Merge bk://linux-dj.bkbits.net/cpufreq

into ppc970.osdl.org:/home/torvalds/v2.6/linux

arch/i386/kernel/cpu/cpufreq/longhaul.c

@@ -38,6 +38,17 @@
 
 #define PFX "longhaul: "
 
+#define TYPE_LONGHAUL_V1	1
+#define TYPE_LONGHAUL_V2	2
+#define TYPE_POWERSAVER		3
+
+#define	CPU_SAMUEL	1
+#define	CPU_SAMUEL2	2
+#define	CPU_EZRA	3
+#define	CPU_EZRA_T	4
+#define	CPU_NEHEMIAH	5
+
+static int cpu_model;
 static unsigned int numscales=16, numvscales;
 static unsigned int fsb;
 static int minvid, maxvid;
@@ -73,9 +84,23 @@ static int voltage_table[32];
 static unsigned int highest_speed, lowest_speed; /* kHz */
 static int longhaul_version;
 static struct cpufreq_frequency_table *longhaul_table;
+static char speedbuffer[8];
+
+static char *print_speed(int speed)
+{
+	if (speed > 1000) {
+		if (speed%1000 == 0)
+			sprintf (speedbuffer, "%dGHz", speed/1000);
+		else
+			sprintf (speedbuffer, "%d.%dGHz", speed/1000, (speed%1000)/100);
+	} else
+		sprintf (speedbuffer, "%dMHz", speed);
+
+	return speedbuffer;
+}
 
 
-static unsigned int calc_speed(int mult, int fsb)
+static unsigned int calc_speed(int mult)
 {
 	int khz;
 	khz = (mult/10)*fsb;
@@ -92,7 +117,7 @@ static int longhaul_get_cpu_mult(void)
 
 	rdmsr (MSR_IA32_EBL_CR_POWERON, lo, hi);
 	invalue = (lo & (1<<22|1<<23|1<<24|1<<25)) >>22;
-	if (longhaul_version==2 || longhaul_version==3) {
+	if (longhaul_version==TYPE_LONGHAUL_V2 || longhaul_version==TYPE_POWERSAVER) {
 		if (lo & (1<<27))
 			invalue+=16;
 	}
@@ -101,8 +126,21 @@ static int longhaul_get_cpu_mult(void)
 
 
 static void do_powersaver(union msr_longhaul *longhaul,
-			unsigned int clock_ratio_index, int version)
+			unsigned int clock_ratio_index)
 {
+	int version;
+
+	switch (cpu_model) {
+	case CPU_EZRA_T:
+		version = 3;
+		break;
+	case CPU_NEHEMIAH:
+		version = 0xf;
+		break;
+	default:
+		return;
+	}
+
 	rdmsrl(MSR_VIA_LONGHAUL, longhaul->val);
 	longhaul->bits.SoftBusRatio = clock_ratio_index & 0xf;
 	longhaul->bits.SoftBusRatio4 = (clock_ratio_index & 0x10) >> 4;
@@ -125,7 +163,7 @@ static void do_powersaver(union msr_longhaul *longhaul,
  * longhaul_set_cpu_frequency()
  * @clock_ratio_index : bitpattern of the new multiplier.
  *
- * Sets a new clock ratio, and -if applicable- a new Front Side Bus
+ * Sets a new clock ratio.
  */
 
 static void longhaul_setstate(unsigned int clock_ratio_index)
@@ -134,22 +172,28 @@ static void longhaul_setstate(unsigned int clock_ratio_index)
 	struct cpufreq_freqs freqs;
 	union msr_longhaul longhaul;
 	union msr_bcr2 bcr2;
+	static unsigned int old_ratio=-1;
+
+	if (old_ratio == clock_ratio_index)
+		return;
+	old_ratio = clock_ratio_index;
 
 	mult = clock_ratio[clock_ratio_index];
 	if (mult == -1)
 		return;
 
-	speed = calc_speed (mult, fsb);
+	speed = calc_speed(mult);
 	if ((speed > highest_speed) || (speed < lowest_speed))
 		return;
 
-	freqs.old = calc_speed (longhaul_get_cpu_mult(), fsb);
+	freqs.old = calc_speed(longhaul_get_cpu_mult());
 	freqs.new = speed;
 	freqs.cpu = 0; /* longhaul.c is UP only driver */
 
 	cpufreq_notify_transition(&freqs, CPUFREQ_PRECHANGE);
 
