diff --git a/drivers/cpufreq/Kconfig.arm b/drivers/cpufreq/Kconfig.arm
index a0b3661d90b0e9..03dea5c6f636df 100644
--- a/drivers/cpufreq/Kconfig.arm
+++ b/drivers/cpufreq/Kconfig.arm
@@ -71,6 +71,15 @@ config ARM_EXYNOS4X12_CPUFREQ
 	  This adds the CPUFreq driver for Samsung EXYNOS4X12
 	  SoC (EXYNOS4212 or EXYNOS4412).
 
+config ARM_EXYNOS4x12_DVFS_HOTPLUG
+	bool "DVFS hotplug driver"
+	depends on ARM_EXYNOS4X12_CPUFREQ && HOTPLUG_CPU
+	default y
+	help
+	  This adds the CPUFreq DVFS hotplug capability for Samsung EXYNOS4x12
+	  SoC (EXYNOS4212 or EXYNOS4412).
+	  
+
 config ARM_EXYNOS5250_CPUFREQ
 	def_bool SOC_EXYNOS5250
 	help
diff --git a/drivers/cpufreq/Makefile b/drivers/cpufreq/Makefile
index fadc4d496e2fc5..4093d66a19d100 100644
--- a/drivers/cpufreq/Makefile
+++ b/drivers/cpufreq/Makefile
@@ -52,6 +52,7 @@ obj-$(CONFIG_ARM_EXYNOS4X12_CPUFREQ)	+= exynos4x12-cpufreq.o
 obj-$(CONFIG_ARM_EXYNOS5250_CPUFREQ)	+= exynos5250-cpufreq.o
 obj-$(CONFIG_ARM_OMAP2PLUS_CPUFREQ)     += omap-cpufreq.o
 obj-$(CONFIG_ARM_SPEAR_CPUFREQ)		+= spear-cpufreq.o
+obj-$(CONFIG_ARM_EXYNOS4x12_DVFS_HOTPLUG)	+= exynos4x12-dvfs-hotplug.o
 
 ##################################################################################
 # PowerPC platform drivers
diff --git a/drivers/cpufreq/exynos4x12-dvfs-hotplug.c b/drivers/cpufreq/exynos4x12-dvfs-hotplug.c
new file mode 100644
index 00000000000000..a3fdece7fc1108
--- /dev/null
+++ b/drivers/cpufreq/exynos4x12-dvfs-hotplug.c
@@ -0,0 +1,348 @@
+/*
+ * drivers/cpufreq/exynos4x12-dvfs-hotplug.c
+ *
+ * DVFS cpu-hotplug driver for Samsung Exynos 4x12 SoCs
+ *
+ * Author: Gokturk Gezer <gokturk@apache.org>
+ *
+ * This program is free software; you can redistribute it and/or modify
+ * it under the terms of the GNU General Public License version 2 as
+ * published by the Free Software Foundation.
+ */
+
+#include <linux/sched.h>
+#include <linux/cpufreq.h>
+#include <linux/cpu.h>
+#include <linux/err.h>
+#include <linux/notifier.h>
+#include <linux/reboot.h>
+#include <linux/suspend.h>
+#include <linux/io.h>
+#include <linux/workqueue.h>
+
+#include <plat/cpu.h>
+
+// tunables
+static unsigned int hotplug_min_cpu_count;
+static unsigned int hotplug_max_cpu_count;
+static unsigned int hotplug_freq_load_tolerance;
+static unsigned int hotplug_tick_interval;
+static unsigned int hotplug_tick_anticipation;
+
+// cpufreq state
+static char governor_name[CPUFREQ_NAME_LEN];
+static unsigned int freq_current;
+static unsigned int freq_min;
+static unsigned int freq_max;
+
+// hotplug state
+static int freq_out_target;
+static int freq_out_limit;
+static int freq_in_target;
+static int freq_in_limit;
+static unsigned int can_hotplug;
+static struct delayed_work hotplug_dynamic_tick_work;
+static struct delayed_work hotplug_fixed_tick_work;
+void (*dynamic_tick_step)(void);
+static unsigned int fixed_tick_cpu_count;
+
+// function declerations
+static void dynamic_hotplug_work();
+static void fixed_hotplug_work();
+static void start_hotplug_dynamic_tick();
+static void start_hotplug_fixed_tick(unsigned int);
+static void stop_hotplug_ticks();
+static void boot_fixed_cores();
+static void cpu_increase();
