+            mhp_altmap.base_pfn = PHYS_PFN(start);
+            mhp_altmap.free = PHYS_PFN(size) - nr_vmemmap_pages;


That change does not belong into this patch.



I kept that change with ppc64 enablement because only ppc64 arch got check 
against
those values in the free path.

Let's make that accounting consistent in patch #4. I think it really belongs in there, especially once we clarify the "free vs. reserved" handling.

--
Cheers,

David / dhildenb

Reply via email to