x86/amd-iommu: Make amd_iommu_iova_to_phys aware of multiple page sizes

This patch extends the amd_iommu_iova_to_phys() function to
handle different page sizes correctly. It doesn't use
fetch_pte() anymore because we don't know (or care about)
the page_size used for mapping the given iova.

Signed-off-by: Joerg Roedel <joerg.roedel@amd.com>