-	dprintk (KERN_INFO PFX "FSB:%d Mult:%d.%dx\n", fsb, mult/10, mult%10);
+	dprintk (KERN_INFO PFX "Setting to FSB:%dMHz Mult:%d.%dx (%s)\n",
+			fsb, mult/10, mult%10, print_speed(speed/1000));
 
 	switch (longhaul_version) {
 
@@ -160,7 +204,8 @@ static void longhaul_setstate(unsigned int clock_ratio_index)
 	 * *NB* Until we get voltage scaling working v1 & v2 are the same code.
 	 * Longhaul v2 appears in Samuel2 Steppings 1->7 [C5b] and Ezra [C5C]
 	 */
-	case 1:
+	case TYPE_LONGHAUL_V1:
+	case TYPE_LONGHAUL_V2:
 		rdmsrl (MSR_VIA_BCR2, bcr2.val);
 		/* Enable software clock multiplier */
 		bcr2.bits.ESOFTBF = 1;
@@ -180,26 +225,18 @@ static void longhaul_setstate(unsigned int clock_ratio_index)
 		break;
 
 	/*
-	 * Longhaul v3 (aka Powersaver). (Ezra-T [C5M])
+	 * Longhaul v3 (aka Powersaver). (Ezra-T [C5M] & Nehemiah [C5N])
 	 * We can scale voltage with this too, but that's currently
 	 * disabled until we come up with a decent 'match freq to voltage'
 	 * algorithm.
 	 * When we add voltage scaling, we will also need to do the
 	 * voltage/freq setting in order depending on the direction
 	 * of scaling (like we do in powernow-k7.c)
-	 */
-	case 2:
-		do_powersaver(&longhaul, clock_ratio_index, 3);
-		break;
-
-	/*
-	 * Powersaver. (Nehemiah [C5N])
-	 * As for Ezra-T, we don't do voltage yet.
-	 * This can do FSB scaling too, but it has never been proven
+	 * Nehemiah can do FSB scaling too, but this has never been proven
 	 * to work in practice.
 	 */
-	case 3:
-		do_powersaver(&longhaul, clock_ratio_index, 0xf);
+	case TYPE_POWERSAVER:
+		do_powersaver(&longhaul, clock_ratio_index);
 		break;
 	}
 
@@ -249,7 +286,6 @@ static int guess_fsb(void)
 
 static int __init longhaul_get_ranges(void)
 {
-	struct cpuinfo_x86 *c = cpu_data;
 	unsigned long invalue;
 	unsigned int multipliers[32]= {
 		50,30,40,100,55,35,45,95,90,70,80,60,120,75,85,65,
@@ -261,62 +297,66 @@ static int __init longhaul_get_ranges(void)
 	unsigned int eblcr_fsb_table_v2[] = { 133, 100, -1, 66 };
 
 	switch (longhaul_version) {
-	case 1:
+	case TYPE_LONGHAUL_V1:
+	case TYPE_LONGHAUL_V2:
 		/* Ugh, Longhaul v1 didn't have the min/max MSRs.
 		   Assume min=3.0x & max = whatever we booted at. */
 		minmult = 30;
 		maxmult = longhaul_get_cpu_mult();
 		rdmsr (MSR_IA32_EBL_CR_POWERON, lo, hi);
 		invalue = (lo & (1<<18|1<<19)) >>18;
-		if (c->x86_model==6)
+		if (cpu_model==CPU_SAMUEL || cpu_model==CPU_SAMUEL2)
 			fsb = eblcr_fsb_table_v1[invalue];
 		else
 			fsb = guess_fsb();
 		break;
 
