Index: linux-2.6.19-rc2-mm1/Documentation/sysctl/vm.txt
===================================================================
--- linux-2.6.19-rc2-mm1.orig/Documentation/sysctl/vm.txt	2006-10-23 14:45:03.341155754 -0500
+++ linux-2.6.19-rc2-mm1/Documentation/sysctl/vm.txt	2006-10-23 15:22:01.482622585 -0500
@@ -198,6 +198,29 @@ and may not be fast.
 
 =============================================================
 
+min_interleave_ratio:
+
+This is available only on NUMA kernels.
+
+A percentage of the free pages in each zone.  If less than this
+percentage of pages are in use then interleave will attempt to
+leave this zone alone and allocate from other zones. This results
+in a balancing effect on the system if interleave and node local allocations
+are mixed throughout the system. Interleave pages will not cause zone
+reclaim and leave some memory on node to allow node local allocation to
+occur. Interleave allocations will allocate all over the system until global
+reclaim kicks in.
+
+The mininum does not apply to pages that are placed using interleave
+based on an address such as implemented for anonymous pages. It is
+effective for slab allocations, huge page allocations and page cache
+allocations. A set of nodes to be interleaved must at least
+contain 3 nodes in order for this mechanism to be activated.
+
+The default ratio is 10 percent.
+
+=============================================================
+
 panic_on_oom
 
 This enables or disables panic on out-of-memory feature.  If this is set to 1,
Index: linux-2.6.19-rc2-mm1/include/linux/mmzone.h
===================================================================
--- linux-2.6.19-rc2-mm1.orig/include/linux/mmzone.h	2006-10-23 14:45:03.348968865 -0500
+++ linux-2.6.19-rc2-mm1/include/linux/mmzone.h	2006-10-23 14:49:11.452433653 -0500
@@ -192,6 +192,12 @@ struct zone {
 	 */
 	unsigned long		min_unmapped_pages;
 	unsigned long		min_slab_pages;
+	/*
+	 * If a zone has less pages free then interleave will
+	 * attempt to bypass the node
+	 */
+	unsigned long 		min_interleave_pages;
+
 	struct per_cpu_pageset	*pageset[NR_CPUS];
 #else
 	struct per_cpu_pageset	pageset[NR_CPUS];
@@ -568,6 +574,8 @@ int sysctl_min_unmapped_ratio_sysctl_han
 			struct file *, void __user *, size_t *, loff_t *);
 int sysctl_min_slab_ratio_sysctl_handler(struct ctl_table *, int,
 			struct file *, void __user *, size_t *, loff_t *);
+int sysctl_min_interleave_ratio_sysctl_handler(struct ctl_table *, int,
+			struct file *, void __user *, size_t *, loff_t *);
 
 #include <linux/topology.h>
 /* Returns the number of the current Node. */
Index: linux-2.6.19-rc2-mm1/include/linux/swap.h
===================================================================
--- linux-2.6.19-rc2-mm1.orig/include/linux/swap.h	2006-10-23 14:45:03.355805338 -0500
+++ linux-2.6.19-rc2-mm1/include/linux/swap.h	2006-10-23 14:49:11.463176686 -0500
@@ -197,6 +197,7 @@ extern long vm_total_pages;
 extern int zone_reclaim_mode;
 extern int sysctl_min_unmapped_ratio;
 extern int sysctl_min_slab_ratio;
+extern int sysctl_min_interleave_ratio;
 extern int zone_reclaim(struct zone *, gfp_t, unsigned int);
 #else
 #define zone_reclaim_mode 0
Index: linux-2.6.19-rc2-mm1/include/linux/sysctl.h
===================================================================
--- linux-2.6.19-rc2-mm1.orig/include/linux/sysctl.h	2006-10-23 14:45:03.365571727 -0500
+++ linux-2.6.19-rc2-mm1/include/linux/sysctl.h	2006-10-23 14:49:11.473919720 -0500
@@ -199,6 +199,7 @@ enum
 	VM_SWAP_PREFETCH=36,	/* swap prefetch */
 	VM_READAHEAD_RATIO=37,	/* percent of read-ahead size to thrashing-threshold */
 	VM_READAHEAD_HIT_RATE=38, /* one accessed page legitimizes so many read-ahead pages */
+	VM_MIN_INTERLEAVE=39,	 /* Limit for interleave */
 };
 
