Merge commit 'cpufreq/next'

6 files changed, 82 insertions, 105 deletions

diff --git a/Documentation/cpu-freq/cpu-drivers.txt b/Documentation/cpu-freq/cpu-drivers.txt
index 43c743903dd7..75a58d14d3cf 100644
--- a/Documentation/cpu-freq/cpu-drivers.txt
+++ b/Documentation/cpu-freq/cpu-drivers.txt
@@ -155,7 +155,7 @@ actual frequency must be determined using the following rules:
 - if relation==CPUFREQ_REL_H, try to select a new_freq lower than or equal
   target_freq. ("H for highest, but no higher than")
 
-Here again the frequency table helper might assist you - see section 3
+Here again the frequency table helper might assist you - see section 2
 for details.
 
 
diff --git a/Documentation/cpu-freq/governors.txt b/Documentation/cpu-freq/governors.txt
index ce73f3eb5ddb..aed082f49d09 100644
--- a/Documentation/cpu-freq/governors.txt
+++ b/Documentation/cpu-freq/governors.txt
@@ -119,10 +119,6 @@ want the kernel to look at the CPU usage and to make decisions on
 what to do about the frequency.  Typically this is set to values of
 around '10000' or more. It's default value is (cmp. with users-guide.txt):
 transition_latency * 1000
-The lowest value you can set is:
-transition_latency * 100 or it may get restricted to a value where it
-makes not sense for the kernel anymore to poll that often which depends
-on your HZ config variable (HZ=1000: max=20000us, HZ=250: max=5000).
 Be aware that transition latency is in ns and sampling_rate is in us, so you
 get the same sysfs value by default.
 Sampling rate should always get adjusted considering the transition latency
@@ -131,14 +127,20 @@ in the bash (as said, 1000 is default), do:
 echo `$(($(cat cpuinfo_transition_latency) * 750 / 1000)) \
     >ondemand/sampling_rate
 
-show_sampling_rate_(min|max): THIS INTERFACE IS DEPRECATED, DON'T USE IT.
-You can use wider ranges now and the general
-cpuinfo_transition_latency variable (cmp. with user-guide.txt) can be
-used to obtain exactly the same info:
-show_sampling_rate_min = transtition_latency * 500    / 1000
-show_sampling_rate_max = transtition_latency * 500000 / 1000
-(divided by 1000 is to illustrate that sampling rate is in us and
-transition latency is exported ns).
+show_sampling_rate_min:
+The sampling rate is limited by the HW transition latency:
+transition_latency * 100
+Or by kernel restrictions:
+If CONFIG_NO_HZ is set, the limit is 10ms fixed.
+If CONFIG_NO_HZ is not set or no_hz=off boot parameter is used, the
+limits depend on the CONFIG_HZ option:
+HZ=1000: min=20000us  (20ms)
+HZ=250:  min=80000us  (80ms)
+HZ=100:  min=200000us (200ms)
+The highest value of kernel and HW latency restrictions is shown and
+used as the minimum sampling rate.
+
+show_sampling_rate_max: THIS INTERFACE IS DEPRECATED, DON'T USE IT.
 
 up_threshold: defines what the average CPU usage between the samplings
 of 'sampling_rate' needs to be for the kernel to make a decision on
diff --git a/Documentation/cpu-freq/user-guide.txt b/Documentation/cpu-freq/user-guide.txt
index 75f41193f3e1..5d5f5fadd1c2 100644
--- a/Documentation/cpu-freq/user-guide.txt
+++ b/Documentation/cpu-freq/user-guide.txt
@@ -31,7 +31,6 @@ Contents:
 
 3. How to change the CPU cpufreq policy and/or speed
 3.1 Preferred interface: sysfs
-3.2 Deprecated interfaces
 
 
 
diff --git a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
index 35dc8fbe92bd..a39ee97c05a5 100644
--- a/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
+++ b/arch/x86/kernel/cpu/cpufreq/powernow-k8.c
@@ -1,3 +1,4 @@
+
 /*
  *   (c) 2003-2006 Advanced Micro Devices, Inc.
  *  Your use of this code is subject to the terms and conditions of the
@@ -823,13 +824,14 @@ static void powernow_k8_acpi_pst_values(struct powernow_k8_data *data,
 	if (!data->acpi_data.state_count || (cpu_family == CPU_HW_PSTATE))
 		return;
 
