Re: [RFC PATCH 3/8]: PVH: memory manager and paging related changes

[Mukesh Rathor <mukesh.rathor@oracle.com>]

On Thu, 16 Aug 2012 15:10:03 +0100
Stefano Stabellini <stefano.stabellini@eu.citrix.com> wrote:

> >  void __init xen_init_mmu_ops(void)
> >  {
> > +       x86_init.paging.pagetable_setup_done =
> > xen_pagetable_setup_done; +
> > +       if (xen_pvh_domain()) {
> > +               pv_mmu_ops.flush_tlb_others = xen_flush_tlb_others;
> > +
> > +               /* set_pte* for PCI devices to map iomem. */
> > +               if (xen_initial_domain()) {
> > +                       pv_mmu_ops.set_pte = xen_dom0pvh_set_pte;
> > +                       pv_mmu_ops.set_pte_at =
> > xen_dom0pvh_set_pte_at;
> > +               }
> > +               return;
> > +       }
> 
> Considering that the implementation of xen_dom0pvh_set_pte is
> native_set_pte, can't we just leave it to the default that is
> native_set_pte?

We can. I had more code in the function that I removed when I mapped
the entire IO space up front. Then, I had lot of debug code there. 
I can change it to native now.