2 files changed, 50 insertions, 32 deletions

diff --git a/kernel/cpuset.c b/kernel/cpuset.c
index a09ac2b9a661..c9837b74ab96 100644
--- a/kernel/cpuset.c
+++ b/kernel/cpuset.c
@@ -2195,7 +2195,7 @@ void cpuset_cpus_allowed(struct task_struct *tsk, struct cpumask *pmask)
         mutex_unlock(&callback_mutex);
 }
 
-int cpuset_cpus_allowed_fallback(struct task_struct *tsk)
+void cpuset_cpus_allowed_fallback(struct task_struct *tsk)
 {
         const struct cpuset *cs;
         int cpu;
@@ -2219,22 +2219,10 @@ int cpuset_cpus_allowed_fallback(struct task_struct *tsk)
          * changes in tsk_cs()->cpus_allowed. Otherwise we can temporary
          * set any mask even if it is not right from task_cs() pov,
          * the pending set_cpus_allowed_ptr() will fix things.
+         *
+         * select_fallback_rq() will fix things ups and set cpu_possible_mask
+         * if required.
          */
-
-        cpu = cpumask_any_and(&tsk->cpus_allowed, cpu_active_mask);
-        if (cpu >= nr_cpu_ids) {
-                /*
-                 * Either tsk->cpus_allowed is wrong (see above) or it
-                 * is actually empty. The latter case is only possible
-                 * if we are racing with remove_tasks_in_empty_cpuset().
-                 * Like above we can temporary set any mask and rely on
-                 * set_cpus_allowed_ptr() as synchronization point.
-                 */
-                do_set_cpus_allowed(tsk, cpu_possible_mask);
-                cpu = cpumask_any(cpu_active_mask);
-        }
-
-        return cpu;
 }
 
 void cpuset_init_current_mems_allowed(void)
diff --git a/kernel/sched/core.c b/kernel/sched/core.c
index e3ccc13c4caa..9c1629c90b2d 100644
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -1263,29 +1263,59 @@ EXPORT_SYMBOL_GPL(kick_process);
  */
 static int select_fallback_rq(int cpu, struct task_struct *p)
 {
-        int dest_cpu;
         const struct cpumask *nodemask = cpumask_of_node(cpu_to_node(cpu));
+        enum { cpuset, possible, fail } state = cpuset;
+        int dest_cpu;
 
         /* Look for allowed, online CPU in same node. */
-        for_each_cpu_and(dest_cpu, nodemask, cpu_active_mask)
+        for_each_cpu_mask(dest_cpu, *nodemask) {
+                if (!cpu_online(dest_cpu))
+                        continue;
+                if (!cpu_active(dest_cpu))
+                        continue;
                 if (cpumask_test_cpu(dest_cpu, tsk_cpus_allowed(p)))
                         return dest_cpu;
+        }
 
-        /* Any allowed, online CPU? */
-        dest_cpu = cpumask_any_and(tsk_cpus_allowed(p), cpu_active_mask);
-        if (dest_cpu < nr_cpu_ids)
-                return dest_cpu;
+        for (;;) {
+                /* Any allowed, online CPU? */
+                for_each_cpu_mask(dest_cpu, *tsk_cpus_allowed(p)) {
+                        if (!cpu_online(dest_cpu))
+                                continue;
+                        if (!cpu_active(dest_cpu))
+                                continue;
+                        goto out;
+                }
 
-        /* No more Mr. Nice Guy. */
-        dest_cpu = cpuset_cpus_allowed_fallback(p);
-        /*
-         * Don't tell them about moving exiting tasks or
-         * kernel threads (both mm NULL), since they never
-         * leave kernel.
-         */
-        if (p->mm && printk_ratelimit()) {
-                printk_sched("process %d (%s) no longer affine to cpu%d\n",
-                                task_pid_nr(p), p->comm, cpu);
+                switch (state) {
+                case cpuset:
+                        /* No more Mr. Nice Guy. */
+                        cpuset_cpus_allowed_fallback(p);
+                        state = possible;
+                        break;
+
+                case possible:
+                        do_set_cpus_allowed(p, cpu_possible_mask);
+                        state = fail;
+                        break;
+
+                case fail:
+                        BUG();
+                        break;
+                }
+        }
+
+out:
+        if (state != cpuset) {
+                /*
+                 * Don't tell them about moving exiting tasks or
+                 * kernel threads (both mm NULL), since they never
+                 * leave kernel.
+                 */
+                if (p->mm && printk_ratelimit()) {
+                        printk_sched("process %d (%s) no longer affine to cpu%d\n",
+                                        task_pid_nr(p), p->comm, cpu);
+                }
         }
 
         return dest_cpu;