-	case 2:
-		rdmsrl (MSR_VIA_LONGHAUL, longhaul.val);
-
-		invalue = longhaul.bits.MaxMHzBR;
-		if (longhaul.bits.MaxMHzBR4)
-			invalue += 16;
-		maxmult=multipliers[invalue];
-
-		invalue = longhaul.bits.MinMHzBR;
-		if (longhaul.bits.MinMHzBR4 == 1)
-			minmult = 30;
-		else
-			minmult = multipliers[invalue];
-
-		fsb = eblcr_fsb_table_v2[longhaul.bits.MaxMHzFSB];
-		break;
-
-	case 3:
-		rdmsrl (MSR_VIA_LONGHAUL, longhaul.val);
-
-		/*
-		 * TODO: This code works, but raises a lot of questions.
-		 * - Some Nehemiah's seem to have broken Min/MaxMHzBR's.
-		 *   We get around this by using a hardcoded multiplier of 5.0x
-		 *   for the minimimum speed, and the speed we booted up at for the max.
-		 *   This is done in longhaul_get_cpu_mult() by reading the EBLCR register.
-		 * - According to some VIA documentation EBLCR is only
-		 *   in pre-Nehemiah C3s. How this still works is a mystery.
-		 *   We're possibly using something undocumented and unsupported,
-		 *   But it works, so we don't grumble.
-		 */
-		minmult=50;
-		maxmult=longhaul_get_cpu_mult();
-
-		/* Starting with the 1.2GHz parts, theres a 200MHz bus. */
-		if ((cpu_khz/1000) > 1200)
-			fsb = 200;
-		else
+	case TYPE_POWERSAVER:
+		/* Ezra-T */
+		if (cpu_model==CPU_EZRA_T) {
+			rdmsrl (MSR_VIA_LONGHAUL, longhaul.val);
+			invalue = longhaul.bits.MaxMHzBR;
+			if (longhaul.bits.MaxMHzBR4)
+				invalue += 16;
+			maxmult=multipliers[invalue];
+
+			invalue = longhaul.bits.MinMHzBR;
+			if (longhaul.bits.MinMHzBR4 == 1)
+				minmult = 30;
+			else
+				minmult = multipliers[invalue];
 			fsb = eblcr_fsb_table_v2[longhaul.bits.MaxMHzFSB];
-		break;
+			break;
+		}
+
+		/* Nehemiah */
+		if (cpu_model==CPU_NEHEMIAH) {
+			rdmsrl (MSR_VIA_LONGHAUL, longhaul.val);
+
+			/*
+			 * TODO: This code works, but raises a lot of questions.
+			 * - Some Nehemiah's seem to have broken Min/MaxMHzBR's.
+			 *   We get around this by using a hardcoded multiplier of 4.0x
+			 *   for the minimimum speed, and the speed we booted up at for the max.
+			 *   This is done in longhaul_get_cpu_mult() by reading the EBLCR register.
+			 * - According to some VIA documentation EBLCR is only
+			 *   in pre-Nehemiah C3s. How this still works is a mystery.
+			 *   We're possibly using something undocumented and unsupported,
+			 *   But it works, so we don't grumble.
+			 */
+			minmult=40;
+			maxmult=longhaul_get_cpu_mult();
+
+			/* Starting with the 1.2GHz parts, theres a 200MHz bus. */
+			if ((cpu_khz/1000) > 1200)
+				fsb = 200;
+			else
+				fsb = eblcr_fsb_table_v2[longhaul.bits.MaxMHzFSB];
+			break;
+		}
 	}
 
-	dprintk (KERN_INFO PFX "MinMult=%d.%dx MaxMult=%d.%dx\n",
+	dprintk (KERN_INFO PFX "MinMult:%d.%dx MaxMult:%d.%dx\n",
 		 minmult/10, minmult%10, maxmult/10, maxmult%10);
 
 	if (fsb == -1) {
@@ -324,10 +364,11 @@ static int __init longhaul_get_ranges(void)
 		return -EINVAL;
 	}
 
-	highest_speed = calc_speed (maxmult, fsb);
-	lowest_speed = calc_speed (minmult,fsb);
-	dprintk (KERN_INFO PFX "FSB: %dMHz Lowestspeed=%dMHz Highestspeed=%dMHz\n",
-		 fsb, lowest_speed/1000, highest_speed/1000);
+	highest_speed = calc_speed(maxmult);
+	lowest_speed = calc_speed(minmult);
+	dprintk (KERN_INFO PFX "FSB:%dMHz  ", fsb);
+	dprintk ("Lowest speed:%s  ", print_speed(lowest_speed/1000));
+	dprintk ("Highest speed:%s\n", print_speed(highest_speed/1000));
 
 	if (lowest_speed == highest_speed) {
 		printk (KERN_INFO PFX "highestspeed == lowest, aborting.\n");
@@ -350,7 +391,7 @@ static int __init longhaul_get_ranges(void)
 			continue;
 		if (ratio > maxmult || ratio < minmult)
 			continue;
-		longhaul_table[k].frequency = calc_speed (ratio, fsb);
+		longhaul_table[k].frequency = calc_speed(ratio);
 		longhaul_table[k].index	= j;
 		k++;
 	}
@@ -426,8 +467,7 @@ static int longhaul_verify(struct cpufreq_policy *policy)
 
 
 static int longhaul_target(struct cpufreq_policy *policy,
-			    unsigned int target_freq,
-			    unsigned int relation)
+			    unsigned int target_freq, unsigned int relation)
 {
 	unsigned int table_index = 0;
 	unsigned int new_clock_ratio = 0;
@@ -442,13 +482,15 @@ static int longhaul_target(struct cpufreq_policy *policy,
 	return 0;
 }
 
+
 static unsigned int longhaul_get(unsigned int cpu)
 {
 	if (cpu)
 		return 0;
-	return (calc_speed (longhaul_get_cpu_mult(), fsb));
+	return calc_speed(longhaul_get_cpu_mult());
 }
 
