Merge branch 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/…

…davej/cpufreq

* 'next' of git://git.kernel.org/pub/scm/linux/kernel/git/davej/cpufreq: (35 commits)
  [CPUFREQ] Prevent p4-clockmod from auto-binding to the ondemand governor.
  [CPUFREQ] Make cpufreq-nforce2 less obnoxious
  [CPUFREQ] p4-clockmod reports wrong frequency.
  [CPUFREQ] powernow-k8: Use a common exit path.
  [CPUFREQ] Change link order of x86 cpufreq modules
  [CPUFREQ] conservative: remove 10x from def_sampling_rate
  [CPUFREQ] conservative: fixup governor to function more like ondemand logic
  [CPUFREQ] conservative: fix dbs_cpufreq_notifier so freq is not locked
  [CPUFREQ] conservative: amend author's email address
  [CPUFREQ] Use swap() in longhaul.c
  [CPUFREQ] checkpatch cleanups for acpi-cpufreq
  [CPUFREQ] powernow-k8: Only print error message once, not per core.
  [CPUFREQ] ondemand/conservative: sanitize sampling_rate restrictions
  [CPUFREQ] ondemand/conservative: deprecate sampling_rate{min,max}
  [CPUFREQ] powernow-k8: Always compile powernow-k8 driver with ACPI support
  [CPUFREQ] Introduce /sys/devices/system/cpu/cpu*/cpufreq/cpuinfo_transition_latency
  [CPUFREQ] checkpatch cleanups for powernow-k8
  [CPUFREQ] checkpatch cleanups for ondemand governor.
  [CPUFREQ] checkpatch cleanups for powernow-k7
  [CPUFREQ] checkpatch cleanups for speedstep related drivers.
  ...

Documentation/cpu-freq/governors.txt

@@ -117,10 +117,28 @@ accessible parameters:
 sampling_rate: measured in uS (10^-6 seconds), this is how often you
 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.
-
-show_sampling_rate_(min|max): the minimum and maximum sampling rates
-available that you may set 'sampling_rate' to.
+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
+To set the sampling rate 750 times as high as the transition latency
+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).
 
 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

Documentation/cpu-freq/user-guide.txt

@@ -152,6 +152,18 @@ cpuinfo_min_freq :		this file shows the minimum operating
 				frequency the processor can run at(in kHz) 
 cpuinfo_max_freq :		this file shows the maximum operating
 				frequency the processor can run at(in kHz) 
+cpuinfo_transition_latency	The time it takes on this CPU to
+				switch between two frequencies in nano
+				seconds. If unknown or known to be
+				that high that the driver does not
+				work with the ondemand governor, -1
+				(CPUFREQ_ETERNAL) will be returned.
+				Using this information can be useful
+				to choose an appropriate polling
+				frequency for a kernel governor or
+				userspace daemon. Make sure to not
+				switch the frequency too often
+				resulting in performance loss.
 scaling_driver :		this file shows what cpufreq driver is
 				used to set the frequency on this CPU
 

arch/x86/include/asm/timer.h

@@ -11,8 +11,8 @@ unsigned long native_calibrate_tsc(void);
 
 #ifdef CONFIG_X86_32
 extern int timer_ack;
+#endif
 extern int recalibrate_cpu_khz(void);
-#endif /* CONFIG_X86_32 */
 
 extern int no_timer_check;
 

arch/x86/kernel/cpu/cpufreq/Kconfig

@@ -87,30 +87,15 @@ config X86_POWERNOW_K7_ACPI
 config X86_POWERNOW_K8
 	tristate "AMD Opteron/Athlon64 PowerNow!"
 	select CPU_FREQ_TABLE
+	depends on ACPI && ACPI_PROCESSOR
 	help
-	  This adds the CPUFreq driver for mobile AMD Opteron/Athlon64 processors.
+	  This adds the CPUFreq driver for K8/K10 Opteron/Athlon64 processors.
 
 	  To compile this driver as a module, choose M here: the
 	  module will be called powernow-k8.
 
