Subject: sched-migrate-disable.patch
From: Thomas Gleixner <tglx@linutronix.de>
Date: Thu, 16 Jun 2011 13:26:08 +0200

Signed-off-by: Thomas Gleixner <tglx@linutronix.de>
---
 include/linux/preempt.h |    8 ++++
 include/linux/sched.h   |   13 +++++--
 kernel/sched/core.c     |   88 +++++++++++++++++++++++++++++++++++++++++++++---
 lib/smp_processor_id.c  |    6 +--
 4 files changed, 104 insertions(+), 11 deletions(-)

--- a/include/linux/preempt.h
+++ b/include/linux/preempt.h
@@ -130,6 +130,14 @@ do { \
 
 #endif /* CONFIG_PREEMPT_COUNT */
 
+#ifdef CONFIG_SMP
+extern void migrate_disable(void);
+extern void migrate_enable(void);
+#else
+# define migrate_disable()		barrier()
+# define migrate_enable()		barrier()
+#endif
+
 #ifdef CONFIG_PREEMPT_RT_FULL
 # define preempt_disable_rt()		preempt_disable()
 # define preempt_enable_rt()		preempt_enable()
--- a/include/linux/sched.h
+++ b/include/linux/sched.h
@@ -1075,6 +1075,7 @@ struct task_struct {
 #endif
 
 	unsigned int policy;
+	int migrate_disable;
 	int nr_cpus_allowed;
 	cpumask_t cpus_allowed;
 
@@ -1445,9 +1446,6 @@ struct task_struct {
 #endif
 };
 
-/* Future-safe accessor for struct task_struct's cpus_allowed. */
-#define tsk_cpus_allowed(tsk) (&(tsk)->cpus_allowed)
-
 #ifdef CONFIG_NUMA_BALANCING
 extern void task_numa_fault(int node, int pages, bool migrated);
 extern void set_numabalancing_state(bool enabled);
@@ -2620,6 +2618,15 @@ static inline void set_task_cpu(struct t
 
 #endif /* CONFIG_SMP */
 
+/* Future-safe accessor for struct task_struct's cpus_allowed. */
+static inline const struct cpumask *tsk_cpus_allowed(struct task_struct *p)
+{
+	if (p->migrate_disable)
+		return cpumask_of(task_cpu(p));
+
+	return &p->cpus_allowed;
+}
+
 extern long sched_setaffinity(pid_t pid, const struct cpumask *new_mask);
 extern long sched_getaffinity(pid_t pid, struct cpumask *mask);
 
--- a/kernel/sched/core.c
+++ b/kernel/sched/core.c
@@ -4871,11 +4871,12 @@ void __cpuinit init_idle(struct task_str
 #ifdef CONFIG_SMP
 void do_set_cpus_allowed(struct task_struct *p, const struct cpumask *new_mask)
 {
-	if (p->sched_class && p->sched_class->set_cpus_allowed)
-		p->sched_class->set_cpus_allowed(p, new_mask);
-
+	if (!p->migrate_disable) {
+		if (p->sched_class && p->sched_class->set_cpus_allowed)
+			p->sched_class->set_cpus_allowed(p, new_mask);
+		p->nr_cpus_allowed = cpumask_weight(new_mask);
+	}
 	cpumask_copy(&p->cpus_allowed, new_mask);
-	p->nr_cpus_allowed = cpumask_weight(new_mask);
 }
 
 /*
@@ -4921,7 +4922,7 @@ int set_cpus_allowed_ptr(struct task_str
 	do_set_cpus_allowed(p, new_mask);
 
 	/* Can the task run on the task's current CPU? If so, we're done */
-	if (cpumask_test_cpu(task_cpu(p), new_mask))
+	if (cpumask_test_cpu(task_cpu(p), new_mask) || p->migrate_disable)
 		goto out;
 
 	dest_cpu = cpumask_any_and(cpu_active_mask, new_mask);
@@ -4940,6 +4941,83 @@ out:
 }
 EXPORT_SYMBOL_GPL(set_cpus_allowed_ptr);
 
+void migrate_disable(void)
+{
+	struct task_struct *p = current;
+	const struct cpumask *mask;
+	unsigned long flags;
+	struct rq *rq;
+
+	preempt_disable();
+	if (p->migrate_disable) {
+		p->migrate_disable++;
+		preempt_enable();
+		return;
+	}
+
+	pin_current_cpu();
+	if (unlikely(!scheduler_running)) {
+		p->migrate_disable = 1;
+		preempt_enable();
+		return;
+	}
+	rq = task_rq_lock(p, &flags);
+	p->migrate_disable = 1;
+	mask = tsk_cpus_allowed(p);
+
+	WARN_ON(!cpumask_test_cpu(smp_processor_id(), mask));
+
+	if (!cpumask_equal(&p->cpus_allowed, mask)) {
+		if (p->sched_class->set_cpus_allowed)
+			p->sched_class->set_cpus_allowed(p, mask);
+		p->nr_cpus_allowed = cpumask_weight(mask);
+	}
+	task_rq_unlock(rq, p, &flags);
+	preempt_enable();
+}
+EXPORT_SYMBOL(migrate_disable);
+
+void migrate_enable(void)
+{
+	struct task_struct *p = current;
+	const struct cpumask *mask;
+	unsigned long flags;
+	struct rq *rq;
+
+	WARN_ON_ONCE(p->migrate_disable <= 0);
+
+	preempt_disable();
+	if (p->migrate_disable > 1) {
+		p->migrate_disable--;
+		preempt_enable();
+		return;
+	}
+
+	if (unlikely(!scheduler_running)) {
+		p->migrate_disable = 0;
+		unpin_current_cpu();
+		preempt_enable();
+		return;
+	}
+
+	rq = task_rq_lock(p, &flags);
+	p->migrate_disable = 0;
+	mask = tsk_cpus_allowed(p);
+
+	WARN_ON(!cpumask_test_cpu(smp_processor_id(), mask));
+
+	if (!cpumask_equal(&p->cpus_allowed, mask)) {
+		if (p->sched_class->set_cpus_allowed)
+			p->sched_class->set_cpus_allowed(p, mask);
+		p->nr_cpus_allowed = cpumask_weight(mask);
+	}
+
+	task_rq_unlock(rq, p, &flags);
+	unpin_current_cpu();
+	preempt_enable();
+}
+EXPORT_SYMBOL(migrate_enable);
+
 /*
  * Move (not current) task off this cpu, onto dest cpu. We're doing
  * this because either it can't run here any more (set_cpus_allowed()
--- a/lib/smp_processor_id.c
+++ b/lib/smp_processor_id.c
@@ -39,9 +39,9 @@ notrace unsigned int debug_smp_processor
 	if (!printk_ratelimit())
 		goto out_enable;
 
-	printk(KERN_ERR "BUG: using smp_processor_id() in preemptible [%08x] "
-			"code: %s/%d\n",
-			preempt_count() - 1, current->comm, current->pid);
+	printk(KERN_ERR "BUG: using smp_processor_id() in preemptible [%08x %08x] "
+	       "code: %s/%d\n", preempt_count() - 1,
+	       current->migrate_disable, current->comm, current->pid);
 	print_symbol("caller is %s\n", (long)__builtin_return_address(0));
 	dump_stack();