On Wed, 2012-03-07 at 10:05 -0800, Yinghai Lu wrote:

quoted

-
-       max_low_pfn_mapped = init_memory_mapping(0, end_pfn << PAGE_SHIFT);
-       max_pfn_mapped = max_low_pfn_mapped;
+       /* max_low_pfn_mapped is updated here */
+       max_pfn_mapped = init_memory_mapping();

 #ifdef CONFIG_X86_64
       if (max_pfn > max_low_pfn) {
-               max_pfn_mapped = init_memory_mapping(1UL<<32,
-                                                    max_pfn<<PAGE_SHIFT);
               /* can we preseve max_low_pfn ?*/
               max_low_pfn = max_pfn;
       }

you may need to move those three lines before
max_pfn_mapped = init_memory_mapping()

otherwise for x86_64, memory from [4G, TOMH) will not be directly mapped.

I'm afraid I don't understand what you mean. The changes in my patch
mean that init_memory_mapping() doesn't work the way it previously did.
It will map all the regions in the e820 table and presumably the top of
memory is contained within one of those regions.

Could you clarify what you think the problem is? Unfortunately I don't
have a test machine with large amounts of RAM so it's entirely possible
I've made a mistake somewhere.