diff --git a/arch/x86/kernel/cpu/mtrr/main.c b/arch/x86/kernel/cpu/mtrr/main.c
index 3d17bc7f06e6..707b6377adf1 100644
--- a/arch/x86/kernel/cpu/mtrr/main.c
+++ b/arch/x86/kernel/cpu/mtrr/main.c
@@ -137,55 +137,43 @@ static void __init init_table(void)
 }
 
 struct set_mtrr_data {
-	atomic_t	count;
-	atomic_t	gate;
 	unsigned long	smp_base;
 	unsigned long	smp_size;
 	unsigned int	smp_reg;
 	mtrr_type	smp_type;
 };
 
-static DEFINE_PER_CPU(struct cpu_stop_work, mtrr_work);
-
 /**
- * mtrr_work_handler - Synchronisation handler. Executed by "other" CPUs.
+ * mtrr_rendezvous_handler - Work done in the synchronization handler. Executed
+ * by all the CPUs.
  * @info: pointer to mtrr configuration data
  *
  * Returns nothing.
  */
-static int mtrr_work_handler(void *info)
+static int mtrr_rendezvous_handler(void *info)
 {
 #ifdef CONFIG_SMP
 	struct set_mtrr_data *data = info;
-	unsigned long flags;
 
-	atomic_dec(&data->count);
-	while (!atomic_read(&data->gate))
-		cpu_relax();
-
-	local_irq_save(flags);
-
-	atomic_dec(&data->count);
-	while (atomic_read(&data->gate))
-		cpu_relax();
-
-	/*  The master has cleared me to execute  */
+	/*
+	 * We use this same function to initialize the mtrrs during boot,
+	 * resume, runtime cpu online and on an explicit request to set a
+	 * specific MTRR.
+	 *
+	 * During boot or suspend, the state of the boot cpu's mtrrs has been
+	 * saved, and we want to replicate that across all the cpus that come
+	 * online (either at the end of boot or resume or during a runtime cpu
+	 * online). If we're doing that, @reg is set to something special and on
+	 * all the cpu's we do mtrr_if->set_all() (On the logical cpu that
+	 * started the boot/resume sequence, this might be a duplicate
+	 * set_all()).
+	 */
 	if (data->smp_reg != ~0U) {
 		mtrr_if->set(data->smp_reg, data->smp_base,
 			     data->smp_size, data->smp_type);
-	} else if (mtrr_aps_delayed_init) {
-		/*
-		 * Initialize the MTRRs inaddition to the synchronisation.
-		 */
+	} else if (mtrr_aps_delayed_init || !cpu_online(smp_processor_id())) {
 		mtrr_if->set_all();
 	}
-
-	atomic_dec(&data->count);
-	while (!atomic_read(&data->gate))
-		cpu_relax();
-
-	atomic_dec(&data->count);
-	local_irq_restore(flags);
 #endif
 	return 0;
 }
@@ -223,20 +211,11 @@ static inline int types_compatible(mtrr_type type1, mtrr_type type2)
  * 14. Wait for buddies to catch up
  * 15. Enable interrupts.
  *
- * What does that mean for us? Well, first we set data.count to the number
- * of CPUs. As each CPU announces that it started the rendezvous handler by
- * decrementing the count, We reset data.count and set the data.gate flag
- * allowing all the cpu's to proceed with the work. As each cpu disables
- * interrupts, it'll decrement data.count once. We wait until it hits 0 and
- * proceed. We clear the data.gate flag and reset data.count. Meanwhile, they
- * are waiting for that flag to be cleared. Once it's cleared, each
- * CPU goes through the transition of updating MTRRs.
- * The CPU vendors may each do it differently,
- * so we call mtrr_if->set() callback and let them take care of it.
- * When they're done, they again decrement data->count and wait for data.gate
- * to be set.
- * When we finish, we wait for data.count to hit 0 and toggle the data.gate flag
- * Everyone then enables interrupts and we all continue on.
+ * What does that mean for us? Well, stop_machine() will ensure that
+ * the rendezvous handler is started on each CPU. And in lockstep they
+ * do the state transition of disabling interrupts, updating MTRR's
+ * (the CPU vendors may each do it differently, so we call mtrr_if->set()
+ * callback and let them take care of it.) and enabling interrupts.
  *
  * Note that the mechanism is the same for UP systems, too; all the SMP stuff
  * becomes nops.
@@ -244,115 +223,26 @@ static inline int types_compatible(mtrr_type type1, mtrr_type type2)
 static void
 set_mtrr(unsigned int reg, unsigned long base, unsigned long size, mtrr_type type)
 {
-	struct set_mtrr_data data;
-	unsigned long flags;
-	int cpu;
+	struct set_mtrr_data data = { .smp_reg = reg,
+				      .smp_base = base,
+				      .smp_size = size,
+				      .smp_type = type
+				    };
 
-#ifdef CONFIG_SMP
-	/*
-	 * If this cpu is not yet active, we are in the cpu online path. There
-	 * can be no stop_machine() in parallel, as stop machine ensures this
-	 * by using get_online_cpus(). We can skip taking the stop_cpus_mutex,
-	 * as we don't need it and also we can't afford to block while waiting
-	 * for the mutex.
-	 *
-	 * If this cpu is active, we need to prevent stop_machine() happening
-	 * in parallel by taking the stop cpus mutex.
-	 *
-	 * Also, this is called in the context of cpu online path or in the
-	 * context where cpu hotplug is prevented. So checking the active status
-	 * of the raw_smp_processor_id() is safe.
-	 */
-	if (cpu_active(raw_smp_processor_id()))
-		mutex_lock(&stop_cpus_mutex);
-#endif
+	stop_machine(mtrr_rendezvous_handler, &data, cpu_online_mask);
+}
 