+
 static int __init longhaul_cpu_init(struct cpufreq_policy *policy)
 {
 	struct cpuinfo_x86 *c = cpu_data;
@@ -457,26 +499,31 @@ static int __init longhaul_cpu_init(struct cpufreq_policy *policy)
 
 	switch (c->x86_model) {
 	case 6:
+		cpu_model = CPU_SAMUEL;
 		cpuname = "C3 'Samuel' [C5A]";
-		longhaul_version=1;
+		longhaul_version = TYPE_LONGHAUL_V1;
 		memcpy (clock_ratio, samuel1_clock_ratio, sizeof(samuel1_clock_ratio));
 		memcpy (eblcr_table, samuel1_eblcr, sizeof(samuel1_eblcr));
 		break;
 
-	case 7:		/* C5B / C5C */
-		longhaul_version=1;
+	case 7:
+		longhaul_version = TYPE_LONGHAUL_V1;
 		switch (c->x86_mask) {
 		case 0:
+			cpu_model = CPU_SAMUEL2;
 			cpuname = "C3 'Samuel 2' [C5B]";
 			/* Note, this is not a typo, early Samuel2's had Samuel1 ratios. */
 			memcpy (clock_ratio, samuel1_clock_ratio, sizeof(samuel1_clock_ratio));
 			memcpy (eblcr_table, samuel2_eblcr, sizeof(samuel2_eblcr));
 			break;
 		case 1 ... 15:
-			if (c->x86_mask < 8)
+			if (c->x86_mask < 8) {
+				cpu_model = CPU_SAMUEL2;
 				cpuname = "C3 'Samuel 2' [C5B]";
-			else
+			} else {
+				cpu_model = CPU_EZRA;
 				cpuname = "C3 'Ezra' [C5C]";
+			}
 			memcpy (clock_ratio, ezra_clock_ratio, sizeof(ezra_clock_ratio));
 			memcpy (eblcr_table, ezra_eblcr, sizeof(ezra_eblcr));
 			break;
@@ -484,15 +531,17 @@ static int __init longhaul_cpu_init(struct cpufreq_policy *policy)
 		break;
 
 	case 8:
+		cpu_model = CPU_EZRA_T;
 		cpuname = "C3 'Ezra-T' [C5M]";
-		longhaul_version=2;
+		longhaul_version = TYPE_POWERSAVER;
 		numscales=32;
 		memcpy (clock_ratio, ezrat_clock_ratio, sizeof(ezrat_clock_ratio));
 		memcpy (eblcr_table, ezrat_eblcr, sizeof(ezrat_eblcr));
 		break;
 
 	case 9:
-		longhaul_version=3;
+		cpu_model = CPU_NEHEMIAH;
+		longhaul_version = TYPE_POWERSAVER;
 		numscales=32;
 		switch (c->x86_mask) {
 		case 0 ... 1:
@@ -518,19 +567,28 @@ static int __init longhaul_cpu_init(struct cpufreq_policy *policy)
 		break;
 	}
 
-	printk (KERN_INFO PFX "VIA %s CPU detected. Longhaul v%d supported.\n",
-					cpuname, longhaul_version);
+	printk (KERN_INFO PFX "VIA %s CPU detected.  ", cpuname);
+	switch (longhaul_version) {
+	case TYPE_LONGHAUL_V1:
+	case TYPE_LONGHAUL_V2:
+		printk ("Longhaul v%d supported.\n", longhaul_version);
+		break;
+	case TYPE_POWERSAVER:
+		printk ("Powersaver supported.\n");
+		break;
+	};
 
 	ret = longhaul_get_ranges();
 	if (ret != 0)
 		return ret;
 
-	if ((longhaul_version==2) && (dont_scale_voltage==0))
+	if ((longhaul_version==TYPE_LONGHAUL_V2 || longhaul_version==TYPE_POWERSAVER) &&
+		 (dont_scale_voltage==0))
 		longhaul_setup_voltagescaling();
 
 	policy->governor = CPUFREQ_DEFAULT_GOVERNOR;
 	policy->cpuinfo.transition_latency = CPUFREQ_ETERNAL;
-	policy->cur = calc_speed (longhaul_get_cpu_mult(), fsb);
+	policy->cur = calc_speed(longhaul_get_cpu_mult());
 
 	ret = cpufreq_frequency_table_cpuinfo(policy, longhaul_table);
 	if (ret)
@@ -563,6 +621,7 @@ static struct cpufreq_driver longhaul_driver = {
 	.attr	= longhaul_attr,
 };
 