 	  For details, take a look at <file:Documentation/cpu-freq/>.
 
-	  If in doubt, say N.
-
-config X86_POWERNOW_K8_ACPI
-	bool
-	prompt "ACPI Support" if X86_32
-	depends on ACPI && X86_POWERNOW_K8 && ACPI_PROCESSOR
-	depends on !(X86_POWERNOW_K8 = y && ACPI_PROCESSOR = m)
-	default y
-	help
-	  This provides access to the K8s Processor Performance States via ACPI.
-	  This driver is probably required for CPUFreq to work with multi-socket and
-	  SMP systems.  It is not required on at least some single-socket yet
-	  multi-core systems, even if SMP is enabled.
-
-	  It is safe to say Y here.
-
 config X86_GX_SUSPMOD
 	tristate "Cyrix MediaGX/NatSemi Geode Suspend Modulation"
 	depends on X86_32 && PCI

arch/x86/kernel/cpu/cpufreq/Makefile

@@ -1,6 +1,11 @@
+# Link order matters. K8 is preferred to ACPI because of firmware bugs in early
+# K8 systems. ACPI is preferred to all other hardware-specific drivers.
+# speedstep-* is preferred over p4-clockmod.
+
+obj-$(CONFIG_X86_POWERNOW_K8)		+= powernow-k8.o
+obj-$(CONFIG_X86_ACPI_CPUFREQ)		+= acpi-cpufreq.o
 obj-$(CONFIG_X86_POWERNOW_K6)		+= powernow-k6.o
 obj-$(CONFIG_X86_POWERNOW_K7)		+= powernow-k7.o
-obj-$(CONFIG_X86_POWERNOW_K8)		+= powernow-k8.o
 obj-$(CONFIG_X86_LONGHAUL)		+= longhaul.o
 obj-$(CONFIG_X86_E_POWERSAVER)		+= e_powersaver.o
 obj-$(CONFIG_ELAN_CPUFREQ)		+= elanfreq.o
@@ -10,7 +15,6 @@ obj-$(CONFIG_X86_GX_SUSPMOD)		+= gx-suspmod.o
 obj-$(CONFIG_X86_SPEEDSTEP_ICH)		+= speedstep-ich.o
 obj-$(CONFIG_X86_SPEEDSTEP_LIB)		+= speedstep-lib.o
 obj-$(CONFIG_X86_SPEEDSTEP_SMI)		+= speedstep-smi.o
-obj-$(CONFIG_X86_ACPI_CPUFREQ)		+= acpi-cpufreq.o
 obj-$(CONFIG_X86_SPEEDSTEP_CENTRINO)	+= speedstep-centrino.o
 obj-$(CONFIG_X86_P4_CLOCKMOD)		+= p4-clockmod.o
 obj-$(CONFIG_X86_CPUFREQ_NFORCE2)	+= cpufreq-nforce2.o

arch/x86/kernel/cpu/cpufreq/acpi-cpufreq.c

@@ -1,5 +1,5 @@
 /*
- * acpi-cpufreq.c - ACPI Processor P-States Driver ($Revision: 1.4 $)
+ * acpi-cpufreq.c - ACPI Processor P-States Driver
  *
  *  Copyright (C) 2001, 2002 Andy Grover <andrew.grover@intel.com>
  *  Copyright (C) 2001, 2002 Paul Diefenbaugh <paul.s.diefenbaugh@intel.com>
@@ -36,16 +36,18 @@
 #include <linux/ftrace.h>
 
 #include <linux/acpi.h>
+#include <linux/io.h>
+#include <linux/delay.h>
+#include <linux/uaccess.h>
+
 #include <acpi/processor.h>
 
-#include <asm/io.h>
 #include <asm/msr.h>
 #include <asm/processor.h>
 #include <asm/cpufeature.h>
-#include <asm/delay.h>
-#include <asm/uaccess.h>
 
-#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "acpi-cpufreq", msg)
+#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, \
+		"acpi-cpufreq", msg)
 