-	control = data->acpi_data.states[index].control; data->irt = (control
-			>> IRT_SHIFT) & IRT_MASK; data->rvo = (control >>
-				RVO_SHIFT) & RVO_MASK; data->exttype = (control
-					>> EXT_TYPE_SHIFT) & EXT_TYPE_MASK;
-	data->plllock = (control >> PLL_L_SHIFT) & PLL_L_MASK; data->vidmvs = 1
-		<< ((control >> MVS_SHIFT) & MVS_MASK); data->vstable =
-		(control >> VST_SHIFT) & VST_MASK; }
+	control = data->acpi_data.states[index].control;
+	data->irt = (control >> IRT_SHIFT) & IRT_MASK;
+	data->rvo = (control >> RVO_SHIFT) & RVO_MASK;
+	data->exttype = (control >> EXT_TYPE_SHIFT) & EXT_TYPE_MASK;
+	data->plllock = (control >> PLL_L_SHIFT) & PLL_L_MASK;
+	data->vidmvs = 1 << ((control >> MVS_SHIFT) & MVS_MASK);
+	data->vstable = (control >> VST_SHIFT) & VST_MASK;
+}
 
 static int powernow_k8_cpu_init_acpi(struct powernow_k8_data *data)
 {
@@ -1046,6 +1048,19 @@ static int get_transition_latency(struct powernow_k8_data *data)
 		if (cur_latency > max_latency)
 			max_latency = cur_latency;
 	}
+	if (max_latency == 0) {
+		/*
+		 * Fam 11h always returns 0 as transition latency.
+		 * This is intended and means "very fast". While cpufreq core
+		 * and governors currently can handle that gracefully, better
+		 * set it to 1 to avoid problems in the future.
+		 * For all others it's a BIOS bug.
+		 */
+		if (!boot_cpu_data.x86 == 0x11)
+			printk(KERN_ERR FW_WARN PFX "Invalid zero transition "
+				"latency\n");
+		max_latency = 1;
+	}
 	/* value in usecs, needs to be in nanoseconds */
 	return 1000 * max_latency;
 }
diff --git a/drivers/cpufreq/cpufreq_conservative.c b/drivers/cpufreq/cpufreq_conservative.c
index 7a74d175287b..7fc58af748b4 100644
--- a/drivers/cpufreq/cpufreq_conservative.c
+++ b/drivers/cpufreq/cpufreq_conservative.c
@@ -42,27 +42,12 @@
  * this governor will not work.
  * All times here are in uS.
  */
-static unsigned int def_sampling_rate;
 #define MIN_SAMPLING_RATE_RATIO			(2)
-/* for correct statistics, we need at least 10 ticks between each measure */
-#define MIN_STAT_SAMPLING_RATE 			\
-			(MIN_SAMPLING_RATE_RATIO * jiffies_to_usecs(10))
-#define MIN_SAMPLING_RATE			\
-			(def_sampling_rate / MIN_SAMPLING_RATE_RATIO)
-/* Above MIN_SAMPLING_RATE will vanish with its sysfs file soon
- * Define the minimal settable sampling rate to the greater of:
- *   - "HW transition latency" * 100 (same as default sampling / 10)
- *   - MIN_STAT_SAMPLING_RATE
- * To avoid that userspace shoots itself.
-*/
-static unsigned int minimum_sampling_rate(void)
-{
-	return max(def_sampling_rate / 10, MIN_STAT_SAMPLING_RATE);
-}
 
-/* This will also vanish soon with removing sampling_rate_max */
-#define MAX_SAMPLING_RATE			(500 * def_sampling_rate)
+static unsigned int min_sampling_rate;
+
 #define LATENCY_MULTIPLIER			(1000)
+#define MIN_LATENCY_MULTIPLIER			(100)
 #define DEF_SAMPLING_DOWN_FACTOR		(1)
 #define MAX_SAMPLING_DOWN_FACTOR		(10)
 #define TRANSITION_LATENCY_LIMIT		(10 * 1000 * 1000)
@@ -182,27 +167,14 @@ static struct notifier_block dbs_cpufreq_notifier_block = {
 /************************** sysfs interface ************************/
 static ssize_t show_sampling_rate_max(struct cpufreq_policy *policy, char *buf)
 {
-	static int print_once;
-
-	if (!print_once) {
-		printk(KERN_INFO "CPUFREQ: conservative sampling_rate_max "
-		       "sysfs file is deprecated - used by: %s\n",
-		       current->comm);
-		print_once = 1;
-	}
-	return sprintf(buf, "%u\n", MAX_SAMPLING_RATE);
+	printk_once(KERN_INFO "CPUFREQ: conservative sampling_rate_max "
+		    "sysfs file is deprecated - used by: %s\n", current->comm);
+	return sprintf(buf, "%u\n", -1U);
 }
 
