cpuset: fix wrong calculation of relax domain level

When multiple cpusets are overlapping in their 'cpus' and hence they
form a single sched domain, the largest sched_relax_domain_level among
those should be used. But when top_cpuset's sched_load_balance is
set, its sched_relax_domain_level is used regardless other sub-cpusets'.

This patch fixes it by walking the cpuset hierarchy to find the largest
sched_relax_domain_level.

Signed-off-by: Lai Jiangshan <laijs@cn.fujitsu.com>
Signed-off-by: Li Zefan <lizf@cn.fujitsu.com>
Cc: Paul Menage <menage@google.com>
Cc: Cedric Le Goater <clg@fr.ibm.com>
Cc: Balbir Singh <balbir@in.ibm.com>
Cc: KAMEZAWA Hiroyuki <kamezawa.hiroyu@jp.fujitsu.com>
Reviewed-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>
Signed-off-by: Linus Torvalds <torvalds@linux-foundation.org>

kernel/cpuset.c

@@ -616,7 +616,7 @@ void rebuild_sched_domains(void)
 		dattr = kmalloc(sizeof(struct sched_domain_attr), GFP_KERNEL);
 		if (dattr) {
 			*dattr = SD_ATTR_INIT;
-			update_domain_attr(dattr, &top_cpuset);
+			update_domain_attr_tree(dattr, &top_cpuset);
 		}
 		*doms = top_cpuset.cpus_allowed;
 		goto rebuild;