Re: [PATCH v6 09/15] efi: use new page table APIs in copy_mapping

[Jan Beulich <jbeulich@xxxxxxxx>]

On 24.04.2020 16:09, Hongyan Xia wrote:
> --- a/xen/common/efi/boot.c
> +++ b/xen/common/efi/boot.c
> @@ -1454,16 +1454,20 @@ static __init void copy_mapping(unsigned long mfn, 
> unsigned long end,
>              continue;
>          if ( !(l4e_get_flags(l4e) & _PAGE_PRESENT) )
>          {
> -            l3dst = alloc_xen_pagetable();
> -            BUG_ON(!l3dst);
> +            mfn_t l3mfn = alloc_xen_pagetable_new();
> +
> +            BUG_ON(mfn_eq(l3mfn, INVALID_MFN));
> +            l3dst = map_domain_page(l3mfn);
>              clear_page(l3dst);
>              efi_l4_pgtable[l4_table_offset(mfn << PAGE_SHIFT)] =
> -                l4e_from_paddr(virt_to_maddr(l3dst), __PAGE_HYPERVISOR);
> +                l4e_from_mfn(l3mfn, __PAGE_HYPERVISOR);
>          }
>          else
> -            l3dst = l4e_to_l3e(l4e);
> -        l3src = l4e_to_l3e(idle_pg_table[l4_table_offset(va)]);
> +            l3dst = map_l3t_from_l4e(l4e);
> +        l3src = map_l3t_from_l4e(idle_pg_table[l4_table_offset(va)]);
>          l3dst[l3_table_offset(mfn << PAGE_SHIFT)] = 
> l3src[l3_table_offset(va)];
> +        unmap_domain_page(l3src);
> +        unmap_domain_page(l3dst);
>      }
>  }

This looks very inefficient - you establish and tear down two mappings
per loop iteration, when really you may end up copying all 512 slots
between the same pair of L3 tables.

Jan