 static ssize_t show_sampling_rate_min(struct cpufreq_policy *policy, char *buf)
 {
-	static int print_once;
-
-	if (!print_once) {
-		printk(KERN_INFO "CPUFREQ: conservative sampling_rate_max "
-		       "sysfs file is deprecated - used by: %s\n", current->comm);
-		print_once = 1;
-	}
-	return sprintf(buf, "%u\n", MIN_SAMPLING_RATE);
+	return sprintf(buf, "%u\n", min_sampling_rate);
 }
 
 #define define_one_ro(_name)		\
@@ -254,7 +226,7 @@ static ssize_t store_sampling_rate(struct cpufreq_policy *unused,
 		return -EINVAL;
 
 	mutex_lock(&dbs_mutex);
-	dbs_tuners_ins.sampling_rate = max(input, minimum_sampling_rate());
+	dbs_tuners_ins.sampling_rate = max(input, min_sampling_rate);
 	mutex_unlock(&dbs_mutex);
 
 	return count;
@@ -601,11 +573,18 @@ static int cpufreq_governor_dbs(struct cpufreq_policy *policy,
 			if (latency == 0)
 				latency = 1;
 
-			def_sampling_rate =
-				max(latency * LATENCY_MULTIPLIER,
-				    MIN_STAT_SAMPLING_RATE);
-
-			dbs_tuners_ins.sampling_rate = def_sampling_rate;
+			/*
+			 * conservative does not implement micro like ondemand
+			 * governor, thus we are bound to jiffes/HZ
+			 */
+			min_sampling_rate =
+				MIN_SAMPLING_RATE_RATIO * jiffies_to_usecs(10);
+			/* Bring kernel and HW constraints together */
+			min_sampling_rate = max(min_sampling_rate,
+					MIN_LATENCY_MULTIPLIER * latency);
+			dbs_tuners_ins.sampling_rate =
+				max(min_sampling_rate,
+				    latency * LATENCY_MULTIPLIER);
 
 			cpufreq_register_notifier(
 					&dbs_cpufreq_notifier_block,
diff --git a/drivers/cpufreq/cpufreq_ondemand.c b/drivers/cpufreq/cpufreq_ondemand.c
index e741c339df76..1911d1729353 100644
--- a/drivers/cpufreq/cpufreq_ondemand.c
+++ b/drivers/cpufreq/cpufreq_ondemand.c
@@ -32,6 +32,7 @@
 #define DEF_FREQUENCY_UP_THRESHOLD		(80)
 #define MICRO_FREQUENCY_DOWN_DIFFERENTIAL	(3)
 #define MICRO_FREQUENCY_UP_THRESHOLD		(95)
+#define MICRO_FREQUENCY_MIN_SAMPLE_RATE		(10000)
 #define MIN_FREQUENCY_UP_THRESHOLD		(11)
 #define MAX_FREQUENCY_UP_THRESHOLD		(100)
 
@@ -45,27 +46,12 @@
  * this governor will not work.
  * All times here are in uS.
  */
-static unsigned int def_sampling_rate;
 #define MIN_SAMPLING_RATE_RATIO			(2)
-/* for correct statistics, we need at least 10 ticks between each measure */
-#define MIN_STAT_SAMPLING_RATE 			\
-			(MIN_SAMPLING_RATE_RATIO * jiffies_to_usecs(10))
-#define MIN_SAMPLING_RATE			\
-			(def_sampling_rate / MIN_SAMPLING_RATE_RATIO)
-/* Above MIN_SAMPLING_RATE will vanish with its sysfs file soon
- * Define the minimal settable sampling rate to the greater of:
- *   - "HW transition latency" * 100 (same as default sampling / 10)
- *   - MIN_STAT_SAMPLING_RATE
- * To avoid that userspace shoots itself.
-*/
-static unsigned int minimum_sampling_rate(void)
-{
-	return max(def_sampling_rate / 10, MIN_STAT_SAMPLING_RATE);
-}
 