+
 static int __init longhaul_init(void)
 {
 	struct cpuinfo_x86 *c = cpu_data;
@@ -580,16 +639,17 @@ static int __init longhaul_init(void)
 	return -ENODEV;
 }
 
+
 static void __exit longhaul_exit(void)
 {
 	int i=0;
-	unsigned int new_clock_ratio;
-
-	while (clock_ratio[i] != maxmult)
-		i++;
 
-	new_clock_ratio = longhaul_table[i].index & 0xFF;
-	longhaul_setstate(new_clock_ratio);
+	for (i=0; i < numscales; i++) {
+		if (clock_ratio[i] == maxmult) {
+			longhaul_setstate(i);
+			break;
+		}
+	}
 
 	cpufreq_unregister_driver(&longhaul_driver);
 	kfree(longhaul_table);

drivers/cpufreq/Kconfig

@@ -69,6 +69,21 @@ config CPU_FREQ_GOV_USERSPACE
 
 	  If in doubt, say Y.
 
+config CPU_FREQ_GOV_ONDEMAND
+	tristate "'ondemand' cpufreq policy governor"
+	depends on CPU_FREQ
+	help
+	  'ondemand' - This driver adds a dynamic cpufreq policy governor.
+	  The governor does a periodic polling and 
+	  changes frequency based on the CPU utilization.
+	  The support for this governor depends on CPU capability to
+	  do fast frequency switching (i.e, very low latency frequency
+	  transitions). 
+
+	  For details, take a look at linux/Documentation/cpu-freq.
+
+	  If in doubt, say N.
+
 config CPU_FREQ_24_API
 	bool "/proc/sys/cpu/ interface (2.4. / OLD)"
 	depends on CPU_FREQ && SYSCTL && CPU_FREQ_GOV_USERSPACE

drivers/cpufreq/Makefile

@@ -5,6 +5,7 @@ obj-$(CONFIG_CPU_FREQ)			+= cpufreq.o
 obj-$(CONFIG_CPU_FREQ_GOV_PERFORMANCE)	+= cpufreq_performance.o
 obj-$(CONFIG_CPU_FREQ_GOV_POWERSAVE)	+= cpufreq_powersave.o
 obj-$(CONFIG_CPU_FREQ_GOV_USERSPACE)	+= cpufreq_userspace.o
+obj-$(CONFIG_CPU_FREQ_GOV_ONDEMAND)	+= cpufreq_ondemand.o
 