 MODULE_AUTHOR("Paul Diefenbaugh, Dominik Brodowski");
 MODULE_DESCRIPTION("ACPI Processor P-States Driver");
@@ -95,7 +97,7 @@ static unsigned extract_io(u32 value, struct acpi_cpufreq_data *data)
 
 	perf = data->acpi_data;
 
-	for (i=0; i<perf->state_count; i++) {
+	for (i = 0; i < perf->state_count; i++) {
 		if (value == perf->states[i].status)
 			return data->freq_table[i].frequency;
 	}
@@ -110,7 +112,7 @@ static unsigned extract_msr(u32 msr, struct acpi_cpufreq_data *data)
 	msr &= INTEL_MSR_RANGE;
 	perf = data->acpi_data;
 
-	for (i=0; data->freq_table[i].frequency != CPUFREQ_TABLE_END; i++) {
+	for (i = 0; data->freq_table[i].frequency != CPUFREQ_TABLE_END; i++) {
 		if (msr == perf->states[data->freq_table[i].index].status)
 			return data->freq_table[i].frequency;
 	}
@@ -138,15 +140,13 @@ struct io_addr {
 	u8 bit_width;
 };
 
-typedef union {
-	struct msr_addr msr;
-	struct io_addr io;
-} drv_addr_union;
-
 struct drv_cmd {
 	unsigned int type;
 	const struct cpumask *mask;
-	drv_addr_union addr;
+	union {
+		struct msr_addr msr;
+		struct io_addr io;
+	} addr;
 	u32 val;
 };
 
@@ -369,7 +369,7 @@ static unsigned int check_freqs(const struct cpumask *mask, unsigned int freq,
 	unsigned int cur_freq;
 	unsigned int i;
 
-	for (i=0; i<100; i++) {
+	for (i = 0; i < 100; i++) {
 		cur_freq = extract_freq(get_cur_val(mask), data);
 		if (cur_freq == freq)
 			return 1;
@@ -494,7 +494,7 @@ acpi_cpufreq_guess_freq(struct acpi_cpufreq_data *data, unsigned int cpu)
 		unsigned long freq;
 		unsigned long freqn = perf->states[0].core_frequency * 1000;
 
-		for (i=0; i<(perf->state_count-1); i++) {
+		for (i = 0; i < (perf->state_count-1); i++) {
 			freq = freqn;
 			freqn = perf->states[i+1].core_frequency * 1000;
 			if ((2 * cpu_khz) > (freqn + freq)) {
@@ -673,7 +673,7 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy)
 
 	/* detect transition latency */
 	policy->cpuinfo.transition_latency = 0;
-	for (i=0; i<perf->state_count; i++) {
+	for (i = 0; i < perf->state_count; i++) {
 		if ((perf->states[i].transition_latency * 1000) >
 		    policy->cpuinfo.transition_latency)
 			policy->cpuinfo.transition_latency =
@@ -682,8 +682,8 @@ static int acpi_cpufreq_cpu_init(struct cpufreq_policy *policy)
 
 	data->max_freq = perf->states[0].core_frequency * 1000;
 	/* table init */
-	for (i=0; i<perf->state_count; i++) {
-		if (i>0 && perf->states[i].core_frequency >=
+	for (i = 0; i < perf->state_count; i++) {
+		if (i > 0 && perf->states[i].core_frequency >=
 		    data->freq_table[valid_states-1].frequency / 1000)
 			continue;
 

arch/x86/kernel/cpu/cpufreq/cpufreq-nforce2.c

@@ -32,7 +32,7 @@
  * nforce2_chipset:
  * FSB is changed using the chipset
  */
-static struct pci_dev *nforce2_chipset_dev;
+static struct pci_dev *nforce2_dev;
 
 /* fid:
  * multiplier * 10
@@ -56,7 +56,9 @@ MODULE_PARM_DESC(fid, "CPU multiplier to use (11.5 = 115)");
 MODULE_PARM_DESC(min_fsb,
 		"Minimum FSB to use, if not defined: current FSB - 50");
 