-/* This will also vanish soon with removing sampling_rate_max */
-#define MAX_SAMPLING_RATE			(500 * def_sampling_rate)
+static unsigned int min_sampling_rate;
+
 #define LATENCY_MULTIPLIER			(1000)
+#define MIN_LATENCY_MULTIPLIER			(100)
 #define TRANSITION_LATENCY_LIMIT		(10 * 1000 * 1000)
 
 static void do_dbs_timer(struct work_struct *work);
@@ -219,28 +205,14 @@ static void ondemand_powersave_bias_init(void)
 /************************** sysfs interface ************************/
 static ssize_t show_sampling_rate_max(struct cpufreq_policy *policy, char *buf)
 {
-	static int print_once;
-
-	if (!print_once) {
-		printk(KERN_INFO "CPUFREQ: ondemand sampling_rate_max "
-		       "sysfs file is deprecated - used by: %s\n",
-		       current->comm);
-		print_once = 1;
-	}
-	return sprintf(buf, "%u\n", MAX_SAMPLING_RATE);
+	printk_once(KERN_INFO "CPUFREQ: ondemand sampling_rate_max "
+	       "sysfs file is deprecated - used by: %s\n", current->comm);
+	return sprintf(buf, "%u\n", -1U);
 }
 
 static ssize_t show_sampling_rate_min(struct cpufreq_policy *policy, char *buf)
 {
-	static int print_once;
-
-	if (!print_once) {
-		printk(KERN_INFO "CPUFREQ: ondemand sampling_rate_min "
-		       "sysfs file is deprecated - used by: %s\n",
-		       current->comm);
-		print_once = 1;
-	}
-	return sprintf(buf, "%u\n", MIN_SAMPLING_RATE);
+	return sprintf(buf, "%u\n", min_sampling_rate);
 }
 
 #define define_one_ro(_name)		\
@@ -274,7 +246,7 @@ static ssize_t store_sampling_rate(struct cpufreq_policy *unused,
 		mutex_unlock(&dbs_mutex);
 		return -EINVAL;
 	}
-	dbs_tuners_ins.sampling_rate = max(input, minimum_sampling_rate());
+	dbs_tuners_ins.sampling_rate = max(input, min_sampling_rate);
 	mutex_unlock(&dbs_mutex);
 
 	return count;
@@ -619,12 +591,12 @@ static int cpufreq_governor_dbs(struct cpufreq_policy *policy,
 			latency = policy->cpuinfo.transition_latency / 1000;
 			if (latency == 0)
 				latency = 1;
-
-			def_sampling_rate =
-				max(latency * LATENCY_MULTIPLIER,
-				    MIN_STAT_SAMPLING_RATE);
-
-			dbs_tuners_ins.sampling_rate = def_sampling_rate;
+			/* Bring kernel and HW constraints together */
+			min_sampling_rate = max(min_sampling_rate,
+					MIN_LATENCY_MULTIPLIER * latency);
+			dbs_tuners_ins.sampling_rate =
+				max(min_sampling_rate,
+				    latency * LATENCY_MULTIPLIER);
 		}
 		dbs_timer_init(this_dbs_info);
 
@@ -678,6 +650,16 @@ static int __init cpufreq_gov_dbs_init(void)
 		dbs_tuners_ins.up_threshold = MICRO_FREQUENCY_UP_THRESHOLD;
 		dbs_tuners_ins.down_differential =
 					MICRO_FREQUENCY_DOWN_DIFFERENTIAL;
+		/*
+		 * In no_hz/micro accounting case we set the minimum frequency
+		 * not depending on HZ, but fixed (very low). The deferred
+		 * timer might skip some samples if idle/sleeping as needed.
+		*/
+		min_sampling_rate = MICRO_FREQUENCY_MIN_SAMPLE_RATE;
+	} else {
+		/* For correct statistics, we need 10 ticks for each measure */
+		min_sampling_rate =
+			MIN_SAMPLING_RATE_RATIO * jiffies_to_usecs(10);
 	}
 
 	kondemand_wq = create_workqueue("kondemand");