 # CPUfreq cross-arch helpers
 obj-$(CONFIG_CPU_FREQ_TABLE)		+= freq_table.o

drivers/cpufreq/cpufreq_ondemand.c

+/*
+ *  drivers/cpufreq/cpufreq_ondemand.c
+ *
+ *  Copyright (C)  2001 Russell King
+ *            (C)  2003 Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>.
+ *                      Jun Nakajima <jun.nakajima@intel.com>
+ *
+ * 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/config.h>
+#include <linux/kernel.h>
+#include <linux/module.h>
+#include <linux/smp.h>
+#include <linux/init.h>
+#include <linux/interrupt.h>
+#include <linux/ctype.h>
+#include <linux/cpufreq.h>
+#include <linux/sysctl.h>
+#include <linux/types.h>
+#include <linux/fs.h>
+#include <linux/sysfs.h>
+#include <linux/sched.h>
+#include <linux/kmod.h>
+#include <linux/workqueue.h>
+#include <linux/jiffies.h>
+#include <linux/config.h>
+#include <linux/kernel_stat.h>
+#include <linux/percpu.h>
+
+/*
+ * dbs is used in this file as a shortform for demandbased switching
+ * It helps to keep variable names smaller, simpler
+ */
+
+#define DEF_FREQUENCY_UP_THRESHOLD		(80)
+#define MIN_FREQUENCY_UP_THRESHOLD		(0)
+#define MAX_FREQUENCY_UP_THRESHOLD		(100)
+
+#define DEF_FREQUENCY_DOWN_THRESHOLD		(20)
+#define MIN_FREQUENCY_DOWN_THRESHOLD		(0)
+#define MAX_FREQUENCY_DOWN_THRESHOLD		(100)
+
+/* 
+ * The polling frequency of this governor depends on the capability of 
+ * the processor. Default polling frequency is 1000 times the transition
+ * latency of the processor. The governor will work on any processor with 
+ * transition latency <= 10mS, using appropriate sampling 
+ * rate.
+ * For CPUs with transition latency > 10mS (mostly drivers with CPUFREQ_ETERNAL)
+ * this governor will not work.
+ * All times here are in uS.
+ */
+static unsigned int 				def_sampling_rate;
+#define MIN_SAMPLING_RATE			(def_sampling_rate / 2)
+#define MAX_SAMPLING_RATE			(500 * def_sampling_rate)
+#define DEF_SAMPLING_RATE_LATENCY_MULTIPLIER	(1000)
+#define DEF_SAMPLING_DOWN_FACTOR		(10)
+#define TRANSITION_LATENCY_LIMIT		(10 * 1000)
+#define sampling_rate_in_HZ(x)			((x * HZ) / (1000 * 1000))
+
+static void do_dbs_timer(void *data);
+
+struct cpu_dbs_info_s {
+	struct cpufreq_policy 	*cur_policy;
+	unsigned int 		prev_cpu_idle_up;
+	unsigned int 		prev_cpu_idle_down;
+	unsigned int 		enable;
+};
+static DEFINE_PER_CPU(struct cpu_dbs_info_s, cpu_dbs_info);
+
+static unsigned int dbs_enable;	/* number of CPUs using this policy */
+
+static DECLARE_MUTEX 	(dbs_sem);
+static DECLARE_WORK	(dbs_work, do_dbs_timer, NULL);
+
+struct dbs_tuners {
+	unsigned int 		sampling_rate;
+	unsigned int		sampling_down_factor;
+	unsigned int		up_threshold;
+	unsigned int		down_threshold;
+};
+
+struct dbs_tuners dbs_tuners_ins = {
+	.up_threshold 		= DEF_FREQUENCY_UP_THRESHOLD,
+	.down_threshold 	= DEF_FREQUENCY_DOWN_THRESHOLD,
+	.sampling_down_factor 	= DEF_SAMPLING_DOWN_FACTOR,
+};
+
+/************************** sysfs interface ************************/
+static ssize_t show_current_freq(struct cpufreq_policy *policy, char *buf)
+{
+	return sprintf (buf, "%u\n", policy->cur);
+}
+
+static ssize_t show_sampling_rate_max(struct cpufreq_policy *policy, char *buf)
+{
+	return sprintf (buf, "%u\n", MAX_SAMPLING_RATE);
+}
+
+static ssize_t show_sampling_rate_min(struct cpufreq_policy *policy, char *buf)
+{
+	return sprintf (buf, "%u\n", MIN_SAMPLING_RATE);
+}
+
+#define define_one_ro(_name) 					\
+static struct freq_attr _name = { 				\
+	.attr = { .name = __stringify(_name), .mode = 0444 }, 	\
+	.show = show_##_name, 					\
+}
+
+define_one_ro(current_freq);
+define_one_ro(sampling_rate_max);
+define_one_ro(sampling_rate_min);
+
+/* cpufreq_ondemand Governor Tunables */
+#define show_one(file_name, object)					\
+static ssize_t show_##file_name						\
+(struct cpufreq_policy *unused, char *buf)				\
+{									\
+	return sprintf(buf, "%u\n", dbs_tuners_ins.object);		\
+}
+show_one(sampling_rate, sampling_rate);