+static void cpu_decrease();
+static void hotplug_deploy(struct cpufreq_policy*);
+
+static void __hotplug_tick_step_freq_track()
+{
+	unsigned int tolerated_freq_in, tolerated_freq_out;
+
+	tolerated_freq_in = freq_max / 100 * hotplug_freq_load_tolerance;
+	tolerated_freq_out = freq_max / 100 * (hotplug_freq_load_tolerance - 20);
+	if (tolerated_freq_out < freq_min)
+		tolerated_freq_out = freq_min;
+
+	if (freq_current >= tolerated_freq_in)
+	{
+		if (freq_out_target > 0)
+			freq_out_target = 0;
+
+		if (++freq_in_target == freq_in_limit)
+		{
+			cpu_increase();
+			freq_in_target = 0;
+
+			if (hotplug_tick_anticipation)
+				freq_out_target = -1 * freq_out_limit;
+		}
+	}
+	else if (freq_current <= tolerated_freq_out)
+	{
+		freq_in_target = 0;
+		if (++freq_out_target == freq_out_limit)
+		{
+			cpu_decrease();
+			freq_out_target = 0;
+		}
+	}
+}
+
+static void dynamic_hotplug_work()
+{
+	(*dynamic_tick_step)();
+
+	start_hotplug_dynamic_tick();
+}
+
+static void fixed_hotplug_work()
+{
+	boot_fixed_cores();
+}
+
+static void start_hotplug_dynamic_tick()
+{
+	schedule_delayed_work_on(0, &hotplug_dynamic_tick_work,
+			msecs_to_jiffies(hotplug_tick_interval));
+}
+
+static void start_hotplug_fixed_tick(unsigned int cpu_count)
+{
+	fixed_tick_cpu_count = cpu_count;
+
+	schedule_delayed_work_on(0, &hotplug_fixed_tick_work,
+			msecs_to_jiffies(500));
+}
+
+static void stop_hotplug_ticks()
+{
+	cancel_delayed_work_sync(&hotplug_dynamic_tick_work);
+	cancel_delayed_work_sync(&hotplug_fixed_tick_work);
+}
+
+static void boot_fixed_cores()
+{
+	int operation_count;
+	unsigned int i,online_count;
+
+	void (*fix_operation)(void) = cpu_increase;
+
+	for(i = 0, online_count = 0; i < 4; i++)
+	{
+		if(cpu_online(i))
+			online_count++;
+	}
+
+	operation_count = fixed_tick_cpu_count - online_count;
+	if(operation_count < 0)
+	{
+		operation_count *= -1;
+		fix_operation = cpu_decrease;
+	}
+
+	for(i = 0; i < operation_count; i++)
+		(*fix_operation)();
+}
+
+static void cpu_increase()
+{
+	unsigned int i;
+
+	if(num_online_cpus() >= hotplug_max_cpu_count)
+		return;
+
+	for(i = 0; i < 4; i++)
+	{
+		if(!cpu_online(i))
+		{
+			cpu_up(i);
+			break;
+		}
+	}
+}
+
+static void cpu_decrease()
+{
+	unsigned int i;
+
+	if(num_online_cpus() <= hotplug_min_cpu_count)
+		return;
+
+	for(i = 3; i >= 0; i--)
+	{
+		if(cpu_online(i))
+		{
+			cpu_down(i);
+			break;
+		}
+	}
+}
+
+static void hotplug_deploy(struct cpufreq_policy * policy)
+{
+	unsigned int cpu;
+
+
+	/*
+	 * no governor, no hot-plug, all cores up
+	 */
+	if (!policy->governor)
+	{
+		stop_hotplug_ticks();
+
+		for_each_cpu_mask(cpu, policy->cpus[0])
+		{
+			if (!cpu_online(cpu))
+				cpu_up(cpu);
+		}
+
+		return;
+	}
+
+	freq_max = policy->max;
+	freq_min = policy->min;
+
+	if( 0 != strnicmp(policy->governor->name, governor_name, CPUFREQ_NAME_LEN))
+	{
+		stop_hotplug_ticks();
+
+		strncpy(governor_name, policy->governor->name, CPUFREQ_NAME_LEN);
+
+		if (0 == strnicmp(governor_name, "performance", CPUFREQ_NAME_LEN))
+		{
+			start_hotplug_fixed_tick(hotplug_max_cpu_count);
+		}
+		else if (0 == strnicmp(governor_name, "powersave", CPUFREQ_NAME_LEN))
