x86, numa: Assign CPUs to nodes in round-robin manner on fake NUMA

commit d9c2d5a "x86, numa: Use near(er)
online node instead of roundrobin for NUMA" changed NUMA initialization on
Intel to choose the nearest online node or first node.  Fake NUMA would be
better of with round-robin initialization, instead of the all CPUS on
first node.  Change the choice of first node, back to round-robin.

For testing NUMA kernel behaviour without cpusets and NUMA aware
applications, it would be better to have cpus in different nodes, rather
than all in a single node.  With cpusets migration of tasks scenarios
cannot not be tested.

I guess having it round-robin shouldn't affect the use cases for all cpus
on the first node.

The code comments in arch/x86/mm/numa_64.c:759 indicate that this used to
be the case, which was changed by commit d9c2d5a.  It changed from
roundrobin to nearer or first node.  And I couldn't find any reason for
this change in its changelog.

Signed-off-by: Nikanth Karthikesan <knikanth@suse.de>
Cc: David Rientjes <rientjes@google.com>
Signed-off-by: Andrew Morton <akpm@linux-foundation.org>

arch/x86/kernel/cpu/intel.c

@@ -284,9 +284,7 @@ static void __cpuinit srat_detect_node(struct cpuinfo_x86 *c)
 	/* Don't do the funky fallback heuristics the AMD version employs
 	   for now. */
 	node = apicid_to_node[apicid];
-	if (node == NUMA_NO_NODE)
-		node = first_node(node_online_map);
-	else if (!node_online(node)) {
+	if (node == NUMA_NO_NODE || !node_online(node)) {
 		/* reuse the value from init_cpu_to_node() */
 		node = cpu_to_node(cpu);
 	}