On Mon, Jul 03, 2006 at 02:01:26PM +0200, Tristan Gingold wrote:
> Le Lundi 03 Juillet 2006 12:27, Isaku Yamahata a écrit :
> > On Mon, Jul 03, 2006 at 10:34:33AM +0200, Tristan Gingold wrote:
> > > Also, note that __cacheline_aligned_in_smp is defined as nothing on
> > > xen/ia64. You'd better either not to use it or define it.
> BTW, why not putting last_vcpu within struct arch_vcpu ?
> "vcpu->domain->arch.last_vcpu[smp_processor_id()] " looks a little bit
> strange...
I don't understand what you're sugessting.
The rid region is per-domain resource with the current implementation,
so it is natural to put last_vcpu[] into struct arch_domain.
If what you suggest is something like
struct arch_vcpu
struct last_vcpu last_vcpu[NR_CPUS]
and
vcpu->last_vcpu[smp_processor_id()],
then it doesn't work correctly.
Please consider the case that two vcpus of a same domain are
running on a same physical cpu.
--
yamahata
_______________________________________________
Xen-ia64-devel mailing list
Xen-ia64-devel@xxxxxxxxxxxxxxxxxxx
http://lists.xensource.com/xen-ia64-devel
|
Home