+		{
+			start_hotplug_fixed_tick(hotplug_min_cpu_count);
+		}
+		else
+		{
+			dynamic_tick_step = __hotplug_tick_step_freq_track;
+			start_hotplug_dynamic_tick();
+		}
+	}
+}
+
+static int hotplug_cpufreq_transition(struct notifier_block *nb,
+		unsigned long val, void *data)
+{
+	struct cpufreq_freqs *freqs = (struct cpufreq_freqs *) data;
+
+	if ((val == CPUFREQ_POSTCHANGE))
+		freq_current = freqs->new;
+
+	return 0;
+}
+
+static int hotplug_cpufreq_policy(struct notifier_block *nb, unsigned long val,	void * data)
+{
+	struct cpufreq_policy * policy = (struct cpufreq_policy*) data;
+
+	if (val != CPUFREQ_ADJUST)
+		return 0;
+
+
+	hotplug_deploy(policy);
+
+	return 0;
+}
+
+static int hotplug_pm_transition(struct notifier_block *nb,	unsigned long val, void *data)
+{
+	switch (val) {
+	case PM_SUSPEND_PREPARE:
+		stop_hotplug_ticks();
+		can_hotplug = 0;
+		freq_out_target = 0;
+		freq_in_target = 0;
+		break;
+	case PM_POST_RESTORE:
+	case PM_POST_SUSPEND:
+		can_hotplug = 1;
+		start_hotplug_dynamic_tick();
+		break;
+	}
+
+	return 0;
+}
+
+static struct notifier_block dvfs_hotplug = { .notifier_call =
+		hotplug_cpufreq_transition, };
+
+static struct notifier_block dvfs_policy_change =
+	{ .notifier_call = hotplug_cpufreq_policy, };
+
+static struct notifier_block pm_hotplug =
+	{ .notifier_call = hotplug_pm_transition, };
+
+/*
+ * Note : This function should be called after intialization of CPUFreq
+ * driver for exynos4. The cpufreq_frequency_table for exynos4 should be
+ * established before calling this function.
+ */
+static int __init exynos4_dvfs_hotplug_init(void)
+{
+	int i, register_result = 0;
+	struct cpufreq_frequency_table *table;
+	unsigned int freq;
+	struct cpufreq_policy policy;
+
+	hotplug_min_cpu_count = 2;
+	if(soc_is_exynos4412())
+		hotplug_max_cpu_count = 4;
+	else
+		hotplug_max_cpu_count = 2;
+	hotplug_freq_load_tolerance = 60;
+	hotplug_tick_interval = 200;
+	hotplug_tick_anticipation = 1;
+
+	freq_out_target = 0;
+	freq_out_limit = 3;
+	freq_in_target = 0;
+	freq_in_limit = 3;
+	can_hotplug = 1;
+
+	table = cpufreq_frequency_get_table(0);
+	if (IS_ERR(table))
+	{
+		printk(KERN_ERR "%s: Check loading cpufreq before\n", __func__);
+		return PTR_ERR(table);
+	}
+
+	for (i=0; table[i].frequency != CPUFREQ_TABLE_END; i++)
+	{
+		freq = table[i].frequency;
+
+		if (freq != CPUFREQ_ENTRY_INVALID && freq > freq_max)
+			freq_max = freq;
+		else if (freq != CPUFREQ_ENTRY_INVALID && freq_min > freq)
+			freq_min = freq;
+	}
+
+	freq_current = freq_min;
+
+	INIT_DEFERRABLE_WORK(&hotplug_dynamic_tick_work, dynamic_hotplug_work);
+	INIT_DEFERRABLE_WORK(&hotplug_fixed_tick_work, fixed_hotplug_work);
+
+	printk(KERN_INFO "%s, max(%d),min(%d)\n", __func__, freq_max, freq_min);
+
+	register_result |= register_pm_notifier(&pm_hotplug);
+
+	register_result |= cpufreq_register_notifier(&dvfs_policy_change,
+			CPUFREQ_POLICY_NOTIFIER);
+
+	register_result |= cpufreq_register_notifier(&dvfs_hotplug,
+			CPUFREQ_TRANSITION_NOTIFIER);
+
+	cpufreq_get_policy(&policy, 0);
+	hotplug_deploy(&policy);
+
+	return register_result;
+
+}
+
+late_initcall(exynos4_dvfs_hotplug_init);