+show_one(sampling_down_factor, sampling_down_factor);
+show_one(up_threshold, up_threshold);
+show_one(down_threshold, down_threshold);
+
+static ssize_t store_sampling_down_factor(struct cpufreq_policy *unused, 
+		const char *buf, size_t count)
+{
+	unsigned int input;
+	int ret;
+	ret = sscanf (buf, "%u", &input);
+	down(&dbs_sem);
+	if (ret != 1 )
+		goto out;
+
+	dbs_tuners_ins.sampling_down_factor = input;
+out:
+	up(&dbs_sem);
+	return count;
+}
+
+static ssize_t store_sampling_rate(struct cpufreq_policy *unused, 
+		const char *buf, size_t count)
+{
+	unsigned int input;
+	int ret;
+	ret = sscanf (buf, "%u", &input);
+	down(&dbs_sem);
+	if (ret != 1 || input > MAX_SAMPLING_RATE || input < MIN_SAMPLING_RATE)
+		goto out;
+
+	dbs_tuners_ins.sampling_rate = input;
+out:
+	up(&dbs_sem);
+	return count;
+}
+
+static ssize_t store_up_threshold(struct cpufreq_policy *unused, 
+		const char *buf, size_t count)
+{
+	unsigned int input;
+	int ret;
+	ret = sscanf (buf, "%u", &input);
+	down(&dbs_sem);
+	if (ret != 1 || input > MAX_FREQUENCY_UP_THRESHOLD || 
+			input < MIN_FREQUENCY_UP_THRESHOLD ||
+			input <= dbs_tuners_ins.down_threshold)
+		goto out;
+
+	dbs_tuners_ins.up_threshold = input;
+out:
+	up(&dbs_sem);
+	return count;
+}
+
+static ssize_t store_down_threshold(struct cpufreq_policy *unused, 
+		const char *buf, size_t count)
+{
+	unsigned int input;
+	int ret;
+	ret = sscanf (buf, "%u", &input);
+	down(&dbs_sem);
+	if (ret != 1 || input > MAX_FREQUENCY_DOWN_THRESHOLD || 
+			input < MIN_FREQUENCY_DOWN_THRESHOLD ||
+			input >= dbs_tuners_ins.up_threshold)
+		goto out;
+
+	dbs_tuners_ins.down_threshold = input;
+out:
+	up(&dbs_sem);
+	return count;
+}
+
+#define define_one_rw(_name) 					\
+static struct freq_attr _name = { 				\
+	.attr = { .name = __stringify(_name), .mode = 0644 }, 	\
+	.show = show_##_name, 					\
+	.store = store_##_name, 				\
+}
+
+define_one_rw(sampling_rate);
+define_one_rw(sampling_down_factor);
+define_one_rw(up_threshold);
+define_one_rw(down_threshold);
+
+static struct attribute * dbs_attributes[] = {
+	&current_freq.attr,
+	&sampling_rate_max.attr,
+	&sampling_rate_min.attr,
+	&sampling_rate.attr,
+	&sampling_down_factor.attr,
+	&up_threshold.attr,
+	&down_threshold.attr,
+	NULL
+};
+
+static struct attribute_group dbs_attr_group = {
+	.attrs = dbs_attributes,
+	.name = "ondemand",
+};
+
+/************************** sysfs end ************************/
+
+static void dbs_check_cpu(int cpu)
+{
+	unsigned int idle_ticks, up_idle_ticks, down_idle_ticks;
+	unsigned int freq_down_step;
+	unsigned int freq_down_sampling_rate;
+	static int down_skip[NR_CPUS];
+	struct cpu_dbs_info_s *this_dbs_info;
+
+	this_dbs_info = &per_cpu(cpu_dbs_info, cpu);
+	if (!this_dbs_info->enable)
+		return;
+
+	/* 
+	 * The default safe range is 20% to 80% 
+	 * Every sampling_rate, we check
+	 * 	- If current idle time is less than 20%, then we try to 
+	 * 	  increase frequency
+	 * Every sampling_rate*sampling_down_factor, we check
+	 * 	- If current idle time is more than 80%, then we try to
+	 * 	  decrease frequency
+	 *
+	 * Any frequency increase takes it to the maximum frequency. 
+	 * Frequency reduction happens at minimum steps of 
+	 * 5% of max_frequency 
+	 */
+	/* Check for frequency increase */
+	idle_ticks = kstat_cpu(cpu).cpustat.idle - 
+		this_dbs_info->prev_cpu_idle_up;
+	this_dbs_info->prev_cpu_idle_up = kstat_cpu(cpu).cpustat.idle;
+
+	up_idle_ticks = (100 - dbs_tuners_ins.up_threshold) *
+			sampling_rate_in_HZ(dbs_tuners_ins.sampling_rate) / 100;
+
+	if (idle_ticks < up_idle_ticks) {
+		__cpufreq_driver_target(this_dbs_info->cur_policy,
+			this_dbs_info->cur_policy->max, 
+			CPUFREQ_RELATION_H);
+		down_skip[cpu] = 0;
+		this_dbs_info->prev_cpu_idle_down = kstat_cpu(cpu).cpustat.idle;
+		return;
+	}
+
+	/* Check for frequency decrease */
+	down_skip[cpu]++;
+	if (down_skip[cpu] < dbs_tuners_ins.sampling_down_factor)
