[PATCH 0/7] [RFC] Sizing zones and holes in an architecture independent manner V2

[Mel Gorman]

On (13/04/06 10:47), Luck, Tony didst pronounce:
> > Double counted a hole here, then went downhill. Does the following fix
> > it?
> 
> Yes, that boots.  What's more the counts of pages in DMA/Normal
> zone match the kernel w/o your patches too.  So for tiger_defconfig
> you've now exactly matched the old behaivour.
> 

Very very cool. Thanks for persisting.

> I'll try to test generic and sparse kernels later, but I have to
> look at another issue now.
> 

When you get around to it later, there is one case you may hit that Bob
Picco encountered and fixed for me. It's where a "new" range is registered
that is inside an existing area; e.g.

add_active_range:    0->10000
add_active_range: 9800->10000

It ends up merging incorrectly and you end up with one region from
9800-10000. The fix is below. 

diff -rup -X /usr/src/patchset-0.5/bin//dontdiff linux-2.6.17-rc1-zonesizing-v6/mm/mem_init.c linux-2.6.17-rc1-107-debug/mm/mem_init.c
--- linux-2.6.17-rc1-zonesizing-v6/mm/mem_init.c	2006-04-13 10:30:50.000000000 +0100
+++ linux-2.6.17-rc1-107-debug/mm/mem_init.c	2006-04-13 18:39:24.000000000 +0100
@@ -922,6 +926,13 @@ void __init add_active_range(unsigned in
 		if (early_node_map[i].nid != nid)
 			continue;
 
+		/* Skip if an existing region covers this new one */
+		if (start_pfn >= early_node_map[i].start_pfn &&
+				end_pfn <= early_node_map[i].end_pfn) {
+			printk("Existing\n");
+			return;
+		}
+
 		/* Merge forward if suitable */
 		if (start_pfn <= early_node_map[i].end_pfn &&
 				end_pfn > early_node_map[i].end_pfn) {