-#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, "cpufreq-nforce2", msg)
+#define PFX "cpufreq-nforce2: "
+#define dprintk(msg...) cpufreq_debug_printk(CPUFREQ_DEBUG_DRIVER, \
+		"cpufreq-nforce2", msg)
 
 /**
  * nforce2_calc_fsb - calculate FSB
@@ -118,11 +120,11 @@ static void nforce2_write_pll(int pll)
 	int temp;
 
 	/* Set the pll addr. to 0x00 */
-	pci_write_config_dword(nforce2_chipset_dev, NFORCE2_PLLADR, 0);
+	pci_write_config_dword(nforce2_dev, NFORCE2_PLLADR, 0);
 
 	/* Now write the value in all 64 registers */
 	for (temp = 0; temp <= 0x3f; temp++)
-		pci_write_config_dword(nforce2_chipset_dev, NFORCE2_PLLREG, pll);
+		pci_write_config_dword(nforce2_dev, NFORCE2_PLLREG, pll);
 
 	return;
 }
@@ -139,22 +141,22 @@ static unsigned int nforce2_fsb_read(int bootfsb)
 	u32 fsb, temp = 0;
 
 	/* Get chipset boot FSB from subdevice 5 (FSB at boot-time) */
-	nforce2_sub5 = pci_get_subsys(PCI_VENDOR_ID_NVIDIA,
-						0x01EF, PCI_ANY_ID, PCI_ANY_ID, NULL);
+	nforce2_sub5 = pci_get_subsys(PCI_VENDOR_ID_NVIDIA, 0x01EF,
+				PCI_ANY_ID, PCI_ANY_ID, NULL);
 	if (!nforce2_sub5)
 		return 0;
 
 	pci_read_config_dword(nforce2_sub5, NFORCE2_BOOTFSB, &fsb);
 	fsb /= 1000000;
 
 	/* Check if PLL register is already set */
-	pci_read_config_byte(nforce2_chipset_dev, NFORCE2_PLLENABLE, (u8 *)&temp);
+	pci_read_config_byte(nforce2_dev, NFORCE2_PLLENABLE, (u8 *)&temp);
 
 	if (bootfsb || !temp)
 		return fsb;
 
 	/* Use PLL register FSB value */
-	pci_read_config_dword(nforce2_chipset_dev, NFORCE2_PLLREG, &temp);
+	pci_read_config_dword(nforce2_dev, NFORCE2_PLLREG, &temp);
 	fsb = nforce2_calc_fsb(temp);
 
 	return fsb;
@@ -174,18 +176,18 @@ static int nforce2_set_fsb(unsigned int fsb)
 	int pll = 0;
 
 	if ((fsb > max_fsb) || (fsb < NFORCE2_MIN_FSB)) {
-		printk(KERN_ERR "cpufreq: FSB %d is out of range!\n", fsb);
+		printk(KERN_ERR PFX "FSB %d is out of range!\n", fsb);
 		return -EINVAL;
 	}
 
 	tfsb = nforce2_fsb_read(0);
 	if (!tfsb) {
-		printk(KERN_ERR "cpufreq: Error while reading the FSB\n");
+		printk(KERN_ERR PFX "Error while reading the FSB\n");
 		return -EINVAL;
 	}
 
 	/* First write? Then set actual value */
-	pci_read_config_byte(nforce2_chipset_dev, NFORCE2_PLLENABLE, (u8 *)&temp);
+	pci_read_config_byte(nforce2_dev, NFORCE2_PLLENABLE, (u8 *)&temp);
 	if (!temp) {
 		pll = nforce2_calc_pll(tfsb);
 
@@ -197,7 +199,7 @@ static int nforce2_set_fsb(unsigned int fsb)
 
 	/* Enable write access */
 	temp = 0x01;
-	pci_write_config_byte(nforce2_chipset_dev, NFORCE2_PLLENABLE, (u8)temp);
+	pci_write_config_byte(nforce2_dev, NFORCE2_PLLENABLE, (u8)temp);
 
 	diff = tfsb - fsb;
 