+		return;
+
+	idle_ticks = kstat_cpu(cpu).cpustat.idle - 
+		this_dbs_info->prev_cpu_idle_down;
+	down_skip[cpu] = 0;
+	this_dbs_info->prev_cpu_idle_down = kstat_cpu(cpu).cpustat.idle;
+
+	freq_down_sampling_rate = dbs_tuners_ins.sampling_rate *
+		dbs_tuners_ins.sampling_down_factor;
+	down_idle_ticks = (100 - dbs_tuners_ins.down_threshold) *
+			sampling_rate_in_HZ(freq_down_sampling_rate) / 100;
+
+	if (idle_ticks > down_idle_ticks ) {
+		freq_down_step = (5 * this_dbs_info->cur_policy->max) / 100;
+		__cpufreq_driver_target(this_dbs_info->cur_policy,
+			this_dbs_info->cur_policy->cur - freq_down_step, 
+			CPUFREQ_RELATION_H);
+		return;
+	}
+}
+
+static void do_dbs_timer(void *data)
+{ 
+	int i;
+	down(&dbs_sem);
+	for (i = 0; i < NR_CPUS; i++)
+		if (cpu_online(i))
+			dbs_check_cpu(i);
+	schedule_delayed_work(&dbs_work, 
+			sampling_rate_in_HZ(dbs_tuners_ins.sampling_rate));
+	up(&dbs_sem);
+} 
+
+static inline void dbs_timer_init(void)
+{
+	INIT_WORK(&dbs_work, do_dbs_timer, NULL);
+	schedule_work(&dbs_work);
+	return;
+}
+
+static inline void dbs_timer_exit(void)
+{
+	cancel_delayed_work(&dbs_work);
+	return;
+}
+
+static int cpufreq_governor_dbs(struct cpufreq_policy *policy,
+				   unsigned int event)
+{
+	unsigned int cpu = policy->cpu;
+	struct cpu_dbs_info_s *this_dbs_info;
+
+	this_dbs_info = &per_cpu(cpu_dbs_info, cpu);
+
+	switch (event) {
+	case CPUFREQ_GOV_START:
+		if ((!cpu_online(cpu)) || 
+		    (!policy->cur))
+			return -EINVAL;
+
+		if (policy->cpuinfo.transition_latency >
+				(TRANSITION_LATENCY_LIMIT * 1000))
+			return -EINVAL;
+		if (this_dbs_info->enable) /* Already enabled */
+			break;
+		 
+		down(&dbs_sem);
+		this_dbs_info->cur_policy = policy;
+		
+		this_dbs_info->prev_cpu_idle_up = 
+				kstat_cpu(cpu).cpustat.idle;
+		this_dbs_info->prev_cpu_idle_down = 
+				kstat_cpu(cpu).cpustat.idle;
+		this_dbs_info->enable = 1;
+		sysfs_create_group(&policy->kobj, &dbs_attr_group);
+		dbs_enable++;
+		/*
+		 * Start the timerschedule work, when this governor
+		 * is used for first time
+		 */
+		if (dbs_enable == 1) {
+			/* policy latency is in nS. Convert it to uS first */
+			def_sampling_rate = (policy->cpuinfo.transition_latency / 1000) *
+					DEF_SAMPLING_RATE_LATENCY_MULTIPLIER;
+			dbs_tuners_ins.sampling_rate = def_sampling_rate;
+
+			dbs_timer_init();
+		}
+		
+		up(&dbs_sem);
+		break;
+
+	case CPUFREQ_GOV_STOP:
+		down(&dbs_sem);
+		this_dbs_info->enable = 0;
+		sysfs_remove_group(&policy->kobj, &dbs_attr_group);
+		dbs_enable--;
+		/*
+		 * Stop the timerschedule work, when this governor
+		 * is used for first time
+		 */
+		if (dbs_enable == 0) 
+			dbs_timer_exit();
+		
+		up(&dbs_sem);
+
+		break;
+
+	case CPUFREQ_GOV_LIMITS:
+		down(&dbs_sem);
+		if (policy->max < this_dbs_info->cur_policy->cur)
+			__cpufreq_driver_target(
+					this_dbs_info->cur_policy,
+				       	policy->max, CPUFREQ_RELATION_H);
+		else if (policy->min > this_dbs_info->cur_policy->cur)
+			__cpufreq_driver_target(
+					this_dbs_info->cur_policy,
+				       	policy->min, CPUFREQ_RELATION_L);
+		up(&dbs_sem);
+		break;
+	}
+	return 0;
+}
+
+struct cpufreq_governor cpufreq_gov_dbs = {
+	.name		= "ondemand",
+	.governor	= cpufreq_governor_dbs,
+	.owner		= THIS_MODULE,
+};
+EXPORT_SYMBOL(cpufreq_gov_dbs);
+
+static int __init cpufreq_gov_dbs_init(void)
+{
+	return cpufreq_register_governor(&cpufreq_gov_dbs);
+}
+
+static void __exit cpufreq_gov_dbs_exit(void)
+{
+	/* Make sure that the scheduled work is indeed not running */
+	flush_scheduled_work();
+
+	cpufreq_unregister_governor(&cpufreq_gov_dbs);
+}
+
+
+MODULE_AUTHOR ("Venkatesh Pallipadi <venkatesh.pallipadi@intel.com>");
+MODULE_DESCRIPTION ("'cpufreq_ondemand' - A dynamic cpufreq governor for "
+		"Low Latency Frequency Transition capable processors");
+MODULE_LICENSE ("GPL");
+
+module_init(cpufreq_gov_dbs_init);
+module_exit(cpufreq_gov_dbs_exit);