 
Index: linux-2.6.19-rc2-mm1/kernel/sysctl.c
===================================================================
--- linux-2.6.19-rc2-mm1.orig/kernel/sysctl.c	2006-10-23 14:45:03.374361478 -0500
+++ linux-2.6.19-rc2-mm1/kernel/sysctl.c	2006-10-23 14:49:11.493452509 -0500
@@ -1022,6 +1022,17 @@ static ctl_table vm_table[] = {
 		.extra1		= &zero,
 		.extra2		= &one_hundred,
 	},
+	{
+		.ctl_name	= VM_MIN_INTERLEAVE,
+		.procname	= "min_interleave_ratio",
+		.data		= &sysctl_min_interleave_ratio,
+		.maxlen		= sizeof(sysctl_min_interleave_ratio),
+		.mode		= 0644,
+		.proc_handler	= &sysctl_min_interleave_ratio_sysctl_handler,
+		.strategy	= &sysctl_intvec,
+		.extra1		= &zero,
+		.extra2		= &one_hundred,
+	},
 #endif
 #ifdef CONFIG_X86_32
 	{
Index: linux-2.6.19-rc2-mm1/mm/mempolicy.c
===================================================================
--- linux-2.6.19-rc2-mm1.orig/mm/mempolicy.c	2006-10-23 14:45:03.382174589 -0500
+++ linux-2.6.19-rc2-mm1/mm/mempolicy.c	2006-10-23 15:32:03.971304687 -0500
@@ -1123,11 +1123,38 @@ static unsigned interleave_nodes(struct 
 {
 	unsigned nid, next;
 	struct task_struct *me = current;
+	struct zone *z;
 
 	nid = me->il_next;
 	next = next_node(nid, policy->v.nodes);
 	if (next >= MAX_NUMNODES)
 		next = first_node(policy->v.nodes);
+
+	/*
+	 * If we are overallocating a zone then restart interleave at some
+	 * other node. We should not loop here since this is a performance
+	 * critical path. Just get another random zone and allocate from it.
+	 */
+	z = &NODE_DATA(nid)->node_zones[policy_zone];
+	if (unlikely(z->free_pages < z->min_interleave_pages)) {
+		int w = nodes_weight(policy->v.nodes) - 1;
+
+		if (w > 0) {
+			/*
+			 * We want a random node in the allowed set while
+			 * skipping the overallocated node.
+			 * get_cycles() gives us a rapidly increasing
+			 * cycle counters that we can use to get a pseudo
+			 * random result without too much effort.
+			 */
+			int n = get_cycles() % w;
+
+			nid = next;
+			next = first_node(policy->v.nodes);
+			while (next == nid || n-- >= 0)
+				next = next_node(next, policy->v.nodes);
+		}
+	}
 	me->il_next = next;
 	return nid;
 }
Index: linux-2.6.19-rc2-mm1/mm/page_alloc.c
===================================================================
--- linux-2.6.19-rc2-mm1.orig/mm/page_alloc.c	2006-10-23 14:45:03.392917617 -0500
+++ linux-2.6.19-rc2-mm1/mm/page_alloc.c	2006-10-23 14:49:11.531541446 -0500
@@ -2053,6 +2053,9 @@ static void setup_pagelist_highmark(stru
 
 
 #ifdef CONFIG_NUMA
+
+int sysctl_min_interleave_ratio = 10;
+
 /*
  * Boot pageset table. One per cpu which is going to be used for all
  * zones and all nodes. The parameters will be set in such a way
@@ -2648,6 +2651,7 @@ static void __meminit free_area_init_cor
 		zone->min_unmapped_pages = (realsize*sysctl_min_unmapped_ratio)
 						/ 100;
 		zone->min_slab_pages = (realsize * sysctl_min_slab_ratio) / 100;
+		zone->min_interleave_pages = (realsize + sysctl_min_interleave_ratio) / 100;
 #endif
 		zone->name = zone_names[j];
 		spin_lock_init(&zone->lock);
@@ -3223,6 +3227,21 @@ int sysctl_min_slab_ratio_sysctl_handler
 				sysctl_min_slab_ratio) / 100;
 	return 0;
 }
+int sysctl_min_interleave_ratio_sysctl_handler(ctl_table *table, int write,
+	struct file *file, void __user *buffer, size_t *length, loff_t *ppos)
+{
+	struct zone *zone;
+	int rc;
+
+	rc = proc_dointvec_minmax(table, write, file, buffer, length, ppos);
+	if (rc)
+		return rc;
+
+	for_each_zone(zone)
+		zone->min_interleave_pages = (zone->present_pages *
+				sysctl_min_interleave_ratio) / 100;
+	return 0;
+}
 #endif
 
 /*