@@ -222,7 +224,7 @@ static int nforce2_set_fsb(unsigned int fsb)
 	}
 
 	temp = 0x40;
-	pci_write_config_byte(nforce2_chipset_dev, NFORCE2_PLLADR, (u8)temp);
+	pci_write_config_byte(nforce2_dev, NFORCE2_PLLADR, (u8)temp);
 
 	return 0;
 }
@@ -244,7 +246,8 @@ static unsigned int nforce2_get(unsigned int cpu)
  * nforce2_target - set a new CPUFreq policy
  * @policy: new policy
  * @target_freq: the target frequency
- * @relation: how that frequency relates to achieved frequency (CPUFREQ_RELATION_L or CPUFREQ_RELATION_H)
+ * @relation: how that frequency relates to achieved frequency
+ *  (CPUFREQ_RELATION_L or CPUFREQ_RELATION_H)
  *
  * Sets a new CPUFreq policy.
  */
@@ -276,7 +279,7 @@ static int nforce2_target(struct cpufreq_policy *policy,
 	/* local_irq_save(flags); */
 
 	if (nforce2_set_fsb(target_fsb) < 0)
-		printk(KERN_ERR "cpufreq: Changing FSB to %d failed\n",
+		printk(KERN_ERR PFX "Changing FSB to %d failed\n",
 			target_fsb);
 	else
 		dprintk("Changed FSB successfully to %d\n",
@@ -327,8 +330,8 @@ static int nforce2_cpu_init(struct cpufreq_policy *policy)
 	/* FIX: Get FID from CPU */
 	if (!fid) {
 		if (!cpu_khz) {
-			printk(KERN_WARNING
-			       "cpufreq: cpu_khz not set, can't calculate multiplier!\n");
+			printk(KERN_WARNING PFX
+			"cpu_khz not set, can't calculate multiplier!\n");
 			return -ENODEV;
 		}
 
@@ -343,7 +346,7 @@ static int nforce2_cpu_init(struct cpufreq_policy *policy)
 		}
 	}
 
-	printk(KERN_INFO "cpufreq: FSB currently at %i MHz, FID %d.%d\n", fsb,
+	printk(KERN_INFO PFX "FSB currently at %i MHz, FID %d.%d\n", fsb,
 	       fid / 10, fid % 10);
 
 	/* Set maximum FSB to FSB at boot time */
@@ -392,17 +395,18 @@ static struct cpufreq_driver nforce2_driver = {
  */
 static unsigned int nforce2_detect_chipset(void)
 {
-	nforce2_chipset_dev = pci_get_subsys(PCI_VENDOR_ID_NVIDIA,
+	nforce2_dev = pci_get_subsys(PCI_VENDOR_ID_NVIDIA,
 					PCI_DEVICE_ID_NVIDIA_NFORCE2,
 					PCI_ANY_ID, PCI_ANY_ID, NULL);
 
-	if (nforce2_chipset_dev == NULL)
+	if (nforce2_dev == NULL)
 		return -ENODEV;
 
-	printk(KERN_INFO "cpufreq: Detected nForce2 chipset revision %X\n",
-	       nforce2_chipset_dev->revision);
-	printk(KERN_INFO
-	       "cpufreq: FSB changing is maybe unstable and can lead to crashes and data loss.\n");
+	printk(KERN_INFO PFX "Detected nForce2 chipset revision %X\n",
+	       nforce2_dev->revision);
+	printk(KERN_INFO PFX
+	       "FSB changing is maybe unstable and can lead to "
+	       "crashes and data loss.\n");
 
 	return 0;
 }
@@ -420,7 +424,7 @@ static int __init nforce2_init(void)
 
 	/* detect chipset */
 	if (nforce2_detect_chipset()) {
-		printk(KERN_ERR "cpufreq: No nForce2 chipset.\n");
+		printk(KERN_INFO PFX "No nForce2 chipset.\n");
 		return -ENODEV;
 	}