On Sat, 2012-03-24 at 14:27 +0400, Anton Vorontsov wrote:

+void clear_tasks_mm_cpumask(int cpu)
+{
+       struct task_struct *p;
+
+       read_lock(&tasklist_lock);
+       for_each_process(p) {
+               struct task_struct *t;
+
+               t =3D find_lock_task_mm(p);
+               if (!t)
+                       continue;
+               cpumask_clear_cpu(cpu, mm_cpumask(t->mm));
+               task_unlock(t);
+       }
+       read_unlock(&tasklist_lock);
+}=20

Why bother with the tasklist_lock at all anymore, afaict you could use
rcu_read_lock() here. This all is called after the cpu is taken down and
marked offline, so its not like new tasks will ever get this cpu set in
their mm mask.