[IA64] fix bug in sn/ia64 for sparse CPU numbering

The kernel's use of the for_each_*cpu(i) macros has allowed for sparse CPU
numbering.  When I hacked the kernel to test sparse cpu_present_map[] and
cpu_possible_map[] cpumasks, I discovered one remaining spot, in
sn_hwperf_ioctl() during sn initialization, that needs to be fixed.
Signed-off-by: NJohn Hawkes <hawkes@sgi.com>
Signed-off-by: NDean Roe <roe@sgi.com>
Signed-off-by: NTony Luck <tony.luck@intel.com>

arch/ia64/sn/kernel/sn2/sn_hwperf.c

@@ -743,13 +743,14 @@ sn_hwperf_ioctl(struct inode *in, struct file *fp, u32 op, u64 arg)
 		if ((r = sn_hwperf_enum_objects(&nobj, &objs)) == 0) {
 			memset(p, 0, a.sz);
 			for (i = 0; i < nobj; i++) {
+				int cpuobj_index = 0;
 				if (!SN_HWPERF_IS_NODE(objs + i))
 					continue;
 				node = sn_hwperf_obj_to_cnode(objs + i);
 				for_each_online_cpu(j) {
 					if (node != cpu_to_node(j))
 						continue;
-					cpuobj = (struct sn_hwperf_object_info *) p + j;
+					cpuobj = (struct sn_hwperf_object_info *) p + cpuobj_index++;
 					slice = 'a' + cpuid_to_slice(j);
 					cdata = cpu_data(j);
 					cpuobj->id = j;