There really is no sense trying to continue if the kzalloc of sysfs_cpus[]
fails in ia64 topology_init.  The code calling into here doesn't check
errors very well, and one ends up with a nonobvious boot failure that
wastes peoples time debugging.
See for example the lkml thread at:
  http://lkml.org/lkml/2006/3/2/215
Since the system is totally dead when this kzalloc fails, not having yet
even booted, might as well announce one's death boldly and plainly.
Signed-off-by: Paul Jackson <pj@sgi.com>
Signed-off-by: Andrew Morton <akpm@osdl.org>
Signed-off-by: Tony Luck <tony.luck@intel.com>
 #endif
 
        sysfs_cpus = kzalloc(sizeof(struct ia64_cpu) * NR_CPUS, GFP_KERNEL);
-       if (!sysfs_cpus) {
-               err = -ENOMEM;
-               goto out;
-       }
+       if (!sysfs_cpus)
+               panic("kzalloc in topology_init failed - NR_CPUS too big?");
 
        for_each_present_cpu(i) {
                if((err = arch_register_cpu(i)))