-	preempt_disable();
+static void set_mtrr_from_inactive_cpu(unsigned int reg, unsigned long base,
+				      unsigned long size, mtrr_type type)
+{
+	struct set_mtrr_data data = { .smp_reg = reg,
+				      .smp_base = base,
+				      .smp_size = size,
+				      .smp_type = type
+				    };
 
-	data.smp_reg = reg;
-	data.smp_base = base;
-	data.smp_size = size;
-	data.smp_type = type;
-	atomic_set(&data.count, num_booting_cpus() - 1);
-
-	/* Make sure data.count is visible before unleashing other CPUs */
-	smp_wmb();
-	atomic_set(&data.gate, 0);
-
-	/* Start the ball rolling on other CPUs */
-	for_each_online_cpu(cpu) {
-		struct cpu_stop_work *work = &per_cpu(mtrr_work, cpu);
-
-		if (cpu == smp_processor_id())
-			continue;
-
-		stop_one_cpu_nowait(cpu, mtrr_work_handler, &data, work);
-	}
-
-
-	while (atomic_read(&data.count))
-		cpu_relax();
-
-	/* Ok, reset count and toggle gate */
-	atomic_set(&data.count, num_booting_cpus() - 1);
-	smp_wmb();
-	atomic_set(&data.gate, 1);
-
-	local_irq_save(flags);
-
-	while (atomic_read(&data.count))
-		cpu_relax();
-
-	/* Ok, reset count and toggle gate */
-	atomic_set(&data.count, num_booting_cpus() - 1);
-	smp_wmb();
-	atomic_set(&data.gate, 0);
-
-	/* Do our MTRR business */
-
-	/*
-	 * HACK!
-	 *
-	 * We use this same function to initialize the mtrrs during boot,
-	 * resume, runtime cpu online and on an explicit request to set a
-	 * specific MTRR.
-	 *
-	 * During boot or suspend, the state of the boot cpu's mtrrs has been
-	 * saved, and we want to replicate that across all the cpus that come
-	 * online (either at the end of boot or resume or during a runtime cpu
-	 * online). If we're doing that, @reg is set to something special and on
-	 * this cpu we still do mtrr_if->set_all(). During boot/resume, this
-	 * is unnecessary if at this point we are still on the cpu that started
-	 * the boot/resume sequence. But there is no guarantee that we are still
-	 * on the same cpu. So we do mtrr_if->set_all() on this cpu aswell to be
-	 * sure that we are in sync with everyone else.
-	 */
-	if (reg != ~0U)
-		mtrr_if->set(reg, base, size, type);
-	else
-		mtrr_if->set_all();
-
-	/* Wait for the others */
-	while (atomic_read(&data.count))
-		cpu_relax();
-
-	atomic_set(&data.count, num_booting_cpus() - 1);
-	smp_wmb();
-	atomic_set(&data.gate, 1);
-
-	/*
-	 * Wait here for everyone to have seen the gate change
-	 * So we're the last ones to touch 'data'
-	 */
-	while (atomic_read(&data.count))
-		cpu_relax();
-
-	local_irq_restore(flags);
-	preempt_enable();
-#ifdef CONFIG_SMP
-	if (cpu_active(raw_smp_processor_id()))
-		mutex_unlock(&stop_cpus_mutex);
-#endif
+	stop_machine_from_inactive_cpu(mtrr_rendezvous_handler, &data,
+				       cpu_callout_mask);
 }
 
 /**
@@ -806,7 +696,7 @@ void mtrr_ap_init(void)
 	 *   2. cpu hotadd time. We let mtrr_add/del_page hold cpuhotplug
 	 *      lock to prevent mtrr entry changes
 	 */
-	set_mtrr(~0U, 0, 0, 0);
+	set_mtrr_from_inactive_cpu(~0U, 0, 0, 0);
 }
 
 /**
diff --git a/include/linux/stop_machine.h b/include/linux/stop_machine.h
index e0f2da25d751..4a9d0c7edc65 100644
--- a/include/linux/stop_machine.h
+++ b/include/linux/stop_machine.h
@@ -27,8 +27,6 @@ struct cpu_stop_work {
 	struct cpu_stop_done	*done;
 };
 
-extern struct mutex stop_cpus_mutex;
-
 int stop_one_cpu(unsigned int cpu, cpu_stop_fn_t fn, void *arg);
 void stop_one_cpu_nowait(unsigned int cpu, cpu_stop_fn_t fn, void *arg,
 			 struct cpu_stop_work *work_buf);
diff --git a/kernel/stop_machine.c b/kernel/stop_machine.c
index e8f05b14cd43..c1124752e1d3 100644
--- a/kernel/stop_machine.c
+++ b/kernel/stop_machine.c
@@ -132,8 +132,8 @@ void stop_one_cpu_nowait(unsigned int cpu, cpu_stop_fn_t fn, void *arg,
 	cpu_stop_queue_work(&per_cpu(cpu_stopper, cpu), work_buf);
 }
 
-DEFINE_MUTEX(stop_cpus_mutex);
 /* static data for stop_cpus */
+static DEFINE_MUTEX(stop_cpus_mutex);
 static DEFINE_PER_CPU(struct cpu_stop_work, stop_cpus_work);
 
 static void queue_stop_cpus_work(const struct cpumask *cpumask,