Commit ce164428 authored Oct 03, 2006 by Christoph Lameter Committed by Linus Torvalds Oct 03, 2006

[PATCH] scheduler: NUMA aware placement of sched_group_allnodes



When the per cpu sched domains are build then they also need to be placed
on the node where the cpu resides otherwise we will have frequent off node
accesses which will slow down the system.

Signed-off-by: Christoph Lameter <clameter@sgi.com>
Acked-by: Ingo Molnar <mingo@elte.hu>
Acked-by: Nick Piggin <nickpiggin@yahoo.com.au>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Linus Torvalds <torvalds@osdl.org>

parent 0feaece9

kernel/sched.c

+4 −3

Original line number	Diff line number	Diff line
		@@ -6349,9 +6349,10 @@ static int build_sched_domains(const cpumask_t *cpu_map)
		> SD_NODES_PER_DOMAIN*cpus_weight(nodemask)) {
		if (!sched_group_allnodes) {
		sched_group_allnodes
		= kmalloc(sizeof(struct sched_group)
		= kmalloc_node(sizeof(struct sched_group)
		* MAX_NUMNODES,
		GFP_KERNEL);
		GFP_KERNEL,
		cpu_to_node(i));
		if (!sched_group_allnodes) {
		printk(KERN_WARNING
		"Can not alloc allnodes sched group\n");