to  IA64 virtualization
>> architecture's intrinsic implementation. Anyway, adding one arg in
>> this function shouldn't impact other architectures. What's your
>> opinion? :) 
>> Xiantao
>> 
> 
> I can't say much about .o vs .ko without further information. 
> Regarding passing the module parameter, you might have your module
> call some function that is only present in the ia64 implementation to
> register itself, instead of calling common code (or, you can have the
> ia64 code call some function in your vmm module to get information
> from it). 

Yes to make it register itself. But if we use current interfaces, I
think it maybe easy,and direct. Otherwise, we have to call kvm_init
first, and call register function. Seems we can integrate them into
kvm_init, and just transfer the module pointer to kvm_init_arch. In this
way, it will simplify the code logic. 
BTW, it doesn't impact other arches, because they can just implement it
as NULL directly. 
Xiantao

-------------------------------------------------------------------------
This SF.net email is sponsored by: Microsoft
Defy all challenges. Microsoft(R) Visual Studio 2005.
http://clk.atdmt.com/MRT/go/vse0120000070mrt/direct/01/
_______________________________________________
kvm-devel mailing list
kvm-devel@lists.sourceforge.net
https://lists.sourceforge.net/lists/listinfo/kvm-devel

Reply via email to