Re: [Xen-devel] [PATCH] Pin vcpu for VMX domain

[Keir Fraser <Keir.Fraser@xxxxxxxxxxxx>]

On 8 Feb 2006, at 10:25, Dong, Eddie wrote:

> > You don't need to __vmpclear at the time the affinity is changed. You
> > can still do it from within arch_vmx_do_resume() -- it is valid to
> > call smp_call_function() there. That avoids having yet another HVM
> > function, and avoids calling yet another arch_* function from common
> > code.
> Good point. Do u mean we check v->arch.hvm_vmx.launch_cpu with
>  v->processor in arch_vmx_do_resume and send IPI if they are not equal?
> If yes, is it worth to exchage common code change with the frequent
> check
> at every resume?
It's not every return to the guest --- it's on every context switch to 
the guest, which will usually be a lot less frequent. It's worth doing 
the check there to have tidier interfaces.
 -- Keir


_______________________________________________
Xen-devel mailing list
Xen-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-devel